Y DE PCR ai de ts dae N ae sur, À $ 1 a О 5 } , р ; | js , d у Eh : y \ vr й Зе: у CLASS у e В . Л ER à UN? E + р REC RUE | A 0 ; À : wy RR 7 Je on nor d s À on ий Ра o, | A ES 3 : Y NAO Fe À | CHAR * mn E 443 Г gid ta A Seat Ac ho НИ р PAS re el en f s 54 PU APR ры DRE FLE ОН : q Е + г pa nade i; i té а oe ag ие eat TRE LR ASE TETE ENS AALACOLOGIA ja International Journal of Malacology Revista Internacional de Malacologia . Journal International de Malacologie Международный Журнал Малакологии Internationale Malakologische Zeitschrift 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: Editorial Assistant: JOHN B. BURCH MARY DUNN University of Michigan, Ann Arbor ANNE GISMANN Maadi, А. В. Egypt 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: J FRANCES ALLEN, Emerita ROBERT ROBERTSON Environmental Protection Agency Washington, D.C. CLYDE F. E. ROPER, President Smithsonian Institution CHRISTOPHER J. BAYNE, President-Elect Washington, D.C. Oregon State University, Corvallis W. D. RUSSELL-HUNTER ELMER G. BERRY, Emeritus Syracuse University, New York Germantown, Maryland NORMAN F. SOHL КЕММЕТН J. BOSS United States Geological Survey Museum of Comparative Zoólogy Washington, D.C. Cambridge, Massachusetts RUTH D. TURNER, Alternate JOHN B. BURCH Museum of Comparative Zoólogy Cambridge, Massachusetts MELBOURNE R. CARRIKER University of Delaware, Lewes SHI-KUEI WU, Vice-President University of Colorado Museum, Boulder GEORGE M. DAVIS, Executive Secretary-Treasurer Institute meetings are held the first Friday in December each year at a convenient place. For information, address the President. Copyright, © Institute of Malacology, 1980 h | 1980 EDITORIAL BOARD J. A. ALLEN Marine Biological Station, Millport, United Kingdom E. E. BINDER Muséum d'Histoire Naturelle Genève, Switzerland A. H. CLARKE, Jr. National Museum of Natural History Washington, D.C., U.S.A. E. S. DEMIAN Ain Shams University Cairo, A. R. Egypt 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 V. FRETTER University of Reading United Kingdom E. GITTENBERGER Rijksmuseum van Natuurlijke Historie Leiden, Netherlands A. N. GOLIKOV Zoological Institute Leningrad, U.S.S.R. A. V. GROSSU Universitatea Bucuresti Romania T. HABE National Science Museum Tokyo, Japan A. D. HARRISON University of Waterloo Ontario, Canada K. HATAI Tohoku University Sendai, Japan B. HUBENDICK Naturhistoriska Museet Góteborg, Sweden A. M. KEEN Stanford University California, U.S.A. R. N. KILBURN Natal Museum Pietermaritzburg, South Africa M. A. KLAPPENBACH Museo Nacional de Historia Natural Montevideo, Uruguay J. KNUDSEN Zoologisk Institut £ Museum Kóbenhavn, Denmark A. J. KOHN University of Washington Seattle, U.S.A. Y. KONDO Bernice P. Bishop Museum Honolulu, Hawaii, U.S.A. C. M. LALLI McGill University Montreal, Canada J. LEVER Amsterdam, Netherlands A. LUCAS Faculté des Sciences Brest, France N. MACAROVICI Universitatea “Al. I. Cuza” lasi, Romania С. МЕЕВ-ВВООК Tropenmedizinisches Institut Túbingen, 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 J. OKLAND University of Oslo Norway Т. OKUTANI National Science Museum Tokyo, Japan W. L. PARAENSE Universidade de Brasilia Brazil J. J. PARODIZ Carnegie Museum Pittsburgh, U.S.A. C. M. PATTERSON University of Michigan Ann Arbor, U.S.A. W. Е. PONDER Australian Museum Sydney А. М. В. POWELL Auckland Institute & Museum New Zealand R. D. PURCHON Chelsea College of Science € Technology London, United Kingdom O. RAVERA Euratom Ispra, Italy N. W. RUNHAM University College of North Wales Bangor, United Kingdom S. G. SEGERSTRALE Institute of Marine Research Helsinki, Finland G. A. SOLEM Field Museum of Natural History Chicago, U.S.A. Е. STARMÜHLNER 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. TOEFOLERO Societa Malacologica ltaliana Milano W. S. S. VAN BENTHEM JUTTING Domburg, Netherlands J. A. VAN EEDEN Potchefstroom University South Africa J.-J. МАМ MOL Université Libre de Bruxelles Belgium N.H. VERDONK Rijksuniversiteit Utrecht, Netherlands В. В. 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, 1980, 20(1): 1-33 PROSERPINOID LAND SNAILS AND THEIR RELATIONSHIPS WITHIN THE ARCHAEOGASTROPODA Fred G. Thompson Florida State Museum, University of Florida, Gainesville, Florida 32611, U.S.A. ABSTRACT The classification of the proserpinoid land snails is reviewed. Two families are recognized, Proserpinidae and Ceresidae new family. The Proserpinidae are confined to the West Indies and include a single genus with eight extant species. The Ceresidae are currently known from Mexico and South America and contain five genera and thirteen species. The Ceresidae are a family of terrestrial diotocardian Archaeogastropoda that have two functional auricles, are hol- opod, lack an operculum and have a lamella barrier within the aperture. The Proserpinidae are more specialized, with a single auricle and an aulacopod foot, but also lack an operculum and have an internal barrier. On the basis of morphological data two families of prosobranch snails can be derived from the Ceresidae: the Proserpinidae and the Helicinidae. The Helicinidae have a single auricle, a holopod foot, an operculum, and lack a lamellar barrier. The operculum is a derived structure that more effectively closes the aperture to intruding objects than does the lamellar barrier in the ancestral groups. The relationship between the Ceresidae and the Helicinidae is clear on the basis of known anatomical data. The relationship between the Proserpinidae and the Helicinidae is less clear. The Proserpinidae and the helicinid subfamily Vianinae have similar radulae which are divergent from the basic type that occurs in other Helicinidae and the Ceresidae. Other anatomical characteristics of the Vianinae are typically helicinid. Probably the similar radulae of the Proserpinidae and the Vianinae are due to convergent evolution for similar trophic activities and do not reflect a close relationship between the two groups. The families Ceresidae, Proserpinidae and Helicinidae comprise the new superfamily Helici- nacea. The Hydrocenidae, which frequently are placed in close association with the Helicinidae, are herein placed in a separate superfamily, Hydrocenacea. The Helicinacea 15 postulated to have evolved from a primitive marine diotocardian ancestor, but not from the Neritacea. INTRODUCTION The classification of proserpinoid land snails into family and subfamily units has sat- isfied few malacologists who have worked with them. The first species to be described were thought to be pulmonate land snails be- cause of the lack of an operculum and the presence of a lamellar barrier within the aper- ture, broadly similar to the lamellar barrier that occurs in several families of pulmonate land snails. Gray (1856) and Bland (1863) estab- lished the relationship of proserpinoids to the prosobranch Helicinidae. Baker (1922, 1926b) and Thiele (1931) gave additional data on the radula, and affirmed the relation- ship of the proserpinoids to the Helicinidae. Only a single review of the proserpinoids has been published within the last century. Boss & Jacobson monographed the West Indian species (1975a) and gave an overview on the classification of mainland taxa (1975b). (1) They treat the group as a subfamily, Proser- pininae, of the Helicinidae and recognize two genera, Ceres, confined to Mexico, and Pro- serpina, including all other mainland and West Indian species. Other authors (Thiele, 1931: 89-91; Wenz, 1940: 447-448; Keen, 1960: 1287-1288) gave various schemes of generic classification but did not treat the species. The recent discovery by the author of two new species of Proserpina in Hispaniola has led to the anatomical examinations of two species and a more critical examination of the shells of other described mainland species. These studies necessitate a reevaluation of proserpinoid classification. Although the ana- tomical information 1$, unfortunately, limited to two species, enough data on the anatomy of the Helicinidae are available to give substan- tial weight to the anatomical criteria used for classification in this paper. This paper con- sists of three sections. The first presents ana- 2 THOMPSON tomical data on some species. The second section deals with the phylogeny of the pro- serpinids and related families. The third sec- tion deals with taxonomic observations on proserpinoids and a synopsis of the species. MATERIALS AND METHODS The anatomical information presented be- low is based upon two species, Ceres nelsoni Dall and Proserpina nitida Sowerby. Also in- cluded are published observations on the radula of Ceres salleana Gray, Linidiella swifti (Bland), and Proserpina (Despoenella) de- pressa (Orbigny) Specimens of Ceres nelsoni were collected at various localities in wet forests over limestone substrates in San Luis Potosi, Mexico (see Distribution under C. nelsoni). Most (UF! 24405, UF 24406) were collected by James Reddell, Texas Tech Uni- versity, while he conducted speleological studies. These specimens were dropped live into 70% isopropanol. One male (UF 24091a) was collected by the author. It was narcotized in water with menthol crystals, killed in Bouin's fixative, and preserved in 70% iso- propanol. Forty-three specimens of Proser- pina nitida were studied. These were col- lected by Glenn Goodfriend at 1.3 mi S. Clarmont, St. Ann Parish, Jamaica, on 1 No- vember 1976 at night. These were drowned in water and preserved in 70% isopropanol. Methods: Dissections were made under 70% isopropanol. The mantle collar and man- tle were removed dorsally to reveal the inter- nal arrangement of the pallial organs, the pal- lial gonoduct, and the lower intestine. Next, the reproductive system was teased free from other organs and removed. Then the dorsal body wall over the head and nape was opened to reveal the central nervous system and the anterior digestive system. Radulae were cleaned in 1% KOH. Radulae for photo- microscopic study were stained with 10% Harris's Haematoxylin. Radulae were also studied with a Cambridge Mark ll scanning electron microscope. Reproductive systems were serial sectioned at 10 um and stained with 10% Harris's Haematoxylin. The nervous systems of Ceres and Proser- рта do not differ from those of the Helicinidae. Thus they are not described in the anatomical section, but are discussed later. Terminology: The Helicinacea differ from Florida State Museum, University of Florida. other Prosobranchia in the structure of the posterior portion of the pallial gonoduct and adjacent organs, for which special terminol- ogy has been used (Thiele, 1902; Bourne, 1911; Baker, 1925, 1926a). V-organ: A peculiar topological configuration formed by the lower end of the primary ovi- duct and the adjacent portion of the pallial oviduct, which combine to form a thick- walled V-shaped structure. The pallial por- tion is called the pedicel. Accessory sperm sac: A small bulb (seminal receptacle II?) on the pedicel. Baker (1925) stated that it probably is homologous with the common reno-pericardial-gonadic duct of ancestral gastropods. Provaginal sac: A thin-walled sac on the side of the vagina just above the vaginal open- ing, and 1$ derived from the vestigial right kidney of ancestral rhipidoglossans (Thiele, 1902; Baker, 1925). Reception chamber: A voluminous chamber lying between and connecting the pedicel, the pallial oviduct, and the vagina; term seminal receptacle is used in this paper to comply with other prosobranch terminol- ogy; however, the bursa copulatrix, pro- vaginal sac, reception chamber, and ac- cessory sperm sac all receive spermato- гоа, so the function of the reception cham- ber is not unique as a receptacle (Baker, 1925, 1926a). Hypobranchial duct: Generally a thin-walled duct leading from the hypobranchial gland into the mantle cavity; in the Helicinidae the vagina opens inside the duct, thus incorpo- rating the structure into the reproductive system. Bourne (1911) and Baker (1925, 1926a), described the histology of the gland and duct. Ureter: Equivalent to renal papilla in other Prosobranchia, and not ureter as occurs in the Pulmonata. Renal papilla is used in this paper. Aulacopod-holopod foot: The side of the foot in the Proserpinidae 1$ circumscribed by a continuous groove originating from the an- terior mucous groove and demarcating a narrow band of tissue bordering the sole. This groove is similar to the aulacopod foot of some Pulmonata. The holopod foot re- fers to the absence of such a demarcating groove. These terms are used as adjectives in this paper and no homology with the Pulmonata 1$ implied. PROSERPINOID LAND SNAILS 3 ANATOMY OF CERES AND PROSERPINA Ceres nelsoni Dall The following observations are based upon twelve preserved specimens. All illustrations are based upon specimens from UF 24405 (see species account for data). External anatomy: Foot holopod, moderate- ly long, broadly spatulate; broadly triangulate in cross-section posteriorly, and bearing long dorsal keel; caudal pore absent. Sole undi- vided longitudinally. Snout projecting beyond it anteriorly; deep anterior mucous groove along anterior margin of foot extending poste- riorly on each side for distance about equal to half the width of sole; sides of foot and snout diffusely mottled with black. Tentacles long, slender and black with black bar connecting them across nape. Eyes on outer side of ten- tacles just above their base. Mantle collar white with diffuse gray along anterior edge. Collar completely surrounding body and bear- ing narrow, free lappet that is confined dorsal- ly within shell; ventrally the lappet expands posteriorly to form thin pad upon which shell URETER RENAL PAPILLA RENAL PAPILLA rests. Respiration facilitated by open mantle cavity which effectively forms a lung. Mantle unpigmented except for a few diffuse patches of gray over the pallial genitalia and lower half of hypobranchial gland. A very short columel- lar cleft extends posteriorly to point just be- hind mantle collar (Fig. 4). Right columellar retractor muscle broad, triangular, and at- tached to shell at posterior end of cleft. Left columellar retractor similar but about half as large. Roof of mantle cavity heavily supplied with large transversely alternating arteries and veins. Blood vessels most abundant pos- teriorly and in mid-region, sparse and smaller just behind mantle collar. Genitalia and rec- tum terminate just behind mantle collar. Hypobranchial gland (Figs. 4, 12) very large, extending along dorsal (right) side of mantle wall from posterior edge of renal cavity to about middle of pallial genitalia. Hypo- branchial duct (Fig. 12) lying between colu- mellar angle of mantle cavity and lower geni- talia, open along anterior mesad half which is densely lined internally with large conical papillae. Pallial complex: Pericardium a large sac “— INTESTINES LEFT AURICLE = LEFT AURICLE ——)]— EFFERENT | | YA £ ) PULMONARY VEIN | VENTRICLE 7 6 VENTRICLE 4 EFFERENT 7 SNE PULMONARY VEIN AORTA 1 PROSTATE П R. COLUMELLAR RETRACTOR PROVAGINAL NE HYPOBRANCHIAL HYPOBRANCHIAL GLAND INTESTINE RIGHT AURICLE 3 RECTUM R. COLUMELLAR | о LN yor 2 —— MANTLE COLLAR 7 FOOT LAPPET N A OVARY FIGS. 1-4. Soft anatomy of Proserpina and Ceres. Fig. 1. Proserpina nitida Sby.—ventral view of pallial organs. Fig. 2. P. nitida—dorsal view of female with shell removed. Fig. 3. Ceres nelsoni Dall—inner view of pallial organs. Fig. 4. C. nelsoni—dorsal view of female with shell removed. Scales in mm. + THOMPSON FIGS. 5-10. Scanning electron micrographs of radulae of Ceres nelsoni Dall (Figs. 5-7) and Proserpina nitida Sby. (Figs. 8-10). Fig. 5. х339. Fig. 6. x133. Fig. 7.x324.Fig. 8. x632. Fig. 9. x316. Fig. 10. x 1270. Fig. 8 is an enlargement from area of arrow in Fig. 9. Fig. 10 is an enlargement from area indicated by arrow in Fig. 8. Symbols: r—rhachidian tooth; a, b, c—A-lateral, B-lateral, C-lateral; d-e—capituliform complex (D-plate & E-plate). PROSERPINOID LAND SNAILS 5 just under ventral side of mantle at columellar angle, overlapping anterior half of kidney and part of loop of intestine, and communicating with renal cavity through small renal-peri- cardial pore lying at base of renal papilla and apex of left auricle. Heart consisting of ventri- cle and two auricles (Fig. 3); left auricle under- lying kidney and receiving efferent pulmonary vein from roof of lung. Right auricle smaller than left, posterior to ventricle and receiving two veins; one from anterior viscera and one from posterior gonadal viscera. Kidney (Fig. 3) broadly bean-shaped and lying along ventral surface of intestine and partially dorsal to left auricle. Renal papilla clearly distinguishable only near anterior end of kidney, terminating as short ovoid papilla discharging into posterior end of mantle cavity. Radula (Figs. 5-7, 11): Central field con- sisting of rhachidian tooth and five lateral teeth. Rhachidian tooth simple, trapezoidal, with broad blunt edge (Fig. 6). Basal ligament thin and membranous, lower margin not clear- ly defined as are basal ligaments of A-, B-, and C-laterals. A-lateral with weak reflection bearing 3 large acuminate cusps, center one ES A 200 um slightly larger. B-lateral with single broad, bluntly pointed central cusp and short blunt cusp on each side. C-lateral with single broad cusp. Capituliform complex consisting of two separate interlocked teeth, the D-lateral and the E-lateral, otherwise referred to as the comb-lateral (D) and the accessory plate (E). Comb lateral (Figs. 5, 11) as in Helicinidae, bearing enlarged acuminate cusps along its rasping edge. Innermost cusp largest, fol- lowed laterally by two or three smaller, nearly equal-sized cusps. Outer edge of comb- lateral interlocking into mesad edge of ac- cessory plate to form capituliform complex, a structure superficially appearing as single tooth. Outer edge of radula containing about 45 marginal teeth on each side (Fig. 7). Mar- ginals with expanded bases that attach to basal membrane. First eight marginals bi- cuspid (Fig. 11); cusps large and rounded; marginals 9-16 tricuspid; outermost margin- als with 5 cusps each. | was unable to deter- mine transition point from 4 to 5 cusps be- cause of torn condition of radula examined. Female reproductive system (Figs. 12, 13): Ovary unpigmented, large and discoidal, lying over anterior (cephalic) half of digestive O 1 МАКС. Y D E FIG. 11. Radula of Ceres nelsoni Dall. R—rhachidian tooth; A, B, C-lateral teeth: D, E-lateral field forming capituliform complex. 6 THOMPSON gland, and folding partially beneath; consist- ing of numerous small lobules discharging into collective tubules leading into ovoid egg sac at mesad side of ovary. Primary oviduct short, thick. Pallial genital system (secondary oviduct and associated organs) with two fe- male openings: vaginal opening receives sperm, oviducal opening passes fertilized eggs. Vaginal opening at posterior right cor- ner of mantle cavity just anterior to primary oviduct, and ventral to hypobranchial duct, not inside hypobranchial duct as in Helicinidae. Vaginal complex (Fig. 13) consists of short papilliform vagina leading into seminal re- ceptacle and bearing on its upper side large, elongate, weakly lobed provaginal sac and smaller ovoid bursa copulatrix. Provaginal sac wrapped over dorsal side of seminal recepta- cle and bursa copulatrix, lying on ventral side of V-organ when in natural position (Fig. 12). Posterior end of secondary oviduct begin- ning with ascending limb of V-organ, which together with accessory sperm sac form a T- shaped structure on top of descending limb (pedicel). Accessory sperm sac a single large bulb, lying on left side of apex of pedicel and about as large as ascending limb. Pedicel short and stocky, entering into thin-walled seminal receptacle continuing anteriad into uterus. Uterus strongly folded externally but without accessory ducts or diverticula. Crys- talline gland absent at base of uterus. Ovi- ducal opening and anus separate but adja- cent just behind mantle collar. Male reproductive system (Fig. 14): Testis similar to ovary in shape and position but with considerably larger lobes. Like ovary, testis partially folded around anterior edge of diges- tive gland. Vas deferens short and thick, en- tering apex of prostate at oblique angle. Apex of prostate forming short elliptical chamber continuous with lower chamber of prostate, and occupying same position as provagina in male Proserpina and Helicinidae, but histo- logically not different from prostate. Ventral surface of prostate strongly folded with trans- verse wrinkles; dorsal surface with elongate field of glandular folds and tubules. Prostate not clearly demarcated into upper and lower segments as in Proserpina and the Helici- nidae. Elongate field of glandular folds along posterior extremity corresponding to limit of prostate-| in helicinids. Lower prostate very short and otherwise not demarcated. Base of prostate forming short voluminous terminal chamber and bearing two short caeca just above male opening. One caecum overlaps the other, so only a single one evident super- ficially. Posterior edge of terminal chamber giving rise to long stout diverticulum ap- pressed against ventral side of prostate and extending to posterior end of prostate-l. Diver- ticulum regularly creased externally into transverse segments throughout its length and having long longitudinal internal folds partially dividing lumen into parallel cham- bers. Diverticulum bearing near its base a short stout appendix. Terminal chamber of prostate and intestine having a common opening. Ceres salleana Gray Gray (1856) gave observations on external morphology of the animal and described and illustrated the radula. Boss & Jacobson (1975a: 61) rejected Gray's data on the basis that they suspected Gray had confused the radula of a Helicina for С. salleana. In light of the information on the radula of C. nelsoni, Gray s data must be accepted. Gray's description and illustration of the radula of С. salleana do not properly depict the structure of the capituliform complex. Aside from this difficulty he demonstrated that the radula of С. salleana is like that of С. nelsoni. Following is a quote of his descrip- tion: “...the central rhachidian tooth is oblong, with a smooth, recurved tip, the 1st and 2nd internal teeth A- and B-laterals rather broader than the central, with a three-toothed recurved tip, the 3rd C-lateral narrow, elongate, with a slight recurved end, the 4th and 5th D- and E-laterals, the capituliform complex much larger, oblong and irregular in shape, the 4th about half the width of the 5th, with 3 or 4 denticles on the inner side of the upper edge; the 5th very large, broad, with a large sub- central reflexed lobe; the lateral marginal teeth are very numerous, subequal, compressed, transparent, with a recurved tip, which in the inner teeth of the series is bifid.” Linidiella swifti (Bland) Thiele (1931: 90) gave a brief description and figure of the radula of L. swifti, which is redrawn (Fig. 24) from his figure. Boss & Jacobson (1975a, b) in their review of the proserpinids overlooked this description. Thiele only briefly described and illustrated the central field, consisting of the rhachidian tooth, the A-, B-, and C-laterals, and the PROSERPINOID LAND SNAILS 7 HYPOBRANCHIAL GLAND | HYPOBRANCHIAL DUCT ACC. SPERM SAC. COPULATORY BURSA PROVAGINAL 7 Se SAC | a Maa RECEPTION TA OVIDUCT CHAMBER 7 EGG SAC RECTUM en J HYPOBRANCHIAL DUCT — 3mm OVIDUCT OPENING GONODUCT PROVAGINAL SAC COPULATORY EIER APPENDIX ACC. SPERM N SAC + V-ORGAN [Seen lu TESTIS an u TERMINAL CHAMBER MU / @ < OPENING N CAECUM VAGINA 13 OVIDUCT 14 FIGS. 12-14. Reproductive system of Ceres nelsoni Dall. Fig. 12. Female system. Fig. 13. Posterior seg- ment of female system with oviduct and associated structures partially separated to show interrelationships of organs. Fig. 14. Male system. 8 THOMPSON capituliform complex. These teeth are basi- cally similar to those of Ceres. The A-, B-, and C-laterals bear about 3 weak cusps along the cutting edge. The D-plate of the capituliform complex is a comb-lateral with about seven distinct acuminate cusps. The innermost cusp is the largest, and the following cusps de- crease in size progressively. Proserpina (Proserpina) nitida Sowerby The following anatomical data are based upon a large series of preserved specimens collected by Glenn Goodfriend 1.3 т! $. Clarmont, St. Ann Parish, Jamaica, 1 Novem- ber 1976. External anatomy: Foot (Fig. 2) long, slen- der, keeled above; sole undivided longitudi- nally; aulacopod, bordered on each side by double row of crenulations. Caudal pore ab- sent, sides lightly spotted with melanophores. Snout white, relatively elongate, separated from foot by deep groove. Pedal gland groove extending around anterior edge of foot and continuing posteriorly with aulacopod groove. Tentacles long, slender, dark gray with light stripe on posterior surface. Eyes at outer base of tentacles. Mantle lappet spotted and mot- tled like foot but more intensely, nearly uni- formly wide, extending posteriorly over edge of shell, and complete around body. Lappet widening over posterior foot and forming pad supporting shell. Both right and left columellar retractors ex- tend into shell for about one whorl. Both bands slender and dilated near their attach- ments to shell. Columellar cleft extending posteriorly for about 34 whorl, separating lower body from upper viscera. Pallial organs: Mantle unpigmented, except over liver and gonad where it is dark gray. Internal organs easily viewed through mantle. Intestine and lower reproductive system ter- minate abut Ya whorl behind mantle collar. Outer lung wall with very sparse network of small veins most concentrated behind collar (pallial organs shown from ventral surface with mantle removed, Fig. 2). Kidney narrowly concentric, forming a semicircular arch just behind pericardium and beneath loop of intes- tine. Renal papilla a low ridge on anterior end of kidney, short, blunt and protruding into posterior end of mantle cavity. Pericardium lying beneath left end of kidney but extending obliquely over right end. Right auricle absent. Ventricle slightly wider but shorter than left auricle, receiving anteriorly efferent pulmo- nary vein, and along left edge two smaller veins from viscera. Radula (Figs. 8-10): Radula not remark- ably different from that of P. depressa as de- scribed by Baker (1926b). Rhachidian tooth simple, parallel-sided and lacking reflection. A-, B-, and C-laterals each with single cusp (radula illustrated, Fig. 9, has anomalous duplicate A-lateral on left side). As in P. depressa, A-lateral smaller than B-lateral. Capituliform complex with very heavy scrap- ing cusp on D-plate. Marginal field consisting of 43 blade-like teeth; first 27 unicuspid with sharp anterior edge (Fig. 9); next six bicuspid; next eight (Fig. 10) tricuspid; outermost two with 4-5 long weak cusps each. Inner mar- ginals consisting of broadly triangular plates thickened at base, twisted posteriad, and re- flected at upper angle to form spatulate blades. Marginal teeth increasing in length laterally through about 28th tooth; then be- coming shorter and narrower at base. Female reproductive system (Figs. 15-16): Ovary very large and circular, occupying al- most entire dorsal surface of digestive gland, consisting of multitude of small, convoluted, compactly coiled lobes which discharge into small converging ducts that lead into a rela- tively large oval egg sac оп base of омагу. Primary oviduct short, thick, extending from egg sac to pallial oviduct where it enters through short limb of V-organ. Hypobranchial gland completely posterior to pallial oviduct, discharging into mantle cavity by a short duct. Pallial oviduct bearing at distal end two bul- bous structures on end of relatively long pedicel; short, cylindrical V-organ on right side of pedicel, and a large bulbous acces- sory sperm sac on left side. Sperm sac analo- gous, not homologous, to accessory sperm sac in Helicinidae. In latter group accessory sperm sac located at end of short duct on side of pedicel; no such structure present in P. nitida. V-organ and accessory sperm sac entering a relatively long pedicel that dis- charges into posterior end of seminal re- ceptacle (Fig. 16). Vagina a short bulbous structure protruding into posterior angle of mantle cavity along left side of pallial oviduct posterior to hypobranchial opening; bearing large weakly lobed provagina and small bursa copulatrix. Provagina wrapped around dorsal side of reception chamber and pedicel; bursa copulatrix lying on ventral side and extending posteriorly. These structures are unwrapped from vagina (Fig. 16) to show interrelation- ships. Seminal receptacle entering long slen- PROSERPINOID LAND SNAILS PROVAGINAL SAC ACCESSORY SPERM SAC SEMINAL RECEP HYPO - BRANCHIAL x ORIFICE OVARY I— PALLIAL OVIDUCT INTESTINE ÁS CRYSTALLINE GLAND HYPOBRANCHIAL GLAND PROSTATE Т SS HYPO- BRANCHIAL ORIFICE VAS DEFERENS \ DIVERTICULUM PROSTATE I < CAECUM 18 = HYPOBRANCHIAL EGG SAC (OVARY REMOVED) V-ORGAN PRIMARY | OVIDUCT ACCESSORY р NN SPERM SAC Q) PEDICEL € GLAND SEMINAL — RECEPTACLE PALLIAL | OVIDUCT —T PROVAGINAL SAC PROVAGINAL SAC | VAS DEFERENS PROSTATE I PROSTATE I 17 7-2 EIS TERMINAL CHAMBER FIGS. 15-18. Reproductive system of Proserpina nitida Sby. Fig. 15. Female reproductive system. Fig. 16. Posterior segment of female system with oviduct interrelationships of organs. Fig. 17. Male reproduc vestigial provaginal sac. and associated structures partially separated to show tive system. Fig. 18. Ventral view of prostate showing 10 THOMPSON der pallial oviduct-Il, bearing at its anterior end elongate crystalline gland. Oviduct and anus close but separate. Male reproductive system (Figs. 17-18): Testis, like ovary, a large circular mass im- bedded on dorsal side of digestive gland; consists of numerous lobes similar to, but much larger than those forming ovary. Vas deferens short and stout, entering end of provagina, which is embedded in posterior end of prostate-l (Fig. 18). Prostate-l strongly folded dorsally and continuing into longer, more slender prostate-Il, which bears a long, clearly demarcated terminal chamber. Diver- ticulum originating at junction of prostate-Il and terminal chamber, lying along ventral sur- face of prostate and extending posteriorly to point where vas deferens enters ргоуадта; diverticulum with short, broad appendix about a third of the length of prostate-Il. As in female, hypobranchial gland in male lying completely posterior to pallial gonoduct and discharging into mantle cavity by short duct. Proserpina (Despoenella) depressa (Orbigny) Baker's (1926b) description of the radula is quoted here for comparison with P. (s.s.) nitida Sowerby: “The rhachidian central... consists of a thin plate with parallel sides. Its anterior edge is weakly notched and has no sign of a reflection or cusp, although its anterior half is slightly thickened. Its posterior edge 1$ very tnin, quite irregular and somewhat pointed. The A-cen- tral (A) is smaller than the B-plate (B) which is the reverse of their relative sizes in the Vianinae. . .” “The D-plate is a T-lateral with a broadly crescentic reflection (about half as deep as wide) and a short, stout, stalk. Under dry lenses its cutting edge appears simple and smooth, but under an oil-immersion objective, the entire upper surface is seen to be beauti- fully striate at right angles to its free margin, which as a result becomes very minutely ser- rate in worn teeth. The E-plate (E) is relatively larger than, but quite similar in structure to that of most Vianinae; its upper one-fourth is very firmly cemented behind the outer portion of the D-lateral.” “... fifty-three to fifty-five uncini are present on each side. The first twenty-two are uni- cuspid; the next three to five are bicuspid; while the outer teeth increase the number of cusps. The innermost marginals consist of a broadly triangular plate which is thickened at the base and twisted posteriorly and reflected at its upper angle so as to form a spatulate blade. The teeth increase in length from the inside out and the blades become larger out to about the 12th tooth. The outer marginals are lingulate and multicuspid; the outermost (40, 55) have broad reflected tips with numerous cusplets ... another specimen has 66 таг- ginals on each side.” MAJOR TAXA AND PHYLOGENY Superfamily relationships Current classifications of the Gastropoda generally recognize three subclasses: Proso- branchia, Opisthobranchia, and Pulmonata. The Prosobranchia in turn are divided into three orders: Archaeogastropoda, Meso- gastropoda, and Neogastropoda (Keen, 1960; Fretter & Graham, 1962; Taylor 8 Sohl, 1962). Golikov & Starobogatov (1975) divided the Prosobranchia into three subclasses equal in rank to the Opisthobranchia and Pulmonata: Cyclobranchia, Scutibranchia, and Pectini- branchia. They placed the Turbinomorpha and the Neritimorpha in the Pectinibranchia along with most other prosobranchs other- wise referred to as the Mesogastropoda and Neogastropoda. However, the Turbinimorpha and the Neritimorpha are more like the Scuti- branchia in most of their anatomical traits and do not conform to their definition of the Pecti- nibranchia. For this reason, in part, the earlier classification of the Prosobranchia into Archaeogastropoda, Mesogastropoda, and Neogastropoda is followed in this paper. The Archaeogastropoda are also referred to as the Diotocardia because of the presence of two auricles on the heart. The Mesogas- tropoda and the Neogastropoda are collec- tively referred to as the Monotocardia be- cause of the presence of a single auricle. The Diotocardia have, in general, paired gills, two kidneys, two columellar retractor muscles, and the anal and genital openings at the posterior end of the mantle cavity. Wastes and reproductive products are liberated into the mantle cavity whence they are conveyed to the outside by excurrent water currents. With the evolution of a conispiral shell there is a strong trend toward reduction of paired organs to single organs because of mechani- cal pressure on the right side of the pallial region due to allometric growth of the left side. Coupled with this allometric growth is a change in the flow of water current into and out of the mantle cavity, so that the location of a gill on the left (incurrent) side and the loca- PROSERPINOID LAND SNAILS 11 tion of excretory openings on the right (excur- rent) side are favored. These trends culmi- nate in the Monotocardia with the evolution of a single (left) auricle, a single (left) gill, a sin- gle (left) kidney, a single (right) retractor muscle, a pallial gonoduct that conveys re- productive products to the anterior right cor- ner of the mantle cavity, and an extension of rectum to the anterior right corner of the man- tle cavity. The mollusks constituting the subject of the paper belong to the Superfamily Helicinacea, which in turn belongs to the Infraorder Neriti- morpha. The Neritimorpha is an infraorder within the Archaeogastropoda. The Helici- nacea are defined as follows: HELICINACEA Thompson, new superfamily Primitive pulmonate archaeogastropods with an exogastric septate shell. Primitively non-operculate. Primitive members with a lamellar barrier partially blocking aperture. More advanced members secondarily oper- culate. Lung a vascularized open mantle cav- ity. Gill and osphradium absent. Reproductive system diaulic or triaulic, with two or three functional openings. Pallial gonoduct well de- veloped. Spermatophores absent. Pallial rec- tum present, conveying waste products to outside of mantle cavity. Hypobranchial gland discharging into mantle cavity via a duct, in- corporated into reproductive system in the Helicinidae. Pedal nerve cords nearly parallel, and bearing primitive lattice-like arrangement of connectives; supra-intestinal nerve absent; zygoneury occurring between pleural ganglia, and almost occurring between pleural and pedal ganglia and between pedal-pedal ganglia, which are only demarcated by narrow zones where connectives would normally be. Radula rhipidoglossate with central field con- sisting of single rhachidian tooth. Lateral field consisting of A-, B-, and C-lateral, and next two teeth (rasping or scraping teeth), that combine to form capituliform complex in which D-lateral is functional rasping or scrap- ing tooth. Within the Neritimorpha the Helicinacea appear to be most closely related to the Neritacea (Neritidae and Septariidae) on the basis of similar radulae, diaulic reproductive systems, and nervous sytems. The Family Hydrocenidae frequently is placed in close relationship with the Helicinidae, but the hydrocenid radula, monaulic reproductive system, and single (right) columellar retractor muscle are so divergent from the more primi- tive anatomical states of the Helicinacea that only a remote relationship can be established on the basis of morphological data (Thiele, 1910; Bourne, 1911; Baker, 1925). The Hy- drocenidae should be placed in a separate superfamily, the Hydrocenacea Troschel, 1856, within the Neritimorpha. The Helicinacea generally have been con- sidered the most advanced group of the Archaeogastropoda because the only infor- mation available on the anatomy of the super- family relates to various species of Helici- nidae, the most specialized of the three fami- lies in the Helicinacea (Isenkrahe, 1867; Bourne, 1911; Baker, 1925, 1926a, 1926b; Thiele, 1931; Boss 8 Jacobson, 1975a). The various anatomical traits were attributed, in part, to the evolution of a conispiral shell. Characters in the Helicinidae supporting that classification are: (1) the absence of a right kidney, (2) the presence of a complete pallial gonoduct, (3) the presence of a rectum ex- tending to the front of the mantle cavity, (4) the absence of a right auricle and, (5) the ab- sence of gills. Within the Archaeogastropoda traits (2) and (3) occur only in the Neriti- morpha but they are present in nearly all Mesogastropoda and Neogastropoda. These traits are not necessarily advanced morph- ological traits consequential of the develop- ment of a conispiral shell as has been sug- gested. An alternative hypothesis is that they are consequences of the evolution of a land snail from a diotocardian marine ancestry. To begin with, neither the Helicinacea nor the Neritacea has a conispiral shell. Basically the shell is limpet-like. Growth occurs in an exo- gastric direction with partial distortion to the right; but as the shell grows, the right wall dissolves away internally and produces a septate shell. The extent to which growth oc- curs is evident externally by the number of volutions produced on the apex; internally the only change that has occurred 1$ an increase in space. The snail's body remains limpet-like. In this respect the shell and body of the Heli- cinacea and the Neritacea are more primitive than the shell and body of the Turbinimorpha which are truly conispiral. Early in the evolution of the Gastropoda the primary gonoduct evolved to empty into the right renal duct, thus incorporating the right kidney into the reproductive system. Adapta- tion to a terrestrial environment requires an albumen coating for the egg which serves as a protective aqueous environment in which 12 THOMPSON the developing embryo can transform without danger of desiccation. This adaptation was accomplished by the evolution of the right kidney into the albumen-secreting provaginal sac of the Helicinacea (Baker, 1925), the pre- cursor of the albumen gland of the Neritacea and higher gastropods. In addition, the evolu- tion of a pallial gonoduct is prerequisite to a terrestrial mode of existence. Archaeogas- tropods, other than the Neritimorpha, have a simplified reproductive system in which eggs and sperm are released at the posterior end of the mantle cavity and are conveyed by water currents to the outside where fertiliza- tion takes place. A terrestrial mode of exist- ence requires the evolution of structural de- vices to facilitate fertilization and ovipositing to replace the water transport mechanisms of more primitive forms. Thus the pallial gono- duct of the Helicinacea is an adaptation for a terrestrial existence. This adaptation would be required for any terrestrial mollusk regardless of its phylogenetic level, and does not neces- sarily reflect a higher phyletic level. Coupled with the evolution of a pallial gonoduct is the evolution of a rectum that conveys waste products outside the mantle cavity. Archaeogastropods, other than the Neritimorpha, are not confronted with the problem of fouling of the mantle cavity for they are aquatic and the mantle cavity is continual- ly flushed by water currents. However, a ter- restrial snail does not have this cleansing mechanism and the evolution of a pallial rec- tum 15 a necessary adaptation to prevent foul- ing of the mantle cavity. As a matter of fact, there would be far greater adaptive pressure to evolve a pallial rectum in terrestrial gas- tropods than in aquatic groups. Loss of the right auricle of the heart has occurred in most species of the Helicinacea, although two auricles still persist in the two most primitive groups within the зирейатиу. In Ceres (Ceresidae) the right auricle is func- tional and is nearly as large as the left. In Hendersonia (Helicinidae, Hendersoniinae) the right auricle is functional but very much reduced in size. п other helicinaceans the right auricle 1$ lost. The loss of the right auricle in more advanced neritimorphs is a conse- quence of the crowding of the right side of the pallial region by the pallial gonoduct. Clearly the loss of a gill is an adaptation for a terrestrial existence, and its absence in the Helicinacea is to be expected. In this connec- tion it should be noted that the gill of some neritids may not be homologous to the gill of other neritaceans. Fretter & Graham (1962: 307) and Bourne (1908: 853) show that in Theodoxus, a freshwater neritid, the gill is in- nervated by the left pleural ganglion, rather than the supraoesophageal ganglion as oc- curs in other archaeogastropods. It may be argued that the gill of Theodoxus is a new structure evolved to accommodate an aquatic existence in a snail that evolved from a gill- less ancestor. This view was favored by Simroth (1896-1907, 1910) and von Ihering (1877). The only difference between the nervous system of Neritacea and that of Heli- cinacea is in the divergence of the pedal nerve cords. In the Neritacea the cords strongly diverge at about a 60-75” angle (Bourne, 1908), which probably is a modifica- tion consequential to the widening of the foot for adhesion to the rock substrate of an aquat- ic environment by a limpet-like snail. In Helici- nacea the pedal cords are nearly parallel (Bourne, 1911; Baker, 1925; this study, Ceres nelsoni, Proserpina nitida), which correlates with the narrow, more mobile foot required for terrestrial movement. The nervous system of Helicinacea, like Neritacea, shows a specialization through zygoneury and loss that make it unlikely that either group could have been ancestral to other orders of Prosobranchia or to the Pul- monata. From the foregoing data, it is apparent that the Helicinacea are a gill-less pulmonate as- semblage of land snails that are properly placed in the Diotocardia. This group has a simplified arrangement of pallial organs due to a reduction in the number of heart cham- bers and excretory organs and the loss of a gill and an osphradium. A pallial gonoduct and rectum were evolved to accommodate terrestrial existence, and the hypobranchial gland is modified to discharge into the mantle cavity through a duct. Primitively this group was non-operculate, protecting the opening of the limpet-like shell with a partial septum and a lamellar barrier (Ceresidae, Proserpinidae). Secondarily, an operculum was evolved to close the aperture (Helicinidae). Which group of marine mollusks was ancestral to the Heli- cinacea 1$ not clear. However, it is apparent that on the basis of the shell, the operculum, the gill, the radula (Baker, 1923b), the heart, and the reproductive system (Fretter 8 Graham, 1962; Bourne, 1908) the Neritacea is not ancestral to the Helicinacea. Internal fertilization through a pallial gonoduct and the complete pallial rectum of the Helicinacea of- PROSERPINOID LAND SNAILS 13 fer advantages that allow these systems to persist in more advanced aquatic groups. They are not required for an aquatic mode of life (as in the Turbinimorpha), but they are required for a terrestrial mode of life. Once evolved they are likely to be retained in ter- restrial or aquatic lineages. FAMILY RELATIONSHIPS WITHIN THE HELICINACEA The Helicinacea include three families, the Helicinidae, Ceresidae, and Proserpinidae. The Helicinidae is further divided into three subfamilies, Helicininae, Hendersoniinae and Vianinae. The latter two subfamilies apparent- ly are natural groups definable by anatomical criteria and shell charcteristics. The Helici- ninae are a heterogenous assemblage that includes several disparate groups. Two of these (‘Ceratodiscinae” and "Stoastoma- tidae”) are separable from the helicinids (s.s.) on the basis of shell and opercular traits. The few observations published on their radula do not show significant differences from Helici- ninae (Pilsbry & Brown, 1910; Baker, 1922; Thiele, 1927). All other aspects of stoastomid and ceratodiscid soft anatomy are unknown, and so they are excluded from further discus- sion in this paper (see Boss, 1972, for a dis- cussion of the subgenera of Stoastoma and Boss, 1973, for a monograph of Cerato- discus). Twenty-three characters are useful for separating families and subfamilies within the Helicinacea and for showing relationships among the groups involved. Because of the structural diversity that occurs within the Heli- cinidae, it is necessary to redefine the family and its two subfamilies Hendersoniinae and Vianinae in order to discuss relationships within the Helicinacea. CERESIDAE Thompson, new family* Type-genus: Ceres Gray, 1856. This family has the following combination of characteristics: SHELL: (1) operculum ab- sent; (2) periostracum present; (3) shell marked with radial sculpture. EXTERNAL ANATOMY: (4) foot holopod; (5) mantle collar not extending out over shell; (6) tentacles long, slender; (7) axial cleft separating last whorl very short, about 0.1 whorl long; (8) heart with two functional, nearly equal-sized auricles; (9) kidney broad, irregularly ovate in shape; (10) hypobranchial gland very long, overlapping posterior half of pallial gonoduct; (11) hypobranchial duct extending length of pallial oviduct; open along lower half. RE- PRODUCTIVE SYSTEM: (12) rectum and pallial gonoduct terminating at mantle collar; (13) vagina opening directly into posterior corner of mantle cavity, not inside hypo- branchial duct; (14) gonad large, flattened, oval in shape; (15) egg sac present at origin of primary oviduct; (16) primary gonoduct very short, thick; (17) prostate not divided into up- per and lower division, provaginal sac absent in males; (18) accessory sperm sac consist- ing of tubular bulb at left end of pedicel oppo- site ascending limb of V-organ; (19) crystal- line gland absent. RADULA: (20) A-, B-, and C-lateral teeth with 2-3 serrated cusps; (21) capituliform complex consisting of а comb- lateral (D-lateral) and accessory plate (E- lateral); (22) accessory plate (E-lateral) with broad wing enveloping end of D-lateral; (23) innermost marginal teeth with three cusps, outer marginals polycuspid. Characteristics 7, 8, 9, 10, 11, and 17 are unique to the Ceresidae. A prior family-group taxon name, Pro- serpinellinae (Baker, 1923a) was proposed for members of this family, based upon the oldest named genus within the group. Con- sidering the scant information available about Proserpinella it is ill-advised to base a family name оп a genus that 1$ so poorly known. There is no marked precedent in malacology for giving priority to the oldest name available for families (Baker, 1956a, 1956b), nor can there be where so many names were spuri- ously founded. Family PROSERPINIDAE Gray, 1847, REDEFINED Type-genus: Proserpina Sowerby, 1839. SHELL: (1) operculum absent; (2) perio- stracum absent; (3) shell smooth, without radial or spiral sculpture. EXTERNAL ANAT- ОМУ: (4) foot aulacopod; (5) mantle collar ex- tending fully or partially out over shell; (6) ten- tacles long and slender; (7) axial cleft about 34 whorl long; (8) heart with single auricle (left); (9) kidney narrowly crescent-shaped; (10) hypobranchial gland short, triangular in shape, confined posteriorly to pallial gonoduct; (11) hypobranchial duct short, opening at posterior *Thompson has already used this name in an abstract (Bulletin of the American Malacological Union for 1979 [published early 1980], p. 63). EDS. 14 THOMPSON end of mantle cavity. REPRODUCTIVE SYS- TEM: (12) rectum and pallial gonoduct termi- nating some distance from mantle collar as in Hendersoniinae; (13) vagina opening directly into posterior mantle cavity, not into hypo- branchial duct; (14) gonad huge, discoidal; (15) egg sac present at origin of primary ovi- duct; (16) primary gonoduct very short, stout; (17) prostate divided into two divisions, pros- tate-l and prostate-Il; provaginal sac vestigial within prostate-l; (18) accessory sperm sac consisting of tubular bulb at left end of pedicel opposite ascending limb of V-organ; (19) crystalline gland present at base of pallial ovi- duct. RADULA: (20) A-, B-, and C-lateral teeth unicuspid; (21) capituliform complex consist- ing of T-lateral (D-lateral) and accessory plate (E-lateral); (22) accessory plate (E-lateral) reduced in size, without wing; (23) innermost marginal teeth unicuspid. Outer marginals with few to several cusps. Characteristics 2, 3, 4, 5, 10, 17, and 19 are unique to the Proserpinidae. Family HELICINIDAE (HELICININAE) Férussac, 1822, REDEFINED Type-genus: Helicina Lamarck, 1799. (For anatomical data see Thiele, 1902; Bourne, 1911; Baker, 1926a.) SHELL: (1) operculum present, concentric; (2) periostracum present; (3) shell marked with radial and/or spiral sculpture. EXTER- NAL ANATOMY: (4) Foot holopod; (5) mantle collar not extending out over edge of shell; (6) tentacles long and slender; (7) axial cleft sep- arating last whorl of body about Y2 whorl long; (8) heart with single auricle (left); (9) kidney narrowly concentric in shape; (10) hypo- branchial gland elongate, overlapping poste- rior end of pallial gonoduct; (11) hypobranchi- al duct not extending beyond posterior half of pallial gonoduct. REPRODUCTIVE SYSTEM: (12) Rectum and pallial gonoduct terminating just behind mantle collar; (13) vagina opening into hypobranchial duct (female diaulic); (14) gonad smaller, elongate; (15) egg sac absent on primary ovary; (16) primary gonoduct rela- tively long; (17) prostate divided into prostate- | and prostate-Il. Provaginal sac absent in males; (18) accessory sperm sac located near middle of right side of pedicel, consisting of small bulbous sac at end of short narrow duct; (19) crystalline gland absent. RADULA: (20) A-, B-, and C-laterals usually with several cusps; cusps frequently reduced or absent on A-lateral; (21) capituliform complex consisting of comb-lateral (D-lateral) and accessory plate (E-lateral); (22) accessory plate with or without wing enveloping end of D-lateral; (23) innermost marginal teeth with 2-3 cusps. Outer marginals polycuspid. Only four traits found in all species exam- ined are unique to the Helicinidae (s.l.) This small number is due to the anatomical diversi- ty within the family and the structural modifi- cations and losses that have occurred within the various phyletic lines. These traits are: (1) operculate, (13) vagina opening into hypo- branchial duct, (14) gonad small and elon- gate, and (16) primary gonoduct moderately long. Within the Helicinidae several trends occur which progress from generalized states, as found in the Hendersoniinae, to modified states, as found in the Helicininae on the one hand and in the Vianinae on the other (Baker 1925, 1926a). These include: (a) modification of operculum from paucispiral type to concen- tric type, (b) increasing complexity of shell sculpture, (c) reduction and loss of right auri- cle, (d) increased length of pallial gonoduct and hypobranchial duct, (e) simplification and elongation of female primary oviduct, (f) trans- location of accessory sperm sacs on pedicel, (g) general reduction of cusps on radular teeth, (h) tendency for D-lateral tooth to change from comb-lateral to T-lateral, and (i) reduction in structural complexity of E-lateral tooth. Characters listed for Helicinidae also char- acterize the subfamily Helicininae. In the fol- lowing two subfamilies, Hendersoniinae and Vianinae, only characters that differ from the Helicininae are given. Subfamily HENDERSONIINAE Baker, 1926a, REDEFINED Type-genus: Hendersonia Wagner, 1905. (For anatomical data see Baker, 1925.) SHELL: (1) operculum paucispiral; (3) shell marked with radial sculpture. EXTERNAL ANATOMY: (7) axial cleft about 1 whorl long; (8) heart with two functional, unequal-sized auricles, right auricle almost vestigial; (11) hypobranchial duct short, opening into poste- rior end of mantle cavity. REPRODUCTIVE SYSTEM: (15) egg sac present at origin of primary oviduct; (16) primary gonoduct mod- erately long; (18) accessory sperm sac con- sisting of several small bulbs on left side of PROSERPINOID LAND SNAILS 15 pedicel opposite ascending limb of V-organ. RADULA: (20) A-, B-, and C-laterals with sev- eral cusps each; (22) accessory plate (E- lateral) with broad wing enveloping end of D- lateral; (23) innermost marginal teeth with three cusps, outer marginals polycuspid. Subfamily VIANINAE Baker, 1922, REDEFINED Type-genus: Viana H. and A. Adams, 1856. (For anatomical data see Isenkrahe, 1867; Baker, 1926a.) EXTERNAL ANATOMY: (6) tentacles short and conical in shape; (9) kidney narrowly crescent-shaped; (11) hypobranchial duct less than half length of pallial gonoduct. RE- PRODUCTIVE SYSTEM: (16) primary gono- duct long. RADULA: (20) A-, B-, and C-lateral teeth usually without cusps; (21) capituliform complex consisting of a T-lateral (D-lateral) and accessory plate (E-lateral); (22) acces- sory plate (E-lateral) reduced in size, without wing; (23) innermost marginal teeth unicus- pid, outer marginals with one or few cusps. FAMILY COMPARISONS Numbers in parentheses refer to the char- acteristics given for the families Ceresidae and Proserpinidae. These two families are alike in only seven charcteristics: (1) they lack орегсша, (6) the tentacles, (13) the opening of the vagina into the mantle cavity, (14) the size and shape of the gonad, (15) the pres- ence of an egg sac on the primary oviduct, (16) the structure of the primary gonoduct, and (18) the location of the accessory sperm sac. They differ in sixteen traits: (2) the peri- ostracum, (3) shell sculpture, (4) the foot structure, (5) the mantle collar, (7) the axial cleft, (8) the number of auricles on the heart, (9) the structure of the kidney, (10), the loca- tion of the hypobranchial gland, (11) the length of the hypobranchial duct, (12) the openings of the pallial gonoduct and rectum, (17) the structure of the prostate, (19) the presence of a crystalline gland, and (20-23) all aspects of the radula. Ceresidae and Helicinidae. The two fami- lies are alike in twelve characteristics: (2) the periostracum, (3) shell sculpture, (4) the foot structure, (5) the mantle collar, (6) the tenta- cles (except Vianinae), (10) the location of the hypobranchial gland (except Hender- soniinae), (12) the openings of the rectum and pallial gonoduct, (19) the absence of a crystal- line gland, and (20-23) similar radula (except Vianinae). The two families differ in eleven characteristics: (1) the presence of an oper- culum, (7) the axial cleft, (8) the number of auricles (except Hendersoniinae), (9) the structure of the kidney, (11) the length of the hypobranchial duct, (13) the opening of the vagina, (14) the size and shape of the gonad, (15) the presence of an egg sac, (16) the structure of the primary gonoduct, (17) the structure of the prostate, and (18) the location of the accessory sperm sac. Proserpinidae and Helicinidae. The two families are alike in eleven characteristics. Five are shared with only the Hendersoniinae. Similarities are: (6) the tentacles (except Vianinae), (7) the axial cleft (Hendersoniinae only), (8) the auricles (except Hender- soniinae), (9) the structure of the kidney, (10) the location of the hypobranchial gland (Hendersoniinae only), (11) the length of the hypobranchial duct (Hendersoniinae only), (12) the openings of the rectum and pallial gonoduct (Hendersoniinae only), and (20-23) the structure of the radula (Vianinae only). The two families differ in nineteen traits: (1) the presence of an operculum, (2) shell sculp- ture, (4) the foot structure, (5) the mantle col- lar, (7) the length of the axial cleft, (10) the location of the hypobranchial gland, (11) the length of the hypobranchial duct, (12) the openings of the rectum and the pallial gono- duct, (13) the opening of the vagina, (14) the size and shape of the gonad, (15) the pres- ence of an egg sac, (16) the structure of the primary gonoduct, (17) the presence of a prostatic provaginal sac, (18) the location of the accessory sperm sac, (19) the presence of a crystalline gland, and (20-23) all charac- teristics of the radula (except Vianinae). From the foregoing data it is clear that the Ceresidae and the Proserpinidae are less closely related to each other than either is to the Helicinidae. Furthermore, recognition of the three groups as separate families is war- ranted by the degree of evolutionary diver- gence that has occurred. The traits that are characteristic of the Cerisidae are primitive morphological states; whereas the traits unique to the Proserpinidae are advanced (derived) morphological states. The traits unique to Helicinidae indicate that it is also an advanced group compared to the Cerisidae but not to the same degree nor in the same lineage as is the Proserpinidae. 16 THOMPSON Morphological traits indicating these phylo- genetic relationships are as follows: The ceresid right auricle is functional and only slightly reduced in size. It persists in the Hendersoniinae as a small, functional vestige. It is completely absent in other helicinids and proserpinids. The ceresid (left) kidney 1$ large and ovate. In the other families it is reduced in size and is crescent-shaped. The vestige of the right kidney (provaginal sac) persists in both sexes in the Proserpinidae. The pro- vaginal sac persists only in females in the other two families. The vagina opens directly into the posterior corner of the mantle cavity in the Ceresidae and the Proserpinidae. In the Helicinidae it is incorporated into the hypo- branchial duct. п Ceresidae the gonad 1$ large and ovate, an egg sac 1$ formed at the base of the primary gonoduct, and the acces- sory sperm sac 15 located on the left side of the pedicel. With regard to the opening of the vagina, the size of the допаа, the presence of an egg sac, and the location of the accessory sperm sac, the Proserpinidae are similar to the Ceresidae in retention of primitive charac- ters as compared to the Helicinidae. In the Helicinidae the gonad 1$ reduced in size, an egg sac is absent, and the accessory sperm sac is translocated to the right side of the pedicel (except in Hendersonia). In the Ceresidae the prostate is undivided. In the other two families it is divided into prostate-I and prostate-Il. In addition, the Proserpinidae has evolved a crystalline gland, de novo, at the base of the reception chamber. The ceresid radula is generalized in all its traits. The central field teeth are heavily cusped, all the marginal teeth are multicusped, and the capituliform complex has a generalized comb- lateral and accessory plate. In Helicinidae a complete transition occurs in cusp reduction, transformation of the comb-lateral to a T- lateral, and simplification of the accessory plate. In the Proserpinidae these changes also are completed. Similarities between the proserpinid and the vianid radula apparently are due to con- vergence, for little morphological similarity otherwise exists between the two groups. On the contrary, greater morphological similarity exists between the proserpinids and the hendersoniines than between the proserp- inids and other groups. The traits unique to the Ceresidae, Proserpinidae and Helicinidae necessitate recognizing these groups as dis- tinct families. The aulacopod foot and the crystalline gland of the Proserpinidae are suf- ficient reasons for separating that family and indicate an extensive degree of divergence from the other two families. MINOR TAXA AND SYSTEMATIC OBSERVATIONS CERESIDAE Thompson, 1980 Type-genus: Ceres Gray, 1856. The Ceresidae are known only from Mexico and South America and contain five genera. Ceres is the only genus that is known ana- tomically. The radula of Linidiella has also been described. It is like the radula of Ceres and unlike the radula of the West Indian Proserpinidae. On the basis of shell structure Linidiella is most similar to Proserpina, but a close relationship (family) between the two 15 not tenable on the basis of their radular differ- ences. Thus Linidiella is tentatively referred to Ceresidae. The other three mainland genera are also provisionally assigned to Ceresidae because their shells are more similar in struc- ture to Linidiella than to Proserpina. Because the radula of Linidiella is a generalized type of helicinacean radula, similarities fo Ceres may only indicate a generalized relationship within the Helicinacea. Additional anatomical data on the South American genera may necessi- tate further division at the family or subfamily level. Dimorphoptychia from the Paleocene of Europe has shell characters that could place it in the Ceresidae. Wenz (1938: 435) places it in a separate subfamily, the Dimorphoptych- inae. Since Dimorphoptychia is known only as a fossil shell, speculation about its rela- tionship to modern groups is highly arbitrary. | find no advantage in uniting it in the same family with Ceres, because shell characters are not absolutely useful for showing relation- ships among modern families (e.g., Pro- serpina and Linidiella). KEY TO THE GENERA OF CERESIDAE 1) Shell 15-25 mm in major diameter; rugosely sculptured above, striate below; strongly keeled at periphery; with six apertural lamellae—two parietal, one columellar and three palatal о nA le A AS A O A Hr Ceres PROSERPINOID LAND SNAILS 17 Shell seldom over 15 mm in major diameter; sculptured with weak growth striations and occasionally microscopic granules or punctations; periphery rounded; aperture with 0-2 lamellae confined to columella and/or parietal wall ................................. 2 Aperture with two lamellae, one on parietal wall and one on base of columella . Staffola Aperture мин теме thant twOwlaMmMella@e ss... 2 vum eee бо eee een 3 Aperture without lamella, although a small denticle may be present on base of columella . 3a) Aperture with single lamella............. 4) Lamella confined to columella ........... 4a) Lamella confined to parietal wall........ Ceres Gray, 1856 Ceres Gray, 1856: 100. Type-species: Caro- colla eolina Duclos, 1834, by subsequent designation (Kobelt, 1879: 203). The shell is characterized by having six lamellae within the aperture: two parietal, one columellar, and three palatal. The shell is strongly keeled, has rugose sculpture on the spire and bears strong growth striations be- low. Three species have been described from eastern Mexico. The anatomy of one, С. nel- soni, and the radula of another, С. salleana, are described earlier in this paper. Ceres eolina (Duclos) Carocolla eolina Duclos, 1834: pl. 30. Odontostoma (Carocolla) eolinum (Duclos), Pfeiffer, 1848: 11. Proserpina eolina (Duclos), Pfeiffer, 1853: 290; Martens, 1890: 44; Martens, 1901: 609. Ceres eolina (Duclos), Gray, 1856: 102; Pfeif- fer, 1856: pl. 35, figs. 23, 24. Type-locality: State of Veracruz, Mexico. Distribution: Mexico, Veracruz: Cerro de Palma, Sierra de Matlaquihahuitl, near Toxpan (Martens, 1901: 609). This is the only locality recorded for this species. Ceres nelsoni Dall Ceres nelsoni Dall, 1898: 27-28: Dall, 1902: 501, pl. 28, figs. 1, 3, 5, 8; Solem, 1954: 7. Type-locality: Pilitla [Xilitla], Potosí, Mexico. Distribution: Known only from the states of San Luis Potosí and Tamaulipas in eastern Mexico. Specimens examined.—San Luis Sant Luis Museum of Zoology, University of Michigan. E See ees A Archecharax n.g. Potosi: Sotano del Rancho de la Barranca, 5km ММЕ Ahuacatlan (UF 24404—1 shell, UF 24405—6 preserved animals); Sotano de Guadelupe, 10 km SW Aquismon (UF 24406 —5 preserved animals, UF 24903—1 shell); 19 km E Xilitla, 350 m alt. (UF 24091a—1 pre- served animal, UF 24901—19 shells, UF 24902—5 shells); 10 km NE Xilitla, 300 т alt. (UF 24900—5 shells); 12km Е Xilitla, 730 m alt (UF 24899—1 shell). Tamaulipas: Solem (1954: 7) records this species from Aserradero del Paraiso, 15 km NNW Chamla (UMMZ2). Ceres salleana Gray Ceres salleana Gray, 1856: 100-102; Pfeiffer, 18/0: 295. Proserpina salleana (Gray), Pfeiffer, 1856: 322, pl. 35, figs. 21, 22; Martens, 1890: 45. Type-locality: Cordera cruz, Mexico. Distribution: Known from localities immedi- ately near the type-locality in Veracruz, Mex- ico: Orizaba; Cerro de Palma, Sierra de Matlaquihuahuitl, near Toxpan (Martens, 1901: 609); Barranca de las Puentes (Mar- tens, 1890: 45). [Cordova], Vera- Staffola Dall, 1905 Staffola Dall, 1905: 202. Type-species by monotypy: Proserpina (Staffola) derbyi Dall, 1905. Staffola is a monotypic genus which is characterized among ceresids by having two lamellae, one on the parietal wall and one projecting downward from the base of the columella. Sculpture consists only of a few incremental striations. However, the only 18 THOMPSON known material of the genus consists of the eroded subfossil holotype of $. derbyi and details about the sculpture cannot be ade- quately determined. The base of the shell has a thin umbilical callus. Keen et al. (1960: 1288) and Boss & Jacob- son (1975a, 1975b) synonymized Staffola with Cyane (=Archecharax). Even though $. derbyi hitherto has been unfigured, Dall's de- scription gives due notice to the columellar lamella and the parietal lamella, which imme- diately separates Staffola. lt was for this rea- son that Dall proposed a subgeneric relation- ship between Staffola and Proserpina, but continued to recognize Cyane (=Arche- charax) as a distinct genus. Additional observations on the holotype of 5. derbyi are necessary to characterize Staffola among other South American genera. The holotype of $. derbyi is redescribed and figured herein. Staffola derbyi (Dall) Proserpina (Staffola) derbyi Dall, 1905: 202. Type-locality: Calcareous banks of the arroyo of the Rio Chico at Paraguassa, State of Bahia, Brazil; holotype: USNM3 185454. Shell (Figs. 19-21): 4.9 mm in major di- ameter; depressed globose, being about 0.55 times as high as wide. Spire weakly convex in outline with sharply pointed apex. Apical whorl slightly raised above succeeding whorl. Shell very thick for its size, strongly callused internally along peristome. Whorls 4.0. Suture not impressed, vaguely apparent along last whorl. Thin callus superimposed on suture forming a rather uniformly narrow spiral band. Body whorl nearly uniformly rounded periph- erally but with tendency to flatten below pe- riphery. Details of surface sculpture of holo- type not clear due to weathered condition of shell, but few weak incremental growth wrin- kles parallel to peristome distinguishable on shoulder of last eighth whorl. Base of shell with thin umbilical callus extending out as far as parietal lamella. Callus bearing weak mi- nute granules, most of which are eroded. Umbilical region indented due to abrupt verti- cal descent of columellar wall of last eighth whorl. Aperture semilunar, with parietal and columellar lamella. Parietal lamella relatively thick and low, about one-eighth whorl long and lying about a third of distance from col- 3National Museum of Natural History, Washington D.C umella to posterior angle of aperture. Colu- mellar lamella projecting obliquely downward as tongue-like projection from columella, con- tinuing into shell for about eighth whorl, and forming narrow bay-like notch at base of col- umella. Columella oblique in frontal view, ridged above and curved forward at base. Dorsal lip deeply indented near suture. Outer lip receding basolaterally, as does basal lip near columella. Peristome sharp-edged but with strong internal callus. Remarks: This species is unusual because of its thick shell. No other ceresid approaches the condition that occurs in $. derbyi. Inas- much as all ceresids live on calcareous rocks, the thickness of the shell in this case cannot be attributed to a factor of the habitat, but al- most certainly 1$ intrinsic. Another peculiarity of the species is the strongly receded dorsal lip that forms a distinct notch near its insertion with the preceding whorl. The depth of the notch is partially obscured in the holotype be- cause the edge of the lip just outside the notch is broken. In Fig. 21 this is reconstruct- ed on the basis of the curvature of the adja- cent non-broken parts of the lip. The holotype of S. derbyi is eroded to the extent that the surface sculpture and details are obscured, and the outline of the suture is only apparent. My figure shows the course of the superim- posed callus and not the underlying suture. It is notable that there is no perceptible impres- sion of a suture. Other ceresids have at least a weakly impressed suture separating the whorls. S. derby ¡ diverges strongly from other spe- cies in the structure of the apertural lamellae. It is unique among South American ceresids in possessing a parietal lamella, or to put it another way, Linidiella and Archecharax are unique by lacking a parietal lamella. The col- umellar lamella is dissimilar in its basic fea- tures to other ceresid genera. In constrast with other genera, the columellar lamella pro- jects obliquely downward as an extension of the columella and forms a narrow U-shaped notch with the basal lip. The lamella resem- bles a tongue-like projection from the colu- mella and curves into the apenture for about a quarter of a whorl. lt appears to be a deriva- tion from the truncate columellar condition such as occurs in Archecharax. It has little similarity to the columellar lamella of Linidiella or Proserpina, in which genera the columella tapers into the basal lip, and the lamella lies at PROSERPINOID LAND SNAILS 19 21 FIGS. 19-21. Staffola derbyi (Dall). (Holotype: USNM 185454). a right angle on the columella about a third of the distance from the parietal wall to the basal lip. Apparently the columellar lamella of Staffola, Linidiella and Proserpina are тае- pendently derived, representing cases of convergent evolution in this structure. Linidiella Jousseaume, 1889 Linidiella Jousseaume, 1889: 256; Baker, 1923: 84. Type-species by subsequent designation (Baker, 1923: 84): Proserpina swifti Bland, 1863. Chersodespoena Sykes, 1900: 136. Type- species by original designation: Despoena (Chersodespoena) стпатотеа Sykes, 1900. Shell with single lamella at base of colu- mella. Columella concave, thickened, and grading into lamella. Sculpture consisting of fine, irregularly spaced incremental striations that become more distinct on periphery and base. Interspersed between striations on base of shell are numerous minute elongate gran- ules which become more concentrated near umbilical region. Umbilical callus not evident; indicated at best by concentrated granular sculpture. Dorsal surface of shell lacking granular or punctate sculpture. Enamel de- posit overlapping suture to form spiral line Iy- ing about midway between sutures. The radula is discussed earlier in this paper. Linidiella contains three species. Two are from the Andes of northern South America and one is from Chiapas, Mexico. They are placed together in Linidiella because each has а spiral lamella at the base of the colu- mella. | suspect that the similarity among the species based on this character is superficial and that two separate lineages are repre- sented. L. sulfureous from Chiapas differs significantly from the two Andean species in that its shell is nearly devoid of granular sculp- ture, and incremental striations are hardly dis- tinguishable. The species are as follows: each is known certainly only from its type locality. 20 THOMPSON Linidiella swifti (Bland) Proserpina swifti Bland, 1863: 16-17; Bland, 1965. 139, НОТ. Cyane swifti (Bland), Thiele, 1927: 90, fig. 65. Type-locality: mountains between Puerto Cabello and Valencia, Venezuela. Miller (1879: 148) listed this species from “Ecuador” on the basis of specimens ob- tained by Higgens. Aside from this record the species is known only from the type-locality. The specimens that | have examined (UF 19053) are merely labeled “Venezuela.” They came from C. T. Simpson and may have been received from Thomas Swift, who originally discovered this species. These specimens are figured (Figs. 22-23) to contrast Linidiella with Staffola and Archecharax. Linidiella cinnamomea (Sykes) Despoena (Chersodespoena) cinnamomea Sykes, 1900: 136-137, fig. Type-locality: between Ayabamba and Santa Rosa, Ecuador. u & Ne eee 5mm 24 Linidiella sulfureous Thompson Linidiella sulfureous Thompson, 1967: 61, figs: 1-3: Type-locality: 8.2 mi. $5. Solusuchiapa, Chiapas, Mexico; 1600 ft. alt. Archecharax Thompson, new generic name Cyane H. Adams, 1870: 376. Non Cyane Felder, 1861; Lepidoptera. Type-species: Cyane blandiana H. Adams, 1870 Etymology: The name Archecharax is de- rived from the Greek arche, first cause, and charax, a pointed stake. It alludes to the struc- ture of the columella. The name is masculine. A genus of the family Ceresidae with the following characteristics (all aspects of its soft anatomy are unknown). Aperture lacking in- ternal lamella; columella truncate, may project forward basally as denticle; periostracum ab- sent. Sculpture particularly noticeable on base and side of whorls, consisting of rather WO и FIGS. 22-24. Linidiella swifti (Bland). Figs. 22-23. Two specimens from Simpson Collection (UF 19053). Fig. 24. Radula redrawn from Thiele, 1931: 90. PROSERPINOID LAND SNAILS 21 regularly spaced incremental striations be- tween and within which numerous small granules occur in radial patterns. Archecharax is immediately distinguishable from other ceresids and proserpinids by hav- ing a truncate columella and lacking lamel- lae within the aperture. Equally striking is the presence of regularly spaced growth striations and radially arranged granular tubercles on the base and sides of the whorls outside the basal callus. Archecharax is known from the foothills and outer mountain ranges of the Andes from Colombia south to Bolivia. Four extant spe- cies are described. These may be separated by the following key: KEY TO THE SPECIES OF ARCHECHARAX 1) Dorsal and ventral surface with conspicuous spiral rows of punctate sculpture ......... la) Shell without spiral or punctate sculpture M an A O A. blandianus (H. Adams) 2) Shell large, 13-15 mm in major diameter; depressed, less than 0.55 times as high as wide . 2a) Shell smaller, less than 10 mm in major diameter; conic-globose, more than 0.60 times as high as wide de nn - SE RE A. orbignyi (Ancey) 3) Shell yellowish with red spiral band about midway between suture and periphery; spire weakly convex in outline ............... ES a A A RE A. cousini (Jousseaume) 3a) Shell uniformly amber colored; spire weakly concave in outline .A. glaeserius new species Archecharax blandianus (H. Adams) Cyane blandiana H. Adams, 1870: 376; pl. 27, 165. 2, 2а. Type-locality: “Eastern Peru.” This species has not been recorded since its discovery. Its original description and fig- ures are deficient in some details. lt is de- scribed and figures herein based upon mate- rial | collected in 1969. Shell (Figs. 25-27): Depressed-helicoid, about 10 mm in major diameter and about 0.59-0.69 times as high as wide. Largest specimen examined with about 5.3 whorls (UF 24359). Spire weakly concave in outline; apex rounded. Body whorl slightly swollen, evenly rounded at periphery; base moderately convex. Suture weakly impressed on last two whorls, not at all on earlier whorls. Umbilicus imperforate, but with a dimple-like impression that lies behind the columellar insertion. Pe- riphery of impression rather abrupt. Proto- conch consisting of 0.6 whorl, set off by a dis- tinct transverse crease. First half whorl of protoconch smooth, subsequent 0.1 whorl with a few weak radial striations; following whorls with incremental striations, within and between which are close spiral striations that may be broken into short linear segments or rows of shallow punctations (Figs. 31, 32). Spiral sculpture most conspicuous near su- ture and on base, weakest around periphery (Fig. 27). Parietal area without apparent cal- lus or deposit; spiral sculpture continuing into aperture undiminished. Aperture semi-lunar; without lamellae. Columella conspicuously thickened, concave and slightly twisted at base, forming very weak forward-projecting denticle. Dorsal lip extending forward and inserted well above periphery of preceding whorl. Outer lip and basal lip nearly straight in lateral profile. Size is highly variable. Measurements in mm for the five largest specimens examined are: Aperture Width Height Width Height Whorls ЧЕ 24355 7.0 4.6 3.2 3.1 5.0 UF 24356 7.7 SA 3.8 4.0 4.6 UF 24357 5.8 4.0 2.6 2.6 4.7 UF 24358 9.8 6.2 4.8 4.2 5.2 UF 24359. 7.3 4.3 3.3 2.8 5.0 Distribution: Known only from eastern Peru п vicinity of Tingo Maria. This is a region of extensive limestone karst formations. Specimens examined: Peru: Huanuco Province; Tingo Maria, 750m alt. (UF 24356.5); 4.7 km S Tingo Maria, 750 т alt. (UF 24350.4); 9.2km S Tingo Maria, 800 m alt. (UF 24357.1); 14.9 кт NE Tingo Мапа, 800 m alt. (UF 24358.2). THOMPSON 22 '80Lx “ze 614 pb Zpx Le 614 ‘eindinos лепиелб pue ¡eipes Buimoys di рицеа 111 aseq jo WAS “(9sepz 4M) (SWepy) snue/pue¡q xeleysayaıy ‘ZE-LE 'SODIW `(666фе qn :э4Аоюн) $эюэ4$ mau sn/ed9e¡b xeseyoayoly ‘05-82 “SOI ‘(6559 AN) (SWEPy) Snueipuejg хелецоэцолу "12-52 “SOI D PROSERPINOID LAND SNAILS 23 Remarks: The distinct spiral punctate sculpture on the dorsal surface of the whorls in A. blandianus is different from the sculptural traits of any other related species, and on the basis of that character alone separate generic status for A. blandianus is justifiable. Perhaps the three other species that | assigned to Archecharax should be placed in a new sub- genus because of the absence of such spiral sculpture. Until better material is available | prefer not to make such an allocation. Archecharax orbignyi (Ancey) Cyane orbignyi Ancey, 1892: 178. Type-locality: Santa Cruz de la Sierra, Bolivia. Distribution: known only from the type- locality. Archecharax cousini (Jousseaume) Proserpinella cousini Jousseaume, 181-182, ples, figs: 15, 16. Type-locality: “Ecuador.” Distribution: Ecuador, but not known from any precise locality. Archecharax glaeserius Thompson, new species Type-locality: Colombia, Departamento Valle, 3km W Atoncelo, 1380 m alt., Holo- type: UF 24355; collected 1 March 1969 by Fred G. Thompson. The type-locality is at the head of a deep ravine in a mountain rain for- est near the top of a mountain range lying west of Dagua and Antoncelo. The unique holotype was found at the base of a huge cal- careous sandstone boulder in the ravine. Shell (Figs. 28-30, 33-36): major diameter about 16 тт; depressed helicoid, being about 0.54 times as high as wide. Spire ele- vated, slightly concave in profile due to ex- pansion of last whorl. Periostracum absent. Shell opaque, glossy, amber yellow, except for white umbilical callus with regularly spaced, narrow darker yellow streaks parallel- ing line of growth, visible through surface gloss. Umbilical area with thin granular callus strongly indented immediately behind the col- umella due to abrupt vertical wall of last whorl at that point (Fig. 36). Whorls 4.7; suture moderately depressed and covered by thin glaze forming narrow transparent zone ex- tending onto preceding whorl and partially obscuring suture. Glaze with numerous small 1887: dimples and pits randomly dispersed over suture area. Protoconch with 1.0 whorl, smooth, elevated above succeeding whorl; very weakly set off from succeeding whorl by faint rest striation. Subsequent whorl smooth but with sparse, fine, incremental striations most noticeable on base and rarely distin- guishable above. Microsculpture on base consisting of numerous minute granules tend- ing to be aligned between incremental stria- tions and arranged in a spiral course. Gran- ules densest on basal callus and disappear- ing near periphery of whorl. Aperture 0.83 times as high as wide, deeply indented by preceding whorl. Lamella absent. Peristome simple, sharp; columella weakly concave, oblique; base truncate, extending forward as weak denticle accentuated by receding basal lip. Denticle does not continue internally as lamella but curves upward uniformly into col- umella. Measurements in mm of the unique holo- type are: width, 15.7; height, 8.4; aperture width, 7.0; aperture height, 5.8. Remarks: The species is most similar in its shell characters to A. cousini (Joussea- ume). Both species have a smooth spire de- void of spirally arranged rows of punctate sculpture, are similar in size and relative height, and both have a small denticle-like projection at the base of the columella. A. glaeserius is immediately separated from A. cousini by the color of its shell and the contour of its spire. A. glaeserius is uniform amber yellow and has a weakly concave spire. A. cousini possesses a red spiral band on a yel- low background, and has a weakly convex spire. A. cousini is known only from its holotype, for which Jousseaume gives only a brief de- scription and an outline illustration. Apparent- ly the relationship between glaeserius and cousini is close, but their differences are suf- ficient to consider them distinct species. Addi- tional collections may show they are subspe- cifically related. Proserpinella Bland Proserpinella Bland, 1865: 157. Type-species by monotypy: Proserpinella berendti Bland, 1865. The genus is characterized by having a smooth, discoidal shell that bears a delicate parietal lamella. Other lamellae are absent. The columella 15 truncate, similar to that in Archecharax. А thin umbilical callus 1$ pres- ent. 24 THOMPSON 36 5mm FIGS. 33-36. Archecharax glaecerius new species (Holotype: UF 24355). Fig. 36. Enlargement showing detail of columellar area. 10 mm scale for 33-35, 5 mm scale for 36. Proserpinella is an obscure genus of mi- nute Mexican land snails; two species have been described. Nothing is known about them other than the descriptions of their shells. Each is known only from its type-locality. Proserpinella berendti Bland Proserpinella berendti Напа, 1865: 157, fig. 2; Strebel, 1873: 11, pl. 4, fig. 5; Martens, 1890: 45. Type-locality: Mirador, Veracruz, Mexico, 3000-4000 ft alt. Proserpinella hannae Dall Proserpinella hannae Dall, 1926: 486-487, pl. 36, figs. 6-8. Type-locality: Maria Madre Island, Tres Marias Islands, Nayarit, Mexico. PROSERPINIDAE Gray, 1847 This family is endemic to the Greater Antil- lean Islands of Cuba, Jamaica, and Hispani- ola, and contains two genus-group taxa, Proserpina Sowerby, 1847, and Despoenella Baker, 1923. Conventionally Despoenella 15 treated as a subgenus of Proserpina. Equally valid reasons can be given for treating it as a separate genus. For the purposes of this pa- per | follow previous authors and treat them as subgenera. However, recognition of them as subgenera or distinct genera on the basis of our current knowledge 15 subjective. Pro- serpina is characterized by having a colu- mellar lamella, two parietal lamellae, and two PROSERPINOID LAND SNAILS 25 palatal lamellae. Proserpina is restricted to Jamaica and contains two species. Despoenella has a columellar lamella and a single parietal lamella. Palatal lamellae are absent. Despoenella contains two species each on Cuba and Jamaica and three on Hispaniola. Until now the only information available dealing with internal anatomy is Bakers (1926b) description of the radula of P. (Despoenella) depressa (Orbigny), the type- species of Despoenella. Data on the soft anatomy of P. nitida Sowerby, the type- species of Proserpina are presented earlier in this paper. These data are the basis of char- acterizing the Proserpinidae as a distinct fam- ily. Two new species of Despoenella are also described. In view of the excellent monograph on the Proserpinidae by Boss & Jacobson (1975a), further discussion of most other spe- cies is not necessary. Calybium from Southeast Asia may be a proserpinid, but it is only known from its shell and radula (see Baker, 1922: 64-65), and its relationship within the proserpinid-helicinid complex remains unclear. Proserpina Sowerby Subgenus Proserpina, $.5. Proserpina Sowerby, 1839: 124; Boss 4 Jacobson, 1975: 67-69. Type-species: Proserpina nitida Sowerby, 1839, by mono- typy. Despoena Newton, 1891: 255. New name for Proserpina Sowerby, 1839, non Proserpi- nus Hubner, 1816, Lepidoptera. Proserpina (Proserpina) nitida Sowerby Proserpina nitida Sowerby, 1839: 124, fig. 274; Boss & Jacobson, 1975a: 69-72, pl. 10, figs. 1-5. Proserpina nitida planulata С. В. Adams, 1851: 174: Type-locality: Jamaica. Distribution: widely distributed throughout the central portion of Jamaica. Proserpina (Proserpina) linguifera (Jonas) Helicina linguifera Jonas, 1839. Proserpina allognoto Jonas, 1846. New name for Helicina linguifera Jonas, 1839. Proserpina pulchra C. B. Adams, 1850: 81. Proserpina linguifera (Jonas), Pfeiffer, 1850: 12, pl. 103, figs. 12-15; Boss & Jacobson, 1975а: 72-74, pl. 10, figs. 6-7. Type-locality: Jamaica. Distribution: known only from St. Elizabeth Parish and Westmoreland Parish, Jamaica. Subgenus Despoenella Baker Odontostoma Orbigny, 1842: 238. Type- species: Odontostoma depressa Orbigny, 1842. Non Odontostoma Turton, 1830, Gastropoda. Despoenella Baker, 1923: 85. New name for Odontostoma Orbigny, 1842, non Odonto- stoma Turton, 1830. Boss & Jacobson, 1975а: 74. Proserpina (Despoenella) globulosa (Orbigny) Odontostoma globulosa Orbigny, 1842: 239, pl. 18, figs. 8-11. Proserpina globulosa (Orbigny), Pfeiffer, 1850: 12, pl. 12, figs. 19-21; Boss & Jacob- son, 1975a: 84-87, pl. 13, figs. 4-6. Type-locality: Interior of island of Cuba. Distribution: Widely disjunct in its distribu- tion in Oriente and Pinar del Rio Provinces, Cuba, and the Isle of Pines. Proserpina (Despoenella) pisum C. B. Adams Proserpina pisum C. B. Adams, 1850b: 108. Boss 4 Jacobson, 1975a: 82-84, pl. 13, figs. 103. Type-locality: Jamaica. Distribution: Confined to western Jamaica where it is found in Westmoreland, St. James, and Trelawny Parishes. Proserpina (Despoenella) depressa (Orbigny) Odontostoma depressa Orbigny, 1842: 238. pl. 18, figs. 4-7. Helicina ptychostoma Pfeiffer, 1848: 12. Proserpina depressa (Orbigny), Pfeiffer, 1853: 291; Baker, 1926b: 451 (radula): Boss & Jacobson, 1975a: 75-78, pl. 11, figs. 1-3. Proserpina depressa rubrocincta (Torre, MS, Aguayo & Jaume, 1947: 88 (nomen nudum); Aguayo & Jaume, 1957: 124, pl. 1, fig. 10. 26 THOMPSON Type-locality: Odontostoma depressa Orbigny: interior of the isle of Cuba; restricted by Aguayo 8 Jaume (1947: 88) to Pan de Guajaiboa, Pinar del Rio, Cuba. Helicina ptychostoma Pfeiffer: Callajabas [=Caya- jabos], Pinar del Rio, Cuba. Proserpina depressa rubrocincta Aguayo 8 Jaume: Los Acostas, Luis Lazo, Pinar del Rio, Cuba. Distribution: widely disjunct; confined to Pinar del Rio Province and Havana Province in western Cuba and Oriente Province т eastern Cuba. Proserpina (Despoenella) bidentata C. B. Adams Proserpina bidentata C. B. Adams, 1850a: 81; Boss 4 Jacobson, 1975a: 79-80; pl. 12, figs. 4—6. Type-locality: Jamaica. Distribution: confined to the John Crow Mountains, Portland Parish, Jamaica. Proserpina (Despoenella) marcanoi Clench Proserpina тагсапо! Clench, 1962: 2, pl. 1, fig. 3; Boss & Jacobson, 1975a: 80-82, pl. 12, figs. 1-3. Type-locality: Colonia Ramfis [=Colonia Majagual], 20km W of San Cristobal, San Cristobal Province, Dominican Republic. Distribution: known only from the type-lo- cality. Remarks: observations are given below with the following species. Proserpina (Despoenella) scudderae Thompson, new species Etymology: this species 15 named Юг Sylvia Scudder, Technician, Florida State Museum, who assisted in field work in the Dominican Republic in 1974. Type-locality: Dominican Republic, Bara- hona Prov., Sierra de Baoruco, 7km NNE Polo, 910 т alt. The type-locality is in a deep limestone ravine. On my first visit in January 1974 the ravine was shaded by a wet mountain forest that was partly planted with coffee. At the time of the most recent visit (January 1977) the ravine was deforested along both sides with cattle pasture on the 4Field Museum of Natural History, Chicago. Academy of Natural Sciences of Philadelphia. Museum of Comparative Zoology, Cambridge, Mass north slope and open coffee grove on the south slope. ' Holotype: UF 24326; collected 18 June 1974 by Fred G. Thompson. Paratypes: UF 24327 (19), UF 24328 (25), UF 24329 (5), FMNH4 195426 (2), ANSP5 (2), MCZ6 288377 (2). Museo Nacional de Historia Natural, Republica Dominica (5). All paratypes are topotypic. Shell (Figs. 37-41): small, 5.2-5.9 mm in diameter; discoidal, adult shells 0.47-0.49 times as high as wide (0.48 in holotype); whorls 4.5-5.0 (4.8 in holotype). Spire de- pressed. Embryonic whorl conspicuously protruding. Last whorl flattened above and rounded below so that periphery of shell lies above middle. Color light greenish-yellow dorsally, lighter below (only dead shells have been collected; live specimens probably are brighter green than the material | have ex- amined). Surface of shell glossy with thin enamel-like wash that overlaps suture and extends about halfway onto previous whorl. Shell fairly translucent, showing regularly spaced, thin incremental lines of growth through outer wash. Dorsal surface and sides of whorls smooth. Ventral surface with thin white basal callus that is very minutely granu- lar. Callus extending outward as arc continu- ing forward from parietal lamella. Columellar margin of umbilical area indented, forming short abrupt wall that causes base to be weakly pitted (Figs. 38, 40). Aperture semi- lunar, equal to or slightly higher than wide, 0.38-0.44 times width of shell (0.39 in holo- type). Aperture with parietal lamella and col- umellar lamella, both extending into aperture about 1/5 whorl (Fig. 41). Parietal lamella located about third distance from columella to posterior angle of aperture. Columellar lamel- la located just above middle of columella and about half as high as parietal lamella. Lip strongly sinuous in outline, strongly receded at periphery and along base. Columella oblique, lying at about 30° to vertical axis of shell. Columella accentuating umbilical pit by having pillar-like thickening between parietal wall and columellar lamella. Measurements of 12 specimens selected to show maximum variations (holotype in paren- theses): height, 2.5-2.8 mm (2.7); width, 5.2- 5.9 mm (5.6); aperture height, 2.1-2.4 тт (2.3); aperture width, 2.1-2.3 mm (2.2). PROSERPINOID LAND SNAILS 27 41 FIGS. 37-41. Proserpina (Despoenella) scudderae new species. Figs. 37-38. Large paratype (UF 24328). Figs. 39—40, small paratypes (UF 24328). Fig. 41. Paratype opened to show lamella (UF 24328). Distribution: this snail has been found only in the Sierra de Baoruco, near Polo, Domini- can Republic, where it was collected on lime- stone outcrops in wet forests. Records in ad- dition to the type-locality are: 6 km NNE Polo, 1000 m alt. (UF 24332); 5km NNE Polo, 990 m alt. (UF 24331); 2 km NNE Polo, 765 m alt. (UF 24333); 3 km SE Polo, 750 т alt. (UF 24330). Remarks: this is a member of the sub- genus Despoenella by virtue of possessing two lamellae within the aperture, a parietal lamella, and a columellar lamella. It differs from most of its subcongeners by its discoidal shape and its protruding embryonic whorl. Its differences from P. planior are described be- low under that species. Adult shells are less than 0.50 times as high as wide with the pe- riphery lying above the middle of the last whorl. Other species of Despoenella, except P. planior, are helicoid or depressed-helicoid in shape with the periphery of the shell lying at the middle of the last whorl, and the embry- onic whorl is not conspicuously elevated above the succeeding whorls. The subgenus Despoenella is divisible into 28 THOMPSON three species groups. One group includes P. pisum C. B. Adams from Jamaica and P. globulosa (Orbigny) from Cuba. They look alike in having globose or subglobose shells (see Boss & Jacobson, 1975a). These two species need not be considered further for comparison with P. scudderae because of their shapes. The second species group in- cludes P. depressa (Orbigny) from Cuba, P. bidentata C. B. Adams from Jamaica, and P. marcanoi Clench. They are alike in having depressed helicoid shells. Differentiation with- in this group is slight. P. depressa differs by its larger size. Adults attain a diameter of about 7-8 тт. Р. bidentata and Р. тагсапо! reach a diameter of about 5 mm. P. bidentata and P. marcanoi are hardly separable. P. bidentata has a weaker indentation at the base of the columella than does P. тагсапо! but other shell differences are nonexistent. They are treated as distinct species because they occur on different islands (Boss & Jacobson, 1975a). A third species group in- cludes P. scudderae and P. planior from His- paniola which differ from other Despoenella by 44 their discoidal shape, with a height/width ratio of less than 0.50, and having protruding embryonic whorls. (See Figs. 42-44 for сот- parisons with marcanoi, Figs. 47-48 for bidentata, and Figs. 45-46 for depressa.) Proserpina marcanoi, P. scudderae, and P. planior are found in Hispaniola. Each is highly restricted in its geographical distribu- tion. P. тагсапо! is known only from its type- locality, Colonia Majagual, San Cristobal Prov., Dominican Republic (formerly known as Colonia Ramfis). This is a small community located on the road from Gambito Garabitos to El Guineo, and is about 12km NW of Gambito Garabitos. The area is mountainous and formerly was covered with wet forest which 1$ replaced with coffee groves. The substrate consists primarily of metamorphic and igneous rocks. There are a few isolated outcrops of highly metamorphosed lime- stones. Р. тагсапо! is known only from the three specimens that comprise the type series. | visited the region of its type-locality on four occasions and was unsuccessful in finding additional specimens. Р. scudderae 1$ 3mm FIGS. 42-44. Proserpina (Despoenella) marcanoi Clench (Holotype: MCZ 188911). PROSERPINOID LAND SNAILS 29 46 47 FIGS. 4546. Proserpina (Despoenella) depressa (Orbigny) (UF 24115). FIGS. 47-48. Proserpina (Despoenella) bidentata (C. B. Adams) (UF 24114). known only from the immediate vicinity of Polo, Sierra de Baoruco, Barahona Prov., Dominican Republic. P. planior is restricted to the Plateau de Rochellois on the Tiburon Peninsula of Haiti. The area formerly was covered with wet forests on a limestone sub- strate, but is now reduced to vegetable gar- dens and a few isolated thickets of brush on limestone outcrops. The extremely isolated and disjunct ranges of these species indicate relictual distributions for the genus on Hispaniola. Each species is confined to a small geographic area, lying at higher, relatively cool and moist elevations on limestone substrates. | have collected at many other placed on Hispaniola that would seemingly comprise suitable habitats for proserpinids, but have not found other popu- lations, in contrast to my experience т Jamaica (1976) where | found proserpinids common in occurrence. Possibly other spe- cies occur on Hispaniola, but their discovery will be extremely fortuitous! Proserpina (Despoenella) planior Thompson, new species Etymology: planior: from the Latin, planus, meaning more flattened, alluding to the dis- tinctive shape of this species compared to other Proserpina. Type-locality: Haiti, Departement du Sud, Plateau de Rochellois, 22 km SW Miragoáne, 930 m alt. Holotype: UF 26566, collected 31 March 1979 by Fred G. Thompson and Richard Franz; Paratypes: UF 26565 (6); same data as holotype; UF 26564 (10), col- lected at type-locality 12 May 1979 by Fred G. Thompson and Kurt Auffenberg. The type-locality 1$ on the south slope of a small knoll covered by a dense thicket of shrubs and small trees. The area once was densely forested but has been cleared for fuel and agriculture. Shells were found among limestone boulders. Shell (Figs. 49-51): minute; adults about 3.8-4.5 mm in diameter. Nearly planispiral 30 THOMPSON 0 1 2 3 4 5 mm EA AA A AAA FIGS. 49-51. Proserpina (Despoenella) planior new species (Holotype: UF 26566). FIGS. 52-54. Proserpina (Despoenella) scudderae new species (Paratype: UF 24327). (Fig. 49); 0.42-0.46 times as high as wide; apical whorl slightly elevated as a sharp nip- ple-like protrusion. Light greenish-yellow when fresh, with a spiral white band on apex due to an internal callus along first 2-3 whorls (Fig. 51). Sides and base thin, subtranspar- ent. Dorsal surface usually opaque due to in- ternal callus and a relatively thick glassy outer deposit that completely covers previous whorls and obscures sutures. Sutural impres- Sion apparent only along last two whorls. About 4.3-4.7 whorls in adult specimens (4.3 PROSERPINOID LAND SNAILS Si in holotype). Surface glossy with a few weak incremental wrinkles. Base with weakly gran- ular circum-umbilical callus. Columellar wall near aperture vertical, forming distinct angle or pit just behind columella (Fig. 50). Aperture bluntly angular at periphery and at baso- lateral margin; baso-columellar angle more distinct; posterior angle of dorsal lip very nar- row and deep due to high insertion of dorsal lip which lies about halfway between periph- ery and sutural impression of previous whorl (Fig. 49). Peristome strongly arched forward along dorsal lip and relatively deeply re- ceded at suture (Fig. 51); lateral lip deeply receded at periphery; basal lip conspicuously arched forward, but not as much as dorsal lip (Fig. 50). Columella slightly oblique. Interior of aperture with parietal and columellar lamellae. Parietal lamella about Уз whorl long and lo- cated at about ‘3 distance from columella to periphery. Columellar lamella about Y2 whorl long and relatively low and thin compared to other species. Measurements in mm based upon seven specimens to show maximum variation in size (holotype in parenthesis): height 1.75-1.97 (1.82); width 3.78—.46 (4.35); aperture height, 1.54-1.68 (1.68); aperture width, 1:47=1.:75:(1:75). Distribution: known only from the type- locality. Remarks: P. planior 15 most closely related to P. scudderae. The two species are similar to each other and differ from all other Pro- serpina by their depressed shapes and pro- truding apical whorls. They differ by several consistent characters. An immature paratype of P. scudderae (Figs. 52-54), comparable in diameter and whorl count, is illustrated for comparison to the holotype of P. planior. P. planior is characterized by its small size, at- taining a diameter of 3.8—4.5 mm, by its lower number of whorls, 4.3-4.7, and by its de- pressed, planular shape, being 0.42-0.46 times as high as wide. The dorsal lip is insert- ed high on the preceding whorl, about halfway between the periphery and the preceding suture, causing the posterior corner of the aperture to be very narrow and deep. The peristome 1$ strongly sinuous with the dorsal and basal lip strongly arched forward and the outer lip strongly receded. The apex bears a Spiral white band and an external callus de- posit that completely covers the preceding whorls. Adult P. scudderae are 4.2-5.9 mm in di- ameter, have 4.5-5.0 whorls and are 0.47- 0.49 times as high as wide. The dorsal lip is inserted about Ys the distance from the pe- riphery to the suture, causing the posterior angle of the aperture to be broader and shal- lower. The peristome is not as strongly curved, the apex is unicolor, and the apical callus overlapping the suture extends only about halfway across the preceding whorls. It may be argued that P. р/апюг and P. scudderae should be treated as subspecies because of their similarities. Such a designa- tion requires evidence that they intergrade, which they do not, either morphologically or geographically. The ranges of the two species are disjunct and are separated by a distance of about 300 km. | have collected at about 200 field stations in the intervening territory and have not encountered other populations of Proserpina. ACKNOWLEDGMENTS Many people have aided me in this study. | am grateful to all for their assistance. James Reddell, Texas Technological University, pro- vided me with preserved specimens of Ceres nelsoni Dall. Preserved animals of Proserpina nitida Sowerby were collected by Glenn Goodfriend. Joseph Rosewater, United States National Museum of Natural History, and Kenneth J. Boss, Museum of Compara- tive Zoology, loaned me specimens in their charges. Richard Franz, Sylvia Scudder, and Kurt Auffenberg, all of the Florida State Mu- seum, assisted me with field work in Hispani- ola. The SEMs, Figs. 5-10, Figs. 31-32, were made by Ms. Scudder. Field work in the Do- minican Republic relating to this project was financed by the Florida State Museum. Field work in Haiti during 1979 was financed by the National Geographic Society. LITERATURE CITED ADAMS, C. B., 1850a, Descriptions of supposed new species and varieties of terrestrial shells, which inhabit Jamaica. Contributions to Conch- ology, (5): 76-84. ADAMS, C. B., 1850b, Descriptions of supposed new species of land shells, which inhabit Jamaica (cont.) Contributions to Conchology, (7): 101-108. ADAMS, C. B., 1851, Descriptions of new species and varieties of the land shells of Jamaica, with notes on some previously described species. Contributions to Conchology, (9): 153-174. 32 THOMPSON ADAMS, H., 1870, List of additional species of land and freshwater shells collected by Mr. E. Bartlett in eastern Peru, with descriptions of new spe- cies. Proceedings of the Zoological Society of London, 1870: 374-377, pl. 27. AGUAYO, C. G. & М. L. JAUME, 1947-1951, Catalogo de los moluscos de Cuba. Havana: 1-725 (mimeographea). AGUAYO, С. С. & М. L. JAUME, 1957, Adiciones a la fauna malacologica Сибапа—1. Memorias de la Sociedad Cubana de Historia Natural, 23: 117-148. ANCEY, С. F., 1892, Descriptions de mollusques nouveaux. Le Naturaliste, ser. 2: 178. BAKER, H. B., 1922, Notes on the radula of the Helicinidae. Proceedings of the Academy of Natural Sciences of Philadelphia, 75: 29-67. BAKER, Н. B., 1923a, Proserpinidae. Nautilus, 36: 84-85. BAKER, H. B., 1923b, Notes on the radula of the Neritidae. Proceedings of the Academy of Natu- ral Sciences of Philadlephia, 75: 117-178. BAKER, H. B., 1925, Anatomy of Hendersonia: a primitive helicinid mollusk. Proceedings of the Academy of Natural Sciences of Philadelphia, 77: 273-303. BAKER, H. B., 1926a, Anatomical notes on Amer- ican Helicinidae. Proceedings of the Academy of Natural Sciences of Philadelphia, 78:29-56. BAKER, H. B., 1926b, The radula of Proserpina. Proceedings of the Academy of Natural Sci- ences of Philadelphia, 78: 449-451. BAKER, H. B., 1956a, Family names in Pulmonata. Nautilus, 69: 128-139. BAKER, H. B., 1956b, Family names for land oper- culates. Nautilus, 70: 28-31. BLAND, T., 1863, On the Family Proserpinacea with descriptions of a new species of the Genus Proserpina. Annals of the Lyceum of Natural History of New York, 8: 13-17. BLAND, T., 1865, Notes on certain terrestrial Mol- lusca, with descriptions of new species. Annals of the Lyceum of Natural History of New York, 9: 155-170. BOSS, К. J., 1972, Minute Jamaican prosobranch gastropods: Stoastoma and its congeners. Breviora, (393): 1-13. BOSS, K. J., 1973, Monograph of Ceratodiscus (Archeogastropods; Helicinidae). Occasional Papers on Mollusks, 3: 253-379. BOSS, К. J. & М. К. JACOBSON, 1975a, Proser- pine snails of the Greater Antilles (Prosobranchia; Helicinidae). Occasional Papers on Mollusks, 4: 53-90. BOSS, К. J. & М. К. JACOBSON, 1975b, Cata- logue of the taxa of the subfamily Proserpinidae (Helicinidae; Prosobranchia). Occasional Pa- pers on Mollusks, 4: 93-104. BOURNE, G. C., 1908, Contribution to the mor- phology of the group Neritacea of the aspido- branch gastropods.—Part |. The Neritidae. Proceedings of the Zoological Society of Lon- don, 1908: 810-887, pl. 46-62. BOURNE, G. C., 1911, Contributions to the mor- phology of the group Neritacea of the aspido- branch gastropods.—Part Il. The Helicinidae. Proceedings of the Zoological Society of Lon- don, 1911: 759-809, pl. 30-42. CLENCH, W. J., 1962, New®species of land mol- lusks from the Republica Dominicana. Breviora, (173): 1-5. DALL, W. H., 1898, A new species of Ceres from Mexico. Nautilus, 12: 27-28. DALL, W. Н., 1902, Illustrations and descriptions of new, unfigured, or imperfectly known shells, chiefly American, in the U.S. National Museum. Proceedings of the United States National Mu- seum, 24: 499-566. DALL, W. H., 1905, A new proserpinoid land shell from Brazil. Proceedings of the Biological So- ciety of Washington, 18: 201-202. DALL, W. H., 1926, Land shells of the Revillagi- gedo and Tres Marias Islands, Mexico. Proceed- ings of the California Academy of Sciences, ser. 4, 15: 467-491. DUCLOS, [Р. L.], 1834, Carocolle. Carocolla. Lamarck... Magasin de Zoologie, 3: classe 5, р 30: FRETTER, V. 8 А. GRAHAM, 1962, British proso- branch molluscs. Ray Society, London. i-xvi, 1- 159: GOLIKOV, А. М. & У. I. STAROBOGATOV, 1975, Systematics of prosobranch gastropods. Mala- cologia, 15: 185-232. GRAY, J. E., 1847, A list of the genera of Recent Mollusca, their synonyms and types. Proceed- ings of the Zoological Society of London, pt. 15: 129-219. GRAY, J. E., 1856, On the position of the genus Proserpina in the system, and a description of its dentition. Proceedings of the Zoological Society of London, 29: 99-102. IHERING, H. VON, 1877, Vergleichende Anatomie des Nervensystems und Phylogenie der Mol- lusken. Leipzig, p. 1-290. ¡SENKRAHE, C., 1867, Anatomie von Helicina titanica. Archiv fúr Naturgeschichte, 33: 50-72, figs. 1-11. JOUSSEAUME, F., 1887, Mollusques nouveaux de la République de l'Equateur. Bulletin de la Société Zoologique de France, 12: 165-186, РЁ 3: JOUSSEAUME, F., 1889, Mollusques du Venezu- ela (Voyage du M. Eugene Simon). Mémoires de la Société Zoologique de France, 2: 232-259, pl. 9. KEEN, A. M., 1960, in Treatise on Invertebrate Paleontology: Mollusca 1, Pt. 1: i-xxili, 1-351, Lawrence, Kansas. KOBELT, W., 1879, Illustriertes Conchylienbuch. Nürenberg. 2: 1-264. MARTENS, E. VON, 1890-1901, Biologia Centrali- Americana: Land and Freshwater Mollusca. London. i-xxvili, 1-709. MILLER, K., 1878-1879, Die Binnenmollusken von Ecuador. Malakozoologische Blätter, 25: 153- 199, pl. 7-8; n.s., 1: 117-199, pl. 4-15. PROSERPINOID LAND SNAILS 33 NEWTON, R. B., 1891, Systematic list of the British Oligocene and Eocene Mollusca in the British Museum (Natural History). London. 1-365. ORBIGNY, A., 1842, Mollusques, in SAGRA, R., Histoire physique, politique et naturelle de l'Ile de Cuba. Paris, 2: 1-264. PFEIFFER, L., 1848, Monographia heliceorum viventium. 1: i-xxxil, 1-484. Leipsig. PFEIFFER, L., in MARTINI & CHEMNITZ, 1850, Systematisches Conchyliencabinet, Die Familie Heliceen. 12: 1-524, pl. 1-161. PFEIFFER, L., 1853, Monographia heliceorum viventium. 3: i-vii, 1-711. Leipsig. PFEIFFER, L., 1856, Descriptions of twenty-seven new species of land snails collected by M. Salle in the State of Veracruz, Mexico. Proceedings of the Zoological Society of London, 24: 318-324, pl. 35. PFEIFFER, L., 1876, Monographia pneumono- pomorum viventium. 4: i-x, 1-479. PILSBRY, H. A. & A. BROWN, 1910, The Mollusca of Mandeville, Jamaica, and its environs. Pro- ceedings of the Academy of Natural Sciences of Philadelphia, 62: 510-535. SIMROTH, H., 1896-1907, Н. G. BRONN's Klas- sen und Ordnungen des Tier-Reichs, 3(2). Mol- lusca: Prosobranchia, 1-1056. SIMROTH, H., 1910, Some remarks with regard to Professor Bourne's monograph on the Neritidae. Proceedings of the Zoological Society of Lon- don, 1910: 27-33. SOLEM, A., 1954, Notes on Mexican mollusks. |. Durango, Coahuila and Tamaulipas, with de- scriptions of two new Humboldtiana. Nautilus, 68: 3-10. SOWERBY, С. B., Il, 1839, A conchological man- ual. London, i-v, 1-130. STREBEL, H., 1873, Beitrag zur Kenntniss der Fauna Mexikanischer Land- und Susswasser Conchylien. 1: 1-69, 7 pl. Hamburg. SYKES, Е. R., 1900, On Despoena стпапотеа, п. sp., and type of a new subgenus, Cherso- despoena, with notes on some allied forms. Proceedings of the Malacological Society of London, 4: 136-138. TAYLOR, D. W. & М.Р. ЗОНЕ, 1962, An outline of gastropod classification. Malacologia, 1: 7-32. THIELE, J., 1902, Die systematische Stellung der Solenogastren und die Phylogenie der Mol- lusken. Zeitschrift für wissenschaftliche Zoologie, 72: 249-466. THIELE, J., 1910, Uber die Anatomie von Hydo- cena cattoroensis Pf. Abhandlungen der Senckenbergischen Naturforschenden Gesell- schaft, 32: 351-358. THIELE, J., 1927, Uber die Gattung Ceratodiscus. Archiv für Molluskenkunde, 59: 155-157. THIELE, J., 1931, Handbuch der systematischen Weichtierkunde: 1: i-vii, 1-778, Stuttgart. THOMPSON, F. G., 1967, A new land snail of the family Proserpinidae from Chiapas, Mexico (Gastropoda: Prosobranchia). Proceedings of the Biological Society of Washington, 80: 61-64. WENZ, W., 1938-1944, Handbuch der Paläo- zoologie: Band 6—Gastropoda, Teil 1: i-xii, 1- 639. Berlin. o o « a oe > a | ра x Rite o Syne Es er A AS. Td 5 : er Dr A Sra” US. pris (met et per MMA u er № oe ate atin on SALER. : TN NOTA se | Ba Al Sc AS В a u Arte (fe Mes AT ter Я $ aa abr: “Ha ae + ¿de LADA i A DENON Me nz BCE A e e A rana Mery gta E Wi ACI À md Аи VOS. A MANMORA = M UPLATE Hi 29540002 rt №, Г “brut réel yy LE клей ВЛ (SAAD Ta ra san AN FA Ae ПЫЛ Г i= ré bi RAP is mo erp eget: + ©. prosa Fig Mw Amón ema? Ore? 4 las Е со iS APA Se MA a4 К ME 408 (rotor TAS № 2e 240 en a ow ceed mul | | AA Me ee I ze a AV ji ods ve. Aria al Corde A, иле Sort DA ull or toseetiskel be ARS a E ; . een 45 a Mos Sa PAT hal. © ih ‹ Е rt cl Serpe? CG №) POS on dl hee an За 1 “i DB ZU > Hal y Are Ре | > : ie 200 of MALACOLOGIA, 1980, 20(1): 35-62 THE ABUNDANCE, DISTRIBUTION AND DIVERSITY OF THE MOLLUSCS OF WESTERN PORT, VICTORIA, AUSTRALIA! N. Coleman? and W. Cuff3 ABSTRACT The mollusc fauna of Western Port, Victoria, Australia (145°20'E; 38°20'S) is described. Between 1973 and 1975, samples were taken according to a stratified random sample design and were used to give regional and bay-wide estimates of mollusc distribution, abundance and diversity. Faunal affinity analysis, using both presence-absence and abundance data, was used to define major faunal assemblages. Fauna-sediment associations were investigated using non-parametric rank correlation tests. More than 1800 individuals belonging to 96 species were collected. Most individuals and species were bivalves. Gastropods were next in abundance, then polyplacophorans and cephalopods. Species providing 2% or more of the individuals were considered dominant. The fourteen dominant species (1 gastropod, 13 bivalves) together provided 82.4% of the individu- als collected. Affinity analysis showed two major faunal assemblages the distributions of which are related to sediment type. Several stations, all situated near the boundaries of different sediment types, showed a particularly wide range of affinities but were assigned to one or other of the assem- blages according to their strongest affinities. One assemblage is associated with medium to coarse, mud-free sand and is distributed mainly in the deeper (>5.5 m) channel areas. Dominant species are the suspension-feeding gastropod Sigapatella calyptraeformis and the suspension-feeding bivalves Lissarca rubricata, Neotrigonia margaritacea, Venericardia bimaculata, Lepton frenchiensis, Notocallista diemen- ensis and Gomphina undulosa. The surface deposit-feeding bivalve Tellina mariae also ap- peared as a dominant because of its abundance at one station in the assemblage. The other assemblage is associated with fine sand and mud and is found in shallow sublittoral (<5.5 m) and intertidal areas. Dominant species are the suspension-feeding bivalves Micro- mytilus francisensis, Cyclopecten favus, Lepton frenchiensis, Mysella donaciformis and Katelysia rhytiphora, and the surface deposit-feeding bivalves Pronucula concentrica, Tellina deltoidalis and Tellina mariae. Although Lepton frenchiensis and Tellina mariae were dominant in both assemblages, they were most widespread in and characteristic of this ‘fine sand and mud’ assemblage. The mollusc fauna of the coarser, mud-free sediments has a greater number of species, higher species diversity and, particularly amongst the epifauna, a greater abundance of suspen- sion-feeders than that of the fine and muddy sediments. INTRODUCTION Victoria has a rich mollusc fauna (listed in Macpherson & Gabriel, 1962). Nevertheless, there is little detailed information on the abun- dance and distribution of species. Qualitative surveys have been made in Port Phillip (Burn, 1966; Macpherson, 1966; King et al., 1971) and Western Port (Smith, 1971; Watson, 1971), two bays adjacent to Melbourne, but previous quantitative data are restricted to one bay-wide survey of Port Phillip (Poore & Rainer, 1974) and one of a small area of Western Port (Coleman, 1976). During 1973- 1974 a survey of the macrobenthos of West- ern Port was undertaken as part of an envi- ronmental study (Ministry for Conservation, 1975; Coleman et al., 1978; Cuff & Coleman, 1979). The survey included the greater part of the bay, and in 1975 samples were taken in those areas not sampled during 1973-1974. This paper describes the abundance, distribu- tion and sediment preferences and diversity of the molluscan fauna of Western Port as shown by the samples taken between 1973 and 1975. \Рарег No. 216 in the Ministry for Conservation, Environmental Studies Series 2Marine Science Laboratories, P.O. Box 114, Queenscliff, Victoria 3225, Australia 3Division of Computing Research, CSIRO, P.O. Box 1800, Canberra, A.C.T. 2601, Australia 36 COLEMAN AND CUFF METHODS Survey area Western Port (Fig. 1) is a marine embay- ment about 60 km SE of Melbourne. Its total area is approximately 1450 km”, but because of French and Phillip Islands the water sur- face area is only 680 km: of which about 40% (270 km”) 15 intertidal. The bay is morphologi- cally complex and includes salt marsh, man- groves, beaches, intertidal rock platforms, tidal flats, subtidal rocky areas, subtidal em- bayment plains and tidal channel systems (Marsden & Майен, 1975; Smith et al., 1975). The sediment of the channels of North and East Arms is predominantly medium to coarse sand4 with less than 5% mud. These sedi- ments are moderately well sorted4 but sorting becomes poorer with greater mean grain size. The sediment of the shallow sublittoral areas and of the intertidal flats is fine sand and mud. Mean grain size is smaller than for the chan- nel sediments and sorting is poorer (Marsden 8 Mallet, 1975). Seagrasses and algae are found through- out the bay. The more exposed areas of the tidal flats support a vegetation chiefly of the AUSTRALIA on 5 10 Km СТМ 0-5-5т 225 Intertidal Western Entrance = $ Bass Strait seagrass Zostera muelleri and Heterozostera tasmanica. On the muddy channel edges and in shallow sublittoral areas the vegetation 1$ mostly a mixture of H. tasmanica and the green alga Caulerpa cactoides plus smaller quantities of other green, brown and red algae. In the deep portions of North and East Arms, turbidity reduces light penetration and plant life is virtually absent (Ministry for Con- servation, 1975; Smith et al., 1975). Temperature, salinity and dissolved oxygen in the bay were recorded during 1973-1974 (Ministry for Conservation, 1975). An annual range of 10-22°С was recorded from the water in the shallowest areas of the bay, to the northeast of French Island. The range was less in areas of deeper water, being 13-20°С in Bass Strait. Annual temperatures recorded from sediments exposed to the air at low tide ranged from 8-27°С at the surface and from 9-29°C at а depth of about 10 cm. Although a number of rivers and streams enter the bay, the effects of freshwater influx were only local. Salinity varied between 30 and 38°/.. and was highest in the late summer. The waters of the bay were generally about 100% saturated with dissolved oxygen, but supersaturation (to 195%) occurred where photosynthesis by Tooradin Hastings 0)... Sa O A FIG. 1. Location and General Features of Western Port, Victoria, Australia. 4See section on sediment analysis for definition of terms. MOLLUSCS OF WESTERN PORT 37 seagrass was vigorous, and depletion (to 80%) occurred in areas where there was de- composing seagrass. Survey design The survey design was that of stratified random sample with approximately propor- tional allocation of samples to strata. Such a design is easy to implement; allows unbiased estimates of species richness (number of species present), abundance (number of indi- viduals of each species) and distribution; and provides both regional and universe esti- mates (Wadley, 1952; Southwood, 1966; Weber, 1973; Cuff & Coleman, 1979). A map of Western Port was divided into three areas corresponding to intertidal, shal- low (<5.5 m) and deeper (>5.5 m) sublittoral areas. These areas were subdivided, on the basis of sediment type, as then known (from aerial photographs and from local boatmen), and location within the bay, into eleven strata (Fig. 2). The area of each stratum was esti- mated and each was allocated a number of stations approximately proportional to its area. Fifty stations, numbered 1701-1750 in accordance with the Marine Studies Group numbering system, were selected, using a random number table to generate grid co- ordinates, and were allocated to strata as shown in Fig. 2. Stratum 4 was not sampled because physical condition (shallowness, heavy swell, surf breaking on shallow banks) in this region made sampling impossible. Stratum 11 was not sampled because propor- tional allocation gave it much less than one station. The numbers of stations allocated to Strata 1, 2, 3, 5, 6, 7, 8, 9 and 10 were 8, 4, 8, 6, 2, 2, 2, 1 and 17 respectively. Collection and treatment of samples All stations were sampled during the sum- mer. Subtidal stations were sampled from a Y ( ma FE | Е — | Se = A EA | ее] ‘À 1709 Ags D > 1703 se) | 1710 DI CH 7 == | ИИ 17067 Ae | Ki = C7170 РЕ ЗЕНИТ 1 < 4 22 1 MEE, e ESE о 5 10 km Yan: $ a £2 =; 1 ne ee | 1711 A) ( #1 2 A É Ye (1 \ eee 701 N ® 1701 Station ES É 7944 E ¡ Hastings 0 )- 1712 yy 1705 K |] 1-11 Strata —, A) И они) 1713 4 4 TL Sr 1725: EVE = DAR French Island er / = 2 R + 1744 ® №1745 - 47 + E ER | ER ++++++++ Phillip Island pi Ss + SÍ Fes 0 1746 < а г вл ЧЕ + + + р Е ZN o 9 FIG. 2. Strata defined and stations sampled т Western Port. 38 COLEMAN AND CUFF research vessel which carries radar capable of locating sample sites to within +30 т. Sta- tions 1706 and 1721-1741 were sampled in November 1973. The Western Entrance sta- tions, 1742-1750, were sampled in November 1975: The intertidal stations, 1701-1705 and 1707-1720, were sampled in January 1974. Sampling was carried out at high tide from a shallow-draught barge. The barge was т radio contact with the research vessel and was positioned by use of the latter's radar. At each station three samples were taken with a 0.1 m? Smith-Melntyre benthic grab. About 50 а of the collected sediment was ге- moved for granulometric analysis and the re- mainder was washed through a sieve with an aperture of 1.003 тт (mesh no. 16 to BS410:1962). The material retained on the sieve was preserved with 5% neutral formalin and taken to the laboratory. In the laboratory, samples were placed in shallow plastic trays, sorted and the animals removed. The portions composed of small particles were placed under an illuminated magnifier or stereomicroscope for sorting. The specimens removed from the samples were counted and identified. Where there was doubt as to whether a mollusc was collected alive or dead the specimen was identified and the shell was then crushed and the presence or absence of the animal determined. Analysis of data For the analyses described in this paper, the three grab samples per station were lumped. Total numbers of individuals and species per station were determined, and species were assigned to feeding types on the basis of information in the literature (Mor- ton, 1958; Macpherson & Gabriel, 1962; Poore & Rainer, 1974). The Shannon-Wiener diversity index, Н’, was calculated as 5 259.10 р; ¡=1 H’ = where p; is the proportion of individuals rep- resented by the ith of s species. Evenness, J’, was calculated as H In s where Н’ is the Shannon-Wiener diversity in- dex and s is the number of species. Using binary (i.e. presence-absence) data, faunal affinity between stations was calcu- lated by comparing all possible pairs of sta- tions according to Czekanowski's formula (Czekanowski, 1913; Clifford & Williams, 1976) 2a/(2a + D+ €) where а is the number of species common to both stations, b is the number of species present at the first station only and с 15 the number of species present at the second sta- tion only. Faunal affinity values were converted to dissimilarity (1-formula) and stations fused using a flexible sorting strategy with B = —0.25 (Lance & Williams, 1967; Clifford & Stephenson, 1975). Using abundance data, faunal affinity was calculated by the Canberra Metric Dissimilar- ity Measure 1 Xi = Хит Хк + Хк where m 15 the number of species in the sam- ples being compared, X, is the number of in- dividuals of the kth species in the ¡th station and Xj, similarly for the jth station. Stations were fused using a flexible sorting strategy with В = —0.25 (Lance 4 Williams, 1967; Clif- ford & Stephenson, 1975). Affinity analysis indicated two faunal as- semblages. Average values for the sediment and faunal characteristics of these assem- blages were tested for significant differences (p < 0.05) by t tests. Where variances were not initially homogeneous, tested by the Fa, test (Sokal & Rohlf, 1969), logarithmic trans- formations made them so and the trans- formed data were used. Tests between assemblages are only ap- proximate as the data for an assemblage do not represent a simple random sample from the population of samples per assemblage. Statistically valid tests of differences between strata were made using one-way analysis of variance tests (Sokal & Rohlf, 1969). Vari- ances were tested for homogeneity by Bart- lett's test (Sokal & Rohlf, 1969) and, where necessary, logarithmic transformation was used. Fauna-sediment associations were investi- gated by non-parametric tests. The Kendall rank correlation coefficient, т (tau) (Siegel, 1956), was calculated for associations be- tween fauna and sediment mean grain size, mud content, sorting and skewness. Where MOLLUSCS OF WESTERN PORT associations with two of the investigated sediment characteristics were significant (p < 0.05), Kendall partial rank correlations, Txy-z were calculated. The partial rank correla- tions cannot be tested for significance, but the magnitude of changes in the rank correlation coefficients when the effects of particular factors are removed gives some indication of the importance of these factors (Siegel, 1956). Sediment analysis Sediments were analysed at the University of Melbourne (Marsden & Mallett, 1974, 1975). Grain size was measured on the ф scale ($ —loge diameter in mm, of sedi- ment particle). Values for ф increase as grain size decreases, thus coarse sand = 0-1 $, medium sand = 1-2 d, fine sand = 2-3 6, very fine sand = 3-4 ф, silt = 4-8 ф and clay = 8-12 $. For each sample the sand and mud (= silt + clay) fractions were separated by wet siev- ing. Size analysis of the sand was by sieving. Pipette analysis of the mud was used to dif- ferentiate silt and clay. Size analysis data were plotted as cumulative per cent by weight against grain size. Percentile values were ob- tained and used to calculate mean grain size ((616 + 650 + 84)/3); sorting ((b84 — &16)/4 + (6/95 — 5)/6.6)), values of which increase as sorting, a measure of clustering of grain size around the mean, becomes poorer; and skewness ((ф16 + 684 — 2650)/2(684 — $16) + (65 +,695 — 2650)/2(695 — &5)) (Folk, 1968). RESULTS Bay-wide estimates of mollusc abun- dance, diversity and distribution Eighteen hundred and forty-three individu- als belonging to 95 species were collected (Table 1). Bivalves predominated both as in- dividuals and as species. Next in abundance were gastropods which, although not highly abundant in terms of individuals, did provide a relatively high proportion of the species col- lected. Chitons and cephalopods were poorly represented in the samples and no scaph- opods were found. The number of individuals per station (Fig. 3; Table 2) ranged from 1 to 208 with a mean of 36.9. Species per station ranged from 1 to 22 with a mean of 5.8. Species diversity ranged from 0 (at stations with only 1 species) to 2.49 with a mean of 1.01. Evenness ranged from 0.39 to 1.00 with a mean of 0.72. Most of the individuals and species col- lected were suspension-feeders (Fig. 3; Table 2). Surface deposit-feeders were also com- TABLE 1. Numbers of individuals and species of molluscs collected in the Western Pont survey. No. of % of all indi- indi- No. of % of all Class viduals viduals species species Polyplacophora 33 1.8 5 5.2 Gastropoda 198 10.7 40 42.1 Bivalvia 1606 87.1 50 51.6 Cephalopoda 6 08 1 leat TABLE 2. Values (mean + SD per station) for sediment and mollusc fauna characteristics bay-wide and in the molluscan assemblages recognised in Western Port. стза = clean medium sand assemblage; fsma = fine sand and mud assemblage. Asterisks denote significant differences (p < 0.05) between assemblages. Mean + SD/station for: Bay-wide cmsa fsma Probability Mean grain size, &' 1:47 = 057 PAT E 1.24 <0.01* % Mud in substratum 2.97 + 14.93 40.95 + 37.64 <0.01* Sorting 0.89 + 0.54 1.25 + 0.71 0.08 No. of individuals 36.9 + 44.3 38.0 + 42.7 35.9 + 46.6 0.16 No. of species 58 + 4.9 (One 56 E 3.5 0.01* Species diversity, Н’ 1.01 = 0.66 ESTE 072 0.81 = 0.53 0.02* Evenness, J' 0.70 = 0.19 0:71 = 023 0:69== 0:15 0.12 No. of infaunal suspension-feeders 15:9) == 20.3 20.2 + 20.4 11.8 =.196 0.14 No. of epifaunal suspension-feeders 5.3. = 16.6 957 = 21.8 Зы 9.2 0.02* Total No. of suspension-feeders 224, 27.1 29.9 + 30.6 149 = 21.5 0.05* No. of surface deposit-feeders 122, =267, 4.7 + 18.7 18: CE 36 0.07 No. of grazers 2. 29:6 PISE STO 1:9 = A0 0.28 No. of predators 0.8 = 1.7 0:8. =.16 0.8 + 18 0.09 No. of scavengers 0.2. = 07 0.3 08 Ost 0:4 0.20 1ф values increase as grain size decrease COLEMAN AND СУЕЕ a Others (E Intauna Surface deposit ` [Y Epitauna mon, 01 10 200 individuals station 5 7 PT 27 РЯ O S D $ iA ' = —— NT D PEER pl ll = > ER y => = ВЕ РЕ: Y == 2 E Scavenge: ( Predators e ЕЙ = == = — m. => \ ==} Е: =: TS Others 2 Grazers [) But = | a = == == E) dE + ЕЕ 29 al fal x = = SA —1 01 10 200 SS = = Е = = HH Y en Y individuals / station — ( A FIG. 3. Number of mollusc individuals and proportion of different feeding types at the sample stations. Circle diameter indicates total number of individuals per station, segment sizes indicate proportions of different feeding types. Clear areas are predominantly deep (>5.5 m) areas with a sediment of medium to coarse sand and little mud (<5%). Hatched areas are predominantly shallow (<5.5 m) or intertidal with a sediment of fine sand and mud. Fig. 3a. Infauna = infaunal suspension feeders; Epifauna = epifaunal suspension feeders; Surface deposit = surface deposit feeders; Others = molluscs of other feeding types. Fig. 3b. Predators = predatory molluscs; Grazers = grazing molluscs; Scavengers = scavenging mol- luscs; Others = molluscs of other feeding types. MOLLUSCS OF WESTERN PORT 41 mon, providing about a third of the individuals collected but only 3% of species. Grazers, predators and scavengers together account- ed for only 8% of individuals but for 48% of species. Species which provided 2% or more of the individuals collected were considered domi- nant (following Bandy, 1958). On a bay-wide basis 14 (14.6%) of the 96 species were dominant, and provided 82.4% of the individ- uals. Most of the dominant species and indi- viduals were suspension-feeders, but the sin- gle most abundantly collected species was the surface deposit-feeder Tellina mariae which contributed almost a fifth of the mol- luscs collected (Table 3). Some dominant species (e.g. Micromytilus francisensis, Gomphina undulosa—Fig. 4) occurred at only a few stations, owing their dominance to their relatively high numbers at these stations. Other dominant species were more widespread, though they occurred less abundantly at any one station (Figs. 4-8; Table 3). Faunal affinity between stations A trellis diagram of faunal affinity calculated from presence-absence data shows two ma- jor groups of stations (Fig. 9). Group 1 con- tains all the stations from strata 2, 7 and 9, most from stratum 1, half from stratum 3 and one each from strata 5 and 10. These stations are in the deeper (>5.5 m) channel areas or immediately adjacent shallows where the sediment is largely medium sand (Tables 2, 5). This station group and its associated mollusc fauna has therefore been character- ised as a clean medium sand assemblage' (cmsa). Group 2 contains all the stations from strata 6 and 8, most from strata 5 and 10 and one from stratum 1. These stations are mostly from shallow sublittoral (<5.5 m) and inter- tidal areas of fine sand and mud and they, and their molluscan fauna, are characterised as a fine sand and mud assemblage’ (fsma). A few stations (e.g. 1708, 1736, 1750), situated near the boundaries between sedi- ment types, show a particularly wide range of affinities but are placed in one or other of the major groups according to their strongest af- finities. Stations 1747, 1727, 1742, 1707, 1739, 174317451708 and 1713 (Fig: 9) show a restricted range of affinities and do not appear as an integral рай of either major group. The affinities of 1747, 1742, 1707, 1743 and 1745 are highest with those stations forming the cmsa (Fig. 9, Group 1), and are considered as part of that assemblage. Sta- tions 1727, 1739, 1703 and 1713 show their TABLE 3. Dominant mollusc species in the Western Port benthic survey. ES = epifaunal suspension feeder; IS = infaunal suspension feeder; SD = surface deposit feeder; cmsa = clean medium sand assemblage: fsma = fine sand and mud assemblage; c = species characteristic of cmsa; f = species characteristic of fsma. No. of stations at which No. of individuals as % of: found as % of: all mol- molluscs molluscs all stations stations Feeding lusc indi- in in stations in in Species type viduals cmsa fsma sampled cmsa fsma Sigapatella calyptraeformis c ES 4.0 TES 0.2 30 58 3 Lissarca rubricata € ES 6.4 12.7 0.2 14 25 3 Neotrigonia margaritacea с IS 2.9 518 0 24 50 0 Venericardia bimaculata c IS 3.2 6.4 0 22 46 0 Notocallista diemenensis с IS 6.9 13.8 0.1 20 З7 5 3.8 Gomphina undulosa € IS 3.4 6.9 0 6 1225 0 Solen vaginoides с IS 4.4 9.0 0 18 37.5 0 Pronucula concentrica f SD 4.9 162 8.6 22 16.6 26.9 Micromytilus francisensis f ES 2.7 0 5.4 6 0 ales Cyclopecten favus f ES (1.5) 0.3 2.6 10 4.2 15.4 Lepton frenchiensis f IS 6.0 5.1 5.9 18 4.2 30.8 Mysella donaciformis f IS 4.8 0.4 9.0 18 4.2 30.8 Katelysia rhytiphora f IS 5.5 1.0 9.9 24 8.3 38:5 Tellina deltoidalis f SD 8.0 0 15.9 14 0 26.9 Tellina mariae f SD 19.3 9.3 29.1 32 8.3 53.9 42 COLEMAN AND CUFF Micromytilus Lissarca Gomphina o 1 10 100 individuals/station FIG. 4. Distribution of the dominant mollusc species Micromytilus francisensis, Lissarca rubricata and Gomphina undulosa at the Western Port sample stations. Circle size indicates the total number of indi- viduals of these species and segment size indicates the proportion of each. (Species have been grouped together on the basis that they may be conveniently ploited on the same figure.) Tellina mariae Venericardia x y Solen 0 1 10 00 individuals/ station FIG. 5. Distribution of Tellina mariae, Venericardia bimaculata and Solen vaginoides in Western Port (see also caption to Fig.: 4). MOLLUSCS OF WESTERN PORT 43 Tellina A deltoidalis L Neotrigonia р Katelysia DA 10 100 individuals /station FIG. 6. Distribution of Tellina deltoidalis, Neotrigonia margaritacea and Katelysia rhytiphora in Western Port (see also caption to Fig. 4). roo || | || Fr Pronucula (A Sigapatella = Lepton zea 10 100 individuals / station Ese UU, e © de FIG. 7. Distribution of Pronucula concentrica, Sigapatella calyptraeformis and Lepton frenchiensis in West ern Port (see also caption to Fig. 4). 44 COLEMAN AND CUFF Mysella » Notocallısta — 0 1 10 100 individuals/station FIG. 8. Distribution of Mysella donaciformis and Notocallista diemenensis in Western Port (see also caption to Fig. 4). strongest affinities with the fsma (Group 2, Fig. 9) and are considered as part of that as- semblage. Dendrograms calculated from binary (Fig. 10) and from aburidance (Fig. 11) data show essentially the same patterns of station affinity and the same associations between station groups and sediment type. The correspond- ence between dendrograms is particularly marked at lower levels of dissimilarity. At a dissimilarity of 1.2 the binary data dendrogram shows seven groups (Fig. 10 A- G). Groups A-D contain the stations from strata 1, 2, 3, 7, 9 and 10 which are grouped in the trellis diagram as the cmsa (Fig. 9, Group 1). Groups E-G contain the stations from strata 5, 6, 8 and 10 grouped in the trellis diagram as the fsma (Fig. 9, Group 2). At a dissimilarity of 1.0 the dendrogram from abundance data shows six groups (Fig. 11, A-D, F-G) similar to those seen at a dis- similarity of 1.2 in the binary data dendro- gram. Groups B in the two dendrograms cor- respond exactly; groups F and G differ by only one and three stations respectively; and the correspondence between the remaining groups, although not as good as for В, Е, and G is still close. Stations 1743 and 1745 which form the smallest group, B, in both dendrograms each contain a single species (Gomphina undulosa) which is found at only one other station (1742). At higher levels of dissimilarity a major dif- ference between the dendrograms is in the separation, in the binary data dendrogram (Fig. 10) of Group A from the remaining groups. This separation is probably a result of group-size dependency (Williams, Clifford & Lance, 1971) and not an indication that group A stations constitute a separate assemblage. Comparison of the trellis diagram with the dendrograms suggests some transference of stations between assemblages. In both dendrograms stations 1703, 1711, 1713, 1714, 1715, 1727 and 1729, appear most closely related to the cmsa, but from the trellis diagram they are considered as part of the fsma. Most of the individuals at these trans- posed stations belonged either to rare spe- cies (which are not characteristic of either assemblage) or were species characteristic of the fsma (viz. Lepton frenchiensis, Mysella donaciformis, Katelysia rhytiphora, Tellina mariae). The trellis diagram therefore seems to give the most appropriate placing of these MOLLUSCS OF WESTERN PORT 45 % CAIN GROUP 1 ST ZUR SHS Nill NE N GROUP 2 UGG N= РАДУ | | FZ 37 40 5 a FIG. 9. Trellis diagram of faunal affinity between stations calculated from presence-absence data using the Czekanowski coefficient. Str = stratum number; Sta stations and for comparative analysis stations were grouped into assemblages as indicated by the trellis diagram. Comparison of faunal assemblages Tests for significant differences between assemblages (Table 2) are only approximate. They indicate that the sediment of the stations in the cmsa 1$ significantly coarser and less muddy than at the fsma stations. Sorting ap- pears poorer in the fsma but the difference between assemblages is not significant at the 5% level. The average number of species and aver- age species diversity per station were signifi- cantly greater in the cmsa, but differences in = station number. evenness were not significant. Differences in ihe number of species per station are also reflected in the total numbers of species per assemblage (Table 4). Only 15 of the 95 spe- cies were common to both assemblages. Of the remainder, 64 (67%) were exclusive to the cmsa. Total H' and J’, calculated for all indi- viduals and species per assemblage, were 3.16 and 0.72 respectively for the cmsa and 2.26 and 0.66 for the fsma. Most (79%) individuals in the cmsa were suspension-feeders. The average number of infaunal suspension-feeders per station did not differ significantly between assemblages, but average values for epifaunal and for total suspension-feeders per station did. Most (51%) individuals in the fsma were COLEMAN AND CUFF 46 ‘UMOUS эле O)-y Sdnoj6 иоце}$ ay) 10} PNW % pue 'p ‘azis шелб ueaw JO} senjeA эбелэлу ‘эбеашеззе pnw pue pues эцц ay] о} WEIBEIP эл ay) зо цоцоэд$и! wos paubisse SUONEIS ay) jo asoy] эле 5лэашпи рашиэрий зиэющеэо9 IHSMOUPY9ZO ay, Buisn вер aduasqe-aduasald LOI} рауепоеэ SUONEIS иээм}э4 Аиище jeune] jo шелболриэа ‘OL ‘914 2:19 С. vr 9-6 9.91 sz 0 0.6 pnw 9 15.5 60.5 И.С BrL Sel alt 2S 1 в GES Tamer 5 ET 113 а ME) lg "y 91 Ob S ZI IE LE №85 9 8 02 BL 6 zt 6 L à bt SL OL Ц 65 65 L th УФ € El 56 6b Lb Ob Bb EZ ZE OS LZ LE $$ SP 96 BZ HE OE SE 9% 22 SZ bz CRA O | N © © © яа-- - + gs NO © © s# N © ы Aysejiwissig 47 MOLLUSCS OF WESTERN PORT ‘UMOUS эле ©—\у sdnoJB иоце}з ay) 10} “pnu % pue ‘ф ‘azis ules6 иеэш 10} зэпел эбелэлу ‘эбе|ашэззе pnw pue pues aulj ay) о} шелбер sia] ayy jo иоцоэд$и! wo paubisse зиоцезз ay) jo э$04} эле $лэащпи рэцшерий ‘aunseen Auejwissıg WOW еллэдие) eu} Buisn eyep ээцерипае шоц pajejno¡eo suone]s uaamjag Апище jeune, jo шеболриэа “Ll “DI 69 с.55 8-41 0-1 Ze O pnw 9% 9.5 v8-7 64-1 OL-L bb-1 aL:t в O 773 a à) Ey ge LE лк © 9 8 эст z St 6 6 PLZ2ZSI LL EL € BE Ф ZL Zr GE OC bb 65 6$ OS LP Ob Bb 96 EZ ZE SE Ze BZ УЕ CE LE pa 95 SZ Erst e à = z- $ ® = N} v- = gnc =) o 8. Or N Aytuejiwissig + Su 48 COLEMAN AND CUFF TABLE 4. Distribution of mollusc species between the molluscan assemblages recognised in the Western Port survey. The first figure indicates the total number or percentage of species in the assemblage. The figure in parentheses indicates the number or percentage exclusive to that assemblage. No. of species found in clean medium sand % of species found in clean medium sand No. of species found in fine sand and mud % of species found in fine sand and mud assemblage assemblage assemblage assemblage Polyplacophora Sr a(S) 6.3 (6.3) 0 (0) ORO) Gastropoda 30 (27) 38 (34.2) 13 (10) 41.9 (32.3) Bivalvia— Infaunal suspension feeders 27 (22) 34.2 (27.9) 9 (4) 29.0 (12.9) Epifaunal suspension feeders 13 (10) 16.5 (12.7) 5) (2) 16.1 (6.5) Deposit feeders 3 (0) 3.8 (0) 3 (0) 9.7 (0) Cephalopoda 1 (0) 183; (0) 1 (0) 3.2 (0) Total 79 (64) 100 (81.1) 311616) 100 (51.7) surface deposit-feeders, but they did not dom- inate the fsma to the extent that suspension- feeders dominated the csma. The difference between assemblages in the average values for surface deposit-feeders at a station ap- proached but was not significant at the 5% level. In both assemblages, most species were suspension-feeders, but the number exclu- sive to the cmsa was five times that exclusive to the fsma (Table 4). The number of species of surface deposit-feeders was the same in both assemblages although they are propor- tionately better represented in the fsma be- cause of the smaller number of species in that assemblage. With one exception the species dominant in any one assemblage are also those which are dominant bay-wide (Table 3). Cyclopecten favus was the only species dominant in an assemblage (fsma) but not constituting 2% or more of total individuals collected. Under the 2% criterion, Lepton frenchiensis and Tellina тапае are dominant in both assemblages, but are best considered as characteristic of the fsma. Both occurred at a greater propor- tion of the stations in the fsma; Tellina was also much more abundant in this assemblage although Lepton was almost equally abun- dant in both the cmsa and the fsma. Stratum analysis Analysis of variance (Table 5) indicated significant differences between strata in mean grain size and mud content, but these differ- ences were not located. The only statistically significant faunal differences found were in the numbers of predators and scavengers. Because these feeding types occurred at very few stations, the importance of these differ- ences 15 uncertain and no attempt was made to locate them. Other statistically significant differences were not found, but the average numbers of species, species diversity and suspension- feeding individuals at a station are higher in those strata (1, 2, 3, 7, 9) which together form the cmsa. The strata which form the cmsa are also those with the lowest average mud con- tent and, in general, the coarsest mean grain size (Table 5). Stratum analysis shows that the occurrence of surface deposit-feeders in the cmsa is largely due to their high incidence in stratum 9. This stratum contained only one station, 1750, which was in a shallow sublittoral area and has a wide range of affinities with stations in both assemblages. In the other strata which form the cmsa, the stations of which are in deeper water and have affinities chiefly amongst themselves, surface-deposit feeders are nearly or entirely absent. Stratum analysis also Suggests the number of epifaunal sus- pension-feeders to be greatest in the East Arm and the Western Entrance (strata 2, 3 and 5). Mollusc-sediment associations The mollusc-sediment associations investi- gated are shown in Table 6. Because mean grain size was measured in ф units, which increase as grain size decreases, a negative correlation between mean ф and faunal at- tributes indicate that the attribute increases as grain size increases. The number of species decreased signifi- 49 MOLLUSCS OF WESTERN PORT e nö hu 0 v 0 0 0 80 + 50 L0 = v0 GO + 50 0 злэбиэлеэ$ JO ‘ON, er + CE 8 0 РЕ = 96 Le + OL 0 6€ + 05 G + 80 pl + 60 siezes6 JO ‘ON ges Al | 10 + GO 0 0 80 + 50 yl + 60 cl + Ot er = Ol sıoyepaud jo ‘ON, Z9OL+ 9'et 16 cle+ 08l 0 c 80L+ S92 prl+ Blt 5'9 = 9¢ 0 0 $лэрэз; изодэр aoeuns JO ‘ON 681+ 9ct (972 vl = OC v8l+ Ove #96 + 085 vipe ale 65Е= Kor 66 = 50 66-641 51эрэз} uoisuadsns jo лэашпи |EJOL 791+ 78 69 260 + Er L6lL+ Gee 796 + 08 $7'05= 5'0е 2'11= #06 96 + Bel ZOc+ 691 Siapaa, uoisuad -sns jeunejul JO “ON |= CV v 0 ¿O + 9+1 0 vel+ 89 Gve+ 661 901+ G9 SEO s19p99} uoisuadsns ¡eunejida jo ‘ON ЕО +690 ¿90 ЕО +FEZO 900 +660 600 +980 6LO +€90 SEO +690 850 +E90 E20 +180 „Г SSSUU8AT 890 +S80 961 GE O +180 LEO + OL r90 =+0!L'L 07'0 *290 680 +171 90 + 9t 0 +91 1 4H ‘Айзлэлр Sal9ads У + ly 81 Le + GE L'e + Gl ge + Sp 61 + €€ Sl + 001 $ + G9 81 + OP saioads jo ‘ON G'I£+ G6c ДЕ £'€c+ SOC GOCH S Lc Gtl+ 7901 Gbb+ L6E ¿8EF+ 0/9 LOLF ECC L'OC+ 861 SIENPIAIPUI JO ‘ом OLOF=SI E 850 6/0 +OO F201 €80 +111 290 = 81 220 +180 020 +060 9r'O +480 Buiuos cc 9£+869€ Cc60 p8ZS+ 169 86 +9/'/ 20% + viel LS Ov+9Z0S 0/9 720% Loe +€9'1 6t | +€8'0 WINES -qns ul pnW %, 280 +8pc ЗС 191 610 +651 LOL +581 v6'l FILE 220 +951 SHO +rrl 990 +it li Ф ezis шелб ueaW. A AAA AE ES (21)o! (1)6 (2)8 (2)2 (2)9 (9)s (8)e (me (8)! JO} UONEIS K—_ A —_—_—_—a_—<—_—_e————— KK /GS + чеэи wneNS en Br SS EN... 'g wınells ul Aluo uONEIS AU 10} pajejnojes элэм Bunuos pue azis uIeIB ueoyy ‘езедз чеэмеа (s0'0 > 9) ээцалаир зиеоциб!5 эоцар SYSUEISY ‘штед$ Jad suo!]e]s jo 1эашпи ay) эзеорш saseyjuased и! SISQUNN ‘ASAINS 204 Ula]sem ay) ul sonsiajoeJeyo euney OSN||OlW pue JUBWIPES JO UNIES ла sisÁjeuy 'S 319W1 50 COLEMAN AND CUFF TABLE 6. Kendall rank correlations and partial rank correlations between sediment and mollusc fauna characteristics in the Western Port survey. 7 = correlation coefficient; p = probability of correlation being significant (p < 0.05); (7) = partial rank correlation with the effect of the other correlated sediment variable removed; NS = not significant. Mean $ Mean & % Mud T= (0:41 р = = 0.001 Sorting т 0:33 p= 01002 Skewness PE 02 No. of individuals NS No. of species т = —0.19 р = 0.049 Species diversity, H' NS Evenness, J’ NS Total no. of suspension feeders NS No. of infaunal suspension feeders NS No. of epifaunal т = -0.24 suspension feeders р= 0.017 (т = —0.19) No. of surface = 0:33 deposit feeders р.=, 0.002 G = 09) Proportion of T= 0:32 suspension feeders p= 0.003 (z= 0:24) Proportion of infaunal NS suspension feeders Proportion of epifaunal т = =0:43 suspension feeders р = 0.001 (7 = —0.39) Proportion of surface п = 10.38 deposit feeders р == 0.001 (r= 0.23) O “y % Mud Sorting Skewness = , Ose) = 0.001 NS т = —0.31 NS NS NS NS NS NS NS 7.023 NS p= 0.020 NS NS NS NS NS NS NS NS NS NS т= 0.20 NS р = 0.039 (= 031) = 0.43) NS NS = <0.001 = 0.43) = 0:23 NS NS = 0.023 —0.11) = —0.26 NS NS = {0/0110 NS T= 10.23 NS p= 0.021 (r = 0.44) OEA NS NS = <0.001 — 037) cantly with increase in mean &. Species di- versity was positively correlated with poorer sorting. Because finer sediments tend to be more poorly sorted, this relationship between species diversity and sorting suggests an in- crease in species diversity in fine sediments. However, по significant association was found between mean d and H'. The number and proportion of epifaunal suspension-feeders increased with decrease in mean ф and poorer sorting. The greater correlation coefficients and probabilities for the associations with mean ф suggest the major association to be with this character- istic. Partial correlation also indicates the major association to be with mean ф. The number of infaunal suspension-feeders was not significantly associated with the sedi- ment characteristics investigated, although the association with mean ф approached sig- nificance (p = 0.07). The association between the proportion of infaunal suspension-feeders and mean ф just failed to reach significance (p = 0.053), but there was a significant nega- tive correlation with mud. The total number of suspension-feeders was not significantly associated with the sedi- ment characteristics measured, but the pro- MOLLUSCS ОЕ WESTERN PORT 51 о 50 100 070 50 100 Mean © Lot 100 % Mud FIG. 12. Occurrence of the dominant infaunal species of the fsma in relation to sediment mean grain size (Mean $) and mud content (% mud). Circle size indicates abundance; +, 1-9 individuals; €, 10-19 individu- als; O, 20+ individuals. a. Pronucula concentrica; b, Lepton frenchiensis; с, Mysella donaciformis, d, Katelysia rhytiphora; €, Tellina deltoidalis; t, Tellina mariae. 52 COLEMAN AND CUFF portion of suspension-feeders was significant- ly greater in coarser, less muddy sediment. Partial correlation indicated the greater as- sociation to be with mean . The most highly significant associations found were those of the number and propor- tion of surface deposit-feeders, both of which increased significantly with increase in mean b and increase in mud content. Partial corre- lation shows the most significant association to be with mud content. The occurrence of the dominant infaunal species in relation to sediment mean grain size and mud content is illustrated in Figures 12 and 13. Mean © % Mud FIG. 13. Occurrence of the dominant infaunal spe- cies of the стза in relation to sediment mean grain size (Mean $) and mud content (% mud). Circle size indicates abundance; +, 1-9 individuals; O, 10- 19 individuals; 9, 20+ individuals. a, Neotrigonia margaritacea; b, Venericardia bimaculata; с, Notocallista diemenensis; 4, Gomphina undulosa; e, Solen vaginoides. Pronucula concentrica (Fig. 12a) occurred at stations with mud contents ranging from 0-100% and in grain sizes ranging from coarse sand to silt. lt occurred most abun- dantly in medium sand with a moderately high mud content. Lepton frenchiensis (Fig. 12b) occurred mainly in fine sand with up to 96% mud, but also at a few stations of medium, mud-free sand. Mysella donaciformis (Fig. 12c) occurred mainly in fine sand with less than 20% mud. Katelysia rhytiphora (Fig. 12d) lives in sediments ranging from coarse to very fine sands and with mud contents from 0-100%, but was most abundant in fine sedi- ments with 10-40% mud. Tellina mariae (Fig. 12f) occurred in sediments with mean grain sizes from coarse to very fine sand and mud contents ranging from 1-100% and was abundant over the whole range of sediments in which it was found. Tellina deltoidalis (Fig. 12e) was slightly more restricted in its occur- rence than Tellina mariae, occurring most abundantly in fine sediments with a mud con- tent of 10-80%. Neotrigonia margaritacea (Fig. 13a) oc- curred in sediments ranging from coarse to fine sand but only at two stations where the mud content was in excess of 5%. Although mostly found in medium sand, it was found in greatest abundance at a station with a sedi- ment of fine sand. Venericardia bimaculata (Fig. 13b) was taken mainly in sediments with less than 5% mud and with mean grain sizes ranging from coarse to fine sand. Notocallista diemenensis (Fig. 13c) was most abundant at a station with 16% mud but was most fre- quently taken from sediments with no mud and a grain size of medium sand. Gomphina undulosa (Fig. 13d) was found only at stations with medium sand and no mud. Solen vagi- noides (Fig. 13e) occurred mainly at stations with little or no mud and in sands ranging from coarse to fine. Like Neotrigonia, it was most abundant at a station with little mud and a mean ф of fine sand. The dominant infaunal species character- istic of the fsma occurred over a much wider range of sediment types than did those char- acteristic of the cmsa. The lack of dominant fsma species in sediments with 40-70% mud (Fig. 12) does not indicate any discontinuity in distribution. Instead, it reflects the fact that only two stations had a mud content in this range. Mysella donaciformis and Tellina deltoid- alis were both most abundantly collected from MOLLUSCS OF WESTERN PORT 53 station 1701, Katelysia rhytiphora and Pronucula concentrica from station 1718, and Solen vaginoides and Tellina mariae from station 1750. In each of these pairs the species are of different feeding types, one being a suspension and the other a deposit feeder. DISCUSSION Survey design and analysis The survey from which these mollusc data are drawn (Coleman et al., 1978) was carried out as part of a multidisciplinary environ- mental study of Western Port (Ministry of Conservation, 1975; Shapiro 8 Connell, 1975). The purpose of the environmental study was to provide data for use in the man- agement of Western Port, and the benthic survey was required to provide information in accord with this purpose. Because species abundance, distribution and community struc- ture throughout much of the bay had not pre- viously been sampled, a bay-wide view of the fauna was required; and because future de- velopment in the bay may be regional, a re- gional knowledge of the fauna was also needed. These two, somewhat contradictory, requirements are satisfied by the chosen sur- vey design of stratified random sample with proportional allocation of samples to strata. It is important to distinguish clearly between the usefulness to benthic surveys of survey statistical procedures and of affinity analysis. Benthic surveys have at least three important aspects: the provision of unbiased universe estimates (of faunal abundance and diversi- ty); the provision of unbiased regional esti- mates; and the allowance of statistically test- able inter-regional (and possibly also tempo- ral) comparisons. These requirements are met by stratified random sampling. But this design requires an a priori subdivision of the survey area. In the absence of a previous sur- vey, this subdivision should be made on the basis of variables (in the present instance sediment type) which are expected to be cor- related with species distribution (Weber, 1973). Stratum analysis allows accurate regional and universe estimates of species abundance and diversity, and also allows statistically valid inter-regional comparisons. It also re- veals regional variations in faunal character- istics (e.g. that the occurrence of deposit feeders in the cmsa is due mainly to their oc- currence in stratum 9) which may be of signif- icance and yet are not so apparent when the more general view, of faunal assemblages, is taken. The use of faunal affinity analysis and the grouping of stations into assemblages pre- sents a more general view of the fauna; shows the manner in which different regions of the bay relate to each other in terms of faunal similarity; emphasizes fauna-sediment relationships; and suggests differences (e.g. fewer species and lower species diversity in fine sediments) which are not so apparent from stratum analysis. Affinity analysis can also suggest the definition of better strata for future survey work. The similarity in faunal assemblages de- fined, in the present survey, from presence- absence and from quantitative data has im- portant practical implications. It supports the view of Moore (1971) that the collection of presence-absence data, which may be ac- complished more rapidly than the collection of quantitative data, may be the most useful ini- tial approach, especially when surveying areas about which little 15 known. Analysis using presence-absence data also has the advantage that when it is between samples taken at different times it may well be unaffected by seasonal or annual changes in individual abundance. This advantage is rele- vant to the present study. All the stations were sampled during the summer but stations 1742-1750 were sampled a year later than the others. Affinity analysis using presence-absence data will be affected by changes in species composition, but there is evidence to show that changes in species composition are less marked than those in individual abundance. Buchanan et al. (1974, 1978) have shown that in benthic communities off the British coast species composition remains relatively stable but the relative abundances of species may change greatly with time. A similar situation 1$ reported for Port Phillip, a bay adjacent to Western Port (Poore & Rainer, 1979). In the present survey there is no evidence to suggest that temporal variation has signifi- cantly affected the results, even though the Western Entrance stations (1742-1750) were sampled a year after the rest. Faunal affinity analyses using both presence-absence and abundance data gave similar results, and in neither case did the Western Entrance sta- tions show any clear tendency to separate 54 COLEMAN AND CUFF from the rest. Stratum analysis also failed to show differences which could be ascribed to temporal changes in species abundance and diversity. Scope and success of the survey Previous quantitative data from Western Port molluscs have been available only for a small area around Crib Point, in North Arm, which was sampled intensively in 1965 (Cole- man, 1976). Most samples were taken from the deeper parts of the channel and showed Neotrigonia margaritacea and Notocallista diemenensis to be dominant. Venericardia bimaculata and Solen vaginoides were also found at these stations but not in sufficient quantity to be considered dominant. The few shallow water stations sampled were char- acterised by Pronucula concentrica, Mysella donaciformis, Katelysia rhytiphora and Tellina mariae. The 1973-1975 survey has shown that the patterns of distribution shown around Crib Point are typical of the bay as a whole, but the extent to which this later survey presents a complete picture of the mollusc fauna will be limited by the extent and method of sampling. The total area sampled in the 1973-1975 survey (15 m”) represents less than one forty- millionth of the area available to benthic mol- luscs (680 km? on the basis of water surface area but more if the surface area of weed fronds, stones, etc. is considered). Because the Smith-Mcintyre grab is primarily an in- fauna sampler, mainly infaunal species were collected. Some epifaunal species, attached to stones and shell fragments, were collected, but no attempt was made to sample the fauna of hard substrata. Salt marshes and man- groves were also omitted from the survey. To estimate the effects of these limitations the number of species collected was com- pared with the number of mollusc species in the archives of the National Museum of Vic- toria. These archives contain molluscs col- lected non-quantitatively from Western Port over many years and include species from salt marsh, rock and mangrove areas (Smith 8 Jepson, 1974). The comparison showed that 9.8% of the chiton, 7.2% of the gastropod, 38.5% of the bivalve and 12.5% of the cepha- lopod species previously recorded from the bay were collected in the present survey. The success of the survey, in terms of the propor- tion of known species found, is much higher if the comparison is only with archival material from areas similar to those sampled in the present work. On this basis approximately 20% of known chiton, 25% of known gastro- pod and 100% of known bivalve species were found. The survey may therefore be assumed to provide a fairly accurate picture of the mol- lusc fauna, especially the bivalves, in those areas sampled. The mollusc fauna of Western Port In Western Port, as elsewhere (Sanders, 1958; Driscoll & Brandon, 1973; Georges, 1973; Lande, 1975), relatively few species are dominant and provide the majority of individu- als. Nevertheless, the degree of dominance shown by species in Western Port is lower than that recorded from many other areas. Only four species provided more than 10% of the individuals in the assemblages in which they were dominant. The most abundant of these was Tellina тапае which accounted for 29.1% of the molluscs in the fsma. In contrast, dominant species in the mollusc faunas de- scribed for the San Pedro Basin contributed up to 46% of individuals (Bandy, 1958); the most abundant species in collections made off the Dutch Coast accounted for 48.5% of indi- viduals (Eisma, 1966); the most abundant molluscs in each of four mollusc assemblages described for Buzzards Bay acocunted for 85.3%, 47.4%, 49.8% and 39% of the indi- viduals in these assemblages (Driscoll 8 Brandon, 1973); the five most abundant mol- luscs in samples from the Borgenfjorden, Nor- way, together provided more than 80% of in- dividuals (Lande, 1975); and in collections off the Delaware Coast, one species (Mytilus edulis) comprised 58% of the bivalves found (Maurer et al., 1979). К 15 generally true that relatively few mol- luscan families provide genera used to char- acterize faunas (Jones, 1964), and most of the dominant species т Western Pont belong to those genera or families which provide characteristic or dominant species in other parts of the world. Various species of Tellina and Nucula have been described as charac- teristic of fine sand and mud areas in Australia (Stephenson et al., 1974), Scotland (Gage, 1972), the Isle of Man (Jones, 1975) and the U.S.A. (Maurer, 1969; Frankenberg, 1971; Driscoll 4 Brandon, 1973; Kinner et al., 1974). Other Tellina, Venus and Масота communi- ties are listed in Thorson (1957). Cyclocard- ит (= Venericardia), Mysella, Ensis, and Solen species are reported as characteristic MOLLUSCS OF WESTERN PORT 55 of sandy substrata in the Middle Atlantic con- tinental shelf region (Maurer et al., 1976), in the San Pedro Basin, California (Bandy, 1958) and off the coast of the Isle of Man (Jones, 1951). Cyclopecten valves are abun- dant in parts of the San Pedro Basin although this abundance is not reflected in samples of the living mollusc population (Wilson 1956). Sigapatella novaezelandiae is commonly found in Lyttelton Harbour, New Zealand, wherever there are shells and stones for at- tachment (Knight, 1974). Although most of the dominant species in Western Port have parallels elsewhere, the occurrence of Neotrigonia margaritacea, a ‘living fossil (Gould, 1968), is an exception. The family Trigoniidae is represented by more than a hundred fossil species found through- out the world, but living species occur only around Australia (Bednall, 1878; Fleming, 1964). N. margaritacea has been recorded from a few localties in south-east Australia, but the present survey is the first to demon- strate quantitatively its presence as a faunal dominant. The distribution of molluscs in Western Port is clearly related to sediment type and there are two major assemblages, one associated with mud-free, medium sand and the other with fine sand and mud. A few stations (e.g. 1708, 1736, 1750) showed a wide range of affinities within both assemblages, a fact ap- parent from the trellis diagram rather than the dendrograms. These stations were those situ- ated near the boundaries of deep and shallow areas and contained species characteristic of coarse and those characteristic of fine sedi- ments. There seems, therefore, to be a mixed, transitional fauna between the deeper and shallower areas, although the stations concerned were assigned, on the basis of their strongest affinities, to one or other of the assemblages as an analytical convenience. Both epifaunal and infaunal diversity were higher in the coarser Western Port sediments. The same is true in other areas and has been attributed to increased habitat heterogeneity in coarse sediments (Nichols, 1970; Gray, 1974; Biernbaum, 1979). Boesch (1973) in- vestigated sandy and muddy areas and found more species in the former. Approximately half the difference in numbers of species was due to the greater abundance of epifauna, mainly attached to shell and polychaete tube fragments, in the sandy areas. Similarly, п Western Port most epifaunal species were collected from stones, shell fragments and bryozoa from the deeper parts of North and East Arms. Surface deposit feeders showed a clear preference for the fine and muddy sediments of intertidal and shallow sublittoral areas in Western Port. Increased mud content is as- sociated with increased organic content (Buchanan, 1958), and in Western Port areas of fine sediment are also those where sea- grasses and algae, living and as detritus, are most abundant. Suspension feeders, as a group, showed less restriction in sediment preference. Al- though most suspension feeding species and individuals occurred in the cmsa, five of the eight dominant species in the fsma were sus- pension feeders. Katelysia rhytiphora, the most widespread of the suspension feeders in the fsma, has a distribution similar to that of Tellina mariae, the most abundant and wide- spread of the deposit feeders. At all but three of the stations occupied by Tellina but not by Katelysia one or both of the other dominant infaunal suspension feeders (Mysella dona- ciformis, Lepton frenchiensis) was found. In addition, the suspension feeding bivalve Anadara trapezia, a large and conspicuous species, though not numerically dominant in the survey was also found in areas inhabited by Tellina. The survey therefore failed to provide evi- dence for trophic group amensalism (Rhoads 8 Young, 1970; Young 8 Rhoads, 1971). The co-occurrence of deposit and suspension feeders in the fsma may in part be due to the presence of seagrass and algae which stabi- lize the sediment. Measurement of the sus- pended matter in water draining from grassed and ungrassed flats shows that it is greater in water draining from ungrassed areas, prob- ably because seagrass traps suspended mat- ter and prevents or reduces resuspension (Brand & Bulthuis, 1976). The co-occurrence of deposit- and suspen- sion-feeders in the fsma may also be due, at least in part, to a wide tolerance of environ- mental conditions by the species concerned. Certainly, the species characteristic of the fsma occurred over a much wider range of sediments than did those characteristic of the cmsa. Similar distributions of the most com- mon surface deposit and suspension feeders have also been found in Manukau Harbour, New Zealand, possibly because of a wide tolerance of environmental conditions by these species (Grange, personal communica- tion). Levinton (1972) suggested that the un- 56 COLEMAN AND CUFF predictability and variable species composi- tion of phytoplankton necessitates that sus- pension feeders remain food generalists, and therefore trophic specialization resulting from competition for food is reduced. In contrast, the constancy and predictability of the food available to deposit feeders may result in trophic specialization because of competition for food. It might be inferred from Levinton's hypoth- esis that there should be more deposit feed- ing than suspension feeding species, but in Western Port the reverse is true. Rather than being related to food supply, the diversity of deposit and suspension feed- ers in Western Рой may be governed by those factors discussed by Franz (1976) who also obtained results contrary to Levinton's hypothesis. The low diversity of the deposit and of the suspension feeders inhabiting fine sediments arises from two factors. The homogenous conditions in these sediments reduces the rate at which species evolve to occupy different niches. п addition, the fine sediments, because they are intertidal ог shallow sublittoral, are those subjected to the greatest temperature fluctuations, and the thermal stress imposed by these fluctuations has further limited species diversity. When dominant species co-occurred at the same station they were of different feeding types. Calow 8 Calow (1975) have shown that morphologically similar freshwater gastro- pods can co-exist because they are physio- logically differentiated, preferentially selecting those food items which they are most capable of digesting. The same is true for hydrobiid snails (Hylleberg, 1976) and possibly for sur- face deposit feeding bivalves for which there is evidence that the efficiency of feeding and assimilation is related to sediment type (Bubnova, 1974). In contrast, it is unlikely that suspension feeders exercise any significant selection over their food intake, both because of their method of feeding and because they are food generalists (Levinton, 1972). The co-occur- rence of dominant suspension feeding spe- cies could therefore lead to large populations competing for the same food resource, an obvious disadvantage in times of food short- age. When the maximal co-occurrences are of species of different feeding types such competition does not occur. Moreover, sur- face deposit feeders may be capable of using the faeces and pseudofaeces (and associ- ated microorganisms) of suspension feeders (Levinton, 1972; Levinton & Lopez, 1977). Divergence in the sediment preferences of different species of deposit and suspension feeders, which might also indicate divergence in preferences for other environmental char- acteristics, may therefore be a means of more evenly distributing dominant species and most effectively partitioning available food resources. APPENDIX One of our referees usefully pointed out that the Shannon-Wiener diversity index and the Canberra Metric dissimilarity measure have each been the subject of recent criticism and review and suggested newer, potentially more useful, measures. The “expected spe- cies diversity measure” (Smith & Grassle, 1977; Smith, Grassle & Kravitz, in press) and the “normalized expected species similarity index” or NESS (Grassle & Smith, 1976) were singled out. Our attempts to comply with this suggestion produced results which are inter- esting, even though they did not reveal new insights about the mollusc fauna of Western Port. Diversity indices Peet (1974) reviewed the measurement of species diversity in terms of heterogeneity in- dices as well as in terms of the component measures, species richness and equitability (evenness). He defined two types of hetero- geneity measures: type |, being those “most sensitive to changes in the rarest species” and type Il, being those “most sensitive to changes in the importance of the most abundant species” (p. 296). The Shannon- Wiener index is apparently the only type | index well known in benthic ecology and hence we used it in this study. We have also included the measure of evenness given in Pielou (1969). Species richness 15 also presented in the paper. Richness was measured in terms of direct species counts, as the number of spe- cies per 0.3 m”. But richness is known to vary with sample size, ¡.e., the number of individu- als in the sample (Peet, 1974). This is not a problem for the analyses done in this paper but would be if one wished to compare spe- cies richness of Western Port to other marine embayments where a different sample size was used. This problem was first addressed MOLLUSCS OF WESTERN PORT 5% by Sanders (1968) who aimed to calculate the number of species expected from each sam- ple if all the samples were reduced to a standard size; Hurlbert (1971) presented a correct formula for calculating the expected number of species. Smith & Grassle (1977) noted that this “expected species diversity measure” allows one, by varying the standard sample size, to stress the abundant or rare species at will and proposed its use as a di- versity measure. The expected species diversity measure 1$ explained in some detail by Smith & Grassle (1977).Suppose we have a finite “population” consisting of К species with nj; individuals of species 1; let п = У; nj. Then the expected number of species in a random sample without replacement of size m individuals is С (n—n;,m) (_ a n>m 4 C (n,m) l max where C represents combination. For large m, the measure 15 sensitive to rare species; for small m, it depends mainly on the dominant species. For m=2, the measure is linearly re- lated to Simpsons diversity index (a well- known, type Il index; Peet, 1974). A further advantage of the measure is that it has a mini- mum variance unbiased estimator (Smith 4 Grassle, 1977; Smith, Grassle & Kravitz, in press). The expected species diversity measure thus has some useful attributes, the major one to us being the flexibility in emphasizing abundant or rare species at will. Unfortunate- ly, in practice the measure failed to provide useful insights into the diversity of the mollusc fauna of Western Port. Because of the ex- treme variation in number of individuals/sta- tion in Western Port, the “individual index” m cannot be set larger than 2 in calculating average diversity per station per assemblage (Table 2), or per stratum (Table 5). In this case (т = 2) the expected species diversity is equivalent to Simpson's diversity index which emphasizes the abundant species only. We think that a type | measure 15 most useful in summarizing faunal diversity in Western Port. (There remains, of course, the possibility of calculating diversity per assemblage or per stratum—ignoring stations—but such calcu- lations would be inadvisable considering that the variation between stations within assem- blages [or strata] is considerable in Western Port.) Dissimilarity measures The reviewer pointed out that in our affinity analyses we used the most robust agglomer- ative method and suggested that we use a method less sensitive to common species. The reviewer suggested that we consider the NESS measure. NESS is a normalized ver- sion of the expected number of species that two random samples of size m (= © | | | BJEJSOOI/ANO ‘© sisuafuege ‘5 186` 010 001 IuleayJe 00 010 001 00 E 29U219/9H 00° vl ‘+ OOF 001 001 L 186° 001 862 | OOF OO! 001 | 001 OOF 808 — — zl En El SI IUOSUOIP ‘D т 126; 001 c96 — 620° — 850' LE we — 694; 001 001 OO} LEZ COL OL 8 LA $/5иэримоу ‘© er = > 501 001 001 OO} 201 € с | Sally виебииАциел ‘D 105 5-153 с-155 ydy ору 2-ydoy L-ydoty $1207 ‘2 ‘614 ul чэлб $лэашпи Ащеэс| au} pue эшеи зэюэ4$ Aq pajeubisap эле зиоцепао ‚SISEIgOIUOH jo suonejndod 81| ul 1901 omdiouÁjod 9| 10} sansobÁzolajay эбелэле pue “omdiowÁjod 1090| jo uoiuodoid ‘заочцэпбэд ajally ‘р JIGVL CHAMBERS 68 010` Ol —-. — = == A A A = = SS — — 850 — 00 OO! OO! 001 OO! 001 O0! O0! 066° £001 066 о = = к — — = = — = 00 — br zo — — 850 — = — — US — . — 850 960: 90 — — — = 2— — O RES — —= — — — 610 957’ Ore — 656 = = = —- — gz — — — — — == — — — —- = —. — — oo — == — BES 557 692° 190 — 620 — 8560 — 185 — 26 — — 850 В lec 6сс Cet ССР Zes Bee 6 _—. OO 656’ 958: = — = — 622 El 228 00S EZ9. — — 850 — 80 — — ose — — = = = — 80 — — — — g€0 — 965 05/ OO! OO! 001 OOF OO! O0 296 00 OOF OOF OO! 296 001 pr —. = = SS SS = za 257°. = — 29 — a u, Fe — == == == = — — — — = — = = — == = —_ — — 00 + — = == 7. —- = =— = go — — 627 052’ в90: 001 296 OO! OO! OO! OO! 001 256 OO! 256 066 066. — — — Tg une == А — — 11S 882 26’ — 80 — — — — — #0 — — — nn “E = === = ce = a de о — = = — — e == = = —- — — 060 == — — 00 — — — — oo а фо — 00: — 00 — Hi — 296 864 8€S 05/ 001 126° 00 OO! 066 о = — 98 001 850° ze zr Orbe — 60 — — — O! 6 OL \Z wur er al GI LL OL 8 12 9 G v о DANS D IUOSUIYIIP ‘9 $изиериоц ‘D 2 su: o] < À D g Q 2 o о 2 ® 3 2 a 2 m n == — = 501 2-de7 001 ОО’ 186 ZOL L-de7 FE = == 501 ypxaH — 60. — 201 pden € с | Sally sn907 вивбииАциел ‘D QE (panuyuos) “| 3718WL 69 GENETIC DIVERGENCE IN FRESH-WATER SNAILS ‘66 0 0} ¡enba Jo иец} 18,2916 Aouanbas e seu ajayje où и OIYdJOWÁJOd pa18pISUO9 si $п9о| Ve ‘66 0 0} Jenba Jo иец} 1812916 Aouanbeu e seu э|э|е où y эцалюшАюо4 palapisuos si snoo0] V2 (8/61) Sisqueyo vol ee, E HF 8/0 vel 281 8950: ISO’ 850’ 660° GEL BELL 201° 820° 130 Ovo 240 970’ 960 020° eo — АизобАголаеч "SAY DOLO 4005: coc ‘осей Bla eee = SS УР. ФИГ Sle 8145 29 656 ele eee 29 ¿de ealydiowAjod 1907 IO ASS ceca ceci DIL = SIE: < 556° 655% EEE, cc: -000’ 990: сес 29 zor ltt Ш zolydiowAjod 1907] — — — — — — — — 00 — == = = + = — = — 96 — oto ore — 001 00:1 Opts 929 295 ¿Ll = 00: 00Е 001. OOF 001: 001. 001 001 86 — 066° 09/° 516 — — 092° SE 010° 882 = = = = == = — — 001 ool — — — — — — — — — — — — — — — — -- 101 — — — 280 — — = — — 610 = = = = == = — — 201 ypS — — — _ — — — — — — 00 — — = = — = — 96 — — — ZI 850 — — — — — — — — — — 126 00+ 001 26 = = = = — — — — — — = == = — = 620) = — — 86 00, — 066° 862 — — — — — — — = — = — — — — 66 — 881° 010° 067 620 — — — — — — — 80 — 00 — — — 001 — 959 — MO 5565 001 5001: 001. OOF 200 L 066: 001 266 001 066 — — — 101 ECOS À EE A A À : аи” OOO! шба == = == — — — 280 — — — — == — — — — 00 — L6 — — — — — = — — — — — 00 — — — — — рб 001 136’ OO! OO! 00! 066° +06 00'| 9S8° 136’ 001 126° 066 27б 00+ 98° +06 066’ 001 — 60 — — — oro 010: — УЕ 610 — 620 — 80 — bSl 280° 010 601 164 — 290 620 — Oto 220° 901° — — _ = = — — — 620° 620 — 16 — G98 €c6 004 — — — — Gil’ est’ = = = — — — — — 001 00, 850° 6200 — 6516 626’ t68 066 988° ¿18 E = == — 620 290 — 220 801 — oo — — — — = — — — gr — — — — — — — 601 — — 610 — 140 — — 201.0 — — 4168 00+ 066 — 276б 06 126° 666 601 = — — — — —- — — — — — — 010 001 620° — — — РЕ рба9 — — — — — — — — — = — — 00 — — — — — 86 — 606 001 00 — — GgE 067 22° 610 = — 29 010° 620 — — — 001 — 960 — — OO MODE SION € OS; “618 196 O0! 001 62S 066° 1/6 00 OO! OO! 201 oor — — — — — — — — — — — — — — — — — 801 aw 70 CHAMBERS ing the number of mm a band runs cathodal to the standard. The reference population was chosen for reasons discussed in Chambers (1978). Null alleles were scored at a locus when no staining was detected for an individ- ual snail. In all cases of пи! alleles, non-null alleles could be scored on the same gel slice, and other slices from the same gel that were stained for a different enzyme showed normal enzyme activity for the same individual. The frequency of null alleles was calculated as the -proportion of individuals in the sample with the null phenotype. Since heterozygous null alleles are not detectable by these methods, the frequencies of null alleles are underesti- mated. Mendelian interpretations of gel staining patterns for all loci was consistent with the electrophoretic analysis of parental and F, adults from two crosses of individuals from the Rainbow River (4) population of G. floridensis and a single cross between С. floridensis individuals from the Rainbow River (4) and Blue Spring (8). The phenotypes of these individuals are given in the Appendix. Three additional successful crosses were not studied electrophoretically since the progeny had not attained adult size by the time of the termination of the study. One of these crosses was between G. floridensis individuals from site 6 and the remaining two involved pairs from site 8. Twelve more crosses were un- successful because of the pairing of individ- uals of the same sex, discovered later; or death of one or both paired individuals before reaching adult size. Parental individuals in these 12 crosses included G. floridensis from sites 4, 6, and 8; and G. vanhyningiana from site 3. No evidence for asexual reproduction was found during these rearing studies since isolated females did not lay eggs until paired with a male. In addition to the controlled crosses de- scribed above, individuals of G. floridensis (sites 4-6, 8 and 10), G. vanhyningiana (sites 1-3), and G. dickinsoni (sites 12 and 13) were reared from eggs layed in laboratory holding aquaria. In all cases, laboratory reared snails were conchologically identical to their field collected parents, except that spires were much less eroded in laboratory reared individ- uals. Since rearing conditions in the labora-' tory appeared to be uniform for all of these snails, development of shell sculpture pat- terns and color are apparently under genetic, and not environmental control. Average genetic identity (/) and distance (D) over all loci, including those monomor- phic, were computed between all samples ac- cording to the method of Nei (1972). These values are presented in Table 2. Standard er- rors of D (Nei & Roychoudhury, 1974) were about one-half of D for samples within the G. floridensis complex, and about one-third of D for more distantly related samples. A dendrogram generated from the genetic distance matrix by the unweighted pair group method (Sneath 8 Sokal, 1973), recommend- ed for use with electrophoretic data by Ме! (1975), 15 presented in Fig. 3. The populations are divided into three major groups: One rep- resenting only G. curvicostata; a second con- sisting of G. athearni (including the reference population) and G. albanyensis; and the third group containing G. floridensis, G. vanhynin- giana, and С. dickinsoni. The last group, the G. floridensis complex, is further subdivided into three clusters: (i) The Rock Spring (1), Juniper Spring (2), and Juniper Creek Gonio- basis(3), which are here considered to be G. vanhyningiana (see discussion below); (ii) the peninsular samples of G. floridensis; and (iii) the Florida panhandle samples of G. flori- densis and G. dickinsoni. Within the final cluster note that G. dickinsoni does not form a distinct group, as the Spring Creek (12) sam- ple is closer to the Chipola (10) and Chocta- whatchee River (15) samples of G. floridensis than to other G. dickinsoni samples. DISCUSSION The discussion will consist of four parts. First, a few general comments will be made on the interpretation and comparative value of electrophoretic data. Next, within drainage system genetic variation in the G. floridensis complex will be considered. Taxonomic rec- ommendations based on the electrophoretic data will then be made. Finally, a gene diver- sity analysis of the electrophoretic data will be presented with the object of evaluating the ef- fect of isolation of populations of the G. flori- densis complex in different drainage systems. Electrophoretic methods Certain biases inherent in electrophoretic methods must be considered when evaluating electrophoretic data. Perhaps the most impor- tant of these is that electrophoresis separates molecules on the basis of charge and confor- mational differences. Different molecules hav- fal GENETIC DIVERGENCE IN FRESH-WATER SNAILS QU Olt OO! 561 S99 989 195° 829 91/` 659’ 328 698 суб’ 158 EC6 156 296 C68 OL ввозить © ele COL’ ПРЕ 596’ 196’ 678 648’ 506’ 202 096’ ELBE £98 188° 096’ AAA CON 6 sısusäuegje 9 19€ 106 OSL’ 80` 201 696 616 ВОР VEZ, 90`[ LES 8r8 +28 Son: СЕР СЕ SO D IUIeay}e "Y £9c c98 098° Feat ва GO CON SOU OUI О LOL 896’ В вес ор ое ZZ 99u919/9H PIS ISE Ive 262° 600° 150’ 8590’ 690’ €S0 68 595 p6e 882 092 c6c 66€ Be bl €0S 88€” УР’ 166’ 166 210 890’ $590’ 560’ 96 08€ ¿82 8/6 153 8bge SSE EEE EL JUOSUIYIIP ‘© OLI = 62628 0955’ TE “086: 800” ©90' 660’ 19 vee 612 95 ble CBC. 066 69€ cl 10505 665° 195’ $506’ 756: 166' 090 970° 191° EIE 96 Occ Ole 68а AT eZee Si 887 tor 895’ 895’ 8b6 856’ 856’ Lv6' 890 BEL 190 Е ФЕ 8st' 902 vie Ole LE GlS c6t Tr Ste: 096’ 8ÿr6 196 796 556’ 9d “296 666 Lcc 961 Loe 805 686’ Ol 657’ 065’ Sve 108’ 1568’ €c8 658 058 0/8 188' Sig sor Voce бег VEG “SEG Убе 8 617’ 117’ Sep €9€ 969 +89 912 08/` OL 692 808 esl 981 “eZ ESE: 850 2Sé 7 SISUBPLIOY ‘© O6€ ler Вер’ vse Stl OSZ €08 Ecsg Les G6/ 198 948` 840 820 202 1020 906 9 LOVE EV MEGA 958“ ‘6VZ 2S 682 ¿08 0У8` 108" 118’ 6568’ 766’ ИСТ: ¿68 09% v9 98 165’ 98€” 9785’ 962 122 ELL 208 018 $898’ ces 988 с78’ 666’ E88 (SIN в PSE 9ce Sze 605’ 9/` 902 + 8% 918 Orel 162 9242 LIB ELL €68' 000” SOO € 085’ 965’ 968 608 Il ОИ ВУ ЭЛ 008’ SEZ 66 84 818: Lil’ €68' 666' soo 2 BuelbumÁyuen 9 Оу 66 155 605 264 Ш 992 092 018 852° 918’ EZZ 618 892° 168' 766’ 966’ | OL 6 OL д vl el Gl: GE LE Ol 8 Je 9 G y € G | O o o D IUOSUIHOIP ‘5) $/5иариоу ‘D eueIBUIUAYUEA ‘9 о сы о 2 © o 2 © 3 о y à | y a Al a/qe | ul se payeubisap aie зиоцетаод “sisego/uos) jo sajdwes 81 10, (¡puoBeip ay) мо!эа) saouejsip pue (1ецобер eu} anoqe) зэцциар! эцаиао ‘с 3718991 m2 CHAMBERS VAN-1 VAN-2 VAN-3 FLOR-4 FLORES FLORES FLOR -8 FEOS FLOR Ao FLORIS DICK -12 FLORA ВК DICK -14 CURV-10 REF -7 | = ATH-10 — ALB-9 L E 1 alt | Es Oo y y | Io, tena tO I = | Dy, A NE 0 D FIG. 3. Dendrogram computed from genetic distances between 18 samples of Goniobasis. Samples are identified by species abbreviation and site number. The reference population (REF) is here referred to G. athearni (see text). ing the same net charge may be indistinguish- able by the particular electrophoretic tech- nique employed. Electrophoresis will there- fore underestimate the amount of divergence between samples (Singh et al., 1976; Coyne, 1976; Coyne & Felton, 1977). Differences rather than similarities between samples should therefore be stressed when interpret- ing electrophoretic data. For a discussion of this and other biases, Lewontin (1974) should be consulted. In spite of problems in interpreting electro- phoretic data, there 1$ a growing body of liter- ature (reviewed by Ayala, 1975) that indicates a correlation between electrophoretically de- termined genetic divergence and taxonomic divergence. The group most extensively studied in this respect is the Drosophila willis- toni group of fruit flies. Ayala et al. (1974) cal- culated average identity values (/) of Nei (1972) for populations at varying degrees of taxonomic distance. Local populations of the same subspecies tended to be quite similar with an average / of 0.970. Populations at in- creasing levels of taxonomic divergence showed decreasing identities. These values were, for samples of different subspecies, 0.795; different semispecies, 0.798; sibling species, 0.563, and nonsibling species, 0.352. Reviews of genetic divergence in other animals find general agreement with these values (Ayala, 1975; Avise, 1976). However, these taxonomic “standard” / values are averages based on a range of values, and studies on other organisms have occasionally detected full species with considerably high- er identities (M. Johnson et al., 1977; D. Johnson, 1978) than the average / of 0.795 between subspecies of the D. willistoni group. The standard error of each calculated / is an- other factor that negatively affects the com- parative value of genetic identities. Although GENETIC DIVERGENCE IN FRESH-WATER SNAILS 73 the average / values at different levels of taxonomic divergence in the D. willistoni group are useful standards for evaluating taxonomic divergence between populations of other organisms, taxonomic decisions based solely on such comparisons will often be er- roneous. Within drainage geographic variation in the G. floridensis complex St. Johns River Drainage: Although they have been referred to two different species (Clench 8 Turner, 1956), the Goniobasis samples from this drainage are genetically very uniform. G. vanhyningiana, as originally described (Goodrich, 1921), differs Нот G. floridensis in its reduced shell sculpture on the adult whorls (Fig. 1B). Clench & Turner (1956) found no record of sympatry between these two species. The Goniobasis from Juniper Spring (2) have the same shell color and sculpture pattern as G. vanhyningiana, but were included with G. floridensis by Clench & Turner. About 6 km down Juniper Creek (3) from the spring, the Goniobasis have the ‘standard G. floridensis sculpture pattern” of spiral cords intersecting costae to form nodules (Fig. 1A, Е). These individuals are also light brown in shell color. Between these sites, the two forms can be found to- gether with a full range of intermediate forms (Fig. 4). The extremely high / values for С. 6 | я Bene ste hom in a daa baba vanhyningiana from Rock Springs (1) and Juniper Spring (2) and Juniper Creek (3) samples (0.994 and 0.995) are well within the range of conspecific populations of the D. willistoni group (Ayala et al., 1974). Especially noteworthy is the virtual identity (/ = 0.999) between the Juniper Spring (2) and Juniper Creek (3) samples which differ so greatly in shell sculpture. All common alleles are shared by all three samples in this drainage except for Aph100, which was detected only in the Rock Springs (1) sample. If the specific status of G. vanhyningiana is to be maintained, it should include the Juniper Spring (2) and Juniper Creek (3) populations. The reproduc- tive relationships of these three populations to G. floridensis from other drainages have not been determined, so the specific status of G. vanhyningiana is tentative. Waccasassa River drainage: Two different shell forms of G. floridensis intergrade in this drainage. À smooth shelled form is found in the Waccasassa River (6). This form has defi- nite costae, but lacks spiral cords. The Wekiva River, a branch of the Waccasassa River, contains a standard sculpture form.At the confluence of these rivers, these forms hybridize, with a full range of intermediate forms (Fig. 5). The / value for samples taken of these two forms is 0.924, which is well above the average / for conspecific popula- tions of different subspecies of the D. willis- toni group. All common alleles were found in 3 10| i 1 Î FIG. 4. Goniobasis from Juniper Creek, about midway between sites 2 and 3. 74 CHAMBERS | 6 as = Pha deba FIG. 5. Goniobasis floridensis from the Wekiva River at US Highway 19. both samples, except that Aph100was absent from the Waccasassa River (6) sample, and there were sharp differences in allele fre- quencies at Est-3 and Me. Morphological dif- ferentiation between the two forms has oc- curred without dramatic genetic differentia- tion. Suwannee River drainage: There are pro- nounced differences at 4 loci between the two G. floridensis samples from the Suwannee River drainage. The samples from the Iche- tucknee River (7) and Blue Spring (8) had no common alleles at three loci (Acph-1, Acph-2, and [ар-1) ard Aph100 had a frequency of 0.875 at Blue Spring and was absent from the Ichetucknee River sample. The apparent lack of gene flow between these two populations can be explained by the tannin-colored acidic waters of the intervening Withlacoochee Riv- er. Mollusks in general do not tolerate acid waters, and Florida Goniobasis seems to have a particularly strong preference for alka- line spring water (Beck, 1965). The overall identity between these two samples is / = 0.808, which is above the average value found between subspecies of the D. willistoni group. Apalachicola River drainage: G. floridensis in this drainage varies in shell sculpture and appears to interbreed with a form referred to G. dickinsoni by Clench & Turner (1956). An unusual population of G. floridensis in Blue Hole Spring (11) is characterized by indi- viduals of large size with heavy costae and weak cords (Fig. 6G-K). The typical G. flori- densis sculpture increases as the spring flows И towards the Chipola River. The G. floridensis found in the Chipola River (10) have the standard sculpture pattern. The samples from sites 10 and 11 share common alleles at all loci except Aph and Lap-1. At these two loci, each of the two samples has a common allele not found in the other sample. Although the allozyme evidence from these two loci argues against extensive gene flow between snails at these two sites, genetic divergence may be magnified by selection operating at these two loci. Since / = 0.933 for these two samples, they are more similar than the average sub- species of the D. willistoni group. It is also possible that hybridization is recent or, as the morphological evidence suggests, is confined mainly to the vicinity of Blue Hole Spring (11). Morphological evidence suggests that G. floridensis and G. dickinsoni are intergrading near the headwaters of the Chipola River. G. dickinsoni is characterized by a lack of shell sculpture. However, some G. dickinsoni shells in the Spring Creek (12) samples have faint indications of spiral cords (Fig. 6C, D). The frequency and expression of standard G. floridensis sculpture increases going down the Chipola River. The electrophoretic data from the Chipola River (10) sample of G. floridensis and the Spring Creek (12) sample of а. dickinsoni are compatible with the inter- grade hypothesis. All common alleles are shared by the two samples except that Lap- 1107 has a frequency of 0.577 and 6Pgd97 has a frequency of 0.106 in the G. dickinsoni sample. Both alleles are absent from the Chipola River sample of G. floridensis. How- GENETIC DIVERGENCE IN FRESH-WATER SNAILS 75 FIG. 6. Goniobasis from the Chipola River drainage and Holmes Creek. А-В, С. dickinsoni, site 13. C-D, С. dickinsoni, site 12. E-F, G. floridensis, site 10. G-K, G. floridensis, site 11. ever, allele frequencies in the Spring Creek (12) sample at Got, Me, and Sdh are more similar to those of Florida panhandle samples (10, 11, 15) of G. floridensis than to other G. dickinsoni samples. The heterozygosity of the Spring Creek (12) sample was twice that of the other G. dickinsoni samples. This would be an expected result of introgression from G. floridensis. Either G. dickinsoni and G. flori- densis are interbreeding, and are therefore conspecific, in the Chipola River; or there has been recent massive introgression between these two species. Choctawhatchee River drainage: Although I values for G. floridensis and G. dickinsoni in the Choctawhatchee River drainage (sites 13-15) are 0.934 and 0.943 and they are sympatric at site 14, there is no morphological evidence for hybridization between them (Fig. 7). The Choctawhatchee River drainage samples of these two species share all com- mon alleles, except at Me where Мет700 has a frequency of 0.490 in the G. floridensis sam- ple and 1$ absent from the G. dickinsoni sam- ples. Taxonomic implications Electrophoretic methods alone cannot gen- erate classification, but they can supply useful data for making subjective taxonomic deci- sions. One of the less subjective applications of these methods is in evaluating the probabil- ity of gene flow between sympatric or con- tiguous populations. For example, electro- phoretic data give strong evidence against gene flow between G. athearni and G. flori- densis in the Ichetucknee River (Chambers, 1978). In the present study, electrophoretic data are consistent with the hypothesis that G. floridensis and С. dickinsoni are inter- breeding in the upper Chipola River. The conchologically different samples of С. vanhyningiana from sites 2 and 3 are so simi- 76 CHAMBERS 5 6 ben ИНН a ИНН | | FIG. 7. Goniobasis from the Choctawhatchee River drainage. A-B, G. dickinsoni, site 14. C-D, G.flori- densis, site 15. lar in electrophoretically-detected allele fre- quencies that they could be considered as separate samples from a single population. Electrophoretic data also appear to be use- ful in placing populations or forms into species groups. For example, each of the three major clusters in Fig. 3 may be thought of as repre- senting a different species group. This order- ing of species groups gives a different overall view of taxonomic affinities than that provided by the pioneering conchological work of Clench & Turner (1956). G. athearni, thought to be very closely related to G. floridensis, is biochemically very different and is closest to G. albanyensis. This G. athearni-G. albany- ensis group is also distinct from G. curvi- costata, which was thought to have affinities to G. albanyensis. The application of the electrophoretic data to populations within the G. floridensis com- plex, which includes G. floridensis, G. dickin- soni, and G. vanhyningiana, is much more subjective. This is an actively evolving as- semblage of closely related forms that dis- plays a range of internal relationships from populations that are clearly interbreeding to those that have apparently evolved reproduc- tive isolation. The complexity of these inter- relationships does not allow placement of all forms neatly into one or more species. A study of reproductive relationships between these forms, as has been done with the equal- ly complex Drosophila willistoni group (Ayala et al., 1974), might most accurately determine the biological species involved. However, giv- en the difficulties and time required for rearing Goniobasis, it is unlikely that such an effort could be completed before extinction of many populations makes the question purely aca- demic. Inferences based on electrophoretic data, as subjective as they might be, are therefore a reasonable alternative. The amount of taxonomic divergence between two populations can be estimated by comparing their genetic identity (/) with the average iden- tity values for various levels of taxonomic di- vergence in the D. willistoni group (Ayala et al., 1974), whose taxonomic divergence 1$ known from reproductive studies. The follow- ing discussion will deal with the relationships of each of the named species within the G. floridensis complex. Goniobasis vanhyningiana forms a distinct cluster within the G. floridensis complex (Fig. 3). The highest / values (0.891 and 0.893) between G. vanhyningiana samples and G. floridensis are with the G. floridensis sample from Rainbow River (4). These values are GENETIC DIVERGENCE IN FRESH-WATER SNAILS Tal: well above the average / value of 0.795 for different subspecies of the D. willistoni group. However, these St. Johns River drainage forms are distinct from other members of the G. floridensis complex in their light brown shell color and chalky white color seen in the shell aperture. G. vanhyningiana and G. floridensis may be a case of reproductive isolation hav- ing been attained between populations with high genetic identities (M. Johnson et al. 1977; D. Johnson, 1978). Although reproduc- tive studies may eventually indicate that G. vanhyningiana is a зупопут of G. floridensis, the distinctiveness of G. vanhyningiana justi- fies its retention as a species at this time. Goniobasis dickinsoni is found only in the headwaters of the Chipola River, Apalachi- cola River drainage, and some adjacent tribu- taries of the Choctawhatchee River drainage. Clench & Turner (1956) justifiably interpret this distribution as being the result of stream capture or mechanical dispersal in this area of low relief. The present study has reported evidence that this species is reproductively isolated from G. floridensis in one part of its range, the Choctawhatchee River drainage, but not in another, the Chipola River. Since they display incomplete reproductive isola- tion, G. dickinsoni and G. floridensis might be considered semispecies (Mayr, 1963). Mayr (1963) has observed that, if speciation is a gradual process, such intermediate cases should exist. An alternative interpretation is that G. dickinsoni in each drainage system has independently evolved similar shell char- acters, and that the population in the Chipola River is a smooth shelled form of G. flori- densis. Preliminary evidence from karyotypic studies, currently in progress, supports this alternative interpretation. There 15 electro- phoretic evidence of independent reduction in the expression of shell sculpture in Gonio- basis in the St. Johns and Waccasassa River drainages (this study) and in Goniobasis proxima in eastern Virginia (Dillon & Davis, unpublished manuscript). Whichever evolu- tionary interpretation 1$ correct, G. dickinsoni and G. floridensis in these two drainages are very closely related, with / values similar to those between local populations of the D. willistoni group. The only certain taxonomic decision that can be made is that G. dickin- soni and G. floridensis are “good” reproduc- tively isolated species in the Choctawhatchee River drainage, even though genetic diver- gence (/ = 0.934 and 0.943) is less than that observed for average subspecies of the D. willistoni group. The populations referred to G. floridensis in this study cannot be recognized as definite conspecifics based on electrophoretic analy- sis. The only populations known to be capable of interbreeding, Rainbow River (4) and Blue Spring (8), have a genetic identity of 0.886, which is above the average value for sub- species of the D. willistoni group. For the other populations, the electrophoretic data al- low several alternative interpretations. The gene diversity analysis presented in the next section indicates that most of the genetic di- versity in the G. floridensis complex 1$ be- tween drainage systems. The G. floridensis of each drainage can then be thought of as dif- ferent evolutionary units or “populations” in the sense of Ehrlich & Raven (1969). An ex- treme taxonomic interpretation based on this finding would be to split G. floridensis into several species, with each species occupying a single drainage system. Another approach would be to recognize each of the two sub- clusters of G. floridensis as different species: One would occupy the Florida panhandle (sites 10, 11, and 15) and the other the Florida peninsula (sites 4-8). This would be an ex- tremely arbitrary scheme since the range of identities between populations of the two pre- sumptive species includes values above and below the average identity of semispecies of the D. willistoni group. It is probably safest to assume that С. floridensis (sites 4-8, 10, 11, 15, and possibly 12) is a single species, with the provision that some reproductive isolation may be discovered within the “species,” most likely between peninsular and panhandle populations. Apportionment of gene diversity in the G. floridensis complex Genetic divergence between samples can be described using Neïs (1975) method of analyzing gene diversity in subdivided popu- lations. This method partitions the total gene diversity into within subpopulation and be- tween subpopulation diversity, and allows the computation of Gsr, the coefficient of gene differentiation. Gs; is the proportion of the total gene diversity (Нт) attributed to the di- versity among subpopulations (Ост), and ap- proximates Wright's Fs7 (Nei, 1975). The terms “population,” “subpopulation,” and ‘colony’ (used below) as used by Nei (1975), simply describe units in a hierarchy and may not correspond to traditional biolog- ical uses of the terms. For example, in one of Nei’s (1975) analyses, “population” refers to 78 CHAMBERS all humans of three major races with each race considered a separate subpopulation. п a separate analysis in the same chapter, “population” refers to all 37 villages of Yanomama Indians, with each village treated as a separate subpopulation. If the 14 samples of the G. floridensis com- plex are treated as subpopulations, Gsr is 0.712. This is a large amount of diversity be- tween samples as compared to @ст values for other organisms (Nei, 1975). If the diver- gent St. Johns River drainage samples (G. vanhyningiana) are omitted, Gr is still 0.654, which 1$ exceeded only by the value of 0.674 for the Ord Kangaroo rat (Dipodomys ordi), the highest С ст reported by Nei (1975). High Gs7 values describe only the relative amount of divergence between samples. The cause of this divergence may or may not in- volve reproductive isolation between sam- ples. Since there is an overall correlation be- tween genetic divergence and the attainment of reproductive isolation, extremely high @ст values, such as that for the G. floridensis complex, indicate the possibility of some re- productive isolation within the complex. Gene diversity can be even more finely par- titioned if the subpopulations can themselves be subdivided into colonies (Nei, 1975). This relationship can be defined as: Hr = Hc ar Des + Dsr, where Нс is the gene diversity within colonies and Dos is the gene diversity between colo- nies within subpopulations. This can be ap- plied to the G. floridensis complex by consid- ering drainage systems as subpopulations and each sample taken within these systems as a colony. The absolute values listed in Table 3 are the result of this analysis. These values are also presented as percentage of the total diversity (Нт). The greatest proportion of the gene diversity in the G. floridensis complex occurs between drainage systems (Gs; = 0.458) and the smallest proportion between samples within drainages (Gcs). The general conclusions are unchanged if the St. Johns drainage samples are omitted from the calcu- lations (Gsr = 0.408, Ges = 0.233). The be- tween drainage proportion of gene diversity is largest even though the within drainage com- ponent includes divergent samples within the Suwannee River drainage and reproductively isolated populations of G. floridensis and G. dickinsoni in the Choctawhatchee River drainage. The relative amount of within drainage genetic divergence in the G. floridensis com- plex seems low relative to that found in sun- fish (Avise & Smith, 1974). This might be ex- pected based on the smaller sizes of drainage systems considered in the Goniobasis study, although the probably more limited within drainage dispersal abilities of Goniobasis should, to some extent, compensate for this. The drainages in the sunfish study were fur- ther subdivided by dams, which are apparent- ly strong barriers to gene flow (Avise & Felley, 1979). A certain minimum amount of dispersal of Goniobasis between drainages is required to explain the occurrence of these snails in dif- ferent river drainages. Geologically related mechanisms, such as stream capture, con- fluence of previously separate rivers by low sea level stands, and connection by under- ground caverns are possible factors. Flooding at the headwaters or mouths of adjacent rivers might provide an occasional avenue of dis- persal. These mechanisms have apparently not occurred recently enough or frequently enough to keep populations in adjacent drainages from diverging genetically. Passive overland dispersal is common in TABLE 3. Summary of gene diversity analysis of samples of the G. floridensis complex occupying various drainage systems (see text). Including St. Johns drainage Percent of Absolute Within samples (Hg) 0.065 Within drainages (Dos) 0.038 Between drainages (D sy) 0.087 Total diversity (H 7) 0.190 Excluding St. Johns drainage Percent of total diversity Absolute total diversity 34.2 0.074 35:9 20.0* 0.048 23:31 45.8** 0.084 40.8** 100 0.206 100 "Value of Gcs x 100 ‘Value of Сост = 100 GENETIC DIVERGENCE IN FRESH-WATER SNAILS 79 many aquatic invertebrates (Maguire, 1963). Such dispersal in many of these animals 1$ facilitated by their possession of dessication- resistant cysts or eggs, which have not been found in Goniobasis. Members of the G. floridensis complex lay their eggs in sand- encrusted clusters on underwater plants, leaves, stones, and other snail shells. One of these items with attached eggs or embryos might occasionally be transported by wind or animal, but lack of dessication-resistant egg adaptations would minimize the probability of successful dispersal by these means. Aquatic gastropods may be transported attached to in- sects or the feet and plumage of aquatic birds (Baker, 1945; Malone 1965a,b; Rees, 1965). However, the available evidence for these mechanisms 15 from pulmonate gastropods, which have the advantages of respiration by means of a lung and are often capable of self- fertilization. Dioecious reproduction т Goniobasis necessitates the transport of a gravid female or successful mating within the lifetime of the transported individual for there to be a chance of gene flow between drain- ages. Goniobasis respires by means of a gill, which means that respiratory efficiency is probably greatly reduced while the animal is removed from water. The reproductive and respiratory systems of Goniobasis and many other pleurocerid snails render them less like- ly overland colonists than most pulmonate snails. The rarity of dispersal between drainages presents a sufficient explanation for the genetic divergence found between members of the G. floridensis complex occupying dif- ferent drainage systems. Land barriers be- tween populations have probably played a major role in promoting genetic divergence and the complex patterns of speciation ob- served in Goniobasis and other pleurocerid snails. ACKNOWLEDGEMENTS | am indebted to T. С. Emmel, M. D. Huet- tel, В. Franz, J. Т. Giesel, M. Kreitman, and F. G. Thompson for their contributions to this study. В. T. Dillon and С. M. Davis kindly al- lowed me to read and cite their unpublished manuscript. Major Jim Stevenson of the Flor- ida Division of Recreation and Parks granted permission to collect Goniobasis in Ichetuck- nee Springs State Park and Florida Caverns State Park. Financial support for this project was provided by the Theodore Roosevelt Memorial Fund of the American Museum of Natural History, a Sigma Xi Grant-in-Aid of Research, and a National Science Founda- tion Grant for Improving Doctoral Dissertation Research in the Field Sciences. LITERATURE CITED AVISE, J. C., 1976, Genetic differentiation during speciation, In: AYALA, F. J. (Ed.), Molecular Evolution. Sinauer Associates, Sunderland, Massachusetts, p. 106-122. AVISE, J. С. & FELLEY, J., 1979, Population struc- ture of freshwater fishes |. Genetic variation in bluegill (Lepomis macrochirus) populations in man-made reservoirs. Evolution, 33: 15-26. AVISE, J. С. & SMITH, М. H., 1974, Biochemical депейс$ of sunfish. |. Geographic variation and subspecific intergradation in the bluegill, Lepomis macrochirus. Evolution, 28: 42-56. AYALA, F. J., 1975, Genetic differentiation during the speciation process. In: DOBZHANSKY, T., НЕСНТ, М. К. 8 STEERE, W. С. (Eds.), Evolu- tionary Biology. Plenum Press, New York. Vol- ume 8: 1-78. AYALA, F. J. (Ed.), 1976, Molecular Evolution. Sinauer Associates, Sunderland, Massachu- setts, 277 p. AYALA, Е. J., HEDGECOCK, R., ZUMWALT, С. 5. & VALENTINE, J. W., 1973, Genetic variation in Tridacna maxima, an ecological analog of some unsuccessful evolutionary lineages. Evolution, 27: 177191. AYALA, Е. J., TRACEY, М. L., HEDGECOCK, D. & RICHMOND, В. C., 1974, Genetic differentiation during the speciation process in Drosophila. Evolution, 28: 576-592. BAKER, F. C., 1945, The Molluscan Family Planorbidae. University of Illinois Press, Urbana, 530: р: BECK, W. M., 1965, The streams of Florida. Bul- letin of the Florida State Museum (Biological Sciences), 10: 91-126. CHAMBERS, S. M. 1977, Genetic divergence dur- ing speciation in freshwater snails of the genus Goniobasis. Ph.D. Dissertation, University of Florida. CHAMBERS, S. M., 1978, An electrophoretically detected sibling species of “Goniobasis flori- densis.” Malacologia, 17: 157-162. CLENCH, W. J. & TURNER, В. D., 1956. Fresh- water mollusks of Alabama, Georgia, and Florida from the Escambia to the Suwannee River. Bul- letin of the Florida State Museum (Biological Sciences), 1: 97-239. COYNE, J. A., 1976, Lack of genic similarity be- tween two sibling species as revealed by varied techniques. Genetics, 84: 593-607. COYNE, J. A. & FELTON, A. A., 1977, Genic 80 CHAMBERS heterogeneity at two alcohol dehydrogenase loci in Drosophila pseudoobscura and Drosophila persimilis. Genetics, 87: 285-304. DARWIN, C., 1860, The Voyage of the Beagle. ENGEL, L.(Ed.), 1962, The Natural History Li- brary edition, Doubleday, Garden City, New York, 524 p. EHRLICH, Р. В. 8 RAVEN, P. H., 1969, Differentia- tion of populations. Science, 165: 1228-1232. GOODRICH, C., 1921, Three new species of Pleuroceridae. Occasional Papers of the Mu- seum of Zoology, University of Michigan, 91: 2- 8. HEDGECOCK, D., 1978, Population subdivision and genetic divergence in the red-bellied newt, Taricha rivularis. Evolution, 32: 271-286. JOHNSON, D. L. E., 1978, Genetic differentiation in two members of the Drosophila athabasca com- plex. Evolution, 32: 798-811. JOHNSON, М. S., CLARKE, В. 8 MURRAY, J., 1977, Genetic variation and reproductive isola- tion in Partula. Evolution, 31: 116-126. LEWONTIN, В. С., 1974, The genetic basis of evo- lutionary change. Columbia University, New York, 346 p. MACARTHUR, R. H. 8 WILSON, E. O., 1967, The theory of island biogeography. Princeton Uni- versity, Princeton, New Jersey, 203 p. MAGUIRE, B., Jr., 1963, The passive dispersal of small aquatic organisms and their colonization of isolated bodies of water. Ecological Mono- graphs, 33: 161-185. MALONE, С. R., 1965a, Killdeer (Charadrius vociferus Linnaeus) as a means of dispersal for aquatic gastropods. Ecology, 46: 551-552. MALONE, С. R., 1965b. Dispersal of aquatic gastropods via the intestinal tract of water birds. Nautilus, 78: 135-139. MAYR, E., 1963, Animal species and evolution. Belknap, Cambridge, Massachusetts, 797 p. MERRITT Rs B., ROGERS, J) Е. & КУБ В 1978, Genic variability in the longnose dace, Rhinichthys cataractae. Evolution, 32: 116-124. NEI, M., 1972, Genetic distance between popula- tions. American Naturalist, 106: 283-292. NEI, M., 1975, Molecular population genetics and evolution. North-Holland, New York, 299 p. NEI, М. & ROYCHOUDHURY, А. K., 1974, Sam- pling variances of heterozygosity and genetic dis- tance. Genetics, 76: 379-390. REES, W. J., 1965, The aerial dispersal of Mol- lusca. Proceedings of the Malacological Society of London, 36: 269-282. SINGH, В. S., LEWONTIN, В. С. & FELTON, А. A., 1976, Genetic heterogeneity within electro- phoretic “alleles” of xanthine dehydrogenase in Drosophila pseudoobscura. Genetics, 84: 609- 629. SNEATH, Р. Н. A. & SOKAL, В. R., 1973, Numer- ical taxonomy. Freeman, San Francisco, 573 р. APPENDIX Electrophoretic analysis of Goniobasis floridensis crosses Electrophoretic phenotypes at 18 loci of parental and F, individuals from three crosses are presented below. In parenthesis are the numbers of F , individuals with the stated phenotype. Cross no. 1: Rainbow River (site 4) x Rainbow River (site 4) Parental ? 3 Acph-1 107 Acph-1 107 Acph-2 103 Acph-2 103 Aldo 94 Aldo 94 Aph 102 Aph 102 Est-1 100 Est-1 100 Est-295 Est-295 Est-3 100 Est-3 100 Got93 Got93 С3ра95 G3pd95 Hexdh 100 Hexdh 100 Lap-1 104 Lap-1 104 Lap-2 102 Lap-2 102 Me 102 Me 102 6Pgd 109 6Pgd 109 Ра! 100 Pgi 100 Pgm 101 Pgm 101 Sdh98 Sah 98 To 100 To 100 RU Acph-1107(12) Acph-2 103(12) Aldo94(12) Aph 102(12) Est-1 100(12) Est-295(12) Est-3 100(12) Got93(12) G3pd95(12) Hexdh 100(12) Lap-1 104(12) Lap-2 102(12) Me 102(12) 6Pgd 109(12) Pgi 100(12) Pgm 101(12) Sdh98(12) To 100(12) 2 Acph-1107 Acph-2 103 Aldo 94 Aph 102 Est-1 100 Est-295 Est-3 100 Got93 G3pd95 Hexdh 100 Lap-1 104 Lap-2 102 Me 102 6Pgd 109 Pgi 100 Pgm 101 Sdh98 To 100 Q Acph-1107 Acph-2 103 Aldo 94 Aph 102 Est-1 100 Est-295 Est-3 100 Got93 G3pd95 Hexdh 100 Lap-1 104 Lap-2 102 Me 102 6Pgd 109 Pgi 100 Pgm 100/101 Sdh98 То 100 GENETIC DIVERGENCE IN FRESH-WATER SNAILS 81 Cross по. 2: Rainbow River (site 4) x Rainbow River (site 4) Parental 6) Acph-1107 Acph-2 103 Aldo 94 Aph 102 Est-1 100 Est-295 Est-3 100 Got 93 G3pd95 Hexdh 100 Lap-1 102/104 Lap-2 102 Me 102 6Pgd 109 Pgi100 Pgm 101 Sdh98 To 100 Acph-1107(8) Acph-2 103(8) Aldo94(8) Aph 102(8) Est-1 100(8) Est-295(8) Est-3 100(8) Got93(8) G3pd95(8) Hexdh 100(8) Lap-1 104(3), Lap-1 102/104(5) Lap-2 102(8) Me 102(8) 6Pgd 109(8) Pgi100(8) Pgm101(8) Sdh98(8) To100(8) Cross no. 3: Rainbow River (site 4) x Blue Spring (site 8) Parental d Acph-195 Acph-2 103/105 Aldo94 Aph 100 Est-1 100 Est-295 Est-3 100 Got93 G3pd95 Hexdh 100 Lap-1null Lap-2 102 Me 102 6Pgd 105/109 Pgi 100 Pgm 100 Sah 98 To 100 E PRE A Acph-195/107(7) Acph-2 103(4), Acph-2 103/105(3) Aldo94(7) Aph100/102(7) Est-1 100(7) Est-295(7) Est-3 100(7) Got93(7) G3pd95(7) Hexdh 100(7) Lap-1 104(7) Lap-2 102(7) Me 102(7) 6Pgd 105/1095), 6Pgd 199(2) Pgi 100(7) Pgm 101(3), Pgm 100/101(4) Sdh98(7) To 100(7) MALACOLOGIA, 1980, 20(1): 83-98 THE GONIOBASIS OF SOUTHERN VIRGINIA AND NORTHWESTERN NORTH CAROLINA: GENETIC AND SHELL MORPHOMETRIC RELATIONSHIPS! Robert T. Dillon, Jr? and George М. Davis? ABSTRACT Three species of Goniobasis inhabiting the upper New River and surrounding drainages, G. proxima, G. semicarinata, and G. simplex, were found to be very distinct using starch gel electrophoresis. Heterozygosity (H) is low, averaging 0.0113 across 12 populations and 15 loci. The average genetic identity (Nei, 1972) between conspecific populations is 0.89, also very low compared to averages recorded in other animals. Much of the difference between conspecific populations is due to fixation of alternative alleles, suggesting that gene flow even among populations connected through water can be quite low, or that selection is high. Three races are provisionally recognized in both G. proxima and G. semicarinata on the basis of differences in isozyme frequencies. Compared to electrophoresis, species identification using shell morphol- ogy alone was found to be unreliable. A population indistinguishable from G. simplex after a multidimensional scaling based on seven shell characters was revealed to be G. proxima using electrophoresis. The unusual shell morphology of this population may result from the introduc- tion of the normally softwater-dwelling G. proxima into a hard-water stream similar to those inhabited Бу С. simplex. И divergence is measured as euclidean distance to species centroid after multidimensional scaling, the amount of population divergence in shell character 1$ corre- lated to the amount of genetic divergence at the .05 level. INTRODUCTION Pleurocerid snails (freshwater proso- branchs) have undergone ап extensive endemic radiation in streams and rivers of southeastern U.S.A. A bewildering variety of shell phenotypes has been the basis for some 500 nominal species (review of Tryon, 1873: Goodrich, 1940, 1941, 1942, 1944). The de- scription of so many phenotypes has posed tremendous problems to taxonomists working with these mollusks. Centers of gastropod endemism have pro- vided areas for the study of adaptive radia- tion, speciation, and divergence. Some ex- amples of such studies involve the pulmonate land snail genera Partula (Murray & Clarke, 1968) and Cerion (Gould & Woodruff, 1978), and the freshwater prosobranch Triculinae (Davis, 1979). This study deals with some members of the pleurocerid genus Gonio- basis in parts of four drainages on the north- ern edge of their area of endemism. The upper New River drains a sparsely populated, mountainous region on the border of Virginia and North Carolina. In an earlier survey (Dillon, 1977), several species of the freshwater snail genus Goniobasis (Pleuro- ceridae) were found common and widespread in small streams and tributaries of the drain- age but absent from the main bed of the New River (Fig. 1). The tendency for Goniobasis populations to diminish downstream is wide- spread and has been shown previously in G. proxima by Foin (1971) and Foin & Stiven (1970). In a survey of a single G. proxima population, Dillon (unpublished) found snail densities over 300 per square meter in a stream only one meter wide. Snail density gradually decreased to less than two per square meter by the time the stream was five meters wide, five kilometers downstream. Dispersal further downstream, between iso- lated populations, probably occurs only when snails are dislodged by flooding. Eggs are cemented firmly onto rocks, and individual snails generally move against the current (Crutchfield, 1966; Krieger & Burbanck, 1976). Although the New River is quite ancient geologically (Ross, 1969), observed Gonio- basis distributions probably date only from the ÎThis work was supported, in part, by a NSF Grant to G. M. Davis, No. DEB 78-01550. Department of Malacology, Academy of Natural Sciences of Philadelphia, and Department of Biology, University of Pennsylvania, Philadelphia, PA 19103-19104, U.S.A. Department of Malacology, Academy of Natural Sciences of Philadelphia, Philadelphia, PA 19103, U.S.A (83) 84 DILLON AND DAVIS HOLSTON R oa en a Jeff Yad YADKIN 1 te EN NEW К Sink SMITH R FIG. 1. Schematized map of the 5 drainage systems in the study area. Darkened symbols indicate presence of Goniobasis: circles, G. proxima; triangles, G. semicarinata; squares, G. simplex. Open circles are sites where no Goniobasis were collected by Dillon (1977). The 12 larger symbols designate the locations of population sampled for this study. The dashed line approximates the southeastern extent of limestone and dolomite. Pleistocene. The heavy rains, high erosion rates, and lowered temperatures that accom- panied worldwide climatic change probably affected the ranges of mountain stream-dwell- ing organisms greatly. Currently, the main body of the New River forms a barrier frag- menting the range of Goniobasis into numer- ous, isolated populations. To some extent these events must have been repeated in other rivers of the southeastern U.S.A., from which scores of pleurocerid species have been described. The erection of barriers to gene flow, while surely not the sole method of animal speciation (Bush, 1975; White, 1978; Futuyma, 1979), has probably played an im- portant role in pleurocerid radiation. One ob- jective of this study is to assess the impact of barriers to dispersal between Goniobasis populations. Electrophoretic investigation of this inter- esting system has been initiated by Cham- bers (1977, 1978, 1980). He has found that genetic divergence among consubspecific populations of Goniobasis in scattered Florida rivers is greater than the divergence т Drosophila populations reported by Ayala et al. (1974). Chambers found heterozygosity quite low in the 18 populations he surveyed and conspecific populations occasionally fixed for alternative alleles at particular loci. This suggests that migration among Gonio- basis populations may be rare. Does migra- tion seem to be as uncommon among Gonio- basis isolated within the same river system as it seems between rivers? The second objec- tive of this study is to assess the amount of genetic similarity within and between popula- tions of the three Goniobasis species found in the upper New River, G. proxima (Say), G. semicarinata (Say), and G. simplex (Say), and compare these populations to conspecific populations in adjacent drainages. СОМОВА$!$ GENETICS AND SHELL MORPHOMETRICS 85 Species identities of North American pleurocerids have been based entirely on shell characters. Many authors (cf. Dazo, 1965) have observed great variability in the shells of pleurocerids and suggested that varying environments are responsible. A great deal of interpopulation variability is ap- parent in the shells of the three Goniobasis species examined in this study. How do shell characters compare with electrophoretic data in distinguishing between species and popu- lations? The third objective of this study 1$ to examine the relationship between electro- phoretic and shell morphological data. METHODS Populations studied Eight populations of Goniobasis were se- lected to represent the geographic and mor- phological range of the genus in the upper New River. Four populations were chosen from surrounding drainages for comparison (Fig. 2). These populations represent three species, Goniobasis proxima, Goniobasis semicarinata and Goniobasis simplex. Spe- cies identifications were based on concho- logical comparisons with Tryon (1873) and collections in the Academy of Natural Sci- ences of Philadelphia. As will be shown later, conchological comparisons are no sure method of determining species status. Col- lection sites are shown in Fig. 1. Populations of G. proxima were given the names CRIP, HILL, IND, JEF, ROCK, and YAD, populations of G. semicarinata were designated MEAD, PINE, ROA, and SINK, and those of G. sim- plex were designated HOLS and WYTH. Locality data for the 12 populations are pre- sented in the Appendix. At least 100 large individuals were random- ly collected at each site in September, 1978. Snails were kept alive in cloth sacks sub- merged under water in a large bucket in order to clean the guts of contents; the water was changed periodically for 24 hours. They were then quickly frozen (around —20°C) and kept in the freezer during the course of the study. Voucher specimens have been deposited in the Academy of Natural Sciences of Phila- delphia; voucher specimen catalog numbers are given in the Appendix. Although the geology of the study area is complex, limestone and dolomite generally underlie regions northwest of the dashed line in Fig. 1. Southeast of the line, the surface FIG. 2. Representative shells from the Goniobasis populations surveyed. a, b, G. simplex; c-h, G. proxima; i-l, G. semicarinata. a—HOLS, b— WYTH, с СЕР, d—HILL, e—IND, f—YAD, g— JEFF, h—ROCK, i—MEAD, J—ROA, k—PINE, |I— УМК. Length of shell “a” 1.65 cm; remaining shells to same scale. See Fig. 1. geology 1$ principally gneiss and schist. Since it has been demonstrated that the distribu- tions of freshwater mollusks can be highly т- fluenced by the effects of limestone on water quality (Shoup, 1943; McKillop 4 Harrison, 1972; Dussart, 1976), alkalinity was meas- ured at each of the 12 sites. Sepkoski & Rex (1974) have found a correlation between alkalinity (‘bicarbonate ion concentration” and calcium concentration, hardness, pH, and total dissolved solids at the .01 level. All of these variables tend to increase in value as a drainage area includes more limestone and dolomite. Alkalinity measurements were made at the riverbank by staining 50 ml of water first with phenolphthalein and then with methyl purple, and titrating with .02N sulfuric acid (American Public Health Association, 1976). DILLON AND DAVIS 86 600' OSp' 2L0' GSIO'C vol S'Sc 900' ect to 8811 $10 ces | 020° oSe | 620° 65/`1 clO ОС 6S 605 LLO” Ore clO ¿co | 510’ 6071 vLo' вет 8Lo' SES I 210 6/1 1 vco' 6191 $10` 0151 910 2021 460’ 966 с 85 9'S! yoo 18< clO cr6 210 £6c | vlO stol 610 8551 510` OLS 210' 9vl'e 59`' S'0c sto 19€' $10` Bell eco" 887 | 910 СЕ | eco" sis"! y¡bua| |эч$/чбиа| einyedy UIPIM [19yS/UIPIM HOUM pay | YIPIM неч/ч6че| печ$ (saalbap) эбие ads (wo) yıpım alnuady (wo) uBue]| jioym Ápog (wo) y¡bua] 119ys nn tr ee 86866 SIOH HLAM VOU ANIS УЗИ 3NId а! 1904 OVA 1131 ам ИН 0 ‘иеэш ay) jo 10118 рлериез ay) $! иваш yes mojaq Апоэла ‘зиоцепадо4 $/5едошо® Z| jo цэеэ илод S|ENPIAIPU! OZ чо paınseaw $лэелецо 194$ 10} SUBAN ‘| 371891 GONIOBASIS GENETICS AND SHELL MORPHOMETRICS 87 Shell morphometrics Twenty large snails were chosen from the collection made at each site for shell morphol- ogy measurements. The seven measure- ments made on each shell are listed in Table 1. Spire angle, aperture length, shell width, and length of body whorl were measured with calipers using the methods of Davis (1969). Shell length was measured as the length of the last three whorls only, since shell apices were often eroded. Aperture width was the maximum distance across the aperture per- pendicular to the measure of aperture length. The width of the third whorl up from the aper- ture was measured. Shell width, aperture length, and third whorl width were converted into ratios for an analysis of shape. Electrophoresis Horizontal starch gel electrophoresis was performed using the procedure described by Ayala et al. (1973). The shells of individual snails were crushed using a pair of pliers and picked from the body with forceps. Then each body was placed in a centrifuge tube with 0.5 ml deionized water and homogenized by sonication. Small tabs of Whatman No. 3 filter paper were dipped in the crude homogenate, blotted, and applied to the gel. Gels were pre- pared using 35g of Electrostarch and 250 ml of one of 3 gel buffers. Table 2 shows the compositions of both the gel buffers and the buffers used in the electrode trays. Gels were run at 35 milliamps or 350 volts, but not exceeding either. TABLE 2. Buffers used in gels and electrode trays. Concentration of ingredients (Molarity). Citric acid (mono- Boric Buffer, pH Tris hydrate) acid NasEDTA Tris-Cit 6 Tray .237 .085 Gel .0083 .0030 TEB 8 Tray .500 .645 .0179 Се! .050 .097 .0018 ТЕВ 9.1 Tray 8 Gel .087 .0087 .0011 After electrophoresis, the gels were sliced and stained for one of 15 enzymes (Table 3). Gel scoring methods were those of Ayala et al. (1973). The following are stain buffers and other standard components of the stains em- ployed along with an abbreviated name for each. “0.1 Tris НС! buffer’—0.1 molar tris (hydroxymethyl) aminomethane (‘Tris’) ad- justed to pH 7.95 with concentrated HCI. “0.2 Tris HCl buffer’—the same but 0.2M Tris. “DH buffer’—0.102M Tris, adjusted to pH 8.4 with concentrated HCl. “Тиз maleate A’— 0.2M Tris, 0.2M maleic acid. “Tris maleate B"—0.2M NaOH. “NAD"—0.05 ml of a 3% solution. “NADP’—0.05 ml of a 2.5% solu- tion. “MgCl>"—1 drop of a 1% solution. “MTT"—-0.05 ml of a 40 mg/ml suspension of 3-(4,5 dimethylthiazolyl-2)-2,5 diphenyl tetra- zolium bromide. “PMS'"—about 0.1 mg of phenazine methosulfate. “Адаг’—10 ml of a 2 g/100 ml solution. “G6PdH” and “PGI"— 5 ul of a 1000 units/ml! solution of glucose-6- TABLE 3. Enzymes surveyed in Goniobasis populations. Enzyme Abbreviation Acid phosphatase Acph Aspartate aminotransferase Aat Glucose-6-phosphate dehydrogenase G6pd Glucose phosphate isomerase Gpi Hexanol dehydrogenase Hexdh Isocitrate dehydrogenase Isdh Leucine aminopeptidase Lap Malate dehydrogenase Mah Mannose-6-phosphate isomerase Mpi Octopine dehydrogenase Odh 6-Phosphogluconate dehydrogenase 6Pgd Phosphoglucomutase Pgm Sorbitol dehydrogenase Sdh Superoxide dismutase Sod Xanthine dehydrogenase Xdh Buffer system Run time (hr.) TEB+951 TEB 9.1 TEB 9.1 Tris-Cit 6 TES 9.1 TEB:8 TEB 9.1 Tris-Cit 6 Tris-Cit 6 Tris-Cit 6 Tris-Cit 6 Tris-Cit 6 TEB 9.1 TEB 9.1 TEB 8 OPPPDD pOUOnNOoOoO BR CO BR CO CO 88 DILLON AND DAVIS phosphate dehydrogenase and phosphoglu- cose isomerase, respectively. Descriptions of the staining procedures em- ployed follow, with standard recipe compo- nents referenced by their abbreviated names above. Most recipes have been modified from Shaw 4 Prasad (1970) and Bush 4 Huettel (1972): Acph—a-napthyl acid phosphate 16 mg, Fast Blue BB salt 20 mg, MgClo. Tris maleate A 2.5 ml, Tris maleate B 1.3 ml, water 6.2 ml, Agar. Aat—L-aspartic acid 400 mg, a-ketoglutaric acid 200 mg. Fast Blue BB salt 300 mg, pyridoxal-5'-phosphate 0.5 mg, 0.2 Tris НС! buffer 100 ml. G6pd—D-glucose-6-phosphate 20mg, NADP, MgCl, MTT, PMS, 0.1 TrisHCI buffer 10 ml, Agar. Gpi—D-fructose-6-phosphate 20mg, G6PdH, NADP, MgClo, MTT, PMS, 0.1 TrisHCI buf- fer 10 ml, Agar. Hexdh—1-hexanol 5 ml, NAD 25 mg, Nitro blue tetrazolium 20 mg, PMS 0.5 mg, DH buffer 100 ml. Isdh—DL-isocitric acid 20 mg, NADP, MgCh, МТТ, PMS, 0.1 TrisHCI buffer 10 ml, Agar. Lap—L-leucyl-B-napthylamide НС! 20 mg, Black К salt 50mg, Tris maleate A 12.5ml, Tris maleate В 6.5 ml, water 81 ml. Mdh—Substrate solution 1.5 ml (containing L- malic acid 13.4 g, 2M Na CO; 49 ml, water to 100 ml, adjusted to pH 7 with Ма›СОз). NAD, МТТ, PMS. 0.1 TrisHCI buffer 10 ml, Agar. Mpi—D-mannose-6-phosphate 10mg, G6PdH, PGI, NADP, MTT, PMS, 0.1 TrisHCI buffer 10 ml, Agar. Odh—(-+ )octopine 8 mg, NAD, MTT, PMS, 0.1 TrisHCI buffer 10 ml, Agar. 6Pgd—6-phosphogluconic acid 10mg, NADP, MTT, PMS, 0.1 Тиз НС! buffer 10 ml, Agar. Pgm—a-D-glucose-1-phosphate 30 mg, a-D- glucose-1,6-diphosphate 0.5 mg, G6PdH, NADP, MTT, PMS, MgCl, 0.1 TrisHCI buf- fer 10 ml, Agar. Sdh—D-sorbitol 250 mg, NAD, MTT, PMS, DH buffer 10 ml, Agar. Sod—lightly colored bands representing this enzyme were most apparent on the darkly stained Hexdh gel. Xdh—Hypoxanthine 10 mg, 0.05M TrisHCI pH 7.5 10 ml (boiled to dissolve Hypoxanthine then cooled to room temperature), КС 15 mg, NAD, MTT, PMS, Agar. Isozyme bands are named according to their mobilities compared to the most common al- lele in WYTH, the reference population. Meth- ods and assumptions are those in Ayala et al. (1973). Analytical methods The program NT-SYS of Rohlf et al. (1972) was used to explore the electrophoretic and morphological relationships among the Goniobasis populations. For the shell mor- phometric study, а principal component analysis was performed on the correlation matrix of standardized means for the seven characters. The factor scores on the first 3 principal components were used as the initial configuration for multidimensional scaling. The scaling was done to maximize goodness- of-fit to the monotonic regression of the dis- tance between populations in three-dimen- sional space and their true distances in seven-dimensional space. Here taxonomic distance was calculated as the square root of the average squared difference between populations across the morphological vari- ables (Sokal, 1961). Finally, principal com- ponent analysis was reapplied to the covari- ance matrix of the new three-dimensional scaled distances (three factors account for 100% of the variance). The analysis of the electrophoretic data was similar to that applied to the shell charac- ters. Genetic distances were calculated ac- cording to Ме! (1972). Then multidimensional scaling was performed to maximize good- ness-of-fit to the regression of genetic dis- tance and distance in three-dimensional space. The initial configuration of the points in the analysis of electrophoretic data was ran- dom. Three-dimensional scaling was followed by principal component analysis as described above. More complete discussion of these methods 15 given by Sneath & Sokal (1973). Here we introduce a method of quantifying population divergence. The amount of diver- gence necessary to explain variation in the phenotypes of n populations is minimized by joining them in a Wagner or Steiner network including 2n —3 segments and п-2 branching points or nodes (Farris, 1970). Thus if evolu- tion has occurred parsimoniously, the Wagner network best reconstructs the evolutionary relationships among a group of populations. But a problem may arise using this method if the object is to compare divergence meas- ured by several criteria, for small differences GONIOBASIS GENETICS AND SHELL MORPHOMETRICS 89 in the measures to be compared may alter the order of linkage in the network. Thus we pro- pose an alternate method of examining di- vergence that may, for small numbers of populations, approximate the Wagner net- work. The species centroid for some group of characters is the point corresponding to the average of all character measurements over some set of conspecific populations. In the case where the number of populations, n, equals three, a Wagner network is formed by connecting the populations to their centroid. The centroid can be seen as a hypothetical population, and the minimum amount of di- vergence necessary to explain observed population variation can be measured as the length of the three segments radiating from the centroid. lt is not necessary to designate the centroid or one of the surrounding points as the ancestral population if only the amount of divergence is of interest, and its direction is immaterial. When п is greater than three, the centroid connected to its surrounding populations is no longer the minimum estimate of population divergence, but it remains a (successively worsening) approximation. In this study we estimate the divergence of each population / as its euclidean distance D to the species centroid: D = (= С - x)2)2 as X\)*) where x; is the coordinate of the population on some axis x, X is the mean coordinate over n conspecific populations, and k is the number of characters or axes. We employ this tech- nique to cases where n = 4 andn = 6. RESULTS Electrophoresis Allelic frequencies at each of the ten poly- morphic loci are presented in Table 4. No vari- ability was found at five loci, Acph, Hexdh, Mah, 6Pgd, and Sdh. Table 4 also shows mean heterozygosities (H) of 15 loci for the 12 populations. Some enzyme assays, particu- larly that for Xdh, were not included until later in the study and thus have small sample sizes. Small samples may also result from reduced enzymatic activity in some popula- tions. Lap activity in particular was occasion- ally absent from individuals in the IND and ROCK populations and was entirely absent in the HILL population. But 12 snails collected from the HILL population in July, 1979, all showed strong Lap activity. They were homo- zygous for the common G. proxima allele, Lap 94. It thus seems possible that synthesis of this enzyme is under environmental control, and that all individuals from the HILL popula- tion may have the capacity to synthesize Lap 94. No “null” allele of Lap is believed to be segregating. Table 5 presents values of Nei’s (1972) genetic identity / and distance D for all pairs of populations, and Fig. 3 depicts these relation- ships graphically. Stress in the multidimen- sional scaling afer 50 iterations was 0.015, and the correlation between taxonomic dis- tance and three-dimensional scaled distance was 0.975. The first two factors account for 53.4% and 45.3% of the variance, respective- ly. The third factor (ungraphed) accounts for only 1.3% of the variance, but ROCK (Rc) is separated by a high factor Ш score and ROA (Ra) is distinguished by a low factor Ш score. Several segments of the minimum-spanning tree (cf. Sneath & Sokal, 1973) are shown connecting the major clusters. For the sake of Clarity the entire tree is not shown, but since the two dimensions graphed account for near- ly 99% of the variance, the tree’s structure is generally obvious. Details of the two major species clusters are shown in a pair of magni- fied insets to Fig. 3. In the inset each popula- tion is connected to its species centroid. The length of these segments is a measure of population divergence in the two dimensions graphed. Shell morphometrics The population means for the seven shell characters measured are presented in Table 1, along with their standard errors. The results of the three-dimensional scaling of these data are shown in Fig. 4. Once again the third fac- tor accounts for a small portion of the variance (5.6%) and is ungraphed. The YAD (Y) and PINE (P) populations both had exceptionally low factor Ill scores and are thus more distinct from conspecific populations than Fig. 4 indi- cates. The two graphed factors accounted for 77.1% and 17.2% of variance, respectively. After 50 iterations, stress was 0.003 and the correlation between taxonomic distance and three-dimensional scaled distance was 0.999. Fig. 5 shows the relationship between two measures of intraspecific divergence, genetic and shell morphometric. Divergence has been DILLON AND DAVIS 90 vs elt Lv Zl Sl 8c 8S el 9 LZ 9 0 ‘ON 00`1 76 00"! 00'1 00'1 001 96 001 86 00"! 00! 00! 66 001 001 oo! de 05 18 Zs er OL УР 89 Or 82 62 09 ol ‘ON L0'0 100 16 00 1 001 66 0 00'L 001 00'1 001 00'1 001 00"! 66 0 001 00! Чр$| 05 сс! 9$ ep GZ by ZS €9 801 $9 09 69 ‘ON +0`0 00"! 6/`0 001 001 16 86 201 L0°0 96 66 0 001 66 0 00'L Le 50° 20 16 200 66 00'1 86 0 001949 УР 06 Or ve 19 tt 29 Lt 19 LZ ct tr 08 ‘ON 910 sol 001 001 00'1 001 001 001 00"! +8 0 001 001 00'1 001 00! р999 09 ez! Or 95 GL УР 89 6r 501 LZ 95 99 ‘ON 001 00'1 001 001 86 001 001 001 001 00'1 001 201 00'! 001 00! Jev SIOH HLAM VOH HNIS OVIN 3NId diH9 1908 OVA Jar GNI TIIH see awAÁzuy e 'SISPIOIUOH jo suonejndod ZI ui 190, SıydıowAjod Je зэаюцэпбэд Sally ‘+ JIGVL 91 GONIOBASIS GENETICS AND SHELL MORPHOMETRICS 6800' 05 001 05 00 1 05 00 E v9 cO 0 86 0 SS S0'0 96'0 9200° GE 00 1 8c! 001 gel 001 99 001 gel 001 900° 9 001 9 001 9 001 9v 00 1 9v 00 E 00 0 95 00 1 9€ 00 1 er 00"! er 001 er 00 1 9200 iv 001 Sl 001 Sl 001 Sl soo S6'0 Gz 00 1 00 0 cl 00 I 9€ 001 vv 00 9c 00 1 Op 001 € OO 79 001 89 001 85 O0 E 85 LO'O 66 0 85 00"! 00ÿ0 El 00 1 сс 00"! 09 001 65 00"! 09 00"! 6150’ cl 00 1 LE 00"! 26 001 Sr ЕО LOO 1<'0 790 26 6/0 10 00'0 00"! 00 1 ¡YA 001 Let Ss! 001 8r 001 09 001 ve 90'0 v6'0 09 00 1 SJENPIAIPU! JO J8QUNU = ‘ом 9200° vl 001 ve 00'l LY 00 1 Le 001 €l 001 (H) AÁysoBÁZOJ919H ‘ON 96 86 001 чрх ‘ON co! 001 POS 56 00: шба ‘ON АЕ clk OL! 60! 80! 901 eo! 00! ЧРО G6 16 001 IdW 92 DILLON AND DAVIS estimated using euclidean distance to the species centroid as outlined in the methods section. All three dimensions were consid- ered, although only the first two have been depicted in Figs. 3 and 4. Since the pair of G. simplex populations sampled were electro- phoretically indistinguishable, G. simplex was omitted from this analysis. The Spearman rank correlation coefficient is .70, significant at the .05 level. Rank correlation is appropri- ate regardless of whether the segment lengths in the morphometric analysis are de- FIG. 3. Multidimensional scaling of Goniobasis populations based on Nei’s (1972) genetic distances. The dashed lines are segments from the minimum-spanning tree. The magnified inset on the left shows details of the Goniobasis proxima cluster, and the inset on the right shows G. semicarinata. Centroids for the two common species are graphed as open circles. Population names are abbreviated as follows: C—CRIP, Hi—HILL, Ho—HOLS, I—IND, J—JEFF, M—MEAD, P—PINE, Ra—ROA, Rc—ROCK, S—SINK, W— WYTH, Y—YAD. GONIOBASIS GENETICS AND SHELL MORPHOMETRICS 93 FIG. 4. Multidimensional scaling of Goniobasis populations based on shell morphology and Sokal's (1961) measure of taxonomic distance. Circles are G. proxima populations, triangles are G. semicarinata, and squares are G. simplex. Centroids are graphed as open circles. Abbreviations as in Fig. 3. TABLE 5. Nei's (1972) genetic identities (below diagonal) and distances (above diagonal) between Gonio- basis populations. HILL IND JEF YAD ROCK CRIP PINE MEAD SINK ROA WYTH HOLS HILL — 065 0:00, 7.051 7.154 .144 402 398 .508 .402 627 4615 IND IÓ .065 .033 .147 PAN, .397 .892 .502 396 .622 .610 JEF 1:00” 937 — 050152 .143 405 .401 .511 .405 .627 .616 YAD 950 .968 .951 — .134 .204 .326 .324 .493 .387 .609 .598 ROCK 860 .863: 859, .875 — 1150 485 .475 .602 .380 .627 .616 CRIP .866 .803 867 .816 .861 =. 510) 2505. +628) .509 627 617 PINE 669" 673, 667 .722 "616 .600 — 0.00 .143 .143 .627 .623 МЕАО 672. 676’ 670" 723.622 .603 1.00 — .140 .136 623" :620 SINK .602 .605 .600 .611 .548 534 .867 .869 — 223 .761 Ton ROA .669 .673 .667 .679 .684 .601 .866 .872 .800 — 627 .623 WYTH .534 .537 .534 .544 .534 .534 .534 .536 .467 .534 — 0.00 HOLS ‚941 543 540 .550 .540 .540 536 .538 469 .536 1.00 == termined by the additive effects of numerous Fig. 3 illustrates that populations of the three genes or slight allometric differences. species form clusters easily distinguishable from one another. Within species, however, a DISCUSSION great deal of genetic divergence has taken place. The mean Ме! identity between con- On the basis of isozyme frequencies, the specific populations, .89 + .06 (one standard three Goniobasis species are quite distinct. deviation), is quite similar to the value of .86 + 94 DILLON AND DAVIS 0.06 0.08 MORPHOLOGIC DISTANCE TO SPECIES CENTROID 0.10 0.12 0.14 0.16 GENETIC DISTANCE TO SPECIES CENTROID FIG. 5. Comparisons of two measures of population divergence, electrophoretic and morphometric. Symbols as in Fig. 4; abbreviations as in Fig. 3. .09 obtained by Chambers (1977) for Florida Goniobasis and strikingly less than the aver- age identities of conspecific animal popula- tions compiled by Avise (1976), generally .95 to .99. The degree of genetic identity of con- specific Goniobasis populations seems more comparable to that of subspecies in other ani- mals. The fixation of alternative alleles 1$ also quite unusual in conspecifics. lt seems prob- able that gene flow between populations of the same species or species-complex can be ex- tremely low, even within a single drainage system. A widely recognised shortcoming of electro- phoresis is the problem of hidden variation. Recently workers have varied gel pH and concentration and employed techniques such as heat denaturation and isoelectric focusing to show that the amount of genetic variation visible using more conventional electrophore- tic techniques is a small fraction of the actual amount (Coyne, 1976; Singh et al., 1976; Johnson, 1977). It is probable, therefore, that the values we report here for genetic similarity are systematically overestimated. However, the genetic similarities compiled by Avise (1976) were generally based on electro- phoretic techniques comparable to ours, so the comparison 1$ valid. The mean heterozygosity (H) across the 12 populations was 0.0113, a very low value compared to those of other animal popula- tions (Selander, 1976). Although many factors can influence the amount of genetic variation within populations, one likely explanation for this extremely low heterozygosity is that many of these Goniobasis populations have experi- enced a severe ‘bottleneck effect” (Nei et al. 1975). И populations are founded by small numbers of individuals and immigration re- mains low, the genetic variability of the found- ing population tends to decrease. There 15 a high probability that alleles will be lost from small populations through chance alone. Vari- GONIOBASIS GENETICS AND SHELL MORPHOMETRICS 95 ability increases by new mutations with time and population growth, but a large number of generations may be required to regain the level of heterozygosity of the founding popula- tion. So even though the Goniobasis popula- tions studied are all currently large, the low genetic variability observed in them may re- sult from a long history of “bottleneck effects.” Another possibility is that the results are due to natural selection imposed by particular en- vironmental conditions. The three species are also fairly distinct on the basis of the seven shell characters con- sidered together (Fig. 4). With the exceptions of SINK (S) and CRIP (C) species occupy separate regions of the figure. Shells from the SINK (S) population of G. semicarinata can be distinguished easily from G. proxima shells by their light brown exteriors, whitish aper- tures and the absence of carination on their whorls. G. proxima shells are dark brown to black, with a dark, often banded aperture and carinate whorls. Fig. 2 illustrates some of these distinctions. However, the СЫР (С) population of G. proxima is quite similar to G. simplex both in the shell characters measured (Fig. 4) and in color and lack of ornamentation (Fig. 2). The CRIP population was in fact con- fused with G. simplex in initial field surveys. A likely explanation for the convergence in shell characters of the CRIP population is that CRIP is the only G. proxima population col- lected from a limestone-draining stream. Fig. 1 shows that throughout the study area, G. proxima inhabits only gneiss/schist areas and other species are found only in limestone areas. Table 6 shows that alkalinities were high where G. simplex and G. semicarinata were collected and low where G. proxima was TABLE 6. Methyl orange alkalinity at 12 Goniobasis stations (phenolphthalein alkalinity in parenthesis). Station Alkalinity (ppm) HILL 11 IND 14 JEF 8 YAD 10 ROCK 6 CRIP 36 PINE 84 (10) MEAD 79 SINK 87 ROA 92 WYTH 80 HOLS 54 collected. The single exception is CRIP, a G. proxima population collected in moderately alkaline water from a creek on the border of the limestone area. With more calcium avail- able, the CRIP population has developed a larger, heavier shell than the typical G. proxima, and resembles G. simplex. The CRIP population has diverged further from conspecific populations in both shell shape and isozyme frequencies than any other population surveyed. Despite the prob- lem that the shells of different species may converge, Fig. 5 shows that divergence т shell morphology is correlated to genetic di- vergence in conspecific populations. This suggests that shell morphology has enough genetic component to be useful as a measure of evolutionary relationships among conspe- cifics. The correspondence between mor- phological and allozyme data has been exam- ined in numerous studies on other groups of animals (Schell et al., 1978). In general, little evidence of relationship has been found. However, no study to date has examined within-species divergence and employed this centroid method. Notice also in Fig. 5 that the data seem bimodally distributed on the electrophoretic axis, with a group of four “outliers.” These four populations, CRIP (C), ROA (Ra), ROCK (Rc), and SINK (S), are fixed or nearly fixed at two of ten polymorphic loci for alleles not present in other conspecific populations. Without a much larger sample of population and a much better understanding of mating systems in these snails it is impossible to know whether the four distinctive populations are different species, subspecies, or geo- graphical variants. Patton 4 Yang (1977), in discussing similar levels of electrophoretic dissimilarity in a species of pocket gopher, noted that divergence in structural gene loci does not necessarily imply reproductive isola- tion. But since the electrophoretic differences are substantial, the populations will be provi- sionally referred to as races. Race A includes HILL, IND, JEF, and YAD in Goniobasis proxima. СЕР is race В of G. proxima, and ВОСК is race С. In Goniobasis semicarinata, PINE and MEAD are race A, SINK is race B, and ROA 15 race С. In this limited sample of Goniobasis popula- tions there seems to be little evidence that geographic distance between populations т- fluences divergence. Two of the populations from outside the upper New River can be considered distinct races, and two are quite 96 DILLON AND DAVIS similar to populations found in the New River drainage. Within the New River, the pair of G. proxima populations separated by the great- est distance, HILL and JEF, are nearly indis- tinguishable. The pair of G. proxima popula- tions geographically closest, IND and CRIP, represent different races. Perhaps the cor- relation between geographic distance and genetic divergence would be evident on a smaller geographic scale. In summary, it has been demonstrated that there is great variability in the amount of di- vergence between conspecific Goniobasis populations judged by either electrophoretic or morphometric criteria. The observed fixa- tion of alternative alleles and the low hetero- zygosities are both consistent with what one might expect if interpopulation gene flow is very low, even between populations isolated only by short distances through water. How- ever, occasionally populations of Goniobasis isolated by great distances remain quite simi- lar, while those in close proximity diverge greatly. In one case an environmental vari- able, the limestone and dolomite in a drain- age, has been implicated in morphometic di- vergence. Doubtless restricted gene flow and selection both play roles in promoting radia- tion and eventual speciation in pleurocerids. A rigorous investigation into the influence of the various agents of pleurocerid evolution will require a large sample of populations, more detailed knowledge of their environ- ments, and a more thorough familiarity with their biology. The anatomy and cytology of Goniobasis are virtually unknown. Our implicit assumptions that each population of Goniobasis within a single creek is randomly breeding, and that each population experi- ences uniform environmental pressures should be tested. Work on Goniobasis cur- rently in progress addresses many of these problems. From a thorough understanding of divergence in this genus of snails inhabiting a restricted geographic area, insight may be gained regarding the process of evolution in isolated populations generally. ACKNOWLEDGEMENTS Special appreciation is due to Shary Dillon, who selflessly volunteered to rise а 5:00 a.m., grind dead snails, and wash the dirty dishes afterward. Appreciation is also expressed to Larry Hornick and Robert Dillon, Sr., for their help with the field work. The Uni- versity of Pennsylvania Biology Department provided some of the computer time and other research support. Dr. T. Uzzell furnished the program to calculate Nei statistics. All electro- phoretic analyses were performed in the labo- ratory of molecular genetics of the Depart- ment of Malacology, ANSP. Figs. 3-5 were prepared by Caryl Hesterman. This study was completed while the senior author held a Na- tional Science Foundation predoctoral fellow- ship. We thank Professor A. J. Cain and Dr. S. M. Chambers for reading and criticizing the paper. LITERATURE CITED AMERICAN PUBLIC HEALTH ASSOCIATION, American Water Works Association, and Water Pollution Control Federation, 1976, Standard methods for the examination of water and wastewater, ed. 14. Washington, 1193 p. AVISE, J. C., 1976, Genetic differentiation during speciation. In: AYALA, F. J. (ed.), Molecular evo- lution. Sinauer, Sunderland, Mass., р. 106-122. AYALA, Е. J., HEDGECOCK, D., ZUMWALT, С. $. & VALENTINE, J. W., 1973, Genetic variation of Tridacna maxima, an ecological analog of some unsuccessful evolutionary lineages. Evolution, 27: 177-191. AYALA, F. J., TRACEY, M. L., HEDGECOCK, D. & RICHMOND, В. C., 1974, Genetic differentiation during the speciation process in Drosophila. Evolution, 28: 576-592. BUSH, G. L., 1975, Modes of animal speciation. Annual Review of Ecology and Systematics, 6: 339-364. BUSH, С. |. & HUETTEL, В. N., 1972, Starch gel electrophoresis of Tephritid proteins. A manual of techniques. International Biological Pro- gramme working group on fruit flies, population genetics, project phase |, 56 р. CHAMBERS, S. M., 1977, Genetic divergence dur- ing speciation in freshwater snails of the genus Goniobasis. Ph.D. dissertation, University of Florida, 59 p. CHAMBERS, S. M.,1978, An electrophoretically detected sibling species of “Goniobasis flori- densis” (Mesogastropoda: Pleuroceridae). Mala- cologia, 17: 157-162. CHAMBERS, 5. M., 1980. Genetic divergence be- tween populations of Goniobasis occupying dif- ferent drainage systems. Malacologia, 20: 63-81. COYNE, J. A., 1976, Lack of genic similarity be- tween two sibling species of Drosophila as re- vealed by varied techniques. Genetics, 84: 609- 629. CRUTCHFIELD, P. J., 1966, Positive rheotaxis in Goniobasis proxima. Nautilus, 79: 80-86. DAVIS, G. M., 1969, A taxonomic study of some species of Semisulcospira in Japan (Meso- gastropoda: Pleuroceridae). Malacologia, 7: 211-294. GONIOBASIS GENETICS AND SHELL MORPHOMETRICS 97 DAVIS, С. M., 1979, The origin and evolution of the gastropod family Pomatiopsidae, with emphasis on the Mekong River Triculinae. Monograph of the Academy of Natural Sciences of Philadel- phia, no. 20: i-viii, 1-120. DAZO, B. C., 1965, The morphology and natural history of Р/еигосега acuta and Goniobasis livescens (Gastropoda: Cerithiacea: Pleuro- ceridae). Malacologia, 3: 1-80. DILLON, R. T., Jr., 1977, Factors in the distribu- tional ecology of upper New River mollusks (Virginia/North Carolina). Undergraduate honors thesis, Virginia Polytechnic Institute and State University, 61 p. DUSSART, G. B. J., 1976, The ecology of fresh- water molluscs in northwest England in relation to water chemistry. Journal of Molluscan Stud- les, 42: 181-193. FARRIS, J. S., 1970, Methods for computing Wagner trees. Systematic Zoology, 19: 83-92. FOIN, T. C., 1971, The distribution pattern of the freshwater prosobranch gastropod Oxytrema proxima (Say). Journal of the Elisha Mitchell Science Society, 87: 1-10. РОМ, Т. C., Jr. 8 STIVEN, A. E., 1970, The rela- tionship of environment size and population pa- rameters in Oxytrema proxima (Say) (Gastro- poda: Pleuroceridae). Oecologia, 5: 74-84. FUTUYMA, D. J., 1979, Evolutionary biology. Sinauer, Sunderland, Mass., 565 p. GOODRICH, C., 1940, The Pleuroceridae of the Ohio River system. Occasional Papers of the Museum of Zoology, University of Michigan, no. 417, 21 p. GOODRICH, C., 1941, The Pleuroceridae of the small streams of the Alabama River system. Occasional Papers of the Museum of Zoology, University of Michigan, no. 427, 10 p. GOODRICH, C., 1942, The Pleuroceridae of the Atlantic coastal plain. Occasional Papers of the Museum of Zoology, University of Michigan, No. 456, 6 p. GOODRICH, C., 1944, The Pleuroceridae of the Coosa River Basin. Nautilus, 58: 40-48. GOULD, 5. J. & WOODRUFF, D. S., 1978, Natural history of Cerion. VIII, Little Bahama Bank—a revision based on genetics, morphometrics and geographic distribution. Bulletin of the Museum of Comparative Zoology, 148: 371-415. JOHNSON, G. B., 1977, Assessing electrophoretic similarity. Annual Review of Ecology and Sys- tematics, 8: 309-328. KRIEGER, K. A. & BURBANCK, W. D., 1976, Dis- tribution and dispersal mechanisms of Oxytrema (= Goniobasis) suturalis Haldeman (Gastropoda: Pleuroceridae) in the Yellow River, Georgia, U.S.A. American Midland Naturalist, 95: 49-63. MCKILLOP, W. B. & HARRISON, A. D., 1972, Dis- tribution of aquatic gastropods across an inter- face between the Canadian shield and limestone formations. Canadian Journal of Zoology, 50: 1433-1445. MURRAY, J. 8 CLARKE, B., 1968. Partial reproduc- tive isolation in the genus Partula (Gastropoda) on Moorea. Evolution, 22: 684-696. NEI, M., 1972, Genetic distance between popula- tions. American Naturalist, 106: 283-292. NEI, M., MARUYAMA, Т. & СНАКВАВОНТУ, R., 1975, The bottleneck effect and genetic variabil- ity in populations. Evolution, 29: 1-10. PATTON, J. |. 8 YANG, 5. Y., 1977, Genetic varia- tion in Thomomys bottae pocket gophers: mac- rogeographic patterns. Evolution, 31: 697-720. ROHLF, F., KISHPAUGH, J. 8 KIRK, D., 1972, NT- SYS: Numerical taxonomy system of multivariate statistical programs. Technical Report, State Uni- versity of New York, Stony Brook, NY. ROSS, R. D., 1969, Drainage evolution and fish distribution problems in the southern Appala- chians of Virginia. In: HOLT, P. С. (ed.), The dis- tributional history of the biota of the southern Appalachians, Vol. |. Virginia Polytechnic Insti- tute and State University Research Division Monograph |, р. 277-292. SCHELL, (GD: BEST, ToL: & KENNEDY; MELE. 1978, Interspecific morphologic variation т kangaroo rats (Dipodomys): degree of con- cordance with genic variation. Systematic Zoology, 27: 34-48. SELANDER, В. K., 1976, Genic variation in natural populations. In: AYALA, Е. J. (ed.), Molecular evolution, Sinauer, Sunderland, Mass., р. 21-45. SEPKOSKI, J. M. & REX, М. A., 1974, Distribution of freshwater mussels: coastal rivers as biogeo- graphic islands. Systematic Zoology, 23: 165-188. SHAW, С. R. 8 PRASAD, R., 1970, Starch gel electrophoresis of enzymes—a compilation of recipes. Biochemical Genetics, 4: 297-320. SHOUP, C. S., 1943, Distribution of freshwater gastropods in relation to total alkalinity of streams. Nautilus, 56: 130-134. SINGH, В. S., LEWONTIN, В. С. & FELTON, А. A., 1976, Genetic heterogeneity within electro- phoretic “alleles” of xanthine dehydrogenase in Drosophila pseudoobscura. Genetics, 84: 609- 629. SNEATH, P. H. А. & SOKAL, В. R., 1973, Numer- ical taxonomy. Freeman, San Francisco, 573 p. SOKAL, R. R., 1961, Distance as a measure of taxonomic similarity. Systematic Zoology, 10: 70-79. TRYON, G. W., Jr., 1873, Land and freshwater shells of North America. Part IV. Strepomatidae. Smithsonian Miscellaneous Collections, no. 253: i-lv, 1-435. WHITE, M. J. D., 1978, Modes of speciation. Freeman, San Francisco, 455 p. APPENDIX The following are locality data for the 12 Gonio- basis populations studied. “Quad” refers to United States Geological Survey topographic maps, 7.5 98 DILLON AND DAVIS minute series. Map coodinates are approximate. The species of Goniobasis collected at each site is followed by the catalogue number of voucher speci- mens deposited in the Academy of Natural Sci- ences of Philadelphia (ANSP). CRIP—G. proxima, ANSP 349362. Cripple Creek at Va. 749 bridge, 8km E of Cedar Springs, Wythe County, VA. Cedar Springs Quad. 36°55'N; 81°19'W. HILL—G. proxima, ANSP 349363. Tiny creek at U.S. 58 bridge, near junction with Va. 820. 4 km E of Hillsville, Carroll Co., VA. Hillsville Quad. 36°46’N; 80°44’W. HOLS—G. simplex, ANSP 349364. Dry Run at Va. 617 bridge, 1.6 km NW of Groseclose, Smythe Co., VA. Rural Retreat Quad. 36"56'N; 81°24’ W. IND—G. proxima, ANSP 349365. Brush Creek at crossroads of U.S. 21 and Va. 701, 4.8 km S of Independence, Grayson, Co., VA. Sparta West Quad. 36°38'N; 81°11’W. JEF—G. proxima, ANSP 349366. Cranberry Creek by U.S. 221 at Co. 1145 bridge, 11 km М of Deep Gap, Ashe Co., NC. Todd Quad. 36°14’М; 81°32'W. МЕАО— С. semicarinata, ANSP 349367. Meadow Creek at Va. 787 bridge, 3.2 km N of Grayson- town, Montgomery Co., VA. Radford South Quad. 37°05'N; 80°33'W. PINE—G. semicarinata, ANSP 349368. Little Pine Run at Va. 100 bridge, near Pine Run Church. Pulaski Co., VA. Fosters Falls Quad. 37°01'N; 80°33’W. ROA-G. semicarinata, ANSP 349369. Mill Creek, .8 km upstream from mouth, .3 km of Bennetts Mill, Montgomery Co., VA. McDonalds Mill Quad. 37°15'N; 80°19’W. ВОСК— С. proxima, ANSP 349370. Tiny branch of Rock Castle Creek at Va. 8 bridge, 4 km NW of Woolwine, Patrick Co., VA. Woolwine Quad. 36°39'N; 80°22'W. SINK—G. semicarinata, ANSP 349371. Sinking Creek at Newport Park, Newport, Giles Co., VA. Newport Quad. 37°20'N; 80°29'W. WYTH—G. simplex, ANSP 349372. Mill Creek downstream from confluence with Нидае Branch by Va. 680, 4.7 km N of Rural Retreat, Wythe Co., VA. Rural Retreat Quad. 36°54'N; 81°18'W. YAD—G. proxima, ANSP 349373. Small creek at Lewis Fork Road Bridge (at crossroads of Co. 1155 and 1156), 3.2 km W of Mount Pleasant, Wilkes Co., NC. Purlear Quad. 36°04’М; 81°29'W. MALACOLOGIA, 1980, 20(1): 99-116 DEVELOPMENT AND FEEDING OF LARVAE OF THE NUDIBRANCH GASTROPODS HERMISSENDA CRASSICORNIS AND AEOLIDIA PAPILLOSA Leslie G. Williams Pacific Marine Station, Dillon Beach, California 94929, U.S.A.' ABSTRACT Preliminary observations of developmental sequence of two aeolid nudibranchs, Hermis- senda crassicornis and Aeolidia papillosa, indicated that planktotrophic larvae of both species hatch from their egg mass with yolk reserves in the stomach and digestive diverticulum. Com- parison of development and feeding of these two species was undertaken in order to evaluate the relative importance of yolk reserves and of phytoplankton to larval nutrition and life history. Early embryology and larval development to hatching is similar in the two species. Cleavage is characterized by the appearance of an asynchronous three cell stage and, otherwise, follows the typical spiral pattern of the Mollusca and results in the formation of a stereoblastula. Gastrulation is characterized by the formation of a sagittal cleft. Subsequent maturation through the transitory trochophore and early veliger stages leads to hatching of free-swimming veliger larvae seven and eight days (14°C) after fertilization for Hermissenda crassicornis and Aeolidia papillosa, respectively. The veliger larvae of both species are dextrally organized and show planktotrophic morphol- ogy. Larval shells of both species are coiled and characterized by six to seven oblique striations on the whorl and inner lip next to the aperture. However, mean shell length of Aeolidia papillosa larvae is significantly greater than that of Hermissenda crassicornis larvae (116 and 102 um, respectively). At hatching each species 15 polytypic with respect to relative yolk reserves and feeding ability. In each case, shell length distributions of yolk-laden and yolk-free larvae were normal with mean shell length of yolk-free larvae being significantly greater than that of their yolk-laden siblings. Feeding ability of recently hatched larvae of each species is determined by the presence of yolk reserves, the size of larvae, and the size of algal cell available for ingestion. With few exceptions, yolk-laden larvae did not ingest algal cells offered as food while their larger, yolk-free siblings did so easily. Mean shell length of newly hatched larvae capable of feeding increased with increased size of algal cell offered as food. For instance, shell length of A. papillosa larvae feeding on Chlorella is significantly less than the shell length of larvae feeding on the larger alga Phaeo- dactylum tricornutum. Larval feeding is governed by two main factors: 1) physical limitation imposed by their mechanical ability to handle large particles, and 2) selection for the largest particles they are capable of ingesting. Larvae of H. crassicornis and A. papillosa are functionally, morphologically, and, in all likeli- hood, morphogenetically planktotrophic. Neither species is unusual in this regard. The presence of yolk reserves at hatching allows a portion of the larval population temporary independence of phytoplankton and, perhaps, increased dispersal. INTRODUCTION Thompson (1967) employed morphological as well as embryological criteria in his defini- tion of planktotrophic, lecithotrophic, and di- rect sequences of development among opis- thobranchs. Planktotrophic larvae develop rapidly from numerous, small (yolk-poor) eggs, and lack certain adult or advanced lar- val characteristics necessary for successful metamorphosis at hatching. Lecithotrophic larvae develop slowly from few, large (yolk- laden) eggs, and emerge from the egg case endowed with somatic features that anticipate the adult mode of life and therefore render them competent to metamorphose. The mor- phological features, which occur in lecitho- tropic larvae but not in premetamorphic plank- totrophic larvae, are a rudimentary radula, an adult kidney vesicle, eyespots, and propodi- Present address: University of Delaware, College of Marine Studies, Lewes, Delaware 19958, U.S.A (99) 100 WILLIAMS um. Direct development 15 accomplished through ametamorphic embryogenesis or metamorphosis of an encapsulated veliger, and results in liberation of a juvenile (Bonar, 1978). The major explicit assumption in Thomp- son's (1958, 1962, 1967) and Tardy's (1970) classification of development types 15 that feeding and subsequent growth are neces- sary prerequisites for metamorphosis of planktotrophic larvae. More recent work on planktotrophic development of Phestilla melanobranchia (Harris, 1975), Doridella obscura (Perron & Turner, 1977), and Hermissenda crassicornis (Harrigan & Alkon, 1978) reaffirms the obligatory feeding of this type of larvae for continued somatic growth and consequent metamorphosis to oc- cur. By comparison, lecithotrophic larvae are initially yolk-laden and presumably capable of metamorphosis without feeding. Lecitho- trophic larvae may feed facultatively if deple- tion of yolk reserves occurs during search for a substratum necessary to induce metamor- phosis. The relationships among larval feed- ing, larval morphology, and metamorphosis are clearly important in assessing the selec- tive pressures which have led to ecologically distinct life history patterns. Preliminary observations of developmental sequence of two aeolid nudibranchs, Hermis- senda crassicornis (Eschscholtz) and Aeolidia papillosa (L.), were interesting be- cause each species produced planktotrophic larvae with yolk reserves in the stomach or digestive diverticulum, a phenomenon more characteristic of lecithotrophic than plankto- trophic larvae. Comparison of development and feeding of these two species was under- taken to assess the importance of yolk re- serves to nutrition and life history of plankto- trophic larvae. The first objective of the study was to document the planktotrophic charac- teristics of the larvae using embryological (developmental sequence and rate) as well as morphological (eyespots, radula, propodium) criteria. The second objective was to assess the importance of yolk reserves through direct observation of functional morphology of feed- ing behavior and through several feeding ex- periments. MATERIALS AND METHODS Collection and culture of egg masses Adult Hermissenda crassicornis was col- lected from beds of Zostera marina Linnaeus located on Lawson's Flat, Tomales Bay, Cali- fornia. Adult Aeolidia papillosa was collected from floating docks at Mason's Marina and the north jetty т Bodega Harbor, California. Adults of both species were isolated in plastic aquaria and supplied with running seawater in the laboratory. Н. crassicornis laid egg masses in the laboratory during the entire year of study (August, 1970 to July, 1971) while A. papillosa egg masses were laid only from May through September 1971, the peri- od of availability of the adult animals. Development was followed for 50 H. crassicornis and for 25 A. papillosa egg masses spawned and cultured in the labora- tory. The extreme abundance of H. crassi- cornis and its spawn on Lawson's Flat al- lowed additional observations on approxi- mately 120 egg masses collected in the field. Egg masses deposited on the sides of aquaria were removed and cultured in 500 ml flasks at 14.0°С. The egg masses and hatched veligers were washed daily on a 45 um mesh screen and returned to clean culture flasks containing filtered seawater. Hatched larvae were fed mixed suspensions from algal cultures of Pavlova lutheri Droop, Isochrysis galbana Parke, Dunaliella tertio- lecta Butcher, Chlorella clone 580, and Phaeodactylum tricornutum Bohlin. Algae were cultured in Guillard 8 Ryther's (1962) f/2 concentration of nutrients at 17°С. Developmental sequence Observations and photographs of develop- mental sequence and of encapsulated larval stages were made on living specimens, with a compound microscope using phase contrast and bright field optics. Egg capsules and embryos were prepared for observation by teasing out a portion of an egg mass and plac- ing it under a cover slip with a drop of sterile seawater. After examination this preparation was either discarded or placed оп moist toweling to prevent dessication and returned to the incubator for later retrieval and re- examination. Larval feeding Functional morphology. Upon hatching, veliger larvae were given suspensions of Pavlova lutheri (4.1 um diameter), /sochrysis galbana (4.2 um), Dunaliella tertiolecta (9.4 um), Chlorella 580 (4.4 um), and Phaeo- dactylum tricornutum (12.5 to 17.8 ит along the longest axis) to determine which of the five DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 101 algal species they would eat. Photographic ob- servations of manipulation of algal particles by the cephalopedal and alimentary cilia were made with a 35 mm single lens reflex camera mounted on a Labolux Leitz equipped with carbon-arc illumination and a trinocular head. Feeding experiments. An experiment was performed to determine the residence time of algal cells in the stomach of Hermissenda crassicornis larvae. Recently hatched larvae were fed a combined suspension of P. lutheri, I. galbana, and D. tertiolecta. After 48 hours the larvae were transferred from the algal sus- pension to clean, filtered seawater. Larvae were then sampled (n = 100) seven times during the 53 hours following their removal from the algal suspension, and examined for the presence of algal cells or pigmentation in the left digestive diverticulum. Preliminary observations of shell length suggested that H. crassicornis were size di- morphic with respect to the presence or ab- sence of yolk reserves at hatching. Therefore, a series of observations of H. crassicornis larvae hatched from the same egg mass was made in order to test the following null hy- pothesis: 1) there is no difference in size be- tween yolk-laden and yolk-free larvae at hatch- ing, 2) there 15 no difference in size between yolk-free larvae and larvae able to feed, 3) there is no difference in size between yolk- laden and nonfeeding larvae, and 4) there 1$ no difference in size between larvae feeding on small algal cells and those feeding on large cells. At the beginning of the experiment 100 recently hatched larvae were examined for the presence of yolk reserves in the left digestive diverticulum. The right digestive diverticulum was not examined since it is not used in digestion of algal cells by nudibranch larvae observed by Thompson (1959) as well as by those used in this study. Independent samples were taken and shell length meas- urements made by ocular micrometer were then recorded for 20 larvae with and 20 larvae without yolk reserves in the diverticulum. Ad- ditional independent and random samples were taken to measure shell length (n = 100), shell width (n = 27), and shell height (n = 39) without consideration to the presence or ab- sence of yolk reserves. Approximately 1000 larvae were than added to each of two flasks containing 500 ml of sterile, filtered seawater containing f/2 concentration of nutrients. Salinity in each flask was 33.07... A suspen- sion of D. tertiolecta was added to one of the flasks while a suspension of Chlorella 580 was added to the other. The flasks were then placed in an incubator at 14.0 + 0.5°С for 72 hr. Continuous illumination was provided dur- ing the experiment with a 40 watt daylight type fluorescent lamp. At the termination of the ex- periment 100 larvae from each flask were ex- amined for the presence of algae in the left diverticulum. Independent samples of larvae were taken from each flask and shell length measured for 20 larvae with and 20 larvae without algae in the left diverticulum. There are several assumptions involved in this experimental design. The first concerns the definition of yolk-laden larvae. The criteri- on 1$ the presence of yolk cells in the digestive diverticulum, though the esophagus and stomach may be free of yolk cells. The reason for this definition is that algal cells ingested by a larvae must enter the digestive diverticulum for digestion. At the end of the experimental period, larvae with algal cells in the stomach or diverticulum were considered to have been feeding. А more restrictive definition of feed- ing ability, and one more consistent with the definition of yolk-laden larvae, would require the presence of algal cells in the diverticulum. However, the less restrictive definition was employed when it was difficult to view the di- gestive diverticulum from the right side of the larvae because the stomach was filled with algal cells. Problems in interpretation of the data arise because yolk-laden and yolk-free larvae were not measured at the end, as they were at the beginning, of the experiment. Thus, there is uncertainty as to the amount of growth of larvae and the degree to which yolk reserves are metabolically diminished during the experiment. Therefore, equating yolk- laden with nonfeeding larvae, and yolk-free with feeding larvae, rests on association with the shell length distributions that characterize each group at the beginning and end of the experiment, respectively. Frequency distributions of shell length were initially tested for normality by graphic tech- niques (e.g., Rankits) described by Sokal 4 Rohlf (1969). Since these methods indicated potentially non-normal distribution of shell length, the more quantitative Kolmogorov- Smirnov statistic was calculated to test for goodness of fit of observed distribution of shell length with the normal distribution (Sokal & Rohlf, 1969). Multiple comparisons of means and tests for significant differences among experimental groups were performed using the Student-Newman-Keuls procedure as summarized by Sokal & Rohlf (1969). The Student-Newman-Keuls procedure is an a posteriori method that permits all possible 102 WILLIAMS comparisons among means, and accommo- dates unequal sample sizes as well. An experiment, similar to that described for H. crassicornis larvae, was performed for recently hatched larvae of Aeolidia papillosa. The four hypotheses, assumptions, and caveats concerning experimental design de- scribed for H. crassicornis apply to A. papil- losa as well. The larvae used in the experiment were newly hatched, and taken from a com- mon egg mass. At the beginning of the experi- ment 100 larvae were examined for the pres- ence of yolk reserves in the left digestive di- verticulum, and independent samples of shell length measured for 50 larvae with and 50 larvae without yolk reserves. Additional in- dependent samples were taken to measure shell length (n = 100), shell height (n = 30) and shell width (n = 30). Approximately 1000 larvae were then added to each of four flasks containing 500 ml of sterile seawater and f/2 concentration of nutrients (S°/ = 33.0). Uni- algal suspensions of Chlorella 580, D. terti- olecta, or P. tricornutum were added to the first three flasks. The fourth flask received a mixed suspension of Chlorella and P. tri- cornutum. The four flasks were then incu- bated at 14.0°С under continuous illumination for 72 hr. At the termination of the incubation period, 100 larvae from each flask were ex- amined for the presence of algal cells or pig- mentation in the left digestive diverticulum. Independent samples of larvae were then withdrawn from each flask and shell length recorded for 30 larvae with and 30 larvae without algal cells in the left diverticulum. Kolmogorov-Smirnov tests for goodness of fit of observed with predicted, normal, distribu- tions of shell length were performed as out- lined by Sokal & Rohlf (1969). Means of ех- perimental groups were compared using the Student-Newman-Keuls procedure (Sokal & Rohlf, 1969). RESULTS Egg masses and gametes Both Hermissenda crassicornis and Aeolidia papillosa have white, spirally coiled egg masses with secondarily twisted egg Strings (Hurst, 1967, Type B). Egg strings consist of individual egg capsules, which are elliptical and smooth in H. crassicornis and are irregularly shaped and wrinkled in A. papillosa. Occasionally the egg capsules do not separate during spawning, but are joined end to end by either a simple constriction in the case of H. crassicornis or by a twisted constriction in the case of A. papillosa. The number of ova per capsule was meas- ured for 17 egg masses deposited by H. crassicornis. There is a significant (р < 0.01) difference among egg masses for the number of ova per egg capsule with a coefficient of intraclass variation of 0.535. The ova of H. crassicornis and A. papillosa are small but significantly (р < 0.001) differ- ent in diameter from one another. The mean diameter and standard deviation of H. crassi- cornis ova is 64.7 + 1.6 um (п = 29) while the mean diameter of А. papillosa ova is 73.7 + 2.2 um (п = 25). Length of the spermatozoa in each species is approximately 150 ит. Developmental sequence The major developmental features and time to hatching of free-swimming veliger larvae of Hermissenda crassicornis and Aeolidia papil- losa are summarized in Table 1. Separation of polar bodies occurs at the animal pole in both H. crassicornis and A. papillosa, and completion of the maturation divisions results in the formation of two polar bodies in each species (Fig. 1). Division of the first polar body occurs shortly after the first TABLE 1. Developmental time to hatching in Hermissenda crassicornis and Aeolidia papillosa cultured in seawater at 14.0°С. Cumulative hours Stage H. crassicornis A. papillosa 1st polar body 15 2.0 2nd polar body 3.4 5.0 1st cleavage 4.9 6.0 2nd cleavage 6.1 7.5 3rd cleavage 7.0 9.5 4th cleavage 9.3 13.0 morula 12.5 25:0 blastula 237 31.0 early gastrula 31.0 45.0 late gastrula 51:0 51.0 trochophore 65.0 82.0 shell gland in- vagination 75.0 88.0 shell gland evagination 83.0 95.0 early veliger 101.0 118.0 Alimentary dif- ferentiation 135.0 160.0 hatching 166.0 195.0 DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 103 cleavage in A. papillosa but not in H. crassi- cornis. The polar bodies remain intact through early veliger stage of development and often through hatching of the larvae. Cleavage 1$ similar for both H. crassicornis and A. papillosa. The first cleavage 1$ holo- blastic and equal with the cleavage plane in the polar axis. The second cleavage is asynchronous, holoblastic and equal, with the cleavage plane in the polar axis at right angles to the first cleavage plane. Asynchrony at this state is characterized by division of one of the blastomeres to form a transitory three cell stage with the second blastomere dividing approximately 10 minutes after the first (Fig. 2). When viewed from the animal pole, divi- sion of the second blastomere is character- ae a D ES ds a RIG: 1. (a) Zygote and first and second polar bodies of Hermissenda crassicornis. (b) Zygote and first and second polar bodies of Aeolidia papillosa. Ar- row points to cleavage furrow of first polar body. 25ит = FIG. 2. Intermediate three cell stage during asyn- chronous second cleavage of Hermissenda crassicornis. ized by counterclockwise displacement of the daughter blastomere into a slightly lower focal plane. Thus the second cleavage appears to be laeotropic. The third cleavage is dexio- tropic, holoblastic and equal. The fourth and ensuing cleavage divisions are typical of the molluscan pattern of alternating spiral cleav- age. Cleavage in each species results in the formation of a stereoblastula. Gastrulation is distinguished by formation of a blastoporal sagittal cleft (Fig. 3). During late gastrulation, the sagittal cleft closes in a zipper-like fashion beginning near the animal pole and terminat- ing with the closure of the definitive blasto- pore at the vegetal pole. Gastrulation results in a stereogastrula. Following gastrulation, the trochophore stage is recognized by development of the prototroch (Fig. 3). In each species, appear- ance of the prototroch is followed rapidly by formation of two posterior anal cells and the formation of the shell gland invagination posteriolaterally on the left side (Fig. 3). The prototroch subsequently differentiates into two rudimentary velar lobes and is followed by evagination of the shell gland (Fig. 4) and by formation of the metapodium. As the shell gland spreads to circumscribe the viscera, the anal cells are displaced anterolaterally to the right side. The developing embryos are early veliger larvae at this point. The shell gland secretes the definitive larval shell and then differentiates into the mantle fold and peri- visceral membrane. 104 WILLIAMS 50 u FIG. 3. Hermissenda crassicornis. Schematics of gastrulation (a), shell gland invagination (b), and spreading of the shell gland over the dorsum (с). Abbreviations: bp., blastopore; s.c., sagittal cleft; s.g., shell gland; s.g. invag., shell gland invagina- tion. Transition to a fully developed veliger larva is characterized by continued shell growth, invagination of the stomodaeum, differentia- tion of alimentary and larval excretory organs, formation of retractor and mantle muscles, deposition of an operculum, and appearance of bristle-like pedal cilia at the junction of the foot and operculum. Larval shells of H. crassicornis and A. papillosa are characterized by oblique stria- tions (Fig. 5) on the whorl and inner lip next to the aperture. These striations are present at the initial formation of the shell. During the early veliger stage, the striations extend from the ventral surface of the whorl to the right dorsolateral surface of shell. Hatching Hatching of mature veligers occurs simi- lary in both Hermissenda crassicornis and Aeolidia papillosa. As larvae near hatching, FIG. 4. Hermissenda crassicornis at trochophore stage. (a) Darkened area shown by arrow is the shell gland invagination. (b) Raised area shown by arrow is the shell gland evagination. they become increasingly active, the capsule membrane becomes thin and flaccid, and in the case of A. papillosa, loses its wrinkled appearance. Larvae beat their velar cilia against the capsule membrane with the ef- fective stroke forward over the velum and the direction of larval movement shell first or posteriorly. Hatching occurs by rupture of the capsule membrane along the interface be- tween it and the shell. Larvae exit by backing through the rupture. Free-swimming larvae Larvae of Hermissenda crassicornis and Aeolidia papillosa show similar gross mor- phological characteristics at hatching. The hyperstrophic larval shell is the coiled, Type 1, DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 105 FIG. 5. Veliger of Hermissenda crassicornis prior to hatching. Arrow points to characteristic striations located on whorl of the shell. variety described by Thompson (1961). The cephalopedal region in each species 1$ char- acterized by moderately sized velar lobes, prominent sub-velar ridges, and bristle-like lateral and apical cilia located at the junction of the foot and operculum. Internal organization is dextral with the anus opening into the right side of the mantle cavity immediately behind the larval kidney. The intestine of both species is distinguished by the presence of superficial lobes along its length. The larvae otherwise conform to Thompson's (1967) definition of plankto- trophic veligers: they lack eyes, radula, propodial rudiment, and adult kidney vesicle (Fig. 6). Larval feeding Functional тогрпоюду. The mechanical treatment of food is similar in Hermissenda crassicornis and Aeolidia papillosa. Patterns of ciliary metachronism in H. crassicornis and A. papillosa agree with those observed by Fretter (1967) for prosobranch veligers. Velar cilia are diaplectic, and beat in a regular, metachronal fashion. If viewed from the ventral side of the veliger, the metachronal 50 u H— = | FIG. 6. Schematic of free-swimming veliger of Hermissenda crassicornis. Note lack of eyes, radula, and the absence of a propodium. Abbrevia- tions: ap. cl., apical cilia; int., intestine; Idd., left digestive diverticulum; Ik., larval kidney; mf., mantle fold; mm., mantle muscles; op., operculum; rdd., right digestive diverticulum; rm., retractor muscles; st., statocyst; stom., stomach; str., striations; vel., velar cilia. waves of both velar lobes move in a clockwise direction with the effective stroke being back- wards over the velum. Velar cilia demonstrate laeoplectic metachronism (Knight-Jones, 1954). The effective stroke of the post-oral сша, on both velar lobes, is towards the mouth. Algal cells trapped by the post-oral cilia are transported to the mouth and either manipulated into the esophagus by oral cilia or are passed onto the anterior base of the foot for rejection. The effective stroke of the pedal cilia is towards the apex of the foot. Algal cells rejected by the oral cilia are con- ducted away from the larva via the long axis of the foot. Pedal cilia also function in rejection of algal cells displaced over the velum but not trapped by the postoral cilia. Cilia along the right side of the foot adjacent to the oper- culum form a strong rejection current for passage of feces and whole undigested algal cells from the anus to the apex of the foot. No apparent functional utilization of the bristle- like cilia located at the sides and apex of the foot was observed. Algal cells in the stomach are rotated against denticulate rods em- bedded in the stomach wall, and manipulated by cilia into the digestive diverticulum (Thompson, 1959). Cellular debris and undi- gested algal cells are transported along an anterior food groove to the confluence of the stomach and intestine (Fig. 7). Metachronal 106 WILLIAMS int. ant. Idd. FIG. 7. Schematics of larval stomach of Hermissenda crassicornis. a) Profile of stomach and diverticulae viewed from right side. Note denticulate rods located in posteroventral wall of stomach. Arrows indicate direction of motion of algal cells during passage through esophagus, diverticulum, and stomach. b) Posterior view of stomach and left digestive diverticulum. Densely stippled vertical region represents anterior food groove, and arrows indicate movement of cellular debris toward confluence of stomach and intestine. Abbreviations: ant. fg., anterior food groove; int. intestine; Idd., left digestive diverticulum. waves in the intestine move from the anus toward the stomach while the effective stroke of the cilia is towards the anus. Intestinal cilia, therefore, have an antiplectic pattern of metachronism (Knight-Jones, 1954). Feces and whole undigested algal cells are moved along the intestine, defecated through the anus and conducted away from the larva along the pedal rejection current. Newly hatched larvae of H. crassicornis and A. papillosa readily ingested cells of Pavlova lutheri, Isochrysis galbana, Chlorella 580, and Dunaliella tertiolecta. The diatom Phaeodactylum tricornutum was readily in- gested by the larvae of A. papillosa but not by those of H. crassicornis. Newly hatched larvae fed voraciously when placed in a suspension of algae, often feeding until the stomach, diverticulum and esophagus become packed with ада! cells. In such in- stances of over-feeding, larvae stopped in- gesting algal cells and began defecating whole undigested cells until the esophagus was Cleared and algal cells were present only in the stomach and the diverticulum. There- after feeding occurred sporadically with most cells passed over the velum or rejected by the Oral cilia. In cases where larvae were not ini- tially overfed, they cleared cells from suspen- sion until both the stomach and diverticulum contained algal cells. Feeding then occurred sporadically. Feeding experiments. Data on the resi- dence time of algal cells in the stomach of H. crassicornis larvae are presented in Fig. 8. Ap- proximately 80% of the larvae that had initially fed on the algal suspension retained algal cells in the left diverticulum for 53 hours after removal from the suspension. Shell measurements of randomly sampled larvae of H. crassicornis and A. papillosa used in feeding experiments show that differ- ences between the two species are mor- phometric rather than morphological (Table 2). Larval shells are cup-shaped at hatching with a ratio of shell length:height:width of 1.28:1.04:1.00 for H. crassicornis and 1.36: 1.01:1.00 for A. papillosa. Frequency distribu- tions of shell length at hatching appear bi- тоаа! (Fig. 9). However, Kolmogorov-Smir- nov tests of deviation of observed distribution from normality (dmax, Table 2) were not signif- icant for either Hermissenda crassicornis (0.1 < р< 0.2) or Aeolidia papillosa (p > DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 107 [OO 90 80 In Gut (%) 70 Larvae With Algae а 20° 3G Time (hours) FIG. 8. Number of larvae of Hermissenda crassicornis with algal cells in stomach or left digestive diverticu- lum plotted against number of hours after their removal from a suspension of Pavlova lutheri, Isochrysis galbana and Dunaliella tertiolecta. 40 50 60 TABLE 2. Means and standard deviations of shell length, shell width and shell height of newly hatched larvae of Hermissenda crassicornis and Aeolidia papillosa. The remaining columns summarize departure from normality (Kolmogorov-Smirnov statistic, dmax), skewness (91), and kurtosis (92). N Х(ит) = 5 dmax 91 92 Hermissenda crassicornis Length 100 102.1 10.03 0.110 +0.251 —0.412 Height 27 83.1 8.77 — = = Width 39 79.7 6.1 — _ = Aeolidia papillosa Length 100 116.0 10.04 0.100 — 0.5683” —0.381 Height 30 85.8 8.27 — — — Width 30 7.42 — — = *Significant at the 0.05 probability level. 0.2). Distribution of shell length for A. papil- losa was significantly skewed to the left (p < 0.05, Table 2). Mean shell length and stand- ard deviation at hatching is 102.1 + 10.03 ит (n = 100) for H. crassicornis and 116.0 + 10.04 um (п = 100) for A. papillosa. Also 40— 60% of newly hatched H. crassicornis and 80-90% of newly hatched A. papillosa larvae have yolk reserves in the stomach or left di- gestive diverticulum at hatching. Frequency distributions of shell length of larvae of H. crassicornis that hatched out with and without yolk reserves are shown in Fig. 10 and distributions of shell length of larvae fed suspensions of Chlorella and D. tertio- lecta are shown in Figs. 11 and 12. Means and standard deviations of shell length for these larvae are summarized in Table 3. Only those larvae that had fed upon D. tertiolecta showed a significant deviation in the observed distribution of shell length from that predicted by a normal distribution (dp ax, Table 3). There are relatively fewer larvae feeding on algal suspensions (41%) than there were yolk-free 108 20] A Hermissenda crassicornis © O = © + 80 90 00 ПО 120 130 140 20] В. Aeolidia papillosa [Ss] D с = O 104 + 54 10) 2 1 80 90 00 ПО 20 130 140 Shell length (um) FIG. 9. Frequencies of shell length of Hermissenda crassicornis and Aeolidia papillosa at hatching (n = 100). A. Feeding larvae (cells present) # larvae m O 90, ¡002101201501 11140 @ O B. Non-feeding larvae (cells absent) # larvae 140 130 ЮО’ МЮ’ "120 Shell length ( um) FIG. 11. Frequency distribution of shell length of larvae of Hermissenda crassicornis fed a suspen- sion of Chlorella 580 (n = 20). WILLIAMS 87 д Yolk-free larvae © E # larvae > 24 О 80 90 100 110 120 130 140 a 67 В Yolk-laden larvae oO = в + 2 4 О 80 90 100 НО 20 130 140 Shell length (um) FIG. 10. Frequency distribution of shell length at hatching for Hermissenda crassicormis with and without yolk reserves (n = 20). 107 A. Non-feeding larvae (cells absent) 81 67 41 21 0 80 # larvae 90 100 10 120 130 140 104 В. Feeding larvae (cells present) 85 64 # larvae > x 100 ПО [20 5.6» AIO Shell length (um) FIG. 12. Frequency distribution of shell length of larvae of Hermissenda crassicornis fed a suspen- sion of Dunaliella tertiolecta (n = 20). DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 109 TABLE 3. Means and standard deviations of shell length (um) of veliger larvae of Hermissenda crassicornis employed in a 72 hour feeding experiment. Samples based on the presence or absence of yolk reserves were taken at the beginning of the experiment. Two groups of larvae were then separated from the culture. One group was fed a suspension of Chlorella 580 while the second group was fed a suspension of Dunaliella tertiolecta. At the end of the experiment shell length was measured for larvae that had ingested cells and that had not ingested cells from each of the suspensions. Percentages (n = 100) of larvae with yolk, Dunaliella or Chlorella cells in the stomach or diverticulum are given in parentheses. Kolmogorov-Smirnov test for de- parture of the observed frequency distribution from normality is summarized under the column labeled dmax- Third and fourth moment statistics indicating either skewness or kurtosis of the observed frequency distribu- tions are Summarized in the columns labeled 91 and go, respectively. N t(hr) Х(ит) = 5 Чтах 91 92 Yolk-free larvae 20 0 108.8 4.19 0.254 +3.386 1.235 Yolk-laden larvae (49%) 20 0 99.3 5.22 0.084 — 0.374 —0.701 Dunaliella fed to larvae cells in stomach (41%) 20 72 108.9 7.96 0.232 +0.698 +0.058 cells not in stomach 20 72 94.4 9.61 0.305* +0.643 —0.753 Chlorella ted to larvae cells in stomach (41%) 20 72 104.4 7.00 0.131 =1.122 +1.745 cells not in stomach 20 72 103.8 9.31 0.152 —0.422 —1.244 “Indicates significance at the 0.05 probability level. TABLE 4. Student-Newman-Keuls test for multiple comparisons among means of shell length (ит) of larvae of Hermissenda crassicornis employed in the feeding experiment outlined in Table 3. Significant differences among means are based on comparison with the Least Significant Range (LSR) statistic calculated accord- ing to the procedures outlined by Sokal & Rohlf (1969). The group means from Table 3 are ranked from lowest to highest. Abbreviations for each group are YF (yolk-free), YL (yoik-laden), CF (Chlorella cells present), CNF (Chlorella cells absent), DF (Dunaliella cells present), and DNF (Dunaliella cells absent). Those means that are not significantly different from one another (p > 0.05) are shown by underscoring. Rank 1 2 3 4 5 6 Group DNF YL CNF CF YE DF Mean (м) 94.4 99.3 103.8 104.4 108.8 108.9 larvae at the beginning of the experiment (51%). Secondly, the relative number of lar- уае feeding on Chlorella was the same as that for D. tertiolecta. Multiple comparisons of means, and tests for significant differences among experimental groups show that at the beginning of the experiment (t = O hr) yolk- free larvae were significantly larger than those with yolk reserves (t-test, р < 0.001, and Table 4). Similarly, larvae that had ingested D. tertiolecta were larger than those that had not fed. There was no significant difference (p > 0.05) between yolk-free larvae at the be- ginning of the experiment and larvae that had ingested cells of either Chlorella or D. tertio- lecta by the end of the experiment. Frequency distributions of shell length for larvae of A. papillosa with and without yolk reserves at hatching are shown in Fig. 13. Distributions of shell length for larvae fed sus- pensions of Chlorella, D. tertiolecta, P. tri- cornutum, and Chlorella + P. tricornutum are shown in Figs. 14, 15, 16 and 17, respectively. The distributions of shell length of A. papillosa larvae with and without yolk reserves are not significantly different (p > 0.05) from a normal distribution while the distributions of shell length for those larvae which cleared Chlor- ella or P. tricornutum from suspension were significantly skewed (p < 0.05) to the left (negative g;,). Mean shell lengths and stand- ard deviations for these larvae are summar- ized in Table 5. Unlike H. crassicornis, there is little difference between the relative numbers of larvae which were yolk-free and those which could clear algae from suspension. The rela- tive numbers of yolk-free and of feeding lar- vae of A. papillosa were much less than those of H. crassicornis. Multiple comparisons of larval shell lengths of the various experi- 110 WILLIAMS 107 A. Yolk-free larvae 81 g 6 2 с == 44 th 24 AE М 80 90 100 НО 120 130 140 107 В. Yolk-laden larvae ® oO = D + 80 90 ООО 120 71507 140 Shell length (um) FIG. 13. Frequency distribution of shell length at hatching of larvae of Aeolidia papillosa with and without yolk reserves (п = 50). mental groups are presented in Table 6. At the beginning of the experiment, yolk-free lar- vae were significantly larger than their yolk- laden siblings (t-test, p < 0.001, and Table 6). Larvae that cleared algae from suspension were significantly larger (p < 0.05) than those that had no algal cells in the stomach or diver- ticulum. There is no significant difference in shell length between larvae with yolk reserves at the beginning of the experiment and those which did not clear D. tertiolecta from sus- pension. Comparisons of shell length of lar- vae that had fed on the various algal suspen- sions show a significant difference between those which had consumed Chlorella and those that had ingested P. tricornutum. In a mixed suspension of Chlorella and P. tri- cornutum, larvae less than 120 um in shell length consumed only Chlorella while larvae greater than 120 шт consumed both species of algae. Larvae ingesting both species of 10 | A. Feeding larvae (cells present) 8 # larvae ^ L >. | of HE 80,99 I00: „,119: 120 SOR Ise # larvae 80, ¿90 100 10 120 130 140 Shell length (um) FIG. 14. Frequency distribution of shell length of larvae of Aeolidia papillosa fed a suspension of Chlorella 580 (n = 30). algae had only P. tricornutum in the digestive diverticulum while Chlorella was found only in the stomach. DISCUSSION Developmental sequence The significant variation in the number of ova per capsule deposited by Hermissenda crassicornis can be explained by the fact that successive spawning by the same individual often results in fewer ova per egg capsule, and in some instances this depletion of ova even results in formation of empty egg cap- sules at the terminal end of an egg string. Furthermore, larger individuals tend to pro- duce more ova per egg capsule than do small individuals. This observation agrees with Har- rigan & Alkon's (1978) demonstration that the DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 111 107 A. Feeding larvae (cells present) 8 д 4 # larvae 80 90 100 lO 120 140 À B. Non-feeding larvae (cells absent) 8 80 90 00 10 120 130 140 Shell length (um) FIG. 15. Frequency distribution of shell length of larvae of Aeolidia papillosa ted a suspension of Dunaliella (n = 30). number of eggs per capsule increases linearly with weight of adult H. crassicornis. Observations of cleavage, gastrulation, formation of early veliger, and veliger mor- phology of Hermissenda crassicornis and Aeolidia papillosa correspond closely with those of other cleioproct aeolids (Hadfield, 1963; Hamatani, 1967; Tardy, 1970; Ras- mussen, 1951). Other workers concerned with early embryology in the Opisthobranchia have described a coelogastrula (Rasmussen, 1951; Rao 8 Alagarswami, 1960). Their illus- trations, however, depict a sagittal cleft as de- scribed and illustrated by Raven (1958). At hatching, each species demonstrates plank- totrophic development as defined by Thomp- son (1967). In each case, ova are small, egg capsules contain multiple embryos, develop- ment time to hatching is rapid, and free-swim- ming veligers lack eyespots, radulae, and propodial rudiments. The larval shell of each species 15 coiled (Type 1 of Thompson, 1961) 101 A Feeding larvae (cells present) 8 © A larvae Bay № ИМ 80 90 ЮО) = 20118071140 107 В Non-feeding (cells absent) # larvae 80 90 00 10 120 130 140 Shell length (um) FIG. 16. Frequency distribution of shell length of larvae of Aeolidia papillosa ted a suspension of Phaeodactylum tricornutum (п = 30). ie т 80 90 lOO ПО 120 130 140 Shell length (um) FIG. 17. Frequency distribution of shell length of larvae of Aeolidia papillosa ted a suspension of Phaeodactylum + Chlorella (n = 30). 112 WILLIAMS TABLE 5. Means and standard deviations of shell length (4m) of veliger larvae of Aeolidia papillosa employed in a 72 hour feeding experiment. Samples based on the presence or absence of yolk reserves were taken at the beginning of the experimental period. Four groups of larvae were then separated from the culture and fed suspensions of Chlorella 580, Dunaliella tertiolecta, Phaeodactylum tricornutum, and Phaeodactylum + Chlorella, respectively. At the end of the experiment shell length was measured for larvae that had ingested cells and that had not ingested cells from each of the suspensions. Percentages (n = 100) of larvae with yolk, Chlorella, Dunaliella, or Phaeodactylum cells in the stomach or diverticulum are given in parentheses. The remaining columns summarize departure from normality (Kolmogorov-Smirnov statistic, dmax), SKew- ness (41), and kurtosis (go) of the observed distributions of shell length. N t{hrn) Хим + 5 max 9: 92 Yolk-free larvae 50 0 125.0 6.66 0.058 —0.367 +0.433 Yolk-laden larvae (91%) 50 0 106.0 9.97 0.093 — 0.261 — 0.647 Chlorella fed to larvae cells in stomach (16%) 30 72 SS 8.60 0.108 —0.962* IO cells not in stomach 30 72 100.7 9.63 0.132 +0.666 0.125 Dunaliella fed to larvae cells in stomach (8%) 30 72 119.1 8.01 0.173 —0.538 — 01635 cells not in stomach 30 72 104.6 5.60 0.187 +0.192 —0.702 Phaeodactylum fed to larvae cells in stomach (6%) 30 72 122.2 5.82 0.118 —0.829* +1.382* cells not in stomach 30 72 110.5 9.67 0.172 +0.232 —1.212 Mixed suspension fed to larvae cells in stomach (17%) 30 72 117.8 8.51 0.144 ES +0.104 cells not in stomach — — — — — — — “Indicates significance at the 0.05 probability level. TABLE 6. Student-Newman-Keuls test for multiple comparisons among means of shell length (ит) of larvae of Aeolidia papillosa employed in the feeding experiment outlined in Table 5. Significant differences among means are based on comparison with the Least Significant Range (LSR) statistic calculated according to the procedures outlined by Sokal & Rohlf (1969). Group means from Table 5 are ranked from lowest to highest. Those means that are not significantly different (р > 0.05) from one another are shown by underscoring. Abbreviation for each group are YF (yolk-free), YL (yolk-laden), CF (Chlorella cells present), CNF (Chlorella cells absent), DF (Dunaliella cells present), DNF (Dunaliella cells absent), PF (Phaeodactylum cells pres- ent), PNF (Phaeodactylum cells absent), and MF (Phaeodactylum or Chlorella present). Rank 1 2 3 4 5 6 7 8 9 Group CNF DNF YL PNF CF MF DF PF YF Mean (ит) 100.7 104.6 106.0 110.5 15:5 117.8 le 122.2 125.0 and sculptured (Fig. 5). Harrigan & Alkon (1978) report an unsculptured larval shell in their study of H. crassicornis. This difference in observations is noteworthy and may reflect geographical variation between populations of H. crassicornis occurring in Monterey Bay, Harrigan 4 Alkon's (1978) collection site, and in Tomales Bay, 180 km north. Larval dimorphism Systematic review of development in the Aeolidoidea reveals the following associa- tions. Features of shell morphology such as shape (inflated vs. coiled), and presence of striations are more conservative at the gener- ic level than morphology of the soft parts (Thompson, 1961). The inflated larval shell js found among pleuroproct aeolidoideans while cleioprocts such as Hermissenda crassi- cornis and Aeolidia papillosa possess only coiled shells. Several developmental types (i.e., planktotrophic, lecithotrophic, encapsu- lated veliger, and ametamorphic) occur com- monly among congeners. For instance, Aeolidiella contains species with plankto- DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 113 trophic, lecithotrophic, and encapsulated veliger larvae (Hadfield, 1963; Hamatani, 1967; Rao & Alagarswami, 1960; Tardy, 1970). Coryphella has species with plank- totrophic and ametamorphic development (Bridges & Blake, 1972; Hurst, 1967; Morse, 1971). Phestilla melanobranchia is plankto- trophic but possesses eyespots which are characteristic of lecithotrophic larvae (Harris, 1975), while P. sibogae is lecithotrophic but possesses great flexibility in its ability to sur- vive in the plankton and metamorphose (Had- field, 1972; Bonar & Hadfield, 1974). Varia- tions in developmental type have been ob- served at the species level for Cuthona nana (Harris, Wright & Rivest, 1975; Rivest, 1978), Tenellia pallida (Eyster, 1979), and Spurilla neapolitana (Clark 8 Goetzfried, 1978). Bonar (1978) points out that the morphology of lar- vae competent to metamorphose and the events which characterize their subsequent metamorphosis are virtually identical for leci- thotrophic and planktotrophic forms. How- ever, premetamorphic larvae may be mor- phologically disparate and show variations that may be ecologically or evolutionarily im- portant. Developmental variations in mor- phology that precede competency may be ecologically adaptive by allowing portions of the larval population to survive seasonal or geographic changes in the environment or enable colonization of new habitats. Such variations may also be evolutionarily adaptive because they provide the population with lar- val phenotypes that may be responsive to the selective pressures that lead to the different developmental types and life history patterns (Strathmann, 1974; Vance, 1973a,b; and the arguments of Underwood, 1974). For in- stance, Clark 8 Goetzfried (1978) theorize that nudibranch species with gamete size and number intermediate to extremes encoun- tered in, for example, planktotrophic and lecithotrophic larvae should be capable of switching developmental types if ecological conditions vary, provided that genetic differ- ences between developmental types are small. They suggest that interruption of adult food supply may induce development of larvae with higher dispersal ability. In a morphological context, larvae of H. crassicornis and A. papillosa are plankto- trophic. At first glance, the polytypic occur- rence of yolk reserves seems to be a lecitho- trophic accommodation to reaching com- petency. If the dimorphism shown for H. crassicornis and A. papillosa truly reflects dif- ferences between lecithotrophic and plankto- trophic larvae, then one would expect to see large, yolky larvae developing from large eggs and small, yolk-free larvae developing from small eggs. However, in the cases of H. crassicornis and A. papillosa, yolk-free larvae are larger than their yolk-laden siblings and both types of larvae develop from uniformly small ova. Clark 8 Goetzfried (1978) have documented shifts in production of lecitho- trophic larvae to planktotrophic larvae by adult Spurilla neapolitana (Delle Chiaje) which had been starved prior to egg mass deposition. The shift from lecithotrophic to planktotrophic larvae was accompanied by a decrease in egg diameter and change from single to multi- ple embryo capsules. The nutritional state of parental H. crassicornis and A. papillosa was not followed during the present study and therefore provides no clue to the cause of dimorphic larvae. The size dependent distri- bution of yolk reserves could result from pre- mature release of slow-developing embryos from egg capsules ruptured by their faster developing, yolk-free siblings. However, yolk- laden siblings of A. papillosa were observed hatching from a common egg capsule which contained по yolk-free larvae. Harrigan & Alkon (1978) have shown that H. crassicornis requires 34 days (14.1°C) to reach competen- cy to metamorphose. Regardless of the mechanism which induces larval dimorphism, the presence of yolk reserves at hatching could significantly decrease the dependency of larvae on phytoplankton if yolk-laden and yolk-free larvae require the same amount of time after hatching to reach competency. If the presence of yolk reserves signals an in- crease in the time which larvae must spend in the plankton before they reach metamorphic competency, then the dependency of larvae on phytoplankton will be decreased while the dispersal area of the larvae will be increased (Strathmann, 1974). The production of slowly developing, widely dispersed, yolk-laden larvae along with faster developing, yolk-free siblings seems consis- tent with the opportunistic life history pattern of Hermissenda crassicornis documented by Birkeland (1974) and Harrigan & Alkon (1978). Adult H. crassicornis occur in a wide range of habitats, are eurytrophic, breed sub- annually, have a short generation time (2.5 months), and have a brief life span of 5-6 months. Larval H. crassicornis can be induced to metamorphose on a variety (at least three species) of hydroids. 114 WILLIAMS Larval feeding Observations concerning treatment of food by veliger larvae of Hermissenda crassicornis and Aeolidia papillosa are, with one excep- tion, in agrement with those of Thompson (1959) for Archidoris pseudoargus, Trin- chesia aurantia and Eubranchus exiguus. The exception noted here 15 the transport of large undigested cells along an anterior food groove (Fig. 7b) to the confluence of the stomach and intestine. This observation is in agreement with that of Fretter & Montgomery (1968) for the treatment of food by 19 species of prosobranch veligers. The initial rate of particle ingestion appears to accommodate the physical capacity of the stomach to retain algal cells. Subsequent oc- casional feeding most likely accommodates the rates of digestion and defecation. Veliger larvae of H. crassicornis are capable of luxury feeding greatly beyond their immediate needs. This point is corroborated by the high residence time of cells in the stomach and diverticulum (Fig. 8). Feeding ability of recently hatched larvae of H. crassicornis is determined by the presence of yolk reserves, the size of the larva, and the size of algal cell available for ingestion. Obvi- ously, those larvae that have their esophagus, stomach, and diverticulum filled with yolk cells are unable to feed. Some larvae, those with their esophagus and stomach free of yolk cells, had yolk cells and algae in the diver- ticulum simultaneously. Thus, the persistence of yolk reserves after hatching may not com- pletely deter yolk-laden larvae from feeding. Those larvae capable of feeding apparently do so in a selective fashion. In the case of H. crassicornis, there is no significant difference in shell length between yolk-free larvae and larvae that ingested Chlorella or D. tertiolecta (Table 4). Yet, there is a difference in percent of yolk-free larvae (51%) and larvae that in- gested either Chlorella or D. tertiolecta by the end of the experiment (41%, Table 3). Fewer larvae were ingesting cells of either species of alga than would be predicted by the percent- age of yolk-free larvae at hatching. This is particularly surprising for two reasons. Some yolk-laden larvae are able to ingest algal cells as described above. Also, one would expect the frequency of yolk-free larvae (i.e., larvae definitely able to ingest algal cells) to increase over time as yolk reserves are consumed. This discrepancy may be explained by the dif- ference in size between those larvae that in- gested Chlorella 580 and those that ingested Dunaliella tertiolecta. The larger larvae ate the larger algal cells (viz., D. tertiolecta), while small larvae ingested small algal cells (viz., Chlorella). Fig. 10 shows that larvae of H. crassicornis less than 103 um in shell length also possess yolk reserves. However, larvae less than 103 um (and presumably yolky) ingested Chlorella (Fig. 11), but for the most part, not D. tertiolecta (Fig. 12). If feeding abil- ity were restricted to only yolk-free larvae, then one would expect larvae less than 103 um to be nonfeeding. On the other hand, most H. crassicornis larvae greater than about 107 um in shell length are yolk-free and would be expected to feed on cells which they are physically able to ingest. This is not the case with Chlorella, where about 50% of the non-feeding larvae were larger than 107 um and were, presumably, yolk-free (Fig. 11). The converse of this situation arises in the case of D. tertiolecta, where most larvae greater than 107 ит had ingested cells (Fig. 12). Thus in each test flask not all those lar- vae capable of ingesting algal cells were suc- cessful, because of limitations imposed by either their mechanical ability to handle large particles (Strathmann, 1971) or by particle size selection (Paulson & Scheltema, 1968). These observations suggest that as larvae grow they may be able to ingest, and there- fore select, proportionately larger particles. Harrigan & Alkon (1978) observed that to de- velop in reasonable numbers to a state com- petent to metamorphose, veligers of H. cras- sicornis require algal cells larger (10-11 um) than the small (5-6 um) cells of /sochrysis and Monochrysis commonly fed molluscan larvae. Similar arguments regarding feeding ability apply to the larvae of A. papillosa. Increasing means of shell length (Table 5) as a function of cell size ingested (Chlorella < Dunaliella < Phaeodactylum) suggest that small larvae are physically incapable of ingesting large algal cells such as P. tricornutum and large larvae select large algal cells. Fig. 13 shows that larvae of A. papillosa less than 105 um in shell length contain yolk reserves while those greater than about 125 ит do not contain yolk reserves. A higher proportion of larvae less than 105 um ingested Chlorella (~16%) than ingested D. tertiolecta (—10%) or P. tricornutum (—3%). Fewer larvae greater than 125 um ingested Chlorella (—10%, Fig. 14) than ingested either D. tertiolecta or P. tri- cornutum (~37%, Figs. 15 and 16). In a mixed suspension of Chlorella and P. tri- DEVELOPMENT AND FEEDING OF NUDIBRANCH LARVAE 115 cornutum, larvae greater than 120 um con- sumed both species of algae. The larger alga, P. tricornutum, may have been consumed first since only P. tricornutum was present in the digestive diverticulum and Chlorella was found only in the stomach. However, random ingestion of particles followed by sorting them by size for passage into the digestive divertic- ulum cannot be discounted as an alternative explanation to size selective ingestion. CONCLUSIONS Larvae of Hermissenda crassicornis and Aeolidia papillosa are planktotrophic in the morphological (Thompson, 1967), morpho- genetic (Bonar, 1978; Harrigan 8 Alkon, 1978) and functional sense of the term. At hatching, larvae of both species are polytypic with respect to the presence of yolk reserves. Feeding ability of recently hatched larvae of both species 15 determined by the presence of yolk reserves, the size of larvae, and the size of algal cell available for ingestion. The larvae of both species either do not need to feed (yolky larvae) when introduced into the plank- ton or can feed greatly beyond their immedi- ate needs (yolk-free larvae) and may, there- fore, survive by sporadic feeding in an envi- ronment where phytoplankton may have spa- tially (Mackas 8 Boyd, 1979; Platt, Dickie 8 Trites, 1970) and temporally (Malone, 1977; Peterson 4 Miller, 1975) patchy distributions. The presence of yolk reserves in the larvae of H. crassicornis and A. papillosa appears to allow a portion of the larvae temporary inde- pendence of phytoplankton food and, per- haps, increased dispersal without the risks of delayed metamorphosis. ACKNOWLEDGEMENTS | thank Drs. James A. Blake, Steven Obrebski, and Edmund H. Smith for providing encouragement and advice while this re- search was being performed at the Pacific Marine Station. Dr. Melbourne Carriker, Dr. Charles Epifanio, Dr. Robert Palmer, Mr. Robert Prezant, Ms. Betsy Brown, and Ms. Lindy Eyster reviewed the manuscript and contributed many helpful suggestions. Final figures were prepared by Mrs. Jan Sick and photographic prints were prepared Бу Mr. Walter Kay. LITERATURE CITED BIRKELAND, C., 1974, Interactions between a sea pen and seven of its predators. Ecological Monographs, 44: 211-232. BONAR, D. B., 1978, Morphogenesis at meta- morphosis in opisthobranch molluscs. In: F.-S. CHIA 8 M. RICE (eds.), Settlement and meta- morphosis of marine invertebrate larvae. Else- vier, New York, p. 177-196. BONAR, D. G. & HADFIELD, М. G., 1974, Meta- morphosis of the marine gastropod Phestilla sibogae Bergh (Nudibranchia: Aeolidacea). |. Light and electron microscope analysis of larval and metamorphic stages. Journal of Experi- mental Marine Biology and Ecology. 16: 227- 255. BRIDGES, С. & BLAKE, J. A., 1972, Embryology and larval development of Coryphella trilineata O'Donoghue, 1921 (Gastropoda: Nudibranchia). Veliger, 14: 293-297. CLARK, K. 8 GOETZFRIED, A., 1978, Zoogeo- graphic influences on developmental patterns of North Atlantic Ascoglossa and Nudibranchia with a discussion of egg size and number in opistho- branchia. Journal of Molluscan Studies, 44: 283-294. EYSTER, L., 1979, Reproduction and develop- mental variability in the opisthobranch Tenellia pallida. Marine Biology, 51: 133-140. FRETTER, V., 1967, The prosobranch veliger. Proceedings of the Malacological Society of London, 37: 357-366. FRETTER, V. & MONTGOMERY, М. C., 1968, The treatment of food by prosobranch veligers. Journal of the Marine Biological Association of the United Kingdom, 48: 499-520. GUILLARD, В. В. L. 4 RYTHER, J. H., 1962, Studies on marine planktonic diatoms. 1. Cyclotella папа Hustedt and Detonula con- fervacea (Cleve) Gram. Canadian Journal of Microbiology, 8: 229-239. HADFIELD, М. G., 1963, The biology of nudibranch larvae. Oikos, 14: 85-95. HADFIELD, М. С., 1972, Flexibility in larval life patterns. American Zoologist, 12: 721. HAMATANI, |., 1967, Notes on veligers of Japan- ese opisthobranchs (7). Publications of the Seto Marine Biological Laboratory, 15: 121-131. HARRIGAN, J. F. & ALKON, D. L., 1978, Larval rearing, metamorphosis, growth and reproduc- tion of the aeolid nudibranch Hermissenda crassicornis (Eschscholtz, 1831) (Gastropoda: Opisthobranchia). Biological Bulletin, 154: 430- 439. HARRIS, L. G., 1975, Studies on the life history of two coral-eating nudibranchs of the genus Phestilla. Biological Bulletin, 149: 539-550. HARRIS, L. G., WRIGHT, L. W. & RIVEST, В. R., 1975, Observations on the occurrence and biol- ogy of the aeolid nudibranch Cuthona nana in New England waters. Veliger, 17: 264-268. HURST, A., 1967, The egg masses and veligers of 116 WILLIAMS thirty northeast Pacific opisthobranchs. Veliger, 9: 255-287. KNIGHT-JONES, Е. W., 1954, Relations between metachronism and the direction of ciliary beat in Metazoa. Quarterly Journal of Microscopical Science, 95: 503-521. MACKAS, P. L. & BOYD, С. М., 1979, Spectral analysis of zooplankton spatial heterogeneity. Science, 204: 62-64. MALONE, T. C., 1977, Light saturated photo- synthesis by phytoplankton size fractions in the New York Bight, U.S.A. Marine Biology, 42: 281-292. MORSE, M. P., 1971, Biology and life history of the nudibranch mollusc, Coryphella stimpsoni (Ver- rill, 1979). Biological Bulletin, 140: 84-94. PAULSON, Т. С. 4 SCHELTEMA, В. S., 1968, Se- lective feeding on algal cells by veliger larvae of Nassarius obsoletus (Gastropoda: Proso- branchia). Biological Bulletin, 134: 481-489. PERRON, Е. Е. & TURNER, В. D., 1977, Develop- ment, metamorphosis, and natural history of the nudibranch Doridella obscura Verrill (Corambi- dae: Opisthobranchia). Journal of Experimental Marine Biology and Ecology, 27: 171-185. PETERSON, W. Т. 8 MILLER, С. B., 1975, Year to year variations in the planktology of the Oregon upwelling zone. Fishery Bulletin, 73: 642-653. PLATT, T., DICKIE, L. M. & TRITES, R. W., 1970, Spatial heterogeneity of phytoplankton in a near shore environment. Journal of the Fisheries Re- search Board of Canada, 27: 1453-1473. RAO, K. V. & ALAGARSWAMI, K., 1960, An ac- count of the structure and early development of a new species of nudibranchiate gastropod, Eolidina (Eolidina) mannarensis. Journal of the Marine Biological Association of India, 2: 6-17. RASMUSSEN, E., 1951, Faunistic and biological notes on marine invertebrates. Il. The egg masses and larvae of some Danish marine gastropods. Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening, 113: 201-249. RAVEN, C. P., 1958, Morphogenesis: the analysis of molluscan development. Pergamon, London, ЗАМ: р: RIVEST, В. R., 1978, Development of the eolid nudibranch Cuthona nana (Alder and Hancock, 1842), and its relationship with a hydroid and hermit crab. Biological Bulletin, 154: 157-175. SOKAL, В. В. 4 ВОНЕЕ, Е. J., 1969, Biometry: the principles and practice of statistics in biological research. Freeman, San Francisco. 776 p. STRATHMANN, В. В., 1971, The feeding behavior of planktotrophic echinoderm larvae: Mechan- isms, regulation and rates of suspension feed- ing. Journal of Experimental Marine Biology and Ecology, 6: 109-160. STRATHMANN, В. B., 1974, The spread of sibling larvae of sedentary marine invertebrates. Amer- ican Naturalist, 108: 29-44. TARDY, J., 1970, Contribution à l'étude des metamorphoses chez les nudibranches. Annales des Sciences Naturelles, Zoologie et Biologie Animale, ser. 12, 12: 299-370. THOMPSON, Т. E., 1958, The natural history, embryology, larval biology and post-larval devel- opment of Adalaria proxima (Alder and Han- cock), (Gastropoda: Opisthobranchia). Philo- sophical Transactions of the Royal Society of London, ser. B, 242: 1-58. THOMPSON, T. E., 1959, Feeding in nudibranch larvae. Journal of the Marine Biological Asso- ciation of the United Kingdom, 38: 239-248. THOMPSON, T. E., 1961, The importance of the larval shell in the classification of the Sacoglossa and the Acoela. Proceedings of the Malaco- logical Society of London, 34: 233-238. THOMPSON, T. E., 1962, Studies on the larval ontogeny of Tritonia hombergi Cuvier (Gastro- poda: Opisthobranchia). Philosophical Transac- tions ofthe Royal Society of London, ser. B, 245: 171-218. THOMPSON, T. E., 1967, Direct development in a nudibranch, Cadlina laevis, with a discussion of developmental processes in Opisthobranchia. Journal of the Marine Biological Association of the United Kingdom, 47: 1-22. UNDERWOOD, A. H., 1974, On models for repro- ductive strategy in marine benthic invertebrates. American Naturalist, 108: 874-878. VANCE, R. R., 1973a, On reproductive strategies in marine benthic invertebrates. American Naturalist, 107: 339-351. VANCE, В. R., 1973b, More on reproductive strat- egies in marine benthic invertebates. American Naturalist, 107: 353-361. MALACOLOGIA, 1980, 20(1): 117-142 OBSERVATIONS ON THE ANATOMY AND LIFE HISTORY OF MODULUS MODULUS (PROSOBRANCHIA: MODULIDAE) Richard S. Houbrick Department of Invertebrate Zoology (Mollusks), National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, U.S.A. ABSTRACT Modulus modulus (Linnaeus), family Modulidae, is a style-bearing marine prosobranch in the superfamily Cerithiacea. It differs from other cerithiaceans by its turbinate shell and certain anatomical features. The basic anatomy of Modulus is that of a typical mesogastropod but the open pallial gonoducts are similar to those of other cerithiaceans. Characteristic anatomical feature $ are the large hypobranchial gland, aphallic males, open pallial oviduct with complex inner aucts, and the forward position of the salivary glands relative to the nerve ring. Moculus modulus lives on marine angiosperm grasses and feeds primarily on diatoms. Ferti"zation is internal and is effected by spermatophores that contain eupyrene and apyrene sperms. The life cycle of a population from Fort Pierce, Florida lasts about one year. Mating occurs in early winter and spawning in spring. Females have complex pallial oviducts and ovipositors. Spawn masses are cylindrical, comprised of gelatinous tubes and deposited on marine grasses. Development is direct. Young snails emerge after three weeks of incubation. Within the Cerithiacea, the Modulidae are phyletically close to the Cerithiidae and Potamididae. INTRODUCTION The family Modulidae Fischer consists of a single genus, Modulus Gray, 1842, and about six species that are largely confined to the shallow waters of tropical and subtropical re- gions. These prosobranchs are epifaunal, style-bearing, microphagous herbivores and are moderately small animals, having shells 10-25 mm in length. Modulus species occur in the fossil record back to the Cretaceous (Dall, 1892). Historically, the family Modulidae has been assigned to the large mesogastro- podan superfamily Cerithiacea Fleming, which is comprised of numerous families of ` herbivorous snails, having open pallial gono- ducts, aphallic males, and shells that are usu- ally high-spired and elongate. Members of the Modulidae, however, differ strikingly from all other cerithiaceans by their top-shaped, trochiform shells. Although the family Modu- lidae has been included in most of the classic iconographies and the western Atlantic spe- cies have recently been monographed by Abbott (1944), there have been no serious anatomical or life history studies of the group. Its standing as a valid family and relationship to other higher taxa in the superfamily has been conjectural. This paper addresses the functional mor- phology, developmental biology and aspects of the life history of Modulus modulus (Linnaeus, 1758), the type-species of the genus. On the basis of these data, the rela- tionship of the Modulidae to other families of the Cerithiacea will be considered. HISTORICAL REVIEW The family Modulidae has been reviewed by A. Adams (1851), Sowerby (1855), Reeve (1865), Tryon (1887) and more recently by Abbott (1944), who limited his work to the western Atlantic. The older monographs were confined to strict conchological taxonomy and did not address relationships above the alpha level. The anatomy of Modulus was first de- scribed by Risbec (1927), who dealt with Modulus candidus Petit, 1853 [= M. tectum (Gmelin, 1791)]. He placed the Modulidae be- tween the Strombidae and Cerithiidae. Al- though Risbec (1927) noted the aphallic con- dition of males, he did not adequately study the reproductive tract and neglected to de- scribe the complex pallial gonoducts. His other anatomical observations are sketchy and sometimes erroneous. Abbott (1944) fig- ured and described the external soft parts of (117) 118 HOUBRICK the head and foot of Modulus modulus and presented a few brief notes on its ecology. To my knowledge, only two papers have considered the developmental biology of Modulus: Lebour (1945: 470—471) has de- scribed the spawn and larvae of Modulus modulus, and more recently Bandel (1976: 258-259) described and figured the spawn and larvae of Modulus modulus and Modulus carchedonius (Lamarck, 1822). Both of these descriptions are brief and the taxonomic iden- tity of the species 15 questionable. This will be discussed in more detail later in this paper. Mook (1977) is the only author to have writ- ten anything about the ecology of Modulus. His study was concerned with the role of Modulus modulus as a control of fouling or- ganisms on marine angiosperm grasses and did not consider the ecology of Modulus. To my knowledge, nothing more has been published about the biology of Modulus spe- cies, despite the fact that some species, such as Modulus modulus, commonly occur in great numbers and are easily collected and observed. MATERIALS AND METHODS Modulus modulus specimens were ob- tained from a large population in a seagrass bed north of Link Port on the west bank of the Indian River lagoon near Fort Pierce, Florida (27°32.1'N, 80°20.9'W). For a more detailed description of this site, see Young et al. (1976). The subtidal site consisted of dense stands of Halodule wrightii Ascherson and occasional beds of Thalassia testudinum König 8 Sims interspersed with sandy patches. Snails from this population were studied during January, February, May and September of 1978 and in January of 1979. Living animals were studied in the field and in sea water aquaria or petri dishes placed under a Wild M-5 stereo dissecting micro- scope. Animals were extracted from their shells and relaxed in a 7.5 percent solution of MgClo, and dissected in an anesthetized state. Carmine particles were used to deter- mine ciliary tracts and a one percent Methyl- ene Blue aqueous solution was used as a vital stain. Stomach contents and fecal pellets were studied under a compound microscope to determine the algae that were ingested. Animals were fixed in Bouin's fluid, embedded in paraffin, sectioned at 9.0 ит, and stained with Basic Fuchsin Picric Indigo Carmine for histological studies. Photomicrographs of sections were taken with a Zeiss photomicro- scope-3. Spermatozoa were fixed in a 2.5 percent gluteraldehyde solution in 0.2 molar Mellonig’s phosphate buffer and were brought up to 50 percent EtOH and air dried on cover slips prior to SEM studies. Spermatophores, eggs and embryos were prepared for critical point drying by fixation in 2.5 percent gluteral- dehyde in 0.2 molar Mellonig's phosphate buffer at pH 7.4. Material was rinsed in dis- tilled water, dehydrated and the critical point drying was done with liquid COs. Spermato- zoa, spermatophores, eggs, embryos, radulae and embryonic shells were studied with a Nova-Scan SEM. Observations on pair- ing, Spawning and feeding were made in the field and in the lab. Eggs and embryos were maintained in culture dishes with daily sea water changes until hatching. Measurements of eggs and embryos were made with an ocu- lar micrometer. In addition to the above work, the anatomy of Modulus tectum (Gmelin, 1791) was stud- ied at Suva, Fiji for comparison with that of Modulus modulus, but this work was only at a superficial level. All measurements are relative to average sized snails (see Table 1). ANATOMY Shell (Fig. 1a-c).—The shell is top-shaped, umbilicate, wider than high, and consists of 5—6 strongly convex, angulate whorls of which the body whorl is disproportionately large. en FIG. 1. Shells, spawn mass and embryos of Modulus modulus; а-с, Apical, apertural and basal views of adult shell with typical sculpture; d-f, Apical, apertural and side view of young shell. Note sharp transition between sculpture of embryonic shell and that of teleoconch (1.5 mm wide); g, Мему hatched snail showing distinctive sculpture (0.5 mm long); h, Five day old snail showing sharp transition in sculpture between embryonic shell and new growth area of outer lip (0.8 mm long); i, Portion of spawn mass critical point dried and fractured to show hyaline capsule surrounding each embryo. Embryonic shell sculpture is visible beneath surface of each capsule; j, Typical spawn mass attached to grass blade (9 mm long); k, Portion of spawn mass with surface debris removed to display encapsulated embryos. 119 BIOLOGY OF MODULUS 120 HOUBRICK The spire is low and turbinate. The suture 15 slightly irregular and moderately impressed. The periphery of each whorl has a prominent spiral cord that forms a keel. Three to four weaker spiral cords are above the peripheral cord. Adapical to the peripheral cord each whorl is sculptured with 12-13 axial ribs that form low, blunt nodules where they cross the spiral cords. The body whorl descends just before the aperture in adults and has 4-5 spi- ral cords on its base sculptured with numer- ous tiny nodules that are aligned to form weak axial riblets. In juvenile shells the body whorl is extremely angulate and keeled (Fig. 19-1. The aperture is ovate and the columella deep- ly concave and terminated with a deep notch or chink that forms a tooth-like lamella. This notch accommodates the pallial tentacles of the inhalant siphon. The outer lip is moderate- ly thin and strongly crenulate, each scallop fitting a pallial tentacle. The inner surface of the outer lip is reinforced with 5—6 spiral ridges. The umbilicus is small but deep and in adults is slightly covered by the columellar fold. The protoconch consists of two whorls that are convex but not angulate, and sculp- tured with 5-6 thin, spiral lirae except for the smooth nuclear tip. Basic shell color is a dirty white but is normally hidden by the periostra- cum and algal epiphytes. Brownish purple splotches occur on the spiral chords and ad- jacent to the suture of the body whorl. The columella is tinged with purple and the colu- mellar notch has a purple spot. Average shell dimensions of the Link Port population were 8.31 mm in length and 9.24 mm in width (see Table 1). The operculum is thin, corneous, circular and multispiral with a central nucleus. It fits snugly into the aperture of the shell when the animal is retracted. The periostra- cum is thin, and tan. Animal: external features (Fig. 2).—The base color of the head, neck and foot regions cm 3mm FIG. 2. A, Modulus modulus, male removed from shell and viewed from left side. B, Female removed from shell and seen from right side. ag, albumen gland; cf, ciliated furrow; cg, capsule gland; cm, columellar muscle; ct, ctenidium; dg, digestive gland; f, foot; hg, hypobranchial gland; in, inhalant siphon; int, intestine; k, kidney; mt, mantle tentacle; op, operculum; ov, Ovary; ovp, ovipositor; pp, propodium; ppf, propodial furrow; pt, pallial tenta- cle; sn, snout; st, stomach; t, tentacle; te, testis. TABLE 1. Analysis of shell dimensions (measurements in mm). ea ee zz esse Length Width Statistic No. % x Sd Range x Sd Range АП shells 61 100 8.31 0.63 6.3-10.1 9.24 0.66 7.7-10.6 Males 27 44.26 8.24 0.78 6.3-10.1 9.09 0.67 7.7-10.6 Females 32 52.46 8.40 0.47 7.5= 9.5 9.94 0.64 9.2-10.5 Parasitized 2 3.27 — — — = — = No., number of snails Sd, standard deviation x, mean BIOLOGY OF MODULUS 121 is pinkish-cream with dusky green and chalky white blotches, spotted with tiny red flecks. The white blotches are composed of fine white dots. Some yellow pigment occurs on the proximal portion of the dorsal surface of the neck and body. The base of the foot 15 more lightly pigmented. The overall appear- ance is a light green or mossy green color, as described by Abbott (1944: 3). When fully extended the foot 1$ slightly smaller than the diameter of the shell. The foot is shield-shaped and begins with a cres- cent-like propodium that has a deep glandular furrow (Fig. 24, ppf) set off anteriorly with a pigmented band of alternating green and white bars and posteriorly with a thin, lightly pigmented area. This furrow is formed by an invagination of numerous, highly vacuolated, glandular cells. Although the muscles of the sole produce monotaxic retrograde waves, the animal moves by jerk-like contractions of the columellar muscle. The head has a short, rounded, dorso- ventrally flattened snout with a bilobed tip bearing a longitudinal slit leading to the mouth (Fig. 14, sn). The dorsal surface of the snout is dark brownish-green and the tip is bordered with alternating white and green blotches and randomly placed red dots. The ventral surface is pinkish. The snout can be extended con- siderably when the animal 15 seeking to right itself but is normally held in a retracted state, even while feeding. The head has two thin tentacles (Fig. 1B, t), about 3 mm in length, in an average-sized animal. The tentacles and head are covered with tiny prominent white pustules. The proxi- mal two thirds of each tentacle is thicker than the tip. When crawling, the tentacles are placed at 45 degrees to the snout and the thin tips are placed on the substratum where they probably serve a sensory function. The eyes are placed on the tentacles where they nar- row, about two thirds the way from the base. Eyes are black and surrounded by a narrow circle of deep yellow pigment. There is a dark horizontal green band of pigment adjacent to each eye. Animals respond quickly to a shad- ow or sudden movement. The dorsal half of the mantle edge 1$ scal- loped and fringed with about 20 thick, white pallial tentacles (Fig. 3D), each having two tiny pink dots on its medial ventral surface. Between each pallial tentacle is a dark green blotch. The mantle edge, as seen in cross section, is trilobed (Fig. 3C). There is a deep ciliated furrow between the lobes. This furrow prevents foreign matter from getting between the shell and mantle by collecting debris which is then moved down the right side of the mantle edge and expelled by the exhalant siphon. The outer lobe, adjacent to the shell, is transparent, microscopically scalloped and has a slight groove at its tip. The inner edge bears the pallial tentacles and is brightly pig- mented. The inhalant and exhalant siphonal portions of the mantle edge are thicker and slightly curved but have no distinguishing morphological characteristics. The ventral half of the mantle edge lacks the lobes and 1$ smooth. A ciliated furrow (Fig. 2B, cf) extends from the exhalant siphonal area down the right side of the foot. It carries fecal pellets and debris entwined in mucus away from the mantle cavity to the bottom of the right side of the foot where they move back under the unattached edge of the operculum and are cast off. The ciliated strip is not readily seen without the application of carmine particles. Marcus & Marcus (1964: 500) described a similar ciliat- ed tract in Cerithium atratum (Born, 1778) and | have noted its presence in other cerithiid species. In females, this ciliated strip is a groove that also functions in transporting egg capsules from the distal pallial oviduct to the ovipositor (Fig. 2B, ovp) with which it is inti- mately associated. The ovipositor is a bul- bous, swollen area that comprises a thickly furrowed flap under which lies a glandular area of vacuolated cells. It is located midway along the ciliated groove on the right side of the foot and has a distinctive darker color and unusual swollen appearance in ripe females. | did not observe the use of this organ during spawning; consequently, its exact function is unknown. Sections of the organ reveal num- erous glandular cells and vesicular elements. Presumably, the ovipositor adds the outer jelly coat that surrounds the egg capsules formed in the pallial oviduct and assists in both the formation of the outer surface and attachment of a completed spawn mass. Mean length of animals removed from their shells is about 9 mm in their natural coiled state and about 15 mm uncoiled. These mean values are the standards against which other anatomical measurements in this paper should be compared. The mantle, as seen in snails extracted from shells, is bright orange and green. The portion covering the pallial gonoducts is whit- ish while that over the intestine and hypo- branchial gland is brilliant green. The mantle 122 HOUBRICK BIOLOGY OF MODULUS 123 covering the ctenidium and osphradium is orange. The mantle cavity is 5-6 mm deep and is broadest at its distal end where the edge flairs up and backwards. The exhalant siphon is marked on the mantle surface by a prominent, rounded ridge that is thick, muscu- lar and curved in a semicircle down towards the medial right part of the mantle edge. The columellar muscle (Fig. 2B, cm) is broad and quite large (2.5 тт long; 3 mm wide) in relation to the animal. It is tightly at- tached for one complete whorl to the inner columella of the shell about one-half whorl's distance from the aperture. The columellar muscle is located at the ventral surface of the snail just behind and on the ventral mantle edge where it extends diagonally backwards in relation to the anterio-posterior orientation of the mantle cavity. This muscle is powerful in Modulus and is used by the animal to re- tract into the shell and to produce quick jerky movements and violent twisting behavior that will be described in more detail later in this paper. The whorls of the animal comprising the stomach, gonads and digestive gland taper off sharply and curl beneath the columellar muscle. The stomach (Figs. 2A,B, st) occu- pies about 1.5 whorls and is colored white, brown and green. The remaining two and a half to three whorls comprise the digestive gland and gonad. The digestive gland (Fig. 2A, dg) is chocolate brown and is almost entirely overlain by the gonad in ripe animals. The ovary (Fig. 2B, ov) is white and extends over the digestive gland in a branching network. The testis (Fig. 2A, te) is yellow-orange and covers most of the digesive gland. Since males are aphallic, the pigmentation of the testis is an easy way to sex ripe individuals. Animals parasitized by trematodes frequently have their gonadial and lower alimentary —— tracts filled with rediae, cercariae and sporo- cysts and have pinkish-colored gonads. Mantle cavity and associated organs (Fig. 3).—At the lower left side of the mantle cavity is the brown-pigmented osphradium (Fig. ЗА, os), a thin, ridge-like structure that is triangu- lar in cross section and about 5 mm long and 0.35 mm wide. The osphradium begins at the proximal end of the mantle cavity and extends forward, adjacent to the ctenidium for about two-thirds of its length. As the osphradium nears the inhalant siphon it sharply turns to form an “S” shape and tapers to an end about 0.25 mm from the mantle edge. It is separated from the ctenidium by an orange-russet pig- mented strip about 0.20 mm wide. Sections of the osphradium show a highly ciliated surface and numerous, darkly staining cells. At its base 15 a thick osphradial nerve that has con- nections with the left mantle nerve. The ctenidium (Fig. ЗА, ct) 15 relatively large, about 6 mm in length and 1.5 mm in width. К extends the length of the mantle cavi- ty and comprises about 125 long finger- shaped, flattened filaments. Each filament is 1.5 mm long and 0.45 mm wide at its attached base and tapers toward the tip. Individual fila- ments are strengthened on the right side (edge) by an internal rod (Fig. ЗВ, г). A ciliated longitudinal band (Fig. 3B, cb) is on each side of the flattened surface of a filament, adjacent to the rod. The narrow edges of each filament are lined with cilia and exceptionally lengthy cilia are on the filament tip (Fig. ЗВ, c/). Thin transverse bands of muscles (Fig. 3B, mb) allow each filament to respond to the stimulus of a probe by quickly retracting and bending. The hypobranchial gland (Fig. 3A, hb) lies to the right and next to the ctenidium. It is about 4.0mm long and 1.0mm wide and highly conspicuous because of its swollen, glandular state and bright russet-green color. FIG. 3. Anatomical features of Modulus modulus. A, Animal removed from shell and mantle cavity opened mid-dorsally. The kidney has been pulled back to expose the kidney opening and heart. B, Individual ctenidial filament. C, Cross section through mantle edge showing trilobed condition and ciliated groove containing debris. D, Dorsal view of mantle edge displaying ciliated groove and inner lobe. E, Stomach opened by dorsal longitudinal cut. Crystalline style has been removed. a, anus; ag, albumen gland; cb, ciliated band; cg, capsule gland; cgr, ciliated groove; cl, long cilia; ct, ctenidium; dd, digestive gland duct; dg, digestive gland; ebv, efferent branchial vessel; es, esophagus; ex, exhalant siphon; fsc, fold emerging from spiral caecum; gs, gastric shield; gspr, groove leading to spermatophore receptacle; gsr, groove leading to seminal receptacle; hb, hypobranchial gland; il, outer lobe of mantle edge; in, inhalant siphon; int, intestine; к, kidney; ko, kidney opening; //b, baffle of lateral lamina; m, mouth; mb, muscle band; me, mantle edge; od, pallial oviduct groove; oes, opening from esophagus to stomach; os, osphradium; ov, ovary; pc, pericardial sac; pod, proximal portion of pallial oviduct; pt, pallial tentacle; r, internal strengthening rod; s, shell; sa, sorting area; sc, reduced spiral caecum; ss, style sac; st, stomach; t,, typhlosole 1; 6, typhlosole 2; ve, ventricle. 124 HOUBRICK Tapering at both ends, it begins at the proxi- mal end of the mantle cavity and rapidly widens, extending longitudinally and ending behind the anus. The hypobranchial gland lies close to and slightly overlaps the intes- tine. It is covered externally with numerous tiny papillae and has a transversely ridged surface that is russet colored. Oval and gob- let-shaped areas appear within the gland when viewed from the surface. If a snail is violently disturbed or if the hypobranchial gland is stimulated with a probe, it exudes copious strands of tiny, globular, mucus-like bodies. These are shot out in salvos from small openings in the surface of the gland un- til the mantle cavity is nearly filled with them. The exudate is rapidly moved by cilia and ex- pelled by the exhalant siphon. The animal will continue to exude globular particles whenever stimulated until the gland is spent. The exu- date is probably used defensively by the snail when under attack. The composition and nature of the exudate is unknown. | have not seen this phenomenon in other cerithiacean snails. Sections of the hypobranchial gland show numerous columnar, vesicular and darkly stained glandular cells at its surface. Beneath these are vesicle-like chambers that appear to store the hypobranchial exudate. No ducts leading from these chambers to the surface of the gland were seen. The intestine (Fig. ЗА, int) is 0.25 mm wide and usually dark colored due to many small, transversely oriented, ovoid fecal pellets that fill it. The pallial epithelium overlying the in- testine is covered with tiny papillae that ex- tend over it from the adjacent hypobranchial gland. The anus, borne оп a large papilla, is about 2.5 mm from the mantle edge (Fig. ЗА, a). The pallial gonoducts are open in both males and females and males are aphallic. The open condition of the gonoducts is best visu- alized as a slit tube that runs the length of the mantle cavity, consisting of inner and outer laminae. The male pallial gonoduct (Fig. 8A, C) is a thin-walled, glandular, open tube while the pallial oviduct of females (Fig. 8B) is a larger, white glandular organ comprising laminae with complex inner chambers and tiny ducts. The functional aspects of both male and female pallial gonoducts are dis- cussed in detail later in this paper, in the sec- tion on the reproductive system. At the proximal end of the mantle cavity lies the anterior wall of the pericardial sac (Fig. 3A, pc) and to its right is the kidney which has a typical slit-like opening on its ventral surface (Fig. ЗА, ko). Alimentary system (Figs. 3A, 4, 5).—This system is typically mesogastropodan and is somewhat like that described by Fretter 4 Graham (1962: 25-32) for Littorina. The mouth lies at the tip of the snout but is usually recessed between the two lobes that comprise the snout apex. At the anterior end of the buc- cal cavity and inserted in its lateral walls are a pair of chitinous jaws (Fig. 4, /) composed of tiny rhomboidal plates. Dissection of the buc- cal apparatus reveals two large radular retrac- tor muscles that arise from the walls of the anterior body cavity and insert on the post- median surface of the buccal mass (Fig. 4, ит, rrm). The buccal mass (Fig. 4, bm) is about 1.65 mm long and 1.24 mm wide. The radular ribbon is 2.25 mm long and 0.35 mm wide, comprises about 68 rows of teeth, and 1$ one- fourth the length of the shell (see Table 2). Itis typically taenioglossate (2+1+1+1+2) and has a rounded, quadrate-shaped rachidian tooth (Fig. 5). The basal plate of the rachidian is smooth and has a long central projection and convex lateral sides. lt is similar to that of Cerithium species and lacks the basal cusps that are seen on potamidid species such as Batillaria. The top of the rachidian tooth 15 markedly convex and has a cutting edge comprised of a large, pointed central denticle that is flanked on each side with 2-3 smaller pointed denticles that diminish in size lateral- ly. The lateral tooth of the radula 1$ trapezoidal in shape, and has a long lateral extension that curves down at its insertion point in the radu- lar membrane. On the basal plate of the later- al tooth there is a long, blunt longitudinal ex- tension. The top of the lateral tooth 15 straight and serrated with a small inner cusp, a sec- ond larger cusp and 3-4 smaller ones. The inner and outer marginal teeth are long, spatulate and serrated at their tips with 5-6 closely set, blunt denticles. The two marginal teeth are virtually identical, and thus differ from those of Cerithium species, which have distinguishable ощег and inner marginal teeth. The radular sac begins ventral to the esophagus and coils in a spiral to the right. There are no esophageal pouches. A large spade-shaped esophageal valve 1$ present. The paired salivary glands (Fig. 4, /sg, rsg) are loosely compacted lobes, lying behind the buccal mass above and beside the anterior esophagus. The left is four-lobed, and twice as large as the right, and extends through the BIOLOGY ОЕ MODULUS 125 FIG. 4. Dissection of the head of Modulus modulus opened by a dorsal longitudinal cut exposing anterior alimentary tract. ae, anterior esophagus; bm, buccal mass; bn, buccal nerve; cpc, cerebal-pedal connective; cplc, cerebral-pleural connective; dfe, dorsal fold of esophagus; dsg, duct of right salivary gland; /, jaw; Icg, left cerebral ganglion; Imn, left mantle nerve; /pg, left pleural ganglion; /rm, left radular retractor muscle; Isg, left salivary gland; me, mid-esophagus; opn, optic nerve; osn, osphradial nerve; pg, right pedal ganglion; ppc, pleural-pedal connective; r, radula; rcg, right cerebral ganglion; rmn, right mantle nerve; rpg, right pleural ganglion; rrm, right radular retractor muscle; rsg, right salivary gland; sec, supraesophageal connec- tive; seg, supraesophageal ganglion; srm, subradular membrane; svc, subvisceral connective; tn, tentacle nerve. 126 HOUBRICK FIG. 5. Radula of Modulus modulus. Left, general view of radular ribbon; right, half row showing cusp arrangent on rachidian, lateral and marginal teeth. Rachidian tooth length = 0.06 mm. TABLE 2. Statistical summary of radular and shell measurements. Statistic No. х Sd Range Shell length 8 875087 8.1- 9.2 Shell width 8 9.75 0.49 9.1-10.2 Radula length 8 22 ОБ 2.0- 2.5 Rows of teeth 8 68.8 2.60 66 —73 No., number of snails Sd, standard deviation X, mean nerve ring a short way down the mid-esopha- gus. The right gland is three lobed, and lies wholly in front of the nerve ring. The ducts arise in front ofthe cerebral ganglia, and enter the buccal cavity at each side (Fig. 4, dsg). The anterior esophagus has a pronounced dorsal food groove, and twists, due to torsion, at the cerebral ganglia; but the dorsal folds do not extend back ventrally as in Littorina. The dark-pigmented mid-esophagus can be seen through the dorsal body wall. It is enlarged into a well developed esophageal gland com- prising numerous deep folds and diverticula, giving it a loosely compacted flocculent struc- ture that easily falls apart under dissection. Begining behind the nerve ring, it tapers rapidly, like the mid-esophagus itself towards the rear of the mantle region. The posterior esophagus is also dark-pigmented, and has 4-6 longitudinal ridges. It opens into the stomach shortly behind the mantle cavity. The stomach (Fig. ЗЕ) is about 4 mm long, occupies about two-thirds of a whorl, and is typically cerithioid in layout. Topographically, it is not unlike that depicted for Turritella by Graham (1938). It has a short, wide style sac and a well-formed crystalline style. If the stomach is opened by a dorsal longitudinal cut, the esophagus is seen to enter at its right anterior end. An opening to the digestive gland (Fig. 3E, dd) and what appears to be a much reduced spiral caecum (Fig. ЗЕ, sc) are both near the esophageal opening. A broad sorting area (Fig. ЗЕ, sa) is in the ventral por- tion of the stomach. A thick cuticular gastric shield (Fig. ЗЕ, gs) lies to the left, directly across from the anterior opening of the intes- tine and style sac. The crystalline style is present in all freshly collected specimens and is a short, transparent dumb-bell shaped rod about 2.8 mm in length. The style sac (Fig. 3E, Ss) from which it emerges, although joined to the intestine, is a separate structure with a blind end. The major typhlosole (Fig. ЗЕ, t,) emerges from the intestinal groove and Curves around to the spiral caecum. A small minor typhlosole (Fig. ЗЕ, to) runs from the sorting area into the intestine. The intestine leaves the stomach at the left anterior end of the style sac region and coils back over the style sac before turning anteri- orly where it enters the mantle cavity. The portion of the intestine that exits from the stomach has a large dorsal typhlosole that is BIOLOGY OF MODULUS 127 gradually lost as the intestine nears the man- tle cavity where it has a smooth interior. It becomes ridged and more glandular in the mantle skirt. Reproductive system (Fig. 8).—Open palli- al gonoducts in both sexes, the lack of a penis in males and the formation of spermatophores by males are fundamental features of the re- productive system in Modulus modulus. The basic groundplan of the pallial gonoducts in Modulus is the same as in other cerithiaceans such as Cerithium, Bittium, Rhinoclavis and Batillaria but differs in the placement and аг- rangement of internal ducts and seminal re- ceptacles in the laminae of the gonoduct. The pallial gonoduct is best visualized as a tube- like duct, extending from the proximal end of the mantle cavity to the mantle edge, and slit open along its longitudinal axis. In both sexes the pallial gonoduct comprises lateral (right) and medial (left) laminae that are fused dor- sally to each other and to the lateral mantle wall. The ventral margins of the laminae are free and open to the mantle cavity. In Modulus both laminae have internal ducts and pouches used for the reception, trans- mission and storage of spermatophores and spermatozoa. Male reproductive tract (Fig. 8).—In males the laminae of the gonoducts are white, thin- walled structures (Fig. 8). The medial (left) lamina is membranous and semi-opaque and its inner surface has numerous, transversely oriented ridges. Near the axis of attachment there are wider, denser and less numerous ridges. The lateral (right) lamina (Fig. 8A, //) is fused along its axis to the mantle and partially on its left side to the epidermis of the visceral hump. It is opaque and thick and its inner sur- face has thick convoluted ridges. The male pallial gonoduct is slightly closed at its proxi- mal end due to the medial lamina folding over the posterior portion of the lateral lamina (Fig. 8A, pmg). This creates a closed pouch filled with numerous glandular axial ridges. This area of the gonoduct is probably the prostate gland (Fig. 8C, pg). Distal to the closed por- tion of the gonoduct is a thicker longitudinal glandular ridge and several wide axial glandu- lar folds that stain darkly with Methylene Blue. Adajcent to these folds are numerous smaller axial folds. This glandular area of the inner surface of the lateral lamina extends to the distal end of the gonoduct and is probably the spermatophore organ. Its exact mode of func- tioning is unknown. | have not seen a devel- oping or a completed spermatophore in the male gonoduct nor the method of transfer to the female, but it probably occurs via siphonal currents as | have described in Cerithium muscarum (Houbrick, 1973). In the aquarium | saw some spermatophores pass out the ex- halant siphon of a male and fall to the bottom. The bright orange testis lies on the outer dorsal surface of the digestive gland. The seminiferous tubules empty into a branching network of microscopic ducts that lead to about 10 vasa efferentia. These empty into the vas deferens that lies on the inside of the visceral coil and appears as a white duct that in ripe males is packed with eupyrene and apyrene sperm. SEM preparations of spermatozoa reveal that a eupyrene spermatozoon (Fig. 6, f) is about 36 um in length, has a pointed acro- some, a short, cylindrical, thick head and con- stricted neck region, and a long midpiece about one-fifth the length of the spermato- zoon. А long narrow flagellum follows the midpiece. Eupyrene spermatozoa taken from the vas deferens are attached by their acro- somes in small clusters. Apyrene spermatozoa (Fig. 6a, g) are larg- er, about 48 um in length, bear six flagella, have a spiral configuration, and very long pointed heads almost one-half the length of a sperm. It is not clear if there is a midpiece; consequently, this long head region may con- sist of head and midpiece. Although no counts were made, there appears to be an equal number of both kinds of sperms, tangled to- gether by the numerous flagella into dense masses. Thus the apyrene spermatozoa may serve to bind the mass together with their flagella. Apyrene sperm move in a slow sinu- ous manner while eupyrene sperm are fast moving. The spermatophores of Modulus are im- mobile, thinly walled, acellular structures of 1-3 mm length. Fresh spermatophores are white, shiny, crescent-shaped and swollen with spermatozoa. One end is round and pointed and the other more flattened and bi- furcate (Figs. 6, 7a-b). A prominent keel ex- tends axially along one side of the spermato- phore (Fig. 6a). The pointed end (Fig. 7c), when lodged in the female gonoduct, lies closest to its proximal end. lf pressure is ap- plied to a fresh spermatophore, sperm emerge from the lower branch of the bifurcat- ed end, where there appears to be a tube-like exit (Fig. 7b). Equal numbers of eupyrene and apyrene spermatozoa are contained in the spermatophore. HOUBRICK 128 FIG. 6. Spermatophore and sperm of Modulus modulus. SEM micrographs; a, Spermatophore, critical point dried and partially collapsed. Note distinctive axial keel; b, Spent spermatophore, freshly extracted from spermatophore receptacle, showing remaining sperm at bifurcate end; c-e, Details of torn portion of critical point dried spermatophore (as seen in Fig. 6a) showing densely packed sperm; f, Eupyrene sperm attached by acrosomes and showing long mid portion; g, Apyrene sperm demonstrating multiflagellate condition and spiral configuration of mid head piece. BIOLOGY OF MODULUS 129 FIG. 7. Spermatophore of Modulus modulus showing details of keel (a), bifurcate end (b), round pointed end (с), and acellular fibrous matrix comprising surface structure (d-e); f, Eupyrene and аругепе sperm extract- ed from seminal receptacle of female. SEM preparations of critical point dried spermatophores reveal a complex surface structure (Fig. 7а-е). Spermatophores are composed of long, axially oriented string-like fibers embedded in a matrix (Fig. 79-е). At the pointed end the fibrous matrix is closely bound together and the external surface 1$ relatively smooth. In the bifurcate part, the ends of individual fibers are free, creating a shaggy appearance (Fig. 7b). The edge of the keel is serrated due to free fiber ends that point away from the pointed tip. SEM pictures of a fractured spermatophore wall show that it is about 5 ит thick and composed of three layers: an outer fibrous layer, a thick middle layer filled with spongy-looking globules of HOUBRICK 130 “ajoe]daoa, jeulwas “Ys ‘21921d9991 ajoydojeunads “ds ‘иебло элоца -oyeuwads ‘ods ‘jonpouob ajew jo ed |ецихола ‘била :puej6 ajejsold “Bd ‘abessed шеш чоприло ‘po ‘eulwe| ¡epa ‘ww “eulue] jeipaw jo dif uy ‘eulwe| jeısje| jo эщеа ‘gy ‘емише! ¡esaye] 1 tejoejdaoas jeuiwes o] бшрез| 810016 ‘sh ‘ejoeydese1 asoydojeunads o] бырез| 210016 ‘/4$6 ‘810016 yonpouob ‘66 ‘Ayneo эциеш jo риа ‘owe ‘Jonpıno ¡en ed эешеэ; jo uonuod jejsıp “podp :Yonpouoß jeijed ajeuw jo uomod jejsıp ‘бшр :риеб ajnsdeo ‘59 ‘еше! ¡eipau jo эщеа wg “eumue] ¡e19]e, jo aoeuns payoeye ‘/е pue иэшпае ‘бе ‘(у yum эледшоэ) Jonajul летриеб Buisodxe хэеа рэпа eulwe| |е'рэш цим jonpouo6 ¡eyed ajen ‘2 'panowaı эозпш 1е|эшп!оэ цим YO] WOA) PAMAIA jonpino jeijjed эешэ; jo иоцеиэ$э19эл эцешецос ‘g (9 чим эледшоэ) рэлоше! э|озпш лецешгпюэ цим YA] шо; реэмеил Jonpouoß ¡enyed эеш jo uoyejuasaldas эцешецос ‘и ‘зпипрош snınpoyy jo зопроиоб |еез ‘8 “Old BIOLOGY OF MODULUS unknown composition, and a thin inner fibrous layer. Female reproductive tract (Figs. 3, 8).— The pallial oviduct is an opaque, cream- colored organ about 5.5 mm long and differs from the male gonoduct by its larger size and longer, swollen laminae. The bases of the laminae stain darkly in sections and are highly glandular. The medial (left) lamina (Fig. 8B, ml) has a thick lower portion composed of transversely oriented glandular folds adjacent to the line of axial fusion to the mantle. There is a construction in the medial lamina, about half-way to its outer edge, separating the low- er swollen, glandular part from the upper, more membranous part. An open slit in the upper portion of the lamina begins at the distal end and becomes a ciliated gutter (Fig. 8B, gspr). The highly ciliated gutter gets deeper and becomes a large, closed, internally ciliat- ed tube-like pouch occupying the proximal third of the medial lamina. Both the ciliated gutter and pouch comprise the spermatophore receptacle (Fig. 8B, spr) which is probably homologous to the “bursa copulatrix” or sperm-collecting pouch in Cerithium and Rhinoclavis (Houbrick, 1974a; 1978). Sec- tions show that the inner walls of the recepta- cle consist of columnar epithelial cells with darker staining glandular cells. Spermatophores are drawn by ciliary cur- rents into the mantle and enter the spermato- phore receptacle pointed end first. An individ- ual spermatophore fills the entire length of the spermatophore receptacle, its bifurcated end sticking out at the distal end. Both the ser- rated edge of the spermatophore keel and the minute free fibrous ends of its surface help to anchor it in the receptacle. Ап individual spermatophore receptacle normally holds one spermatophore but occasionally two may be found. Sperm emerge from the lower, bifurcate end of the spermatophore and move into an- other ciliated groove formed by a fold arising from a baffle on the middle part of the free edge of the medial lamina. This forms a long, deep, highly ciliated slit with an open edge facing and slightly overhanging the free edge of the lateral lamina. Sperm leaving the spermatophore are probably transferred by this ciliated groove to the ciliated gutter (Fig. 8B, gsr) of the lateral lamina leading to the seminal receptacle in the upper portion of the lateral lamina (Fig. 8B, sr). Spermatophores slowly dissolve within the spermatophore re- ceptacle and may be recovered in various stages of disintegration. 131 The lateral lamina (Fig. 88) has the same layout as the medial lamina: i.e., a thick glandular, attached axial base separated from the upper membranous, free part of the lami- na by an axial constriction. The upper mem- branous part of the lateral lamina also has the same groundplan as the membranous part of the medial lamina, but is mostly fused to the columellar muscle. Not far from the distal end of the lateral lamina, an open slit develops into a long ciliated gutter (Fig. 8B, gsr), that proximally becomes a closed pouch. This pouch is the seminal receptacle (Fig. 8B, sr) and is distended with sperm during the repro- ductive season, when its inner walls are glandular and villous. The morphology of the gutter and seminal receptacle of the upper lateral lamina is com- plicated by an involution and doubling back of the upper lateral, membranous wall of the ciliated gutter. This involution in the wall of the gutter assumes the form of a large flap of tis- sue (Fig. 8B), covering the top of the open slit and gutter at the upper part of the distal third of the gonoduct. The flap probably serves as a baffle which may prevent spermatophores from entering the seminal receptacle instead of the spermatophore receptacle. It may also guide released sperm into the seminal recep- tacle. The thick, lower, inner portions of both laminae form the albumen gland and the capsule gland. The albumen gland (Fig. 8B, ag) is located in the proximal part of the axial base of the oviduct. The capsule gland (Fig. 8B, cg) is a loosely compacted darkly staining area located in the middle portion of the later- al lamina. The pale green ovary consists of numer- ous, tiny, compact lobes. A thin-walled ovari- an duct, about 0.4 mm wide, extends from the ovary and runs forward along the ventral side of the viscera where it passes under part of the kidney and lies adjacent to the intestine and pericardial cavity before entering the mantle cavity. Oocytes from the ovarian duct are about 0.15 mm in diameter. Nervous system (Fig. 4).—The nervous system is similar to that of Cerithium and Rhinoclavis and is typically cerithioid. It is not as highly condensed as in other mesogastro- pods such as the rissoaceans. If one uses the “RPG” ratio of Davis et al. (1976: 263) (length of the pleurosupraesophageal connective di- vided by the sum of the lengths of the right pleural ganglion) to determine the state of condensation, a mean value of 0.59 is ob- tained which is a little above an intermediate 132 HOUBRICK TABLE 3. The RPG ratio for Modulus modulus. This ratio is the length of the pleuro-esophageal connective divided by the sum of the lengths of the supra-esophageal ganglion, pleuro-supraesopha- geal connective and right pleural ganglion. No. x Sd Range if 0.59 5.2] No., number of snails Sd, standard deviation x, mean 0.54—0.70 value (Table 3). As Davis pointed out, the higher the value the “looser” and more primi- tive is the nervous system. Distinctive char- acters are the lightly pigmented cerebral, pleural and pedal ganglia which are covered with tiny dots of tan color. The origin of the tentacular nerve is also ligntly pigmented and is slightly swollen. The cerebral ganglia are moderately fused to the pleural ganglia and have virtually no connectives (Fig. 4, rcg, rpg, Icg, Ipg). The cerebral-pedal connective is also short, about 0.25 mm long. The pedal ganglia are large, about 0.50 mm in length and deeply embedded in the muscle of the foot. The cerebral ganglia are of equal size (0.50 mm) and joined by a very short com- missure. The pleural ganglia are slightly smal- ler, each about 0.40 mm long. A schematic representation of the ganglia and their rela- tionship to other organs of the head are shown in Fig. 4. Circulatory and excretory systems.—There is nothing particularly distinctive about either of these systems in Modulus. Both kidney and heart are typically monotocardian and not un- like those of Littorina, described in detail by Fretter & Graham (1962: 34-35). It is note- worthy that the anterior aorta is very wide as it passes over the mid-esophagus. REPRODUCTIVE BIOLOGY The percentages of males and of females and statistics relevant to their shell measure- ments are presented in Table 1. Females are more numerous than males. Gametes have been described in detail in the sections on the reproductive tract. Most reproductive activity takes place from late winter through spring (February through May). Pairing.—Males become ripe in mid-winter and produce spermatophores from January through May. Females begin spawning in the spring. In May, pairing of males and females is frequent. The male's foot is attached to the female’s shell so that the two edges of their mantle cavities are adjacent. Spermato- phores leave the mantle cavity of males by the exhalant siphon and presumably enter the female’s mantle cavity via her inhalant siphon because males lack a penis or other organ of intromission. Not all spermatophores are suc- cessfully introduced: some drop to the bottom of the aquarium. The same phenomenon was seen in Cerithium muscarum (Houbrick, 1973) and may be due to artificial lab condi- tions. Spermatophores in the female's mantle cavity are moved by ciliary currents over the head-foot into the open ciliated gutter and spermatophore receptacle in the medial lami- na of the pallial oviduct. The flap on the distal part of the lateral lamina and the second cili- ated gutter of the medial lamina overhanging the lateral lamina probably prevent premature entry of the spermatophore into the seminal receptacle. Spawn.—Deposition of spawn was first seen in May, although spawning may have occurred in March and April when | was away from the study site. Spawn is deposited on grassy blades of marine angiosperms, such as Halodule, in worm-like gelatinous tubes about 14.10 mm in length, and 2.13 mm in width (Fig. 1k). Each spawn mass contains an average of 121 eggs. Within the gelatinous tubes fertilized eggs are enclosed in hyaline- capsules of 0.50 mm diameter, each capsule containing a single egg, 0.15 mm in diameter (Fig. 11, К). Although a few eggs contain tera- © tological embryos, there are no nurse eggs. Table 4 presents more statistics about egg masses. Individual spawn masses are cylindrical ex- cept for a narrow, flattened surface at the point of attachment to a grass blade. They are axially attached to grass blades (Fig. 1j) but occasionally curve or have a spiral configura- tion. The tough parchment-like outer wall of a spawn mass is frequently covered with fine sand grains and detrital particles, making it TABLE 4. Statistical summary of spawn mass di- mensions and of embryos per spawn mass. Statistic (n = 8) x Sd Range Length (mm) 14.1 5.05 9.3- 24.2 Width (mm) 2.13 0.26 1.8= 2.5 No. ofembryos 121 43.33 70 -169 x, mean Sd, standard deviation BIOLOGY ОЕ MODULUS 133 somewhat opaque. Internally, a spawn mass is viscous and comprises a matrix of spiral, gelatinous strands that contain tear-shaped, jelly-filled compartments each of which holds an egg capsule. In cross section, the entire spawn mass appears to be two gelatinous strands thick. There is no internal cavity evi- dent. Spawn masses are deposited in vast num- bers in the grass beds at Link Port. There 1$ an average of 360 spawn masses, comprising 43,560 embryos, per square meter of grass bed. Development.—Only a brief descriptive ac- count of the developmental process 1$ pre- sented below because a more detailed study was beyond the scope of this research. Development of Modulus modulus is direct. Each spawn mass contains capsules with embryos in various stages of development. That portion of the spawn first deposited by the female has more advanced embryos than the latter part, which may contain capsules having embryos in early cleavage stages. Thus, within a single spawn mass one may find fertilized eggs, early cleavage stages, blastula and gastrula stages and early veliger embryos, progressively arranged from one end to the other. Fertilized eggs and early cleavage stages are the same pale green color as eggs that emerge from the oviduct. Freshly laid spawn may be easily identified by the pale green embryos whereas older spawn masses lack this color. Embryos become light tan when they attain the mid-veliger stage. To study development, newly deposited spawn masses, laid on the evening of May 9, 1978 were placed in petri dishes with sea water and the growth of embryos from the first portion of each spawn mass was monitored until they hatched from their capsules. Early cleavage stages are about one-fourth the diameter of the egg capsule. As embryos develop into blastulas and gastrulas they be- come larger and continue to grow throughout the veliger and hatching stages until the embryonic shell occupies the entire capsule (Fig. 1j). Capsule diameter remains constant throughout all developmental stages. Cleavage 1$ rapid; the 8-cell stage was at- tained within 3 hours after deposition. Blastula and gastrula stages were attained on May 10 and May 11. By May 12, a shell gland and prototroch were present, indicating the begin- ning of the veliger stage. On May 15 early veliger stage embryos with cap-like proto- conchs, ornamented with the beginnings of a distinctive spiral sculpture, were present. A large, yolky digestive “anlage” was also pres- ent as well as the early pedal structure. Al- though early velar lobes were present, no stomodaeum was seen. Embryos at this stage begin to spin within their capsules. On Мау 16 the start of the second whorl of the protoconch was noted and two large, darkly pigmented statocysts appeared at the base of the head-foot region. By May 17, velar lobes were well-developed, tiny black eyespots, small cephalic tentacles and a stomodaeum appeared, and spinning of embryos was more rapid. On May 18 the foot was better defined and the protoconch comprised two whorls. No operculum was seen. The pale green color, so indicative of earlier embryonic stages, be- gan to fade and was completely gone by May 19 when embryos were a light tan color. Spin- ning of embryonic veligers, due to ciliary beat of the velum and foot, was more rapid but sharp jerky movements began and some- times individuals would periodically stop and resume spinning in a reverse direction. On May 19 the foot was further enlarged and an operculum was present. The larval heart could be seen pulsating within the mantle cavity. The external morphology of embryos ге- mained the same for the next five days, but internal development was obscured by their opaque color. By May 24 the late veliger stage was attained and embryos almost filled their capsules. Statocysts were no longer visi- ble, cephalic tentacles were large and the eyes enlarged. Embryos frequently stopped spinning as they probed the capsule wall. Velar lobes began to disappear, and the embryos began to look like tiny snails. Hatch- ing of juveniles began on May 27 and con- tinued through June 2 until the spawn mass was empty. Five other monitored spawn masses underwent similar development with- in the same time. Incubation of embryos lasts from 18-24 days. In general the Link Port population takes about 2.5 to 3.5 weeks to undergo direct development. Hatching.—Young snails escape as the capsule wall splits and breaks арай due to pressure from the foot and snout of the en- capsulated snail. Young snails do not leave the interior of the spawn by any definite route or exit but escape randomly. As they emerge from their capsules they crawl about the in- terstices of the spawn mass which becomes more viscous and begins to slowly disinte- grate. After hatching, young snails graze on 134 HOUBRICK detritus of the outer wall of the spawn mass and gradually move off onto grass blades. п May, thousands of juvenile snails, from 0.5- 1.8mm in diameter were observed in the grass beds (Fig. 1g-h). Although obviously of different size classes due to different hatching times and the long spawning period, these young snails may be considered to form one large group of young. Newly hatched snails quickly twist and turn as they explore the en- vironment. Young snails each have large black eyes, a typical radula and a long tactile propodium that bears a heavily ciliated groove along its anterior margin. Growth.—New growth of the protoconch occurs rapidly. Within a few days, a change in sculpture separates the embryonic shell from the new shell growth (Fig. 1h). This is marked by axial shell sculpture that lacks the spiral elements of the protoconch. The aperture be- comes angulate and flaring due to the ap- pearance of the median keel that is so indica- tive of adult shells. By mid June young snails have added a much larger whorl sculptured with spiral cords and a prominent median keel and have shells that range from 1.3-2 mm in diameter. The outer lip is sharply angulate and the distinctive columellar notch of adults is present (Fig. 1d— f). Observations were not made during the mid-summer months, but in September young Modulus were abundant in the grass beds. No living adults were found but numerous adult shells occupied by hermit crabs were seen. Adolescent Modulus, 2-5 mm in size, were covered with thick filamentous green algae and were difficult to detect in the grass beds. Shells were thick and had typical adult sculpture except for their outer lips which were thin due to recent growth. Animals had adult pigmentation and the internal anatomy appeared normal except for the incipient pallial gonoducts. The ctenidium was slightly smaller and comprised only 80 filaments. The pleuro-esophageal connective was slightly longer than in adults; consequently, the RPG ratio of adolescent snails is higher, about 0.70. This indicates a looser concentration of the nerve ring. The ganglia probably become more condensed as they grow larger. By December, snails have nearly reached adult size and males are ripe. Females be- come ripe in January. The pallial oviducts are developed but ovaries are just beginning to ripen. By spring (late February-early March) Modulus is reproductively mature and the shell has reached its maximum size. Adults appear to die after spawning, when many empty adult shells may be found. A summary of these data and other ob- servations indicate that the Link Port popula- tion of Modulus modulus has a life cycle of one year. Although the spawning period 1$ long and results in various cohorts of young, these overlap to form one large group that develops during spring and grows quickly throughout the year, reaching maturity in late winter-early spring, when spawning occurs. Developmental stages occur throughout the spring with the subsequent emergence of the new progeny and death of adults. These conclusions are supported by ex- amination of large monthly samples of benthic animals taken from other sites in the Indian River during the Indian River Study conducted by the Harbor Branch Consortium in 1973- 1974 (see Young et al., 1974). Growth stages of Modulus from these samples fit the general pattern given above. lt thus appears that populations of Modulus from other areas of the Indian River estuary have a similar life cycle that lasts about one year. ECOLOGY Modulus modulus is one of the more com- mon prosobranchs associated with marine grassbeds in Florida and the Caribbean. As an abundant primary consumer it is an impor- tant factor in the trophic structure of this eco- tope. lt is thus surprising that so little was known about its anatomy and ecology. General observations.—A thorough study of the autecology of Modulus was not at- tempted; nevertheless, some ecological ob- servations made during this study will provide information for future workers. The study site consisted of dense stands of Halodule wrightii Ascherson, and two other less common angiosperm grasses, Syringod- ¡um filiformis Knetz and Thalassia testudinum Kónig 8 Sims, all covered with dense epi- phytic growth which traps detritus. The entire site is rich with detritus, and the water 1$ fre- quently turbid with suspended particulate matter. The salinity in this habitat, normally 337.0, undergoes considerable variation sometimes within a short period of time due to heavy rainfall. Modulus occurs on the grass blades and occasionally on the substratum. The popula- tion studied lives at a depth of about one BIOLOGY ОЕ MODULUS 135 TABLE 5. Epiphytic algae growing on shells of Modulus modulus (* = dominant). Phaeophyta *Sphacelaria furcigera Kútzing Cyanophyta Microcoleus lyngbyaceus (Kútzing) Crouan Callothrix crustacea Thuret Rhodophyta Goniotrichum alsidii (Zanardini) Howe Chlorophyta Enteromorpha linguleta J. Agardh Cladophora sp. meter and is never exposed, even at extreme low tides. Shells of living Modulus are normally densely covered with algal epiphytes. A list of these is given in Table 5. The dominant epi- phyte is Sphacelaria furcigera Kützing. It is noteworthy that this alga does not grow on the sea grasses whereas all of the other epi- phytes on Modulus occur also on the grasses. Living Modulus snails frequently have an unidentified colonial hydroid growing on the bases and peripheries of their shells. The slipper shell, Crepidula fornicata (Linnaeus) may also occur on the base of the shell and barnacles occasionally are found on the shell top. Empty shells are common and are fre- quently utilized by the hermit crab Pagurus bonairensis Schmitt and the sipunculid Phascolion cryptus Hendrix. Food and feeding.—Modulus modulus is an active browser that engulfs microphytic algae and detrital particles. Like other micro- phagous, style-bearing mesogastropods it feeds more or less constantly. Mook (1977: 136) presented evidence that the grazing ac- tion of Modulus may retard the accumulation of fouling organisms by dislodging their newly settled larvae. Modulus thus aids in keeping the surfaces of seagrass blades clear and available as a substratum for microphytic algae. Stomach contents of freshly-collected spec- imens contained sand grains, occasional foraminifer tests, numerous diatoms, algae, detrital particles, and fragments of larger filamentous macro-algae. The bulk of the algal contents comprises diatoms and of these, the dominant species is Melosira moniliformis, а relatively large diatom. An analysis and identification of stomach con- tents is presented in Table 6. Fecal pellet analysis shows that detrital particles, sand, complete diatoms, and diatom fragments pass through the gut. The most common un- broken diatoms in fecal pellets were Nitzschia and Navicula species. The evidence of stom- ach, gut and fecal pellet contents leaves no doubt that Modulus ingests primarily diatoms. Analysis of the dominant diatoms suggests that larger diatoms are preferred to smaller ones. Although ingested food is not neces- sarily what is digested by the snail, it is prob- able that diatoms are the chief source of en- ergy. Gut contents of very young Modulus had no appreciable diatom content. This in- dicates that the young are feeding on different plant food. Associations and predators.—The most common prosobranchs co-existing with Modulus modulus on seagrasses are Cerith- jum muscarum Say and Bittium varium (Pfeiffer), both cerithiaceans and also style- bearing, algal-detritus feeders. Several other common snails found on the seagrass are the carnivores Mitrella lunata (Say) and Hami- noea elegans (Gray). The blue crab, Са/тесе$ sapidus Rath- bun was the only predator observed feeding on Modulus and many shells with chipped apertures point to heavy crab predation. Very few drilled shells of Modulus were found, sug- gesting that predation by naticid snails is in- significant. This is not surprising because Modulus is generally found on the grass blades and not as frequently on the substrate. Large rays were frequently seen in the grass beds and these along with other fish such as the sheepshead, Archosargus probato- cephalus Walbaum, are suspected as chief predators. The numerous young snails ob- served in the spring are thin-shelled and small and are probably eaten by a variety of pre- dators. Behavior—Modulus is a slow-crawling grazer and does not demonstrate a wide vari- ety of behavior. lt moves with retrograde, monotaxic muscular waves and sudden jerky motions. Although its normal habitat is on the blades of seagrass it will move down onto the substratum when weather conditions cause estuary waters to be rough. When irritated or attacked, Modulus strong- ly twists itself, turning its shell back and forth with the columellar muscle, as if to dislodge a predator. If the irritation continues, the hypo- branchial gland exudes a mass of sticky parti- cles that probably discourages predators. The animal will then withdraw completely into its shell, about 5 mm beyond the edge of the 136 HOUBRICK TABLE 6. Analysis of algal stomach contents of Modulus modulus. Mean size (M) Taxon Diameter Height Form Jan. Feb. Class BACILLARIOPHYCEAE Order Centrales Suborder COSCINODISCINEAE Family MELOSIRACEAE Melosira moniliformis O. F. Muller Melosira sulcata Ehrenberg Order PENNALES Suborder ARAPHIDINEAE Family FRAGILARIACEA Synedra sp. (fragments only) Striatella unipunctata Agardh (fragments only) Rhabdonema sp. 9.8 Suborder BIRAPHIDINEAE Family NITZSCHIACEAE Nitzschia Sp. 5:2 Family NAVICULACEA Mastogloia crucicula (Grunow) Cleve Mastogloia sp. Navicula sp. Family SURIRELLANCEAE Camplyodiscus sp. 95 chains e > - chains - + = chains + aa — chains — + x chains ES _ — 61 solitary = + + 10 solitary = solitary + — — solitary = = + solitary = _ — *Dominant species +Present —Absent outer lip, until the operculum snugly fills the aperture. Modulus reacts violently with the same twisting movements when it is exposed to secretions given off by other wounded Modu- lus snails, and will move rapidly away. This behavior has been documented in other marine and freshwater snails by Snyder (19675-1971): DISCUSSION A consideration of the interrelationships of higher taxa is contingent upon the amount and quality of comparative data available. Many familial definitions are based only on shell characters and there are few compre- hensive anatomical studies upon which to rely. | have used the anatomical data | found available for cerithiaceans, although this was frequently incomplete and/or contradictory. Thus my conclusions, while based on ana- tomical evidence, are tentative and partially speculative. On the basis of anatomy, | believe that the Modulidae should be regarded as a distinct family within the Cerithiacea. On balance, Modulus species share more anatomical characters in common with members of this superfamily than with any other group. They appear to be closest to the Cerithiidae and Potamididae. My reasons for these conclu- sions are developed in the following discus- sion. Phylogenetic relationships.—Modulus spe- cies differ conchologically from other ceri- thiaceans by their turbinate shape and um- bilical notch. The family is a small one (one genus and about 6 species) in comparison with other cerithiacean families such as the Cerithiidae, Potamididae, Dialidae, Turritel- BIOLOGY ОЕ MODULUS 137 laidae and Vermetidae, all comprising numer- ous genera and species. Risbec (1927: 17) believed that the family Modulidae was intermediate between the Cerithiidae and Strombidae and cited a num- ber of characters that he said were shared with each group. His remark that both cerithiid and Modulus species have short anterior siphons is incorrect: two genera of cerithiids, e.g. Rhinoclavis and Pseudovertagus, are characterized by long siphonal canals (Houbrick, 1978). Risbec (1927) also errone- ously reported that Modulus and the ceri- thiids lacked salivary glands. As | have dem- onstrated in this paper and others (Houbrick, 1974a, 1978) both Modulus and all cerithiids heretofore studied have salivary glands. Bright (1958: 135) has reported salivary glands in a potamidid, Cerithidea, and | have observed them in Batillaria. | donot agree with Risbec's (1927) opinion that Modulus has close affinities with the Strombidae. While his observation that the anterior position of the eyes on the tentacles of Modulus is shared with the strombs is correct, his citation of a crystalline style as a unique shared character is incorrect because a style is characteristic of most algal-detrital feeders in the Mesogastro- poda. He was apparently unaware that many cerithiaceans have styles. Risbec's (1927) citation of an osphradium with indistinct lamel- lae as a shared character between the Modu- lidae and Strombidae is accurate, but among the Cerithiinae, the Potamididae also have a similar osphradium (personal observation). The similarity of the radula between Modulus and Strombus species is superficial: any simi- larity is probably due to convergence of gen- eralized taenioglossate radulae adapted for feeding on epiphytic algae and detritus. The reproductive tract of Modulus is very different from that of Strombus. The open condition of the pallial gonoducts in both sexes and a lack of a penis in males are conservative cerithi- acean characters not seen in the Strombidae. Although comparison may be made be- tween the sudden jerk-like motions of a crawl- ing Modulus and those of Strombus, the movement of the former is more like that of other cerithiids, only more pronounced, and is in no way similar to the jumping motions of Strombus. Aside from the turbinate shape of the shell, Modulus also has some distinctive anatomi- cal features. Among these are short, stout pallial tentacles that reach extreme develop- ment in M. tectum. Pallial tentacles are also present in cerithiid genera such as Cerithium, Rhinoclavis, Pseudovertagus, Clypeomorus and Bittium, but in relation to the animals' body they are never of the same size as are those of Modulus. Another distinctive feature of the external anatomy 1$ the forward position of the eyes on the tentacles. In the Cerithi- idae, Potamididae, Turritellidae and Vermet- idae, the eyes are located at the bases of the cephalic tentacles. Although the Strombidae are similar to the Modulidae in regard to eye placement, this does not necessarily indicate close relationship, as noted before. An unusual feature of the mantle cavity in Modulus modulus is the large and highly glandular hypobranchial aland. The ability of this gland to exude salvos of sticky mucoid particles and discharge them via the exhalant siphon 1$, to my knowledge, unrecorded for other cerithiaceans. | assume that this behav- ior, coupled with quick twisting movements of the body, is a deterrent to predators. Although the hypobranchial gland of M. tectum is not exactly the same, the identical twisting behav- ¡or was noted. Another noteworthy feature of the mantle cavity of Modulus is the sinuously twisted distal portion of the osphradium that ends near the entrance of the inhalant siphon. The osphradium is a ridge-shaped structure much like the osphradium seen in the mem- bers of the Potamididae, while in species of the Cerithiidae, it is bipectinate. The placement of the salivary glands and their ducts anterior to the cerebral commis- sure is a noteworthy feature of Modulus. The salivary ducts do not appear to pass through the nerve ring, but sections show that they begin very close to it. Although this arrange- ment is unlike that of many monotocardians, it is shared by the cerithiid genera Cerithium, Rhinoclavis, Pseudovertagus and Clypeo- morus (personal observation) and has been described by Davis et al. (1976: 276) in mem- bers of the rissoacean families Assimineidae, Truncatellidae, Bithyniidae and Hydrobiidae. In Modulus, the passage of a portion of the left salivary gland through the nerve ring and partially behind the cerebral commissure, shows that the Modulidae stand in an inter- mediate position among the Mesogastropoda in regard to this trait. Some Cerithium species also have a similar arrangement of the left salivary gland. Bright (1958: 134) found that the salivary glands and ducts of the potami- did, Cerithidea californica (Haldeman, 1840) were located behind the nerve ring next to the “crop” (esophageal gland). He noted that the 138 HOUBRICK left gland was the largest and that the salivary ducts were highly convoluted and partially embedded in the connective tissue of the “preesophagus” (anterior esophagus). It is not clear from his statement that the ducts pass through the nerve ring but his figure indi- cates this is the case. Thus, while most meso- gastropods have their salivary glands located behind the cerebral commissure they lie an- teriorly in some rissoaceans and have an in- termediate position in the Modulidae and in many Cerithiidae. This supports Davis' (1976: 276) position that location of salivary glands and their ducts is a poor character to differen- tiate mesogastropods from stenoglossan neogastropods. The combination of a well-developed esophageal gland and a crystalline style in Modulus, although thought to be unusual in mesogastropods (Fretter & Graham, 1962: 220), is a common condition shared with cerithiids | have examined in the genera Cerithium, Rhinoclavis, Pseudovertagus and Clypeomorus. Sections of this large gland in Modulus clearly show numerous and deep glandular lateral outpouchings arising from the mid-esophagus, leaving no doubt about its function. The nervous system of Modulus differs from that of Cerithium or Rhinoclavis in lack- ing a large siphonal ganglion, but this may be due to the short siphon of Modulus. It is inter- esting to compare the RPG ratio of Modulus with that of other marine mesogastropods. Davis et al. (1976: 267) presented a table of RPG ratios which compared selected hydro- biid, rissoid and littorinid taxa. As mentioned previously, the higher the value of the RPG ratio the less concentrated are the ganglia of the nerve ring and presumably the more prim- TABLE 7. The RPG ratio for selected cerithiid, modulid and potamidid taxa. (This ratio is the length of the pleuro-esophageal connective divided by the sum of the lengths of the supraesophageal gangli- on, pleuro-supraesophageal connective and right pleural ganglion). Taxon RPG Cerithiacea Cerithiidae Cerithium lutosum Menke 0.59 Cerithium atratum (Born) 0.59 Modulidae Modulus modulus (Linnaeus) 0.59 Potamididae Batillaria minima (Gmelin) 0.77 itive the nervous system. п Table 7, | com- pare the ratios of selected cerithiacean taxa. The mean value for Modulus is the same as in Cerithium, 0.59, but both of these taxa have lower values than the potamidid, Batillaria minima, which has a value of 0.77, closer to the littorinid value. Modulus has one of the more complex pal- lial oviduct systems found in the Cerithiacea. The open pallial gonoducts of some Cerithi- acea have been surveyed by Johansson (1953; 1956), Fretter (1951), Fretter 8 Graham (1962: 625) and Houbrick (1971; 1974a, 1977, 1978). Although all of the spe- cies studied have a basic groundplan of open pallial gonoducts, those of Bittium and Cerithiopsis are very complex in organization. However, the allocation of Cerithiopsis within the Cerithiacea is uncertain. The layout of the pallial gonoducts of Modulus are even more complex and unusual. While one can interpret the pallial gonoducts of the Cerithiacea from a functional viewpoint, it 15 difficult to relate these open systems to each other in a com- parative systematic manner. As Johansson (1953: 8) pointed out, pallial gonoducts some- times differ considerably, even in closely re- lated species. This has been seen in several rissoaceans (Johansson, 1953) and | have found many differences in the arrangement of the sperm gutter, bursa copulatrix and semin- al receptacle in different species of Cerithium, Clypeomorus and Rhinoclavis. Too little is known of the arrangement of the gonoducts in other members of the Cerithiacea to discuss their comparative anatomy satisfactorily; moreover, the epithelial origin of pallial gono- ducts renders any homologies suspect. Thus, any attempt to decide what is a primitive or derived state would be premature and purely speculative. Nevertheless the unique ground- plan of the pallial gonoducts of Modulus clear- ly separates the Modulidae from other ceri- thiacean families and 1$ a reliable discriminat- ing character. | have observed spermatophores in other cerithiid members of the genera Cerithium, Rhinoclavis, and Gourmya and suspect that this method of sperm transfer is characteristic of the group. Spermatophores have also been described by Dazo (1965) in the freshwater cerithiacean Goniobasis. The spermatophore of Modulus has a more complex surface structure and shape than those | have seen in the cerithiid species. The general physiog- поту of the sperm of Modulus and in particu- lar the elongate midpiece of the eupyrene BIOLOGY ОЕ MODULUS 139 spermatozoon, are comparable to those de- scribed for other cerithiids, e.g. Cerithium (Tuzet, 1930; Houbrick, 1973), Bittium and Cerithiopsis (Fretter & Graham, 1962). Interspecific comparisons.—Aside Нот the observations presented in this paper and those recorded on the anatomy of Modulus tectum by Risbec (1927) (cited as Modulus candidus Petit), nothing is known of the ana- tomy or ecology of other Modulus species. | was able to study Modulus tectum in Fiji and offer the following observations for compari- son with M. modulus. Modulus tectum differs from M. modulus by living on hard substrates in coral reef habitats. It is a much larger snail and not abundant. Modulus tectum clamps tightly on dead coral rubble when disturbed and is dislodged with difficulty. It has a trans- parent operculum through which may be seen the foot, brightly colored with large orange spots and irregular white blotches on a black pigmented background. These bright colors and the eye-like spots may startle predators who are able to remove snails from the rocks. Males are smaller than females and both sexes have highly colored soft parts. The snout is dark brown and spotted with white while the tentacles are reddish and papillate. The eyes are located near the distal ends of each tentacle, as in M. modulus. The foot is reddish brown and covered with whitish blotches and small brown dots. A large pro- podial furrow is present. The foot is much broader than in M. modulus and serves to keep the animal tightly clamped to the sub- strate. The mantle edge has long pinnate or multi-branched papillae in contrast to the simpler mantle papillae of M. modulus. The hypobranchial gland is smaller than that of M. modulus. When irritated, M. tectum does not emit salvos of sticky mucoid particles as does М. modulus. Modulus tectum, however, makes the same violent twists as M. modulus when irritated, but the former also clamps down on the substratum to discourage preda- tors. The thick, open pallial oviduct of Modulus tectum is highly glandular and the outer lami- na is internally more convoluted with glandu- lar folds than that of M. modulus. The male pallial gonoduct has a large yellow glandular area located on the inner lamina adjacent to the columellar muscle. This may be homolo- gous to the structure seen in M. modulus and is probably a spermatophore-forming organ. Females of M. tectum have a large ovipositor on the mid right side of the foot that is identical to that | have described in M. modulus. Within the buccal cavity of M. tectum, lying anterior to the nerve ring, are paired salivary glands with ducts that empty into the anterior esoph- agus. The radula is very much like that of M. modulus. Behind the nerve ring the esoph- agus rapidly widens and a large, wide, choco- late-colored esophageal gland is present. The large black kidney is covered with a network of fine white branch-like blood ves- sels. The testis is light green and a large vas efferens runs from the testis to the pallial gonoduct along the inside of the whorls. п females the ovary is dark green and filled with large green ova that suggest direct develop- ment may take place. Green ova were also seen in M. modulus. Reproductive biology.—Spawn masses of Modulus modulus are deposited at Link Port at the same time as those of Cerithium muscarum. Both species are abundant on seagrasses, undergo direct development, and their spawn is somewhat similar. Thus, spawn masses of both species may be confused and future workers should be alerted to this fact. They can be easily separated with careful ex- amination. The spawn of Modulus is deposit- ed along the axis of a seagrass blade as a cylindrical tube with smooth outer walls and resembles a caterpillar, whereas that of Cerithium muscarum is a more irregular, disk- like mass with a lumpy outer surface. | have figured the spawn of Cerithium muscarum elsewhere (Houbrick, 1973: 880; 1974a: 78). The embryos and young snails of the two species are also different. Cerithium embryos do not have the pale green color of Modulus embryos during their earlier stages of devel- opment. The embryonic shells and proto- conchs of newly hatched Cerithium muscar- um are characterized by the purple color of the outer lip and umbilical region. The aper- ture is oval while in Mogulus it is more angu- late. Finally, the embryonic shell and proto- conch of Cerithium muscarum lack the spiral striae so characteristic of Modulus. The reproductive mode of the Link Port population of Modulus modulus is direct. It is interesting to note that Lebour (1945: 470- 471), in Bermuda, and Bandel (1976: 258) in Santa Marta, Colombia, each observed indi- rect development in Modulus modulus. These observations are contrary to what | have seen and present a discrepancy for which there are several possible explanations: 1) Both Lebour (1945) and Bandel (1976) were mistaken about the identity of their specimens; 2) What 140 HOUBRICK has been called Modulus modulus in the stenohaline environments of Bermuda and the Caribbean is a separate species; 3) Modulus modulus has two developmental modes. The first explanation is unlikely because Modulus carchedonius, which has planktonic larvae, has not been recorded from Bermuda and Bandel (1976) described the spawn of both species of Modulus at Santa Marta. The second explanation, while possible, does not seem to be likely. Although consid- erable variation in shell sculpture occurs be- tween populations of Modulus modulus, there are intergrades between the various morphs indicating that there is but a single species. Abbott (1944: 4) believed that these differ- ences were within the normal range of varia- tion and regarded them as one species. | have examined the extensive collections at the National Museum of Natural History and concur with him. | have examined the anato- my and radulae of the Caribbean forms and find no differences. | believe that the third explanation is the most probable, and that Modulus modulus has a reproductive strategy comprising two developmental modes, direct and indirect. These modes are correlated with euryhaline and stenohaline environments, respectively. The utilization of direct development by estu- arine populations such as those at Link Port may provide better protection against sudden changes in salinity or exposure and thus en- hance the maintenance and survival of the embryos. Although not a common phenome- non, there are seven documented cases of a single species having both direct and indirect developmental modes. These have been re- viewed by Robertson (1974: 227). Intraspecif- ic variation of developmental modes has also been recorded among other invertebrates such as echinoderms and polychaetes. Modulus carchedonius is the only other species for which eggs and larvae have been described. This species has free-swimming planktonic larvae. The spawn masses of this species differ from those of M. modulus by having an inner cavity formed by the outer walls of the egg mass. The spawn masses of M. carchedonius contain many more embry- os (7,000) that hatch in 5-6 days as free swimming veligers (Bancel, 1976: 259-260, fig. 12,a,:D): The populations of Modulus modulus that Lebour (1945) and Bandel (1976) observed had smaller and more numerous eggs, and a short incubation period of 5-7 days prior to hatching, typical of indirect development. In these populations the egg capsules dissolved and the veligers swam into the hollow center of the spawn mass and then to the outside. Spawn masses of the Link Port population did not have hollow centers. The longer develop- mental time of the Link Port population 1$ simi- lar to those | have observed in Cerithium lutosum and C. muscarum, euryhaline spe- cies which also have direct development (Houbrick, 1973, 1974b). The developmental mode of Modulus modulus is much like those of Cerithium muscarum Say and Cerithium lutosum Menke (Houbrick, 1973). The veliger stage of Modu- lus modulus was reached in five days, about the same rate observed in C. muscarum, and the total incubation period of the former (2-3 weeks) is likewise similar to those of the Cerithium species cited above. Cerithium muscarum shares many of the reproductive and developmental patterns seen in Modulus modulus such as open pallial gonoducts, spermatophores, dimorphic sperm, spawn masses and direct development. Moreover, both species are abundant and occur together in the same habitat. This is probably ex- plained by a common reproductive anatomy imposed by similar phyletic origin as well as convergence due to similar ecology. ACKNOWLEDGEMENTS Most of this study was done at the Smith- sonian's Fort Pierce Bureau, Fort Pierce, Flor- ida. | thank Dr. Mary Rice for her assistance while using these facilities. | extend special thanks to Julie Piraino for her help with labo- ratory equipment. | also thank Joseph Mur- doch and Dr. John Pilger for their help both in the lab. and field. Pat Linley, of Harbor Branch Foundation, assisted with histological prepa- rations. Grateful thanks are extended to Ron Mahoney, Robert Gibson and Penny Hall of Harbor Branch Foundation, who kindly identi- fied the algae for me. | thank Dr. Uday Raj, University of The South Pacific, Suva, Fiji, for lab. space and assistance in collecting. June Jones kindly typed portions of the manuscript. | also wish to thank Susann Braden, SEM La- boratory, National Museum of Natural History, for assisting with operation of the SEM. Victor Krantz, of the National Museum of Natural History Photography Lab, took pictures of BIOLOGY ОЕ MODULUS 141 shells. Cathy Lamb assisted with proofing the manuscript, and Mary Parrish typed the final draft. This 15 Smithsonian's Font Pierce Bu- reau contribution number 41. LITERATURE CITED ABBOTT, R. T., 1944, The genus Modulus in the western Atlantic. Johnsonia, 14: 1-6, 2 pl. ADAMS, A., 1851, A monograph of Modulus, a genus of gasteropodous Mollusca, of the family Littorinidae. Proceedings of the Zoological So- ciety of London for 1850, 18: 203-204. BANDEL, K., 1976, Observations of spawn, em- bryonic development and ecology of some Caribbean lower Mesogastropoda. Veliger, 18: 249-271. BRIGHT, D. B.; 1958, Morphology of the common mudflat snail, Cerithidea californica. Bulletin of the Southern California Academy of Science, 57: 127-139. DALL, W. H., 1892, Contributions to the Tertiary fauna of Florida with especial reference to the Miocene Silex-beds of Tampa and the Pliocene beds of the Caloosahatchee River. Transactions of the Wagner Free Institute of Science, Phila- delphia, 3: 1-947, 35 pl. DAVIS, G. M., KITIKOON, V., & TEMCHAROEN, P., 1976, Monograph on “Lithoglyphopsis” aperta, the snail host of Mekong River schisto- somiasis. Malacologia, 15: 241-287. DAZO, B. C., 1965, The morphology and natural history of Pleurocera acuta and Goniobasis livescens (Gastropoda: Cerithiacea: Pleuro- ceridae). Malacologia, 3: 1-80. FRETTER, V., 1951, Observations on the life history and functional morphology of Cerithiopsis tubercularis (Montagu) and Triphora perversa (L.). Journal of the Marine Biological Association of the United Kingdom, 29: 567-586. FRETTER, V. & GRAHAM, A., 1962, British proso- branch molluscs, their functional anatomy and ecology. Ray Society, London, 755 p. GRAHAM, A., 1938, On a ciliary process of food- collecting in the gastropod Turritella communis Risso. Proceedings of the Zoological Society of London, ser. A, 108: 453-463. HALDEMAN, S. S., 1840, A monograph of the freshwater Limniades and other freshwater uni- valve shells of North America, No. 1: 1-16, 5 pl. HOUBRICK, R., 1971, Some aspects of the anato- my, reproduction, and early development of Cerithium nodulosum (Вгидшеге) (Gastropoda Prosobranchia). Pacific Science, 24: 560-565. HOUBRICK, R., 1973, Studies on the reproductive biology of the Genus Cerithium (Gastropoda: Prosobranchia) in the western Atlantic. Bulletin of Marine Science, 23: 875-904. HOUBRICK, R., 1974a, The genus Cerithium in the western Atlantic. Johnsonia, 5: 33-84. HOUBRICK, R., 1974b, Growth studies on the genus Cerithium (Gastropoda: Prosobranchia) with notes on ecology and microhabitats. Nautilus, 88: 14-27. HOUBRICK, R., 1977, Reevaluation and new de- scription of the genus Bittium (Cerithiidae). Veliger, 20: 101-106. HOUBRICK, R., 1978, The family Cerithiidae in the Indo-Pacific, Part 1: The genera Rhinoclavis, Pseudovertagus and Clavocerithium. Mono- graphs of Marine Mollusca, No. 1: 1-30. JOHANSSON, J., 1953, On the genital organs of some mesogastropods: Cerithium vulgatum Brug., Triphora perversa (L.) and Melanella (Eulima) intermedia (Cantr.). Contributions to the Prosobranchia. Zoologiska Bidrag fran Uppsala, 30: 1-23, pl. 1. JOHANSSON, J., 1956, On the anatomy of Tympanotonus fuscatus (L.), including a survey of the open pallial oviducts of the Cerithiacea. Atlantide Report, No. 4: 149-166, pl. 1. LEBOUR, M. V., 1945, The eggs and larvae of some prosobranchs from Bermuda. Proceed- ings of the Zoological Society of London, 114: 462-489. LINNAEUS, C., 1758, Systema naturae per regna {па naturae. Editio decima, reformata. Stock- holm, vol. 1, Regnum animale, 824 p. MARCUS, E. & MARCUS E., 1964, On Cerithium atratum (Born, 1778). Bulletin of Marine Sci- ence, 14: 494-510. MOOK, D., 1977, Studies on fouling invertebrates in the Indian River, Florida 2: Effects of Modulus modulus (Prosobranchia: Modulidae). Nautilus, 91: 134-136. REEVE, L. A., 1865, Conchologia Iconica: or illus- trations of the shells of molluscous animals. Vol. 15, Cerithium, 20 pl. + index. RISBEC, J., 1927, De l'anatomie de trois strombidés et du Modulus candidus, Petit qu'on trouve vivant sur la côte de la presqu'île de Noumea. Annales du Musée d'Histoire Naturelle de Marseille, 21: 1-18, pl. 28. ROBERTSON, R., 1974 Marine prosobranch gastropods: larval studies and systematics. Thalassia Jugoslavica, 10: 213-238. SNYDER, N. 1967, An alarm reaction of aquatic gastropods to intraspecific extract. Cornell Ex- periment Station Memoir 403, 122 p. SNYDER, М. & SNYDER, H., 1971, Pheromone- mediated behavior of Fasciolaria tulipa. Animal Behaviour, 19: 257-268. : SOWERBY, G. B., 1855, Thesaurus Conchyliorum, or monographs of genera of shells. Vol. 2, Cerithium: 847-859, pl. 176-186. TRYON, G. W., 1887, Manual of conchology; structural and systematic; with illustrations of the species. Ser. 1, 9: Cerithium, р. 127-149, pl. 20- 29. TUZET, O., 1930, Recherches sur la sperma- togenese des prosobranches. Archives de Zoologie Expérimentale et Générale, 70: 95- 229, 12 pl. YOUNG, D., AVENT, R., BRIEL, L., ELSEMAN, N., 142 HOUBRICK GORE, R., JONES, R., KERR, G., SEIBERT, H., VAN ZWECK, O., 8 WILCOX, J., 1974, Indian River study. First annual report, Harbor Branch consortium. Unpublished manuscript, 2 vols., 385 p. YOUNG, D., BUZAS, M., 8 YOUNG, M., 1976, Species densities of macrobenthos associated with seagrass: A field experimental study of pre- dation. Journal of Marine Research, 34: 577- 592. MALACOLOGIA, 1980, 20(1): 143-151 ON PATRO AUSTRALIS WITH COMPARISONS OF STRUCTURE THROUGHOUT THE ANOMIIDAE (BIVALVIA) C. M. Yonge Department of Zoology, University of Edinburgh, West Mains Road, Edinburgh EH9 3JT, Scotland ABSTRACT Study of specimens of Patro australis has involved further consideration of the Anomiidae,! in particular the extent of the differences between the more primitive Pododesmus (Monia) and the more modified Anomia. These are now extended to include comparisons of visceral structure. In Pododesmus the visceral mass surrounds the centrally placed byssal apparatus, the two (very similarly sized) gonads arranged in the same longitudinal plane. The greater effects of lateral compression in Anomia involve restriction of the visceral mass to posterior and ventral surfaces of the byssal apparatus with great reduction of the left gonad but extension of the right gonad into both right and left mantle lobes. The distinctive features of Heteranomia are seen to include enclosure of the base of the foot within the left gonad. Patro australis resembles Anomia in most respects, the major differences being conchological, but is adapted for life on a more uneven surface and under more turbid conditions. INTRODUCTION In a recent general account of the super- family Anomiacea (Yonge, 1977), very signifi- cant differences in structure (although not in habit, both living closely applied to flat sur- faces) were found between species of what proved to be much the less structurally spe- cialized Pododesmus (Monia) and those of the much more highly modified Anomia. Heteranomia with much the same mode of life differs from both but primarily in ctenidial structure. Enigmonia, however, despite its remarkably modified form and unique mode of life—a kind of bivalve limpet spending much time out of water on leaves and stems in the extremely damp atmosphere of mangrove swamps—has the same basic structure as Anomia. Patro australis (Gray) was also examined but only by way of empty shells obtained from the British Museum (Nat. Hist.). This species is stated by Iredale (1939) to occur all around the north of Australia but Beu (1967) extends this south to Victoria on the east. The obser- vations of both were confined to shells, Beu stating of P. australis that “It appears that these forms occupy some situation where the shell is able to grow unrestricted.” Personal conclusions were to the same effect, the area of byssal attachment being somewhat more limited than in Anomia spp., the valves much more irregular with the posterior margins raised clear of the surface to facilitate egress of faeces and especially pseudofaeces. A drawing was made showing this suggested posture (Yonge, 1977, fig. 21). More recently four preserved specimens of this species have been obtained from the Australian Museum, Sydney. The locality and habitat, given on the label, are “Saibai Vil- lage, Saibai, Torres Strait, N. Queensland, muddy sand and rock flat in front of village. Low tide 7 July, 1976.” This species was orig- inally described by Gray (1847) from the col- lections of the “Fly” as Anomia australis and then renamed by him in his review of the “Anomiadae” (Gray, 1849) as Patro elyros, this corrected to Patro australis by lredale (1939). All taxonomic data have been based on shell characters and the initial purpose of this study was to check differences in internal structure between this and other anomiids. This has inevitably involved some re-exami- nation and comparisons of structure in Pododesmus cepio, Anomia ephippium, A. simplex and Heteranomia squamula (the four species earlier examined) with results which extend previously published conclusions (Yonge, 1977). Further data are provided about the nature and significance of structural modifications in the Anomiidae. TAs restricted to exclude Placunanomia and Placuna (Yonge, 1977). (143) 144 УОМСЕ ABBREVIATIONS USED IN THE FIGURES a anus ad adductor adc catch muscle of adductor ааа quick muscle of adductor aol anterior outer ligament layer april anterior pedal retractor, left aul auricle, left aur auricle, right bn byssal notch br byssal retractor by byssus cr crurum cs crystalline style sac ctl ctenidium, left ctr ctenidium, right f foot hg hypobranchial gland il inner ligament layer isl boundary of inner shell layer m mouth mbn membrane around byssal notch mi mantle isthmus obn opening of byssal notch 09! oral groove, left ogr oral groove, right pbr posterior byssal retractor pol posterior outer ligament layer pl labial palps, left pr labial palps, right u umbo V ventricle vg visceral ganglion vma visceral mass, anterior vmp visceral mass, posterior vmv visceral mass, ventral SHELL STRUCTURE The shell in the Anomiacea 15 described by Taylor, Kennedy 8 Hall (1969) as consisting in the main of calcitic, foliated structure, the restricted inner shell layer (Fig. 3, isl) is of агадоп с, complex crossed-lamella structure, the muscle scars which they surround “leave a trace of aragonite, prismatic myostracum through the complex crossed-lamellar layer.” Their observations were largely confined to species of Anomia but Beu (1967) finds that in the three species of Patro which he examined the here somewhat thicker right (under) valve is prismatic, a notable distinction from Anomia. He also notes that the shell in all these species is more inflated than in Anomia or Pododesmus (those of Enigmonia and Placuna even more flattened). He noted that the shells were usually “regularly subcircular with the left (upper) valve “slightly saddle- shaped.” His largest specimen was 60 mm long with the right valve 2 mm thick. Of the four specimens here available, all were rough- ly circular in outline with the greatest diam- eters ranging from 18 to 45 mm. All were in- flated. One, shown in Fig. 1, was attached to a stone to the rounded side of which it con- formed although projecting marginally. The right valves of all were internally convex (Fig. 6) indicating a general tendency to settle on rounded rather than flat surfaces. The upward extensions of the posterior margin, although present, were not so well marked in these specimens as they were in the shell originally figured. The upper surface of the left valve (Fig. 2) has what Beu describes as a “regular fine sculpture of radial ribs, of which every tenth is stronger than the others.” Internally (Fig. 3) the restricted white inner shell layer down the centre is largely occupied by three large mus- cle scars, the most ventral that of the ad- ductor, the other two those of the divided posterior byssal retractor. The larger (more dorsal) one (br), as in Anomiidae generally, is larger than that of the adductor (ad). In addi- tion, just below the anterior margin of the resilifer, there is the scar of the small left an- terior pedal retractor. Conditions here are similar to those in Anomia although, as Beu FIG. 1. Patro australis,animal attached to rounded surface. РАТНО (ANOMIIDAE) 145 points out, the three major scars are some- what more vertically arranged; the anterior retractor scar (арп) 15 also somewhat larger. In Pododesmus the byssal retractor is only rarely divided. In Heteranomia the adductor and posterior pedal retractor scars merge into one as pointed out by Winckworth (1922) who separated the genus on this basis together with the distinctive ctenidia described by Ridewood (1903) and Atkins (1936). The characteristic deep notch on the right valve 1$ distinctly smaller in Patro than in spe- 3cm FIG. 2. P. australis, left valve, outer surface (from Yonge, 1977). lem р ñ e isl FIG. 3. P. australis, left valve, inner surface show- ing ligamental attachments (inner layer hatched, outer layers black) with inner shell layer containing three major muscle scars. bn 5mm FIG. 4. P. australis, inner surface of dorsal region of right valve showing byssal notch with crurum and attachment of inner ligament layer. cies of Pododesmus and Anomia. Moreover, as shown in Figs. 4—6 the notch is always wide open which is not true for either of the other genera where it is often completely closed by approximation, and sometimes fu- sion, of the anterior margin with the crural area. There also the calcified byssus may be- come fused with the margins of the notch (or foramen as it may become). There is no evi- dence that this happens in Patro where the animal has a more restricted area of attach- ment. The subcrural groove on the under face of that structure (permitting dorsal extension of byssal fibres) which is so very well marked in Pododesmus (Yonge, 1977) is absent in Patro and there is no attachment of a right anterior pedal retractor to the inner base of the crurum. The only muscle scar on the right valve is that of the adductor. In all three re- spects Patro resembles Anomia. LIGAMENT Owing to the extent to which the left valve overarches the right valve dorsally, the liga- ment in the Anomiidae is vertically instead of laterally disposed. It is topographically ex- tended horizontally, parallel to the substrate. 146 УОМСЕ 5mm FIG. 5. P. australis, intact animal, under (right) view of byssal region showing widely open byssal notch with united anterior and posterior outer ligament layers on under (outer) side of crurum. L 1cm j FIG. 6. P. australis, right valve viewed from dorsal aspect showing convex form also crurum with com- plete oval-shaped ligament, position of union of anterior and posterior outer layers indicated by broken line. As previously shown (Yonge, 1977) this over- arching by the left valve involves a “supra- dorsal” extension of the mantle margins at both ends of the ligament. This results in se- cretion of shell outside the ligament and con- sequent displacement of the umbo from the margin (Fig. 7). At the same time the anterior and posterior outer (lamellar) ligament layers are bent back topographically below the inner ligament layer secreted by the mantle isthmus (Figs. 3-6). Conditions on the right (under) valve are profoundly influenced by the presence of the relatively enormous anomiid byssal notch which, as shown in Figs. 4 and 6, stretches within and to the anterior of the ligamental region. In consequence on this valve the resil- ifer surface occupies the summit of a unique anomiid type of chondrophore known as a crurum. In side view this is straight both in Patro (Fig. 4) and Anomia, unlike Pododes- mus where it is convex. Although there cannot be supradorsal extension of the shell on this side, the outer ligament layers (secreted by epithelia which stretch between the valves) are inevitably bent in an oval around the top- ographically under side of the crurum. The re- sultant form of the ligament is best realized by viewing a right valve from the dorsal aspect (Fig. 6). The flat resilifer surface of the crurum bears the inner ligament layer (il) on its upper (i.e. interior) surface and the united anterior and posterior outer layers (ad, pol) on the lower (i.e. outer) surface. In life these layers are, of course, continuous with those on the upper, left valve shown in Fig. 3. Conditions, however, are somewhat more advanced in Patro than Anomia, indeed there is an interesting gradation of ligamental struc- ture starting with that in Pododesmus and in- dicated semi-diagrammatically in Fig. 7. In Pododesmus (А) initial supradorsal extension of the mantle margins and so of the outer valve layers they secrete is followed by their retreat with later decay of this region. This involves breakdown of the shell marginal to mi | С FIG. 7. Diagrams comparing ligamental conditions in— A, Pododesmus, incomplete supradorsal ex- tension followed by decay and exposure of dorsal surface of ligament; В, Anomia, complete supra- dorsal extension with fusion of shell but not of outer layers of ligament; C, Patro, complete fusion of liga- ment as well as shell. РАТВО (ANOMIIDAE) 147 the umbo and of the upper regions of the ex- posed ligament (described and figured т Yonge (1977) and indicated in Fig. 7A). In Anomia (B) the initial supradorsal growth of the mantle margins endures with union of the tissues and so persistence and increase of shell marginal to the umbo. The ligament re- mains enclosed but the anterior and posterior outer layers do not unite; there is always an ap- preciable gap between them. This is not the case in Patro (С) (or in Enigmonia and Heter- anomia) where there must be more complete union of the mantle margins resulting in dis- appearance of the line of union between the supradorsal regions of the shell which per- sists in Anomia (B). Fusion of the outer liga- ment layers produces the complete com- pressed oval shown in Fig. 6. The anterior outer ligament layer is appreciably the longer (FigS 36): Because not specifically considered previ- ously (Yonge, 1977), the periostracum needs mention. In Placuna this becomes separated from the inner and outer layers which form the primary ligament to produce a highly signifi- cant secondary ligament. This extends along the new hinge line evolved in association with the change in that genus from byssal cemen- tation to unattached life on mud. lt is con- cerned with alignment of the valves, the pri- mary ligament solely with provision of the opening thrust. But in Patro, as in all the Anomiidae, although the periostracum loses contact with the ligament when the mantle margins overarch to unite temporarily or per- manently, it retains contact with the margin of the left valve which now extends around the entire periphery, periostracum everywhere forming its outermost layer. There is no pro- duction of, or need for, a secondary ligament. INTERNAL STRUCTURE Comparison between Anomia and Podo- desmus. Before the smaller differences be- tween Patro and Anomia can profitably be discussed, more needs to be said about the differences between Anomia and Pododes- mus. There has been some unfortunate con- fusion here because species of these two genera have been examined from different aspects and never directly compared except incompletely by the writer (Yonge, 1977). Anomia ephippium has been the subject of detailed anatomical studies by Lacaze- Duthiers (1854), Pelseneer (1891, 1911) and Sassi (1905) with further observations on A. glabra by Jackson (1890), A. achaeus by Pelseneer (1911) and A. cytaeum by Tanaka (1955). Structure in this genus has been thoroughly studied but there are no observa- tions in life. All these workers apparently con- sidered species of Pododesmus (Monia) as essentially similar in internal structure. Sepa- ration of the genera was based purely on conchological differences. The state of affairs 1$ just the opposite with Pododesmus. No study of internal anatomy has been made of any species but detailed observations have been made in life of the exposed mantle cavity with figures showing the organs and the course of ciliary currents on ctenidia, palps, mantle and visceral mass in Pododesmus cepio (= Monia machro- chisma) (Kellogg, 1915; Yonge, 1977) and for “Monia” squama by Atkins (1936) who also examined and figured Heteranomia squamula (see below). During the years Atkins worked on these animals at Plymouth she reports never seeing specimens of A. ephippium and clearly assumed that it did not differ signifi- cantly in structure from “Monia.” The present author also failed to see living species of Anomia, relying in his work on the Anomiacea on preserved specimens of A. ephippium from the west of Scotland and of A. simplex from the Atlantic coast of North America. Ma- jor differences were found between the two genera with Pododesmus decidedly the less modified and these were displayed in tabular form (Yonge, 1977, Table 1, p. 495). However the full extent of these differences was not appreciated. These, largely affecting the distribution of the visceral mass, became apparent in the course of the present study and are best ap- preciated by reference to Fig. 8. After careful removal from the shell, preserved specimens of small Pododesmus cepio and Anomia simplex were embedded in 20% gelatin. After some hardening in 10% formalin these were cut horizontally through the centre of the foot and so of the byssal retractor (A, B) with the ventral portion later cut transversely thus passing through the ctenidia and ventral half of the byssal retractor, single or divided (Aj, B,). Differences in the disposal of the visceral mass, especially of the gonad, then became apparent. Differences in the degree of supradorsal extension have already been noted (Fig. 7, A, B) while the presence of the very large hypo- branchial glands separating the ctenidia in 148 УОМСЕ + vmp ___“< HR ha Ts) vmv Ay By FIG. 8. Pododesmus cepio (A) and Anomia simplex (В), horizontal (A, В) and transverse (Ay, By) sec- tions through gelatin-embedded specimens (for de- tails see text). Digestive diverticula indicated by fine, and gonad by coarse, stippling. Pododesmus but absent in Anomia is further demonstrated in these sections (Figs. 8A,, hg; 8B;,). In the latter, also, the ctenidia are seen to be united in the middle line by tissue instead of by the ciliary junctions that connect those of Pododesmus for a short distance anterior to the hypobranchial glands. As al- ready figured (Yonge, 1977), the ctenidia which in Pododesmus, as in bivalves general- ly, pass symmetrically to the right and left palps, in Anomia and Enigmonia are asym- metrically disposed. As shown for the similar Patro in Fig. 9, three demibranchs make func- tional contact with the left (upper) palps and only one with the right palps. This difference also is shown in Fig. 8 (cf. A & B, ctl, ctr). These gelatin sections further reveal that in Pododesmus the visceral mass, including the gonads, although laterally much compressed, remains in the mid-line more or less sym- metrically disposed around the byssal appa- ratus. Held in position between right and left anterior pedal retractors, the mouth continues to be centrally placed, the gut surrounded by the digestive diverticula extending along the posterior surface of the byssal apparatus with the anus as usual projecting at the end of a short rectal extension on the hind surface of the adductor. The very long separate style sac, characteristic of the Anomiacea, extends into the substance of the right mantle lobe. Full effect of lateral compression is shown by the gonads, although remaining approximately the same size, the left gonad extends around the anterior and then ventral surfaces of the byssal apparatus (Fig. 8A, vma; Ay, vmv), in- itially passing between the left ctenidium (ctl) and the foot (f). The right gonad surrounds the digestive diverticula on the posterior side reaching down below the level of the byssal apparatus and meeting, but not uniting with, the left gonad. The two gonads therefore en- circle the central mass of the byssal appa- ratus on the anterior (vma), ventral (vmv) and posterior (vmp) sides. There is the minimum of change from the normal bilateral symmetry of the Bivalvia. Conditions are very different in Anomia. As described in A. ephippium by Lacaze- Duthiers (1854), Pelseneer (1891) and in greatest detail by Sassi (1905) the effects of lateral compression are much greater. The visceral mass is almost entirely concentrated on the posterior side. The gonads are ex- tremely asymmetrical, that on the left greatly reduced and confined to the dorsal side of the byssal apparatus, while that on the right is hypertrophied and extended by way of three connexions from the visceral mass, widely throughout the right mantle lobe. According to Sassi (1905) the former opens dorsally into the left kidney, which encircles the byssal ap- paratus, the latter into the right kidney pos- tero-ventrally. Examination of A. simplex largely confirms these statements. With loss of the right an- terior pedal retractor the mouth moves from the central to a more posterior position (as shown for Patro in Fig. 9A, m). The rectum gets caught up with the right gonad so that the anus becomes attached to the right mantle lobe instead of projecting freely into the man- tle cavity. As shown in Fig. 8B, the visceral mass does not extend anteriorly between ctenidia and foot. It is confined to the posterior and ventral region (vmp, vmv). It is largely PATRO (ANOMIIDAE) 149 occupied by the hypertrophied right gonad which extends beyond it mainly into the right, but also to some extent into the left, mantle lobe. The common origin of the gonadial tis- sue in both lobes, coming from the ventral region of the visceral mass is clearly shown in Fig. 88,. This extension of the right gonad into the left lobe, the left gonad much re- duced, represents a unique attempt to re- establish functional bilateral symmetry in these extremely asymmetrical bivalves. There is no evidence that this occurs in A. ephip- pium but this may be because only young animals have been examined. As noted later there is some evidence that penetration into the left mantle lobe is beginning to take place in one specimen of Patro, all of which were small. Conditions are very different in other anomiaceans, in Enigmonia the left gonad is much the larger (Bourne, 1907) while in Placuna it is lost (Hornell, 1909). Comparisons between Heteranomia and other genera. Ridewood (1903) in his survey of lamellibranch gills showed that “Anomia aculeata, Muller” (= Anomia squamula Linnaeus) has unreflected ctenidial filaments. He therefore associated it with Dimya in a dis- tinctive order Dimyacea. Its obvious anomiid affinities were later recognized by Winckworth (1922) who erected the genus Heteranomia for its accommodation. Recent examination of the filaments in the Dimyidae (Yonge, 1978) shows the ctenidial resemblance to be super- ficial, the inner filaments of the two sides be- ing united in different manners. In the Dimyidae also, unlike Heteranomia, the outer demibranchs make no functional contact with the mantle surface. No work has been done on internal structure in Heteranomia but Atkins (1936) described and figured super- ficial anatomy. lt was earlier shown (Yonge, 1977) that Heteranomia resembles Anomia in posses- sion of a straight crurum without a right an- terior pedal retractor. The rounded byssal notch has more resemblance to that of Pododesmus in possession of a very small subcrural groove. Supradorsal fusion of the tis- sues is complete and with the outer ligament layers united as they are in Patro. There 1$ also a symmetrical secondary union of the mantle lobes posterior to the ligament (i.e. un- like Anomia, Patro and Enigmonia where it is asymmetrical) and, as shown by Atkins, this union extends over the greater part of the ex- halant region unlike the other genera. Further examination in the light of the pres- ent work shows that although the right pedal retractor is lost the mouth remains in much the same position as in Pododesmus; there 15 a large left anterior retractor as figured by Atkins. The gonads are disposed essentially as in Pododesmus without any intrusion of the right gonad into the mantle. A unique fea- ture is the enclosure of the basal half of the foot in the left gonad. This probably reduces pedal activity in cleansing and may be corre- lated with the greater sweep of the unreflect- ed ctenidia. The adductor is smaller than in other anomiids, its scar, as noted by Winck- worth (1922), blending with that of the very much larger, undivided byssal retractor. In brief, in its anatomy Heteranomia has af- finities with Pododesmus (visceral mass, position of mouth), with Anomia (straight crurum, complete supradorsal fusion of tis- sues), with Patro (union of outer ligament layers) together with features peculiar to itself (unreflected ctenidia, enclosure of base of foot within right gonad). This is a very well defined genus. Structure in Patro australis. The general appearance after removal from the shell and viewed from the right (under) and left sides is shown in Fig. 9A, В. Structure 1$ similar to that of Anomia and Enigomia with the three demi- branchs associated with the left labial palps (pl), and the mouth (m) well over on the right side following loss of the right anterior pedal retractor. The anus (a) at least in this small specimen (because conditions might change with growth) is not adherent to the right man- tle lobe. The left anterior pedal retractor (aprl) is larger than in Апотиа spp. The visceral mass is confined to the posterior and ventral sides of the central byssal apparatus with the right gonad extending widely throughout the right mantle lobe. In this the attenuated style sac (cs) describes a complete circle. Only in one, rather larger specimen, was there some indication of extension of this gonad into the left mantle lobe. In the absence of a pericardium, the ventri- cle (v) with the asymmetrical entering auricles (aur, aul) are freely exposed. Viewed from the right side (B) the four muscles producing the scars shown in Fig. 3 are apparent. As throughout the Anomiacea, only a small sec- tion of the adductor is composed of non- striated catch muscle (adc). Adduction 1$ achieved by way of the much larger divided byssal retractor (br, pbr) the quick muscle of the adductor (adq) being responsible for ejec- tion of pseudofaeces by way of the somewhat 150 УОМСЕ I = KA a j > ада L 1cm B FIG. 9. P. australis, specimen removed from shell and viewed A, from right (under) side and B, from left side. Arrows indicate supradorsal extension at each end of mantle isthmus resulting in fusion above. Digestive diverticula and gonad indicated as before. raised posterior region of the mantle cavity. Although only seen after preservation when much contracted, the foot does appear to be smaller than in Anomia sp. and so of possibly less importance in cleansing. The major differences between Patro and Anomia are conchological, the prismatic character of the right valve, the smaller and more open byssal notch (Fig. 5), the complete union of the outer ligament layers (Fig. 7C) and the uneven convexity of both valves (Fig. 6). The area of byssal attachment is some- what smaller but not less calcified. This, to- gether with the obvious ability of the valves to conform to irregular surfaces, supports the opinion of Beu (1967) that species of this genus are adapted for life on more irregular surfaces than those of other anomiid genera. A greater capacity for dealing with sediment is also indicated. Thus, although regarded as a subgenus of Anomia in the Treatise on Inver- tebrate Paleontology, there is good evidence for regarding Patro as a distinct genus, very closely related to Anomia but with its species capable of exploiting the possibilities of life on more irregular substrates and under more turbid conditions. ACKNOWLEDGEMENTS Thanks are due to Dr. Winston Ponder, Curator of Molluscs at the Australian Mu- seum, Sydney, for supply of the specimens here studied. The writer has also to thank his wife for much technical and secretarial help, the Natural Environment Research Council for financial assistance (Grant GR3/1380) and Professor J. M. Mitchison F.R.S. for pro- vision of facilities in the Department of Zool- оду, University of Edinburgh. LITERATURE CITED ATKINS, D., 1936, On the ciliary mechanisms and interrelationships of Lamellibranchs. Part 1. Some new observations on sorting mechanisms in certain lamellibranchs. Quarterly Journal of Microscopical Science, 79: 181-308. ВЕЧ, А. G., 1967, Notes on Australasian Anomiidae (Mollusca, Bivalvia). Transactions of the Royal Society of New Zealand, 9: 225-243. BOURNE, G. С., 1907, On the structure of Aenigma aenigmatica, Chemnitz: a contribution to our knowledge of the Anomiacea. Quarterly Journal of Microscopical Science, 51: 253-295. GRAY, J. Е., 1847, In J. В. JUKES, Narrative of the surveying voyage of H.M.S. Fly, Vol. 2, Appen- dix, p. 362. Boone, London. GRAY, J. E., 1849, On the species of Anomiadae. Proceedings of the Zoological Society of Lon- don, 16: 113-124. HORNELL, J., 1909, Report upon the anatomy of Placuna placenta, with notes upon its distribu- tion and economic uses. Report Marine Zoology Okhamandal, 1: 43-97. РАТНО (ANOMIIDAE) loi IREDALE, T., 1939, Mollusca. Part 1. Scientific Reports, Great Barrier Reef Expedition (1928- 29), British Museum (Natural History), 5: 209- 425. JACKSON, R. T., 1890, Phylogeny of the Pelecy- poda, the Aviculidae and their allies. Boston Society of Natural History, Memoir 4: 277-400. KELLOGG, J. L., 1915, Ciliary mechanisms of lamellibranchs with descriptions of anatomy. Journal of Morphology, 26: 625-701. LACAZE-DUTHIERS, H., 1854, Mémoire sur l'organisation de l'anomie (Anomia ephippium). Annales des Sciences Naturelles, Zoologie. ser. 4, 2: 5-35. PELSENEER, P., 1891. Contribution à l'étude des Lamellibranches. Archives de Biologie, 11: 147- 312: PELSENEER, P., 1911, Lamellibranches de l'expédition du Siboga. Partie anatomique. Siboga-Expeditie, Monogr. 53a: 1-125. RIDEWOOD, W. G., 1903, On the structure of the gill of the Lamellibranchia. Philosophical Trans- actions of the Royal Society of London, ser. B, 195: 147-284. SASSI, M., 1905. Zur Anatomie von Anomia ephippium. Arbeiten aus dem Zoologischen Institute, Wien, 15: 81-96. TANAKA, J., 1955, Anatomical observations on Anomia cytaeum Gray. Bulletin of the Marine Biological Station Asamushi, 7: 109-119. TAYLOR, J. D., KENNEDY, W. J. 8 HALL, A., 1969, The shell structure and mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea Bulletin of the British Museum (Natural History), Zoology, Suppl. 3: 1-125. WINCKWORTH, R., 1922, Note on the British spe- cies of Anomia. Proceedings of the Malacologi- cal Society of London, 22: 23-26. YONGE, C. M., 1977, Form and evolution in the Anomiacea—Pododesmus (Monia), Anomia, Patro, Enigmonia (Anomiidae); Placunanomia, Placuna (Placuniidae fam. nov.). Philosophical Transactions of the Royal Society of London, ser. B, 276: 453-523. YONGE, С. M., 1978, On the Dimyidae with special reference to Dimya corrugata Hedley and Basili- omya goreaui Bayer. Journal of Molluscan Studies, 44: 357-375. Ch 2 й. сы du ‚8 SM TA es 0 COS nn “| » Ze MA À МЭ ASMOY ww ia Al ath но 2081 / 2 Wat О tancia A ES 5 dé AO ieee: oh 2410 PTLARE à men fie si a a a ел AT | . Me erin = y» Gh } er _ ws 1% A L y AA 8 MI do r $ | A x ecrot ri re} ram } 1 ayer ne, if oy eee y 8 : AA! Bi ¡A LA de” у Г ‘ uds ie — «ee D EZ pie +> ALO (On FS wr hive Ig Mel E pe E Е a | P | à u MALACOLOGIA, 1980, 20(1): 153-160 EPISODIC GROWTH IN GASTROPODA Robert M. Linsley! and Mahdokht Javidpour? ABSTRACT A study of the growth lines of shells from the families Cassidae, Cymatiidae, Bursidae and Muricidae indicates that shell growth in these forms 15 not a continuous event, but occurs in abrupt episodes of growth. During these growth spurts the animal may be very vulnerable and possibly the animal's behavior is modified to reduce the vulnerability. The growth of Cassis apparently represents the most rapid deposition of aragonite to be found in the phylum Mollusca. INTRODUCTION The helical form has been adopted by many different groups of organisms, primarily be- cause it allows growth to be continuous with- out modification of overall form (Raup, 1966; Thompson, 1917). Most frequently this form will be functional without modification through- out the adult life of the organism. However, a few gastropod genera have evolved shell forms which are not continuously repeating, but are functional only at periodic stages dur- ing their growth. In some instances (i.e. Biplex Perry) the snail has even abandoned the helical form and evolved a compressed variation of а helix. In other instances (Cymatium Róding and Distorsio Róding) the helical form 15 retained, but the axis of coiling changes for each growth episode, resulting in a shell with very distorted volutions. Still others (Phalium Link, Cassis Scopoli, Bursa Róding and many Muricidae) retain a regular helical coil, but interrupt the helix with thick- ened, reflexed, varix-forming apertural modi- fications. The functional forms repeat them- selves at various intervals in these genera (Wrigley, 1934). In Murex the functional form frequently repeats about every one-sixth or one-third of a volution, although this is an ex- tremely variable feature in this family. In Biplex repetition 1$ every one-half volution, in most species of Cassis, Cymatium and Dis- torsio approximately every two-thirds volution (Laxton, 1970). Shell formation must thus oc- cur in these amounts as the intermediate forms are non-functional. In a limited sense, shell accretion in these mollusks somewhat resembles growth of the arthropods. Firstly, the forms are discontinu- ous from one stage to another and secondly, while the animal is secreting the earliest shell layers, it is very vulnerable to predation. It is anticipated that many of these animals seek protection by burial during these unprotected stages, only reemerging when the shell has been sufficiently thickened to be of significant protective value. Because of this vulnerability, many have adopted aperture-edge modifica- tions such as the shell reflected back on itself (Fig. 1) or greatly flared to strengthen that weakest portion of the shell (Vermeij, 1976, 1977). Thirdly, it might prove interesting to in- vestigate the physiology of these mollusks for we anticipate that there must be storage of calcium salts in the mantle in preparation for the sudden growth needs found in these forms. lt is presumed that aragonite deposi- tion is very rapid during the early periods of new growth. DESCRIPTION OF EXAMINED THIN SECTIONS We have examined thin sections of most genera that have pronounced varices. These include the genera Cassis Scopoli, Cymatium Rôding, Distorsio Róding, Biplex Perry and various muricids. Thin sections of all speci- mens examined show at least superficially the expected shell layers of periostracum, ostra- cum and hypostracum. However the growth lines do not cross the shell layers at an oblique angle as they do in most mollusc shells. Rather the growth lines all parallel the outermost surface of the shell as well as the 1Department of Geology, Colgate University, Hamilton, New York 13346, U.S.A. University of Teacher Education, Tehran, Iran. (153) 154 LINSLEY AND JAVIDPOUR boundaries between various shell layers. From this evidence we conclude that shell- secretion in varix-bearing gastropods is not a constant process of accretion with the mantle acting as a moving conveyor belt and altering its secretion from periostracum at the mantle margin to the various calcareous deposits as the cells move away from the margin. Instead we suggest that shell secretion is a very abrupt episode of growth. The nature of the growth line suggests that the animal extends quite far from the old aperture whereupon the mantle secretes a broad layer of periostracum and then follows it with the various calcareous layers. Thus the change from ostracum to hypostracum is not a matter of topology for a single cell but an abrupt secretory shift for the entire mantle. Some of the genera studied offer further in- sight into the nature of this episodic growth so a genus by genus discussion seems appro- priate. DESCRIPTION OF EXAMINED SPECIES Genus Cassis Scopoli (Figs. 1, 2) We have studied sections of the shells of three species, Cassis tuberosa (Linné), C. flammea (Linné) and C. madagascariensis (Lamarck). These shells were collected in the Bahamas in January, 1975. The helmets are large shells which have modified their basic helical shell form into a subtriangular shape by the formation of a thickened outer lip and a thick parietal inductura to form a flattened shield-like area around the aperture. This provides the shell with considerable stability on the substrate even when the animal is re- tracted into its shell. This sub-triangular shape is further emphasized in many species by the formation of a large “dorsal” node or row of nodes directly opposite the aperture. The aperture is periodically abandoned and re- mains in the older portions of the shell as a varix. Typically the varices are about two- thirds of a volution apart. Thus, a single growth episode must span two-thirds of a vo- lution very quickly in order to remain func- tional. Since some helmets are very large (in ex- cess of thirty centimeters long) this initial de- posit of aragonite may represent the greatest and most rapid secretion of calcareous mate- rial in the phylum Mollusca. Cassis has developed one feature relating to growth which we believe may be unique to the genus. Not only does it secrete two-thirds of a volution of the outer lip in one episode, but it also secretes one-third volution of a new inner whorl at the same time. This inner whorl is separated from the preceding whorl through much of its growth, thus forming a disjunct inner surface which completely envelopes the old varix and old siphon in an interior space. Since specimens of Cassis frequently pos- sess a lush growth of epibionts this secretion of a new inner whorl covers over these en- crustations and shields the body of the gas- tropod from them during the abrupt growth spurt. We believe that Cassis buries itself in sand during the formation of the new shell, but can offer only indirect evidence to support this be- lief. When fresh shells are cut open this inner space between the new inner surface and old shell is filled with sediment which suggests that it may have been enclosed there at the time of formation of the new shell layer. Al- though this supports our contention that this growth takes place while Cassis is buried, this space 1$ open to the exterior through an open- ing (a “pseudoumbilicus”) behind the si- phonal canal and the most recent varix. It is, therefore, possible that sand entered this cavity after its formation, but the area is so tightly packed with similar looking sediment that we suspect the sand to be a primary fea- ture. One specimen was given to us by Porter Kier (National Museum of Natural History, Washington, D.C.) who found it buried in sedi- ment in the Caribbean. The last two thirds of a volution are literally paper-thin (0.5 mm) and was probably just thickening its shell prior to re-emergence. Rhoads & Morse (1971) have pointed out how burial in anoxic sediments can inhibit calcium deposition in mollusks. Obviously this defensive behavior would only be possible for animals whose shell form would allow burial and who live in areas where the sediments are well oxygenated. Cassis frequently buries into the sediment as it searches for echinoids, and lives in areas of coarse oolitic sands which would presumably be well oxygenated. Therefore, we see burial during the growth phase as a viable protective behavior for Cassis, but certainly not for all other varix- forming gastropods. It is presumed that Cassis emerges from the sand while the shell is still moderately thin and thus the formation of a thickened outer lip GASTROPOD EPISODIC GROWTH 155 FIG. 1. Transverse thin section and schematic interpretation through the outer lip of Cassis tuberosa (Linné), showing reflection of the shell layers to form a double thickness of shell. BEOUNEC DOS wn — FIG. 2. Transverse thin section and schematic interpretation through the parietal shield of Cassis tuberosa (Linné). In the reconstruction shell layers 1-3 are of the varix (the outer lip of the preceding resting phase). Layers 4—6 are shell layers of the final volution with the layers above the varix belonging to the final whorl, while those below the volution are part of the parietal shield. 156 LINSLEY AND JAVIDPOUR would help protect the animal from predation, particularly by crabs (Vermeij, 1976, 1977). At the mantle edge where the new outer lip will be formed, the mantle reflects back on itself to secrete a double thickness of shell material. The outer lip maintains this growth style throughout the existence of the lip, con- stantly accreting a double shell layer at the outer lip at the same time that the parietal shield is accreting (Figs. 1, 2). An examination of the outside of the shell reveals features that resemble growth lines, but close inspection shows they they do not possess the imbri- cated structure normally associated with growth lines, but instead are rounded, conflu- ent structures that we call “pseudo-growth lines.” Genus Cymatium Róding (Fig. 3) Cymatium echo (Kuroda 8 Habe) has only a single varix preceding the thickened outer lip (Laxton, 1970). In the area between the varix and the outer lip the periostracum 1$ ex- tended into lines of hair-like extensions six times. The lines of periostracal hairs are close together near the varix and progressively more widely spaced as the outer lip is ap- proached. From this we infer that the periostra- cal layer is formed by six abrupt extensions of the mantle with the periostracal hairs delineat- ing each extension. We would expect a thin layer of calcareous material to be deposited under the periostracum to rigidify it soon after the deposition under the periostracum. This process would then be repeated after short intervals, each periostracal secretion being broader than the preceding one, until the final extension is more than a quarter of a volution. Harold Lewis (personal communication, 1977) has reported seeing individuals of Cymatium in this final stage and reports that the perios- tracal layer was flexible enough to be pulled across the aperture to partially close it off. Nevertheless, examination of thin sections of Cymatium echo demonstrates that these short calcareous layers of these subepisodes are tissue thin, for they are not visible in cross-section. Again, the major impression in the cross-sectional view (Fig. 3) are thin sheet-like laminae that parallel the outer sur- face of the shell, showing that, for the most part at least, shell material is being deposited in continuous sheets over two-thirds of a volu- tion with thin extensions past the preceding varix. Varix formation 15 quite different from that found in Cassis. The outer lip is formed by ап outward bulge of the outermost shell layer and 1$ thickened by subsequent addition of thicker layers of aragonite (Fig. 3). It is ap- FIG. 3. Transverse thin section of Cymatium lotorium Linné and schematic presentation of the varix. The schematic presentation traces six well-defined growth lines within the varix and subsequent whorl. GASTROPOD EPISODIC GROWTH 157 parent that aragonite is deposited in the area of the thickened ощег lip long after it has ceased elsewhere. Many species of Cymatium have morpho- logic features of the shell which are non-func- tional except at periodic intervals when the ощег lip is restored. In all species examined, the varices (and thus the major growth epi- sode) are roughly every 240° + 15°. The major non-continuous feature in many species, such as Cymatium femorale (Linné), С. lotorium (Linné) and C. hepaticum (Róding) consists of an inclination and rotation of the axis of coiling relative to the original axis of coiling. This strange growth mode allows the plane of the aperture to converge upon being parallel to the original axis of coiling. This lowers the center of gravity of the shell (Linsley, 1977) and assures that the modest spire does not extend very high above the substrate. These animals live on unconsolidated sediments. where the modest spire would provide a good lever for dislodging the specimens if it pro- truded upwards too Пай above the substrate. Genus Distorsio Réding (Fig. 4) Distorsio shows many features that are similar to Cymatium. The growth episodes constitute two-thirds of a volution and are accompanied by inclination and rotation of the shell axis for each depositional event. Inter- mediate growth positions would obviously be non-functional, with the spire held much too high off the substrate, presenting a high cen- ter of gravity and easy dislodgement of the animal from the soft substrate. The presence of periostracal hairs again suggests the oc- currence of subepisodes (about twenty are suggested in D. reticulata). A careful exami- nation of the shell beneath the periostracum does not show any reflections of the subepi- sodes, and examination of the thin sections (Fig. 4) again suggests that shell deposition occurred in sheets parallel to the outer sur- face of the shell. Varix formation is again distinctive in Dis- torsio. The initial formation of the outer lip oc- FIG. 4. Transverse thin section and schematic interpretation of Distorsio anus (Linné). Shell layer 1 in the schematic diagram represents the former growth phase and varix, while layers 2 and 3 represent the final two-thirds of a volution showing the method of formation of the outer lip and the parietal shield deposited over the varix. 158 LINSLEY AND JAVIDPOUR curs by a figure “S” flaring of the outermost (ostracal) shell layers (Fig. 4). Subsequent shell deposition, primarily of the hypostracal shell layers, thickens the outer lip dispropor- tionately, causing an inward bulge of shell material at the position of the future varix. Genus Biplex Perry (Fig. 5) Our understanding of this genus is based on examination of the shell and thin sections of a single species, Biplex perca (Perry). This is an unusual species in that it has completely abandoned the logarithmic growth form in favor of a very flattened approximation thereof (Fig. 5). This flattening is emphasized by the construction of flanges near the position of the outer lip which results in varices every 180”. Nowhere in the class Gastropoda is there а shell that is so patently non-functional except at the 180” positions. Examination of the thin sections again suggests that growth is indeed episodic, and that 180° of shell is secreted in sheets and then progressively thickened in- wards. Varix formation is accomplished by a blade- like Outpouching of the outer shell layer (Fig. 5) which results in sub-parallel layers of shell enclosing a hollow core. It is presumed that the mantle initially occupies this core area. As new shell layers are deposited in this space the mantle retreats more rapidly than shell deposition occurs with the consequence that hollow spaces are left between subsequent infilling shell layers. The resultant form of this shell is well adapted to lying flat on a soft substrate with the flanges disposed to protect the protruding soft parts and at the same time resist having the shell sink into the ooze. Family Muricidae da Costa (Figs. 6, 7) The majority of the members of the family Muricidae have episodic growth. Those that do are typically characterized by varices, usually placed every third of a volution or every sixth of a volution. Abrupt growth is at- tested to by the fact that of three hundred specimens of Chicoreus florifer (Reeve) in the Colgate collection, every one has а fully formed outer lip, some very thin, some fully filled in with subsequent inner deposits. None of the three hundred specimens was collected with a lip formed in between varices. The growth rates of a number of muricids have been studied under laboratory condi- tions and certainly the intermittentness of the growth process has been noted. Spight & Lyons (1974) plotted the growth of juveniles of Ceratostoma foliatum (Gmelin) and noted growth ceased in between varix formation in more mature individuals. MacKenzie (1961) reported that immature individuals of Eupleura caudata (Say) deposited one-half volution in three weeks and then spent four weeks reinforcing it. Inaba (1967) observed Chicoreus asianus for sixty days and ob- served that the snail grew less than half the time it was under observation and that it fasted while making all modifications to its shell. Pterynotus trialatus (Sowerby) and Forreria belcheri (Hinds) were also observed to undergo episodic growth by MacGinitie & MacGinitie (1949). And lastly, Abbott (1954) notes that some species of Murex require two days to grow the shell between varices. Apparently, muricids which have been ob- served under laboratory conditions do not at- tempt to hide during these growth periods. In- deed, we would not expect them to under FIG. 5. Transverse thin section and schematic representation of Biplex perca Perry. Note the flattened helical form and the mode of varix formation. GASTROPOD EPISODIC GROWTH 159 FIG. 7. Transverse thin section of Murex troscheli (Lischke) and schematic diagram of varix. natural conditions either. For one thing, many muricids, such as those with spires or exten- sive varices have shell forms that would be poorly suited to a burrowing life mode. Sec- ondly, since many muricids live in muds which may be anoxic, burrowing might well inhibit calcite deposition (Rhoads & Morse, 1971). Thin sections were made of Chicoreus flor- fer (Reeve), Murex troscheli (Lischke), Chicoreus axicornis (Lamarck), Phyllonotus ротит (Gmelin), Murex tribulus (Linne), Chicoreus ramosus (Linne), Marchia pelluci- dus (Reeve), and Pteropurpura macroptera (Deshayes). All are similar in that they show aragonite deposition in layers paralleling the outer surface from which it is inferred that original growth is very rapid, extending the whorl 1/3 or 1/6 volution. This phase is then followed by continued thickening of existing shell until the next growth spurt is com- menced. With a few exceptions, the formation of the 160 LINSLEY AND JAVIDPOUR outer lip of the muricids follows the general plan found in Cymatium. п most instances (i.e. Phyllonotus) the outermost shell layer forms an initial outward bulge which is then thickened by the addition of layers inside of the initial layer. A variation on this general theme is found in Pteropurpura macroptera where the initial shell layer merely flares out- ward without the inward bend (Spight 4 Lyons, 1974). Subsequent deposition merely parallels this initial form to cause the resulting flange. Pterynotus pellucidus combines these two techniques to form its outer lip. The initial deposits resemble those of Pteropurpura macroptera in having an outward flare. After this portion 15 thickened slightly a second phase is initiated where there is now an in- ward bend more reminiscent of that found in Murex pomum. PRESUMED CONSEQUENCES OF EPISODIC GROWTH We suggest that there may well be physio- logical and behavioral adaptations in organ- isms that utilize episodic growth. At the very least we expect feeding to cease while modifi- cations are made to the shell, as has been reported for Chicoreus asianus (Inaba, 1967). Where possible, we expect the animals will seek protection during this phase of vulnera- bility either by pulling the periostracum across the aperture (т Cymatium) or by burial (in Cassis). We also suggest that there may well be a buildup of calcium salts in the mantle prior to the growth episode in preparation for the need for quick deposition of large sheets of aragonite. Lastly, there may be some adapta- tion of the columellar retractor inside relating to the growth episode. lt would be of interest to know if muscle migration is abrupt and dis- continuous so as to be in the most advantage- ous placement during the long periods of no growth or whether it migrates steadily through- out the quiescent phase. SUMMARY AND CONCLUSIONS Within the class Gastropoda, a number of species have evolved shell forms character- ized by varices. These varices not only repre- sent resting phases” in shell growth, but pro- vide the shell with a geometry that makes in- tervarical shell forms non-functional. Exami- nation of thin sections of these shells indi- cates that growth from one varix to the next 15 indeed very rapid, for growth lines parallel the outer shell surface as though sheets of aragonite were deposited, each underlying the preceding sheet. It is further inferred that each growth episode probably consists of sub-episodes so that, though growth from one varix to the next is very rapid, it does not occur in a single event. It is also suggested that this growth mechanism must necessitate physio- logical and behavioral adaptations by these organisms. LITERATURE CITED ABBOTT, R. T., 1954, American Seashells. New York: Van Nostrand, xiv and 541 p. INABA, A., 1967, Growth of Chicoreus asianus. Venus, 26: 5-7. LAXTON, J. H., 1970, The relationship between the number of varices and total shell length in some New Zealand Cymatiidae (Gastropoda: Proso- branchia) and its ecological significance. Veliger, 13: 127-134. LINSLEY, R. M., 1977, Some “laws” of gastropod shell form. Paleobiology, 3: 196-206. MACGINITIE, С. E. 4 MACGINITIE, N., 1949, Natural History of Marine Animals. New York: McGraw Hill, 473 p. MACKENZIE, С. L., Jr., 1961, Growth and repro- duction of the oyster drill Eupleura caudata in the York River, Virginia. Ecology, 42: 317-338. RAUP, D. M., 1966, Geometric analysis of shell coiling: general problems. Journal of Paleontol- ogy, 40: 1178-1190. RHOADS, D. С. 8 MORSE, J. W., 1971, Evolution- ary and ecologic significance of oxygen-deficient marine basins. Lethaia, 4: 413—428. SPIGHT, Т. М. & LYONS, H., 1974, Development and functions of the shell sculpture of the marine snail Ceratostoma foliatum. Marine Biology, 44: 77-83. THOMPSON, D'A. W., 1917, On Growth and Form. Cambridge, xv and 793 p. VERMEIJ, С. J., 1976, Interoceanic differences in vulnerability of shelled prey to crab predation. Nature, 260: 135-136. VERMEIJ, С. J., 1977, Patterns in crab claw size: the geography of crushing. Systematic Zoology, 26: 138-151. WRIGLEY, A., 1934, English Eocene and Oligocene Cassididae, with notes on the nomen- clature and morphology of the family. Proceed- ings of the Malacological Society of London, 21: 108-130, pl. 15-17. MALACOLOGIA, 1980, 20(1): 161-194 THE GIANT WHITE CLAM FROM THE GALAPAGOS RIFT, CALYPTOGENA MAGNIFICA SPECIES NOVUM Kenneth J. Boss and Ruth D. Turner Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts 02138, U.S.A. ABSTRACT During DSRV/ALVIN cruises to the Galapagos Rift and the East Pacific Rise in 1977 and 1979, dense populations of large white clams were found associated with the thermal vents. On the basis of shell characters, M. Keen identified them as a species of Calyptogena, family Vesicomyidae. These clams exceed 260 mm in length and are the largest known living members of this family, though some fossil vesicomyids are comparable in size. Morphologically they closely resemble C. pacifica and C. kilmeri, the only species of Calyptogena which have been studied anatomically. Specimens ranging from 34.5 to 263 mm in length obtained during the 1977 and 1979 cruises are here described as a new species, C. magnifica, most closely related to C. elongata from off the coast of southern California, a species described by Dall on the basis of shells only. In Appendix 1, C. Berg and R. D. Turner describe living specimens, noting the pink-purple iridescences of the mantle, the yellow-brown, wrinkled periostracum, the short siphons which do not extend beyond the valves, and the large iridescent-pink protrusible foot. The red blood pigment, a haemoglobin, gives the visceral mass a red appearance. The gills in large specimens are mottled red-brown with purple lines on the ventral margin while in small specimens they are a uniformly pinkish cream. In Appendix 2, K. Boss provides an annotated checklist of ten fossil and seven living species of Calyptogena. One species is from the eastern Atlantic, off Africa; two are from the western Caribbean and 14 from the eastern and northern Pacific and Peru to Japan. INTRODUCTION Lonsdale (1977) reported unusually dense macrobenthic communities, including large white clams, associated with thermal anomalies along the deep-sea spreading centers of the Galapagos Rift. Samples from these peculiar ecosystems were obtained during DSRV/ALVIN Cruise 90 (Corliss et al. 1979) and during a biological investigation (ALVIN Cruise 102, legs 8 and 9) in January- February 1979 (Grassle et al., 1979) and Cruise 103, leg 5, in November-December 1979. The unusual habitat and organisms have received considerable attention in the public media, including an illustrated popular account by Corliss & Ballard (1977) and Bal- lard 8 Grassle (1979). The big white clams, which provided the nickname “Clambake” for one of the hot vents, caused great excitement among oceanographers and marine biologists. They were referred to the genus Calyptogena by Keen (1977a, b). The large size (all speci- mens from the 1977 cruise were about 200 mm in length) and unusual habitat of these clams led us to believe that they repre- sented a new species closely related to C. modioliforma (Boss), a species from the Caribbean based on a unique specimen, and C. elongata Dall from off southern California known only from three specimens all less than 50 mm in length. When DSRV/ALVIN revisited the Galapa- gos Rift in 1979, several smaller specimens were collected which produced a growth series ranging from 34.5 mm in length (which is smaller than the type-specimen of C. elongata) to 240 mm in length. Though the type-locality of C. elongata is some hundreds of miles north of the northernmost specimens collected by ALVIN, these widely separated populations show relatively minor morpho- logical differences in the shells except the size of the valves and the ligament.1. The dis- covery of new populations of Calyptogena as TTurekian, Cochran 8 Nozaki (1979) have calculated the age of the shells of this species to be between 6.5 and 830 years old. (161) 162 BOSS AND TURNER exploration of the East Pacific Rise prog- ressed northward (Fig. 13), and the variability shown by many species in this genus has led us to consider the ‘hot vent’ clams as a new species most closely related to C. elongata. This 15 a tentative assignment for, though we are secure in our assignment on the basis of the shells, the anatomy of the soft parts of C. elongata and other species of Calyptogena 15 essential before we can positively determine the relationship of these species. SYSTEMATICS AND DESCRIPTIONS Family Vesicomyidae Dall 8 Simpson 1901 Description. Shell of adults from less than 10 to over 200 mm in length, ovate to elon- gate in outline, inequilateral and equivalve, usually without gape; shell substance aragonitic with homogeneous inner and outer layers without tubulations or cross-lamella- tions; periostracum present, usually well de- veloped; umbos prosogyrous, sometimes partially enrolled; lunule and escutcheon vari- able, present or absent. Ligament external, opisthodetic, and parivincular. Dentition with variably formed, more or less subumbonal cardinal teeth diverging or subundulate be- neath the beaks. Adductor muscle scars sub- equal and with pedal retractor scars at their dorsal inner margin. Pallial line sometimes broadened, usually entire and sometimes with variously formed posterior ‘pallial sinus.’ Pedal gape extending from anterior adductor muscle to fusion of mantle folds to form the short posterior siphons. Foot strong, with byssal gland. Remarks. The family Vesicomyacidae was instituted by Dall & Simpson (1901) and later used by Dall (1908). Keen (1969) corrected the spelling to Vesicomyidae but dated it from Dall (1908). Thiele (1934) apparently, and possibly correctly, did not recognize the family Vesicomyidae. He placed Calyptogena Dall, 1891 in the Carditidae and Vesicomya in Kelliellidae Fischer, 1887 (nomen correctum Dall, 1900 pro Kellyellidae Fischer, 1887, see Keen, 1969). However, according to Boss (1969b), Callocardia atlantica Smith, 1885, the type-species of Vesicomya Dall, 1886 may prove to belong in the genus Kelliella and, if so, the famiy name Kelliellidae will take precedence. Alteration of familial and super- familial classification is out of place in the present context, but evidence is accumulating to show that the more primitive heterodont veneroids have been excessively divided at the higher phyletic ranks (Yonge, 1969). Fur- ther, the difficulty in assigning many taxa currently included in the Vesicomyidae is re- flected by their earlier placement in such families as the Arcticidae (= Cyprinidae), Carditidae, Kelliellidae and Veneridae. The systematics of the family Vesicomyidae is beset with difficulties because there is at present no satisfactory diagnosis which would exclude the constituent taxa from all other heterodonts based on shared derived char- acter states. Among the reasons for this dif- ficulty are the rarity of samples, their consid- erable variability, the lack of anatomical data, and the ill-defined boundaries of the numer- ous families of heterodont bivalves with which vesicomyids have been associated. Most species listed by Lamy (1920), Odhner (1960) and Boss (1969a) are repre- sented by few specimens or by mere frag- ments. Large ontogenetic series are usually lacking; many species are known from single localities, usually at considerable depths. Vesicomyids, particularly Calyptogena, are highly variable (Kanno, 1971). The unreliabil- ity of certain conchological features is shown by two obviously conspecific specimens of an undescribed vesicomyid taken in the same dredge haul in the North Atlantic. They are the same size, texture, color, sculpture and shape, but one possesses, and the other lacks, a fully formed, impressed lunule, a fea- ture sometimes used to distinguish genera. Genus Calyptogena Dall, 1891 Calyptogena Dall, 1891: 189 (type-species, by monotypy, Calyptogena pacifica Пай, 1891). ?Pleurophosis van Winkle, 1919: 23 (type- species, Бу monotypy, Pleurophopsis unioides van Winkle, 1919). ?Pleurophophis van Winkle. Cossmann, 1920: 29, error for Pleurophopsis, van Winkle, 1919. ?Pleurophoropsis Cossmann, 1920: 29, nomen vanum for Pleurophopsis van Winkle, 1919. Phreagena Woodring, 1938: 50 (type- species, by original designation, Phreagena lasia Woodring, 1938). Ectenagena Woodring, 1938: 51 (type- species, by original designation, Calyptogena elongata Dall, 1916). Akebiconcha Kuroda, 1943: 14 (уре- species, by monotypy, Akebiconcha kawa- murai Kuroda, 1943). ?Hubertschenckia Takeda, 1953: 85 (type- species, by original designation, Tapes ezoensis Yokoyama, 1890). CALYPTOGENA MAGNIFICA FROM GALAPAGOS НЕТ 163 Description. Shell white, usually chalky, heavy, and more or less elongate, length be- ing about 1.5 to 2 or more times the height. Sculpture usually of irregular growth lines, most apparent peripherally. Beaks anterior to middle, generally in anterior third or quarter. Ligament external, opisthodetic, rather strong and resting on variously developed nymphal callosities. Generally without lunule; escutch- eon weak or absent to relatively strong. Pallial line broad and generally with shallow posteri- or indentation or ‘pallial sinus.’ Valves often constricted mesially, forming gena or cheeks. Periostracum usually dehiscent, dull, various- ly developed. Dentition irregularly developed, consisting principally of subumbonal cardinal elements without well developed distal lateral dentition. In left valve, subumbonal cardinal tooth more or less curved, D-shaped, with central socket between the two dental ele- ments, which may be referred to as the ante- rior ramus (or anterior dorsal cardinal tooth) and posterior ventral ramus (or posterior ven- tral cardinal tooth). Posterior subumbonal cardinal tooth or irregular keel or ridge radiat- ing posteriorly from beneath umbo, just ven- tral to the beginning of the nymphal callosity, ligament and posterior dorsal margin. Right valve, with subumbonal cardinal teeth, con- sisting of variously developed D-shaped ele- ments, which may be separated distally into two dental elements (Fig. 10, Cb; Boss, 1968: figs. 16-17) and of a more or less curved ven- tral tooth, with an irregular D-shaped socket which separates these dental elements. Addi- tionally, posteriorly radiating nymphal callosity subtending the ligament on posterior dorsal margin may appear as subobsolete, ridge-like dental element. Animal with pedal gape extending from anterior adductor muscle to fusion of mantle folds beneath siphons. Foot strong, variously pointed and with poorly developed byssal gland and groove; apparently non-byssate in adult. Mantle edge thickened, broad anterior- ly, and forming an embayment or “раша! sinus” posteriorly in vicinity of short, separate, incurrent and excurrent siphons. Ctenidia homorhabodic without distinguishable food groove, at least in C. magnifica; inner and outer demibranchs present; extensive dorsal extension of ascending lamella of outer demi- branch present. Labial palps reduced, obso- lete and lip-like. Stomach with style-sac not differentiated from midgut. Remarks. The genus Calyptogena and C. pacifica Dall were instituted by Dall (1891) in the family Carditidae. He maintained this placement (1895a: 541; 1903a: 70; 1903b: 1410) and was followed by other authors, notably Lamy (1922: 349), Grant & Gale (1931: 278) and Thiele (1935: 838). Calypto- gena, however, is not confamilial with the Carditidae because, though the shells are aragonitic, the outer and inner layers in Calyptogena are homogeneous while in the Carditidae they are cross-lamellar and have a dense system of tubulations (Oberling & Boss, 1970; Taylor, Kennedy & Hall, 1973). Additionally, carditids usually have strong radial sculpture and crenulate valve margins, are byssate in the adult stage, lack a formed incurrent siphon, have the ventricle beneath the rectum, and have a tendency to brood the young (Boss, 1968; Yonge, 1969). Dall's placement of Calyptogena pacifica in the Carditidae may have been based on the superficial resemblance of its hinge with that of Cardita affinis Sowerby from the Pacific coast of Mexico (Boss, 1968). Woodring (1938) indicated the vesicomyid affinities of Calyptogena based on hinge structures and established the genera Ecte- nagena (type-species, C. elongata Dall, 1916) and Phreagena (type-species, Р. lasia Woodring, 1938). He subsequently indicated that Phreagena was a synonym of Calypto- gena (Winterer & Durham, 1962; Boss, 1968; Woodring, personal communication). The type-species of Ectenagena is closely related to C. pacifica. Okutani (1966b) suggested that Calyptogena and Akebiconcha (type- species, A. kawamurai Kuroda), a Japanese genus thought to belong to the Cyprinidae (= Arcticidae, see Boss 1969b), were confamili- al. The homology of the hinge elements of the arcticids and vesicomyids is difficult to deter- mine because well-differentiated lateral teeth are not found in Calyptogena and there is great infraspecific variation of all dental ele- ments (Fig. 10; Boss, 1968: figs. 16-17 and 19-20). Akebiconcha and Calyptogena can- not be separated on the basis of valve shape and dentition (Boss, 1968). There are also many anatomical differ- ences between Са/урюдепа and the single living species of Arctica (Saleuddin, 1964; Zatsepin & Filatova, 1961). Arctica islandica has a laterally compressed, hatchet-shaped foot, the gut is long and elaborately coiled, and the labial palps are distinct and large. In con- trast, Calyptogena and other known vesi- comyids have greatly reduced labial palps, a short gut and a pointed conical foot. The tenuous inclusion of Pleurophopsis 164 BOSS AND TURNER and its synonyms in the synonymy of Calyp- togena is based on the exceptionally large specimens of an unnamed species of Pleuro- phopsis from supposed Oligocene deposits in Colombia (USNM 11253, “No. 33, from a point % mile north of junction of Arroyo- Piedras Palmar and Palmar-Molinero road on the Palmar-Molinero road. Plane table station 245 of Link and White”). The coordinates of sta. 245 are 10°40’N; 75°03'W, a point sever- al miles south of Barranquilla, Colombia. These specimens exceed 200 mm in length and, externally, appear very close to Calyp- togena but the hinge structure is unknown. Species associated with Pleurophopsis in Oligocene deposits of Peru include Solemya and Vesicomya, further evidence of a rela- tionship between Calyptogena and Pleuro- phopsis (Olsson, 1931). The type-species of Pleurophopsis from Trinidad, originally thought to be Oligocene but probably Plio- сепе (Woodring, personal communication), is similar to Calyptogena in outline and has two cardinal teeth in each valve (Woodring, 1938). Hubertschenckia Takeda, based on an Oligocene fossil, is tentatively considered a synonym of Calyptogena because Keen (1969) and Habe (1977) included it in the Vesicomyidae and Kanno (1971) infers that it is closely related to Calyptogena. Although Habe (1977) placed the genus Adulomya Kuroda, 1931 in the Vesicomyidae and though some species that were once re- ferred to it (e.g. Adulomya chitanii Kanehara, 1937) are species of Calyptogena (see Kanno, 1971: 80-82, text-figs. 10-12, and pl. 7, figs. 5, 6a-b, and pl. 17, fig. 12), we have not considered it a synonym of Calyptogena because the type-species of Adulomya, A. uchimuraensis Kuroda, is supposedly edentu- lous and is not a vesicomyid (Kanno 4 Ogawa, 1964: 285). According to Cox (1969), Adulomya belongs in the Solemyidae. Most species of Calyptogena, both living and fossil are found in the Pacific Ocean, oc- curring from Japan to the Gulf of Alaska and south to off South America. In the Atlantic they are known from the Caribbean and off the coast of Africa. They occur from the Oligocene to the Recent and in depths rang- ing from about 100 m to over 2600 m. (Ap- pendix 2 lists the species currently referable to the genus Calyptogena.) Subgenus Ectenagena Woodring, 1938 Ectenagena Woodring, 1938: 51 (type- species, by original designation, Calyptogena elongata Dall, 1916). Description. Shell similar to Calyptogena, $.5., sometimes exceeding 200 mm in length. Escutcheon generally not demarcated nor well developed. Dental configuration as in Calyptogena except teeth more or less blunt and comparatively shorter; hinge plate com- paratively less extensive and thinner. Right valve lacking anterior dorsal cardinal element, probably resulting from reduction of dorsal ramus of D-shaped subumbonal cardinal tooth (Fig. 10, Eb and Fb). Ctenidia with strong interlamellar septa. Remarks. Boss (1968) used Ectenagena as a genus but, considering the intraspecific variation of Calyptogena, s.s. and the paucity of distinguishing traits, we follow Keen (1969) in considering it a subgenus of Calyptogena. As we are placing the white clam from the Galapagos Rift in the subgenus Ectenagena we include here a description of С. elongata, the type-species of the subgenus. Calyptogena (Ectenagena) elongata Dall, 1916 Figs. 10E, 11, 12А-С Calyptogena elongata Dall, 1916: 408 (Albatross Sta. 4432, off Point Loma, Cali- fornia, in 275 fathoms [8 mi. S. of Brockway Point, Santa Rosa Id., Channel Ids.];? holo- type, USNM 110774); 1921: 32, pl. 3, fig. 3; Oldroyd, 1924: 116, pl. 22, fig. 6; Grant & Gale, 1931: 279; Oinomikado & Kanehara, 1938: 73; Otatume, 1942: 198; Okutani, 1957: 28; Okutani, 1962: 23; Bernard, 1974: 18, non C. elongata Ozaki, 1958. Ectenagena elongata (Dall). Woodring, 1938: 51, fig. 2c; Boss, 1968: 739, 744, figs. 25, 28; Bernard, 1974: 19. Calyptogena (Ectenagena) elongata (Dall). Keen, 1969: N664,fig. E138, 8a, b. Range. Known only from the type-locality— Albatross Sta. 4432, 8 miles S. of Brockway “The three specimens are catalogued separately. In the original description Dall gave the locality for the holotype (USNM 110774) as Albatross Station 4432, off Point Loma, California in 275 fathoms and the specimen is so labeled in the USNM. A paratype (USNM 205888) 15 labeled as from the same station but the depth is given as 183 fathoms. The third specimen, a paratype (USNM 209309), also from Albatross Station 4432 is labeled as off Santa Rosa, Santa Barbara Island, California in 270-280 fathoms. Recourse to the Dredging and Hydrographic Records for 1904 and 1905 (published 1906) give the station data as indicated under Range. CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 165 Point, Santa Rosa Island, Channel Islands in 272-270 fathoms [500 т]. Specimens examined. Holotype and two paratypes (only known specimens). Description. Shell white, elongate, elliptical, equivalved but inequilateral, periostracum yellow-brown, largest known specimen 44 mm in length, 17.5 mm in height and 10 mm in width. Umbos low, small, pointed and located on anterior Ya of valves. Anterior margin of valves rounded, anterior dorsal margin 1% total dorsal margin and sloping from the umbos at an angle of about 20°. Ventral margin long, nearly straight. Posterior margin rounded, posterior dorsal margin 34 total dorsal margin sloping from umbos at an angle of about 10°. Valves smooth, sculptured only with rather conspicuous incremental growth lines. Escutcheon and lunule absent. Ligament moderate in size (based on area of attach- ment, the ligament proper was missing in the type specimens) extending about 12 posterior dorsal margin. Periostracum thin and uniform over entire valve. Interior of valve porcelaneous white, mus- cle scars and pallial line impressed. Anterior adductor scar rounded anteriorly, nearly straight posteriorly, well impressed. Anterior pedal retractor triangular in outline, impressed and located near the dorsal, posterior margin of the anterior adductor muscle. Posterior ad- ductor sub-elliptical, lightly impressed, co- extensive with the posterior pedal retractor anteriorly. Pallial muscle scar only lightly im- pressed. Measurements (mm). length height width 44 1725 10.0 USNM 110774: holotype 38.8 16.2 09.3 USNM 205888: paratype 38.7 16.2 USNM 209309: paratype Remarks. Calyptogena elongata is known only from the shells of the three small speci- mens which constitute the type-series and which may be the young of a much larger species. Though most species in the Vesico- myidae are small, those in the genus Calyp- togena such as C. pacifica Dall, C. modioli- forma (Boss) and C. ponderosa Boss are larger. Specimens of these species may reach 100 mm or more in length. Dall (1916) related C. elongata to C. pacif- ica, and Boss (1968) stated that C. modioli- forma was most closely related to, and was the Atlantic homolog of C. elongata from which it differed in being larger and higher. Calyptogena elongata is probably most closely related to C. magnifica, differing in having a more uniform persistent periostrac- um, more anteriorly placed umbos, a smaller ligament, in lacking the “shelf” beneath the ligament so prominent in C. magnifica and in having a similar but more delicate hinge area. See also Remarks under C. magnifica. Unfor- tunately, until the soft anatomy of C. elongata, and for that matter other species of Calypto- gena, is known, it is impossible to state defi- nitely the relationship of these species. Calyptogena (Ectenagena) magnifica, Boss & Turner, species novum Figs. 1-9, 10F-G, 11, 12D-F, 13 Types. Holotype, Mollusk Department, Mu- seum of Comparative Zoology, Harvard Uni- versity, no. 288500, from Galapagos Rift vent ALVIN Dive 717. Paratypes have been de- posited in the Museum of Comparative Zool- ogy (MCZ), the Division of Mollusks, National Museum of Natural History (United States National Museum, USNM), the Department of Malacology of the Academy of Natural Sci- ences of Philadelphia (ANSP), The Depart- ment of Invertebrate Zoology, Los Angeles County Museum (LACM), Mollusca Section of the British Museum (Natural History) (ВММН), Museum National d'Histoire Naturelle, Paris (MNHNP) and the Invertebrate Collection of the Scripps Institution of Oceanography (SIO) and are all from ALVIN dives as follows: 727 (MCZ: USNM); 879 (MCZ); 887 (MCZ); 888 (MCZ); 892 (MCZ); 895 (MCZ); 896 (MCZ); 981 (SIO); 983 (MCZ); 984 (MCZ; ANSP; BMNH; LACM; MNHNP; SIO; USNM); 986 (MCZ); 991 (MCZ). Type-locality. The holotype 15 from ALVIN Dive 717 at 0°47.9'N; 86°08.5'W in 2495 m. Paratypes are from other vents on the Gala- pagos Rift taken on ALVIN dives listed under Specimens examined and from ALVIN Dive 981 on the East Pacific Rise vent at 20°50'N; 109°06’М/ in 2600 m off Mexico. Range. This species is apparently confined to the area of the thermal vents along the Galapagos Rift and the East Pacific Rise from 21°N south to 0%47'N in depths ranging from 2445 to 2680 m. Species examined. Mexico: ALVIN 3Specimens identified from pictures only show that С. magnifica was also seen on ALVIN Dive 909, 20°51.9’N; 109°4.4’W in 2645 т; ALVIN Dive 915, 20°51'N; 109°04.9'W in 2655 т; ALVIN Dive 917, 20°49.9’N; 109°04.8'W in 2655 m (all about 200 miles off Punta Mita, Mexico) as well as ALVIN Dive 733 (see Fig. 1A) from 0°47.3'N; 86°07.8'W in 2496 m. 166 BOSS AND TURNER CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 167 Dive 981, 20°50’М; 109°06’W т 2600 m (about 20 miles W of Punta Mita, Mexico) (1 entire specimen); Galapagos Islands (all on the Galapagos Rift, about 200 miles NE of San Cristobal Island): ALVIN Dive 727, 0°47.4'N; 86°08.9’W, in 2680 m (2 specimens); ALVIN Dive 879, 0°48.18'N; 86°04.11’W, in 2495 т (1 specimen); ALVIN Dive 887, 0°48.5'N; 86°09.1'W т 2488 т (2 speci- mens); ALVIN Dive 888, 0°47.07'N; 86°08.5'W, in 2478 т (1 specimen); ALVIN Dive 892, 0°48.3'N; 86°13.8’W, in 2454 m (4 specimens); ALVIN Dive 895, 0°47.9'N; 86°09.3’W, in 2480 m (1 specimen); ALVIN Dive 896, 0°48.23'N; 86°13.6’W, in 2445 m (1 specimen); ALVIN Dive 983, 0°48.24’М; 86°13.47'W in 2450 т (2 specimens); ALVIN Dive 984, 0°48.24'N; 86°13.47'W in 2450 m (30 specimens); ALVIN Dive 986, 0°47.89'N; 86°9.21'W in 2492 m (2 specimens); ALVIN Dive 991, 0°47.89'N; 86°9.21'W in 2492 m (3 specimens). Description. Shell white, thick, _ brittle, chalky in texture, slightly gaping, equivalve, elongate, and subelliptical; present speci- mens reaching 240 mm in length, 110 mm in height and 60 mm in width (Figs. 2-4). Valves inequilateral; umbos low, abraded in adult, and located on anterior third of valve. Anterior margin of valves rounded, ventral margin long, variable, ranging from nearly straight to rather strongly concave medially; posterior margin broadly rounded, posterior dorsal margin about 24 total dorsal margin, descend- ing from umbo at angle of about 12”; anterior dorsal margin about Уз length and descending at angle of about 26” from umbo (Fig. 11). Valves nearly smooth, sculptured with irregu- lar growth ridges interspersed with fine irregu- lar growth increments (Fig. 3A). Escutcheon and lunule not developed. Ligament massive, extending length of posterior dorsal margin, opisthodetic and parivincular with strong periostracal layer, thick calcareous outer layer and thin inner layer. Periostracum dark brown, coextensive with ligament dorsally, forming thick “ruffled” band along anterior margin in many specimens, extending poste- riorly to about midway along the ventral mar- gin, but reduced to traces on posterior margin and along growth ridges (Fig. 4C-E). Interior of valves porcelaneous white, mus- cle scars and pallial line impressed. Anterior adductor scar rounded anteriorly, irregular posteriorly and deeply impressed. Anterior pedal retractor deeply impressed, elongate, irregular in outline and located slightly dorsal and posterior to anterior adductor scar (Fig. 3B, C). Posterior adductor scar irregularly subelliptical, rounded posteriorly and coex- tensive with small rounded posterior pedal retractor anteriorly. Ventral pallial muscle scar broad anteriorly, becoming narrower over midportion of valve (disc) and broadening again posteriorly where its breadth and slight indentation suggest а “pallial sinus.” Fine, closely spaced scars extending dorsally at right angle from ventral pallial line indicating variable mantle attachments (Fig. 3B). Series of small scars just ventral to hinge line mark- ing dorsal mantle attachments. Distinct, but difficult to discern, scars in umbonal cavity on inner medial surface of cardinal plate marking insertion of ctenidial retractor or elevator muscle (Figs. 5 [no. 11], 8C). Ligament. (Figs. 2A, ЗВ, С, 4А, В, 12D-F). Large strong, external opisthodetic; extending from umbo posteriorly to posterior pedal re- tractor muscles, subtended by elongate nymphal callosities and underlain by highly differentiated fused mantle isthmus. Perios- tracal layer elastic, horny, differentiated into thin, blackish outer portion and horn-colored inner portion, and becoming yellowish poste- riorly. Outer layer thick, calcareous (some- what disintegrated anteriorly in specimen dis- sected) with closely-spaced dorso-ventral cleavage planes. Inner layer thin, immediately ventral to outer calcareous layer and formed by mantle isthmus. For a discussion of the structure of the bivalve ligament see Yonge (1978a). 4A specimen we have not seen was reported by Keen (1977b) to be 250 mm in length; another collected in 1979 by Mr. G. Ellis, an ALVIN Pilot, was 263.5 mm long. IA FIG. 1. A) Expired hot vent at the site referred to by Corliss & Ballard (1977) as Clambake Il (047.3'N; 86°07.8’W; 2496 т). ALVIN Dive 733, with large numbers of dead Calyptogena magnifica Boss & Turner. В) Habitat shot of an active hot vent (referred to as Clambake |) showing a few living Calyptogena magnifica nestled among large numbers of mussels and with a few large galatheid crabs running around and over them (0°47.4'N; 86°08.9’W; 2680 m), ALVIN Dive 727. (Photographs courtesy В. Ballard, Woods Hole Oceano- graphic Institution.) 168 BOSS AND TURNER — FIG. 2. Calyptogena magnifica Boss & Turner. A) Dorsal view of apposed valves. B) Inner view of right valve. С) Inner view of left valve. (Scale is in mm.) Specimens from Clambake |, 0°47.4'N; 86°08.9°W at 2680 m, taken on ALVIN Dive 727. Specimens at Woods Hole Oceanographic Institution. (All pictures by Woods Hole Oceanographic Institution Photographic Laboratory.) CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 169 О FIG. 3. Calyptogena magnifica Boss & Turner. A) Outer view of right valve. В) Inner view of right valve showing external ligament, muscle scars, hinge. C) Inner view of apposed valves to show ligament and hinge when valves are gaping. C) Inner view of apposed valves with the left valve broken showing position of hinge and ligament when valves are closed, and anterior pedal retractor muscle scars. Specimen in A, B and D is 240 mm in length (paratype MCZ 288499) from ALVIN Dive 727. Specimen in С is 190 mm in length (holotype MCZ 288500) from ALVIN Dive 717. (Photographs by A. Coleman, Photographic Laboratory, Museum of Comparative Zoology.) 170 BOSS AND TURNER CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 171 Hinge teeth (Figs. 10F-G, 12E-F). Denti- tion irregular and somewhat worn in larger specimens. Hinge plate concomitantly thick- ened but rather small for large shelled indi- viduals. Left valve with D-shaped subum- bonal cardinal tooth consisting of weak, more or less straight, ridge-like, upcurled dorsal anterior ramus or tooth radiating from umbo, ventral ramus or tooth with two blunt points; excavated D-shaped socket between; weak ridge-like keel or posterodorsal cardinal tooth radiating posteriorly from umbo toward nymphal callosity. Right valve with diverging subumbonal cardinal dentition and lacking an anterior dorsal cardinal element which prob- ably represents a reduction of the D-shaped umbonal cardinal tooth in Calyptogena, s.s. Posterior cardinal tooth blunt to shelf-like and ventral cardinal element more or less sharply keeled, upcurled, and ridge-like; elements separated by D-shaped socket. Somewhat excavated between teeth and anterior dorsal margin. Measurements (mm). height ALVIN length atumbo width Dive 240 110 60 727 Paratype 208.7 87.1 887 Paratype 196.5 86.1 888 Paratype 196.1 88.5 887 Paratype 190 79.4 717 Holotype 188.5 88.4 895 Paratype 180.0 78.2 53.3 879 Paratype 179.5 76.8 45.9 727 Paratype 130.6 57.2 36.8 896 Paratype 82.5 36.5 22.0 892 Paratype 56.7 25:5 984 Paratype 54.8 24.8 984 Paratype 54.2 237, 984 Paratype 34.5 16.8 984 Paratype Animal. Description of morphology of soft parts based on two preserved specimens available for study; the details of internal anatomy based on single specimen (for notes on living specimens see Appendix 1). Mantle and siphons. Mantle lobes bilateral- |у symmetrical, with unusual thickenings both anteriorly and posteriorly (Fig. 5, nos. 2, 36, 54 & 55) firmly attached to shell ventrally by broad pallial muscles and dorsally by series of small muscles just ventral to mantle isthmus. Mantle cavity open ventrally from anterior ad- ductor muscle posteriorly to base of incurrent siphon. Mantle with outer fold producing cal- careous shell layers; periostracal glands on inner surface of outer fold giving rise to exten- sive, thick periostracum, particularly at anteri- or end (Fig. 4C, D), and middle fold with short papillae (Fig. 5, no. 3; Fig. 6A). Inner mantle folds fused anteriorly over anterior adductor muscle to mantle isthmus; posterior fusions of inner lobe forming incur- rent and excurrent siphons (Figs. 6E, 7C) and extending over dorsal surface of posterior ad- ductor muscle to mantle isthmus. Pallial muscles particularly extensive ventral to anterior adductor muscle and in region of siphons (Fig. 5, nos. 36, 55). Anterior thick- ened region of mantle heavily vascularized and glandular. Siphons separate (Fig. 5, nos. 32, 34; Figs. 6E, 7C); incurrent siphon narrow, elliptical, and with numerous (about 40) short papillae on its inner margin; excurrent siphon rounded and lacking papillae. Rather large, rounded “papilla” ventral to incurrent siphon. Muscles and foot. Anterior and posterior ad- ductor muscles (Fig. 5, nos. 4 & 29) sheathed with heavy connective tissue, and composed of antero-ventral “catch” portion and larger mostly postero-dorsal “quick” portion; fibers form discrete bundles (Fig. 6C, D). Anterior pedal retractor muscles (Fig. 5, no. 6) arising in postero-dorsal portion of foot and extending anteriorly through visceral mass to insert on valves beneath hinge plate just pos- terior and dorsal to anterior adductor muscle. Posterior pedal retractor muscles (Fig. 5, no. 26) arising in antero-dorsal portion of foot and inserting on valves adjacent to dorsal anterior portion of posterior adductor muscle. Foot large and composed of two distinct portions, graded dorsally into wall of visceral mass and composed of several discrete layers of longitudinal and oblique muscles. —_ FIG. 4. Calyptogena magnifica Boss & Turner. Holotype MCZ 288500 from ALVIN Dive 717 (see also Fig. 3C). A) Dorsal view of gaping valves, after removal of body. B) Dorsal view of closed valves to show ligament and dorsal extension of the periostracum. C) Ventral view of apposed valves to show width of the specimen and the development of the periostracum along the ventral margin. D) Anterior view of apposed valves to show slight anterior gape and the development of the ruffled periostracal layers along the anterior margin. E) Posterior view of apposed valves to show width of specimen and lack of periostracum. (Photograph A by A. Coleman, Photographic Laboratory, Museum of Comparative Zoology; B-E, Бу В. D. Turner.) 172 BOSS AND TURNER 56 55 54 Sa 529517350 FIG. 5. Semidiagrammatic sketch of the anatomy of Calyptogena magnifica Boss & Turner. The valve, mantle, ctenidium and outer layers of body wall on the left side have been removed and a section of the foot cut away. 1. shell; 2. thickened outer edge of mantle; 3. band of sensory papillae; 4. anterior adductor muscle; 5. labial palp (upper); 6. anterior pedal retractor; 7. mouth; 8. cerebral ganglion; 9. esophagus; 10. cerebro-visceral connective; 11. ctenidial retractor or elevator muscle; 12. stomach; 13. hinge tooth; 14. opening of duct of digestive gland into stomach; 15. intestine; 16. digestive gland; 17. broadened section of intestine with typhlosole; 18. periostracal layer of ligament; 19. outer layer of ligament; 20. mantle; 21. pericardium; 22. ventricle; 23. auricle; 24. pericardial gland; 25. kidney; 26. posterior pedal retractor muscle; 27. branchial пегуе; 28. visceral ganglion; 29. posterior adductor; 30. pallial nerve; 31. anus; 32. excurrent siphon; 33. fusion of inner mantle lobe to form incurrent siphon; 34. incurrent siphon lobes; 35. outer circumpallial nerve; 36. section of muscular portion of posterior mantle which forms the “pallial sinus”; 37. smooth margin of ctenidium; 38. inner circumpallial nerve; 39. torn edge of ctenidial attachment; 40. de- scending lamella of outer demibranch of right ctenidium; 41. ctenidial vein; 42. descending lamella of inner demibranch of right ctenidium; 43. torn edge of ctenidial attachment; 44. ascending lamella of inner demi- branch of right ctenidium; 45. thickened spongy glandular area of mantle; 46. “ducts” in mantle; 47. posterior pedal nerve; 48. mid-gut; 49. ventral pedal nerve; 50. pedal ganglion; 51. statocyst nerve; 52. statocyst; 53. papillose foot; 54. circumpallial “vessel”; 55. thickened anterior mantle; 56. anterior pallial nerve. Ventral portion of foot pointing anteriorly and sub-conical in shape (triangular in side view), strongly rugose and papillate in preserved specimens. Histologically, outer layer highly glandular and inner portion forming complex of crossing muscle fibers. Byssal groove dis- tion of foot; small, byssal gland located at junction with heel; specimens lacking byssus. Ctenidia. Large, thick, homorhabdic, non- plicate, covering entire visceral mass from pericardial cavity to ventral portion of foot, and composed of large inner, and small outer cernible along mid-ventral line of rugose por- demibranchs, both with descending and — FIG. 6. Calyptogena magnifica Boss & Turner (compare these halftones with the drawings in Figs. 5 and 7). A) Opened specimen still partially attached to the valves showing the thickened mantle margin, the papillae, periostracal groove and periostracum. B) Specimen removed from shell, with the left mantle lobe turned back, and with left ctenidium still in place. C) Ventral view showing relation of ctenidia to posterior adductor muscle: a, outer demibranch and b, inner demibranch of left ctenidium; c) inner demibranch and d, outer demibranch of right ctenidium; e, posterior adductor muscle showing discrete muscle bundles; f, opening to excurrent siphon. D) Ventral view of anterior end showing: a) anteror adductor muscle; b, mouth; c, dorsal lip; d, ventral lip; e, outer demibranch of left ctenidium; f) inner demibranch of left ctenidium; д, foot. E) Posterior view showing: a, excurrent siphon; b, incurrent siphon; с, sensory knob; d, fused inner mantle lobe; e, outer mantle lobe; f, posterior adductor muscle; g, muscular portion of posterior mantle. (Photographs by R. D. Turner.) CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 173 BOSS AND TURNER 174 — N SE ККИ < N N ue fe Er ASR < CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 175 ascending lamellae (Figs. 7a, В, ВА, В). In preserved specimens, gills contracted dorso- ventrally, producing antero-posterior ridges and a “herring-bone” appearance to filaments and their chitinous rods (Fig. 8A). Paired demibranchs weakly fused to each other posteriorly and weakly attached to the si- phonal septum distally, thus separating small epibranchial or anal chamber from large in- frabranchial chamber. Ascending lamellae of outer demibranchs fused to visceral mass an- teriorly and to mantle posteriorly; strong inter- lamellar septa uniting lamellae; filaments fused by numerous interfilamentar junctions (Figs. 7A, B, 8B). Ventral margin of both demibranchs appearing smooth and showing no evidence of food groove (Fig. 8A) in pre- served speciens. Digestive system. Mouth small, rounded, located just posterior to the anterior adductor muscle. Labial palps greatly reduced, and consisting of small non-plicate ridges (Fig. 6D), representing vestiges of inner palpal lamellae. Dorsal (= anterior) palpal ridge, co- extensive with the ctenidium, fusing with distal edge of ascending lamella of outer demi- branch; ventral (= posterior) palpal ridge fus- ing with inner demibranch. Mouth opening into short, thin-walled esophagus (Fig. 5, no. 9) leading to thin- walled elongate stomach about four times di- ameter of esophagus (Fig. 5, no. 12). Large, paired digestive diverticula with numerous secondary dichotomising ducts opening into latero-ventral anterior third of stomach via large short ducts. Short combined midgut- style sac (though no style detected) extending from stomach postero-ventrally to thin-walled intestine which recurves sharply dorso- anteriorly paralleling midgut to about midway over stomach where it turns postero-dorsally as flattened ribbon-like structure and passes through visceral mass to pericardial cavity; posterior half of this section of intestine with thickened walls and ventral typhlosole (Fig. 5, и no. 17). Rectum thin-walled in pericardial cavity, surrounded by extensive anterior aorta and muscular ventricle, passing through kid- neys and between posterior pedal retractor muscles. Весит ribbon-like posteriorly, im- bedded in sheath of posterior adductor mus- cle and extending over dorsal surface of pos- terior adductor muscle to terminate at papil- late anus in epibranchial chamber near the opening of excurrent siphon. Stomach of specimen examined empty, ex- cept for clear whitish mucous-like material; rectum also empty and flattened throughout its length. Circulatory and ехсгеюгу systems. Реп- cardium large, elongate and located postero- dorsally on surface of visceral mass and pos- teriorly between kidney and pedal retractor muscles. Pericardial walls somewhat thick- ened but transparent and elaborated ventrally into extensive pericardial glands of two dis- tinct types (Fig. 5, no. 24). Ventricle thick- walled, muscular, surrounding rectum and broadly furrowed dorsally. Anterior aorta also thick walled, extending anteriad from ventricle and surrounding intestine as it enters peri- cardium. Auricles large, paired, thin-walled, triangular, opening into ventricle mid-laterally via small ostia and collecting blood from elon- gate ctenidial sinuses. Remaining portions of circulatory system not discerned but appear- ing to consist of many open sinuses. Kidneys occupying space between peri- cardium and posterior adductor muscle and enveloping posterior pedal retractor muscles. Paired reno-pericardial apertures opening at postero-ventral extremity of pericardium and leading into ventral proximal portion of kidney. Dorsal distal portion of kidney sac-like and apparently interconnected medially. External opening of kidney into suprabranchial cavity not observed. Nervous system. Ganglia and nerves easily discerned (Figs. 5, 9). Cerebral ganglia (Fig. 5, no. 8) situated beneath anteroventral sur- FIG. 7. Calyptogena magnifica Boss & Turner. A) Diagrammatic sketch of section through ctenidium. 1. inner demibranch; 2. outer demibranch; 3. attachment membrane of outer demibranch; 4. outer surface of ascending lamella of outer demibranch; 5. cut edge of demibranch; 6. outer surface of descending lamella of outer demibranch; 7. attachment area and blood vessel; 8. interlamella septum; 9. attachment membrane of inner demibranch; 10. ascending lamella of inner demibranch; 11. descending lamella of inner demibranch (see also Fig. 8A, B). B) Diagrammatic enlarged section of demibranch. 12. interlamellar septum; 13). interfilamental space; 14. filaments; 15. interfilamental junctions. C) Diagrammatic sketch of posterior siphonal area (see also Fig. 6E). 1. posterior adductor muscle; 2. right outer mantle lobe; 3. fusion of inner mantle lobes; 4. anus; 5. excurrent siphon; 6. incurrent siphon; 7. papillae around incurrent siphon; 8. “sensory knob” or “button”; 9. open mantle cavity; 10. muscular portion of mantle attachment; 11. left outer mantle lobe. 176 BOSS AND TURNER CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT WAT. face of anterior pedal retractor muscle and directly above mouth, closely juxtaposed and connected by very short supraesophageal commissure. Pallial nerves arising from an- terior portion of cerebral ganglia and giving off branches to anterior adductor muscle. Neither buccal ganglia nor labial palp nerves ob- served. Cerebro-visceral connectives arising from posterior lateral regions of cerebral ganglia and coursing posteriorly along sides of visceral mass to connect with visceral ganglia anterolaterally. Pleural ganglionic thickenings present on the cerebro-visceral connectives near the junction of esophagus with stomach. Cerebro-pedal connectives arising from anterior lateral region of cerebral ganglia and passing posteroventrally into foot and enter dorsal surface of partially fused, but distinctly bilobate pedal ganglion. Branches innervating intrinsic foot muscles and deep portions of anterior pedal retractor muscle arising from connectives. Fused pedal ganglia giving rise to three pairs of nerves: 1) ventral pedal nerves dividing into medial and lateral rami serving distal papillose portions of foot; 2) posterior pedal nerves bifurcating into medial and lateral branches and innervating posterior intrinsic muscle, deep portions of posterior pedal retractor muscle and byssal gland; and 3) long anterior nerves terminating in hollow bulb-like statocyst with refractive granular statolith. Paired visceral ganglia, partially fused, and located on anterior sur- face of posterior adductor muscle. Branchial nerves arching anteriorly from lateral portion of visceral ganglia and recurving to innervate axes of ctenidia. Visceral ganglia giving rise posteriorly to large pallial nerves with numer- ous branches to posterior adductor muscle and dorsal mantle. Pallial nerve ventrally bi- furcating to form outer and inner circumpallial nerves, sending numerous discrete branches to excurrent and incurrent siphons and asso- ciated sensory structures as well as extending along ventral mantle margin. Two small pro- tuberances on ventral surface of posterior ad- ductor muscle and closely associated with visceral ganglia probably represent abdomin- al sense organs. Remarks. As a member of a family of typi- cally burrowing infaunal bivalves, Calypto- gena magnifica is an unusual species not Only for its size but for its remarkable ability to exploit a unique epifaunal niche, nestled in crevices and among mussels surrounding the abyssal hot vents. This new species, on the basis of shell characters, is most closely related to C. elongata but differs in attaining greater size, having the umbos located more posteriorly and having a much larger ligament and longer nymphal callosity (see Fig. 12 for a compari- son of these characters). In addition, the periostracum of C. magnifica, even in very small specimens does not remain on the disc though it may be present as a large “ruffle” along the anterior margin (see Fig. 4C-D). п C. elongata the periostracum is thinner and persists over the entire valve (compare Fig. 12A and 12D). In addition when comparing specimens of the same size the cardinal den- tition is much coarser though similar in C. magnifica. This new species is also closely related to С. modioliforma, but differs in being much larger, having the umbos more posterior, be- ing more elongate and having the perios- tracum on the margin of the valves more high- ly developed. The dentition of С. modioliforma tends to be more blunt, thickened, and rela- tively more extensive than т С. magnifica. In addition, as С. magnifica increases in size, the subumbonal cardinal tooth tends to curl upwards (Fig. 10F-G). Anatomically, C. magnifica resembles C. pacifica and C. kilmeri—the only species of Calyptogena for which we have any knowl- edge of the soft parts (Bernard, 1974). The elaborations of the mantle folds, comparable to those in other heterodont bivalves (Yonge, 1957) appear to be even more strongly differ- entiated т С. magnifica than in С. pacifica or C. kilmeri. Sensory papillae along the anterior mantle folds and an extensive periostracal border anteriorly are characteristic of C. magnifica (Figs. 4C, D, 5, 6A, B). The mantle thickenings in the anterior ven- tral region of C. magnifica are more highly — ————— FIG. 8. Calyptogena magnifica Boss & Turner. A) Closeup of posterior end of external surface of left ctenidium to show smooth ventral margins and texture of demibranchs (see also Figs. 6 and 7). B) Dorsal view of left ctenidium to show: a, outer surface of ascending lamella of outer demibranch; b, connecting membrane to mantle; с, inner surface of descending lamella of outer demibranch; d, ctenidial axis with blood vessel and muscle; e, inner surface of descending lamella of inner demibranch; f, inner surface of ascending lamella of inner demibranch (see also Fig. 7A, B). C) Closeup of ctenidial muscle (see also Fig. 5). 178 BOSS AND TURNER LA. branches to mantle margin -------- to anterior adductor muscle a FE anterior pallial nerve Les oe = supraeosphageal commissure -------cerebral ganglia ------ pleural gandionic thickenina ----- cerebro-pedal connective - ---- пегуе to statocyst ee statocyst ----fused pedal ganglia ----ventral pedal nerve ----posterior pedal nerve ——-cerebro-visceral connective --branchial nerve ¿visceral ganglia + 7° r===>posterior pallial nerve to posterior adductor muscle ---to pedal retractor muscle ---outer circumpallial nerve ---- inner circumpallial nerve 4--}---to posterior mantle muscle --branches to excurrent siphon ~branches to incurrent siphon A} --Z4--branch to inner mantle margin /A-to outer mantle mantle margin 47 CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 179 ligament escutcheon FIG. 10. The hinge and ligament of Calyptogena. A-C) Species in Calyptogena, s.s. Note small ligament and the presence of an escutcheon. D-G) Species in the subgenus Ectenagena. Note the presence of a large ligament and the lack of an escutcheon. A a, b) С. (Calyptogena) ponderosa Boss. Holotype, Oregon |, sta. 1426, 29°07'N; 87°54'W, about 77 mi. S of Mobile Bay, Gulf of Mexico, in 1097 m. Ва, b) С. (Calypto- gena) pacifica Dall. Syntype, Albatross sta. 3077, Clarence Strait, Dixon Entrance, Alaska, 55°46’М; 132°24'W, in 580 m. Young specimen 30.5 mm long. С а, b) /bid. Adult specimen 47.5 mm long. D а, b) С. (Ectenagena) modioliforma Boss. Holotype, Pillsbury sta. 394, 9°28.6'N; 76°26.3'W; Golfo del Darien, 66 mi. NNE of Punta Caribana, Colombia, in 421-641 m. E a, b) C. (Ectenagena) elongata Dall. Holotype, off Point Loma, California, in 486 т. Specimen 44 mm long. F а, b) С. (Ectenagena) magnifica Boss & Turner, Galapagos Rift, ALVIN dive 892. Specimen 82.5 mm long. С a, b) С. (Ectenagena) magnifica Boss & Turner, Galapagos Rift, ALVIN dive 896. Specimen 130.6 mm long. ug un FIG. 9. Diagrammatic sketch of the nervous system of Calyptogena magnifica Boss & Turner from the dorsal aspect. (See Fig. 5 for the lateral view of the essential elements of the nervous system.) 180 BOSS AND TURNER 38.8 paratype 44mm holotype FIG. 11. Inset outlines of a graded series of valves of С. magnifica Boss & Turner obtained from the Galapagos thermal vents (in solid lines) as well as the holotype and paratype of C. elongata Dall (in dashed lines) to show the gradual increase in the curvature of the ventral margin with increase in size and the proportional difference in shell length and height position of the umbos in C. magnifica and C. elongata.. developed than in either C. pacifica or C. kilmeri, and may be similar in function to the so-called ‘pallial gills’ or Mantelkiemen of the Lucinacea (Duvernoy, 1853; Semper, 1880; Pelseneer, 1911; Allen, 1958) or to the pallial glands of carditaceans, which function in the cleansing of the mantle cavity (Pelseneer, 1911; Harry, 1966; Allen, 1968; Yonge, 1969). п Calyptogena, however, neither special transverse folds nor definitive pallial blood vessels leading directly to the auricles, as in the lucinids, are apparent though some kind of secondary respiratory function may be surmised. It seems possible that this region in C. magnifica even though it is not between the outer and middle mantle folds (Figs. 4C, D, 6A), is developed for the massive produc- tion of periostracum, which may protect the anterior end of the valves against the environ- ment of the hot vents or assist, perhaps, in maintenance of position. The region might also be similar in function to the anterior man- tle thickenings or pallial mucous glands found in the Saxicavacea (Yonge, 1971). Bernard (1974: 13) described as hyper- trophied the posterior portion of the thickened mantle musculature of С. pacifica. The im- pression on the valves of this extensive poste- rior development 15 variable and has been called, by numerous authorities, the ‘pallial sinus.’ The siphons of vesicomyids are not greatly extendable and the radiating fan-like siphonal retractor muscles which form the pal- lial sinus in so many families of bivalves (Pelseneer, 1891; Boss & Kenk, 1964) are not present. The so-called ‘pallial sinus’ is a dubious character for taxonomic discrimination in Calyptogena or other vesicomyids, but in a group with few differentiating features, this character may have to be employed. As can be seen from Figs. 2B, C, 3B, the ‘pallial sinus’ is only a reflection of the extent, in an anterior-posterior axis, of the thickening or hypertrophy of the edge of the mantle. The ‘pallial sinus’ of C. magnifica is 14% of total shell length while in C. pacifica it is 22%. Un- fortunately, when the internal surface of the shell is glossy, the configuration and extent of the ‘pallial sinus’ is difficult to determine. The sensory papillae of the mantle edge are protrusible posteriorly in the siphonal re- gion and anteriorly in the vicinity of the maxi- mum periostracal development. The fused in- ner lobe of the mantle ventral to the incurrent siphon forms an extrusible velum in C. pacifi- ca and C. kilmeri (Bernard, 1974). The same appears to be true for C. magnifica. The “sensory knob” beneath the incurrent siphon (Figs. 6E, 7C) is similar to the “sensory but- ton” in Thyasira (Allen, 1958: 435, fig. 10b). In Calyptogena, as with many bivalves (Yonge, 1962), the byssus may be functional only during settlement and in the young stages. The foot, capable of protrusion be- yond the valves, probably functions in loco- motion and positioning. In contrast to the rath- er smooth pedal integument found in С. pacif- ica and C. kilmeri (Bernard, 1974), the foot of C. magnifica is highly rugose, roughened, or subpapillose with the distal portion containing glandular structures. Since most species of CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 181 x Reaca. _ N AA ern A | ye $ FIG. 12. Comparison of Calyptogena elongata Па! and С. magnifica Boss & Turner. А-С) Calyptogena elongata, paratype USNM 205888 (Albatross sta. 4432, off Point Loma, California, in 275 fathoms, 8 mi. $ of Brockway Point, Santa Rosa Id., Channel Ids.(see also footnote 2). A) Outer view of valves to show position of the umbos in the anterior Ya of the valves and retention of the periostracum. В) Inner view of valves to show hinge and position of the relatively weakly impressed muscle scars. C) Close-up of hinge area to show small ligamental area (specimen 38.8 mm long). D-F) Calyptogena magnifica, paratype MCZ, ALVIN Dive 984, 0°48.24'N; 86°13.47'W in 2450 m. Galapagos Rift. A) Outer view of valves showing lack of periostra- cum except on the edge of the valves, the position of the umbos at the anterior "Уз of the valves and the large ligament extending the length of the posterior dorsal margin. B) Inner view of valves showing the large cardinal teeth, the large ligament and the well-impressed muscle scars. C) Close-up of the hinge area to show the large ligament and the shelf bordering it (specimen 34.5 mm long). (Photographs by R. D. Turner and C. B. Calloway.) 182 BOSS AND TURNER FIG. 13. Dense set of large white clams (Calyptogena magnifica Boss & Turner). Photograph taken on ALVIN Dive 909, 20°51.9’N; 109°4.4’W, in 2645 m, about 200 mi. W of Punta Mita, Mexico. (Photograph by Bruce Luyendyk, University of California, Santa Barbara.) Calyptogena are assumed to burrow, the dif- ferentiation of the foot of C. magnifica may be a specialization for living on hard substrates nestled in cracks and crevices or among other organisms, especially large mytilids also found around the hot vents. The reduction of the labial palps to mere lip-like folds above and beneath the mouth is typical of Calyptogena and probably of other vesicomyids (e.g. Callogonia) (Boss, 1969a; Bernard, 1974). Thiele and Jaeckel (1931), describing V. striata, noted: “Die Mundlappen sind schmal und ziemlich kurz.” Thiele (1935) incorporated that observation in his remarks on the family Kellyellidae (sic) in which he in- cluded Vesicomya. However, reduction or loss of labial palps, usually correlated with re- duced selection of particulate food, is a sec- ondarily derived feature convergent in several distinct lineages of bivalves: lucinids (Thiele, 1886: 247, fig. 23; Purchon, 1939; Allen, 1958), limids (Stuardo, 1968); dimyids (Wal- ler, 1978; Yonge, 1978b), mytilids (Yonge, Goreau 8 Goreau, 1972) and teredinids (Turner, 1966). Calyptogena shows a reduced palp config- uration similar to that of Phacoides (Allen, 1958: fig. 39) and is unlike other primitive heterodonts such as Astarte (Saleuddin, 1965) or the isocardiacean Glossus (Owen, 1953) with fully developed palps. The ctenidia of vesicomyids were originally described by Dall (1895a: 505; 1895b: 696) as being protobranch-like, thick and fleshy, lacking both interlamellar septa and a com- pleted anal or epibranchial chamber. He thought that the gills, though fused distally to the siphonal septum, were not fused to each other. Ridewood (1903: 224-226, fig. 23) dis- agreed with Dall and showed that the gills of V.stearnsi were eulamellibranch with small outer demibranchs and closely packed fila- ments. He stated that in V. stearnsi the ascending lamellae of the outer demibranchs were not fused to the mantle nor were the gills fused to each other posteriorly but pointed out that the gills were so rigid that they would ‘readily come away from adjacent parts even if organically united to them.’ Such a configu- ration is certainly not true of C. magnifica. We were able to see the connections before be- ginning the dissection; however, these quickly CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 183 separated, leaving no evidence of connec- tions. The interlamellar septa in V. stearnsi are weak, and the ctenidia readily separate into plates. In contrast, the interlamellar septa in С. magnifica (Fig. 7B) are strong and hold the lamellae of the demibranchs together. Ac- cording to Bernard (1974), neither C. pacifica nor С. kilmeri possess interlamellar septa. Ridewood (1903) concluded that Vesi- comya resembled Lucina rather than the Protobranchia but it is probable that the simi- larities are, in part, convergent as both these genera have long geological histories, espe- cially lucinids which date from the Paleozoic. In fact the ctenidia are rather poor distinguish- ing familial taxobases in lamellibranch bi- valves. Ridewood himself (1903: 186) was unable to give diagnoses for the numerous suborders he employed, much less discrimi- nating features of individual families. The placement and gross structure of the kidney of С. magnifica conforms with the type found in eulamellibranch bivalves as de- scribed by Odhner (1912) and is similar to that of Kelliella miliaris (Clausen, 1958). According to White (1942), pericardial glands function to excrete acids from the blood either through blood sinuses in the mantle or through the wall of the pericardium and auricles. Products discharged into the pericardial cavity pass through the reno- pericardial aperture and are extracted by the kidneys. In С. magnifica the pericardial glands, consisting of differently colored moieties, lie mostly on the floor of the peri- cardium; the ventricle is tubular and em- braces the rectum. This morphology is unlike that of the Carditidae, with which vesicomyids have sometimes been placed in that the ven- tricle of carditids lies mostly beneath the rec- tum (White, 1942). In Arctica, both White (1942 [as Сургта]) and Boltzmann (1906) noted the extensive de- velopment of the pericardial gland within the pericardium and anteriorly over the visceral pedal mass in the mantle. However, unlike that of Calyptogena, the ventricle in Arctica is short and rectangular; the renopericardial apertures are on the posteroventral wall of the pericardium. In its general anatomy, the nervous system of Calyptogena magnifica does not differ sig- nificantly from most eulamellibranchs (Lammens, 1969; Bullock & Horridge, 1965). Unlike the statocysts of Kelliela miliaris, which are close to or incorporated in the pedal ganglion (Clausen, 1958), the statocysts in Calyptogena are large, well-developed and located in the foot some distance from the pedal ganglia (Fig. 5). Specimens available to us were not in proper condition for detailed work on the re- productive system and no previous descrip- tive work is known; however, sections taken in the visceral mass just anteriad the pericardi- um and dorsad the stomach showed the pres- ence of numerous, large, yolky oocytes, measuring 150-195 um in greatest diameter, in germinal vesicle stage. On the basis of anatomical evidence, C. magnifica can be separated from C. pacifica and C. kilmeri by the development of strong interlamellar septa between the lamellae of the demibranchs. Lesser features in C. mag- nifica include the glandular structure on the surface of the foot, the more extensive devel- opment of the anterior portion of the edge of the mantle and the greater proliferation of the periostracum. Possibly the pericardial glands may differ from those of C. pacifica which are prolonged anteriorly (Bernard, 1974), but it is not known whether they enter the visceral mass or mantle tissue. APPENDIX 1. Description of living specimens of Calyptogena magnifica Boss & Turner with notes on their distribution and ecology Carl J. Berg and Ruth D. Turner During the 1979 Biological expedition to the thermal vents of the Galapagos Rift (LULU- ALVIN Cruise 102, leg 9 and Cruise 103, leg 5) we made observations on living Calypto- gena magnifica to supplement the description of the species by Boss & Turner based on preserved specimens. We made observations of specimens in situ from ALVIN, of speci- mens maintained in the aquarium aboard LULU, while taking samples for histological and electrophoretic studies, while preparing valves for age determinations, and while pre- serving the soft parts for anatomical studies. Additional data on distribution, orientation and behavior of the large white clams were ob- tained from discussions with other scientists who had visited the sites in ALVIN as well as from videotapes and photographs taken on the various dives. During these two series of dives 61 living Calyptogena magnifica were collected, 15 on Cruise 102 and 46 on Cruise 103. Specimens ranged in size from 34.5 to 241 mm in length, the largest ones (188- 184 BOSS AND TURNER 241 mm in length) were taken at the vent named “Mussel Bed” (0°48.5'N; 86°09.0'W in 2480 m and the smaller ones (34.5-151 mm in length) from “Rose Garden” (0°48.3’М; 86°13.8’W in 2450 т). The smaller size of the white clams, the lighter color and smaller size of the mussels combined with the abundance and great size of the vestimentiferans at “Rose Garden” suggest that this may be the younger and more active of these two vent areas. Although large groups of empty clam shells were seen at most of the vents, the living clams were scattered and in small groups, usually of ten or fewer. They were nestled among dead clams, living mussels or in rock crevices, usually oriented in a nearly vertical position, the iridescent pink mantle and the openings of the siphons often show- ing between the slightly gaping valves. This suggests that larval clams settle in the pres- ence of other clams and/or in response to warm water issuing from the fissures in the lava rock and that, as they grow, they main- tain a position with the foot probing for warmth. The clams were never observed with the valves more than slightly gaping and they would often remain closed for long periods of time, particularly in the aquarium. The mantle of living specimens is an iridescent purple- pink, the papillae on the incurrent siphon yel- lowish. In situ observations showed that the siphons opened and closed frequently, that they could be extended slightly but that they never extended beyond the valves. They are capable, however, of forcibly ejecting material a distance of at least 30 cm from the siphonal aperture. А yellowish, transparent, tough and stretchy, plastic-like periostracum extends from the periostracal groove of the mantle margin out over the valve protecting the grow- ing edge of the shell. When removing the soft parts from the valves it was necessary to cut this with scissors so firmly was it attched to the outside of the valve and to the mantle. It appeared as a brown wrinkled border along the margins of the valves, being thin posteri- orly and increasing to multiple ruffles at the anterior end. An isolated specimen, possibly disturbed by ALVIN activity was observed lying on its side at the edge of a cluster of clams. The specimen, approximately 150 mm in length, had extended its foot about half the length of its shell but showed little activity except weak probing movements during the half hour it was in view (while the pilot was picking up microbiological experiments for return to the surface). The foot was lighter but similar in color to the mantle and appeared smooth. Specimens brought to the surface, however, all appeared to have a strongly rugose foot, a condition which might be due partly to con- traction and/or reduction in pressure. The rugosities were blister-like and when punc- tured released a clear fluid. The foot of a specimen kept for two days in the aquarium, maintained at about 2°C, became less rugose, and the foot of one maintained under pressure became smooth, suggesting the possibility that these clams have some ability to adapt to changes in pressure. The proximal half of the foot, visible only after the shell has been opened and the mantle and gills turned back, is smooth light brown and streaked with dark red blood vessels. Small clams are light- er in color and have a cream-colored foot reminiscent of Mercenaria. The ctenidia in large, newly collected spec- imens are light brown with pronounced zigzag longitudinal streaks of red and finer, more-or- less vertical mottlings of reddish-brown. These markings are perhaps a reflection of the 'herring-bone” configuration of the gill filaments described by Boss 8 Turner. The ventral margins of the ctenidia were a purple- pink and there was a grayish line at the junc- tion of the inner demibranch with the visceral mass. The gills of small clams were a uniform cream color. Many clams had notches on the posterior ventral margin of the outer gills, the number varied from one to three and was not necessarily the same on both sides of the animal. We have no explanation for these notches which occurred on clams found at all vents from which material was collected. The large white clams, unlike the mussels commonly found around the vents, have red blood, which gives a red appearance to the animals when the shell is opened. The red pigment is intracellular, and is a haemoglobin but it has not been fully characterized as yet. The pericardial region has a dark red colora- tion which extends laterally down to the ctenidia. No pumping of the heart was ob- served, nor did the blood spurt when th clams were cut or when blood samples were taken. The non-muscular portion of the mantle was thin, nearly transparent and reddish-brown. Small specimens are much lighter in color than the larger ones but the red color of all specimens increased with time as they re- mained in the aquarium. CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 185 The position of the gonads, though typical for bivalves, cannot be determined by general examination of the body surface even in living specimens, probably due to the overall red coloration of the animal. Dissection and his- tological examination of specimens collected in February 1979 showed ova in all stages of development and yolky eggs 309 ит in di- ameter. Eggs found lying free in a jar with a preserved clam ranged from 364-482 ит in diameter. These data suggest that C. mag- nifica releases large yolky eggs. The clams did not react when crabs or shrimp crawled over their shells and touched the mantle but they closed slowly when han- dled by the submarine's manipulator. Crabs are obviously not predators on the larger clams though they may feed on young specimens. Octopus have been seen at the vents and are suspected predators on the larger clams. APPENDIX 2. Annotated list of fossil and living species currently referable to Calyptogena Kenneth J. Boss ?Calyptogena akanudaensis Tanaka Calyptogena akanudaensis Tanaka, 1959: 119, pl. 2, figs. 1-8; Hanzawa, Asano 4 Takai, 1961: 219 (type-locality, cliff along the moun- tain-side, about 1.5 km E of the Nishikibe Ele- mentary School, Shigamura, Higashi- Chikuma-gun, Nagano Prefecture, Bessho Formation, Мюсепе; Holotype, no. 510; para- types, nos. 524-531, Geological Institute, Faculty of Education, Matsumoto Branch, Shinshu University); Okutani, 1966b: 301). Remarks. Despite considerable efforts, we have not been able to obtain the original ref- erence and follow Okutani (1966b: 301) in considering this species, though initially re- ferred to Calyptogena, as a questionable member of the genus. Range. Fossil in Miocene of Japan. Calyptogena (Calyptogena) chitanii (Kanehara) Adulomya chitanii Kanehara, 1937: 19-20, pl. 5, figs. 1, 6, 7, 8, 9 (localities, Ashikaya- Zawa, Sekinami-mura; Takai and Nakosono- seki, Sekimoto-mura; Hatanaka Shizuzaku, Izumi-mura; Nishinakada, Watanabe-mura; geological horizon, Mizunoya Shale, Kamen- owo Shale, Yunagaya Series; Shirado Series; [Jóban coal field, Fukushima Prefecture, Japan, Miocene; for type-specimens, see Hanzawa, Asano & Takai, 1961: 211); Aoki, 1954: 31-32 as “Adulomya,” pl. 1, figs. 9, 10, 11 (localities, cliff of small valley, Dónosaku, Kamikatayose, Kabeya and small cliff of path- side in small valley of Kamikatayose, Kabeya. Kabeya Formation [lshimori district, Jóban coal field, Fukushima Prefecture, Japan; Мосепе]); Hanzawa, Asano 8 Takai, 1961: 211 (type-locality, Ashikaya-zawa, Sekinami- mura, Jóban coal field, Fukushima Prefec- ture, Mizunaya Shale, Miocene; holotype and paratype, Geological Survey, Japan); Kamada, 1962: 3941 (localities, Shisawa, Nakoso City. Kamenoo Formation; Fukuda, Sekinami, Kitaibaraki City. Kamenoo Forma- tion; Nagako, Nishikimachi, Nakoso City. Kamenoo Formation; Kanegasawa, Hisano- hama-machi. Kamenoo Formation; Tango- zawa, Shiroyama, Taira City. Honya Forma- tion; northern cliff of Taira railroad station, Shiroyama, Taira City. Honya Formation. Yagawase cliff, Taira City. Honya Formation; Tsuruga-machi, lino, Taira City. Misawa For- mation [Jóban coal field, Fukushima Prefec- ture, Japan, Miocene)). Akebiconcha chitanii (Kanehara). Kanno 8 Ogawa, 1964: 285-286, pl. 1, figs. 17-18 (up- per part of the Takinoue Formation [Hokkaido, Japan]); Kanno, 1967: 401-402, pl. 1, figs. 9-11, 15 (Itsukaichi [machi group] Basin, Tokyo Prefecture; Miocene). Calyptogena chitanii (Kanehara). Kanno, 1971: 80-82, pl. 7, figs. 5, 6a-b; pl. 17, fig. 12, text-figs. 10, 11, and 12 (Kayak Island [Alas- ka], Yakataga Formation [upper middle Mio- cene or upper Miocene)). Remarks. We include this species in Calyp- togena following its placement there by Kan- no (1971), who figured the hinge, showed the range of its allometric change in shape, docu- mented its extensive occurrence in the fossil record, and contrasted it with both С. elongata and C. pacifica. Probably referable to Calyptogena, s.s., С. chitanii is similar to С. pacifica, especially in regard to the dentition of its left valve (contrast Kanno, 1971: text-fig. 10, nos. 2, 3 with Fig. 10, Ba, Ca). Represent- ing a Miocene ancestor of Calyptogena in the northern Pacific Basin, С. chitanii is apparent- ly separable from C. pacifica only by its more narrowly elongate and arcuate form; further distinctions might be made in regard to the development of the escutcheon and nature of the dentition of the right valve, if suites of 186 BOSS AND TURNER specimens were available for comparative analysis. Range. Miocene (upper middle or upper) in Alaskan and Japanese formations [Kamenoo Formation, Itsukaichi-machi group, Нопуа Formation, Kabeya Formation, Takinoue Formation, and Yakataga Formation] (see Kanno, 1971). Calyptogena (Ectenagena) elongata Dall Remarks. See text, p. 164. Calyptogena (Calyptogena) Каматига! elongata (Ozaki) Akebiconcha kawamurai elongata Ozaki, 1958: 123, “pl."5, -figs. 3, 4, pl.+6, figs. 1-5 (cotype from Ebishima (type-locality, small islet off Inuwaka, Tyósi, Na-arai Formation (Lower Pliocene) and paratype from roadside cutting in front of Electric Car Station of Zinmuzi, Miura Peninsula, Kanagawa Pre- fecture; Ikego Formation (Lower Pliocene) non Dall, 1916; Hanzawa, Asano & Takai, 1961: 221 (syntype from Ebishima, off Inu- waka, Choshi City, Na-arai Formation (Mio- cene); Shikama, 1962: 53; Okutani, 1966b: 301 (Pliocene). Remarks. This is a synonym of C. kawa- murai, С. nipponica and probably С. soyoae, and, since it was originally described т Akebiconcha, it became, when transferred to Calyptogena, a secondary homonym of C. elongata Dall; the population named by Ozaki (1958) constitutes only an individual variation (Shikama, 1962). Calyptogena (?Calyptogena) gibbera Crickmay Calyptogena gibbera Crickmay, 1929: 93, 1 fig. (type-locality, lowest bed of the Santa Barbara Formation, on Deadman's Island, near San Pedro, California [Pliocene]; Wood- ring, 1938: 51 (early Pleistocene silt of Dead- man's Island [Arnold's Pliocene; Alex Clark's Timms Point Formation]); Woodring, Bram- lette & Kew, 1946: 83. Remarks. When naming this species, Crick- may (1929) compared it with C. elongata and С. pacifica, though the hinge was not de- scribed and therefore a subgeneric placement is uncertain; however, С. gibbera very much resembles C. pacifica, especially in its size, proportions and gross outline (52 x 29 x 15 mm) and is probably synonymous with C. pacifica; it differs from С. elongata in not be- ing narrowly elongate. Range. Pleistocene, Deadman's Island, California. Calyptogena (Calyptogena) kawamurai (Kuroda) Akebiconcha kawamurai Kuroda, 1943: 17, text figs. 1-3 (type-locality, off Odawara, Sagami Bay, in about 100 fathoms); Habe, 1952: 159, pl. 22, figs. 20-31; Hatai & Nisi- yama, 1952: 33 (as synonym of Calyptogena nipponica); Habe, 1961: 122, pl. 55, fig. 16 (100-200 m, Sagami Bay); Okutani, 1957: 28; Ozaki, 1958: 124, pl. 3, figs. 1-3, pl. 4, figs. 1-3, pl. 5, figs. 1, 2 (Kashima-Nada); Okutani, 1962: 23 (700-750 m, Sagami Bay); Okutani, 1966b: 301; Shikama, 1962: 53, pl. 3, figs. 6a-d, 7a-c (about 46 miles east to southeast of Chöshi, 200-230 fathoms); Habe, 1968: 179, pl. 55, fig. 16 (100-600 m in Sagami Bay to Kashima-Nada, Honshü); Habe, 1977; 237, pl. 50, figs. 34. Calyptogena kawamurai (Kuroda). Ber- nard, 1974: 18. Remarks. The original Japanese descrip- tion of C. kawamurai by Kuroda (1943) has been rendered into English by Ozaki (1958). The nomen is a synonym of C. nipponica as indicated by Hatai & Nisiyama (1952: 33). In all probability, C. nipponica will come to be recognized as the senior synonym of not only C. kawamurai, and C. elongata Ozaki, but of C. soyoae as well. Range. Recent from Kashima-Nada and from off Chóshi to Sagami Bay, in 100 to 750 m. Calyptogena (Calyptogena) kilmeri Bernard Calyptogena (Archivesica) kilmeri Bernard, 1974: 17-18, text-figs. 1B, 2B, 3B and 4E (type-locality, off west coast of Moresby Island, Queen Charlotte Islands, British Columbia, Canada, in 1170 m). Remarks. Bernard (1974) provides several additional localities besides the type-locality. Because of the similar nature of its dentition and the absence of any reliable distinguishing anatomical traits from C. pacifica (see Ber- nard, 1974: 18), we place C. kilmeri in Calyp- togena, s.s. although we have not included Archivesica in the synonymy of Calyptogena. Bernard apparently utilized Archivesica for C. CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 187 kilmeri because of its relatively thin shell, stat- ing ‘Archivesica ...thinner — shelled... modioliform ...small pallial sinus.” Archi- vesica, based on its type-species Callocardia gigas Dall, from the Gulf of California, may indeed be thinly shelled but it also can be thick and heavy (e.g. USNM 266874). Previously, Boss (1967; 1968) related Archivesica gigas to such species as Vesi- comya caribbea Boss from the Caribbean Sea, V. chuni Thiele & Jaeckel from off West Africa, V. winckworthi Prashad from the East Indies and V. leeana Dall from North Carolina to Tobago in the western Atlantic. One might also include: V. angulata Dall from Panama Bay, V. longa Thiele 8 Jaeckel from the Gulf of Guinea and V. suavis from off Lower Cali- fornia. Thiele & Jaeckel (1931) placed V. chuni in Callogonia, a procedure subsequent- ly followed by Boss (1969a: 254; 1970: 69), who suggested that Archivesica might be considered a synonym of Callogonia. Until we have a better understanding of the supraspecific categories of this group and al- though Archivesica may fall into the synony- my of Calyptogena, we presently consider it a synonym of Callogonia, and conserve that as a subgenus of Vesicomya for Archivesica gigas and its relatives as mentioned above. This group, which may have shells of variable thicknesses, usually shows inflation of the valves, a short prominent ligament with a concomitant .subtending nymphal callosity; a demarcated escutcheon 1$ lacking and the pallial sinus, or posterior sinuosity of the palli- al line is slight but usually noticeable, and sometimes rather pointed or angled above. Range. Living from British Columbia, Canada, to northern California (53°-40°N), in 549-1464 m. Calyptogena (Calyptogena) lasia (Woodring) Phreagena lasia Woodring, 1938: 50, pl. 5, figs. 3, 4, text-fig. 2a (type-locality, Standard Oil Co., Baldwin No. 73, Montebello field, depth 3340-3358 feet, United States Geo- logical Survey locality 13864, Repetto Forma- tion, Lower Pliocene, Los Angeles Basin; holotype, USNM 496097). Calyptogena lasia (Woodring). Winterer & Durham, 1962: 295, 302, 307, 308 (Ventura Basin, Los Angeles County); Boss, 1968: 739. Remarks. Woodring (1938) lists numerous additional localities for С. /asia in the Repello and Pico formations of the Los Angeles Basin and discusses the relationship of this species with other vesicomyids, especially C. pacifica and С. elongata. Winterer & Durham (1962) add several other occurrences of С. lasia in the Ventura Basin. Woodring (personal com- munication) concurred in the synonymy of Phreagena and Calyptogena. Range. Fossil in Lower Pliocene of Los Angeles and Ventura Basin, California. ?Calyptogena longissima (Yokoyama) Cucullaea longissima Yokoyama, 1925: 20, pl. 3, fig. 1 (type-locality, Shigarami); Maki- yama, 1958: pl. 27. Calyptogena longissima (Yokoyama). Hatai & Nisiyama, 1952: 56 ([Shigarami] Pliocene, Shigarami [a short distance N of Shimoso- yama, Shigarami-mura, Kami-Minochi-gun, N. Nagano, 36°40'N; 13804 'E]; holotype, GT по. ?); Hanzawa, Asano & Takai, 1961: 233 (holotype, Geological Institute, University of Tokyo; Shigarami, Nagano Prefecture, Shigarami Formation, Pliocene); Okutani, 1966b: 301. Remarks. We have followed several differ- ent Japanese authorities (Hatai 4 Nisiyama, 1952; Hanzawa, Asano & Takai, 1961), in placing this species in Calyptogena, s.l., with some doubt as to the propriety of this assigna- tion (Okutani, 1966b: 301). The species is based on an internal cast, measuring 115 x 55 х 35mm, roughly shaped like a Calypto- gena. The pallial line is distinctly visible, rela- tively wide, strongly impressed, and weakly sinuate posteriorly. Range. Fossil in Japan, Shigarami Forma- tion, which was considered Pliocene by Hatai & Nisiyama (1952: 332) and Hanzawa, Asano & Takai (1961); Okutani (1966b) cited its oc- currence as Miocene. Calyptogena (Ectenagena) modioliforma (Boss) Ectenagena modioliforma Boss, 1968: 742-746, figs.10, 21-24, 26-27 (type-locality, Pillsbury station 394, 9”28.6'N; 76°26.3'W, Golfo del Darien, 66 miles NNE of Punta Caribana, Colombia in 42-641 m; holotype, MCZ 256973). Remarks. As noted earlier in the text, we no longer consider Ectenagena of generic rank and follow Keen (1969) in placing it as a sub- genus of Calyptogena; С. modioliforma is the Caribbean homolog of C. elongata and a close living relative of C. magnifica. Range. Known only from the holotype. 188 BOSS AND TURNER Calyptogena moraiensis (Suzuki) Unio moraiensis Suzuki, 1941: 55, pl. 4, figs. 2-5 (type-locality Morai hard shale for- mation in Pliocene); Hanzawa, Asano & Takai, 1961: 293 (Holotype and paratype, Oil- well no. 2, Shunbetsu, Ishikari-machi, Atsuta- gun [sic], Hokkaido, Morai hard shale, Plio- cene, Geological Institute, University о! Tokyo). N Calyptogena moraiensis (Suzuki) Otatume, 1942: 437 (Morai, Atuta-mura, Atuta-gun, Ishikari Province, Mio-Pliocene); Okutani, 1962: 23; Okutani, 1966b: 301. Remarks. This is recognized as a synonym of Japanese fossils of Mio-Pliocene age from Hokkaido referred to С. pacifica Dall (Otatume, 1942; Okutani, 1962; 19665). Range. Mio-Pliocene of Morai shale, Hokkaido, Japan. Calyptogena (?Calyptogena) nipponica Oinomikado & Kanehara Calyptogena nipponica Oinomikado & Kanehara, 1938: 677-678, pl. 21, figs. 1-5 (type-locality, Ushigakubi Bed, Lower Plio- cene, Higashiyama Oil-field, Niigata Pre- fecture; range: Kubiki Bed, Niigata Prefecture [Mio-Pliocene]; Katsuura Bed, Chiba Pre- fecture [Pliocene]; Otatume, 1942: 197 + 199; Hatai & Nisiyama, 1952: 33 (Ushigakubi Low- er Pliocene [Miocene]. On the eastern bank of the Maekawa, about 1.2 km $ of the vil- lage office at Nakamura, about 200 т W of the shrine at Shigebuko, Nishitani-mura, Koshi-gun, Niigata. Nagaoka, 37°24.30'N; 138°59'E; holotype and paratype destroyed; also Kubiki Miocene. The wall of the water well, 33 m deep from surface at the primary school, Nakanosawa, Higashiyama-mura, Koshi-gun, М; [Kojiya, 37°19'N; 138°52’Е] paratype destroyed (is Akebiconcha cf. А. kawamurai Kuroda); Okutani, 1957: 218; ltoigawa, 1958: 251 (Teradomari Formation Kubiki Group, Upper Miocene, Higashiyama Oil-field, Niigata Prefecture); Hanzawa, Asano & Takai, 1961: 219, holotype and para- type, eastern bank of the river [Mae-kawa], about 1.2km south of the village office of Nishitani-mura at Nakamura, Nishitani-mura, Koshi-gun, Niigata Prefecture, Ushigakubi Formation, Pliocene; Okutani, 1962: 23 as Akebiconcha? nipponica (Oinomikado and Kanehara); Okutani, 1966b: 301, Pliocene. Remarks. Oinomikado & Kanehara (1938) established this species on specimens from several Japanese localities of Miocene and Pliocene age. It is quite probable that, as indi- cated by Hatai & Nisiyama (1952), C. kawamurai Kuroda, 1943 and C. k. elongata Ozaki, 1958 are synonyms of C. nipponica. We suggest that even C. soyoae may be considered in this lineage and might prove to be synonymous if critical comparable suites were available. Apparently, there are conflicting opinions concerning the primary type-material of C. nipponica. Hatai & Nisiyama (1952) noted that the holotype & paratype were destroyed, while Hanzawa, Asano & Takai (1961) cited both a holotype & paratype! Range. Fossil species from the Miocene and Pliocene of Japan in Ushigakubi Bed, Kubiki Bed, and Teradomari Formation of the Kubiki Group in Niigata Prefecture and from the Katsuura Bed, Chiba Prefecture. Calyptogena (Calyptogena) pacifica Dall Calyptogena pacifica Dall, 1891: 190 (type-locality, Albatross Station 3077, off Dixon Entrance, Alaska, in 322 fathoms; 1895b: 713, pl. 25, figs. 4, 5 (holotype, USNM 122549); 1903a: 700, 712 (Pliocene and Recent); 1903b: 1435-1436 (Pliocene of Los Angeles, California; Recent, in Clarence Strait, southeastern Alaska); 1916: 408; 1921: 32 (Clarence Strait, Alaska to Santa Barbara Channel, California); Lamy, 1922: 349; Oldroyd, 1924: 116; Crickmay, 1929: 93; Grant € Gale, 1931: 278-279, pl. 13, figs. 1За- b (Pliocene, Los Angeles; blown out of big Amalgamated Gas well, from depth of 2500 feet with Dendraster interlineatus Stimpson, Wolfskill Lease, Salt Lake oilfield, near Bever- ly Hills, Los Angeles County (coll. by J. O. Lewis, 1912); Thiele, 1935: 848; Otuka, 1937: 231 (Wakimoto Bed, Pliocene of Oga Penin- sula, Akita-ken, Japan); Oinomikado & Kanehara, 1938: 678; Woodring, 1938: 50-52 (in Clarence Strait, Alaska, in 322 fms.; off Santa Cruz in 506-680 fms.; off Santa Rosa Island in 30-41 fms.); Otatume, 1942: 198, pl. 16, figs. 1-12 (Morai Bed of Ishikari Oil-field, Hokkaido, Mio-Pliocene); Okutani, 1957: 28; Okutani, 1962: 23; Okutani, 1966b: 301, pl. 27, figs. 1, 3; Boss, 1968: 739, figs. 16, 17, 19, 20; Keen, 1969: N664, fig. E 138, 11a, b; Boss, 1970: 70; Oberling & Boss, 1970: 82; Kanno, 1971: 81; Habe, 1977: 237. Calyptogena (Calyptogena) pacifica Dall. Bernard, 1974: 11, text figs. 1A, 2A, 3A, 4A-D (anatomy; localties off British Columbia, Canada, and northern California). CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 189 Remarks. This, the type-species of Calyp- togena, is probably the best known, most widely distributed species of the genus. Ber- nard (1974) studied its anatomy and showed that there are but few minor features to distin- guish it from closely related species. Unio moraiensis Suzuki was a nomen used for populations in the Morai Beds of Japan which have been referred to C. pacífica (Otatume, 1942; Okutani, 1962). Certain fos- sil species in the Americas might well fall into the synonymy of C. pacifica or at least consti- tute a portion of its geological lineage, namely, C. gibbera Crickmay and C. panamensis Olsson. The species might even be cosmo- politan since С. valdiviae Thiele 8 Jaeckel really is hardly distinguishable except for its provenance, off West and South Africa. Range. According to published reports, this species is known as a fossil from Mio-Plio- cene times in the Morai Bed of Ishikari Oil- field, Hokkaido, Japan (Otatume, 1942), in the Pliocene of the Oga Peninsula, Akita Pre- fecture, Japan, as well as in Los Angeles, California (Grant & Gale, 1931). Living sam- ples are known from Clarence Strait, Alaska to off southern California in 55-1244 m. Calyptogena (?Calyptogena) panamensis Olsson Calyptogena panamensis Olsson, 1942: 33(185), pl. 2 (15), figs. 2, 3 (type-locality, Punta Burica [sandstone], Costa Rica and Panama; lower Pliocene and uppermost Miocene). Remarks. From the internal view of the left valve (Olsson, 1942: pl. 2 (15), fig. 2), the dental configuration of С. panamensis sug- gests a relationship to C. pacifica and thus to the subgenus Calyptogena rather than Ectenagena. The samples of C. panamensis appear to be an in situ thanatocenosis and are associated with other sometimes deeper water, offshore forms (e.g. Solemya, Thyasira). Range. Fossil in lower Pliocene and upper- most Miocene of Punta Burica, Costa Rica and Panama (Pacific Coast). Calyptogena (Calyptogena) ponderosa Boss Calyptogena ponderosa Boss, 1968: 737- 742, figs. 9, 11-15, 18 (type-locality, M/V Oregon | station 1426, 29°07'N; 87 54'W, about 77 miles south of Mobile Bay, Gulf of Mexico, in 600 fathoms (1,097 m); additional localities, Pillsbury station 364, 9°28.7'N; 76°34.3'W, Golfo del Darien, 63 miles NNE of Punta Caribana, Colombia, in 933-961 т. Pillsbury station 394, 9°28.6'N; 76°26.3’W, Golfo del Darien, 66 miles NNE of Punta Caribana, Colombia, in 421-641 m. Pillsbury station 391, 10°03.0’N; 76°27.0'W, Golfo del Darien, 69 miles SSW of Cartagena, Colom- bia in 1417-1767 m) Oberling & Boss, 1970: 81-90, 2 figs. 1 pl.; Boss, 1970: 70. Remarks. As previously indicated (Boss, 1968), C. ponderosa, though similar to C. pacifica, has not only a thicker, heavier shell but a strong posterior radial ridge; additionally the dentitions differ. Range. Living in the Caribbean Sea and Gulf of Mexico, in 421-1767 m. Calyptogena (Calyptogena) soyoae Okutani Calyptogena soyoae Okutani, 1957: 27, pl. 1, figs. 1, 4, 5a—b (type-locality, 6 miles WSW of Jógashima [islet], eastern part of Sagami Bay, Honshú, at 750m muddy bottom, 35°05.1'N; 139°33.3’E); Habe, 1961: 122, pl. 55, fig. 17; Okutani, 1962: 22, pl. 2, figs. 10— 11, pl. 4, fig. 14 (in 710-770 т, Sagami Bay); Habe, 1968: 179, pl. 55, fig. 17 (750 т in Sagami Bay and off Bósó Peninsula, Honshú); Boss, 1970: 70; Bernard, 1974: 18; Habe, 1977: 237 (in 750-1000 m, Sagami Bay). Akebiconcha soyoae (Okutani). Okutani, 1966a: 11, pl. 1, fig. 7 (in 1005-1020 m, Sagami Bay); Okutani, 1966b: 300, pl. 28, figs. 1-2. Remarks. Okutani (1957) introduced this name for specimens between 50 and 140 mm in length taken by the R/V Soyo-Maru in Sagami Bay at a depth of 750 m; he remarked on its similarity to C. pacifica, C. elongata, and C. nipponica and also noted a relation- ship with Akebiconcha kawamurai as well but suggested that the mesial constriction of the valves was diagnostic for C. soyoae. In 1962, he presented a key to the species, C. soyoae, C. nipponica, C. elongata and C. pacífica, wherein one of the diagnostic features was size, now known to be highly variable. He (1966a, b) transferred C. soyoae to Akebi- concha and placed the genus in the Arctici- dae (as Cyprinidae). However, Akebiconcha seems to be indistinguishable from Calypto- gena (Boss, 1968) and the genus does not show familial affinities with the Arcticidae (see Remarks in text on the genus Calyptogena). 190 BOSS AND TURNER Were large series of C. soyoae and C. nipponica available, it is quite possible that C. soyoae would prove to be synonymous with C. nipponica (and its synonyms, C. kawa- murai and С. К. elongata Ozaki, see Наа! 8 Nisiyama, 1952). Range. Living in Sagami Bay, Honshú, be- tween 750-1020 m. Calyptogena (Calyptogena) valdiviae (Thiele & Jaeckel) Vesicomya valdiviae Thiele & Jaeckel, 1931: 229 (71), pl. 9 (4), fig. 101 (Valdivia stations 33 and 103). Calyptogena valdiviae (Thiele & Jaeckel). Boss, 1968: 742; Boss, 1970: 69-70, figs. 3, 4, 22, 23 (type-locality designated, Valdivia station 33, 24°35.3’N; 17°4.7'W, about 140 miles (23 km) off Morro Garnet, Rio de Oro, West Africa, in 2500 m; additional locality, Valdivia Station 103, 35°10.5'$; 23°2’E, about 72 miles (120 km) south of Knysna, Republic of South Africa, just off Agulhas Bank, in 500 m). Remarks. As noted in Boss (1968: 742; 1970: 70), С. valdiviae closely resembles С. pacifica and may be separable on such fea- tures as minor differences in outline (C. valdiviae has a more convex ventral margin) and sculpture (C. valdiviae has less weakly expressed growth lines and concentric lira- tions). Range. Living (?). From off west Africa to off south Africa in 500-2500 m. ACKNOWLEDGEMENTS For receipt of specimens collected off the Galapagos during ALVIN Cruise 90 in 1977 we are grateful to J. Corliss, Chief Scientist, and for specimens collected in the same area during ALVIN Cruise 102, leg 9, in 1979, we thank С. Berg and F. Grassle, Chief Scientist; R. Ballard, Chief Scientist on leg 10 of the 1979 cruise kindly supplied pictures and data on populations seen around vents on the East Pacific Rise off Mexico during ALVIN dives 915 and 917. We are grateful to J. Rosewater for the loan of type specimens and collections from the National Museum of Natural History (USNM). The 1977 expedition was supported by a National Science Foundation grant to Oregon State University. The 1979 expedition was supported by NSF Grant No. OCE-08855 to Harvard University. Special thanks are due W. Woodring and D. Wilson for help in finding specimens of fossil Calyptogena in the USNM and to C. M. Yonge and an anonymous reader who criti- cized the text. This article is contribution Number 2 of the Galapagos Rift Biology Expedition. LITERATURE CITED ALLEN, J. A., 1958, On the basic form and adapta- tions to habitat in the Lucinacea (Eulamelli- branchia). Philosophical Transactions of the Royal Society of London, ser. В, 241: 421—484. ALLEN, J. A., 1968, The functional morphology of Crassinella mactracea (Linsley) (Bivalvia: Astartacea). Proceedings of the Malacological Society of London, 38: 27-40. AOKI, S., 1954, Mollusca from the Miocene Kabeya Formation. Jóban coal-field, Fukushima Pre- fecture, Japan. Science Reports of the Tokyo Kyoiku Daigaku (formerly Tokyo Bunrika Daigaku), Tokyo University of Education, Geo- logical and Mineralogical Institute, Sec. C., 3(17): 2341, 2 pl. BALLARD, В. D. & GRASSLE, J. F., 1979, Return to oases of the deep. National Geographic, 156(5): 689-705. BERNARD, Е. R., 1974, The genus Calyptogena in British Columbia with a description of a new spe- cies (Bivalvia, Vesicomyidae). Venus, 33: 11-22. BOLTZMANN, H., 1906, Beiträge zur Kenntnis der Perikardialdrüse der Lamellibranchiaten. Arbeiten aus dem Zoologischen Instituten der Universität Wien, 16: 313-324, 1 pl. BOSS, K. J., 1967, A new species of Vesicomya from the Caribbean Sea (Mollusca: Bivalvia: Vesicomyidae). Breviora, No. 266, 6 p. BOSS, К. J., 1968, Biological Investigations of the Deep Sea, 40. New species of Vesicomyidae from the Gulf of Darien, Caribbean Sea (Bivalvia; Mollusca). Bulletin of Marine Science, 18: 731- 748. BOSS, K. J., 1969a, Systematics of the Vesicomy- idae (Mollusca: Bivalvia). Malacologia, 9: 254- 255. BOSS, К. J., 1969b, Nomenclatorial note: Arctic- idae versus Cyprinidae (Mollusca; Bivalvia). Nautilus, 83: 61-64. BOSS, К. J., 1970, Redescription of the Valdivia Vesicomya of Thiele and Jaeckel. Mitteilungen aus dem Zoologischen Museum in Berlin, 46: 67-84. BOSS, K. J. 8 KENK, V. C., 1964, Anatomy and relationships of Temnoconcha brasiliana Dall. Occasional Papers on Mollusks, Harvard Uni- versity, 2: 325-343, pl. 55-60. BULLOCK, Т. & HORRIDGE, A., 1965, The struc- ture and function of the nervous system in in- vertebrates. Freeman, San Francisco, 2 vols., xx + 1719 p. CLAUSEN, C., 1958, On the anatomy and history CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 191 of the Eulamellibranch Kelliella miliaris (Philippi) with observations on the ciliary mechanisms in the mantle cavity. Nytt Magasin for Zoologi, 6: 144-175, 5 pl. CORLISS, J. В. 8 BALLARD, В. D., 1977, Oases of life in the cold abyss. National Geographic, 152(4): 441-453. CORLISS, J. В. et al., 1979, Submarine thermal springs on the Galápagos Rift. Science, 203 (4385): 1073-1083. COSSMANN, M., 1920, [Review.] Revue Critique de Paléozoologie et de Paléophytologie, 24: 29. COX, L. R., 1969, Superfamily Solemyacea H. Adams 8 A. Adams, /n Moore, R. C., ed., Treatise on Invertebrate Paleontology, Part N, vol. 1, Mollusca 6, Bivalvia: 241-243. CRICKMAY, C. H., 1929, On a new pelecypod, Calyptogena gibbera. Canadian Field-Naturalist, 43: 93, 1 fig. DALL, W. H., 1886, Reports on the results of dredg- ing, under the supervision of Alexander Agassiz, in the Gulf of Mexico (1877-78) and in the Carib- bean Sea (1879-80), by the U.S. Coast Survey steamer “Blake,” Lieut.-Commander С. D. Sigsbee, U.S.N. and Commander J. В. Bartlett, U.S.N. commanding. XXIX. Report on the Mol- lusca, Pan 1, Brachiopoda and Pelecypoda. Bulletin of the Museum of Comparative Zoology, 12: 171-318, pl. 1-9. DALL, W. H., 1891, On some new or interesting West American shells obtained from the dredg- ings of the U.S. Fish Commission steamer “Alba- tross” in 1888, and from other sources. Proceed- ings of the United States National Museum, 14(849): 173-191, pl. 5-7. DALL, W. H., 1895a, Contributions to the Tertiary fauna of Florida, with especial reference to the Miocene silex-beds of Tampa and the Pliocene beds of the Caloosahatchie River. Tertiary mol- lusks of Florida. Part Ill. A new classification of the Pelecypoda. Transactions of the Wagner Free Institute of Science of Philadelphia, 3: 485— 570. DALL, W. H., 1895b, Scientific results of explora- tions by the U.S. Fish Commission steamer “Albatross.” No. XXXIV. Report on Mollusca and Brachiopoda dredged in deep water, chiefly near the Hawaiian Islands, with illustrations of hitherto unfigured species from northwest America. Proceedings of the United States National Museum, 17: 675-733, pl. 23-32. DALL, W. H., 1903a, Synopsis of the Carditacea and of the American species. Proceedings of the Academy of Natural Sciences of Philadelphia, 54: 696-716. DALL, W. H., 1903b, Contributions of the Tertiary fauna of Florida with especial reference to the silex-beds of Tampa and the Pliocene beds of the Caloosahatchie River, including in many cases a complete revision of the generic groups treated of and their American Tertiary species. Part VI. Concluding the work. Transactions of the Wagner Free Institute of Science of Philadel- phia, 3:1219-1654. pl. 48-60. DALL, W. H., 1908, The Mollusca and the Brachi- opoda. Bulletin of the Museum of Comparative Zoology, 43: 415419, pl. 6, 16, 17, 18. DALL, W. H., 1916, Diagnoses of new species of marine bivalve mollusks from the northwest coast of America in the United States National Museum. Proceedings of the United States Na- tional Museum, 52(2183): 393—417. DALL, W. H., 1921, Summary of the marine shell- bearing mollusks of the northwest coast of Amer- ica, from San Diego, California, to the Polar Sea, mostly contained in the collection of the United States National Museum, with illustrations of hitherto unfigured species. Bulletin of the United States National Museum, 112: 1-217, pl. 1-22. DALL, W. H. & SIMPSON, С. T., 1901, The Mol- lusca of Porto Rico. U.S. Fisheries Commission Bulletin, 20: 351-524, pl. 53-58. DREDGING AND HYDROGRAPHIC RECORDS of the U.S. Fisheries Steamer Albatross for 1904 and 1905, 1906, Bureau of Fisheries Document по. 604, Washington: 1-80. DUVERNOY, M., 1853, Mémoires sur le systeme nerveux des Mollusques Acephales. Mémoires Académie des Sciences, Paris, 24: 3-312. GRANT, U. S. & GALE, Н. R., 1931, Catalogue of the marine Pliocene and Pleistocene Mollusca of California and adjacent regions. Memoirs of the San Diego Society of Natural History, 1: 1036 p., 32 pl. GRASSLE, J. R., BERG, С. J., CHILDRESS, J. J., GRASSLE, J. P., HESSLER, В. R., JANNASCH, Н. J., KARL, D. M., LUTZ, В. A., MICKEL, T. J., RHOADS, D. C., SANDERS, H. L., SMITH, К. L., SOMERO, С. N., TURNER, В. D., TUTTLE, J. H., WALSH, Р. J. & WILLIAMS, А. J., 1979, Galápagos ‘79: initial findings of a deep-sea bio- logical quest. Oceanus, 22(2): 2-10. HABE, T., 1951-1953, Lyonsiidae, Poromyidae, Arcticidae and Gaimardiidae in Japan. /llustrated catalogue of Japanese shells, 21: 153-160, pl. 22. HABE, T., 1961, Coloured illustrations of те shells of Japan. Hoikusha, Osaka, vol. 2: 133 p. HABE, T., 1968, Shells of the Western Pacific т color. Vol. Il. Revised edition. Hoikusha, Osaka, 233 p. HABE, T., 1977, Systematics of Mollusca in Japan. Bivalvia and Scaphopoda. Tokyo, 372 p. HANZAWA, S., ASANO, K. 8 TAKAI, F., 1961, Catalogue of type-specimens of fossils in Japan. Paleontological Society of Japan, 25th Anniver- sary Volume, 422 p. HARRY, H. W., 1966, Studies on bivalve molluscs of the genus Crassinella in the Northwestern Gulf of Mexico: anatomy, ecology and system- atics. Publication of the Institute of Marine Sci- ence, University of Texas, 11: 65-89. HATAI, К. 4 NISIYAMA, S., 1952, Check list of Japanese Tertiary marine Mollusca. Science Reports of the Tóhoku University, Sendai, Japan, ser. 2, special vol. 3: 464 p. ITOIGAWA, J., 1958, Molluscan fossils from the Niitsu, Higashiyama and Takezawa oil-fields, 192 BOSS AND TURNER Niigata Prefecture, Japan. Memoirs of the Col- lege of Science, University of Kyoto, ser. B 24(4): 249-263, 2 pl. KAMADA, Y., 1962, Tertiary marine Mollusca from the Jóban coal-field, Japan. Paleontological Society of Japan, Special Paper, 8: 187 p. 21 pl. KANEHARA, K., 1937, Miocene shells from the Jóban coal-field. Bulletin of the Imperial Geo- logical Survey of Japan, 27(1): 1-21, 5 pl. KANNO, S., 1967, Mollusks of the Itsukaichi Basin. Shibaba Memorial Volume: 401-402, pl. 1, figs. 9-11, 15. KANNO, S., 1971, Tertiary molluscan fauna from the Yakataga District and adjacent areas of southern Alaska. Palaeontological Society of Japan, Special Paper, 16, 154 p., 18 pl. КАММО, S. & OGAWA, H., 1964, Molluscan fauna from the Momijiyama and Takinoue districts, Hokkaido, Japan. Science Reports of the Tokyo Kyoiku Daigaku [Tokyo University of Education], Section C, Geology, Mineralogy and Geogra- phy, 8(81): 269-294, 4 pl. KEEN, A. M., 1969, Family Vesicomyidae Dall, 1908 In MOORE, R. C., ed., Treatise on Inver- tebrate Paleontology, Part М, vol. 2, Mollusca 6, Bivalvia: 663-664. KEEN, A. М., 1977a, A deep-water paradox. Ап- nual Reports of the Western Society of Mala- cologists, 10: 10. KEEN, А. М., 1977b, New sea-floor oasis. Veliger, 20: 179-180. KURODA, T., 1931, Fossil Mollusca In HOMMA, Е. Geology of central part of Shinano Province (Shinano Chubu Chishitsu-Shi). Part 4, 90 p., 1-13. KURODA, T., 1943, Akebiconcha, a new pelecy- pod genus. Venus, Japanese Journal of Mala- cology, 13: 14-18. LAMMENS, J. J., 1969, The morphology of the central nervous system of Macoma balthica (L.). Archives Néerlandaises de Zoologie, 19: 105-127. LAMY, E., 1920, Révision des Cypricardiacea et des Isocardiacea vivants du Muséum d'Histoire Naturelle de Paris. Journal de Conchyliologie, 64: 259-307. LAMY, E., 1922, Révision des Carditacea vivants du Muséum National d'Histoire Naturelle de Paris. Journal de Conchyliologie, 66: 218-276, 12 figs., pl. 8; 289-368. LONSDALE, P., 1977, Clustering of suspension- feeding macrobenthos near abyssal hydro- thermal vents at oceanic spreading centers. Deep-Sea Research, 24: 857-863. MAKIYAMA, J., 1958, Matajiro Yokoyama's Terti- ary fossils from various localities in Japan. Paleontological Society of Japan, Special Paper 4, 6 p., pl. 25-57. OBERLING, J. J. 8 BOSS, К. J., 1970, Observa- tions on the shell structure of Calyptogena (Vesicomyidae; Bivalvia; Mollusca). Revue Suisse de Zoologie, 77: 81-90, 1 pl. ODHNER, N. H., 1912, Morphologische und phylogenetische Untersuchungen Uber die Nephridien der Lamellibranchien. Zeitschrift fur wissenschaflichen Zoologie, 100: 287-391. ODHNER, М. H., 1960, Mollusca. Report for the Swedish Deep-sea Expedition, 2 (Zoology 22): 365—400, 2 pl. OINOMIKADO, T. & KANEHARA, K., 1938, A new species of Calyptogena from the Higashiyama oilfield, Niigata-ken, Japanese Journal of the Geological Society of Japan, 45(539): 677-678, ВР OKUTANI, T., 1957, Two new species of bivalves from the deep water in Sagami Bay collected by the R. V. Soyo-Maru. Bulletin of the Tokai Re- gional Fisheries Research Laboratory, 17: 27-30, pl. 1. OKUTANI, T., 1962, Report on the archibenthal and abyssal lamellibranchiate Mollusca mainly collected from Sagami Bay and adjacent waters by the R.V. Soyo-Maru during the years 1955- 1960. Bulletin of the Тока! Fisheries Research Laboratories, no. 32: 1-40, 5 pl. OKUTANI, T., 1966a, Archibenthal and abyssal Mollusca collected by the R.V. Soyo-Maru from Japanese waters during 1964. Bulletin of the Tokai Regional Fisheries Research Laboratory, 46, 32 p., 2 pl. OKUTANI, T., 1966b, Identity of Calyptogena and Akebiconcha (Bivalvia, Cyprinidae). Venus, 24: 297-303, pl. 27-28 (English summary). OLDROYD, |. S., 1924, Marine shells of the Puget Sound and vicinity. Publications of the Puget Sound Geological Station, University of Wash- ington, Seattle, 4: 1-272, 49 pl. OLSSON, А. A., 1931, Contributions to the Tertiary Paleontology of Northern Peru: Part 4. The Peruvian Oligocene. Bulletins of American Paleontology, 17(63): 97-264 (1-168), pl. 13-33 (1-21). OLSSON, A. A., 1942, Tertiary and Quaternary fossils from the Burica Peninsula of Panama and Costa Rica. Bulletins of American Paleontology, _ 27(106): 1-82, pl. i-xii. OTATUME, K., 1942, On the оссштепсе of fossil Calyptogena from the Isikari Oil-field, Hokkaido. Journal of the Geological Society of Japan, 49(590): 435—437, pl. 16. OTUKA, Y., 1937, Occurrence of Calyptogena pacifica Dall from Wakimoto on Oga Peninsula. Journal of the Geological Society of Japan, 44(522): 231, 1 fig. OWEN, G., 1953, On the biology of Glossus humanus (L.) (Isocardia cor Lam.). Journal of the Marine Biological Association of the United Kingdom, 32: 85-106. OZAKI, H., 1958, Stratigraphical and paleontologi- cal studies on the Neogene and Pleistocene formation of the Tyoshi District. Bulletin of the National Science Museum of Tokyo, 42: 1-182, pl. 1-24. PELSENEER, P., 1891, Contributions à l'étude des CALYPTOGENA MAGNIFICA FROM GALAPAGOS RIFT 193 lamellibranches. Archives de Biologie, Paris, 11: 147-312, pl. 6-23. PELSENEER, P., 1911, Lamellibranches de l'expedition du Siboga. Partie anatomique. Siboga Expeditie Leiden, Monographie 53a, 125 p., 26 pl. PURCHON, R. D., 1939, Reduction of the ctenidia in the Lamellibranchia. Nature, 144: 206. RIDEWOOD, W. G., 1903, On the structure of the gills of the Lamellibranchia. Philosophical Transactions of the Royal Society of London, ser. B, 195: 147-284. SALEUDDIN, A. S. M., 1964, Observations on the habit and functional anatomy of Cyprina islandica (L.). Proceedings of the Malacological Society of London, 36: 149-162. SALEUDDIN, А. S. M., 1965, The mode of life and functional anatomy of Astarte spp. (Eulamelli- branchia). Proceedings of the Malacological Society of London, 36: 229-257. SEMPER, K., 1880, Die natúrlichen Existenzbed- ingungen der Thiere. Bd. 1. SHIKAMA, T., 1962, On some noteworthy shells from off Choshi, Chiba Prefecture. Science Reports of the Yokohama National University, Sec. Il, 8: 29-56, pl. 1-3. SMITH, E. A., 1885, Report on the Lamellibranchi- ata collected by H.M.S. Challenger. Report of the Scientific Results. Voyage of H.M.S. Chal- lenger. Zoology, 13: 1-341, 25 pl. STUARDO, J. R., 1968, On the phylogeny, taxon- omy and distribution of the Limidae (Mollusca: Bivalvia). Ph.D. thesis, Harvard University, i-ii, 1-327, 27 pl., 24 maps. SUZUKI, K., 1941, Three new species of non- marine shells from the Tertiary formation of Hokkaido and Karahuto. Japanese Journal of Geology and Geography, 18: 53-58, pl. 4. TAKEDA, H., 1953, The Poronai Formation (Oligocene Tertiary) of Hokkaido and South Sakhalin and its fossil fauna. Studies on Coal Geology. No. 3. Geological Section. The Hokkaido Association of Coal Mining Technolo- gists, Sapporo, Japan, p. 85. TANAKA, K. [Y?], 1959, Molluscan fossils from Central Shinano. Nagano Prefecture Japan (Рай 1). Fossils from Akanuda limestone. Journal Shinshu University, 8: 115-127, pl. 1-3. TAYLOR, J. D., KENNEDY, W. J. & HALL, A, 1973, The shell structure and mineralogy of the Bivalvia: Il: Lucinacea-Clavagellacea. Conclu- sions. Bulletin of the British Museum of Natural History, Zoology, 22: 253-294, 15 pl., 33 figs., 22 tables. THIELE, J., 1886, Die Mundlappen der Lamelli- branchiaten. Inaugural-dissertation. Leipzig. Wilhelm Engelmann. 34 p., 2 pl. Zeitschrift fur wissenschaftliche Zoologie, 45: 239-272. THIELE, J., 1934-1935, Handbuch der system- atischen Weichtierkunde. Jena, Fischer, 2: 779- 1022. THIELE, J. & JAECKEL, S., 1931, Muscheln der deutschen Tiefsee-Expedition. Wissenschaf- liches Ergebnisse deutschen Tiefsee Expedi- tion, 21: 161-268 (1-100), pl. 6-10 (1-5). TUREKIAN, К. K., COCHRAN, J. К. 8 NOZAKI, Y., 1979, Growth rate of a clam from the Galapagos Rise hot spring field using natural radionuclide ratios. Nature, 280: 385-387. TURNER, R. D., 1966, A survey and illustrated catalogue of the Teredinidae. Museum of Com- parative Zoology, Cambridge, 265 p., 64 pl. WALLER, T. R., 1978, Morphology, morphoclines and a new classification of the Pteriomorphia (Mollusca: Bivalvia). Philosophical Transactions of the Royal Society of London, ser. В, 284: 345- 365, 2 figs. WHITE, K. M., 1942, The pericardial cavity and the periocardial gland of the Lamellibranchia. Proceedings of the Malacological Society of London, 25: 37-88. WINKLE, K. VAN, 1919, Remarks on some new species from Trinidad. Bulletins of American Paleontology, 8(33): 19-27, pl. 3. WINTERER, Е. L. & DURHAM, D. L., 1962, Geol- ogy of southeastern Ventura Basin, Los Angeles County, California. United States Geological Survey Professional Paper 334-H: 275-366, 6 pl. WOODRING, W. P., 1938, Lower Pliocene mol- lusks and echinoids from the Los Angeles Basin, California. United States Department of the In- terior, Geological Survey Professional Paper 190: 67 p., 9 pl. WOODRING, W. P., BRAMLETTE, М. М. & КЕМ, W. S. W., 1946, Geology and paleontology of Palos Verdes Hills, California. United States Geological Survey Professional Paper 207: 145 pora pl: YOKOYAMA, M., 1925, Tertiary Mollusca from Shinano and Echigo. Journal of the Faculty of Science of the Imperial University of Tokyo, Sec. Il, 1(1): 1-22, pl. 1-7. YONGE, C. M., 1957, Mantle fusion in the Lamelli- branchia. Pubblicazioni della Stazione Zoologi- ca di Napoli, 29: 151-171. YONGE, C. M., 1962, On the primitive significance of the byssus in the Bivalvia and its effects in evolution. Journal of the Marine Biological As- sociation of the United Kingdom, 42: 113-125. YONGE, C. M., 1969, Functional morphology and evolution within the Carditacea. Proceedings of the Malacological Society of London, 38: 493- 527. YONGE, С. М., 1971, On functional morphology and adaptive radiation in the bivalve superfamily Saxicavacea (Hiatella (= Saxicava), Saxicavella, Panomya, Panope, Cyrtodaria). Malacologia, 11: 1—4. YONGE, С. M., 1978a, Significance of the ligament in the classification of the Bivalvia. Proceedings of the Royal Society of London, ser. В, 202: 231- 248. YONGE, C. M., 1978b, On the Dimyidae (Mollusca: 194 BOSS AND TURNER Bivalvia) with special reference to Dimya cor- ZATSEPIN, V. |. & FILATOVA, Z. A., 1961, The rugata Hedley and Basiliomya goreaui Bayer. bivalve mollusc Cyprina islandica (L.), its geo- Journal of Molluscan Studies, 44: 357-375. graphic distribution and role in the communities YONGE, С. M., GOREAU, Т. Е. & GOREAU, N. I., of benthic fauna. Transactions of the Institute of 1972, On the mode of boring in Fungiacava Oceanology, Academy of Sciences, USSR, 46: eilatensis (Bivalvia: Mytilidae). Journal of the 201-216. Zoological Society of London, 166: 55-60, 2 pl. | 1 я a $ $ on Г ps y р у 7 Rae | . . EU >> | 7 A ee у 7 u a 29 à В = CT e u в = й ha AR: : Tres рысь ns un rodaja 9 A O 4 | AUR CAR МУЖ Dn Ne Br « MNT AAC ых А epee) lien Ce и E A eran ye ae аа vo oy ча ME + Pi re alfa paced Dre у A a hn rela a Ce ni |. © a ANA Le и tre oe му ML уно» ую Eur E IAS e» ut ja Кв off a.) 1 In fl My n ACL yer Га A Bore u u J 3% Qe } = ae Ad Te й wei ad | LA mye a A A = ap: ql es? m=: <, CS Ш = 2 Ue: eL LE Y Ta ANR ve e | A | РФ у De e ns a ÚS va” in = Me ye YO га 1.0 0 i 112 Die у НЫ : 7 = La oe В ty I р а We Y | DAA 4) 7 u + . V nm “ - o f e on ay - u = oo. o A a e = 7 Dé a — o = o = ru u м ry O u = a 5 = = EJ 1! u = o > te | o PS us > i О o . u = у = o > Y | y О o e a = o “ ® o u as = » u 7 4 } Prone Van — 4 = u - т 19(2) INSTRUCTIONS FOR AUTHORS MALACOLOGIA publishes original studies on the Mollusca that are of international in- terest and are of high scholarly standards. Both descriptive and experimental research results are acceptable provided they are pri- marily or exclusively concerned with the phy- lum. Contributions include long monographs as well as moderately short research papers. Brief papers are not acceptable. MALACO- LOGIA provides a forum for such different as- pects of malacology as anatomy, comparative physiology, ecology, medical malacology, paleontology and systematics. Papers of only biochemical or physiological interest should be submitted elsewhere. Review articles are more appropriately submitted to Malaco- logical Review (P.O. Box 801, Whitmore Lake, Michigan 48189, U.S.A.). All manu- scripts submitted are reviewed by at least 2 malacologists. Articles are accepted with the firm understanding that they have not been submitted or published elsewhere in whole or in part. Manuscripts may be in English, French, German or Spanish, and should follow MALA- COLOGIA style. They must contain a concise but informative Abstract summarizing not only the content but the results. Papers in lan- guages other than English should include a translation of the Abstract into English. Au- thors desiring their abstracts translated into other languages must provide these. Care should be taken to include all necessary for- eign accents. Manuscripts must be typed on one side of good quality white paper, double- spaced throughout, with ample margins, and are to be submitted in triplicate. Illustrations are likewise to be in triplicate (the 2 copies may be photocopies, etc.). Tables, figure cap- tions and all footnotes are to be grouped (in this order) at the end of a manuscript, and all Ms pages (including the Abstract) are to be numbered sequentially. Avoid internal page references (which have to be added in page proof). Make the hierarchy of headings within the text simple and consistent. Suggest an abbreviated running title to be used at the top of each right hand page. Contributors in English are asked to use the Council of Biology Editors (CBE) Style Man- ual (Ed. 3, 1972), obtainable for $6.00 from the American Institute of Biological Sciences, MALACOLOGIA 1980 1401 Wilson Boulevard, Arlington, Virginia 22209, U.S.A. MALACOLOGIA follows most of the recommendations in this Manual. In particular, simplified particles such as the fol- lowing are used: numbers above ten should not be written out except at the beginning of a sentence; percentages following a number are expressed as %, and abbreviations of measures (after a number): mm, ml, kg, etc. have no period (full stop), nor an “s” in the plural. Note that the international symbol for micron is now um, not u. Illustrations must be carefully prepared and so planned that they can be printed in 1 column or the full width of a page of the jour- nal. The maximum size of a printed figure is 13.5 x 20.0 cm (preferably not as high as this so that the caption does not have to be on the opposite page). Drawings and lettering must be in dark black on white, blue tracing, or blue-lined paper. Lines and dots should be thick enough to allow reduction by Y2 or Y. This should be taken into consideration also in relation to the lettering. Letters and num- bers must not be less than 2 mm in height, preferably larger, after reduction. Several drawings or photographs may be grouped together to fit a page, but drawings are not to be grouped with photographs. Photographs are to be glossy and high contrast. All illustra- tions are to be numbered sequentially as fig- ures (not grouped as plates), and are to be arranged as closely as possible to the order in which they are first cited in the text (Each fig- ure must be cited in the text.) All original illus- trations should be mounted, numbered, la- beled or lettered and ready for the engraver. Scale lines are required for all figures and should be convenient lengths (e.g., “200 um,” not “163 ит”). Magnifications in captions are not acceptable, and neither are photographic reductions of line drawings. Captions should summarize what is shown in an illustration, and should not duplicate additional information given in the text. Each lettered abbreviation labeling an individual feature in a figure must either be explained in each caption (listed alphabetically), or be grouped in one alphabetic sequence in a sec- tion near the beginning of the text (use the latter method if many abbreviations are re- peated on different figures). Tables are to be used sparingly, and should be planned to fit 1 or 2 columns on 1 раде. Each table must be submitted double- spaced throughout on a separate manuscript page. Do not use vertical lines. All References cited in the text must be list- ed (bibliographies including uncited items are unacceptable). Each reference should be cited accurately (the Editors will spot check for accuracy) and should be in the style used in recent issues of MALACOLOGIA—except that beginning with Vol. 16 journal titles will be cited complete and unabbreviated. For all manuscripts submitted henceforth, disregard the abbreviations in MALACOLOGIA, 1972, 11(2): 415-426. The journal uses the amper- sand (8) for “and”; “et al.” may be used in the text, but not in the References. In addition to citations but should relate by author, date and sulted by other will also require that voucher specimens from — other kinds of research be deposited in mu- о Es) o a =. кан N ИЯ 3 Na | AN EA A & A \ о = y e Vol. 20, No. 1 SAT MALAGOLOGIA "IS CONTENTS F. G. THOMPSON Proserpinoid land snails and their relationships within the Archaeo- — $ oido e 0 Whe cea A O О SENN DEP > N. COLEMAN and W. CUFF The abundance, distribution and diversity of the molluscs of Western = Ports Victorias Alstalla dle e NG FA S. M. CHAMBERS Genetic divergence between populations of Goniobasis (Pleuro- ceridae) occupying different drainage systems ................. eee В. T. DILLON, Jr. and G. M. DAVIS $ The Goniobasis of southern Virginia and northwestern North Caro- | lina: genetic and shell morphometric relationships .................. A L. G. WILLIAMS Development and feeding of larvae of the nudibranch gastropods = Е Hermissenda crassicornis and Aeolidia papillosa ............... reat 2 В. $. HOUBRICK Observations оп the anatomy and life history of Modulus modulus (Prosobranthia: Modulidae)i леч RES ne C. M. YONGE On Patro australis with comparisons of structure throughout the 5 Añomiidae (Bivalvia)... igs ae 33 CA UE EN 2 Nee A 4 x‘ R. M. LINSLEY and M. JAVIDPOUR Episodic growth: in Gastropad >. ол AR у К. J. BOSS and В. D. TURNER The giant white clam from the Galapagos Rift, Calyptogena mag- RTE nifiCa SPECIES OVA he LASER RE RS Po N MOTOS > С. J. BERG and В. D. TURNER A APPENEDOMN AN. e SE O eE 18: bi K. J. BOSS Appendix 2 ии da cd A eae a pros RS PEER LE M MUS. COMP. ZOOL. _ 1981 LIBRARY JUN 1 91981 HARVARD UNIVERSITY \LACOLOGIA Revista Internacional de Malacologia Journal International de Malacologie N № Международный Журнал Малакологии Internationale Malakologische Zeitschrift 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, А. R. Egypt Editorial Assistants: MARY DUNN СВЕТСНЕМ В. EICHHOLTZ 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: J FRANCES ALLEN, Emerita Environmental Protection Agency Washington, D.C. CHRISTOPHER J. BAYNE, President-Elect Oregon State University, Corvallis ELMER G. BERRY, Emeritus Germantown, Maryland KENNETH J. BOSS Museum of Comparative Zoólogy Cambridge, Massachusetts JOHN B. BURCH MELBOURNE R. CARRIKER University of Delaware, Lewes GEORGE M. DAVIS, Executive Secretary-Treasurer ROBERT ROBERTSON CLYDE Е. E. ROPER, President Smithsonian Institution Washington, D.C. W. D. RUSSELL-HUNTER Syracuse University, New York NORMAN F. SOHL United States Geological Survey Washington, D.C. RUTH D. TURNER, Alternate Museum of Comparative Zoólogy Cambridge, Massachusetts SHI-KUEI WU, Vice-President University of Colorado Museum, Boulder Institute meetings are held the first Friday in December each year at a convenient place. For information, address the President. Copyright, © Institute of Malacology, 1981 1981 EDITORIAL BOARD J. A. ALLEN Marine Biological Station, Millport, United Kingdom E. E. BINDER Muséum d'Histoire Naturelle Genève, Switzerland А. Н. CLARKE, Jr. National Museum of Natural History Washington, D.C., U.S.A. E. S. DEMIAN Ain Shams University Cairo, A. R. Egypt С. 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 V. FRETTER University of Reading United Kingdom E. GITTENBERGER Rijksmuseum van Natuurlijke Historie Leiden, Netherlands A. N. GOLIKOV Zoological Institute Leningrad, U.S.S.R. A. V. GROSSU Universitatea Bucuresti Romania T. HABE National Science Museum Tokyo, Japan A. D. HARRISON University of Waterloo Ontario, Canada K. HATAI Tohoku University Sendai, Japan B. HUBENDICK Naturhistoriska Museet Góteborg, Sweden A. M. KEEN Stanford University California, U.S.A. В. М. KILBURN Natal Museum Pietermaritzburg, South Africa M. A. KLAPPENBACH Museo Nacional de Historia Natural Montevideo, Uruguay J. KNUDSEN Zoologisk Institut £ Museum Kóbenhavn, Denmark A. J. KOHN University of Washington Seattle, U.S.A. Y. KONDO Bernice P. Bishop Museum Honolulu, Hawaii, U.S.A. C. M. LALLI McGill University Montreal, Canada J. LEVER Amsterdam, Netherlands A. LUCAS Faculté des Sciences Brest, France N. MACAROVICI Universitatea “Al. |. Cuza” lasi, Romania C. MEIER-BROOK Tropenmedizinisches Institut Túbingen, Germany (Federal Republic) H. K. MIENIS Hebrew University of Jerusalem Israel J. Е. MORTON The University Auckland, New Zealand R. NATARAJAN Marine Biological Station Porto Novo, India J. OKLAND University of Oslo Norway T. OKUTANI National Science Museum Tokyo, Japan W. L. PARAENSE Universidade de Brasilia Brazil J. J. PARODIZ Carnegie Museum Pittsburgh, U.S.A. C. M. PATTERSON University of Michigan Ann Arbor, U.S.A. W. F. PONDER Australian Museum Sydney A. W. B. POWELL Auckland Institute & Museum New Zealand R. D. PURCHON Chelsea College of Science & Technology London, United Kingdom О. ВАУЕНА Еигают Ispra, Italy N. W. RUNHAM University College of North Wales Bangor, United Kingdom S. G. SEGERSTRALE Institute of Marine Research Helsinki, Finland G. A. SOLEM Field Museum of Natural History Chicago, U.S.A. Е. STARMÜHLNER 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. TOFEOLETTO Societa Malacologica Italiana Milano W. $. $. 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, 1981, 20(2): 195-204 ETUDE DE L'EFFET DU “GROUPEMENT” DES INDIVIDUS CHEZ THEBA PISANA (MOLLUSQUE GASTEROPODE PULMONE STYLOMMATOPHORE) Maria Lazaridou-Dimitriadou! et Jacques Daguzan Laboratoire de Zoologie Générale et d'Ecophysiologie, U.E.R. des Sciences Biologiques, Avenue du Général Leclerc, 35042 Rennes Cedex, France RESUME Des expériences sont réalisées, au laboratoire, afin de préciser l’action du ‘groupement des individus” de Theba pisana (Müller) (Gastéropode Pulmoné dunicole) nés au laboratoire, sur leur croissance, leur reproduction et leur mortalité. Plus le nombre d'individus groupés est important, plus le taux de croissance est faible et plus la mortalité est élevée. Par contre, il s'avère qu'une densité animale est nécessaire pour que cette espèce puisse se reproduire normalement. INTRODUCTION ET OBJET DU TRAVAIL Nous savons depuis longtemps, qu'à cer- taines périodes de l'année, les escargots, sur- tout dunicoles, stamassent sous forme de “grappes” plus ou moins importantes. Ces re- groupements temporaires et périodiques peuvent être provoqués par certains facteurs tels que l'élévation de la temperature du sol ou la force du vent (Astre, 1921: Engel, 1957). De plus, à ces facteurs abiotiques peuvent s'adjoindre des facteurs d'ordre “social” (Le Masne, 1952: Bigot, 1967). Ces groupements sont considérés comme primitifs et inco- ordonnes, et si une inter-attraction existe, il n'y a pas de coordination ou d'organisation reelle (Le Masne, 1952). Il ne s’agit probable- ment pas de ‘simples foules” liées à des fac- teurs physicochimiques indépendants des in- dividus rassemblés dont le groupement dure autant que s'exerce l'influence extérieure (Rabaud, 1927, 1929). Au sein de ces groupes, il semble exister une attraction mutuelle qui serait due à des stimuli tactiles et chimiques, et qui se manifesterait après la formation de l'agrégat d'individus, rendant ainsi ce dernier plus dense. Ce groupement a certainement un rôle de protection vis-à-vis des conditions climatiques défavorables (Allee, 1928). Plusieurs auteurs ont déjà étudié pour les Pulmonés aquatiques (Wright, 1960; Eisen- berg, 1966, 1970; Mooij-Vogelaar et col. Laboratoire de Zoologie, Université de Thessaloniki, Grèce. 2Synonyme: Euparypha pisana. 1970, 1973; Chevalier et col., 1974; Thomas & Benjamin, 1974... .) et les Pulmones ter- restres (Wolda, 1963; Yom-Tov, 1972; Chevallier, 1974; Williamson et col., 1976; Oosterhoff, 1977... .) les conséquences que peut entrainer le “groupement” des individus sur leur croissance, leur productivité et leur mortalité. Enfin, Bigot (1967) étudie les ras- semblements animaux de Theba pisana (Müller)? (Gasteropode Pulmoné) aux de- partements des Bouches-du-Rhöne, du Gard et du Vaucluse, mais il ne s'occupe pas des effets du “groupement” sur la croissance, la productivité et la mortalité de cette espèce. Sur les dunes de Penvins, situées sur le littoral atlantique агтопсат (Morbihan, France), on remarque qu'à certaines périodes de l’année, les individus de Theba pisana se présentent temporairement sous forme de groupes plus ou moins denses. L'étude de la distribution spatiale de Треба pisana au cours de l’année présente une répartition de type “еп agrégats” (indice de Taylor: о? = 2.29(x)'45)3 et o2 > x et 1 < b — +. Sion considère séparément les valeurs de la vari- ance relative (A2 = o2/X),3 on remarque qu'elles sont plus accentuées pendant la période d'accouplement (par exemple, A? = 7.5 le 18 juillet 1977; ce jour-là le nombre maximal d'animaux qui étaient sur une grappe, était de 36) et les journées sèches et ventées pendant lesquelles les animaux se ramassent en grappe sur des plantes abritées (par ex- Ou o* = la variance, X = le moyen et À? = la variance relative. (195) 196 LAZARIDOU-DIMITRIADOU ЕТ DAGUZAN emple, À2 = 11.2 le 6 avril 1978; ce jour-là le nombre maximal d'animaux mesuré sur une grappe était de 52) (Lazaridou-Dimitriadou, 1978). | De plus, sur une zone ди cordon dunaire а végétation clairsemée, le nombre moyen d'animaux par metre carré variait de 1 a 5 et de 5 á 9 sur une autre zone a végétation plus dense, sur le méme cordon et pendant la période mai 1977-mai 1978. Pendant la même période mais dans l'arriere-pays, où changeait considérablement la pédologie et ой la végétation était tres dense, le nombre moyen d'animaux par т? variait de 110 а 1200. Зиг ce dernier biotope la répartition de Theba pisana était toujours de type “еп agre- gats” (par exemple, le À? = 18.8 le mai 1977; а ce jour-lá le nombre maximal d'animaux, qui étaient sur une grappe, était de 115). Enfin, on constate que le nombre d'oeufs pondus de Theba pisana dans l’arriere-pays variait de 40 à 70 et de 10 à 30 sur le cordon dunaire. Pitchford (1954) pense que le nombre d'oeufs pondus par Theba pisana varie de 40 à 50 et il note que Taylor compte environs 60 oeufs, mais déposés en trois moments différents. A la suite de ces constatations il nous a paru intéressant d'examiner dans des condi- tions expérimentales précises si le “groupe- ment” des individus de Theba pisana observé sur le terrain a une action sur la croissance et la productivité de cette espece. PROTOCOLE EXPERIMENTALE Cette étude est réalisée au laboratoire, sur des animaux marqués (а l'aide d'une pastille de plastique, numérotée et collée sur la co- quille), ayant le méme áge et élevées а partir de pontes ramenées des dunes de Penvins. Les escargots sont élevés ensemble et nourris avec de la laitue, pendant deux mois. Puis, on sélectionne parmi eux des individus ayant sensiblement la même taille (même valeur du grand diamètre de la coquille). Dans des enceintes de plexiglass4 (18 x 12 x 7.5cm) contenant du sable sur une hauteur de 3 cm, nous plaçons des animaux à des densités croissantes (progression géométriques): isolés, groupés par 2, 4 ou 8 (Tableau 1). De cette manière, si les effets du groupement des individus sont des ‘effets de masse’ (Grassé, 1946, 1968) leur intensité augmentera d'un type de groupement а un autre. La température varie entre 12° et 22°C, selon la saison, et l'humidité relative journa- lière oscille entre 50 et 95% selon l'heure d'arrosage. Les escargots sont nourris avec de la laitue, lavée à l'eau, fournie en excès et dont la quantité est proportionnelle au nombre d'individus par boîte. De plus, l'apport de calcaire se fait sous forme de maërl lave, broyé, séché à 105°C et disposé dans deux des coins opposés des enceintes d'élevage. Tous les 2 ou 3 jours, la nourriture est changée, les fèces retirées et chaque boîte d'élevage nettoyée, selon les normes établies par Herzberg (1965). De même, tous les 15 jours, chaque animal est pesé à l'aide d'une balance “Mettler” au 1/100 de gramme et sa coquille est mesurée (grand diamètre de la coquille (GD), et diamètre de son ouverture) à l'aide d'un pied à coulisse au 1/10 de mm. Enfin, chaque jour, on note les évènements essentiels qui peuvent se manifester: ac- couplement, ponte, éclosion, mortalité.5 RESULTATS Chaque mois on teste ГПотодепейе des mesures obtenues pour des “types de groupement” (1, 2, 4 ou 8) grâce à l'analyse de la variance. Les différences enregistrées entre les divers essais réalisés avec un même “type de groupement” n'étant pas significa- TABLEAU 1. Caractéristiques des expériences réalisées dans les diverses enceintes. Nbre. des enceintes Nbre. d'individus Espace disponible par Nbre. d'individus d'élevage utilisées par enceinte individu (en cm) par m2 8 8 202 370 5 4 405 185 5 2 810 93 5 1 1620 46 ELEFANT ar зы a SS E E IRL ILE RE SD EE EE 4Le couvercle des enceintes est fait d'une toile de nylon à mailles carrés de 1 mm de côté. SAvec la collaboration technique de Mlle. Marie-Madeleine Le Piver, aide biologiste C.N.R.S. 197 “GROUPEMENT” ОЕ ТНЕВА PISANA = 82761 JOlAUEF 01 000! vel 0821 TE > — 8461 JSIAUEF 9 00'0 + 00'OL OO + 5911 84 0 + SZ'Z! = — = 1461 219898 9 00'0 + 00°01 r00 + 991 8/0 + SL! 0€'6 €0'L + 00/1 1461 81QUSAON 91 LEO + OF OL +0 0 + 6/1 УГО + 55'81 ¿50 + 0/6 60'0 + FIL Le O + Sc Zk 1461 9.99120 95 LEO + OF OL 800 + b0'¢ НЕО + 55'81 ¿S0 + 0/6 vO'0 + Stl 200 = 96'91 1461 91400 с 15`0 + 05'01 ОНО + PEL УГО + 55'81 6C 0 + 5/`6 90'0 + StL +0 0 + 00/1 1161 asquajdas pz LEO + OF OL 80'0 + 781 УЕО + 55'81 Le O + Oc OL Le O + SI 1 9/ 0 + EC 81 216, aique]des 9 Sc'0 + OV OL 60 0 + 96'L 8c 0 + 99°81 ce 0 + OF OL 810 + SLL c9 0 + 95'81 1461 MOV cc ECO + be OL |1 `0 + 061 ce'0 + Oral Sc'0 + BL OL © 0 Zt 050 + 9081 4461 MOV 9 co 0 + OL OL LL'O + 561 ce DO + 85'81 vz0 + EL OL 91'0 + 6811 09`0 + cc 8l 1461 JOIIING ce Sc'0 +086 ЕГО + S8'L 8£ 0 + OZ ZI 8c 0 + v6'6 ДНО + 981 ISO + €E8 21 1461 JOIN Z LEO + cv'6 6r0 + cSt 59`0 + 09'91 930 + 95'6 GL'O + EDL Gt 0 + 2/91 1461 unf cc 8t0+ct8 ctO + G60 Ze 0 + 80+1 cc0++tt8 60 0 + 860 С7`0 + vb vl 1461 uImP 9 90 0 + c6 + 100 + LLO 200 + 86'Z 600 + L/'p LO'0 + OL'O СО = 78`1 1461 IBN 9 $э10$! зприирш Sap $э4поле) snpinipuı г эр Sednoin) Ly'O + 6bv'6 [0 + tel S9'0 + /t 91 9€'0 + 08'8 ct 0 + 160 bvO'l + 06 rl 1461 MOV cc 0C 0 + 79'6 ОНО = $91 55`0 + Bl Zt Sc'0 + 8/8 80 0 + 001 vty'O + pr Ss! 1461 nov 9 [10 + 99'6 0:0 + 091 55`0 + 70`11 110 + 66'8 60'0 = Zab 16`0 + 64'S! 1461 JOIING сс 910+ 796 01`0 + 9911 veo + LZ OL 91'0 + 138 800 + lek GEO + vr Sl 1461 Inf Z 6LO+cCLE6 600 + 651 cet O = 26 St! УС. 0 + 078 800 + 80'L €v0 + pS pl 1461 unf cc УЕО + 078 90 0 + 16'0 0c'0 + cE VL ECO + c9/ 900 + 040 4£ 0 + 94 21 1461 ump 9 400 + p4'p LO'O + LLO OLO + €8Z 90'0 + 6£'p 10`0 + 60 0 400 + bY'Z 1461 е/ 9 Ws + ww ue p Ws + 16 ue 4 Ws + uu ua gd Ws = ww ua p US Био Us + uu ue q) sejeq SNPIAIPUI y эр sadnoJe) при!ри 8 ap Sadnoin) ‘euesId eqay/ 2эцэ ээие$$10л лпа| Ins SNPIAIPUI sap ‚диэшедпо.б эр sadÁ),, зиэлашр Sep ээцэпции зиешезиоо хпеао!6 sjeynsey ‘г NVAIGWL 198 LAZARIDOU-DIMITRIADOU ЕТ DAGUZAN 0.05), il est alors possible de regrouper les résultats concernant chaque groupement. Dans la série relative au groupement “type 4 individus” et “type 2,” nous ne tenons pas compte d'une enceinte expérimentale, dans laquelle, il y a eu apparition de microorga- nismes indéterminés. Influence du groupement des individus sur leur croissance. Evolution du grand diametre de la coquille (GD) et du poids total (P) de l'animal pendant la durée de l'expérience Grâce aux nombreuses mesures ef- fectuées, il est possible de suivre les évolu- tions du grand diamètre de la coquille (Tableau 2) et du poids total de chaque escargot pour chaque “type de groupement” (Tableau 2). On constate que la taille maxi- male est d'autant plus réduite et la vitesse avec laquelle cette taille est atteinte d'autant plus diminuée que l'effectif est plus important. En plus, les moyennes du grand diamètre de la coquille des individus obtenus, re- spectivement pour le groupement “type 8” au-delà du 22 juillet 1977 et pour le groupe- ment “type 4” au-delà du 6 août 1977 peu- vent аррагайге а priori comme aberrantes, mais sont tout simplement dues au fait que les individus qui vivent le plus longtemps sont aussi les plus petits et que par conséquent la moyenne devient alors beaucoup plus faible (Tableau 2). Le méme phénomene s'observe en ce qui concerne l'évolution du poids total moyen des individus (Tableau 2), mais il est accentué par l'existence de la ponte (types 2, 4) ou de la mortalité constatée pour le groupement “type 8.” De plus, nous consta- tons qu'au début du mois d'aoút, les individus isolés perdent du poids, phénomène proba- blement dû à une diminution de l’activité et de la consommation alimentaire des animaux, le moment que ces individus pondent beaucoup plus tard. Afin de voir s’il existe une différence signifi- cative entre les taux de croissance relative du grand diamètre de la coquille (GD) en fonc- tion du diamètre de son ouverture (d), des individus appartenant aux divers types de groupement, on utilise la loi d'allométrie simple d'Huxley et de Teissier (1936) préci- nisée en particulier pour les Gastéropodes Prosobranches par Daguzan (1975) et dont la formule en coordonnées logarithmiques est: log GD = a (log d) + logb (ou a, b: constantes). Cette droite ainsi obtenue, ou axe majeur réduit, a une pente (a) qui représente le taux de croissance relative du GD en fonction du d. | est possible de comparer éventuellement les pentes des droites d'allométrie obtenues en admettant que: difa > 30 difa : la différence entre les pentes est très significa- tive (seuil à 99%) : différence significative (seuil à 95%) : différence non significa- tive 2difa < difa < 30 dita difa < 20 difa (Où difa = |aj—aji|; одна = (0a)? + (вар? et ба» Taj: les écart-types des a de chaque formule). Tout d'abord, on constate que quel que soit le type de groupement des escargots, il existe une tres bonne corrélation (r) entre le GD et le d (Tableau 3); on note aussi une allométrie majorante de la croissance du grand diamètre de la coquille par rapport à celle du diamètre de l'ouverture car a > 1 (Tableau 3). Enfin, il existe une différence significative entre les pentes des axes majeurs réduits des indi- vidus appartenant au ‘groupement type 8” et TABLEAU 3. Principaux paramètres de la croissance relative du grand diamètre de la coquille (GD) en fonction du diamètre (d) de l'ouverture chez Theba pisana; (еп coordonees logarithmiques). a, b = constantes; г = coefficient de correlation; а, GD = moyens; N = effectif; « = écart-type, correspondant au coefficient de sécurité 68.3%. Aaa OA gg DIE AC IE ES Groupement Groupement Groupement Animaux Paramètres “types 8” “types 4” “types 2” isolés а +d, 1.032 + 0.016 1.086 + 0.016 1.090 + 0.022 1.133 + 0.042 logb + 0.208 = 0.014 OMG = 0:015 0.159 = 0.020 0.117 = 0.040 r 0.985 0.986 0.986 0.977 1099 + ciogd 0.874 + 0.123 0.918 + 0.114 0.940 + 0.120 0.944 + 0.112 11109 0127 1.158 + 0.124 1.184 + 0.131 1.187 == 01127 М 134 126 68 35 amg PEN I EEE “GROUPEMENT” ОЕ ТНЕВА PISANA 199 TABLEAU 4. Comparaison des droites d'allométrie représentant la croissance chez Theba pisana. (С: groupement; Чиа: |aj—ajl et одна: У (0a)? + (a;)°) Comparaison des pentes des droites Différence entre Ecart-type de la différence entre Signification de la d'allométrie les pentes (difa) les pentes (0 gifa) différence Gg - G, 0.101 0.045 Significative (au seuil de 95%) Gg 65 0.058 0.027 Significative (au seuil de 95%) Gg — Ga 0.054 0.023 Significative (au seuil de 95%) Gr Gi 0.047 0.045 Non significative G4 — Go 0.004 0.028 Non significative Go - G;y 0.043 0.047 Non significative celle des escargots des autres “types de groupement” (Tableau 4). De plus, en compa- rant les valeurs de a on observe que plus l'effectif est important et plus la vitesse de croissance est faible (Tableau 3). Etude de la variabilité de la croissance ob- servée chez les individus appartenant au sein d'un même “type de groupement” Au sein d'un groupement donné, on note qu'un certain nombre d'individus présente une taille beaucoup plus petite que les autres. Ainsi, pour chaque type de groupement étudié, on calcule le “coefficient de retard de croissance” en effectuant la différence entre le grand diamètre moyen (GD), de la coquille des individus les plus grands appartenant à ce groupement et le grand diamètre moyen (GD), de la coquille des plus petits escargots rencontrés également dans ce type de groupement (Fig. 1). De cette étude, on constate que plus le nombre d'individus groupés est important et plus le “coefficient de retard de croissance” est grand. De plus, quel que soit le type de groupement, ce “coefficient de retard de croissance” augmente en fonction de Гаде des individus, passe par un maximum, puis diminue par le fait que la croissance devient pratiquement nulle chez les adultes et que les individus en retard de croissance continuent a s'accroítre. Influence du “groupement” des individus sur leur reproduction Des observations effectuées régulièrement permettent d'établir, pour chaque type de groupement, un tableau regroupant les va- leurs moyennes des événements essentiels concernant la reproduction des individus (ponte, éclosion, mort etc.) (Tableau 5). Tout d'abord, on remarque que quelquefois les escargots s'accouplent deux fois avant de pondre. Dans certains cas, Ну a des animaux qui pondent deux fois; souvent le nombre total des oeufs de deux pontes dépasse légère- ment celui d'une seule ponte importante. L'effet du groupement semble avoir une action sur la ponte et sa précocité. Pour le groupement “type 8 individus” aucune ponte n'est observée. Les premières pontes ont lieu chez les animaux du groupement “type 4,” mais la date moyenne de la première ponte de tous les escargots utilisés montre que ceux du groupement “type 2” pondent neuf jours avant les individus du “type 4” et 68 jours avant ceux du “type isolés” (Tableau 5). Bien que le grand diamètre moyen des indivi- dus au moment de la première ponte soit d'autant plus petit que l'effectif est plus im- portant, le test de l'analyse de leurs variances ne présente pas une différence significative (P > 0.05) (Tableau 5). Les pontes qui par la suite éclosent, représentent l'ensemble du groupement “type 4” (100%), 50% du “type 2” et 0% du “type isolés” (Tableau 5). Les individus isolés pondent au maximum 8 oeufs, dont le poids n’a pas été détermine afin d'éviter de les détruire, bien que Wolda & Kreulen (1973) signalent que les manipula- tions des oeufs n'entraínent aucune perturba- tion sur leur développement et leur éclosion. En plus, des coupes histologiques effectuées sur une de leurs pontes montrent que les oeufs sont stériles; quant aux autres pontes, elles se desséchent dans les mois suivants. Quant aux pontes des autres types de groupe- 200 LAZARIDOU-DIMITRIADOU ЕТ DAGUZAN coefficient de retard de croissance {уре4 ео wok a et N, 31 48 63 18 9 jours passes du début de l'expérience FIG. 1. Importance du ‘coefficient de retard de croissance” du grand diamètre de la coquille selon le type de “groupement des individus” chez Theba pisana (Müller). individus vivants еп 4 100 БК sa jours passés du début de l'experience Fig. 2. Courbes de survie des individus de Theba pisana (Müller) en fonction du “type de leur groupement.” “GROUPEMENT” ОЕ ТНЕВА PISANA 201 TABLEAU 5. Principaux résultats concernant l'influence de différents “types de groupement” sur la repro- duction et la mort de Theba pisana. CR I A A A о, A A Test de l'analyse de la variance entre Groupement Groupement les différents Paramètres “type 4” “type 2” Animaux isolés “types de groupement” Date moyenne de la première ponte (en jours) + écart-type 29 Juillet + 14 20 Juillet + 11 8 Octobre + 28 P < 0.01 Grand diamètre moyen (en mm) + écart-type IAE 17.8 15 18/2120 0 P > 0.05 Nombre moyen d'oeufs + écart-type 78.5' = 19.4: 38:5 + 23:9 ЗБ Р < 0.01 Poids moyen d'oeufs (en g) + écart-type 0.58 = 0.17 0.37 = 0.04 — Р < 0.05 Durée moyenne entre ponte—éclosion (en jours) + écart-type ИО 72 270 7.6 — P > 0.05 Pourcentage des pontes qui éclosent. 100% + 0% 50% + 5% 0% + 0% Р < 0.05 Durée тоуеппе entre ponte-mort des animaux (en jours) + écart-type 2352 10.7 31:3 = (6:3 20/0167 P > 0.05 ment, le nombre moyen et le poids moyen d'oeufs pondus sont d'autant plus grands que l'effectif est plus important (Tableau 5). Plus la ponte est tardive et plus la durée de vie normale de l'animal est prolongée. Dans се cas les animaux atteignent un grand diamètre plus important. Enfin, que la ponte soit précoce ou tardive, la durée de vie de l'animal, de la ponte à la mort, ne présente pas une différence significative (P > 0.05) selon le test de l'analyse de la variance (Tableau 5). Influence du groupement des individus sur leur mortalité Les courbes de survie des individus en fonction de leur type de groupement (Fig. 2) montrent que la mortalité est particulièrement précoce et important chez les groupes de 8 individus, alors qu'elle est très faible pour les individus isolés. DISCUSSION Tout d'abord il faut noter que les densités utilisés au laboratoire présent une situation moyenne entre les densités animales trouvées sur le cordon dunaire d’une part, et dans l’arrière-pays d'autre part. On constate que le groupement a un effet doublement inhibiteur sur la croissance de Theba pisana: d'une part en réduisant la taille maximale et d'autre рай en abaissant la vitesse avec laquelle cette taille est atteinte. Cette relation inversée (taille-densité) co- incide avec celle observée chez la même espèce par Bigot (1967) aux départements des Bouches-du-Rhône, du Gard et du Vaucluse. Le même phénomène est signalé chez Aeolidiella alderi par Chevalier et col. (1974) et chez Cepaea nemoralis par Wolda (1969) et Williamson et col. (1976). Cette superiorité de la taille chez les animaux isolés ou appartenant aux faibles rassemblements de Трера pisana, est expliquée par Bigot (1967) comme “. . . un simple effet de compe- tition alimentaire (diminution de la ration à cause du gonflement de la population) ou d'une diminution du temps d'alimentation (la séparation des individus en grappe est moins rapide que celle d'individus isolés). Il est peu probable qu'il faille incriminer l’action défavo- rable d'un effet de groupe (Chauvin, 1952). De plus, Eisenberg (1970) a montré chez Lymnaea elodes que la relation inversée entre la densité animale et la taille des indi- vidus disparaît avec l’additionnement de nourriture. Au laboratoire, comme toutes les enceintes 202 d'élévage contiennent la тёте quantité de Ca, sous forme de máerl—le manque de Ca peut freiner la croissance (Thomas et col., 1975) —et comme la nourriture est toujours fournie en excès et qu'elle est en plus de la méme qualité pour un jour donné dans toutes les enceintes d'élévage, on peut suggérer que c'est soit Гезрасе disponible (Tableau 1) soit les interactions chimiques ou comporte- mentales (Williamson et col., 1976) soit les deux qui entrainent ces differences de croissance au sein de divers “types de groupements.” On note, en plus, que plus la densité est forte et plus la mortalité est précoce et im- portante. И est possible que dans l'espace disponible les catabolites rejetés par les escargots (par exemple urée) et par les mi- croorganismes associés avec eux ou les métabolites solubles des feuilles de laitue, d'une part inhibent la croissance et d'autre part entrainent la mortalité (Simpson et col., 1973; Thomas & Benjamin, 1974). Wright (1960) suggere que les phéromones en petites quantités peuvent étre avantageux pour les animaux, mais en concentrations importantes limitent la densité de la popula- tion. Mais, les résultats précedents ne permettent pas de déterminer par quel moyen le groupement intervient dans la croissance et la mortalité; c'est plutót une étude qui nous montre les effets du groupement et non pas les agents déterminants. Mais, d'apres ce qu'il est généralement admis, l'intensité s'accroissant avec la densité, on serait plutôt en présence d'un effet de masse (Grassé, 1946). A la suite, on note qu'au sein d'un groupe- ment donné, un certain nombre d'individus présentent une croissance beaucoup plus lente que les autres; on le présente avec un coefficient qu’on appelle “le coefficient de retard de croissance.” Plus l'effectif est im- portant et plus le coefficient de retard de croissance est grand. Le même phénomène est observé chez Biomphalaria glabrata par Thomas & Benjamin (1974) et chez Aeolidi- ella alderi par Chevalier et col. (1974). En tenant compte du protocole expérimental, nous ne pensons pas que le facteur limitant soit la quantité de la nourriture (toujours fournie en excès), mais probablement la qualité de la nourriture ingerée par chaque individu de chaque enceinte d'élévage. Се type de competition, dite “scramble” par Nicholson (1955), se caractérise par le fait que le “compétiteur” le plus efficace est celui LAZARIDOU-DIMITRIADOU ET DAGUZAN qui peut obtenir et utiliser la partie de la nour- riture qui a la valeur la plus nutritive. Ainsi, plus l'effectif est important, et plus difficile est l’accessibilite par les individus les moins effi- caces a la partie de la nourriture la plus nutri- tive. Eisenberg (1966) a d'ailleurs montré chez Lymnaea elodes que la qualité de nour- riture est nécessaire pour la croissance maxi- male. Enfin, on constate que le groupement des individus a une action sur la précocité de la ponte, le nombre d'oeufs et l'éclosion. Tout d'abord, il faut rappeler que les animaux du groupement “type 8” meurent sans pondre et que les animaux isolés pondent très peu d'oeufs stériles (l'autofecondation n'arrive pas chez cette espece) et beaucoup plus tard que les individus des autres “types de groupe ment.” Quant aux “types 2 ou 4,” nous consta- tons que les escargots du “type 4” ont au moment de la ponte un grand diamètre plus petit que celui des individus du “type 2,” que le nombre et le poids d'oeufs de leurs pontes sont plus grands et que tous leurs oeufs éclo- sent. La différence de taille des individus au moment de la ponte, constaté aussi chez Aeolidiella alderi par Chevalier et col. (1974), peut être expliquée par la différence de la vitesse de croissance que montrent les deux “types de groupement.” L'influence du groupement sur la productivité est montrée chez divers Pulmonés par Wolda (1963), Steen (1967), Steen et col. (1973), Mooij- Vogelaar & Steen (1973), Eisenberg (1970), mais dans leurs expériences c'est la quantité ou la qualité de la nourriture qui provoquent ces différences. Dans notre cas, nous pen- sons que ce sont les échanges tactiles ou chimiques (phéromones) qui les provoquent; il semblerais, d’ailleurs, que la densité opti- male au point de vue productivité est celle du groupement “type 4” (185 individus/m2). TRAVAUX CITES ALLEE, W. G., 1928, Studies in animal aggrega- tions: mass protection from fresh water for Procerodes, a marine turbellarian. Journal of Experimental Zoology, 50: 239318. ASTRE, G., 1921, Sur la nature physique de la cause qui, dans les dunes maritimes, régit l'ascension des Mollusques le long des tiges а faible feuillage. Bulletin de la Société d'Histoire naturelle de Toulouse, 49: 162-164. BIGOT, L., 1967, Recherches sur les groupements des Gastéropodes terrestres: la constitution des “grappes.” Vie et Milieu, 18(1C): 1-27. CHAUVIN, R., 1952, L'effet de groupe. Colloque “GROUPEMENT” ОЕ ТНЕВА PISANA 203 C.N.R.S., “Structure et Physiologie des Sociétés animales,” 34: 81-90, Editions C.N.R.S. CHEVALIER, F., MOCQUARD, J. P. & TARDY, J., 1974, Croissance et precocite d’Aeolidiella alderi (Cocks), (Mollusque Nudibranche): effets du groupement et de la température. Bulletin de la Société Zoologique de France, 99: 601-621. CHEVALLIER, H., 1974, Les grands Arion de France (Mollusca, Pulmonate): Taxonomie, biogéographie, écologie, polymorphisme. Croissance et cycle biologiques. These, Docto- rat de l'Université, Paris VI. DAGUZAN, J., 1975, Contribution à l'écologie des Littorinidae (Mollusques Gastéropodes Proso branches), recherches écophysiologiques chez quatre especes: Littorina neritoides (L.), Littorina saxatilis (Olivi), Littorina littorea (L.) et Littorina littoralis (L.) These, Docteur es-Sciences, Rennes, 400 рр. EISENBERG, В. M., 1966, The regulation of densi- ty in natural populations of the pond snail Lymnaea elodes. Ecology, 47: 890-906. EISENBERG, R. M., 1970, The role of food in the regulation of the pond snail Lymnaea elodes. Ecology, 51: 680-684. ENGEL, H., 1957, Okologisch-Faunistische Stu- dien im Rhóne-Delta, unter besonderer Berücksi- chtigung der Mollusken. Bonner Zoologische Beitraege, 8: 5-55. GRASSE, P. P., 1946, Sociétés animales et effet de groupe. Experientia, 2: 77-82. GRASSE, P. P., 1968, Introduction. Colloque C.N.R.S., “L'effet de groupe chez les animaux,” 173: 11-14, Editions C.N.R.S. HERZBERG, R., 1965, Crowding as a factor in growth and reproduction of Helix aspersa. Amer- ican Zoologist, 5: 254. HUXLEY, J. $. & TEISSIER, G., 1936, Termino- logie et notation dans la reproduction de la croissance relative. Comptes Rendus des Séances de la Société de Biologie et des ses filiales, 121: 934-936. LAZARIDOU-DIMITRIADOU, M., 1978, Contribu- tion а Гесорпузююде d'un Gastéropode Pulmoné dunicole: Euparypha pisana (Müller) du littoral armoricain. These. Docteur 3eme cycle, Rennes, 279 p. LE MASNE, G., 1952, Classification et caractéristi- ques des principaux types de groupements soci- aux réalisés chez les Invertébrés. Colloque C.N.R.S., “Structure et Physiologie des Sociétés animales,” 34: 19-70, Editions C.N.R.S. MOOIJ-VOGELAAR, J. W., JAGER, J. С. & VAN DER STEEN, W. J., 1970, The effect of density changes on the reproduction of the pond snail Lymnaea stagnalis (L.). Netherlands Journal of Zoology, 20: 279-288. MOOIJ-VOGELAAR, J. W. & VAN DER STEEN, W. J., 1973, Effects of density on feeding and growth in the pond snail Lymnaea stagnalis (L.). Proceedings of the Koninklijke Nederlandse Akademie van Wetenschappen, Series C, 76: 61-68. NICHOLSON, A. J., 1955, An outline of the dy- namics of animal population. Australian Journal of Zoology, 2: 9-65. OOSTERHOFF, L. M., 1977, Variation in growth rate as an ecological factor in the landsnail Cepaea nemoralis (L.). Netherlands Journal of Zoology, 27: 1-132. PITCHFORD, G. W., 1954, Eggs of Theba pisana. Journal of Conchology, 24: 22. RABAUD, E., 1927, Les passemblements des males d'Halictes. Bulletin biologique de la France et de la Belgique, 61: 163-185. RABAUD, E., 1929, Phénomene social et Societés animales. Bulletin biologique de la France et de la Belgique, 63: 377-398. SIMPSON, A. W., THOMAS, J. D. & TOWNSEND, C. R., 1973, Social behaviour in the freshwater pulmonate snail Biomphalaria glabrata (Say). Behavioural Biology, 9: 731-740. STEEN, W. J. VAN DER, 1967, The influence of environmental factors on the oviposition of Lymnaea stagnalis (L.) under laboratory condi- tions. Archives Néerlandaises de Zoologie, 17: 403-468. STEEN, W. J. VAN DER, JAEGER, J. C. & TIEMERSMA, D., 1973, The influence of food quality on feeding, reproduction and growth in the pond snail Lymnaea stagnalis (L.) with some methodological comments. Proceedings of the Koninklijke Nederlandse Akademie van Weten- schappen, Series C, 76: 47-60. THOMAS, J. D. & BENJAMIN, M., 1974, The ef- fects of population density on growth and repro- duction of Biomphalaria glabrata (Say) (Gas- teropoda: Pulmonata). Journal of Animal Ecol- оду, 43: 31-50. THOMAS, y. D., GOLDSWORTHY, С. J. & ARAM, В. H., 1975, Studies on the chemical ecology of snails: the effect of chemical conditioning by adult snails on the growth of juvenile snails. Journal of Animal Ecology, 44: 1-28. WILLIAMSON, P., CAMERON, В. А. D. & CAR- TER, M. A., 1976, Population density affecting adult shell size of Cepaea nemoralis (L.) Nature, 263: 496-497. WOLDA, H., 1963, Natural populations of the poly- morphic landsnail Cepaea nemoralis (L.). Ar- chives Néerlandaises de Zoologie, 15: 381-471. WOLDA, H., 1969, Stability of a steep cline in morph frequencies of the snail Cepaea nemoralis (L.). Journal of Animal Ecology, 38: 623-635. WOLDA, H. 4 KREULEN, D. A., 1973, Ecology of some experimental populations of the landsnail Cepaea nemoralis (L.). И Production and survival of eggs and juveniles. Netherlands Journal of Zoology, 23: 168-188. WRIGHT, C. A., 1960, The crowding phenomenon in laboratory colonies of freshwater snails. Annals of Tropical Medicine and Parasitology, 54: 244- 232. YOM-TOV, T., 1972, Field experiments on the effect of population density and slope direction on the reproduction of the desert snail Trochoidea (Xerocrassa) seetzeni. Journal of Animal Ecol- ogy, 41: 17-22. 204 LAZARIDOU-DIMITRIADOU ЕТ DAGUZAN ABSTRACT EFFECTS OF “CROWDING” ON GROWTH, MORTALITY RATE AND REPRO- DUCTION OF THEBA PISANA (MULLER) (GASTROPODA PULMONATA) Maria Lazaridou-Dimitriadou and Jacques Daguzan This article deals with the effect of “crowding” on growth, reproduction and mortality of Theba pisana (Múller) hatched in the laboratory. From a series of experiments, it is determined that the higher the number of grouped individuals the smaller the growth rate and the higher the mortality rate. On the other hand, it seems that a certain population density is necessary for the possibility of this species to reproduce efficiently. MALACOLOGIA, 1981, 20(2): 205-216 THE OSPHRADIAL COMPLEX OF TWO FRESHWATER BIVALVES: HISTOLOGICAL EVALUATION AND FUNCTIONAL CONTEXT Louise Russert Kraemer Department of Zoology, University of Arkansas, Fayetteville, Arkansas 72701, U.S.A. ABSTRACT From studies based largely on two species of freshwater bivalves, Lampsilis ventricosa ovata (Barnes), a unionacean, and Corbicula cf. С. fluminea (Müller), a sphaeriacean, osphradia are seen to be located within the roof of the exhalant chamber just under the visceral ganglion. In turn the visceral ganglion is positioned anteroventral to the posterior adductor muscle. Extensive serial section study of the osphradial tissues of C. fluminea reveals each osphradium to be comprised of (1) a characteristic epithelium of columnar epithelial cells which lack a basement membrane, but which are profusely innervated by neuronal fibers from (2) clusters of nerve fibers which parallel the base of the epithelium and also send fibers into the visceral ganglion, and (3) clusters of neuronal soma near the base of the epithelium, which send fibers into the visceral ganglion. The clusters of nerve fibers do not constitute a distinct osphradial nerve, nor do the neuronal soma constitute a discrete osphradial ganglion. Three-dimensional appearance of the osphradia is that of two narrow, pie-wedge-shaped organs located at right angles to the longitudinal axis of the animal. The narrowing tips of the organs approach but do not touch in the midline of the ventral surface of the visceral ganglion. Neuroanatomical context of the osphradia is examined in order to test some assumptions implicit in a current hypothesis that bivalve osphradia may function as particle-size sensors which serve to regulate activity of gill cilia. Neuroanatomically, the osphradia are closely associated with nerves which supply the posterior adductor muscle and with nerves which supply kidney tissue. The osphradia do not have a close anatomical association with branchial nerves or with the ganglionated roots of those nerves. A simple but repeated error may account for the inchoate state of literature on bivalve osphradia. А misrepresentation of osphradia in bivalves as inverted from their normal position may stem from an effort by early workers to homologize bivalve osphradia with those of gastro- pods. The error may be related to an incorrect assumption by some authors that bivalve osphra- dia are associated with the roof of the inhalant rather than the exhalant chamber. New hypothe- ses are needed concerning the function of bivalve osphradia. Hypotheses should incorporate the new information presented here concerning the precise anatomical site, histological organiza- tion, three-dimensional structure and neuroanatomical context of these paired organs. On the basis of findings presented here, it may be that bivalve osphradia function as light sensors which regulate seasonal behavior or reproductive physiology. Alternatively, it might be that bivalve osphradia function not only in control of fluid movement through the exhalant chamber, but also in adduction of the shell valves. INTRODUCTION A word taken from the Greek, “osphradion” for “strong scent,” has for years been applied to a single or paired, putative sense organ which occurs commonly among aquatic gas- tropods, near their visceral ganglion at the outer side of the gill (Lankester, 1883). Hyman (1967) reviewed findings of workers who noted that the gastropod osphradium takes varied form: as an elongate ridge or swelling, as a pitted swelling, as a raised patch of pigmented or unpigmented epithe- lium, aS a row of warts, or as resembling a (205) well differentiated, miniature bipectinate gill. Careful comparative histological study of gas- tropod osphradia has been done, especially by Demal (1955); and ultrastructure studies have been made by Anderson (1963), Simp- son (1971), Benjamin (1971), Crisp (1973), Alexander & Weldon (1975), and Newell & Brown (1977). Yonge (1947) analyzed structural and func- tional context of the osphradium in aspido- branch gastropods. Yonge opines (personal communication and 1947) that osphradia were originally tactile receptors which de- tected the amount of sediment in the water; 206 KRAEMER and that osphradia function as chemorecep- tors in carnivores like the neogastropods Buccinum and Nassarius. Physiological and behavioral studies of gastropod osphradia seem to be primarily those of Michelson (1960), Brown 8 Noble (1960), Kohn (1961), Bailey 8 Laverack (1963), Carr (1967), Jahan-Parwar et al. (1968), Bailey 8 Benja- min (1968), Stinnakre 8 Tauc (1969), von Baumgarten et al. (1968), Jahan-Parwar et al. (1969), Townsend (1973), Phillips (1975), Катагат (1976), and Sokolov & Kamardin (1977). In bivalved mollusks neither the anatomical, histological nor physiological attributes of the osphradium seem to be clearly understood. Many authors (e.g. Yonge, 1947; Bayne et al. 1976) speculate that the osphradium may not be homologous with that of a gastropod. Es- sentially histological studies of Freidenfelt (1904), Dakin (1910, 1928) and Stork (1934) provide limited information on the organiza- tion of the bivalve osphradium. Some confu- sion in the literature is mirrored in the fact that Bullock & Horridge (1965) have reprinted Dakin’s (1910) histological drawing of Ensis, and seem unaware that it is labelled just as Dakin drew it, upside down. Stork’s (1934) histological drawings show the same incorrect interpretation, as Stork consistently drew sec- tions of osphradia of bivalves upside down. Apparently there has been misrepresentation of the anatomy and anatomical context of the bivalve osphradium which has seriously af- fected a postulational-deductive process in studies of the organ’s structure and function. This matter will be discussed further below. No physiological work on bivalve osphradia has been found in the literature reviewed for this study. Some workers have hypothesized that the bivalve’s osphradium may aid in the regulation of ciliary activity in the gills (e.g. Aiello & Guideri, 1964). Some workers who have done experimental studies to determine physiological bases of ciliary activity in bivalve gills are Grave & Schmitt (1925) on Anodonta, Mya and Lampsilis, Setna (1930) on Pecten, Lucas (1931) on Mytilus and Megalonaias, Bulbring et al. (1953), Aiello (1960) on Mytilus edulis, Gosselin et al. (1962), and Aiello & Guideri (1964) on Mytilus edulis. Results and conclusions drawn from these studies will be evaluated below. In this paper | hope to clarify the structural and functional context of osphradia in certain freshwater bivalves. To do this | will (1) de- lineate the precise location and orientation of the osphradia in two species of freshwater bi- valved mollusks, Lampsilis ventricosa ovata (Barnes) and Corbicula cf. С. fluminea (Muller); (2) characterize in detail the histo- logical organization of the osphradia; (3) sum- marize the three-dimensional anatomy of the osphradia as revealed by serial sections; (4) examine the neuroanatomical context of the osphradia, especially as it clarifies objection to a current hypothesis that bivalve osphradia may function to regulate gill cilia activity; and (5) present alternative hypotheses for the role of osphradia in bivalve mollusks. MATERIALS AND METHODS The present study of freshwater bivalve osphradia was preceded and accompanied by prolonged behavioral observations of living bivalves in their natural habitat and in labora- tory aquaria. From behavioral studies of L. ventricosa and of C. fluminea, it was possible to develop an understanding of the structures of the exhalant and inhalant chambers and siphons, and their relation to siphoning, shell valve adduction and locomotion in living ani- mals. Some of these findings are described elsewhere (Kraemer, 1970, 1977). Several dozen neuroanatomical dissec- tions of relaxed tissues, both fresh and pre- served, were made of the two bivalve species. Histological material used included visceral ganglion and associated pallial tissues of five L. ventricosa, and at least 15 whole, serially sectioned C. fluminea, 2mm to 20 mm long. Serial sections were mostly sagittal, though some were transverse, some frontal, and were made from animals sacrificed in July, August, September, January and March. Animals of L. ventricosa used in this study were primarly from the White River and tribu- taries in northwestern Arkansas, and from the Arkansas River in central Arkansas and from the Buffalo River in northwestern Arkansas. C. fluminea used were from the Arkansas River in central Arkansas and from the Buffalo River in northwestern Arkansas. All animals were relaxed in Nembutal solution, fixed in Bouin’s fluid, and preserved in 70% ethanol. An aniline blue variation of Mallory’s triple staining technique was used (Schmitz, 1967). Photomicrographs were made with a 35 mm Camera in conjunction with a Leitz Ortholux microscope equipped for bright-field transmit- ted light, and with a 35 mm Wild MKa 1 cam- era in conjunction with a Wild M5 stereo- microscope. BIVALVE OSPHRADIA 207 LIST OF ABBREVIATIONS IN FIGURES A anus AA anterior adductor muscle AS edge of exhalent siphon BG basal granule BM basement membrane BN _ branchial nerve BS edge of inhalant siphon BV blood vessel С cilia CC ciliated cell CT connective tissue CVC cerebrovisceral connective E epithelium EC exhalant chamber ES exhalant siphon F foot G gill GRB ganglionated root of branchial nerve IS inhalant siphon LO location of osphradium LOG left outer gill LP labial palp М nerve fibers NA nerve supplying posterior adductor muscle NK nerve supplying kidney NP neuropil ОЕ osphradial epithelium OG ganglionated osphradial tissue ON osphradial nerve fibers PA posterior adductor muscle PN pallial nerve RBC roof of branchial chamber REC roof of exhalant chamber АС roof of inhalant chamber RPT right pallial nerve trunk 5 пеигопа! зота SP soma of osphradial neuron terminating in distal “pore” U umbo VG visceral ganglion RESULTS Location and orientation of the osphradia in L. ventricosa and C. fluminea Osphradia are located in a sheath of loose connective tissue, along with the butterfly- shaped visceral ganglion and its other asso- ciated structures, just anteroventral to the posterior adductor muscle and within the roof of the exhalant canal (Fig. 1, LO, VG). It is FIG. 1. Schematic diagram of a freshwater bivalve with left valve and left lobe of mantle removed to show location of visceral ganglion (VG) and of osphradia (LO) within the roof of the exhalant chamber (REC). Note that the ventral surface of the osphradia is exposed to excurrent fluids and not to incurrent fluids. important to note that the floor of the exhalant canal, which 1$ also the roof of the branchial cavity, is formed by the fusion of the dorsal portions of the demibranchs. Ortmann care- fully examined the branchial roof in unionids (“najades”). He referred to it as the dia- phragm and described it thus (1911: 288): “posteriorly to the abdominal sac (visceral mass) and foot, the inner laminae of the two inner gills unite in the median line of the body. ... By this union, together with the connection of the outer laminae of the outer gills with the mantle, a complete separation of the branchial chamber from the posterior part of the suprabranchial canals (cloacal cham- ber) is effected by a septum or diaphragm, which forms a horizontal division, from which the gills hang down... .” The floor of the exhalant chamber is thus the roof (“diaphragm”) of the inhalant or branchial cavity. Therefore, the visceral ganglion and its closely associated osphradia are directly exposed to fluid moving not through the incurrent siphon of the bivalve animal (as has been stated in the literature), but to fluid moving through the exhalant canal. That the foregoing is not generally understood by biologists is underscored, for example, by the fact that “osphradium” is defined as fol- lows by Pennak (1964: 365): “osphradium. Small sensory area in the incur- rent siphon of pelecypod and gastropod mol- lusks; presumably a chemoreceptor sensitive to incoming water.” 208 KRAEMER In contrast, in evident paraphrase of Yonge (1947), the obvious paradox of the bivalve osphradium's actual anatomical site, and the supposed function investigators have hy- pothesized for it was stated by Charles (1966: 504): “The possible function of these organs (osphradia) is obscure. As far as testing the respiratory current is concerned, water has to pass through the ctenidia before reaching these organs, and the sensory tentacles fring- ing the inhalent opening would be better suited for this function.” To visualize the location of the freshwater bivalve's osphradia, one may examine the т situ dissection of the visceral ganglion of L. ventricosa (Fig. 2). Osphradial tissues are appended in two triangle-shaped regions to the ventral surface of the ganglion, as.will be described in detail in the histological discus- sion below, and as shown in the diagrammatic insert for Fig. 2. When an organism's fresh tissues are ex- amined from the ventral aspect with a dissect- ing microscope, the membranous tissue comprising the roof of the exhalant chamber in the region immediately adjacent to the ventral surface of the visceral ganglion, may show a pinkish cast, and may bulge very slightly. There is no obvious manifestation of the osphradium, however, in fresh tissue. His- tological study of serial sections of the ani- mal is required to verify that the aforemen- tioned pinkish area is indeed the vicinity of the osphradial tissues. Histological organization of the osphradia of C. fluminea As shown in Fig. 3, each osphradium is comprised of three distinct kinds of tissues: the osphradial epithelium (OE), clusters of osphradial nerve fibers (ON), and gangli- _опаеч tissue (OG). The osphradial epithelium, even at low magnification, has a distinctive appearance. Published figures and histological descrip- tions of certain bivalve osphradia (e.g. Dakin, 1910 for Pecten; Stork, 1934 for Cardium edule, Маска subtruncata, Sphaerium FIG. 2. Visceral ganglion (VG) of female freshwater bivalve, Lampsilis ventricosa ovata, in situ, viewed from ventral aspect. Left and right mantle lobes and left and right gills have been spread apart. Roof of inhalant (branchial) chamber has been cut and removed. Roof of exhalant chamber has been removed to expose ventral surface of the visceral ganglion. Insert: diagram indicating location of osphradia (LO) on ventral surface of visceral ganglion. BIVALVE OSPHRADIA 209 in il EC 0:80) ых 0; ‹ 17g 104 Gatos a cose У FIG. 3. Зет! -Фадгаттайс, summary sketch showing sagittal section of osphradial tissues (OE, OG, ON) of the bivalve, Corbicula cf. C. fluminea, and some related structures described in this article. corneum, Pisidium henslowanum) show a histologically discrete epithelium on the dorsal surface. It is obvious that a similar peculiar epithelium does indeed constitute the ventral boundary of the osphradial tissue of C. fluminea. Furthermore, examination of thou- sands of serial sections of these clams allows verification that the osphradial epithelium 1$ found nowhere else among the clam's tis- sues. Osphradial epithelial cells are at once dis- cernible from neighboring cells as (i) smaller and more crowded, (ii) columnar, (iii) having oval nuclei with their long axes at right angles to the surface. Under oil immersion, typical osphradial epithelial cells also (iv) have a granular cytoplasm and (v) lack a basement membrane (Figs. 3, 4, OE). In lieu of a basement membrane, small clusters of nerve fibers and groups of neu- ronal soma closely parallel the base of the osphradial epithelium, and branch extensively among the cells of the osphradial epithelium. In their descriptions of gastropod osphradia, other authors (e.g. Demal, 1955) detail an “osphradial nerve” and ап “osphradial ganglion” as conspicuous, well-developed entities. In the bivalve osphradium examined here the nerve fibers are not organized into a distinct osphradial nerve. Similarly, the 8 to 10 clusters of neuronal soma associated with each osphradium are not grouped into an organized ganglion. The histological figures of other authors show no clearly defined nerve or ganglion as part of the bivalve osphradium, even though they are so labeled. | suspect that serial sections of other bivalve osphradia would show neuronal tissue similar to that of C. fluminea. Innervation of the osphradial epithelium is by means of the fibers mentioned above which parallel its base and which send abundant, branching, naked nerve fibers to outline or obliquely cross the surface of many of the epithelial cells, seeming to form a neuronal reticulum there. Contrary to what other workers have stated (e.g. Dakin, 1910), os- phradial epithelial cells are not triangular. They may, however, have seemed to other authors to be triangular, as so many nerve fibers do lie obliquely across the columnar epithelial cells. Most of the nerve fiber end- ings do not seem to penetrate the distal epithelial surface. A few osphradial epithelial cells seem to be extensions of underlying neuronal soma. These soma extend through the epithelium to terminate in a distal “роге” (Fig. 4C, SP). In sagittal sections near the proximal end of each osphradium, the organ typically appears as shown in Figs. 3 and 4. Sagittal sections near the distal end of the organ show a much more extensive osphradial epithelium, os- phradial nerve fibers, and osphradial gangli- onated tissue (Fig. 5, OE). Also at the distal end of the organ, the visceral ganglion itself 210 KRAEMER Be NOS С FIG. 4. A. Low-power photomicrograph of sagittal section of visceral ganglion of Corbicula cf. С. fluminea and proximal end of osphradium. B. Enlargement of Fig. 4A, showing detail of osphradium. C. Semi- diagrammatic sketch of sections comparable to Figs. 4В, С, showing the several tissue elements comprising the osphradium (OE, OG, ON) and relation of the tissue elements to the visceral ganglion (VG). BIVALVE OSPHRADIA 211 SR Te] MAA A ENS E prono > AU EE A a 01mm ‘ FIG. 5. Photomicrograph of lateral sagittal section through visceral ganglion (VG) and osphradium (OE, OG, ON) of Corbicula cf. C. fluminea. Note that the visceral ganglion appears appressed to the osphradial epithelium from the point marked OE to the far right side (anterior) of the picture. seems to be closely appressed to the os- phradial epithelium, and thus to the roof of the exhalant chamber (Fig. 4, VG). Gastropod osphradial epithelium has been evaluated by others (Demal, 1955; Bullock & Horridge, 1965; Hyman, 1967) as consisting of groups of one or another kind of cell. Yonge noted that (1947: 511) “...the epithelium contains mingled sensory, mucous and cili- ated cells...” in aspidobranch gastropods. Neither of the foregoing findings correspond to observations made for bivalve osphradial epithelium in this study. The typical osphradial epithelium for C. fluminea has already been described above. However, there are at least two other kinds of cells comprising the epi- thelium of the rest of the exhalant chamber roof in C. fluminea. Both kinds resemble cell types ascribed to osphradial epithelium in gastropods by other workers: (1) goblet- shaped cells (referred to as mucocytes by Demal, 1955, and others); and (2) rectangu- lar, ciliated cells (Figs. 3, CC, 6A, CC, 6B). The latter show a line of very slender cilia emanat- ing from an apparent row of basal granules in the long axis of the cell. The ciliary row may or may not originate at the distal surface of the cell. Most of the ciliated cells are scattered along the exhalant chamber roof, usually pos- terior to the osphradial epithelium. Some of these cells are also to be found anterior to the osphradium. 0.015 MM FIG. 6. А. Photomicrograph of roof of exhalant chamber of Corbicula cf. C. fluminea, posterior to the osphradium, showing characteristic, typical rec- tangular ciliated epithelial cells (CC). B. Sketch showing detail of ciliated cells from Fig. 6A. Note the conspicuous basement membrane (BM) in both 6A and 6B. The three-dimensional anatomy of the osphradia As revealed by extensive serial section study, the three-dimensional appearance of the osphradia in C. fluminea is that of two narrow, pie-wedge-shaped organs located at right angles to the longitudinal axis of the clam (Fig. 2, LO). The narrowing tips of the wedge- shaped organs approach but do not touch in the midline of the ventral surface of the 212 KRAEMER FIG. 7. Photomicrograph of midsagittal section of visceral ganglion of Corbicula cf. С. fluminea. Osphradial tissue is not evident here, as this region is between the two osphradia. visceral ganglion. A sagittal series of micro- scopic sections verifies that a distance of about 25 um separates the two osphradia from each other in the midline. No contiguous or other osphradial tissue is found in mid- sagittal sections of this region (Fig. 7). The wedge-shaped osphradia in a mature clam may measure about 100 um from their tips near the midline to their lateral aspect. Each osphradium tapers from about 80 um at its lateral limit (Fig. 5), to 10 um at its medial tip. Medially, each osphradium is thinnest (about 15 um), being comprised of little more than a short strip of epithelial cells and a few nerve fibers. At its wide, lateral margin, each osphradium contains many more neuronal soma and fibers, and attains a thickness of about 50 um (Fig. 5). Neuroanatomical context of the osphradial tissues As noted in the introduction of this paper, prior investigators have failed to determine the function of any bivalve osphradium. Some limited physiological investigations have seemed to other workers to implicate bivalve osphradia in regulation of ciliary activity of the gills (Ае!о & Guideri, 1964). To test some assumptions implicit in this hypothesis, it is important to determine the neuroanatomical context of osphradial tissue т С. fluminea in some detail. The neuroanatomical context of osphradial tissues was examined in careful serial section study. Anterior and medial to the osphradial region (Fig. 3), there are at least a dozen tiny nerves which extend from the visceral gan- glion to tissues of the kidney (Fig. 3, NK). On a level with the osphradial structures, several nerves emanate from the dorsal surface of the visceral ganglion, the largest supplying poste- rior adductor muscle tissue (Fig. 3, NA). In the region of its association with the osphradium, the visceral ganglion has many neuronal soma. The axons of these soma may extend from the dorsal cortex of the ganglion across the central neuropil, into or through the ventral cortex of the ganglion and thence into the osphradial tissues. Anterior and ventrolateral to the osphradial complex, a large extension of the visceral ganglion on each side closely adheres to the roof of the exhalant chamber, and rounds the ventral curve (Fig. 8, GRB) where the ex- halant chamber roof becomes the floor of that chamber (also the roof of the branchial chamber, or the “diaphragm” of Ortmann, 1911). On each side this ganglionated exten- sion gives rise to a branchial nerve. Each branchial nerve courses along the branchial shelf just under the shelf's simple epithelium and basement membrane, and just dorsal to a prominent longitudinal muscle. At the poste- rior end of the roof of the branchial chamber (also the floor of the exhalant chamber), each branchial nerve terminates in apparent junc- tion with the aforementioned, cislateral, longi- tudinal muscle. From the foregoing study it does not seem likely that the osphradia of C. fluminea func- tion in regulation of gill cilia. It seems more likely that the osphradia are related to function of the posterior adductor muscle or to the kidney. Implications of these neuroanatomical observations will be considered further below. DISCUSSION This study shows that: 1) The osphradia of certain freshwater bi- valves are located along with the visceral BIVALVE OSPHRADIA 213 FIG. 8. Photomicrograph of sagittal section of Corbicula cf. C. fluminea, showing the ganglion- ated root of a branchial nerve (GRB) as it enters the branchial shelf (diaphragm). Region shown 15 ante- rior to the visceral ganglion proper, and is some distance from the osphradium. Proximally (and dorsally) the branchial root is connected to the visceral ganglion. Distally (and ventrally) the ganglionated root becomes the branchial nerve which courses of the posterior end of the branchial shelf. Relate to location of right branchial nerve shown in Fig. 2. ganglion with which they are closely associ- ated, anteroventral to the posterior adductor muscle and within the membrane which forms the roof of the exhalant siphon. Thus it is fluid within the exhalant canal which moves freely over the highly innervated epithelial surface of the osphradia. This finding confirms ana- tomical relationships described earlier for Corbicula sp. (Kraemer, 1977). 2) The osphradia of these bivalves are paired, and extend on each side of the mid- ventral surface of the visceral ganglion later- ally and anteriorly to terminate in a small pro- jection of the exhalant chamber roof. 3) Each osphradium is a complex of three parts: (a) the osphradial epithelium, which lacks a basement membrane, and is com- prised of small, abundantly innervated colum- nar cells; (b) groups of nerve fibers immedi- ately subjacent to the osphradial epithelium, with many fibrous connections ventrally to the osphradial epithelium and dorsally to the visceral ganglion; (c) several clusters of neuronal soma, some of which extend from the visceral ganglion proper into the tissue space between the visceral ganglion and the underlying osphradial epithelium, and some of which seem to be merely specialized re- gions of the visceral ganglion. Many of these neuronal soma have long processes which extend as nerve fibers among cells of the osphradial epithelium. 4) Examination of the neuroanatomical context of the osphradia indicates: (a) on a level with the osphradial structures, nerves emanate from the dorsal surface of the visceral ganglion to supply the posterior ad- ductor muscle; (b) immediately anterior and medial to the osphradial structures, many tiny nerves extend from the visceral ganglion to the kidney; and (c) further anterior and later- ally from the osphradial structures, a pair of large branchial nerves may be traced from their ganglionated bases at the visceral ganglion, into the roof of the inhalant chamber (also the floor of the exhalant chamber). Neuroanatomically the osphradia are there- fore more closely associated with the poste- rior adductor muscle and even with the kidney than they are with branchial nerves. Many of the above findings are summarized graphically in Fig. 3. Some of the histological findings seem generally to corroborate related findings in a marine bivalve Mytilus edulis (e.g. Lucas, 1931). Other results of the pres- ent study cast doubt on the hypothesis that bivalve osphradia are particle-size detectors related to nervous control of gill cilia (e.g. Aiello & Guideri, 1964). In many gastropods, the osphradium 1$ often located within the inhalant siphon, and is separated from the visceral ganglion by a dis- tinct osphradial nerve and osphradial gan- glion (e.g. Demal, 1955). By contrast, in the present study of bivalves, the paired osphra- dia are intimately associated with the antero- ventral surface of the visceral ganglion itself. In evaluating the neuroanatomical context of these bivalve osphradia, it is apparent that the osphradia are much more closely associated with nerves innervating the posterior adductor muscle and the kidney than with branchial nerves. The branchial nerves in turn are closely associated with the visceral ganglion, both via ganglionated bases, and via tiny nerves emanating from the surface of the 214 visceral ganglion itself. The branchial nerves are not closely associated with the osphradia. While various authors have proposed func- tions for bivalve osphradia (e.g. Yonge, 1947), the literature reviewed for this study offers no physiological evidence for a chemo- receptor or mechanoreceptor role for the bi- valve osphradium. Aiello & Guideri (1964), who recorded changes in rate of beat of gill cilia in response to experimental stimulation of the visceral ganglion of Mytilus edulis, con- cluded that the osphradium was involved. They argued that Bailey 8 Laverack (1963) had recorded impulses from the branchial nerve of a gastropod in apparent response to stimulation of its osphradium. Aiello 4 Guideri erred. Bailey 8 Laverack recorded impulses from the visceral ganglion, not from the branchial nerve. While a few authors have correctly de- scribed the general location of these organs within the exhalant chamber (e.g. Yonge, 1947; Charles, 1966), many others have not. It may be that a simple, but repeated error can account for the inchoate state of literature on bivalve osphradia. Frequent citations of cer- tain studies of bivalve osphradia (e.g. Dakin, 1910; Stork, 1934), have perpetuated a mis- representation of bivalve osphradia which views the osphradia as inverted from their actual position. lt seems likely that this error stems from an effort by early workers to homologize bivalve osphradia with those of gastropods. The error is serious, because it has accompanied the assumption (implicit in many of the histological and physiological studies to date), that the bivalve osphradia are associated with the roof of the inhalant chamber, and thus with the incurrent or branchial siphon. As shown above (Figs. 1, 3), the bivalve osphradia are actually associ- ated with the roof on the exhalant chamber. Finally, it seems that new hypotheses are needed regarding the function(s) of the bi- valve osphradia. New hypotheses should in- corporate information presented here con- cerning the precise anatomical site, histologi- cal organization, accurate three-dimensional structure and neuro-anatomical context of these paired organs. New hypotheses might consider the possibility that bivalve osphradia are anatomically more suited to being “eyes” than a “nose.” This is not a fanciful sugges- tion. With its location deep within the animal, its very intimate association with the largest ganglion and with its thickly innervated epi- thelium, | think a bivalve osphradium histo- logically resembles the pineal gland of the KRAEMER vertebrate brain. Zoologists are generally well aware of light as an environmental stimulus to the pineal gland which enables migratory birds to respond with behavioral changes, for example, In bivalve mollusks an analogous phenomenon may occur. Elvin (1978: 646) noted that in certain intertidal mussels ”... light reaching the mantle cavity . . . (was) found to be predominantly of long wave- length. . . . (There is) a positive correlation be- tween thermal shocks enhanced by light, dis- appearance of neurosecretory material, and the release of oocytes.” Might it be that the extensively innervated epithelium of the osphradium is a light sensor which functions in the regulation of seasonal behavior (e.g. Kraemer, 1970, on three species of Lampsilis) or reproductive physiology of the bivalve? (Regarding the latter, histological and neuroanatomical details of the herma- phroditic process in C. fluminea have been worked out by Kraemer, 1978). The new hypotheses might also incorporate the possibility that functional links to kidneys and to adductor muscles may implicate the bivalve osphradia not only in control of fluid movement through the exhalant chamber, but also in adduction of the shell valves. ACKNOWLEDGEMENTS Thanks are due to Profs. C. M. Yonge, George M. Davis and Carol C. Jones, who critically reviewed the manuscript, and to Prof. Eric Russert Kraemer who contributed help- ful suggestions in the revision of the manu- script. | thank Frances Waite for her skillful drawing of the Lampsilis dissection in Fig. 2, and James Hawkins for his expert assistance in the drafting of Figs. 1, 3, 4C and 6B. LITERATURE CITED AIELLO, E. L., 1960, Factors affecting ciliary activ- ity on the gill of the mussel Mytilus edulis. Physio- logical Zoology, 33: 120-135. AIELLO, E. L. & GUIDERI, G., 1964, Nervous con- trol of ciliary activity. Science, 146: 1692-1693. ALEXANDER, C. G. & WELDON, M. W., 1975, The fine structure of the osphradial leaflets in Conus flavidus. Marine Biology, 33: 247-254. ANDERSON, E., 1963, Cellular and subcellular organization of the osphradium of Busycon. Proceedings of the Sixteenth International Con- gress of Zoology, 2: 280 (abstract). BAILEY, D. F. & BENJAMIN, P. R., 1968, Anatomi- cal and electrophysiological studies on the gastropod osphradium. Symposium of the Zoo- logical Society of London, 23: 263-268. BIVALVE OSPHRADIA 215 BAILEY, D. F. & LAVERACK, М. F., 1963, Central nervous responses to chemical stimulation of a gastropod osphradium. Nature, 200: 1122-1223. BAILEY, D. F. 4 LAVERACK, M. F., 1966, Aspects of the neurophysiology of Buccinum undatum L. (Gastropoda). I. Central responses to stimulation of the osphradium. Journal of Experimenta! Biol- ogy, 44: 131-148. BAUMGARTEN, R. VON, JAHAN-PARWAR, B. 8 SMITH, M., 1968, Effect of olfactory and osmotic stimulation of the osphradium on neurosecretory nerve cells in Aplysia. Mental Health Research Institute (University of Michigan), Annual Re- port, 1968: 20-21. BAYNE, В. L., WIDDOWS, J. 8 THOMPSON, В. J., 1976, Physiology |. т BAYNE, В. L. (ed.), Marine Mussels: Their Ecology and Physiology. Cambridge University Press, New York, р. 207- 260. BENJAMIN, P. R., 1971, On the structure of the pulmonate osphradium. Part 2. Ultrastructure. Zeitschrift für Zellforschung und mikroskopische Anatomie, 118: 168-189. BROWN, А. С. & МОВЕЕ, В. G., 1960, Function of the osphradium in Bullia (Gastropoda). Nature, 188: 1045. BULBRING, E., BURN, J. H. & SHELLE, H. J., 1953, Acetylcholine and ciliary movement in the gill plates of Mytilus edulis. Proceedings of the Royal Society of London, ser. B, 141: 445. BULLOCK, Т. G. 4 HORRIDGE, С. A., 1965, Struc- ture and Function in the Nervous Systems of Invertebrates, Il. Freeman, San Francisco and London. CARR, W. E. S., 1967, Chemoreception in the snail Nassarius. Biological Bulletin, 133: 90-127. CHARLES, G. H., 1966, Sense organs (less Cephalopoda). /n WILBUR, К. L. & YONGE, С. М. (eds.), Physiology of Mollusca, Il. Academic Press, New York, р. 455-521. CRISP, M., 1973, Fine structure of some proso- branch osphradia. Marine Biology, 22: 231-240. DAKIN, W. J., 1910, The visceral ganglion of Pecten, with some notes on the physiology of the nervous system and an inquiry into the innerva- tion of the osphradium in the Lamellibranchiata. Mitteilungen aus der Zoologische Stazione zu Neapel, 20: 1-40. DAKIN, W. J., 1928, The anatomy and phylogeny of Spondylus, with a particular reference to the Lamellibranch nervous system. Proceedings of the Royal Society of London, ser. В, 103: 337- 354. DEMAL, J., 1955, Essai d’histologie comparée des огдапез chémorecepteurs des gastéropodes. Mémoires de l’Académie Royale de Médécine de Belgigue, 29: 1-82. ELVIN, D. W., 1978, Environmental and internal control of oogenesis in Mytilus. American Zoolo- gist, 18: 646 (abstract). FREIDENFELT, T., 1904, Uber den feineren Bau des Visceral Ganglions von Anodonta. Acta Universitatis Lundensis Sectio II Medica Mathematica Scientiae Rerum Naturalium 2, 40(5): 1-28. GOSSELIN, R. E., MOORE, K. E. & MILTON, A. S., 1962, Physiological control of molluscan gill cilia by 5-hydroxytryptamine. Journal of General Physiology, 46: 277. GRAVE, C. & SCHMITT, F. O., 1925, A mecha- nism for the coordination and regulation of ciliary movement as revealed by microdissection and cytological studies of ciliated cells of mollusks. Journal of Morphology, 40: 479-510. HYMAN, L. H., 1967, The Invertebrates: Mollusca I. McGraw-Hill, New York. JAHAN-PARWAR, B., VON BAUMGARTEN, R., HUKUHARA, T. & SMITH, M., 1968, Effects of osmotic stimulation of the osphradium upon the activity of neurosecretory cells in the abdominal ganglion of Aplysia californica. Proceedings of the International Union of Physiological Sci- ences, 7(24): 214 (abstract). JAHAN-PARWAR, B., SMITH, M. & VON BAUM- GARTEN, R., 1969, Activation of neurosecretory cells in Aplysia by osphradial stimulation of sea- weed oyster shrimp. American Journal of Physi- ology, 216(5): 1246-1257. KAMARDIN, N. N., 1976, Response of osphradium of mollusk Lymnaea stagnalis to different oxygen concentrations in water. Journal of Evolutionary and Biochemical Physiology, 12(5): 427-428. (English translation of Zhurnal Evolyutsonnoi Biokhimii i Fiziologii). KOHN, A. J., 1961, Chemoreception in gastropod molluscs. American Zoologist, 1: 291-308. KRAEMER, L. R., 1970, The mantle flap in three species of Lampsilis (Pelecypoda: Unionidae). Malacologia, 10: 225-282. KRAEMER, L. R., 1977, Aspects of the functional morphology of the mantle/shell and mantle/gill complex of Corbicula (Bivalvia: Sphaeriacea: Corbiculidae). Bulletin of the American Malaco- logical Union, 1977: 25-31. KRAEMER, L. R., 1978, Corbicula fluminea (Bi- valvia: Sphaeriacea): the functional morphology of its hermaphroditism. Bulletin of the American Malacological Union, 1978: 40-49. LANKESTER, R., 1883, Mollusca. т Encyclo- paedia Britannica ХУ! (9th edition). Scribners and Sons, New York. LUCAS, A. M., 1931, The distribution of the branch- ial nerve in Mytilus edulis and its relation to the problem of nervous control of ciliary activity. Journal of Morphology, 51: 195-205. MICHELSON, E. H., 1960, Chemoreception in the snail, Australorbis glabratus. American Journal of Tropical Medicine and Hygiene, 9: 480—487. NEWELL, P. F. & BROWN, A. C., 1977, The fine structure of the osphradium of Bullia digitalis (Gastropoda: Prosobranchia). Malacologia 16: 197-205. ORTMANN, A. E., 1911, Monograph of the naiades of Pennsylvania. Memoirs of the Carnegie Mu- seum, 8: 1-347. PENNAK, R. W., 1964, Collegiate Dictionary of Zoology. Ronald Press, New York. PHILLIPS, D. W., 1975, Localization and electrical activity of the distance chemoreceptors that mediated predator avoidance behavior in 216 KRAEMER Acmaea limatula and Acmaea scutum (Gas- tropoda: Prosobranchia). Journal of Experimental Biology, 63: 403—412. SCHMITZ, E. H., 1967, Visceral anatomy of Gammarus lacustris Sars (Crustacea: Amphi- poda). American Midland Naturalist, 78: 1-54. SETNA, S. B., 1930, The neuro-muscular mecha- nism of the gill of Pecten. Quarterly Journal of Microscopical Science, 73: 365-391. SIMPSON, L., 1971, Light and electron microscopic structure of the osphradium in the pulmonate gastropod, Helisoma. American Zoologist, 11: 697 (abstract). SOKOLOV, V. A. & KAMARDIN, М. N., 1977, The relation of impulse frequency in the osphradial nerve to the concentration of oxygen and inulin in liquid passing over the osphradium of the pond snail. Leningrad University, Leningrad, USSR. Vestnik Leningradskogo Universiteta Biologie, Biologii Geografii i Geologi (1). 1977: 87-90. STINNAKRE, J. & TAUC, L., 1969, Central neuronal response to the activation of osmore- ceptors in the osphradium of Aplysia. Journal of Experimental Biology, 51: 347-361. STORK, H. A., 1934, Beitráge zur Histologie und Morphologie des Osphradiums. Archives Neerlandaises de Zoologie, 1: 71-99. TOWNSEND, С. В., 1973, The role of the osphradium in chemoreception by the snail Biomphalaria glabrata. Animal Behaviour, 21:549-556. YONGE, С. М., 1947, The pallial organs in the Aspidobranch Gastropoda and their evolution throughout the Mollusca. Philosophical Trans- actions of the Royal Society of London, B, 232: 443-578. MALACOLOGIA, 1981, 20(2): 217-253 GENETIC RELATIONSHIPS AMONG RECENT UNIONACEA (BIVALVIA) OF NORTH AMERCIA! George M. Davis and Samuel L. H. Fuller Academy of Natural Sciences of Philadelphia, Nineteenth and the Parkway, Philadelphia, PA 19103, U.S.A. ABSTRACT The purposes of this paper are to determine why there has been so little agreement among classifications of North American Unionacea, to test the Heard & Guckert (1971) assumptions that the number of marsupial demibranchs and length of breeding season serve to define higher taxa, to examine the congruency among major classifications of North American Unionacea, and to establish a classification resulting from a synthesis of data derived from molecular genetics, comparative anatomy, and zoogeography through time. Immunoelectrophoretic studies of 52 species belonging to 27 genera were conducted. We scored the percent difference between pairs of taxa. Data were analyzed with multivariate techniques of the NT-SYS program. Emphasis was placed on results of multidimensional scal- ing, ordination, minimum spanning tree, and subsets. On the basis of our results we determined that in the Nearctic Unionacea there are one family (Unionidae) and three, genetically very distinct subfamilies: Margaritiferinae, Anodontinae, and Ambleminae. The three subfamilies are clearly defined morphologically and immunologically. The Ambleminae are further divided into four tribes: Gonideini, Amblemini, Pleurobemini, Lampsilini. It is clear that both tetragenous and ectobranchous taxa have evolved in various clades. The ectobranchous genus Elliptio and tetragenous genus Fusconaia are closely related in the Pleurobemini, the ectobranchous genus Cyclonaias and tetragenous genus Quadrula are closely related in the Amblemini, and the tetragenous Gonidea is more closely related to the Lampsilini (which are ectobranchous) than to the Pleurobemini or Amblemini. The ecto- branchous state has undergone parallel evolution, as have different lengths of breeding season. Our classification and that of Ortmann (1910a) have the greatest congruence. We consider these classifications to reflect real clades more closely than other systems do, because both are based on all of the data available. We consider the other classifications to be artificial in that they are based on conchology alone or on the unjustified weighting of one or two key characters. We differ from Ortmann and all previous workers in establishing the Anodontinae as a taxon of equal standing with the Margaritiferinae as a second group and with all other North American Unionidae in the Ambleminae as a third. INTRODUCTION North American unionacean bivalves (unios, freshwater mussels, naiades) com- prise one of the most diverse radiations of macroinvertebrates seen today in fresh water. There are about 50 nominal genera, which include over 225 species and subspecies (Heard & Guckert, 1971; Burch, 1973, 1975). Unios have dominated streambeds in terms of biomass and numbers of individuals, but de- creasingly in this century. Centers of endem- ism and high species diversity are found in the eastern United States, e.g. the Ohio, Tennes- see and Coosa-Alabama river drainages. Numbers of sympatric species literally paved the large-river shoals in the early 19th cen- tury. For example, Conrad (1834) reported the richest of all known localities, a section of the Tennessee River that later became known as Mussel Shoals. The shoals contained some 70 species packed valve to valve. Even though the diversity and abundance of the unionid fauna stimulated a more than 80 year continuous outflow of systematic and taxonomic literature concerned with higher- category relationships among unionids (re- view by Heard 8 Guckert, 1971), there is little agreement among classifications today. Dis- parity among major classifications (reviewed in Appendix 1) seems to occur because 1) different sets of characters were used by dif- ÎThis work was supported by National Science Foundation grant DEB 78-01550. (217) 218 DAVIS AND FULLER ferent investigators; 2) a monothetic basis for classification was used by some; 3) there is the high probability that one or more “key” character-states has undergone parallel or convergent evolution, and 4) too few morpho- logical characters with unique character- states exist (or have been found) that would enable a satisfactory comparison of taxa. Classifications based primarily on shell characters persist to the present (Frierson, 1927; Modell, 1942, 1949, 1964; Haas, 1969 a, b). Some early classifications were based on additional characters of the soft parts, e.g. gill structure, marsupium, and glochidia (Simpson, 1896, 1900, 1914; Sterki, 1898, 1903). Ortmann (1910a, 1911, 1912b, 1916) extended the work of Simpson and Sterki by increasing the number of characters derived from soft-part morphology and integrated all available morphological data (i.e. on shell and soft parts). Use of shell characters for classifi- cation above the species level eventually was rejected. Hannibal (1912) stated that shell characters were of no use in establishing taxa above the generic level. Hannibal was fol- lowed by Heard & Guckert (1971), who stated: “... we have subjectively elected to ignore one entire array of characters (i.e., conchological features) and to suggest soft- part anatomy and reproductive habits as pre- eminent in describing phylogenies.” Heard & Guckert (1971) especially weighted two characters involving reproduction. These are the number of demibranchs used as the marsupium and the length of the breeding season. For example, they recognized two families, Amblemidae and Unionidae, on the basis that species with four marsupial demi- branchs (tetragenous) belong to the former family while those with only the outer two demibranchs marsupial (ectobranchous) be- long in the latter family. They created sub- families on the basis of whether taxa are bradytictic (i.e. long-term breeders, retaining glochidia except in the Nearctic summer) or tachytictic (i.e. short-term breeders, retaining glochidia only in the Nearctic summer). We initiated our work on higher-category relationships among North American unios in order to test the validity of the Heard & Guckert assumptions and classification. Be- cause there is so little agreement among the major classifications, we suspected that one or more key character-states has undergone convergent or parallel evolution. We also sus- pected that one cannot excessively weight characters having to do with reproductive strategies. For example, a_ strategy of bradytixis might occur again and again in dif- ferent radiations of unionids. Likewise, the tetragenous and ectobranchous conditions feasibly could occur in different radiations. We noted that brooding young in the pallial ovi- duct of mesogastropods has arisen inde- pendently in several families of different superfamilies (Fretter & Graham, 1962). Our suspicions are not without basis. For example, Fusconaia masoni (Conrad) was placed in the genus Elliptio by Haas (1969a) and relegated to Pleurobema (Lexingtonia) by Johnson (1970). Elliptio and Pleurobema were considered Unionidae by Ortmann (1910a, 1911, 1912b, 1919) and Heard & Guckert (1971). However, F. masoni is tetra- genous and thus belongs to Fusconaia (Ful- ler, 1974). Except for the one character-state difference, one finds little difference between F. masoni and various species of Elliptio and Pleurobema. However, if the Heard & Guckert classification were followed, the species would have to be transferred from the ecto- branchous Unionidae to the tetragenous Amblemidae on the basis of that one charac- ter-state. We further suspected that unionid species have too few unique morphological character- states to permit an adequate phenetic or cladistic analysis of relationships. We admit that there may be more morphological char- acters, but these have yet to be discovered. Such characters probaby would have to be found by detailed comparative anatomical studies of internal organ systems. Because of the dearth of unique morphological character- states, we established a program to assess relationships on the basis of molecular genetics. In this paper we present a higher classification of North American Unionacea based on immunoelectrophoretic data and what morphological, paleontological, and zoogeographical data are available. In so doing, we assess not only the relationships among unionid taxa, but also the relative values of different approaches to unionacean taxonomy used to structure the various major classifications. MATERIALS AND METHODS Species studied Fifty-two species, representing 27 genera, were studied (Table 1). These species were TABLE 1. Fifty-two species of North American Unionacea alphabetized by the code designations used in this UNIONACEAN GENETIC RELATIONSHIPS study. Localities are given with ANSP catalog numbers. Code * = type-species + = monotypic genus Species Anodonta cataracta (Say) Actinonaias carinata (Barnes) Anodonta imbecillis (Say) Anodonta implicata (Say) Amblema perplicata (Conrad) Amblema plicata (Say) Alasmidonta undulata (Say)* Anodonta wahlametensis (Lea) Carunculina parva (Barnes)* Cyclonaias tuberculata (Raf.)*, + Cumberlandia monodonta (Say)*,+ Elliptio buckleyi (Lea) E. complanata (Lightfoot) kk E. crassidens (Lam.)* E. icterina (Conrad) E. lanceolata (Lea) жж* жж* Е. waccamawensis (Lea) Fusconaia Cf. flava (Raf.) F. ebena (Lea) F. flava (Raf.) Е. masoni (Conrad) Gonidea angulata (Lea)*,+ Glebula rotundata (Lam.)*, + Lasmigona costata (Ва+.)* Lampsilis claibornensis (Lea) Leptodea fragilis (Raf.) Lampsilis hydiana (Lea) Ligumia nasuta (Say) Lampsilis ovata (Say)* L. radiata (Gmelin) Ligumia recta (Lam.)* Lampsilis splendida (Lea) L. teres (Raf.) L. ventricosa (Barnes) Margaritifera falcata (Gould) Megalonaias gigantea (Barnes)* Margaritifera hembeli (Gould) M. margaritifera (L.)* Proptera alata (Say)* Pleurobema cordatum (Raf.) Plectomerus dombeyanus (Val.)*,+ Proptera purpurata (Lam.) Ptychobranchus subtentum (Say) Quadrula apiculata (Say) Q. cf. quadrula (Raf.)* Q. cylindrica (Say)** Quincuncina infucata (Conrad) Quadrula pustulosa (Lea) Tritogonia verrucosa (Raf.)*, + Uniomerus tetralasmus (Say)* Villosa delumbis (Conrad) V. iris (Lea) ““type-species of Orthonymus ‘used for analysis of conspecific populations Locality Gloucester Co., New Jersey Clark Co., Arkansas Jenkins Co., Georgia Hartford Co., Connecticut Rapides Parish, Louisiana Clark Co., Arkansas Hartford Co., Connecticut Modoc Co., California Rapides Parish, Louisiana Hancock Co., Tennessee Hancock Co., Tennessee Putnam Co., Florida Gloucester Co., New Jersey Sussex Co., New Jersey Barnwell Co., South Carolina Barnwell Co., South Carolina Wayne Co., Pennsylvania Hancock Co., Tennessee Jenkins Co., Georgia Jenkins Co., Georgia Barnwell Co., South Carolina Jenkins Co., Georgia Columbus Co., North Carolina Rapides Parish, Louisiana Greene Co., Alabama Hancock Co., Tennessee Jenkins Co., Georgia Modoc Co., California Rapides Parish, Louisiana Hancock Co., Tennessee Lowndes Co., Mississippi Hancock Co., Tennessee Rapides Parish, Louisiana Burlington Co., New Jersey Hancock Co., Tennessee Sussex Co., Delaware Clark Co., Arkansas Barnwell Co., South Carolina Rapides Parish, Louisiana Clark Co., Arkansas Oregon Rapides Parish, Louisiana Rapides Parish, Louisiana Schuylkill Co., Pennsylvania Hancock Co., Tennessee Clark Co., Arkansas Rapides Parish, Louisiana Clark Co., Arkansas Hancock Co., Tennessee Evangeline Parish, Louisiana Rapides Parish, Louisiana Hancock Co., Tennessee Crawford Co., Georgia Clark Co., Arkansas Rapides Parish, Louisiana Rapides Parish, Louisiana Jenkins Co., Georgia Hancock Co., Tennessee ANSP voucher no. 333526 341958 333563 334650 334560 341939 334649 345880 334564 335048 341956 334427 333527 334428 333296 339430 333566 333565 339967 334563 340626 335049 333564 339965 334556 335047 335046 334558 334251 335029 339342 340628 334432 334557 339339 334553 334426 334867 335040 340629 334555 390630 335045 339670 334562 335041 334539 340946 339671 334561 333569 335050 220 chosen among all those collected from vari- ous localities across the country because they were representative of each of the famil- ies and subfamilies recognized by Ortmann (1910a, 1911, 1912b, 1916, 1919), Modell (1942, 1949, 1964), and Heard & Guckert (1971). We collected and chose type-species whenever possible. We were able to study 18 type-species of the 27 genera (66.7%). These are marked with an asterisk in Table 1. Shells of each population are maintained as voucher specimens in the Academy of Natural Sci- ences of Philadelphia (ANSP); catalog num- bers and locality data are given in Table 1. Preparation of antigens and antigen bank Foot muscle and gravid gill were used as a source of proteins. Gravid gills were carefully inspected for the presence of unionicolid mites in order to ensure that antigen prepara- tions were not contaminated with mite anti- gens. Tissues were pooled from 12 to 60 indi- viduals according to their size. The gravid gill extract essentially equaled glochidial extract because gill filament tissue added very little in terms of protein mg in comparison to the yield from the glochidia. In preparing extracts, 300 mg blotted foot tissue (cleaned of gonad) were homogenized in 1.5 cc buffer. Homo- genization was accomplished by first subject- ing the mixtures to a Waring Blender for two minutes and then to sonication-homogeniza- tion (via a Polytron) for two minutes per 30 ml. The homogenate was centrifuged at 6900 x g for 20 minutes; the supernatant was lyophil- ized (1 ml per 2 ml ampule). All prelyophiliza- tion operations were carried out at 1-3°C. In this manner 100 to 300 ampules of lyophilized extract from each population were prepared for storage in freezers (at —20°C). The protein content of each lyophilized batch was determined by the folin reagent test (Daughady et al., 1952). Antisera Two rabbits (New Zealand white, virgin, female, 7-8 Ibs) initially were used per mussel population. Lyophilized antigens were recon- stituted with normal saline and injected sub- scapularly with an equal volume of Freund's complete adjuvent. There were two injection series (days 1, 3, 5, and 7; rest 3 weeks; re- peat the series). Each injection contained 2 mg protein. We bled out the rabbits by heart puncture 4 days after the last injection. DAVIS AND FULLER Titre and Serum Quality—Antiserum quality was tested by immunoelectrophoresis (see below). Sera were kept and used in experi- ments if 10 or more precipitin arcs resulted in the homologous reaction. If an antiserum was discarded for producing too few antigen- antibody precipitin systems, two or more rab- bits were used to produce specific antisera. Controls—Each rabbit was bled from the ear prior to the first injection with antigen; the serum was tested for reactivity to molluscan antigen. Absorption—An antiserum was absorbed with a heterologous antigen by adding 0.8 ml antiserum to an ampule of lyophilized antigen, swirling, leaving at room temperature for 30 minutes, refrigerating for 30 minutes, and centrifuging for 20 minutes (6900 x 9). Immunoelectrophoresis—The procedures used are those of Davis (1969) and Davis & Suzuki(1971), with some adjustments. The 2% agar noble contained 0.45% NaCl, 1:10,000 merthiolate, and half-strength barbital-acetate buffer of pH 8.2. Full-strength buffer con- tained 5.4g Na-barbital, 4.3 g sodium ace- tate, and 58.2 ml 0.1 М НС per liter. Protein concentrations of antigens were adjusted to 6 mg/ml. Direct current of 6- 8 v/cm across the slides was sustained for one hour. Analysis of immunological data Twelve slides were used in each experi- ment, of which two were controls, i.e. the homologous system with unabsorbed serum. We determined the number and position of each precipitin arc by comparing the experi- mental slides with control slides. In each ex- periment we absorbed the serum of the refer- ence population (homologous system) with a heterologous antigen so that there were five sets of absorbed sera. The two wells punched in the agar on each slide were loaded with homologous and heterologous antigens, re- spectively. Absorbed antisera were used in the slots of the 10 non-control slides. Lack of arcs between the slots and the wells with the heterologous antigens indicated complete absorption. The number of arcs between the slot and the well with the homologous anti- gens indicated the number of anitgens unique to the reference species. The position of the arc identified the antigen. We scored the percent difference between taxa. The average number of precipitin arcs was 12 with a range of 10 to 14. We analyzed 221 UNIONACEAN GENETIC RELATIONSHIPS I 0 SLE 0'57 GE S'Ot 0'6p 0'0p 0'Zp S'9p 0'0p OVS 979 0'06 0'c9 vl ejoe/e]e9 виороиу 0 S'c€ SG 0'6c SI A 5'55 995 S'c€ Sic 999 95 0'05 0'19 el sıqwnj3p ESOJIIA 0 g'8c 915 GLE GSE S'Sc GLE 0'0€ S'sp 0'06 S'Sp 19 cl ejeje е/э}40/4 0 Sve 99 595 O'ZC 595 GLE S'09 S'09 Sr 0'59 LL ¡Áep9nq ondı3 0 09. oz! 9'96 Gee O'EC 0'8S Oops Ss tp 0'S9 OL eJe/Nn9/8qN} sereuojoAd 0 5'06 0'€l S'6c G'ec 0'0S S'99 ger 029 6 eajuebib sereuojebayy 8 eınıpenb е/плрепо JE Iuosew “Y 9 BABY BIEUOISNZ 0 0‘0S 0'57 Gir 019 5 вэиршиИэ еплрепо y виороиош вриеиэдшп) € в/эуиеблеш “y с aqway ‘и L в1еэ/е) елэддиеблеи/ '(елэзцие jo $1э$ FL x зиэбцие jo $1э$ FL) exe} | jo чозиедшоо ssoJo ul зиэбцие Jo ээцэлаир jugoied эбелэле ay] Бим!6 хщей ‘с 3718VL DAVIS AND FULLER 222 Ly 55 Oc 55 Ge 85 vs Lp 55 lp 09 Le 55 8 0 91 Le Lv 9! Lp Ov 193 Lv Se Ov 05 55 99 59 05 Lv 05 05 81 55 Se 55 8 81 99 Lv 05 05 M3 Lp 9! 0€ 55 Se GZ €9 Lv Lv 05 05 81 55 8 55 91 0 05 91! 05 Op 13 Lv 55 05 55 Se 8S cL os 05 05 Ov Le Sc 0 55 91 81 99 Se 05 OV 13 Lv 91! Oc Se Se 55 59 55 Se 05 05 6 Oc 8 55 0 81 Ly Sc Lv 09 93 Se Ge 05 91 Sc 8? CL 55 91 Ly 0€ Le Oc Se Lv 91 95 85 0 05 OV 19 Lv Se Oc Se Sc 8p 59 Lv 0 Ir 05 81 9c 55 05 91 Le 05 91 95 Or BW Lv 0 Oc Se 55 09 06 Lv Lv 05 Op Sy Ov Sc 05 Lv Le 85 Sc 05 Ov AL ON эм OE 91 Se 85 ON Lv Lv SNES ONE 22% 55 ON: ON ЭМ 5 85 9! ON OV ID 55 Se Oc Se 91 85 59 Lv 8 Lv Ov Le 55 Se Lv 91 LC 05 91 95 OV Pd Lv 91 0 8 Se 05 59 55 Se Lv Or St Oc Sc Lv Sc Sv 05 91 05 OV 20 Sc 55 OL 91 Se 8S cl 05 Ge Lp 05 Le 55 Lv 05 55 95 99 8 05 05 eo 55 91 Oc 0 Se 8r 59 Lv 91 Lv 05 Le Oc 91 Ly Sc Le Ly 8 Lv 05 qq) Lv E74 OV 91! 55 05 CL Ly 8 Ir 05 0 LC 91 Lv 55 95 99 Se Lv OV uy 55 Lv 05 Se Se 99 €9 55 55 Lv Ov 6 0 91 05 Sc LC Ly Se Lp OV 993 Lv Se 05 55 55 Lv CL Lv Se Lv 05 6 55 91 55 91 81 85 Sc 05 05 44 Lv 91! OV Lv 55 85 18 05 Lp 05 05 Le Ov Se 55 05 95 99 05 05 05 914 91 91 0€ 55 91 05 59 55 91 05 05 81 9 Se Sc Sc 81 Lv 55 Lp OV dy 55 Ly OV 05 55 SL €9 05 Ly Ir 0 Le 55 05 Lp Sc ЭМЗ Lv 05 05 eo 05 95 05 05 Lp 91! Gt 91 05 v9 OZ Sv Ov 05 05 Ir 59 0 89 9S 05 no 85 05 05 85 Lv 55 95 0 99 95 05 vs Ov Lv 05 05 Sv 55 05 9S 05 UN 85 05 05 85 Lv 0 9€ 0 Sl 09 05 Sy 9p 05 95 Ly 59 Se 8S 05 09 ши 05 05 05 99 05 05 0 Se Lp GZ 04 Sy 55 95 05 05 59 55 85 95 OZ JN ап AL 20 чо Ed UN м UN BW On) = eo 19939 199 ¿ITA Ga SnD 19, Many, ov san Le Oc 6I gst Zt 9+ St vl el cl LE 01 6 8 Z 9 5 v € С | -ads eJas!uy EE AA RS A A Е ‘эаенеле вер OU = ЭМ ‘| э1ае1 ul иэлмб $! зиоцеле.аде о} Áay '(зиэбцие) зэюэд$ ZG pue елэзцие |2 10} (ээцэлаир juaoad) вер мен “€ 3719VL 223 UNIONACEAN GENETIC RELATIONSHIPS Ov Ol 03 05 Ov 05 09 05 05 05 05 05 05 05 05 Ov 05 05 05 05 05 Ov Ov Ov 05 05 05 05 Ol 8S 6S 05 85 8S 99 05 85 99 99 GZ 8S 05 05 99 05 05 Lv 99 Lv 8S 05 05 99 05 05 Lp 05 85 18 69 CL 59 59 59 CL CL 59 59 vs 59 59 CL CL CL 59 CL 59 Cl CL 18 18 CL 59 59 59 cl CL 05 Lv Lv 05 05 05 Lv 05 Lv 8S 09 09 Lv 55 05 05 Lv Ly Lv Lv 05 Lv Lv 05 05 05 09 Ly 05 9 55 55 55 97 97 OV Ov 9+ 09 55 9 Ov 55 OV 9v 55 9 Ov 9 95 55 9 55 97 9 56 Oc 9c 55 91 55 91 Lp 95 05 95 Ly 95 Lv ON 55 ON Lv 55 Sc 55 Se ON 91! 55 91 55 Lv 55 55 Se Lv ON 8S 05 gl Sc 55 55 85 05 Lv Lv 8S Lv 55 55 55 55 Lv 55 Sc Lv Se 55 55 Lp Ly 55 55 55 55 Ge 05 05 Op 0€ N] ¡dy do 19Y 917 MY 97 пу div Iv ov qa Sd IA PA паз ed u] 11 n u7 07 $7 197 7 19 do 9d in 224 the relationships among taxa by using multi- variate analysis. Computations were made using the NT-SYS program (Rohlf et al. 1972) at the Uni-Coll Corporation, Philadel- phia, or via a remote job-entry station to the Sun Oil Corporation Computer (IBM 370/168 VS2). Initially, three types of matrices were used: 1) the Mainardi (1959) immunological distance was used as a distance coefficient, 2) the distance between taxa was used as a distance coefficient (Table 2), and 3) and OTU x antiserum matrix was made where the 52 OTUs were antigens (i.e. species or popu- lations of a given species) and the 21 antisera were treated as characters (Bashford et al., 1968). The percent arcs unique to the homolo- gous system = percent difference that was used as a distance coefficient (Table 3). In the first two matrices, comparisons were made where there was an antiserum for each species. In the analysis we standardized by rows (antisera) in order to produce a matrix of transformed distance coefficients. We em- ployed the minimum spanning tree (MST) and “subsets” components. Character correla- tions were subjected to Principal Component Analysis (PCA) with components extracted until eigen-values became less than 1.0. A transposed matrix of the first three PCs with their character loading was post-multiplied by the standardized matrix in order to yield a matrix of OTU projections in the PCA space (Rohlf et al., 1972). The resulting PCA-based configuration portrays distance-ordered rela- tionships well, but tends to distort close-rela- tive relationships, which often are of critical interest to taxonomists (Rohlf, 1970; Webster, 1975). OTU locations in the 3- DAVIS AND FULLER dimensioned PCA space were used as the initial configuration for a nonmetric multidi- mensional scaling (MDS) placement of taxon- omic distances between OTUs (Kruskal, 1964). OTU configurations were adjusted after scaling by PCA analysis on а variance- covariance matrix obtained from the MDS- coordinates in order to realign the major trends of the variation in the reduced configu- ration space with the coordinate axes, while maintaining the accuracy of between-OTU distances in the ordination space (Rohlf et al., 1972). Distances between OTUs in the PCA- and MDS-spaces were found and compared with the matrix (cophenetic) correlation coef- ficient. We placed no reliance on cluster analysis and phenograms to illustrate relationships. We emphasized ordination and MDS that are freed from the constraints of phenogram con- struction. Comparison of taxa Experiments were conducted by using foot muscle antigens in order to determine to what extent we could find differences between unionids and non-unionacean clams and be- tween different conspecific populations of the same species. Results would be important bench marks for assessing differences be- tween species. We also compared different populations of the same species by using glochidial antigens. We compared three species of unionids with three species of marine bivalves by using the foot-muscle system. The comparisons in- volved four different taxonomic orders (Table 4). TABLE 4. Classification according to Morton (1971) of marine species compared with the unionacean species used to test immunological congruity on the basis of foot-mussel antigens. A xq _ _ E __ ____$»>-_e yp_EEE__R—R q qv»>»xIIEOEOEEO EN m ee ee eng Class Bivalvia Order: Anisomyaria Superfamily: Mytilacea Genus and species Order: Schizodonta Superfamily: Unionacea Genus and species Order: Heterodonta Superfamily: Veneracea Genus and species Order: Adapedonta Superfamiy: Myacea : Geukensia demissa (Dillwyn) : Anodonta cataracta Say Elliptio complanata (Lightfoot) Elliptio icterina (Conrad) : Mercenaria mercenaria (Linné) Genus and species: Mya arenaria Linné OO o pp II A a UNIONACEAN GENETIC RELATIONSHIPS 225 Annotations and terminology Annotations, indicated by superscript num- ber with a taxon, are given in numerical order in Appendix 2. Definitions of terms concerning breeding season and marsupial conditions are given in Appendix 3. RESULTS Comparison of unionids with marine species As is seen in Table 5, only one or two anti- gens were shared in common by unionids and TABLE 5. Congruity of marine and unionid species in percentage differences. Antisera Marine species Ac? (10) Ec (12) Ei (12) Mercenaria mercenaria 80 92 84 Geukensia demissa 80 92 92 Mya arenaria 80 92 92 ( ), number of preciptin arcs. Coded names given in Table 1. TABLE 6. Comparison of different populations of the same species on the basis of glochidial and foot-muscle antigen-antisera systems. Percentage difference is given. Gravid Gill Antisera Populations Ec2(12) El (14) Ec3* 33 — El2 — 21 El3 — 14 Foot-Muscle Antisera Populations Ac? (10) Ec 0 Ec3 0 Ec4 0 — — 0 Ec5 Ac EI2 — 0 г = 0 El3 — — 0 ( ) number of precipitin arcs. “coded names given in Table 1. species of other bivalve orders. Accordingly, the immunological comparisons among unionid taxa involve antigens that are pri- marily (> 85%) unique to the Unionidae. Comparisons among populations of the same species К was possible to demonstrate 14% to 33% difference among populations of the same species by using glochidial antigen-antibody systems; it was not possible to discriminate among populations of the same species by using foot-muscle systems (Table 6). Be- cause the foot-muscle systems were the more conservative, investigations reported here were based on foot-muscle systems. Differ- ences among taxa are differences above the conspecific population level. Comparing species by using foot-muscle antigens An initial multivariate assessment was made where а comparison involving an anti- serum for each species was possible. This initial comparison involved 14 species. We used the Mainardi (1959) immunological dis- tance and the average percent difference as distance coefficients. We abandoned use of the Mainardi distance coefficient because the results of ordination by using this distance were more distorted (r = 0.809) than results with the average distance (Table 2, г = 0.922). The results of ordination based on the aver- age distance and the first two principal com- ponents are given in Fig. 1. The first two com- ponents accounted for 89.50% of the data. The correlation between the matrix of taxon- omic distances and distances in the 3-dimen- sional MDS was excellent, ¡.e., 0.922; the stress was 0.213. As can be seen from Fig. 1, there are three widely separated groups of taxa: 1) species considered on classical grounds to be Margaritiferidae (i.e., species of the nominal genera Margaritifera and Cumberlandia), 2) the single species of Anodonta, and 3) the cluster including Elliptio, Fusconaia, Megalonaias, Proptera, Quadrula, and Vil- losa. C. monodonta and Margaritifera mar- garitifera are in the same set; Cyclonaias tuberculata and Quadrula cf. Q. quadrula are in a set and more closely allied to each other than either is to Q. cylindrica. Megalonaias gigantea and F. masoni are т a set. Elliptio and Fusconaia are closely associated. 226 ANODONTA CATARACTA +0.6 Q77Sp! +0.4 VILLOSA DAVIS AND FULLER +0.8 CYCLONAIAS TUBERCULATA - 10 | ELLIPTIO BUCKLEYI = 11 FUSCONAIA - | Е MASONI MEGALONAIAS PROPTERA QUADRULA SP. GIGANTEA - 9 ALATA = ih CYLINDRICA - 5 8 DELUMBIS - 13 MARGARITIFERIDAE: 1-4 4 С. MONODONTA M. MARGARITIFERA + 9 | | -0. N FUSCONAIA: 6,7 | ; | 0.8 M. HEMBELI M. FALCATA FIG. 1. Ordination diagram in two dimensions showing relationships among 14 taxa via use of the minimum spanning tree and subsets. The data are based on cross comparisons using 14 sets of antisera and antigens. See text for details. Results of comparing 52 OTUs x 21 anti- sera are shown in Fig. 2. The ordination in- volving the first two principal components ac- counted for 92.84 percent of the data. The correlation between the matrix of taxonomic distances and distances in the 2-dimensional MDS was excellent, ¡.e., 0.946; the stress was 0.205. We again found three widely separated groups of species. These groups are essen- tially those seen in Fig. 2: 1) the Margaritifera group (quadrant |), 2) the Anodonta group (quadrant IV), and 3) a large mass of taxa linked together primarily in quadrants ll and Ill. A series of subsets encloses the species in the Margaritifera cluster. Cumberlandia monodonta is in a subset with M. hembeli; those two are in a large set with M. margariti- fera. А! four species are clustered in an in- clusive set. In the Anodonta cluster A. cataracta is in a subset with A. imbecillis. A. implicata is in another subset with Lasmigona costata. Because so many taxa are grouped in the third cluster and because this cluster is so distinct from the Anodonta and Margaritifera groups, we did a multivariate analysis of only those species and corresponding antisera from that third cluster. This involved a subset of the database (Table 7) of 15 antisera and 40 OTUs. We omitted data for Quincuncina infucata because we had too few data for this comparison. In this reduced set there were only 19 comparisons of 600 for which we had no data. The results of ordination involving the first two principal components are shown in Fig. 3. Only 78.97 percent of the data are represented. The correlation between the matrix of taxonomic distances and distances in the 3-dimensional matrix was 0.913; the stress was 0.378. Three closely allied clusters are seen (Fig. 3) in quadrants |, И-Ш, and IV. Genera within these clusters are listed in Table 8. Two of the genera are found in two clusters: Amblema and Fusconaia. A. plicata is in cluster 1; A. UNIONACEAN GENETIC RELATIONSHIPS 227 I .8 .7 .6 .5 ET «Ес SS 4 / \ > à x = 3 № Pi h 2 р Mg\\ Ра Ар! La Fir y Qbb À И 1 Ei, нА |9 AE АСР Ps} Vi Ct s \ Lre | 1 A: 2 Gr 4 N Plpu 1 2 N — YE Lh Ln Е: ь TR Срд \ EDR AUX Eu AS x A AMBLEMINAE I .8 MARGARITIFERINAE N FIG. 2. Ordination diagram in two dimensions showing relationships among 52 taxa via the minimum spanning tree and subsets. The data are the percent difference when 52 taxa were analyzed using 21 sets of antisera. See text for details. Abbreviations are explained in Table 1. perplicata, cluster 2. Three species of Fusconaia are in cluster 1; one, in cluster 3. Four species of two genera are emphasized in Table 8 because the species are not grouped in the same major cluster. Distance coefficients (d.c.) are given for those three species that are the least distant from each of the species in question. The closest relation- ship of A. plicata is with Elliptio buckleyi (d.c. = 0.620), its next is to Plectomerus dombeyanus of the 2nd cluster (d.c. = 0.97), and the third is to A. perplicata (2nd cluster; d.c., 1.095) or Glebula rotundata (3rd cluster, d.c. = 1.095). Note that in Fig. 3 the MST ties Glebula and Plectomerus together and A. perplicata to Plectomerus. On the basis of the interrelationships that revolve around Plectomerus, it is appropriate to consider the two species of Amblema to be part of cluster 2: The situation with Fusconaia ebena of clus- ter 3 is different. Closest relationships of F. ebena are with Lampsilis, Ptychobranchus, and Glebula, all of cluster 3. F. flava of cluster 1 has its closest relationship with Elliptio of clus- ter 1 (Table 8). We suspect experimental error in testing relationships with F. ebena (see Appendix 2, point 1). Таха that are grouped in subsets are listed in Table 9 with the taxonomic distances among them. There are only seven such sub- sets, and only two of these are found in clus- ter 3. The closest relationship in a subset in- volves Plectomerus dombeyanus and Megalonaias gigantea (d.c. = 0.588). By use of the MST, groupings of taxa are shown that suggest relationships (Fig. 3) that will be discussed later. 1) In cluster 1 there are no clear associations of classically de- fined species of Elliptio or Fusconaia that re- sult in an Elliptio cluster or a Fusconaia clus- ter. 2) Plectomerus, Megalonaias, and Amblema form a grouping in cluster 2. 3) Quadrula, Tritogonia, and Cyclonaias are linked in cluster 2. All species of Lampsilis are linked in cluster 3, and Villosa and Proptera 228 DAVIS AND FULLER TABLE 7. Raw data (percent difference) for 15 antisera and 40 species (antigens). Key to abbreviations is given in Table 1. NC = no data available. Spe- cies Ct Eb Ecr El РБ ЕТ! Ga Са 41 МС 41 50 33 27 0 Ар 33 18 25 25 26 18 30 Fe 50 36 33 25 40 27 50 Fi 25 18 33 16 33 9 50 РЫБ 25 27 50 16 0 9 40 Em 25 36 41 16 26 0 50 Qbb 8 27 41 16 20 27, 50 Qa 8 36 50 41 33 27 50 Qc 16 45 41 25 20 45 40 Fd 16 27 41 25 33 27 40 Ту 25 27 50 25 40 45 40 Mg 16 27 50 33 26 18 50 Ct 0 36 41 25 20 27 30 Ec 25 18 33 8 20 9 30 ЕЁ 25 18 33 0 25 27 40 Er. 16 0 33 8 33 18 30 Ем 41 18 33 25 33 18 50 Ecr 16 27, 0 8 33 27, 50 Ut 25 45 16 41 26 27 20 Pc 33 № 16 25 20 0 40 Ср 33 45 41 33 33 36 30 Gr 33 18 41 33 26 36 50 Lt 33 2 41 41 46 45 40 Lei 41 36 33 33 33 27 50 [$ 41 36 33 16 26 36 50 Lo 33 27 41 33 33 36 40 Lh 33 36 33 16 26 36 30 Antisera Lo Mg Pa link to L. radiata. 4) L. radiata is central in the Lampsilis cluster. 5) Gonidea is a distinct subgroup of cluster 3, far removed from other species of that cluster (d.c. = 1.068). DISCUSSION Evolutionary trends: primitive and derived character-states In the evolution of freshwater bivalves, the ecological transition from the sea through estuaries into rivers necessitated the survival of larval forms. The free-swimming veliger larvae had to be retained in the parent in order to prevent their destruction by being swept downstream or by osmotic shock. Two strategies evolved to accommodate retention of the veliger larvae; brooding young to the juvenile stage and brooding young to an early pre-pediveliger parasitic state. The two larval retention strategies correlate with the size of the breeding adult for reasons documented by Hoagland (1975). Where one UNIONACEAN GENETIC RELATIONSHIPS 229 TABLE 8. Listing of genera in each of the three clusters shown in Fig. 3. Distance coefficients are given showing the three species closest to species of those two genera apparently located in two different clusters. Coded names are given in Table 1. Distance coefficients ik 2. 3. Quadrant | (cluster 1) Amblema plicata Eb (0.620) Pd (0.970) Apl (1.095) Gr (1.095) Elliptio Fusconaia flava Pleurobema Uniomerus Quadrant IV (cluster 2) Amblema perplicata Cyclonaias Megalonaias Plectomerus Quadrula Tritogonia Ea (0.816) Е! (0.850) El (0.870) Pd (0.893) Qbb (1.042) Ct (1.042) Quadrants II-III (cluster 3) Actinonaias Carunculina Fusconaia ebena Glebula Gonidea Lampsilis Ligumia Leptodea Ptychobranchus Proptera Villosa Lel (0.993) Ps (1.005) Gr (1.125) TABLE 9. Taxa included in the smallest subsets together with their taxonomic distance coefficient. Ranking is by lowest to highest taxonomic distances. The smaller the distance, the closer the relationship. Taxonomic distance Species pairs 0.588 Plectomerus dombeyanus х Megalonaias gigantea 0.620 Amblema plicata x Elliptio buckleyi 0.631 Elliptio icterina x E. lanceolata 0.658 Ligumia nasuta x Actinonaias carinata 0.697 Lampsilis radiata x Ptychobranchus subtentum 0.777 Fusconaia flava x Elliptio waccamawensis 0.870 Quadrula pustulosa x Q. cylindrica niche dimension is small body size, a propor- tionally small amount of energy is available for reproduction, few young are produced, and these are brooded to the juvenile state; there is a high probability of individual survivorship of the young. When body size is large, repro- duction is delayed until large body size is at- tained, and then proportionally large amounts of energy are available to produce numerous young that are released at an early larval stage; there is a low probability of survivorship to the young. Native freshwater bivalves of North Amer- ica are of two types: Unionacea, which are 230 DAVIS AND FULLER I „— ~ GONIDEINI e ES e A ARES: a Se 7 N 2 \ a A о CLUSTER 3 LAMPSILINI AMBLEMINI 0 3 / / Y N / / N и CLUSTER 2 FIG. 3. Ordination diagram in two dimensions showing relationships primarily among those taxa clustered in quadrants II and Ш in Fig. 2 using only antigens and antisera pertaining to those taxa. Relationships are clarified by use of the minimum spanning tree and subsets. See text for details. Abbreviations are explained in Table 1. large-bodied (most adult shells exceeding 8 cm length), and Sphaeriacea, which are small-bodied (adult shells usually are less than 12 mm in length). Sphaeriidae brood their young to the juvenile stage, whereas Nearctic Unionidae typically brood young to the glochidial (parasitic) stage. The Unionidae are sedentary as adults, and numerous spe- cies often live sympatrically side by side in the same river bed. Because of this sedentary life and the brooding of the young, it is not sur- prising that morphological characters serving to distinguish among species are few and that those soft-tissue characters that are useful in- volve structures of the demibranchs for hous- ing the brooding young and structures of the mantle margins and pseudosiphonal regions, which interface with the aquatic environment for the purposes of pumping water and food into the animal, expelling water and waste, and getting the glochidia to the appropriate host. Morphological character-states were con- sidered primitive when they represented the simplest condition. Derived morphological character-states are those showing increased organization, complexity, and specialization. We follow Ortmann (1910a, 1911, 1912b), Heard & Guckert (1971), and Heard (1974) in considering Margaritifera to have the most primitive groundplan of all unionaceans. We suggest that taxa with this type of groundplan probably gave rise to all other Recent unionaceans. Primitive character-states are designated with “P” in tables 10 and 11. The most de- rived states are designated “S”. The consid- eration that the direction of evolution is from primitive to derived as defined here is con- sistent with the facts that Margaritifera is ancient, known from the Cretaceous, and has a Holarctic distribution, including representa- tion in Southeast Asia (Heard, 1974; Smith, 1977). Representative of the most derived and specialized taxa, Lampsilis is known from the Oligocene and is endemic in North Amer- ica. In Margaritifera, demibranch lamellae are held apart by randomly arranged trunks of interlamellar connective tissue. The ctenidia UNIONACEAN GENETIC RELATIONSHIPS 231 TABLE 10. Morphological character-states serving to define subfamilies of Unionidae employed in this paper. Margaritiferinae *1. № true septa—P *2. No water tubes—P *3. Excurrent aperture entire—P *4. Diaphragm grossly incomplete—P *5. No additional connective tissue at distal margin of marsupial demibranch—P *6. Glochidia with irregular teeth *7. Glochidia without numerous spines—P 8. Glochidia subspherical *9. Glochidia smallS—P 1. With true septa (parallel to gill filaments)—S *2. With water tubes tripartite—S 3. With supra-anal opening—S 4. Diaphragm slightly incomplete—S *5. Additional connective tissue at distal margin of marsupial demibranch—S *6. Glochidia with hooks—S *7. Glochidia with numerous spines—S *8. Glochidia subtriangular *9. Glochidia large 5—S Ambleminae . Water tubes present, not tripartite Diaphragm slightly incomplete * © © O O1 PB U D . Glochidia without hooks** or teeth . Glochidial shape variable . Glochidia medium sized5 . Glochidia without numerous spines . With true septa (parallel to gill filaments) With supra-anal opening, but excurrent aperture sometimes entire No additional connective tissue at distal margin (= ventral margin) of marsupial demibranch * most distinguishing character-states ** except Proptera 5, see Appendix 2 P, Primitive S, derived, specialized thus lack water tubes, and the eggs and/or larvae are incubated in a flaccid sac. All four demibranchs are marsupial. When feeding and respiring, the animal exhibits a wide даре between the posterior ends of the valves, which leaves the soft tissues within vulnerable to disturbances from without. Associated with this gape is an extraordinary development of muscular arborescent papillae at the incurrent mantle aperture. Margaritifera lacks a sepa- rate supra-anal opening (i.e. there is no sub- division of the excurrent mantle aperture by fusion of the opposing mantle margins), and there is no clear demarcation of the anterior boundary of the incurrent aperture. Finally, at the posterior end of the gills, the diaphragm is incomplete and formed only by the ctenidia. The glochidia are tiny (about 50 um long) and hookless (Baker, 1928). The sac-like marsupia, tetragenous condi- tion, and posteriorly gaping valves are condi- tions associated with low species diversity and ecological restrictions to streams of peb- ble-cobble substrate with rapid flow of highly oxygenated water. That sac-like marsupia in- volve all demibranchs means that gravid gills are loaded with eggs and embryos, a condi- tion that must interfere with respiration. Be- cause there is little supporting storage struc- ture within the gill to assure efficient packag- ing and protection, there probably is some vulnerability of the eggs and embryos to mechanical damage, especially as the poste- rior animal, including the gills, is exposed due to the wide shell gape and the lack of mantle sutures helping to protect the posterior region from the outside environment. Morphologically derived character-states in other unionids involve increasing complexity of the interlamellar gill tissue and modifica- 232 DAVIS AND FULLER TABLE 11. Morphological features characterizing the four tribes of the Ambleminae employed in this paper. Gonideini * O OP © D = . Perforated septa—P . No specialized mantle structures—P Shells smooth—P Lampsilini 1. Ectobranchous—S . Septa not perforated—S . Tetragenous (mostly or all—P) or ectobranchous (perhaps some)? . Marsupia not confined to restricted regions of the demibranchs—P . Marsupial water tubes do not extend beyond distal margins of demibranch lamellae—P . Marsupia confined to restricted region of the demibranchs—S . Marsupial water tubes extend beyond distal margins of demibranch lamellae—S 2 3 *4. Many taxa with specialized mantle structures (flaps, caruncles, etc.)—S 5 6 . Shells mostly smooth Pleurobemini Septa not рейогаеа . No specialized mantle structures * DONADN= . Shells smooth Amblemini . Septa not perforated** . No specialized mantle structure пром — . Ectobranchous (mostly) or tetragenous . Marsupia rarely confined to restricted regions of the demibranchs Marsupial water tubes do not extend beyond the distal margins of the demibranch lamellae . Tetragenous (mostly) or ectobranchous . Marsupia not confined to restricted region of the demibranchs . Marsupial water tubes not extending beyond distal margins of the demibranch lamellae *6. Shells heavily sculptured (few exceptions)—S * distinguishing character-states(s) ** except for marsupium of Megalonaias P, primitive S, derived, specialized tions of the mantle margin. Also, there are trends of reduction in the number of marsupial demibranchs and development of specialized regions of the gill for incubation of young. There are several derived character-states of great importance. First, the scattered margaritiferoid interlamellar connectives were increased numerically. The advantage of more connectives probably is to increase the internal strength and stability of the gill and, therefore, the safety of its contents. Second, the connectives were organized into continu- ous walls (septa) that served to define and separate linear series of adjacent water tubes within the gills. Septa are perforate or im- perforate. The perforate condition probably 1$ the more primitive (Heard, 1974). Septa prob- ably greatly increased structural support for the gills. Third, septa were aligned “vertical- ly,” ¡.e., parallel to the gill filaments. This ver- tical attachment along the filament strength- ens the septum, and, as a simple exercise in geometry will demonstrate, less space within the demibranch is needed for parallel orienta- tion of septa and filaments than is occupied by identically spaced septa oriented obliquely to the filaments. A reduction in space occupied by interlamellar tissue presumably would facilitate gas exchange. Fourth, a portion of the gill was set aside as a permanently modi- fied marsupium whose interlamellar septa be- came thicker and more closely spaced than those in non-marsupial parts of the gill. This further reduction in the extent of the marsupium presumably facilitated respiration additionally. Indeed it may have been neces- sitated by proliferation of interlamellar tissue, at least in the case of very active mussels. There 15 a strong association between ге- duced marsupial size and the need for energy (and thus for oxygen). In what we regard as the most advanced Nearctic unionids (i.e., UNIONACEAN GENETIC RELATIONSHIPS 233 Lampsilis and its allies) are found the least extensive marsupia. These are consistently the most active of mussels, not only in terms of locomotion, but also because of the move- ments of specialized structures on the post- basal mantle margin (flaps, caruncles, etc., which are important in reproduction). In any case, thicker, closer-spaced septa in the marsupium strengthen it further and thus pro- vide added protection for its contents (eggs and/or larvae). These developments were accompanied by modifications of the bivalve hydrodynamic (water pumping) system. Modifications ap- parently were necessitated because develop- ment of the vertical water tube meant an in- crease in the distance that larvae would have to travel in order to escape the marsupium to the external environment; marsupial contents vertically evacuate the water tube and then perpendicularly traverse the excurrent pallial chamber before emission to the waterway through the excurrent mantle aperture. The necessary increase in propulsive hydrody- namic pressure was created by realizing or at least approximating a “closed” hydrodynamic system within the adult female mussel. Sev- eral devices were possible: stronger muscular adduction of the valves, close fit of the valves, increased fusion between apposing mantle margins, and/or posteriad extension of the diaphragm. Morphology, immunology, and a new classification The ordination diagrams (Figs. 2, 3) with MST and subsets indicate the classification given in Table 12. For three reasons, we argue that there are one family and three sub- families. First, we see only two directions of morphological change from the primitive Margaritifera type, i.e. to the derived Anodonta and Lampsilis types. These are progressive changes within a single morpho- logical groundplan. Few morphological changes, involving increased complexity, are needed to progress from a Margaritifera-type morphology to an anodontine type or to an amblemine type. We do not see abrupt dif- ferences among the three groundplans such as exist between the marine Cardiidae and Tridacnidae of the superfamily Cardiacea, for example, or between tne marine families Pteriidae, Malleidae, and Pectinidae of the superfamily Pteriacea (see Yonge & Thomp- son, 1976). Second, immunologically there are three distinct clusters, which correspond to the three morphologically defined Nearctic groups within the unionid morphological groundplan; the Margaritifera-type, the amblemine type, and the anodontine type (Tables 10 and 11). (Cladistic relationships among these types will be presented later.) We believe that immunologically, as well as morphologically, the three groups have equal weight. They might be interpreted as three families or as three subfamilies of a common family. Of all the antigens discovered during our analyses only one or two were not unique to the freshwater mussels we used. This sug- gests strong immunological cohesion of this group. The average genetic distances among the three mussel subgroups were close to 50%. This reinforces the conclusion (above) that the three groups are not far apart geneti- cally. Therefore we conclude that the three groups are best regarded as subfamilies within one family. The taxonomic results are Unionidae: Margaritiferinae, Anodontinae, and Ambleminae. The greatest difference is between the Anodonta and Margaritifera groups; the least between the Amblema and Margaritifera groups. What is the relationship between immuno- electrophoretic genetic distance, as present- ed here, and taxonomic hierarchy? The rela- tionship is not a simple one; there is no direct correspondence. Classifications traditionally have been based on comparative morphol- ogy. Increments of change in the taxonomic hierarchy follow discrete changes in morpho- logical groundplans. Pronounced changes in morphology and behavior can occur rapidly with respect to geological time (Stanley, 1979). These changes, presumably under the control of regulatory genes, may involve few genetic changes involving regulatory genes, yet be pronounced enough to impress taxo- nomists that the taxon in question belongs in a different higher-category taxon from that of the taxon most closely related to the one in question. Such morphological change may not be accompanied by an equal amount of change in structural proteins. The now ciassic example is one involving man and chimpan- zee. These animals are classified in different families on the basis of considerable morpho- logical and behavioral divergence. However, the molecular genetic distance between man and chimpanzee is very small, essentially equal to the genetic distance among sibling DAVIS AND FULLER 234 "suoyejouuy ‘с xipuaddy ээс 9'£'z idiosiadns "LONBINNSSEIO Jan 3 SIAEQ ay, ul диэшеэв9 эмзиэб Buuepisuos pue ‘рэдпо.б uaaq sey и yoıym цим елэиэб Jayjo eu} Buuepisuos иохе} oueueBeudns juasayip e o] бибиоеа snuab ce ‘зиезэр 10} Xe} ass ‘0} aalbe jouueo эм иохе} e 10 ‘UOXE} JNO о} juajeamba jou si yey) uoxe] эзэцабелап$ ‘| ‘Saqli} JO Jaquunu ‘99 ‘1 ‘зэнишие}апз jo saquinu ‘‘эе ‘| ‘за!ише} “019 y A дд ————_—_—_—____ BJOBIEIEO uopouebAY Joeunuapnosy X виороиу Jaeunuopouy x] в/иобо]111 вшэипэито в/плрепо sereuojebayy seleuojoAg фэвеиипарепо ИЛ Jeepluolun ‘9 sn/au10]99/d 2uwa/quy {эемшаашу ‘Л se/euado4 5плэшоший 0101/13 Jeeuruondi]3 ‘|A BWAQOINI|d BIEUOISNZJ {эешшэдолэ ‘A aeuisdwe7 Al euobiwse7 EJUOPIWSEIW Jaeunuopiuwsejy 111 faepiuondi|3 'g E9PIUOH {эеициорорпэ$ "|| эвииалуиеблеи\ '| Jaepuanuebien y (2961 ‘676! 'Zr6L) ISPON 5е/гио/ебаи/ faeuipereuojebayy ‘Л eıuobozu] виэипэито в/п.репо SN/9U0]99/d „EIEUOISNZ вша/дшу faeuimajquy ‘|A Joeulepiuos :A +oepiwaquy ‘9 »/Á9/y9NQ sereuadog Joeuipeieuadog 'Al snsawoiun ewaqoinald 0101/13 xSPIBUOIDAD | эешедо.па|а ‘||| seulisdue7 ‘| aeunuopouy | Jaepiuolun 'g +oeunpuejlaquino ‘|| эевииэлуилеблеи\ || sepuaynueBien y (1/61) нэхопэ 3 pueay 2!/иобо]и1| виэипэито в/п.репо snwÄuoyuo SE/PUO/2Â9 SNIBWO}IA/g se/euo¡ebay euwaquy «SNJIWOIUN) ~PWAQOINAald „BIEUOISNZ «0101/13 +3euluolun “A Jaeulapivoo Al faeurjisdwe7 ‘|| aeunuopouy ‘|| эециэциеблеи\ | аэершоип * (9161 '*0161) чиешио (4aded siy)) олэп- 3 Sineg в/иобоу11 севизипэито в/плрепо 20 =) snuAuoyuo (SNJALWOJDA|H ‘see SEIPUO/9Â9 -uojebay +) ewejquty uwwagwy ‘+ гплэшоий 2uaQqo/nald BIEUOISNZ ondı13 ее ПЕ € E9PIUOH IUI8PIUO) ‘2 PSOJJIA SNYIUBIGOYIA}Y e/9]d0/4 enunbr7] espojda7 sıısdwe7 e/nga/9 euljnounseg SP/BUOUNOY usdwe7 ‘| aeul|isdwe7 ‘||| ¡UIWa¡quiy ‘+ |UIWAQOINa|q ‘€ IUI8PIU0) ‘г ¡unisdwe7 ‘| эемшеашу ‘||| euobiwse7 EJUOPIWSE/W виороиу aeunuopouy (|| aeunuopouy ‘|| (eıpueuagwndg +) elanuebiey aeULayWebey | SPUIJSJNUEDIeN | эершоий y эершоший y 98/61 ‘IE Ja sineg EXT (8/61) ‘1e 19 sineg jo Анше}ап$ Jo aqu ayı 0] Juajenınba si sone ue yo Аишезап$ au] и раз! Jou эле елэиэ9 ‘зээцалашр pue зэциенилв Buimoys зиоцеэциз ею Jofeu jo vosuedwon ‘21 3181 UNIONACEAN GENETIC RELATIONSHIPS 235 species of other organisms (King & Wilson, 1975). However, there 15 strong evidence that, once lineages have begun to diverge and are reproductively isolated, there 1$ a rather regu- lar increase in molecular genetic distance with increasing time (Fitch, 1976; Sarich, 1977). Further, there seem to be rapid and slow rates of certain loci in protein evolution that can be detected electrophoretically. The contribution of rapidly evolving loci should be completed in some six million years, while further increases in genetic distance are contributed by the slowly evolving loci (Sarich, 1977). Given the divergence of Margaritifera, Anodonta, and Ambleminae at least by the late Cretaceous (i.e. 60 million years ago), one would reason- ably expect considerable genetic distance among these naiad taxa. In the few immunoelectrophoretic studies such as this one, species of different gastro- pod families, but of the same superfamily, have differed from 40% to 80% and most have differed by 50% to 80% (Davis & Suzuki, 1971; Davis, 1978). Using allozyme electro- phoretic analyses, and considering changes in | between different hierarchial levels in other organisms (Davis et al., 1980) we have found that genetic similarity between Margaritifera and our Ambleminae was greater (| = 27%) than we would expect if these taxa belonged to different families. In summary, given the 47% immuneolectro- phoretic genetic distance between our Margaritiferinae and Ambleminae and con- sidering the great age of divergence of these taxa, we think it reasonable to consider them to belong to a single family, the Unionidae. This is supported by the cohesiveness of the unionid morphological groundplan, which in- cludes a single larval type (the glochidium). The three subfamilies are cohesive and distinct immunologically and morphologically. Morphological character-states that aid in dis- tinguishing among these taxa are marked with an asterisk in Tables 10 and 11. The Margaritiferinae have been discussed above in terms of character-states that have been considered primitive and that in some cases are unique to taxa of this subfamily. The Anodontinae are defined, in part, by unique derived character-states, indicated in Table 10. These states have been discussed previously, for the most part (Ortmann, 1910b, 1911, 1912b; Heard, 1975). Known Anodontinae have an extraordinary type of glochidium, whose shell is subtriangular in lateral outline, usually large and powerful, and armed with spined hooks at the apex of the ventral margin. The powerful, well armed anodontine glochidium can sever soft host tissues (e.g. the gill filaments of fishes) and prematurely fall from the host to the stream- bed, where it will die. On the other hand, this type of glochidium fares well on tougher tis- sues (e.g. the fins and even the scales of fishes) and thus occupies a metamorphic niche that has been rarely exploited and perhaps chiefly vacated by weaker types of unionid glochidia. The anodontine marsupium exemplifies Simpson's (1900, 1914) group Homogenae, yet it is of a unique type. During gravidity a marsupial water tube adopts a tripartite con- struction: secondary septa develop parallel to the inner and outer demibranch lamellae. The object presumably is to facilitate additional gas exchange for the marsupial contents. The dorsal margins of gravid water tubes are sealed by a film of connective tissue, whose purpose presumably 1$ prevention of the escape of eggs and the premature escape of larvae. With one known exception, the egg mass is loosely structured. Premature loss of ova and larvae to the excurrent mantle cham- ber would be a great threat were it not for the dorsal tissue. Gravid anodontine marsupia are greatly and uniquely swollen. This 15 facilitated by another unique feature, the development along the distal margins of the outer demi- branchs of additional connective tissue during gravidity. This device permits the apposing lamellae to separate and move apart. The en- tire phenomenon necessitates the presence of the secondary water tubes (above) and probaby 1$ a response to the need for space in order to incubate competitively large numbers of offspring, which are themselves exception- ally large, as noted earlier. The typically loose structure of the egg and larval masses is caused by the great size of the marsupial contents, which cannot pack together so securely as can those small ova and glochidia of other unionid groups. On the other hand, the loose mass is an advantage because the adult does not have to overcome the inertia of a large mass during expulsion of marsupial contents. Instead, the glochidia can be pumped out singly or in small numbers. The advantage is of even greater value for very thin-shelled, low-density species, whose poor valve adduction and fit weaken the hydrodynamic system. The Ambleminae are like the Anodontinae in that the water tubes parallel the gill fila- 236 DAVIS AND FULLER ments; the posterior mantle is not united, but drawn together by the diaphragm, to effect functional separation of incurrent and excur- rent apertures; the excurrent aperture is closed above, which effects a supra-analopen- ing separate from an anal one; and the dia- phragm is almost complete and is formed en- tirely by the ctenidia. These two subfamilies differ in that most Ambleminae: Amblemini and some Ambleminae: Pleurobemini are tetragenous, whereas Anodontinae are ecto- branchous; amblemine glochidia are hookless and are of various shapes and sizes but never so large as the anodontine glochidia or hooked in the same way; and amblemine water tubes are undivided at all times. We distinguish four subgroups of Amblemi- nae: Gonideini, Lampsilini, Amblemini, and Pleurobemini (Table 11). There are few unique morphological character-states serv- ing to define these tribes. The Gonideini have perforated septa and are tetragenous; the Lampsilini have specialized marsupial (in most taxa) and postbasal mantle modifica- tions (many taxa) and are ectobranchous; the Pleurobemini have smooth shells and no spe- cialized marsupial features or postbasal man- tle structures and mostly are ectobranchous; and the Amblemini have heavily sculptured shells and mostly are tetragenous. We have not employed the subfamily name Unioninae because we do not know how Unio $.5. Of Europe relates to North American Unionidae. Ortmann (1912b) did not consider Unio s.s. to be equivalent to North American taxa that have similar morphology of shell and soft parts. Heard & Vail (1976) provided an excellent account of the morphology of Unio. Unio differs from our Ambleminae by having glochidia with hooks, perforated marsupial septa and imperforated non- marsupial septa, subtriangular and medium- sized glochidia. Unio and allied taxa have a smooth shell, undivided water tubes, and ectobranchous marsupium. The genetic rela- tionship of Unio s.s. to our Ambleminae and Anodontinae must be determined before we can consider the use of this taxon name for a higher-category taxon below the level Unionidze. However, the groundplan of the Eurasian Unio is generally so like our Ambleminae: Pleurobemini type that we pre- serve the traditional usage of the name at the family level. Comparison of classifications Eleven classifications are given in Appen- dix 1. These classifications represent three very different types of approach, and much is to be learned from their study. An historical account of unionid classification is given in Appendix 4. The three approaches that have been used are based on 1) conchology only, 2) selected use of a few key characters, 3) use of all data available and asking questions about how character-states evolved and about the relationships of character-states to environments. Of all the classifications prior to our own, only Heard & Guckert (1971) clearly stated the basis for their classification. Ac- cordingly, we will first contrast our classifica- tion with theirs and, by so doing, address vari- ous problems raised т the Introduction. We reject the Heard & Guckert classifica- tion for three reasons. First, they used the character “number of marsupial demibranchs” to establish families. They used the Amblemidae to accommodate all non-margaritiferine unionids with four marsupial demibranchs; those taxa with two outer demibranchs marsupial were consid- ered Unionidae. On the basis of immuno- logically derived relationships, it is clear that the tetragenous condition has undergone parallel evolution and that reduction to two marsupial demibranchs has occurred at least four and possibly five times, i.e. at the origin of the Anodonta clade, during the evolution of recent Ambleminae, and within the lineages of the Amblemini and Pleurobemini; it possi- bly has occurred in the Gonideini. It should not be surprising that reduction of marsupial demibranchs occurred several times once greater efficiency in the hydro- dynamic system had been achieved. Any reduction in the space taken up with marsu- pial function would mean increased efficiency in respiration. Specializations in how glochidia are incubated and delivered to their hosts cor- relate with different reproductive strategies. Second, they created two new subfamilies on the basis of length of breeding season: the Megalonaiadinae and the Popenaiadinae within their families Amblemidae and Unionidae, respectively. The new subfamilies contained only taxa that are bradytictic, whereas the other subfamilies of their Amblemidae and Unionidae are tachytictic. An examination of Fig. 3 clearly shows that there is no separate clade that separates Megalonaias from other taxa or so-called Popenaias buckleyi from its congeners (Elliptio). Heard & Guckert (1971) created their new subfamily concept Popenaiadinae primarily because of information about E. buckleyi (see Fuller, 1975). We do not, how- ever, have biochemical data for Popenaias UNIONACEAN GENETIC RELATIONSHIPS 237 роре! (Lea), so this subfamily has not been fully invalidated (but see Fuller, 1975). It clearly is unacceptable to create higher taxa on the basis of length of breeding season. As with the question of how many demibranchs and water tubes bear glochidia, the question of length of breeding season seems more correlated with reproductive strategy than with diverging clades. This character (length of breeding season) in most cases should not be used to assess taxonom- ic relationships among unionid taxa. Third, they created the subfamily Cumber- landiinae to provide a subfamily for the monotype Cumberlandia monodonta (Say), which 15 the only margaritiferine sensu nostra of the Mississippi basin and is confined to it. This rank was based on a single character state, the proliferation of interlamellar connec- tive tissue approximating true septa. We argue that Cumberlandia (described by Оптапп, 1912a) is not deserving of sub- family rank, and that it may best be consid- ered a synonym of Margaritifera. On the basis of immunological data С. monodonta is in the same subset with М. margaritifera (Fig. 2). The MST shows that C. monodonta is inter- mediate in relationship between M. margariti- fera and M. hembeli. M. falcata is more dis- tantly related to М. margaritifera than is Cumberlandia monodonta. We consider these relationships to be accurate as shown because we had antisera for each of the four species and thus could make the appropriate cross-comparisons. What distinguishes Cumberlandia mono- donta from the other three species is the somewhat continuous, oblique “septa” of the former in contrast to the patternless inter- lamellar connectives of the latter. We con- sider this small modification of the “septum” to be indicative of a species difference in the Margaritifera complex and thus we consider Cumberlandia, on all data available, to be a synonym of Margaritifera. We see no reason to use subgeneric rank at all. We are especi- ally confident of our conclusions because we have tested and examined every nominal Nearctic margaritiferine species: Cumber- landia monodonta of the Mississippi basin, Margaritifera hembeli of the Gulf drainage, М. margaritifera of the northern Atlantic drain- age, and M. falcata of the Pacific drainage. We compared our classifications with those of Ortmann (1910a, 1911, 1912a, 1916, 1919) and Modell (1942, 1949, 1964), as well as Heard € Guckert (Table 12). These classi- fications were chosen because Ortmann's work stands for the totally synthetic approach; Modell's is a sprawling classification based on shell characters; and the Heard & Guckert classification purposely ignores conchology and 15 essentially monothetic with some atten- tion to reproductive characters. It is clear that there is closest agreement between our classification and that of Ortmann, especially his earliest work (1910a). In his later work Ortmann (1911, 1912b, 1919) elevated the group of Margaritifera to familial status. Accordingly, in scoring diver- gence of other classifications from ours, we used a range for Ortmann’s classification based on his earlier and later schemes. We arbitrarily gave 5 points for each family or sub- family that is not equivalent to our comparable family or family concept (Table 12, t) and one point for a genus that has been placed with another genus (genera) that belongs to an- other suprageneric taxon in our classification (Table 12, *). Ortmann’s score is 19 or 29; the Heard & Guckert score, 48; Modell scores 55. On this basis, Modell's classification is the least satisfactory. Modell (1942, 1949, 1964) established a strictly conchological classification of three families and 10 subfamilies with Nearctic mem- bers. Heard & Guckert (1971) and Heard (1974) were mostly correct in stating that shell char- acters typically do not correlate with soft-part characters. For the most part shell characters do not correlate with anatomical characters or genetic data, also we have discussed the reasons for rejecting a two- or three-family classification. Gonidea is not closely related to Margaritifera even though the shells are somewhat similar. Alasmidonta and Anodonta are closely related genera of the same sub- family, and neither is closely related to Lampsilis or Elliptio. Numerous other objec- tions to Modell’s scheme could be raised. We have discussed our far fewer objections to the Heard & Guckert classification. Our classifica- tion is closest to Ortmann’s (1910a) original arrangement, i.e. one family and the sub- families Margaritaninae (= Margaritiferinae), Unioninae, Anodontinae, and Lampsilinae. Indeed, the combination of his North Ameri- can Unioninae, Lampsilinae, and (1916) Gonideinae equals our Ambleminae. More- over, his Lampsilinae and Gonideinae have exact cognates in our amblemine tribes Lampsilini and Gonideini. Our classification thus differs from Ort- mann's in three significant ways. First, his 238 DAVIS AND FULLER Unioninae comprised diverging clades, i.e. our Amblemini and Pleurobemini. Ortmann lacked the biochemical tools necessary to reveal genetic relationships and parallel evolution that are crucial to our concepts Pleurobemini and Amblemini. Second, by raising what we consider tribes to subfamily rank he implied greater morphological and genetic divergence among groups than we think justifiable. Third, no previous author recognized that the Anodontinae are as dis- tinct and separate a group as indeed they are. While the anodontine taxa have unique morphological character-states that set them apart, these do not appear to present as great a magnitude of difference in comparison with the comparable character-states of our Ambleminae as in comparison with those of the Margaritiferinae. In other words, the Anodontinae have advanced beyond the Mar- garitiferinae about as far as the Ambleminae have done, but divergently. Given the evolving differentiation of the gills as efficient marsupial chambers, one could argue on the basis of morphology that our Amblemini and Pleurobemini are the most primitive of Nearctic non-margaritiferinae taxa; they lack the complex mantle structures of some of our Lampsilini, and many species have heavy shells, some resembling those of Margaritifera. The anodontine species could have been considered (and indeed, have been so considered by some authors) as de- rived from our Ambleminae with advanced specialization that included a reduction from heavy to light shell, reduction of hinge, and further marsupial development to yield the tripartite water tube. This definitely has not been the case. Considering all classifications and in sum- mary we can make several points. 1) Wher- ever a monothetic basis for classification has been used, the classification places closely related taxa into artificial groupings. 2) Ort- mann's approach 15 superior because he used all data available. He was interested not simply in a utilitarian classification, but also in obtaining answers to why and how morphol- ogy and habits yielded the amazing unionid diversity in North America. 3) The Heard & Guckert (1971) study is of particular value be- cause it purposefully set up a classification following a stated approach. They provided clear-cut concepts that are amenable to es- tablishing hypothesis that can be tested. 4) Heard & Сискей were mistaken in “subjec- tively electing . . . to ignore one array of fea- tures (i.e., conchological features)” and in overemphasizing such reproductive aspects as number of marsupial demibranchs and length of breeding season. Relationships within the Ambleminae We discuss at some length relationships among certain taxa within the Ambleminae because of our extensive immunological data and the availability of a certain amount of anatomical data. 1. Gonidea—The placement of this genus in a taxonomic hierarchy has continuously been a problem to malacologists. On the basis of the shell it would be considered a magaritiferine. Modell (1942, 1949, 1964) considered this to be the case and relegated Gonidea to his Pseudodontinae of the Margaritiferidae. Ortmann (1916) considered Gonidea a member of the Unionidae: Gonideinae. Heard & Guckert (1971) placed Gonidea in their Amblemidae because the genus is tetragenous. They preserved Ortmann’s (1916) Gonideinae for it because of its perforated septa. Heard (1974) reported that Megalonaias has perforated septa and considered them characteristic of primi- tive tetragenous taxa, such as Gonidea, Megalonaias, and Pseudodon. Gonidea is immunologically more closely related to our core Ambleminae (Fig. 3). Gonidea diverges away from the Amblemini and Pleurobemini and can in no way be con- sidered a member of the Margaritiferinae. With the Anodonta group removed from Ortmann’s equal ranking of Anodontinae, Gonideinae, Unioninae, and Lampsilinae, we see that Gonidea deserves equal rank with Ortmann’s generic groupings around Lampsilis (our Lampsilini) and around Elliptio and Quadrula (his Unioninae, our Pleuro- bemini and Amblemini). Because Gonidea has vertical septa and a complete diaphragm, in contrast to the primitive margaritiferine groundplan, we consider its perforations a primitive condition that has been sustained in the Gonideini and, by parallel evolution, in Megalonaias of the Amblemini. It is possible that this west coast North American genus is most closely related to the tetragenous genera of Asia (Heard, 1974). Further investi- gations are necessary to assess such a sug- gested relationship. Should this proposed link to Asia be correct, it is probable that the Gonideini would deserve subfamiy ranking. UNIONACEAN GENETIC RELATIONSHIPS 239 2. The Elliptio-Fusconaia-Pleurobema prob- lem. Immunologically and morphologically there is little basis for taxonomically sepa- rating these genera. Pleurobema and Elliptio are ectobranchous, and Fusconaia is tetra- genous. However, species assigned to these taxa do not sort into two or three immuno- logically separate clusters. Clarification of re- lationships within the Pleurobemini will de- pend on molecular genetic studies using taxa only in this group, plus more sets of Pleuro- bemini antisera. As seen in Tables 13 and 14, when the 12 traditional genera of Pleuro- bemini and Amblemini are compared by using eight characters, Fusconaia differs from Elliptio in three character-states and from Pleurobema in two. Fusconaia differs from both genera in having brightly colored tissues and in being tetragenous. Elliptio differs from both Fusconaia and Pleurobema in having simple, not dentritic, incurrent papillae. We shall retain these genera until such a study has been completed. The genera are tentatively defined as follows: Elliptio, ecto- branchous, simple incurrent papillae, shells more or less elongate with beaks placed well anterior and not prominent; Pleurobema, ectobranchous, dentritic incurrent papillae, shells subtriangular to rhomboid with beaks anterior or subanterior and prominent; and Fusconaia, tetragenous, dentritic incurrent papillae, shell much as in Pleurobema. However these taxa are defined in the fu- ture, we note that so-called Elliptio, Fusconaia, and Pleurobema are very closely related genetically. 3. Linkages in the Lampsilini—The т- munological data (Figs. 2, 3) show that 1) the closest relationship of the Anodontinae to the Ambleminae is via the genus Lampsilis, spe- cifically L. teres, 2) L. teres, Ptychobranchus subtentum, and L. radiata form the core of related taxa from which other taxa fan out. 3) The Pleurobemini and Amblemini are closely related to each other and there does not ap- pear to be an extensive divergence among genera of these tribes. 4) Gonidea ties into L. teres via Carunculina parva. Because the relationships indicated by the MST represent genetic relationships among living taxa, and as new taxa are studied and added to the data matrix, one would expect shifts in relationships from those seen in Figs. 2 and 3. Accordingly, one should not consider the overall MST patern to represent evolution- ary pathways. For example, Lampsilis teres did not evolve from Lasmigona costata. Also, as additional anodontine taxa are studied, the linkage between the Anodontinae and Ambleminae might not be between Lasmigona costata and Lampsilis teres. Re- alizing that with additional data there will be shifts in associations of taxa along the MST, we can still say quite a lot about general rela- tionships among groups of genera in the Ambleminae. It is clear that the amblemine clade is ancient and that the genus Lampsilis, of all taxa within this clade, is most closely related to the Anodontinae. 4. The Amblema-Plectomerus-Megalonaias complex—Immunological data indicate a close relationship among species of these three genera. Megalonaias and Plectomerus are especially closely allied within the same subset (Table 10). There is a remarkable piece of morphological evidence that corrobo- rates the implied close genetic relationships of these three taxa. Arborescent incurrent papillae are characteristic of our Amblemini (Quadrula s.s. and Quincuncina have dentritic papillae); incurrent papillae of the simple type do occur, but only in the trio of genera in ques- tion. In short, these form a natural group with- in the Amblemini. Plectomerus and Megalonaias are im- munologically so closely related as to suggest congeneric status. There are data supporting and against congeneric status. The support- ing data would include Amblema with them in a common genus because all three have large, strong, thick, heavy shells with plicate sculpture (three different character-states). In view of the above data, these taxa are allied on the basis of four distinct character-states (3 shell, 1 soft-part) additional to those serv- ing to define the Lampsilini (in addition to the close immunological relationship). Differences occur. Megalonaias possesses a somewhat unusual beak sculpture (i.e. it persists until after adult sculpture has begun and thus intermingles with it). Megalonaias exhibits perforate gill septa, at least in the gravid female (Heard, 1974). Both character- states are considered primitive. Unfortunate- ly, Plectomerus has not been studied ade- quately in regard to these character-states. Amblema beak sculpture is separate from the disc of the adult shell; it has no known perfor- ate septa. Given the total evidence available, given the comparisons in Tables 13 and 14, and considering the morphological changes one expects to see in adaptive radiation (Davis, 1979), we consider it worthwhile to make a 240 DAVIS AND FULLER TABLE 13. Comparison of 12 genera of the Lampsilinae: Pleurobemini and Amblemini by using eight char- acters and 21 character-states. [Q. (Orthonymus) cylindrica represents a discrete genus here because it differs so greatly from Quadrula s.s.] оо À © © = O 3 AS] iS © © © 5 $ 8 GPS $ Е 3 Enno Ei Seh SOUS ES il Se DI a En; eng $ Poy Go Bors ово Е A O A 1. Incurrent papillae simple (0) dendritic (1) arborescent (2) 0 0 0 1 2 2 2 1 0 1 1 1 2. Excurrent papillae weak or absent (0) well developed (1) 1 1 1 1 0 0 0 1 1 1 0 3. Tetragenous (0, 1)* 1 1 1 1 1 0 1 1 0 1 0 0 4. Supra-anal opening (0, 1) 1 1 1 1 1 0 1 1 1 1 5. Septa not perforated (0) weak, but imperforate (1) perforate (2) 0 2 0 0 0 0 0 0 0 0 0 6. Tissues brightly colored (0, 1) 0 0 0 0 1 0 0 0 0 1 0 0 7. Sculpture smooth (0) plicate (1) pustulate (2) IL 102% 2. 2 2 2 0 0 0 0 8. Shell pustulate: in chevron pattern (1) random (2) 1 row large pustules (3) 2 rows pustules (4) 0 0 OF 2/4 73 2 2 1 0 0 0 0 еек кк A A A ПБЯ" НЕЕ ЕЕ Е A A *given only (0, 1), O—does not have, 1—has the character state. **juvenile only. TABLE 14. Comparison of genera given in Table 13 by the number of shared character states. 7) 7 Ф © © © ) о © © y 5 “ 5 E © = 9 Е 3 Ep oe oes I a oes Вы 5 Diy cig SARS SUSE Moe HAS? MC: Q > © © o ra о = а О 5 © Е D © 3 y Sale. = = 3 © = RN об © Cy «ЧЕ o u a a =) Amblema 8 7 8 5 3 2 4 5 6 5 5 3 Megalonaias 8 7 4 2 1 3 3 5 4 4 3 Plectomerus 8 5 3 2 4 5 6 5 5 3 Quadrula s.s. 8 3.5 3 5 6.5 4.5 5.5 5.5 3:5 Q. cylindrica 8 4 6 4 2 4 2 2 Cyclonaias 8 6 3 2 1 3 3 Tritogonia 8 5 3 3 3 3 Quincuncina 8 4 5 5 3 Elliptio 8 5 T4 5 Fusconaia 8 6 4 Pleurobema 8 6 Uniomerus 8 UNIONACEAN GENETIC RELATIONSHIPS 241 hypothesis that these three taxa are con- generic and that the synonymy 15: Amblema Rafinesque, 1820 + Plectomerus Conrad, 1853 + Megalonaias Utterback, 1916 5. The Cyclonaias-Quadrula-Tritogonia complex—We studied at least five Amblemini taxa that have complex pustulate shell sculp- ture: Quadrula spp., Q. (= Orthonymus) cylindrica, Cyclonaias tuberculata, Tritogonia verrucosa, and Quincuncina infucata. Ad- mittediy some species traditionally assigned to Quadrula have smooth or nearly smooth shells. However, as noted in the discussion of Q. cylindrica (below), the genus Quadrula has yet to be defined with precision. Also, subgenera such as О. (Bullata) Frierson [rep- resented by Q. pustulosa (Lea)] may or may not have validity. Quadrula cylindrica differs from the other species of Quadrula we studied in 4.5 mor- phological character-states (shell and soft parts) (Tables 13, 14). Q. cylindrica has arborescent incurrent papillae such as occur in the Margaritiferinae and Gonidea of the Ambleminae; these papillae are considered to represent a primitive character-state. Excur- rent papillae are absent (contrast Quadrula s.s. and Quincuncina); tissues are various shades of browns and blacks (contrast un- colored tissues of other genera studied here except Fusconaia and Margaritifera). Other differences involve shell sculpture. On the basis of molecular genetics (immunological distance coefficients), the closest relation- ships are with other species of Quadrula [Q. pustulosa (0.870); Q. cf. quadrula (Qbb, 0.902); О. apiculata (1.04) and then Plectom- erus (1.04)]. Because Q. cylindrica differs from Quadrula s.s. in three anatomical (soft part) character-states, we consider О. cylind- rica to typify a distinct genus, Orthonymus. Cyclonaias, a monotypic genus, closely re- sembles Quadrula conchologically. Of the Amblemini genera, only Cyclonaias, Orthony- mus, and Tritogonia have arborescent papillae and no (or poorly developed) excurrent papillae. Cyclonaias differs from all other Amblemini studied by us in being ectobranch- ous and having an entire excurrent aperture. The immunological distance coefficients among C. tuberculata and the five closest species are, in increasing order: Plectomerus (0.780); Quadrula cf. quadrula (Qbb, 0.826): Megalonaias (0.966); О. pustulosa (1.04); and Q. apiculata (1.05). No immediate ge- netic relationship to Orthonymus or Tritogonia is indicated. Because Cyclonaias differs from Quadrula in three morphological character- states and from Amblema (plus synonyms) in at least six character-states (Tables 13, 14), we shall maintain Cyclonaias as a discrete genus. Tritogonia is maintained as a separate genus because it has arborescent papillae, not in Quadrula s.s., and a different shell sculp- ture (Table 13). Its closest immunological re- lationships are with Quadrula pustulosa (0.870) and Orthonymus cylindrica (0.219) and then with non-Amblemini, e.g. Elliptio lanceolata (1.257) and Lampsilis teres (1.263). 6. Lampsilini—The Lampsilini are unique among the Unionidae in that the marsupial water tubes extend beyond the distal margins of the demibranch lamellae; the marsupia show externally marked sulci, not the smooth pads as in the homogenous taxa (tetra- genous or ectobranchous); and discrete areas of the outer demibranch are marsupial in the great majority of species. It is reasonable to assume that in the evolu- tion of the Lampsilini there were independent origins of some of the marsupial types and that some developed from others. For ex- ample, it is improbable that the mesogenous condition (Appendix 3) was modified to pro- duce the heterogeneous condition, because the two are structurally different and occupy different parts of the ощег demibranchs. Nevertheless, there is a progression from primitive to specialized character-states. Presumably the most primitive state 15 longenous: the entire length of the demi- branch is marsupial, and the distad distension of the water tubes is slight. This is not a very successful state; only two genera have been assigned to the Lampsilini: Longenae (a pos- sible subtribal concept): Friersonia and Cyrtonaias. We have not studied these genera immunologically. Additional information about them occurs in Ortmann (1912b), Heard 8 Guckert (1971), and Fuller (1975). A condition possibly derived from the longenous is the ptychogenous state. Here the marsupium extends the full length of the demibranch, but only the ventral portions of the water tubes are marsupial. Additional space for incubation of larvae is created by a distad distension of the water tubes that is greater than that in the Longenae: the tubes 242 DAVIS AND FULLER are somewhat distended laterally and from front to rear, which causes a folded (‘‘ptycho- genous”) condition such that the lower border of the demibranch is furbelowed. Only Ptycho- branchus, the only genus of Ptychogenae, has this condition (Fig. 3). The eschatigenous condition (in the lone genus of Eschatigenae, Dromus, not studied here) resembles the ptychogenous type in be- ing limited to the ventral border of the demi- branch, but is unique in consisting of a series of several discontiguous sacs formed by a distad distension of the marsupial water tubes that exceeds that in ptychogenous mussels. The mesogenous condition (in the Mesogenae, Obliquaria and Cyprogenia, not studied here) involves great distad distension of several contiguous water tubes in the mid- dle of the demibranch; the distensions exceed those in the eschatigenous marsupium, and in Cyprogenia they are so long that they must coil in order to remain within the mantle cavity and thus be protected by the shell. The heterogenous condition is restriction of the marsupium to the posterior (or even the postbasal) portion of the demibranch. All other Lampsilini are Heterogenae. Results of our immunological analysis of Heterogenae are represented in Fig. 3. The eventual addition of other taxa to the analysis doubtless will change this portrayal in some respects and will permit greater confidence in all the results of that time. At present, how- ever, the picture has some features that are gratifyingly in keeping with morphological evi- dence; there are, also, some relationships that are mystifying. п the former category, there is, for example, the radiation of several Lampsilis and Villosa from L. radiata. This is not surprising, because of the similarity of the two genera and because this species 1$ not an advanced member of the genus (the mantle Нар is essentially ribbon-like, unlike the fully developed piscine type seen in L. ovata). Also of note is that L. teres is not a part of this radiation, because its postbasal mantle mar- gin questionably forms a flap of the sort ex- emplified by L. ovata, the type-species of the genus and a member of the radiation from L. r. radiata, and because its beak sculpture, also, is atypical of the genus. A further inter- esting aspect of L. teres is its immunological alliance to Ligumia recta (Rafinesque), which it resembles morphologically so much that it long was the accepted practice to place Ligumia in Lampsilis. One concludes that in at least some cases conchological evidence is more meaningful than has been recognized in many years. The opposite point is indicated in some other cases. For example, Ligumia recta and L. nasuta have been considered congeneric because of their similar shells, but they are not closely allied in our immunological analysis. We cannot be confident that we fully under- stand these two species' relationship. We feel even more uncertainty about our results con- cerning Lampsilis hydiana. This species seems morphologically to be related to (or even part of) the sprawling Lampsilis r. radiata complex, but immunolologically it not only is not part of the radiation centered in that subspecies, but also lies a great genetic dis- tance from it. Entirely unexpected results, such as these, strongly suggest the need (and some directions) for further study. These remarks about relationships within the Lampsilini serve to illustrate some of the apparent strengths and weaknesses of our analysis. The same point can be made about the indicated relationships between the Lampsilini and other tribes and subfamilies. We think it significant that Ptychobranchus (Ptychogenae) is both the genus of Lampsilini studied by us that has been considered most generalized by some (e.g., Ortmann, 1912b) and the one that serves as the immunological connector to the Ambleminae: Pleurobemini and (though Lampsilis teres) to the Ano- dontinae. Similarly, the pathway between the Lampsilini and the Ambleminae: Amblemini lies through Glebula, a monotypic, general- ized genus in the Heterogenae. We by no means anticipate that these details would remain unmodified in the event of an analysis of a larger number and variety of taxa, but we find it intuitively satisfying that rather unspe- cialized animals are the connections of the present scheme. Adaptive radiation and success Changes from primitive to derived charac- ter-states presumably indicate entrances into new adaptive zones and the establishment of new groundplans that engendered adaptive radiation. In considering the success of groups with various groundplans, we are con- cerned with 1) the extent of a given adaptive radiation, ¡.e. the number of species radiating with a given morphological groundplan; 2) the geographic range and abundance of these species, and 3) the competitive ability of the UNIONACEAN GENETIC RELATIONSHIPS 243 species that enables coexistence with other, sympatric unionid species. As discussed previously, the critical factors for unionid success appear to involve aspects of reproduction and respiration that depend on hydrodynamic efficiency; the critical factor for increased hydrodynamic efficiency is the bivalve diaphragm. The diaphragm is a col- lection of tissue that more or less separates the incurrent and excurrent portions of the mantle cavity. The unionid diaphragm is in- complete, i.e. the separation of incurrent and excurrent mantle cavities is imperfect and a tightly closed hydrodynamic system is thus impossible. In the Margaritiferinae this diffi- culty is exacerbated because the gill extends posteriad far short of the posterior mantle margin and only the gill bars effect separation of the two cavities. As a result, there is leak- age between them, which must cause a physiological disadvantage, but also corre- lates well with other primitive aspects of mar- garitiferine morphology, namely, the large foot and gaping valves. The latter two features obviously are re- lated, and they serve further to weaken the margaritiferine hydrodynamic system. On the other hand, the large foot helps in negotiating the gravels and rock interstices favored by margaritiferines. The apapillose character- state of the excurrent posterior mantle aper- ture is considered by us to be the primitive condition and perhaps correlates with a weak pumping system because papillae would im- pede exit of waste particles and larvae ex- pelled by the weak excurrent stream. By comparison to other unionid groups the Margaritiferinae are not very successful. They are holarctic with representation in southeast Asia. They have a fossil record extending from the upper Cretaceous (Haas, 1969b). However, they are few species (five or six), which apparently belong to only one genus. Most of the species are restricted to cool, highly oxygenated water and gravel or rocky substrate. The species are most frequently found in soft-water upland streams without other species of unionids. The Anodontinae have some unique char- acter-states. These complement any deci- sion, based on whatever kind of evidence, that Anodonta and its kin are a distinct unionid group. The subfamily is holarctic in distribu- tion, as is the Margaritiferinae The wide- spread distribution patterns suggest that the two subfamilies predate some groups of the Ambleminae that are entirely restricted to North America. The Anodontinae have had a far greater success than have the Margariti- ferinae. The genus Anodonta is represented by several nominal subgenera (including, no doubt, at least some legitimate biological en- tities). The Nearctic is the area of greatest anodontine survival and speciation, as ap- parently is the case for the Margaritiferinae. The Anodontinae may have been successful elsewhere, as well, as is suggested by the great similarity of Alasmidonta arcula (Lea) of the Altamaha River, George, U.S.A., to Unio languilati of China (see Johnson, 1970, and Heude, 1875). The Anodontinae are similar (yet hardly identical) to the Margaritiferinae, but vastly different from other Nearctic Unionidae, in ex- hibiting almost no hitherto discernible generic differences of soft-tissue anatomy. Soft-tissue diversification has been the key to success of the Ambleminae, even though there have been some correlative conchological adjust- ments. However, evolution among the Anodontinae appears to have involved essen- tially only the shell. Accordingly, in North America (where genera that are morphologi- cally anodontine are numerous) there exists a conchological range from the heavily hinged and completely dentate Lasmigona com- planata through the paper-thin and edentu- lous Anodonta imbecillis. The genus Alasmidonta (perhaps including Unio langu- ilati) and its complex of at least five nom- inal subgenera represent an intermediate step in this evolutionary progression. The conchological characters of this genus т- clude pseudocardinal dentition and more or less well developed lateral teeth. Our point is that this group includes character variation that probably is too great to justify inclusion in a single genus. For example, one such spe- cies, Alasmidonta (Prolasmidonta) heterodon (Lea) recently had its subgenus (of which the species is Ortmann’s (1914) monotype) raised to generic level (Fuller, 1977). The cor- rect systematic placement of this species is most uncertain, but it remains symbolic of the difficulties in classifying morphologically equivocal animals whose genetic affinities have not been immunologically well estab- lished. The conchological diversity and the soft- tissue conservatism of the Anodontinae have been reviewed. This peculiar combina- tion of trends in characters probably justifies Our Suspicion that this subfamiy's morpho- logical features are mainly unique and war- 244 DAVIS AND FULLER rant unusual taxonomic treatment. However, this standpoint does not exhaust the roster of anodontine peculiarities. The modern Anodontinae are more species- rich and morphologically diversified than the modern Margaritiferinae, but this serves to dramatize the apparent pattern of anodontine differential extinction—or lack of initial suc- cess. Several of the Nearctic anodontine genera are monotypic, and the list probaby will increase as a result of further research because some of the other genera have nu- merous monotypic nominal subgenera, some of which probaby deserve generic rank. Only Anodonta itself has speciated with much suc- cess, and only this genus exhibits wide eco- logical and geographical ranges. Anodontine failure strengthens the supposition of the sub- family's antiquity and early derivation from other Unionidae. Subtribal groups of Ambleminae: Lampsilini that are based on the longenous, ptycho- branchous, eschatigenous, and mesogenous marsupial types include only six genera, of which Obliquaria and Dromus are monotypic. Cyprogenia, Friersonia, and Cyrtonaias pres- ently include at most two species each. These marsupial conditions and corresponding sub- tribes are not correlated with success as measured by large radiations of species or numerous genera. Cyrtonaias tampicoensis (Lea) and the monotype Obliquaria reflexa Rafinesque are successful in the Gulf of Mexico drainage of Texas and Mexico and in the Gulf drainage and the Mississippi River basin, respectively, but the other species of these groups probably never have had geo- graphically successful ranges. More specifi- cally, several of these species have been ге- stricted to the Cumberlandian and/or Ozark- ian faunas (see van der Schalie 8 van der Schalie, 1950). Dromus and Сургодета are limited to one or both biogeographical prov- inces. lt probably is significant that the Longenae are geographically separated from the others of these unsuccessful groups. The Heterogenae are successful. Their success correlates with the marsupial restric- tion to the posterior section of the demibranch (see p. 242). One quarter of the naiad species recognized as having invaded or reinvaded the Canadian interior basin since the most recent (Wisconsin) glaciation are hetero- geneous Lampsilini (Clarke, 1973). As an- other example, the Lampsilis radiata complex probably is the geographically most widely ranging group of Nearctic naiades. In order to accomplish its geographic range, the complex must have wide ecological tolerances, as well. The Heterogenae include most of the genera of Lampsilini. However, even within this, the most specialized and by far the most successful Lampsilini group, there are differ- ent degrees of morphological development and of success. There is a morphological gra- dient corresponding to the joint theme of great- ly reducing the amount of outer demibranch that is marsupial and of locating the marsupi- um at the posterior end of the demibranch. The more specialized heterogeneous genera have a swollen reniform marsupium (when charged) in the postbasal corner of the outer demibranch. The nearby postbasal mantle margin is modified in various ways that serve as attractants for piscine hosts of unionid larvae. For example, the posbasal margins of Lampsilis are piscine in character; the implication is that predatory (or merely grazing) fish species will attack the “prey” represented by the mussel's mantle margins and will be showered with glochidia if, as is often true in the case of heterogenous genera, discharge of parasitic larvae is through the marsupial wall and not through the excurrent mantle aperture. These morphological adjustments have been accompanied by ethological adapta- tions, as well. For example, the female of some (perhaps all) Lampsilis is able to orient herself so that her marsupium's proximity to the host fish is optimized and the movement of her posbasal mantle margins (piscine flaps) are capable of attracting a host. Whether or not all Lampsilini: Heterogenae can coordinate with potential hosts is not known, but complementary structural and be- havioral strategies are clear. The incorpora- tion of behavioral factors into the reproductive process not only probably is the key to the success of the Heterogenae, but also pro- vides a key to classification of the group. The fact that the postbasal portion of the outer demibranch is marsupial defines this group, but there are other variables that are of use in classification, e.g. pigmentation, size and shape of the egg mass, and lamellar cover- age of the egg mass. An example of a problem in a classification that uses such characters is Unio ochraceus Say. This species was long classified as a Lampsilis, which it clearly is not, because it has no mantle flaps, as recognized by Morri- son (1975), who considered this species a UNIONACEAN GENETIC RELATIONSHIPS 245 Leptodea. Bereza & Fuller (1975) pointed out that the number and structure of the egg masses of this species are not similar to those of Leptodea. No one has yet proposed a generic name for this species. As a group, however, the Heterogenae are character-poor. This not only has created taxonomic problems, but also makes tracing the group's radiation very difficult on morpho- logical grounds alone. Nevertheless, immuno- logical evidence is somewhat compensating. Zoogeography through time The Unionacea are known with some de- gree of authority from the Triassic (review by Walker, 1910; Modell, 1942; Haas, 1969b). They are perhaps known from the upper Devonian (Smith, 1977). The Unionacea were widely spread in Pangaea; the presumably primitive family of Hyriidae of Australasia and western South America remained in Gond- wanaland continents; the Unionidae, es- sentially in Laurasian continents. African Unionidae are either due to an original Gond- wanaland stock or derived from a later inva- sion from Eurasia (see Heard, 1974). The greatest diversity of naiades today is found in the Atlantic drainages of the Old and New Worlds. The implication is that the area of initial radiation of the ancestors of modern naiades lay in that portion of Pangaea where the Atlantic rift began in the Mesozoic. Wheth- er or not the Unionacea (a nearly global group) and the Mutelacea (a Gondwanaland element) have a derivative relationship is un- resolved, though not at issue here. The essential identity of soft-tissue plan in all Unionacea suggests that a common stock existed in Pangaea. The acquisition of four or five derived morphological character-states separating the more specialized non-margari- tiferine Unionidae from the Margaritiferinae must have occurred before the breakup of Pangaea, as is evidenced by the modern dis- tribution of Margaritiferinae, Anodontinae, and African and Asian taxa related to North Amer- ican Ambleminae. The Margaritifera group is of Laurasian origin and has a modern relict distribution in Laos and the Holarctic. It is inconceivable that this ephemeral, at present largely unsuccess- ful group, confined essentially to uplands, could have achieved its present distribution entirely by post-Pangaean land bridges. Walker (1910) argued persuasively that Margaritifera evolved in Asia and reached western North America via the Bering land bridge in the Miocene or early Pliocene, and that Margaritifera reached eastern North America via the Greenland bridge. Pangaean- Asian origin of Margaritifera subsequently af- fected by plate tectonics and dispersal over Pliocene to Pleistocene land bridges was endorsed by Smith (1977). The Anodontinae, also, are Holarctic and confined to the northern hemisphere. Only Anodonta is known with certainty to be repre- sented in Eurasia. There is a pronounced conchological similarity of certain Nearctic Alasmidonta to at least one species of east- ет Asia, which accordingly 15 considered anodontine. The geographic distribution of Anodontinae 15 in Europe, Asia (Oriental zoo- geographic province), and North America; this indicates a widespread Laurasian distri- bution. The modern proliferation of anodontine genera is in North America. Only Anodonta is widespread, commonly encountered, species- rich, and biologically adaptable. There are at least three, chronologically differing interpre- tations of the occurrence of Eurasian Anodonta (or very similar forms): 1) The anodontine radiation, including Anodonta, was complete prior to the breakup of Laurasia; 2) the greatest anodontine cladogenesis oc- curred in North America, but Anodonta was widespread in Laurasia before North America separated from the pangaean supercontinent; 3) Anodonta spread to Asia from eastern North America prior to the rise of the Rocky Mountains and subsidence of the Bering land bridge. Walker (1910) considered Nearctic Anodonta of the Pacific drainage to be of Asiatic origin. Heard (1974) considered primi- tive progenitors of modern Anodontinae (e.g., Strophitus) to have originated in Asia and spread via the Bering land bridge into North America. Following Walker, it is highly prob- able that Anodonta, as ancient as Margariti- fera, had its origin in the same pangaean re- gion as Margaritifera, and dispersed via the same general routes. Some taxa in our Ambleminae questionably originated in the Triassic age and certainly existed in the Cretaceous. Margaritifera and the Anodontinae are known with certainty from the Cretaceous. Simpson (1896) noted a “remarkable similarity” between the unionid faunas of North America and southeastern Asia, plus the Tertiary faunas of both Europe 246 DAVIS AND FULLER and Asia. Walker (1910) stated that there 15 “no doubt but that the characteristic (unionid) fauna of North America is descended from the Upper Cretaceous species, which then lived” in certain western U.S.A. states, as is evi- denced by the fossil record. Walker (1910) noted the strong resemblance of Oriental Unionidae and those of North American Cretaceous to early Tertiary North American fossil unionids. Given the evidence, one rea- sonably assumes that the breakup of Pangaea did isolate a segment of early unionid stock in North America and that these isolates gave rise to most of the current endemic North American fauna. Only much later did some Asian stock reach western North America via the Bering bridge or east- ern North America from Europe. Members of the Ambleminae: Pleurobemini and Gonideini have morphological affinities to certain African and Eurasian taxa. These are reviewed by Heard & Guckert (1971) and Heard (1974); a few will be mentioned here. Brazzaea anceyi Bourguignat, of Africa, was grouped in the Gonideinae (Heard € Guckert, 1971) because it had been reported (Bloom- er, 1931a) to be tetragenous, with distinct supra-anal opening, and with continuous, but |} Ambleminae ————— Gonideini Lampsilini Pleurobemin Amblemini perforated septa. Lamellidens marginalis (Lamarck) from India is ectobranchous, yet with perforated septa (Bloomer, 1931b). Be- cause tetragenous or ectobranchous taxa may occur in any tribe, we provisionally place this taxon in the Gonideini. Heard & Guckert listed several Southeast Asian taxa with perforated septa that they considered Amblemidae and we provisionally consider as Gonideini. The tribe Lampsilini is uniquely North American and, with the possible exception of a morphologically somewhat Lampsilini ele- ment in the Pacific drainage (Dwight Taylor, personal communication), is entirely confined to the Atlantic drainages. It is probable that the Lampsilini radiation occurred only since the complete independence of North America. Of the five morphologically defined sub- Lampsilini groupings of taxa that have been proposed, four are comparative failures, but the fifth, the Heterogenae, dominates the en- tire Atlantic drainage faunas in terms of num- bers of genera and species and in terms of ecological success. This great success is at- tributed to the sum of morphological character- states that are unique to the Lampsilini in gen- eral and to the Heterogenae in particular. + Anodontinae{ }Margaritiferinae4 UNIONIDAE FIG. 4. Cladogram portraying relationships among unionid taxa. Numbered points are discussed in the text, p. 247. UNIONACEAN GENETIC RELATIONSHIPS 247 Cladistic relationships A cladogram (Fig. 4) was constructed on the basis of our immunological results, mor- phology, the fossil record, and zoogeography. As implied by the cladogram, there was di- vergence within proto-unionid stock before Gondwanaland split up in the late Mesozoic. The proto-unionid stock would have had the generalized, Margaritifera-type anatomy and would have been tetragenous. Divergence gave rise to proto-Margaritifera (point 1, Fig. 4) and to a lineage that gained some morphologi- cal advances, ¡.e. development of septa and water tubes рага!е! to the gill filaments, crea- tion of the diaphragm and supra-anal aperture. The septa probably were perforated (point 2, Fig. 4). The unionids with these derived morpholog- ical character-states diverged before Gond- wanaland split up and yielded yet again two lineages. One of these, the proto-Anodontinae (point 3, Fig. 4), developed tripartite water tubes, became ectobranchous, and developed hooks on the glochidia. One eventual taxon (Strophitus) retained perforated septa. The other lineage (point 4, Fig. 4) remained tetragenous and had undivided water tubes with hookless glochidia and perforated septa. This lineage diverged, yielding proto- Gonideini (point 5, Fig. 4) prior to the breakup of Pangaea. This clade is primarily tetra- genous and has perforated septa. There was, also, rapid divergence that formed the lineages of the 1) proto-Amblemini (point 6, Fig. 4), where the taxa are primarily tetragenous, one species group has perforat- ed septa, and several species have arbores- cent incurrent papillae, as in the Margariti- ferinae and the lineage of the 2) Pleurobemini (point 7, Fig. 4), where the taxa primarily are ectobranchous, without perforated septa, and without arborescent papillae. Last, the proto-Lampsilini (point 8, Fig. 4) evolved; they are uniquely North American, totally ectobranchous, and with the most spe- cialized character-states of marsupial devel- opment and mantle modifications. The cladogram is consistent with the ordina- tion diagrams based on immunological data given in Figs. 2, 3. ACKNOWLEDGEMENTS We acknowledge with great thanks the technical assistance of Mrs. Caryl Hesterman. We thank Dr. William H. Heard, Mr. Daniel J. Bereza, Mr. Malcolm F. Vidrine, and Dr. Dwight Taylor for supplying us with various populations of clams. Mrs. Harriet Davis and Ms. Karen Snow, also, provided technical as- sistance. We especially acknowledge Dr. Wayne Moss for his help in the initial stages of the multivariate analyses. LITERATURE CITED BAKER, F. C., 1928, The freshwater Mollusca of Wisconsin. Part Il. Pelecypoda. Wisconsin Geo- logical and Natural History Survey Bulletin 70(2): 495 p., 53 pl. BASHFORD, N. L., BUTLER, J. E., LEONE, С. A. & ROHLF, Е. J., 1968, Immunological comparisons of selected Coleoptera with analyses of relation- ships using numerical taxonomic methods. Sys- tematic Zoology, 17: 388-406. BEREZA, D. J. & FULLER, $. L. H., 1975, Notes on “Lampsilis” ochracea (Say) (Mollusca: Bivalvia). ASB [Association of Southeastern Biologists] Bulletin, 22(2): 42. BLOOMER, H. H., 1931a, On the anatomy of Brazzaea anceyi Bourguignat. Proceedings of the Malacological Society of London, 19: 228- 233. BLOOMER, H. H., 1931b, А note on the anatomy of Lamellidens marginalis, Lamarck and 1. thwaithesii, Lea. Proceedings of the Malaco- logical Society of London, 19: 270-272. BURCH, J. B., 1973, Freshwater Unionacean clams (Mollusca: Pelecypoda) of North America. United States Environmental Protection Agency. Biota of Freshwater Ecosystems; Identification Manual 11, 176° p: BURCH, J. B., 1975, Freshwater unionacean clams (Mollusca: Pelecypoda) of North Amer- ica. Revised Ed. Malacological Publications, Hamburg, Michigan, 204 p. CLARKE, A. H., 1973, The freshwater mollusks of the Canadian interior basin. Malacologia, 13: 1- 509. CONRAD, T. A., 1834, New fresh water shells of the United States with colored illustrations, and a monograph of the genus Anculotus of Say; also a Synopsis of the American naiades. Phil- adelphia, Pennsylvania, р. 1-76. DAUGHADY, W. H., LOWRY, O., ROSEBROUGH, N. & FIELDS, W., 1952, Determination of cere- brospinal fluid proteins with the folin phenol reagent. Journal of Laboratory and Clinical Medicine, 39: 663-665. DAVIS, G. M., 1969, Electrophoretic, immunologi- cal, and biological properties of a population of Semisulcospira transmitting Paragonimus west- ermani in Japan. Japanese Journal of Parasitol- ogy, 18: 93-119. DAVIS, G. M., 1978, Experimental methods in mol- luscan systematics. In: FRETTER, V. & PEAKE, J., eds., Pulmonates, Vol. 2A: 99-169. Academic Press, London. 248 DAVIS AND FULLER DAVIS, G. 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: ых, 1-120. DAVIS, С. M., FULLER, $5. L. H. & HESTERMAN, C., 1978, Toward a definitive higher classifica- tion of North American Unionacea. Bulletin of the American Malacological Union, Inc., for 1977, р. 85 (abstract). DAVIS, G. M., HEARD, W., FULLER, S. L. H. 4 HESTERMAN, C., 1981, Molecular genetics and speciation in Elliptio, and its relationships to other taxa of North American Unionidae (Bivalvia). Biological Journal of the Linnean Society of London (in press). DAVIS, G. M. 4 SUZUKI, S., 1971, Mouse ascites fluid as a source of antibody against molluscan antigens. Veliger, 13: 207-225. FITCH, W. M., 1976, Molecular evolutionary clocks, In: AYALA, J., ed., Molecular Evolution. Sin- auer, Sunderland, Massachusetts, U.S.A., p. 160-178. FRETTER, V. & GRAHAM, A., 1962, British Proso- as Molluscs. Ray Society, London, р. хм, 1-755. FRIERSON, L. S., 1927, A classification and anno- tated check list of the North American naiades. Baylor University Press, Waco, Texas, 111 р. FULLER, 5. L. H., 1974 [1973], Fusconaia masoni; (Conrad 1834) (Bivalvia: Unionacea) in the Atlan- tic drainage of the southeastern United States. Malacological Review, 6: 105-117. FULLER, 5. L. H., 1975, The systematic position of Cyrtonaias (Bivalvia: Unionidae). Malacological Review, 8: 81-89. FULLER, 5. L. H., 1977, Freshwater and terrestrial mollusks, p. 143-194. In: COOPER, J. E., ROB- INSON, $. S. 4 FUNDERBURG, J. B., eds. Endangered and Threatened Plants and Ani- mals of North Carolina. North Carolina State Mu- seum of Natural History, Raleigh, p. 1444. HAAS, F., 1969a, Superfamilia: Unionacea. In: Das Tierreich. Eine Zusammenstellung und Kennzeichnung der resenten Tierformen. Lief. 88: i-x, 1-663. de Gruyter, Berlin. HAAS, F., 1969b, Superfamily Unionacea. In: Treatise on Invertebrate Paleontology. MOORE, R. C., ed. Part N, Mollusca, 6: Vol. 1 (of 3): Bi- valvia, Unionacea, p. N411-N470. HANNIBAL, H., 1912, A synopsis of the Recent and Tertiary freshwater Mollusca of the California Province, based upon an ontogenic classifica- tion. Proceedings of the Malacological Society of London, 10: 112-211, pl. 5-8. HEARD, W. Н., 1974, Anatomical systematics of freshwater mussels. Malacological Review, 7: 41-42. HEARD, W. H., 1975, Sexuality and other aspects of reproduction in Anodonta (Pelecypoda: Unionidae). Malacologia, 15: 81-103. HEARD, W. Н. 4 GUCKERT, В. H., 1971 [1970], А re-evaluation of the recent Unionacea (Pelecy- poda) of North America. Malacologia, 10: 333- 355. HEARD, W. Н. & VAIL, V. А., 1976, The systematic position of Unio сайег (Pelecypoda: Unionoida: Unionidae). Zoologica Africana, 11: 45-58. HEUDE, P. M., 1875, Conchyliologie fluviatile de la province de Nanking. Paris, Savy, pt. 1, pl. 1-8. HOAGLAND, К. E., 1975, Reproductive strategies and evolution in the genus Crepidula (Gastrop- oda: Calyptraeidae). Unpubl. Ph.D. thesis, Har- vard Univ. JOHNSON, В. I., 1970, The systematics and zoo- geography of the Unionidae (Mollusca: Bivalvia) of the southern Atlantic slope region. Bulletin of the Museum of Comparative Zoology, 140: 263— 450. KING, M. C. 8 WILSON, A. C., 1975, Evolution at two levels in humans and chimpanzees. Sci- ence, 188: 107-116. KRUSKAL, J. B., 1964, Multidimensional scaling by optimizing goodness of fit to a поптешс hy- pothesis. Psychometrika, 22: 1-27. LEA, |., 1858, Descriptions of embryonic forms of thirty-eight species of Unionidae. Journal of the Academy of Natural Sciences of Philadelphia, 4: 43-50. LEA, |., 1863, Descriptions of the soft parts of one hundred and forty-three species and some embryonic forms of Unionidae. Journal of the Academy of Natural Sciences of Philadelphia, 5: 401-446. MAINARDI, D., 1959, Immunological distance among some gallinaceous birds. Nature, 184: 913-914. MODELL, H., 1942, Das natúrliche System der Najaden. Archiv für Molluskenkunde, 74: 161- 191. MODELL, H., 1949, Das natúrliche System der Najaden. 2. Archiv für Molluskenkunde, 78: 29- 48. MODELL, H., 1964, Das natúrliche System der Najaden. Archiv fúr Molluskenkunde, 93: 71- 126. MORRISON, J. P. E., 1955, Family relationships in the North American freshwater mussels. Ameri- ican Malacological Union Annual Reports, 1955: 16-17. MORRISON, J. P. E., 1975, Maryland and Virginia mussels of Lister. Bulletin of the American Maiacological Union, 1974, р. 36-39. MORTON, J. E., 1971, Molluscs. Hutchinson, Lon- don, 244 p. ORTMANN, A., 1910a, A new system of the Unionidae. Nautilus, 23: 114-120. ORTMANN, A. E., 1910b, The marsupium of the Anodontinae. Biological Bulletin, 19(3): 217. ORTMANN, A. E., 1911, A monograph of the naiades of Pennsylvania [Parts |, Il]. Memoirs of the Carnegie Museum, 4: 279-347, pl. 80-89. ORTMANN, A. E., 1912a, Cumberlandia, a new genus of naiades. Nautilus, 26: 13-14. ORTMANN, А. E., 1912b, Notes upon the families UNIONACEAN GENETIC RELATIONSHIPS 249 and genera of the naiades. Annals of the Carne- gie Museum, 8: 222-365, pl. 18-20. ORTMANN, A. E., 1914, Studies in naiades. Nautilus, 28: 4147. ORTMANN, A. E., 1916, The anatomical structure of Gonidea angulata (Lea). Nautilus, 30: 50-53. ORTMANN, A. E., 1919, A monograph of the naiades of Pennsylvania. Part Ill. Systematic account of the genera and species. Memoirs of the Carnegie Museum, 8: 1-384, 21 pl. ROHLF, Е. J., 1970, Adaptive hierarchical cluster- ing schemes. Systematic Zoology, 19: 58-82. ВОНЕЕ, E. J., KISHPAUGH, J. & KIRK, D., 1972, NT-SYS; Numerical Taxonomy System of Multi- variate Statistical Programs. State University of New York, Stony Brook, N.Y. SARICH, V. M., 1977. Rates, sample sizes, and the neutrality hypothesis for electrophoresis in evo- lutionary studies. Nature, 265: 24-28. SCHALIE, H. VAN DER & SCHALIE, A. VAN DER, 1950, The mussels of the Mississippi River. American Midland Naturalist, 44: 448-466. SIMPSON, C. T., 1896, The classification and geographical distribution of the pearly fresh-water mussels. Proceedings of the United States Na- tional Museum, 18: 295-343. SIMPSON, C. T., 1900, Synopsis of the naiades, or pearly fresh-water mussels. Proceedings of the United States National Museum, 22: 501-1044. SIMPSON, C. T., 1914, A descriptive catalogue of the naiades or pearly freshwater mussels. Pri- vately published by Bryant Walker, Detroit, Mich- igan, xi, 1540 p. SMITH, D.G., 1977, The distribution of the Margariti- feridae: a review and a new synthesis. Bulletin of the American Malacological Union, Inc. for 1976, p. 42. STANLEY, S., 1979, Macroevolution. Freeman, San Francisco, 132 p. STERKI, V., 1898, Some observations on the geni- tal organs of Unionidae, with reference to classi- fication. Nautilus, 12: 18-21, 28-32. STERKI, V., 1903, Notes on the Unionidae and their classification. American Naturalist, 37: 103- 113: WALKER, B., 1910, The distribution of Margaritana margaritifera (Linn.) in North America. Proceed- ings of the Malacological Society of London, 9: 126-145, pl. 2. WEBSTER, R., 1975, Intuition and rational choice in the application of mathematics to soil system- atics. Soil Science, 119: 394-404. YONGE, С. М. & THOMPSON, T. E., 1976, Living Marine Molluscs. Collins, London, 288 p. APPENDIX 1. A list of some important unionacean classifications. Note the profound influence of Ortmann's work on most subsequent systems. TT A TATI A AA A AA A AAA AAA AAA AAA A Simpson (1900, 1914) Unionidae Margaritiferidae Unioninae Unionidae Heterogenae Gonideinae Digenae Unioninae Mesogenae Anodontinae Ptychogenae Lampsilinae Eschatigenae Diagenae Homogenae Tetragenae Ortmamn (1910a, 1911, 1916) Hannibal (1912) Margaritiferidae Unionidae Unioninae Anodontinae Lampsilidae Lampsilinae Propterinae Quadrulidae Quadrulinae Pleurobeminae Frierson (1927) Unionidae Margaritiferinae Unioninae Alasmidontinae Anodontinae Lampsilinae Modell (1942, 1949, 1964) Margaritiferidae Margaritiferinae Unionidae Quadrulinae Rectidentinae Anodontinae Elliptionidae Pleurobeminae Elliptioninae Alasmidontinae Ambleminae Lampsilinae Morrison (1955) Margaritiferidae Unionidae Unioninae Alasmidontinae Anodontinae Amblemidae Ambleminae Lampsilinae 250 DAVIS AND FULLER APPENDIX 1. (Continued) A AAA A A A AAA ШЕЕ ЕЕ nmel Haas (1969a,b) Heard 8 Guckert (1971) Clarke (1973) Margaritiferidae Margaritiferidae Margaritiferidae Unionidae Margaritiferinae Unionidae Unioninae Cumberlandiinae Ambleminae Quadrulinae Unionidae Anodontinae Alasmidontinae Unioninae Alasmidontini Anodontinae Pleurobeminae Anodontini Lampsilinae Popenaiadinae Lampsilinae Anodontinae Lampsilinae Amblemidae Gonideinae Ambleminae Megalonaiadinae Davis et al. (1978) Unionidae Margaritiferinae Anodontinae Lampsilinae Lampsilini Gonideini Elliptionini Amblemini Davis € Fuller (this study) Unionidae Margaritiferinae Anodontinae Ambleminae Lampsilini Gonideini Pleurobemini Amblemini APPENDIX 2. Annotations. 1. Fusconaia ebena (Lea) Shells of the species that we call F. ebena conform to the type-concept of F. ebena. We cannot, at this time, explain the close relation- ship of this species to Lampsilis (Fig. 3). We must obtain fresh F. ebena, examine the anatomy, and retest the relationship. We sus- pect either experimental error in this case or a species with Lampsilini anatomy within a Fusconaia-appearing shell. 2. Uniomerus In the first analysis of taxa here relegated to the Ambleminae (matrix of 15 antisera x 41 sets of antigens), Uniomerus tetralasmus was linked by the minimum spanning tree to Quincuncina infucata (see annotation no. 3). Because we were missing 46.6% of the data for О. infucata in the matrix of cross сотрап- sons, we reran the analysis without the data for Q. infucata. We had anti-Q. infucata anti- sera and data for all but three comparisons (no data for Leptodea fragilis, Villosa delumbis, Elliptio buckleyi). In the reanalysis (15 x 40 matrix), U. tetralasmus was, on examination of the matrix of distance coefficients, most clearly related to 1) Elliptio buckleyi (.749), 2) E. com- planata (.893), and 3) Fusconaia flava (.982). On this basis Uniomerus 15 classified as a genus of the tribe Pleurobemini. 3. Quincuncina We did not have data for 7 of the 15 com- parisons of the marix of 15 antisera x 41 sets of antigens (OTUs). Given this lack of data, the closest relationships seen in the matrix of taxonomic distances were: Elliptio crassidens (.726), Leptodea fragilis (.786), Elliptio lan- ceolata (.891), Tritogonia verrucosa (.895), Quadrula apiculata (.896). The minimum spanning tree showed connections of L. fragilis, E. crassidens, U. tetralasmus, and Q. apiculata to all other taxa in the Amblemini and through L. fragilis to all other taxa in the Lampsilini. The net result indicates that Quincuncina UNIONACEAN GENETIC RELATIONSHIPS 251 should be provisionally placed т the Amblemini. Verification of this placement is dependent on filling in the missing data per- mitting a more precise analysis of relation- ships. The placement in the Amblemini agrees with the grouping of genera in the Ambleminae sensu Heard 4 Guckert (1971) if one excludes Fusconaia (we have no data for Elliptoideus) but includes Megalonaias which does not deserve separate ranking in a sub- family (Megalonaiadinae). 4. Hooked glochidia There are two types of hooked glochidia. The large single pair of hooks at the periphery of the glochidial shells of Anodontinae are not homologous with the two pairs of hooks, one pair at each side of the glochidial shells of Proptera. The hook morphology is quite differ- ent in the two taxa. 5. Glochidium size We used a glochidial index (Gin) for size, where the Gin = the height of the glochidium (Hmm) x the length of the glochidium (Lmm). The glochidium of the Margaritiferinae is small (Ortmann, 1911; Baker, 1928). “Small” is de- fined as Gin = < .0036. The glochidia of the Anodontinae are “large”: the average Gin = about 0.1000. The range is from 0.078 in Alasmidonta to 0.1225 in Anodonta corpu- lenta. Most species have а Gin > 0.0900. The glochidia of the Ambleminae are “medium” sized where the average Gin = about 0.047. The smallest was that of Quadrula quadrula (Gin = 0.007; note that Gin of О. pustulosa was 0.0736); the largest was that of Cyclonaias tuberculata (Ст = 0.0867). Most had a Gin between 0.02 and 0.06 (16 of 24 = 66.7%). None was as small as seen in the Margaritiferinae; only two species (8%) had a glochidium size as large as the smallest glochidium size of the Anodontinae (Cyclonaias tuberculata and Megalonaias gigantea of our Amblemini). 6. Change in nomenclature Lampsilinae was changed to Ambleminae; Elliptionini was changed to Pleurobemini for reasons of nomenclatural priority (see Heard & Guckert, 1971, and Haas, 1969a,b). APPENDIX 3. Glossary of terms. In the follow- ing definitions the noun is followed by its adjec- tive in parentheses. Bradytixis (bradytictic): long term breeder; re- tains larvae in demibranchs except in Nearc- tic summer. Diagenae (diagenous): ectobranchous group whose ovisacs are transverse to the demi- branchs (only in Strophitus of Anodontinae). Digenae (digenous): ectobranchous; two outer demibranchs are marsupial. Ectobranch (ectobranchous): digenous; de- fined above. Eschatigenae (eschatigenous): sub-tribal taxon of Lampsilini where the lower part of the posterior region of the demibranch is marsupial. Demibranch not folded; eschati- genous state. Heterogenae (heterogenous): subtribal taxon of Lampsilini where the posterior section of the demibranch is marsupial; heterogenous state. Homogenae (homogenous): entire outer demibranch loads with glochidia forming a smooth pad; Anodontinae; Ambleminae, Gonideini, Pleurobemini, Amblemini, and Lampsilini: Longenae (in part). Longenae (longenous): subtribal taxon of Lampsilini where the lower region of the demibranch is marsupial; longenous state. Mesogenae (mesogenous): sub-tribal taxon ofthe Lampsilini where the middle section of the demibranch is marsupial. Ptychogenae (ptychogenous): sub-tribal tax- on of the Lampsilini where the lower part of outer demibranch is marsupial and folded. Tachytixis (tachytictic): short-term breeder; retains larvae in demibranchs only in Nearc- tic summer. Tetragenae (tetragenous): four demibranchs are marsupial and homogenous. APPENDIX 4. Historical account of unionid classification. One should consult Heard & Guckert (1971) for additional historical information. Lea (1858, 1863), although using an erro- neous and simplistic classification of his own devising, nevertheless wrote and illustrated many soft-tissue descriptions and thus was 252 DAVIS AND FULLER the first to develop this category of data. Had Lea not overlooked the possibility that his ob- servations could be applied to a revolutionary new type of classification, he might have be- come the important figure in the history of naiad systematics; instead, that mantle eventually fell to Simpson and ultimately to Ortmann. Sterki (1898, 1903) partly suc- ceeded where Lea had missed his opportuni- ty; he recognized that soft-tissue characters could be important in unionid classification, but he did not exploit this realization, perhaps because of his greater interest in the Sphaeriidae. He indicated that unionids should be classified on the basis of characters involving reproductive structures such as the marsupial demibranchs, the specialized marsupial areas of some demibranchs, the glochidial morphology, and duration of breed- ing season. Simpson (1900, 1914) published not only the first comprehensive account of global naiad systematics, but also the first naiad classification that purposely incorporated soft- tissue data. Moreover, his classification ar- ranged taxa according to marsupial charac- ters, thus preparing the way for more sophis- ticated work by Ortmann. Finally, Simpson's work is especially important for our study be- cause so many of his observations (some of them unique and no longer replicable because of extinction) concern Nearctic unionids. Simpson’s works not only were pro- digious, but also marked the turning point in the history of studying freshwater mussels. They pointed the way from totally inadequate 19th century conchological schemes towards Ortmann’s future classifications. Simpson's classification involved a single family and subfamily (Unionidae: Unioninae for Nearctic naiades), plus numerous further subdivisions, of the same rank, which today can be construed as tribes. The great weak- ness of the classification is that it is primar- ily monothetic, based on where the gills are loaded with glochidia in gravid females, and that his goal was a utilitarian classifica- tion. For example, Simpson (1900, 1914) was aware of the essential morphological peculi- arities of the Margaritiferinae, but classified them in his tribe “Homogenae” with all other naiades of his acquaintance that exhibit a marsupium occupying the entirety of the outer demibranch. Ortmann (1910a) was the first to ask funda- mental questions about how the organisms related to themselves and to their environ- ments. He was the first to make a synthesis of all data available while questioning how morphological structure related to function. He integrated data from shell, soft tissues, behavior, and environments. His result (1910a) was an original classification of one family and three subfamilies (Unionidae: Margaritiferinae [= Margaritaninae in those days], Unioninae, Lampsilinae). Subsequent- ly, Ortmann (1911) raised his “Margaritani- nae” to family rank and (1916) created an- other unionid subfamiy, Gonideinae, for Gonidea angulata (Lea) of the Pacific drain- age of North America. These were Ortmann's last (and only) changes of family-group taxa in comparison to his (1910a) original scheme. Ortmann correctly interpreted the unique morphological character-states that set apart the higher taxa that include the groups of 1) Margaritifera, 2) Anodonta, 3) Lampsilis, 4) Gonidea. His grouping in the Unioninae (our Pleurobemini and Amblemini) included taxa with four as well as two marsupial demi- branchs. The marsupium is not confined to restricted region of the gills as in his Lampsili- nae, and taxa do not have unique mantle structures below the branchial openings as in many Lampsilinae. It is with Ortmann’s Unioninae that we find, as did Heard & Guckert (1971), need for re-evaluation. Most subsequent classifications involve alternate interpretations of the groups of Lampsilis and Anodonta. Hannibal (1912) recognized four families of Nearctic naiades (Appendix 1). His Unionidae is a partial sub- scription to Simpson's Homogenae; the marsupia in his subfamilies Unioninae and Anodontinae are homogeneous. His Unioni- nae comprise taxa in our Ambleminae: Pleuro- bemini (partim); his Anodontinae essentially are Ortmann's and ours. His Lampsilidae are our Ambleminae: Lampsilini. He created a subfamily for Proptera, presumably because of that genus’ “ax-head” shaped glochidium. His Quadrulidae equals our Amblemini (partim) and Pleurobemini (partim). His Quadrulinae probably equals our Amblemini; his Pleurobeminae, our Pleurobemini (partim). In summary of Hannibal’s contribution, he anticipated our division of Ortmann’s Unioninae into district groups of Pleurobemini and Amblemini. Overall, however, his system is one of gross taxonomic inflation. For ex- ample, there is no justification for a higher category based on Proptera (Fig. 3). Frierson (1927) divided the Nearctic UNIONACEAN GENETIC RELATIONSHIPS 253 Unionidae into five subfamilies. Only two items of his arrangement differ significantly from ours. His Unioninae is that of Ortmann and depends on the Eurasian concept of Unio. As Heard & Guckert (1971) have shown, this concept does not adequately ac- commodate the relevant New World naiades, which we interpret as the tribes Amblemini and Pleurobemini. Our second objection to Frierson's arrangement is his Alasmidontinae. We consider the relevant genera as evolu- tionary stages within a single subfamiy Anodontinae (Fig. 2). Modell (1942, 1949, 1964) is an atavism to 19th century conchology. He created a highly controversial scheme that is monothetic, i.e. based almost solely upon a single discrimi- nant, beak sculpture. Heard & Guckert (1971) have fully discussed the artificiality of the Моде! classification. Remarkably, our Anodontinae, which we seemingly rightly regard as an integrated group both morpho- logically and immunologically are distributed by Modell between two families and sub- families, the Unionidae: Anodontinae and the Elliptionidae: Alasmidontinae. We reject Modell's classification. Morrison's (1955) classification is primarily based on the monothetic notion that the na- ture of the glochidial shell is the key to naiad classification. He opted for a three-family ar- rangement (Appendix 1). The Unionidae are taxa with hooked glochidia and divided into three subfamilies: Alasmidontinae, Anodonti- nae, Unioninae. Morrison’s Amblemidae (our tribes Amblemini, Pleurobemini, Lampsilini) are equal to our Ambleminae minus Gonidea; his Amblemini are, excepting the European Unio, equal to Ortmann’s Unioninae. Morrison's classification is rejected be- cause it is taxonomically inflated, separates morphologically and immunologically allied groups, exhibits the problems of a classifica- tion based on monothetic concepts, and, as in many of his published ideas about naiades, is supported by little or no evidence. His work is, however, laudable because very often he employed ecological information in framing his ideas. Haas (1969a) and Clarke (1973) published Classifications that are essentially rearrange- ments of Morrison’s (1955). wi À rep au ñ pe" = Щи ФА ем» à ULT аа E e. u N о Margo pre dos + pur me on or “rT Ott E DL RL = VEAIS р рва se, Ele bat > kc qt a | por e me Ze ee o par lil e dk Мы ФУМ 7 ies 5 de ¿rea Ov an МВ ре e reo ata | ADOS: сеты e a E ниц ar dr «¿rd мафия < nee | À étre a Аве fo e: Ola ga do cl en A AA AA О pi ont, ven u | | ar: rire = атм E ee a a | MEN mae пой A р | 2 y = As ya ch. Я 2 Beth dd рб КАЛИЯ РАМЕ ar cis ar UG age? ges Y ei | aras «api, | A AS e AA 4 an у A es | ES > MALACOLOGIA, 1981, 20(2): 255-266 INTERPOPULATION VARIATION IN CALCAREOUS AND PROTEINACEOUS SHELL COMPONENTS IN THE STREAM LIMPET, FERRISSIA RIVULARIS' W. D. Russell-Hunter, Albert J. Burky2 and R. Douglas Hunter3 Department of Biology, Syracuse University, Syracuse, New York 13210, U.S.A.; and the Marine Biological Laboratory, Woods Hole, Massachusetts 02543, U.S.A. ABSTRACT Ten natural populations of the North American stream limpet, Ferrissia rivularis, were studied in upstate New York, in a set of localities whose waters have a 15-fold range of dissolved calcium (4.6 to 67.6 mg/liter) and also range from oligotrophy to eutrophy. Shell component analyses (calcium carbonate, total organic carbon, and total nitrogen) are reported both as component mass-fractions (mg/g or ug/g dry weight) and as values for a “standard limpet” shell of 3.5 mm aperture length (AL). More than twofold differences occur between populations in all three components, with relatively little variation occurring within each population. Expressed “рег standard limpet,” СаСОз values for different populations range from 0.8 to 1.97 mg with no direct relationship to environmental dissolved calcium. Nominal “concen- tration ratios” of body calcium to environmental calcium range from 1,953:1 to 29,130:1. Values for total organic carbon (9.13 to 21.0 ug) and total nitrogen (2.7 to 6.69 jg) in the shells parallel each other, all C:N ratios being relatively uniform (3.0:1 to 3.4:1), and indicating that the non- calcareous components are largely proteinaceous. Although alternative hypotheses predict an inverse or a direct relationship between the organic and the calcareous components, neither 1$ shown by these populations. It appears that genetic controls of shell secretion for the two major components are inde- pendent, and that chance dispersal has resulted in some “rather inappropriate shells in certain habitats.” This irregular variation in Ferrissia is first discussed in relation to other patterns of shell component relationships known for other freshwater molluscs, including direct relationship of the mass of shell calcium carbonate to the dissolved calcium available as in Lymnaea peregra and Laevapex fuscus and the apparent “regulation” producing standard shell weights in Lymnaea palustris and Physa gyrina. The results are then discussed in relation to assessment of radio- nuclide pollution using molluscan shells from fresh waters and in their more general relationship to modes and rates of evolutionary change in freshwater faunas. INTRODUCTION Natural populations of freshwater pulmo- nate snails show extensive infraspecific physiological variation between populations (Russell-Hunter, 1964, 1978). Aspects of this in growth, fecundity and respiration have been reported (Burky, 1970, 1971; Hunter, 1975a, b; McMahon, 1973, 1975a, b; Russell- Hunter, 1953, 1961, 1964) and its evolution- ary significance discussed (Russell-Hunter, 1964, 1978). There can also be interpopula- tion variations in shell components in several species, and the present report concerns these in the freshwater limpet, Ferrissia rivularis (Say). In freshwater pulmonates—as in the major- ity of molluscs—the secreted shells have two principal components: a meshwork of protein fibers (the organic matrix) and crystalline cal- cium carbonate (Degens, Spencer & Parker, 1967; Jones, 1969; Russell-Hunter, Meadows, Apley & Burky, 1968; Russell-Hunter, Burky & Hunter, 1970). The latter is secreted in greater part after active uptake directly from environ- mental water, and in a lesser fraction after assimilation from food. In the euryoecic spe- cies, Lymnaea peregra, the thickness (and mass) of the calcareous shell varies with the calcium available in the waters (Hubendick, 1947; Russell-Hunter, Burky & Hunter, 1970). Thus, it appears that Lymnaea peregra ex- 1This investigation was supported by National Science Foundation research grants GB-36757, and DEB-7810190 to Dr. W. D. Russell-Hunter. 2Present address: Department of Biology, University of Dayton, Dayton, Ohio 45409. Present address: Department of Biological Sciences, Oakland University, Rochester, Michigan 48063. (255) 256 RUSSELL-HUNTER, ВУАКУ AND HUNTER pends about the same energy on shell- making no matter what the environmental hardness. A second pattern found in Lymnaea palustris (Hunter, 1975b) and in Physa gyrina (Hunter & Lull, 1977) involves somewhat more “regulation”: over a consid- erable range of environmental calcium val- ues, populations have shells of approximately “standard” weight at all growth stages. The case of the stream limpet, Ferrissia rivularis, in natural creek populations of upstate New York is strikingly different from both of these patterns. Although these creek waters vary over 15-fold in dissolved calcium, the highly significant differences in shell calcium found to exist between populations are not correlat- ed (Russell-Hunter, Apley, Burky 8 Meadows, 1967; Russell-Hunter, Burky 8 Hunter, 1970). Anabolic concentration ratios appeared to range from 1,609:1 to 10,615:1 and there was other circumstantial evidence of physiological races. These data on shell calcium content (for limpets from six creeks and one lake) re- ported in these two earlier notes require some correction, as a result of the improved meth- ods described below, but the significant dif- ferences and lack of environmental correla- tion remain as claimed. More recently, we have measured total organic carbon and total nitrogen in shells of limpet growth stages from ten natural populations, and published a pre- liminary abstract on seven of them (Russell- Hunter, Burky 8 Hunter, 1970). Subsequently, the shell calcium content for the same ten populations was redetermined. We now re- port environmental water conditions and the calcium, organic carbon and nitrogen con- tents of limpet shells for nine creeks and for Oneida Lake. After computing these values in terms of “standard” limpets to allow more di- rect comparisons, we discuss several hy- pothetical relationships which might be ex- pected to affect variation in shell components, compare the available data for other species, and end by briefly reviewing the nature and significance of this kind of interpopulation physiological variation in freshwater molluscs. MATERIALS AND METHODS The freshwater basommatophoran limpet, Ferrissia rivularis (Say), is ubiquitous in ap- propriate stream habitats in northeastern North America. In upstate New York, this spe- cies lives in waters ranging in calcium content from 4.6 to 67.6 mg per liter (total hardness values range from 25 to 243 mg calcium car- bonate per liter), and we have collected regu- lar population samples for other biometric studies from 53 localities. The ten localities providing populations of limpets for the pres- ent study are (in order of decreasing hard- ness): Limestone Creek (LC) near Manlius, Canandaigua Outlet (CO) at Alloway, Chit- tenango Creek (CC) at Cazenovia, a section of the shore of Oneida Lake (OL) at Shackle- ton's Point, Chenango River (CR) at Randalls- ville, Big Bay Creek (BBC) near Central Square, Fish Creek (FC) below Westvale, Slocum Creek (SC) at West Monroe, Black Creek (BC) above Cleveland, and Morgan's Hill Creek (MHC) near Truxton. All of these localities are in the central or “upstate” sec- tion of New York State (exact latitudes, longi- tudes, quadrangles and county references can be provided on request). CR and MHC are in the drainage system of the Susque- hanna River which eventually empties into Chesapeake Bay. The waters at CO drain into the Clyde division, and LC, CC, BC, FC, SC, BBC and Oneida Lake (OL) itself into the Oneida division, of the Seneca-Clyde-Oneida drainage which passes by way of the Oswego River into Lake Ontario and then on to the St. Lawrence. The environmental concentrations of dis- solved calcium and magnesium were ana- lyzed by an EDTA (ethylenediaminetetraace- tate) titration, and total hardness also deter- mined chemically at the same time. Inde- pendently the average total hardness was de- termined from conductivity measurements of samples made on every visit throughout the year. The aperture length of each limpet shell (AL) was measured by stage micrometer in 0.1 тт class intervals (Russell Hunter, 1961). Weights of shell calcium carbonate (and of “ash-free” tissue dry weights) were determined on whole limpets (starved for 48 hours). Analyses of shell organic carbon and nitrogen were run on limpet shells from which the tissues had been removed. For dry weights and shell weights, two procedures were fol- lowed. Selected individual limpets were oven- dried at 98°С to constant weight, then trans- ferred to a muffle-furnace at 475°C for 105 min. This provided an ash weight (almost en- tirely shell СаСО. in starved or laboratory-fed animals), a total dry weight, and by subtrac- tion an ash-free dry weight (or tissue weight). Other individuals were oven dried at 98°С to constant weight and then treated with an ex- FERRISSIA SHELL COMPONENTS 257 cess of 12% nitric acid (8.5% HNO,) and then washed and redried, giving two dry weights (whole limpet and tissue) and by subtraction a value for dissolved calcium carbonate. Earlier studies on limpets (Russell Hunter, Apley, Burky & Meadows, 1967) had utilized 3% nitric acid (2.2% HNO;) which gives success- ful decalcification in other snails (Hunter 4 Lull, 1976, see also Richards & Richards, 1965). In Ferrissia, the results of muffle- furnace ashing could not be reconciled with those for decalcification with 3% nitric acid. A series of trials with limpets from BC and MHC (and with stocks of Helisoma trivolvis, see Russell-Hunter 8 Eversole, 1976) showed that significantly higher values for shell cal- cium resulted from decalcification with 8.5% HNO;, and that these agreed with values ob- tained by ashing. [For Black Creek limpets, the following linear regressions of “total shell calcium carbonate” (S) in mg were computed: by 2.2% HNO:, $ = -1.44 + 0.635 AL (r- value of 0.95); by 8.5% HNO3, $ = —1.95 + 0.927 AL (r-value of 0.98); and by 475°С ash- ing, S = —1.92 + 0.933 AL (r-value of 0.99).] Two sets of calcium analyses (each on ten limpets from BC) by the chloranilic acid meth- od gave values closely corresponding to the 8.5% HNO, regression (Dr. Christopher H. Price, unpublished). A series of additional tests revealed that a maximum of only 1.3% of the “total shell calcium carbonate” resulting from our standard 475°С ashing could not subsequently be dissolved by 8.5% HNO, (Dr. Jay S. Tashiro, unpublished). There were no significant systematic differences between ashing whole animals, and ashing shells alone. With larger limpets (approximately 4.0mm AL and larger), less than 3% of the “ash” weight was attributable to the limpet bodies when these were separated from the shells; and, with smaller sizes of limpets suit- ably starved, there was no detectable differ- ence between “ash” vales for whole limpets with shells and values for shells alone. Al- though we have data from ashing for three of the populations discussed here (BC, MHC, and CC), the results presented in detail below for the ten populations and used in subse- quent computations are all derived from de- calcification with 8.5% HNO3. Total organic carbon was determined on batches (selected by aperture length) of lim- pet shells using a wet oxidation colorimetric method (Russell-Hunter, Meadows, Apley & Burky, 1968). Values for smaller limpet shells (e.g., at AL = 2.2 mm) had to be determined on batches of 24-27 individual shells of each cohort size. Analyses of total combined nitro- gen on selected batches of shells were made on a Coleman model-29 semiautomatic nitro- gen analyzer which employs a modified micro- Dumas method as described by Gustin (1960). Subsequent computational methods mostly utilized linear regressions. For some kinds of comparisons, shell CaCO,, tissue dry weight, shell C, and shell N can be computed in terms of a “standard” limpet of modal size (3.5 mm AL), as read off from regressions for each of these components against shell size. Aper- ture length (AL) in limpets such as Ferrissia rivularis and Laevapex fuscus is a better measure of size (age) than similar shell meas- urements on other planorbid or turbinate snail species. Finally, all population samples used in the analyses were collected in early sum- mer (thus avoiding any early spring complica- tions from overwinter degrowth, see Russell- Hunter & Eversole, 1976) and, as already noted, all samples involving tissues had been starved for 48 hours (thus avoiding the com- plications of inorganic gut contents, see Hunter & Lull, 1976). RESULTS In Tables 1-4 and in Fig. 1, data are ar- ranged from left to right in order of increasing calcium concentration of the habitat waters (from MHC to LC). The conditions of the abiotic environment are set out in Table 1, including concentrations of dissolved calcium and mag- nesium, the average pH, and the altitude. In Table 1 are also set out assessments of the trophic state of the habitats, of the limpet den- sities, and of the pattern of life-cycle involved. As with other freshwater molluscs, there can be infraspecific variation in the number of generations per year in different populations of Ferrissia rivularis (Burky, 1971), and in similar pulmonates including ancylid limpets (Russell-Hunter, 1964, 1978, and references therein). In the present set of ten populations, six have a simple annual life-cycle, two (CC and LC) have two generations with incom- plete replacement (that is, representatives of both spring-born and late-summer born gen- erations survive overwinter), and two (CR and CO) have two generations with complete re- placement (that is, only the second or late- summer generation overwinters to breed in the next year). The results of analyses of shell components RUSSELL-HUNTER, BURKY AND HUNTER 258 LEHRER zeit $91 46 L 866° 8£6' 108' 008' ВЕТ 631 vel (Bw) зэдши prepue]s/LO geo 11eys $19 6rZ Lez 669 969 959 Leg 869 679 0/9 (6/6w) ма edu! /Logeo 119ys 509: 0571 815 092° 959° vor Les vOL’ e66' za (Bw) £Q980 иэцз чеэи z0'€ 07€ 98 2 > ele 982 262 Loe 21126 89'€ (ww) y¡Bua] зэдши ueayy сре See Orie. Bec Сирс créel ДУбЕ 65-00 9Эу0е «= GE (ww) эбиел чбие| jaduur7] 9 9 S 9 S 9 v 1 8 9 sayoyeq jo лэашпм 6+ Se Lt Or 82 09 ze Lt GG 81 зэдши JO лэдшпм on 09 99 10 HO 999 93 os 94 HOW ayıs uoe/ndod 0 ха} ass ‘иоцеиваха лэцип} JOY ‘мод мэм ajejsdn ul suena eISSI1184 yo зиоцепдо4 ¡e/mjeu Ud} 10} Ju8]U09 ajeuoques шпюеэ [Jays ‘с FIAVL à IZ 9 Iz | De | | | | | (92 ‘Iz ‘L) 1094 sed sjadul] jo зиоцелэие) W H/1 7 7 N/A и N H = H (7 ‘W ‘H) Aisuep зэдши элцеледшо) W Е N 3 3 N O N O O (O ‘W ‘3) juawssesse olydoi 509 017 081'L 0/8 OLL'E ose Ges gee 697 JA (¡ana eas элоде y) эрпииу 90'8 10'8 10'8 0€'8 52`8 621 612 621 E YA Hd эбелэлу 661 Eve 951 Srl ee! 65 pS 8p 6€ Ge (1811/0989 Bw) sseupieu [201 FL 1781 OL ol VS SY 6+ 67 ze 60 (1811/6u1) wnıssußeu panjossıq 929 199 8'vy zer 8`65 Br gel € cl rol 9'y (18111/Bu1) шпюеэ pealossiq on O9 92 10 yO 998 93 os 98 OHW aus uonyejndog eS ‘Quewsesedei ‘|Z ‘ээ9шоэи JO ‘9z ‘aye|du109 цим JesÁ sad зиоцелэцэб ом} 10 ‘| ‘suoiyesquab ¡enuue ajBuis лэциэ элец SuONe|Ndod зэдши ay} pue :99 PUB yO Je зиоцелэиэб |епиие OM] 10; иэлм6 эле pue mo] = 7 ‘шигрэш = W ‘чбщ = H эле saisuap Jedi] au} “o1do10610 = O ‘эчаодозэш = y ‘эцаодпе = y :эле Ayanonpoid э1490 JO SJUAUISSASSE AY)) PAIPNIS SAM з/ле/пли BISSIJSY JO SUONEINDOA |елщец элэцм JOA мэм a]e3sdn ul sa1]1¡290| US] JO} SUOIJIPUO9 |[езиэшиоллиЗ ‘| FIGVL 259 FERRISSIA SHELL COMPONENTS A AAA A 202 A IO A AS 012 68 c air 96€ 82.5 8L'+ ¿ve pore 699 Z0'p (Br) зэдши рлерие}$/м 1194S 0'€ 0'€ ee Ze ve Ze Ze oe ge Ze оцел N:9 IIS 28"! 08'2 сне sce 9E°2 99'€ 26€ Gee Lys 86° (6w/67) ма IIayS/N IIaUS СУ vez He? ГЕ 20€ 662 092 s62 299 6/`5 (Br) иэболии ¡pays ues) 66'€ LL'E LL2 rd 6€'€ 80'€ 962 9L'€ ¿ve Gre (ww) yBue] зэдши ueayy YES 021 05901 OSE? 0521 St Orel Er SrO0e avez (ww) эбиел щбиз| jedur 9 S v v 9 S S 9 8 S seu9}eq jo лэашпм OL! 18 cS 9 сб 89 92 98 611 69 sjadul] JO JOQUINN Sil O9 39) 10 HO 998 93 os 98 OHW ayıs uoneindog ха} ээ$ ‘иоцеие!Аха лэцип} 104 ‘од MAN ajejsdn ul зме/пли EISSI1184 yo suoyeindod jeinyeu из} 10} оцел N:9 pue juajuoo иэбодии ||э4$ ‘+ 37891 L'OL €L'6 OAI КС! ELL 1 Ир 601 o Lz gel (Br) э4ши prepues/9 119ys ors 05'8 88`9 КН 70`8 zul ус ZO igi 15`6 (бш/б7) ма 1194yS/9 иеч$ 601 607 vr Zt L'OL 95'8 8/'8 20`6 ele 9°01 (Br) uoqueo oiueBio ¡pays ueeW v9'€ LL'E 89'€ grE ove 862 LO'E 126 9v'€ 67€ (ww) цубиа| зэдши ueayy $'6-91 6751 96-06 Brr2 буре Gb Grrl 99-94 Sle 9-96 (ww) abues ue] jaduur7 9 G G y G 9 9 9 8 p seyojeq JO лэдшпм ФЕ! 18 GL GG DD 69 ZL 98 621 95 syodwı JO JoquunN ON O9 99 10 нэ 9998 94 96 94 OHW ayıs uoneindog ха} Gas ‘иоцеиеаха лэцип} 104 “JOA MON ajejsdn ul зме/пли EISSI84 jo зиоцепдо4 jeunyeu из} 10} juajuo9 uOqueo эшеблю |IeyS ‘€ 371991 260 25 ® 20 oF | = | u | E 7) | A | с 4 Eis | a | | A < | | | a ez | | ® N u | | ® | 5 a RO a 10 | | | A A Ey | | X x = er x uu : Гы < | = | w Гы | Гы о =E WwW a x ac be le [о her INS) = lS Го Ms le ee es ls ¡ES Го Г < ® Ine | oO | O ee | a 5 Mon wiles. Sang o lO | = | © |” pa | © | | Y / 7 | | / val A | / / ZEN 7 | о BE ON | / Las ee wer / | / Ya 0 We if | А / | AS / / ГУ / RUSSELL-HUNTER, BURKY AND HUNTER A 2.0 | | | ab | | | sea | | E | | a | | = a | | | < Mie hee 2 0 ig: | + 2 Don во | N ИЕ | x m a Го | о el 5 = |u MER 15 В и : 2 AE a NE eee Ne г 2 BL NÉ E | © TES + NO 0.5 À fr ш IE |S u E I = | < ls о Го | O [о as / 7 \ I A / \ | / 4 EL aes i a! 7 / / ‘on 0 YE \ Я 2 м 7 \/ ENVIRONMENTAL DISSOLVED CALCIUM mg/l FIG. 1. Summary of shell component data for ten natural populations of Ferrissia rivularis in upstate New York. Shell carbon values (filled circles) are shown in ug organic carbon per standard shell (of 3.5 mm limpet derived from regression values). Shell nitrogen values closely parallel these (the noncalcareous component appears to be pure protein). Shell CaCO; values (filled triangles) are shown in mg per standard shell. Note that the values are only ranked in order of increasing concentrations of environmental dissolved calcium; actual values for calcium content (mg/liter) being shown to scale only on the abscissa. (shell calcium carbonate, total organic carbon, and total nitrogen) are set out in Tables 24. They are shown in the lower lines of the tables computed both as component fractions (mg/g) of dry weight (of limpets for CaCO3, and of shell for organic C and for total N), and as values for the “standard limpet” of 3.5 mm AL. The 32 regression equations of components against shell size are not set out here, but can be made available to any interested investi- gator. As shown in Table 2, there can be a twofold difference in mass of calcium carbonate (FC at 0.80 and CC at 1.97 mg) in the standard limpet. There is clearly no direct relationship to environmental dissolved calcium. Further, computaton of (somewhat arbitrary) “concen- tration ratios” between environmental calcium content and the calcium contents of whole limpet tissues (as wet weights—not shown in Table 2) show that these can range from 1,953 (LC) to 29,130 (МНС). This 15 clearly an interpopulation variable of considerable bio- energetic significance. As shown in Table 3, the organic carbon content of Ferrissia shells also shows a more than twofold range, whether expressed as dry weight component fractions (mg/g) or as values for the “standard” limpet. The highest carbon content is found in the shells of the BC population, the lowest in CO (or LC depend- ing on method of computation). Carbon val- FERRISSIA SHELL COMPONENTS 261 ues show no direct relationship with environ- mental dissolved calcium, and neither a direct nor an inverse relationship to shell calcium. Further, there is no obvious relationship of shell organic carbon content with the as- sessed productivity of the environment, or even with the productivity of the limpets them- selves (which can be crudely assessed in terms of generations times densities, rather than simple densities). However, organic car- bon content does correlate rather closely with total shell nitrogen. As shown in Table 4, the total shell nitrogen content can vary over nearly a threefold range. The highest nitrogen content is again at BC while the lowest (by both computations) is at LC. The dry weight component values for nitro- gen (line 7, Table 4) can be used with those for organic carbon (line 7, Table 3) to produce carbon:nitrogen (C:N) ratios for the limpet shells. These are all relatively uniform in the range 3.0:1 to 3.4:1 (an average value for pure animal protein would be C:N = 3.25:1, Brody, 1945; Russell-Hunter, 1970), and we can conclude that the noncalcareous com- ponent of these limpet shells is largely pro- teinaceous. In fact, the individual shell nitro- gen values are closely parallel to those for shell organic carbon, and require no further separate discussion. Accordingly, Fig. 1 pre- sents the relationship between only three vari- ables for the ten population-sites: the mean values for dissolved calcium in the environ- mental waters, the noncalcareous compo- nents of the limpet shells as micrograms organic carbon per standard shell, and the calcareous component as milligrams CaCO, per standard shell. DISCUSSION Infraspecific (interpopulation) physiological variation in growth, fecundity, life-cycle pat- tern and respiration, as reported for many freshwater molluscs, appears in a number of cases to be based on distinct genotypes (“physiological races”). In such cases, trans- fer experiments between population-sites have shown that particular patterns of fecun- dity or ratios of shell dimensions are retained by “foreign” stocks introduced to other sites where the “native” snails have different char- acteristics. In other cases, notably of growth rates and size at maturity, stocks transferred to waters of markedly different hardness or trophic conditions take on similar external characteristics to those shown by the natural population at the sites. Apart from genetic controls of shell secre- tion, variation in the two principal components of the shells of freshwater snails—the crystal- line calcium carbonate and the organic matrix—could be seen as depending upon available energy for shell-making (reflecting trophic conditions) and available precursors for components (reflecting on the one hand, available environmental calcium, and on the other hand, trophic conditions once again). A starving snail would not secrete much shell protein. Our present studies cannot discrimi- nate among the noncalcareous components of the limpet shell, all shell protein (including both organic matrix fibers and periostracal sheets, and perhaps encompassing a poly- saccharide fraction) being expressed as total organic carbon or as total combined nitro- gens. Future work may allow both finer bio- chemical discrimination and more specific structural allocation. Recent X-ray diffraction studies (Weiner 8 Traub, 1980) have con- firmed that the fibers of the organic matrix are a silk-like B-fibroin protein, as earlier sug- gested by studies on amino acid residues by Degens and his associates (Degens, Spencer & Parker, 1967; Ghiselin, Degens, Spencer 8 Parker, 1967; Degens, 1976), and by ultra- structural studies (Jones, 1969). Similarly, it is almost certain that the polysaccharide fraction found in certain mollusc shells is chitin. For the limited purposes of this discussion, secre- tion of all components of the organic fraction of the shell can be regarded as energy-con- suming and dependent on trophic input. Populations of molluscs with calcareous shells are found in fresh waters with more than 100-fold range in concentrations of dis- solved calcium (Boycott, 1936; Russell- Hunter, 1964, 1978). Several workers have presented clear experimental evidence from laboratory cultures of direct effects of calcium concentration on the growth and fecundity of snails (Williams, 1970b; Harrison, Williams & Greig, 1970; Thomas, Benjamin, Lough & Aram, 1974). Correlations of environmental calcium with field distribution patterns and abundance have been demonstrated for sev- eral snail species (Williams, 1970a; McKillop & Harrison, 1972; Dussart, 1976, 1979). In temperate regions of the world, extremely soft waters (calcium concentrations <3 mg/liter) can support only about 5% of the molluscan species of the region, moderately soft waters (Са < 10 mg/liter) can support about 40%, 262 RUSSELL-HUNTER, BURKY AND HUNTER intermediate waters (10 to 25 mg/liter) can support up to 55%, with hard waters (Ca >25 mg/liter) being required for the rest (Boycott, 1936; Macan, 1950; Russell-Hunter, 1957, 1964, 1978). However, it is noteworthy that most of those species tolerant of low cal- cium could survive in, and are sometimes found in, harder waters (Russell-Hunter, 1964). Although Dussart (1976) had claimed that environmental calcium level was a major determinant of field abundance for several molluscan species, his more recent multiple regression analyses (Dussart, 1979) suggest that in some species, correlation is with other cations associated with water hardness rather than with calcium itself. Future experimental work may show that exclusion of “soft-water species” of molluscs from certain waters of high mineral content does not result from high calcium content as such. The present paper reports populations of the freshwater limpet Ferrissia rivularis in waters with a nearly fifteen-fold range (4.6 to 67.6 mg/liter) of calcium content. The ex- tremely euryoecic freshwater snail Lymnaea peregra can undoubtedly colonize an even wider range. In terms of organic productivity, waters supporting freshwater snails can again vary widely. Lymnaea peregra 1$ found in the most oligotrophic mountain lakes, but also can occur in eutrophic (even hypertrophic, or mildly polluted) waters. Again the range of Ferrissia rivularis (Table 1) is somewhat less but still extensive. Direct environmental ef- fects on variation in shell components in F. rivularis are not apparent (Fig. 1). The lightest shells in terms of calcium occur not in the three populations from the softest waters but at the two somewhat harder sites, the heavi- est shells in waters of intermediate hardness (CC), and the overall range of apparent con- centration ratios runs from 1,953:1 to 29,130:1. If ratios of secretion of shell calcium and shell protein both depend upon levels of en- ergy turnover, one might have expected a direct relationship between the two com- ponents (and possibly some relationship to the general trophic state of each habitat). Our ten limpet populations do not show this (Fig. 1). On the other hand, one could hypothesize that the adaptive need for a certain level of mechanical protection by the shell could re- sult in an inverse relationship between shell calcium and shell protein. [In certain land snails of the tropical rain forest, relatively un- calcified shells have unusually massive pro- teinaceous layers, and in certain freshwater sphaeriid clams there are supportive data for a similar inverse relationship (Burky, Benja- min, Catalano & Hornbach, 1979)]. Again, our ten natural populations of F. rivularis show no evidence of such an inverse relationship (Fig. 1). К should be noted that in content of shell protein and of shell calcium (as even more clearly in other measurable characters), the variation within the majority of single popula- tions is very much /ess than the range of vari- ation for the species as a whole. It seems that genetic controls of shell se- cretion for the two major components are т- dependent, and that the chances of genetic dispersal among the isolated creek popula- tions of this limpet have resulted in an irregu- lar distribution of shell forms. Obviously this anomalous variation in components found in Ferrissia rivularis differs from the patterns found in other freshwater snails. In Lymnaea peregra, the mass of calcium carbonate in the shell varies directly with cal- сит available. The shell component differ- ences in another North American freshwater limpet, Laevapex fuscus, are markedly less than those discussed for Ferrissia, but shell calcium content increases with calcium con- centration (McMahon, 1973, 1975a). As re- gards the calcareous component then, in Laevapex and Lymnaea peregra, the shells of variable mass in different populations could result from similar energy expenditures in shell-making. A third pattern of relationship between shell calcium and the environment has been demonstrated for Lymnaea palustris (Hunter, 1972, 1975b), where a survey of fourteen population-sites showed that the ratio of shell calcium to whole animal dry weight changes little throughout growth, and does not vary greatly between populations. This proved true over a wide range of calcium concentrations of environmental water, and represents a “regulation” unusual in a spe- cies which shows great interpopulation varia- tion in other respects. Physa gyrina (Hunter & Lull, 1977) appears to show similar “regula- tion.” Hunter & Lull (1976, 1977) also studied natural populations of Physa integra and Helisoma anceps, and for these two species, there was no relationship between shell cal- cium to tissue ratios (which varied greatly from population to population) and the calci- um concentrations of their environmental waters. In this “irregular” variation these spe- cies resembled the populations of Ferrissia rivularis described above. However, Hunter & FERRISSIA SHELL COMPONENTS 263 Lull (1977) claim that a similar ranking be- tween these two species in seven population- sites where they co-exist is evidence for a possible relationship to trophic conditions. Unfortunately, there are no data on shell pro- tein from these species-populations of the sort we have presented for Ferrissia. Thus, at least four “patterns” of shell cal- cium relationship can be discerned in the data already available on freshwater pulmonates. These are: first, a direct relationship as in Lymnaea peregra and Laevapex fuscus be- tween shell calcium and environmental hard- ness; secondly, a seeming “regulation” of calcareous shell secretion, as in Lymnaea palustris and Physa gyrina, resulting in shells of standard weight for size categories with each species; thirdly, a relationship between shell calcium secretion and general bio- energetic turnover (or trophic) rates, as was claimed for Physa integra and Helisoma anceps; and fourthly, great variation between (but not within) populations, as in Ferrissia rivularis, reflecting an irregular distribution of genetic “forms” neither obviously clinal nor adaptive (as claimed here). The fifth possible relationship—an inverse relationship between shell calcium and environmental calcium— has never been found in a pulmonate nor in other freshwater gastropods. However, Agrell (1949) found, in populations of the freshwater bivalve Unio tumidus in Sweden, that shell weights decreased from oligotrophic to eutrophic waters. For two other species of Unio and one of Anodonta, he found increas- ing shell weights with “rising trophic degree,” essentially as claimed for Physa integra and Helisoma anceps. “Over-compensation” in the form of in- creased calcium storage in populations in low calcium environments is better documented in plants, where pairs of closely-related species and subspecies have long been known to oc- cur (Tansley, 1917; Salisbury, 1920; De Silva, 1934). Closer parallels to our possible physio- logical races in the limpet Ferrissia, irregularly distributed with no obvious geographic clines, are provided by the economically important grass, Festuca ovina (Bradshaw and Snaydon, 1959), and the ecologically impor- tant microorganism, Azotobacter (Bullock, Bush & Wilson, 1960). In the case of Festuca, population differentiation has resulted т “races,” termed edaphic ecotypes, which when cultivated in a range of calcium levels show significantly different responses т growth rates and patterns. In many pulmonates, shell shape is another infraspecific variable. Extensive data have been collected on shell biometrics in Ferrissia rivularis from a set of 53 populations in a vari- ety of localities in upstate New York of varying trophic conditions and water hardness, and with different degrees of isolation from each other (Russell-Hunter and Nickerson, unpub- lished; Russell-Hunter, 1978). Certain shell- ratios (including isometric “roundness” т marginal growth) seem to be rather rigidly genetically determined, while others (allo- metric “steepness” of the cone) reflect local growth patterns as modified by trophic condi- tions. None of the present data on interpopu- lation variation in shell components can be directly related to either kind of shell-ratios as determined for this set of populations. Despite all this, it seems reasonable to hypothesize that, in Ferrissia, differences be- tween populations in both shell-calcium se- cretion and shell-protein synthesis and secre- tion are under independent genetic control, and that the different genomes are “irregular- ly” (not necessarily adaptively, nor in geo- graphic clines) distributed among the isolated creek populations as a result of the stochastic element introduced by passive dispersal of single propagule individuals. Short reviews of evidence on passive dispersal of freshwater molluscs are provided by Rees (1965) and Russell-Hunter (1978). Two consequences of this kind of infraspecific variation remain to be discussed: one related to assessments of radionuclide pollution in fresh waters, and the other to broader aspects of evolutionary processes in freshwater animals. Molluscs of various sorts have been pro- posed as indicators of environmental radio- contamination with strontium-90 (Nelson, 1964; Rosenthal, Nelson & Gardiner, 1965) which they found to be accumulated indis- criminately with calcium. Likins, Berry & Posner (1963) claimed relative discrimination against strontium in Australorbis glabratus, but Van Der Borght and Van Puymbroeck (1964, 1966) demonstrated active uptake di- rect from environmental water of both calcium and strontium in pulmonates such as Lymnaea stagnalis, L. auricularia and Planorbarius corneus. Further, their work showed that 80% of shell calcium gained by growing snails comes directly from the water, and only 20% indirectly through their food. Thus, it becomes important that some fresh- water mollusc species have populations which differ significantly from each other in 264 the extent to which they do concentrate cal- cium from the environment. It is worth noting that the range of apparent concentration ratios (1,953:1 to 29, 130:1) for Ferrissia popu- lations noted above do not necessarily meas- ure calcium transport costs through succes- sive physiological “compartments” on the way to the shell, nor do they assess relative bioenergetic expenditures by the stocks in any actuarial framework of assimilation and growth. They are sufficiently different, how- ever, to suggest that any proposed use of a freshwater mollusc species in a biological assay of strontium-90 pollution would have to be based upon a set of stocks known to be genetically uniform in such aspects of calcium metabolism. Finally, these data on interpopulation varia- tion in the calcareous and proteinaceous shell components in Ferrissia, and the subsequent deductions regarding their bases in an “ir- regular” distribution of genetic units by the chances of passive dispersal in the establish- ment of isolated creek populations can be added to lists of other types of physiological variation already documented for freshwater molluscs. All provide evidence that the rates and modes of evolutionary change which have worked to produce present (and past) freshwater molluscan faunas, differ markedly from those which have operated for similar animals living in the sea or on land. As more fully discussed elsewhere (see, for example, Russell-Hunter, 1961, 1964, 1978), a domi- nant characteristic of the freshwater environ- ment is that the transience of most freshwater habitats in time, along with their spatial limita- tions and discontinuity, results in animal spe- cies distributed with much small-scale and short-term isolation of populations. Compara- tively little full speciation occurs in freshwater animals like gastropod molluscs, since this is prevented by sufficient gene exchange result- ing from limited and rare transfers of individ- uals between populations by passive dis- persal. In contrast, much infraspecific inter- population variation can and does occur be- cause of the short-term, smaller scale isola- tion of discretely panmictic population-units. Where the high levels of interpopulation varia- tion in some physiological or morphological characteristic have been investigated in detail there is always one consistent feature. The amount of variation found within any single population is always much less than the range of that character's variation for the species as a whole. Considering certain other features of RUSSELL-HUNTER, BURKY AND HUNTER the physiology of freshwater molluscs (partic- ularly respiration), one also encounters stocks with exceptionally high levels of adaptive plasticity. All of these features of freshwater molluscs reflect the environmental discontinu- ities of time and space during their evolution. The irregular distribution of shell components found in these creek populations of Ferrissia again reflects the peculiarities of dispersal, gene-flow, and evolutionary rates in such freshwater habitats. ACKNOWLEDGMENTS The senior author is particularly indebted to Dr. Jay Shiro Tashiro for much help with the necessary redeterminations of the calcareous component in limpet samples. We are most grateful to Dr. Richard P. Nickerson for some additional organic determinations, and we must also thank Drs. Christopher H. Price and Arnold G. Eversole, for help with the field work and in the laboratory. The senior author 1$ also grateful to Drs. Robert A. Browne, Jay Shiro Tashiro, and David W. Aldridge, and to Wendy W. Lull and Perry Russell-Hunter, for help checking data transcriptions and com- putations; and to T. Braun, Catherine M. Herrity, Ann Evans, and Colette O'B. McMahon for help in manuscript preparation. REFERENCES CITED AGRELL, |., 1949, The shell morphology of some Swedish unionides as affected by ecological conditions. Arkiv för 200/041, 41A(15): 1-30. BOYCOTT, A. E., 1936, The habitats of fresh-water Mollusca in Britain. Journal of Animal Ecology, 5: 116-186. BRADSHAW, А. D. 8 SNAYDON, R. W., 1959, Population differentiation within plant species in response to soil factors. Nature, 183: 129-130. BRODY, S., 1945, Bioenergetics and Growth. Reinhold, New York, 1023 p. BULLOCK, С. L., BUSH, J. A. 4 WILSON, Р. W., 1960, Calcium requirements of various species of Azotobacter. Proceedings of the Society for Experimental Biology and Medicine, 105: 26-30. BURKY, A. J., 1970, Reverse acclimation at low temperature in the stream limpet, Ferrissia rivularis (Say). Acta Cientifica Venezolana, 21: BURKY, A. J., 1971, Biomass turnover, respiration and interpopulation variation in the stream lim- pet, Ferrissia rivularis (Say). Ecological Mono- graphs, 41: 235-251. FERRISSIA SHELL COMPONENTS 265 BURKY, A. J., BENJAMIN, М. A., CATALANO, D. М. & HORNBACH, D. J., 1979, The ratio of cal- careous and organic shell components of fresh- water sphaeriid clams in relation to water hard- ness and trophic conditions. Journal of Mol- luscan Studies, 45: 312-321, DEGENS, E. T., 1976, Molecular mechanisms on carbonate, phosphate and silica deposition in the living cell. Topics in Current Chemistry, 64: 1- 112: DEGENS, E. T., SPENCER, D. W. & PARKER, R. H., 1967, Paleobiochemistry of molluscan shell proteins. Comparative Biochemistry and Physi- ology, 20: 553-579. DE SILVA, B. L. T., 1934, The distribution of “Calcicole” and “Calcifuge” species in relation to the content of the soil in calcium carbonate and exchangeable calcium, and to soil reaction. Journal of Ecology, 22: 532-553. DUSSART, G. B. J., 1976, The ecology of fresh- water molluscs in North West England in relation to water chemistry. Journal of Molluscan Studies, 42: 181-198. DUSSART, G. B. J., 1979, Life cycles and distribu- tion of the aquatic gastropod molluscs Bithynia tentaculata (L.), Gyraulus albus (Müller), Planorbis planorbis (L.) and Lymnaea peregra (Müller) in relation to water chemistry. Hydro- biologia, 67: 223-239. GHISELIN, М. T., DEGENS, Е. Т., SPENCER, D. W. 8 PARKER, В. H., 1967, A phylogenetic sur- vey of molluscan shell matrix proteins. Breviora of the Museum of Comparative Zoology, 262: 1-23. GUSTIN, G. M., 1960, A simple rapid automatic micro-Dumas apparatus for nitrogen determina- tion. Microchemical Journal, 4: 43-54. HARRISON, А. D., WILLIAMS, N. V. 8 GREIG, G., 1970, Studies on the effects of calcium bicar- bonate concentrations оп the biology of Biomphalaria pfeifferi (Krauss) (Gastropoda: Pulmonata). Hydrobiologia, 36: 317-327. HUBENDICK, B., 1947, Die Verbreitungsverhált- nisse der limnischen Gastropoden in Súd- schweden. Zoologiska Bidrag frán Uppsala, 24: 419-559. HUNTER, R. D., 1972, Energy budgets and phys- iological variation in natural populations of the freshwater pulmonate Lymnaea palustris. Ph.D. thesis, Syracuse University, 115 p. (Dissertation Abstracts, 33, 5066-B, order no. 73-9534). HUNTER, R. D., 1975a, Growth, fecundity, and bioenergetics in three populations of Lymnaea palustris in upstate New York. Ecology, 56: 50- 63. HUNTER, R. D., 1975b, Variation in populations of Lymnaea palustris in upstate New York. Ameri- can Midland Naturalist, 94: 401-420. HUNTER, R. D. & LULL, W. W., 1976, A compari- son of two methods for estimating the weight of inorganic materials in molluscs. Malacological Review, 9: 118-120. HUNTER, В. 0. 8 LULL, W. W., 1977, Physiologic and environmental factors influencing the cal- cium-to-tissue ratio in populations of three spe- cies of freshwater pulmonate snails. Oecologia, 29: 205-218. JONES, T., 1969, Aspects of calcification in gas- tropods. Proceedings of the Malacological Society of London, 38: 549-550. LIKINS, В. C., BERRY, E. С. & POSNER, A. S., 1963, Comparative fixation of calcium and stron- tium by snail shell. Annals of the New York Academy of Science, 109: 269-277. MACAN, T. T., 1950, Ecology of fresh-water Mol- lusca in the English Lake District. Journal of Ani- mal Ecology, 19: 124-146. MCKILLOP, W. В. & HARRISON, A. D., 1972, Dis- tribution of aquatic gastropods across an inter- face between the Canadian Shield and lime- stone formations. Canadian Journal of Zoology, 50: 1433-1445. МСМАНОМ, В. F., 1973, Respiratory variation and acclimation in the freshwater limpet, Laevapex fuscus. Biological Bulletin, 145: 492-508. МСМАНОМ, В. F., 1975a, Growth, reproduction and bioenergetic variation in three natural popu- lations of the freshwater limpet Laevapex fuscus (C. B. Adams). Proceedings of the Malacologi- cal Society of London, 41: 331-351. MCMAHON, В. F., 1975b, Growth, reproduction and life-cycle in six Texas populations of two species of freshwater limpets. American Mid- land Naturalist, 94: 174-185. NELSON, D. J., 1964, Biological vectors and reser- voirs of strontium-90. Nature, 203: 420. REES, W. J., 1965, The aerial dispersal of mol- luscs. Proceedings of the Malacological Society of London, 36: 269-282. RICHARDS, N. J. 4 RICHARDS, S. W., 1965, Ef- fect of decalcification procedures on the dry weights of benthic invertebrates. Limnology and Oceanography, 10: 469-471. ROSENTHAL, С. M., Jr., NELSON, О. J. 4 GARDI- NER, D. A., 1965, Deposition of strontium and calcium in snail shell. Nature, 207: 51-54. RUSSELL HUNTER, W., 1953, On the growth of the fresh-water limpet, Ancylus fluviatilis Müller. Proceedings of the Zoological Society of Lon- don, 123: 623-636. RUSSELL HUNTER, W., 1957, Studies on fresh- water snails at Loch Lomond. Glasgow Univer- sity Publications, Studies on Loch Lomond, 1: 56-95. RUSSELL HUNTER, W., 1961, Life cycles of four freshwater snails in limited populations in Loch Lomond, with a discussion of infraspecific varia- tion. Proceedings of the Zoological Society of London, 137: 135-171. RUSSELL HUNTER, W., 1964, Physiological as- pects of ecology in non-marine molluscs, р. 83- 126 т WILBUR, К. М. & YONGE, С. M. eds. Physiology of Mollusca, Vol. 1. Academic Press, New York and London. RUSSELL-HUNTER, W. D., 1970, Aquatic pro- ductivity. Macmillan, New York, 306 p. 266 RUSSELL-HUNTER, ВУАКУ AND HUNTER RUSSELL-HUNTER, W. D., 1978, Ecology of freshwater pulmonates, p. 335-383 in FRET- TER, V. 8 РЕАКЕ, J. F., eds., Pulmonates, Vol. 2A: Systematics, Evolution and Ecology. Aca- demic Press, New York and London. RUSSELL-HUNTER, W. D. 4 EVERSOLE, A. G., 1976, Evidence for tissue degrowth in starved freshwater pulmonate snails (Helisoma trivolvis) from tissue, carbon, and nitrogen analyses. Comparative Biochemistry and Physiology, 54A: 447-453. RUSSELL-HUNTER, W. 0., BURKY, А. J. 4 HUNTER, R. D., 1970, Interpopulation variation in shell components in the stream limpet, Ferrissia rivularis. Biological Bulletin, 139: 402. RUSSELL HUNTER, W., APLEY, М. L., BURKY, А. J. & MEADOWS, В. T., 1967, Interpopulation variations in calcium metabolism in the stream limpet, Ferrissia rivularis (Say). Science, 155: 338-340. RUSSELL-HUNTER, W. D., MEADOWS, R. T., APLEY, M. L. 8 BURKY, A. J., 1968, On the use of a “wet-oxidation” method for estimates of total organic carbon in mollusc growth studies. Pro- ceedings of the Malacological Society of Lon- don, 38: 1-11. SALISBURY, E. J., 1920, The significance of the calcicolous habit. Journal of Ecology, 8: 202- 215: TANSLEY, А. G., 1917, On competition between Galium saxatile (G. hercynicum Weig.) and Galium sylvestre Poll. (G. asperum Schreb.) on different types of soil. Journal of Ecology, 5: 173-179. THOMAS, J. D., BENJAMIN, M., LOUGH, A. 4 ARAM, R. H., 1974, The effects of calcium in the external environment on the growth and natality rates of Biomphalaria glabrata (Say). Journal of Animal Ecology, 43: 839-860. VAN DER BORGHT, O. & VAN PUYMBROECK, S., 1964, Active transport of alkaline earth ions as physiological base of the accumulation of some radionuclides in freshwater molluscs. Nature, 204: 533-534. VAN DER BORGHT, O. & VAN PUYMBROECK, S., 1966, Calcium metabolism in a freshwater mollusc: quantitative importance of water and food as supply for calcium during growth. Nature, 210: 791-793. WEINER, S. & TRAUB, W., 1980, X-ray diffraction study of the insoluble organic matrix of mollusk shells. F.E.B.S. Letters, 111: 311-316. WILLIAMS, N. V., 1970a, Studies on aquatic pul- monate snails in Central Africa. 1. Field distribu- tion in relation to water chemistry. Malacologia, 10: 153-164. WILLIAMS, N. V., 1970b, Studies on aquatic pul- monate snails in Central Africa. 2. Experimental investigation of field distribution patterns. Mala- cologia, 10: 165-180. MALACOLOGIA, 1981, 20(2): 267-289 THE ARENOPHILIC RADIAL MANTLE GLANDS OF THE LYONSIIDAE (BIVALVIA: ANOMALODESMATA) WITH NOTES ON LYONSIID EVOLUTION! Robert S. Prezant College of Marine Studies, University of Delaware, Lewes, Delaware 19958, U.S.A. ABSTRACT Two of the three marine genera of Lyonsiidae (Lyonsia and Entodesma) possess small, multicellular glands that line at least a portion of their mantle edge. The glands, which are evenly dispersed along the entire mantle edge of Lyonsia but may regress quantitatively with overall growth in some Entodesma, are modifications of the inner epithelium of the outer mantle fold. The gland system in both genera is composed of a central, club-shaped gland of two basic cell types and a surrounding secretory sheath. The latter is derived from the outer fold epithelium in Entodesma and the middle fold epithelium in Lyonsia. Central gland cells and surrounding sheath cells show a complex ultrastructure typical of secretory cells. Roughly triangular or ovoid medial cells of the central gland and flat surrounding sheath cells secrete a меаКу acidic mucopolysaccharide over a thick glycoprotein layer produced by tall flask-shaped cells of the central gland. In Lyonsia, glands open into the periostracal groove and secrete above the periostracum. In Entodesma they open distal to the groove and the secretion periodically pene- trates and emerges above the periostracum. The secretion from Entodesma may possess a protease that allows the localized dissolution of periostracum. The mucoid secretion is active in adhesion of foreign particles to the exterior surface of the shell, and may play a role in stabilizing or protecting the thin-shelled Lyonsia or juvenile or smaller sized species of Entodesma. The mucopolysaccharide component of the bilayered secretion may act as a lubricant during release of the more viscous glycoprotein, or as an initial adhesive following release. In Entodesma this polysaccharide layer may also form a protective shield buffering other tissues from potential enzymatic activity. Based upon structure and location of arenophilic glands, as well as shell ultrastructure, general anatomy and specific habits, an early split in lyonsiid phylogeny is hy- pothesized. From an ancestral, free-living Lyonsia-like bivalve, two branches diverged; one branch produced Lyonsia, and the second the Entodesma lineage. Mytilimeria, the third marine lyonsiid genus, lost the arenophilic glands, and is an offshoot of the Entodesma stock. INTRODUCTION The molluscan mantle edge has provided an important vehicle for divergent evolution and resulting diversity within the phylum. The wide variety of functions that this organ has developed (Dakin, 1928; Fretter & Graham, 1954; Thompson, 1960; Hillman & Shuster, 1960, 1966; Hodgkin, 1962; Hillman, 1964, 1968, 1969; Gilmour, 1967; Muscatine, 1967; Carriker, 1972; Prezant, 1979a) is a repre- sentation of its phylogenetic significance. The role of the mantle in molluscan evolution has been discussed by Stasek 8 McWilliams (1973). Part of this phylogenetic progression has included the development of clusters of mucocytes along specific regions of the mantle. Hillman & Shuster (1960, 1966) and Hillman (1964, 1969) found two different areas of the mantle edge of Mercenaria mercenaria specialized for the secretion of mucoid products and suggested that they function in cleansing the mantle cavity, and in the production and elaboration of shell ma- terial, respectively. Hypobranchial glands of numerous prosobranchs secrete copious amounts of mucus (Ronkin & Ronkin, 1951; Ronkin, 1952), as do accessory mantle folds of many chitons (Stasek 8 McWilliams, 1973). In some scaphopods, such as Dentalium, the anterior mantle rim is swollen with subepi- thelial mucocytes that actively aid in binding extraneous materials and debris for expulsion (Stasek 8 McWilliams, 1973). Mucoid se- cretions play other roles in molluscs ranging University of Delaware, College of Marine Studies, Contribution No. 153. (267) 268 PREZANT from pedal lubricants in gastropods to food carriers and condensors in bivalves, to court- ship in many terrestrial slugs (Hyman, 1967). Mucoid secretions seem particularly im- portant among many bivalves of the subclass Anomalodesmata. Thracia phaseolina, a burrower in shelly gravels, produces mucus- lined inhalant and exhalant tubes that pene- trate the surface and allow access of func- tional feeding and respiratory currents from its burrow (Yonge & Thompson, 1976). Similar mucus-lined tubes are formed by Cochlo- desma praetenue, but the exhalant siphon does not penetrate the substratum surface and instead lies in a horizontal plane as does the animal (Allen, 1958). Lyonsia hyalina has a series of small, multicellular glands lining the mantle edge that secrete a mucoid sub- stance over the periostracum (Prezant, 1979a). This secretion, in conjunction with numerous shell spinules that cover the shell exterior (Prezant, 1979b), function in ad- hesion of foreign material to the shell. Similar glands, almost certainly of similar adhesive function and termed “radial mantle glands” (Allen & Turner, 1974), were previously found in some deep-sea Verticordiidae by Allen 8 Turner (1974). For the purposes of this paper, these organs are termed arenophilic radial mantle glands based upon their role in sand or foreign particle adhesion. Many other species of Lyonsia as well as Entodesma, among the Lyonsiidae, have sediment adhering to their periostracum; it has been hypothesized that they, too, pos- sess arenophilic mantle glands (Prezant, 1979a). The third genus of marine Lyonsiidae, Mytilimeria, may also possess some modifi- cation of the mantle and periostracum that al- lows a tight adhesion within its tunicate host. Yonge (1952) suspected that this adhesion might be a reflection of attachment to a still fluid or “sticky” periostracum. Prezant (1980a) found the periostracum of М. nuttalli to be covered with small crater-like pores that may play a role in this adhesion. Contrary to the sessile habits of Mytilimeria, Lyonsia is a mobile bivalve found partially buried in fine sediments where it is loosely secured by a few, thin byssal threads. Ento- desma, containing the largest and thickest shelled members of the Lyonsiidae, occurs nestled within crevices along rocky shore- lines, ог among tunicates, sponges or algal holdfasts. Members of each genus show some indication of an adhesive periostracum although it is most evident in the sand- covered Lyonsia. Prezant (1979a) noted that the sand cover of L. hyalina serves three purposes: protection of the thin shell, camou- flage, and increased stabilization in the sub- stratum due to increased external surface area and added weight. The ecologic diver- gence of the Lyonsiidae may parallel the development of arenophilic mantle glands. Yonge (1952) envisaged ап orthogenetic lineage within the Lyonsiidae, originating in the Eocene (Yonge & Morton, 1979), based upon intrafamilial morphology and life styles, from a free-living Lyonsia to a more sedentary Entodesma to a sessile Mytilimeria. The present study has two main goals: first to ex- plore the distribution, structure, and function of arenophilic radial mantle glands among the Lyonsiidae; and second, to examine the ques- tion of lyonsiid evolution with respect to func- tional morphology and arenophilic mantle glands. MATERIALS AND METHODS The following live specimens of lyonsiids were used in this research: Lyonsia floridana from fine to medium grained shelly-sand in shallow waters at Blind Pass, Sanibel, Flori- da; L. californica and Mytilimeria nuttalli from the Venice, California coast; L. hyalina from a muddy sand substratum at a depth of about 15 m in Delaware Bay, Delaware; and Ento- desma saxicola, found nestled within rocky crevices just subtidally along Shaw Island, Washington. М. nuttalli was found embedded in compound tunicates, probably Eudistoma psammion. Previously preserved specimens (usually formalin fixed and ethanol preserved) which were examined included: L. gouldii from San Diego, California; L. pugetensis from Chignik Bay, Alaska; Entodesma fretalis from Corral Bay, Chile; E. chilensis from Corral Bay, Chile (Zenker’s fluid fixed, ethanol pre- served); E. beana from Vieques Island; E. patagonica from Argentina; and E. saxicola from British Columbia. Entodesma fretalis and E. chilensis were both found among the tuni- cate Pyura chilensis. Non-lyonsiid Anoma- lodesmata examined for comparative pur- poses included: Pandora gouldiana collected live нот Delaware Bay; Репр/ота fragile col- lected live from Massachusetts Bay; and for- malin fixed-ethanol preserved specimens of Pandora inaequivalvis and Cochlodesma praetenue from Isle of Cumbrae, Scotland; and Laternula truncata from Loo Bay, Lubang Island, Philippines. LYONSIID MANTLE GLANDS 269 Live animals were maintained for short periods of time prior to fixation on running seawater tables with recirculating, sand fil- tered water at 12°С and 30 ppt salinity. Some were fed Thalassiosira pseudonana and /so- chrysis galbana. For histological purposes, live animals were relaxed in a 7% magnesium chloride solution, and fixed in either Zenkers or Hollande Bouin's fluid. Small specimens were fixed in toto while small parts of the mantle of larger specimens were fixed separately. Samples were fixed for 18-24 hr and then washed overnight in running tap water, dehydrated in a consecutive series of increasing concentra- tions of ethanol and embedded in polyester wax. Sections were cut at 5-10 ит, mounted on albuminized slides and stained with either Heidenhain's iron or Groats hematoxylin counterstained with eosin Y, Heidenhain's Azan or a modification of the Pantin trichrome (Prezant, 1979a). Specimens used for histo- chemical analysis of polysaccharides were fixed in Rossman’s fluid or a mixture of 9 parts absolute ethanol and 1 part formalin. For pro- tein histochemistry, specimens were fixed in 10% formalin buffered to pH 7.2 with sodium phosphate. Specimens for histochemistry were embedded in paraffin or polyester wax and sectioned at 6-10 um. The histochemical tests, along with their specificity, are noted in Table 1. Only living specimens were used for histo- chemistry or transmission electron micro- scopic observations. These included Lyonsia californica, L. floridana, and Entodesma saxi- cola although some polysaccharide histo- chemistry was performed on other alcohol fixed species. Small pieces of the mantle were fixed in cold Anderson's glutaraldehyde fixative in a sodium phosphate buffer at pH 7.2 for one hour for ultrastructural examina- tion. Following primary fixation, specimens were washed in several changes of phos- phate buffer solution, post-fixed in 2.0% osmium tetroxide in a phosphate buffer, also TABLE 1. Histochemical tests used in analysis of arenophilic radial mantle glands and gland secretions of lyonsiid bivalves. Test 2% toluidine blue in 1% sodium borate periodic acid Schiff (PAS), alcoholic metachromatic substances mucopolysaccharides periodate reactive material Specificity Reference Humason, 1972 Pearse, 1968 Barka & Anderson, 1965 potential mucins PAS/diastase 1% alcian blue, pH 1.0 1% alcian blue, pH 2.5 0.1% alcian blue, pH 5.7 at critical electrolyte levels: 0.1 М MgChb -4H50 0.2 М MgCh:4H0 0.4 М MgClo-4H2O 0.5 М MgClo-4H2O 0.8M MgCh:4H0 1.0 М MgClo-4H2O 1% alcian blue/PAS, aqueous glycogen sulfated mucosubstances weakly acidic mucosubstances periodate reactive material, alcinophilic mucosubstances Pearse, 1968 Pearse, 1968 Pearse, 1968 below 0.3 M sulfated mucins and Scott et al., 1964 glucosaminoglucuronoglycans containing carboxyl groups, 0.2 M and above only sulfated mucosub- stances (chondroitin sulfates up to 0.5 M, some sulfomucins and heparin up to 1.0 M) Pearse, 1968 Pearse, 1968 Thompson, 1966 (all but most strongly acidic) ninhydrin Schiff mercury-bromphenol blue mercury orange dihydroxy-dinaphthyl-disulfide mercury orange/thioglycerol dihydroxy-dinaphthyl-disulfide/thioglycerol protein-bound amines proteins sulfhydril groups sulfhydril groups disulfide groups disulfide groups Barka & Anderson, 1965 Thompson, 1966 Barka & Anderson, 1965 Barka & Anderson, 1965 Barka & Anderson, 1965 Barka & Anderson, 1965 Barka & Anderson, 1965 270 at pH 7.2, for another hour, washed in buffer, dehydrated in an acetone or ethanol series and embedded in either Spurr's low viscocity embedding medium or Epon 812. Thin sec- tions were cut on a Porter-Blum MT-1 ultra- microtome using glass or diamond knives and stained with Sato lead and uranyl acetate. Thin sections were examined on a Philips EM 201 transmission electron microscope at an accelerating voltage of 80 kV. Lyonsiid valves were dehydrated in an ethanol series up through several changes of absolute and placed in a 60°С oven for 5-10 days for ultrastructural examination of the periostracum and any mantle secretions over the periostracum. They were then mounted with silver paint on aluminum stubs, coated with a thin layer of carbon and gold in a Denton Vacuum 515 Evaporator, and ex- amined on a Philips PSEM 501 scanning electron microscope at accelerating voltages of 15-30 kV. In order to examine potential replacement of sediments over the periostracum, live specimens of Lyonsia californica, L. floridana, Entodesma saxicola and Mytilimeria пива! (removed from its ascidian host), from various size classes, were cleaned of all adhering particles and placed in a clean, fine sand sub- stratum on the seawater tables. These specimens were fed a mixture of Thalassio- sira pseudonana and /sochrysis galbana every other day for a period of five weeks and showed slow, but evident growth. After this time, specimens were examined for newly adhering particles. All figures are light micrographs taken on a Wild-M20 compound microscope unless otherwise noted. RESULTS General anatomy of the arenophilic gland system Of the three genera of marine Lyonsiidae (Prezant, 1980b), only Lyonsia and Ento- desma possess arenophilic radial mantle glands. The glands of these two genera are histologically and cytologically similar. They are readily distinguishable, however, based upon location with respect to the periostracal groove, and origin of a thin, at times almost membranous, surrounding epithelial sheath (Fig. 1). The gross anatomy of arenophilic glands of PREZANT all species of Lyonsia that have been ex- amined is similar to that of L. hyalina (Prezant, 1979a). The glands, which always occur joint- ly in the mantle of Lyonsia as one т the left outer fold lying opposite another in the right, are modifications of the inner epithelium of the outer fold. These regions consist of a central, club-shaped ridge of secretory cells in the longitudinal axis of the wall of a mantle т- pocketing, surrounded by a thin extension of the middle fold. The extension is confluent with the distal part of the central gland and merges with the central gland along its inner, longitudinal border (Figs. 1, 12, 13). This epi- thelial sheath of cells forms an effective coat around all but the apical, external end of the FIG. 1. Diagrammatic representation of the two types of arenophilic glands found in lyonsiids. 1a represents the gland as found in Lyonsia, and 1b, that found in Entodesma. Open arrow in 1b points out the split in periostracum at the region of areno- philic gland secretion in Entodesma. Dark, solid layer represents the periostracum and the stippled layer the gland secretion. С, arenophilic gland; i, inner mantle fold; m, middle mantle fold; o, outer mantle fold; P, periostracum secreting cells. 1a is after Prezant 1979a, fig. 2. LYONSIID MANTLE GLANDS 271 FIG. 2. Lyonsia floridana. Scanning electron micro- graph of single thread of mucoid secretion from arenophilic gland overriding concentric periostracal ridge. Horizontal field width = 850 um. central gland. Both the central gland cells and the surrounding sheath cells are secretory. The apical end of the central gland in Lyon- sía ореп$ into the periostracal groove, behind the periostracum-secreting cells, via an in- vagination of the outer fold epithelium. The secretion from the glands 15 released as thin, sometimes ramifying, threads above the periostracum (Fig. 2). A full description of the mantle inpocketing and associated neck cells through which the gland secretes is given elsewhere (Prezant, 1979a). Briefly, the inner epithelium of the outer fold invaginates to form a pocket with recessed lateral borders (Fig. 3). Numerous, thin muscle fibers, origi- nating from the inner pallial musculature, at- tach to the neck region of the individual glands and function in retracting these pro- trusible organs. The arenophilic glands of Lyonsia form small papillations along the FIG. 3. Lyonsia floridana. External, recessed pocket of arenophilic gland, frontal section. Portion of central gland at arrow. Zenker's fluid, Pantin modification. Horizontal field width = 45 um. FIG. 4. Lyonsia californica. Protruding arenophilic gland, frontal section. Zenker's fluid, Groat's hemotoxylin. Horizontal field width = 120 um. frontal section. Formalin, ethanol, Groat’s hematoxylin. Horizontal field width = 590 um. FIG. 5. Lyonsia pugetensis. Densely situated arenophilic glands along mantle edge, FIG. 6. Lyonsia californica. Arenophilic glands along posterior siphonal epithelium, cross-section. Zenker's fluid, Pantin modification. Horizontal field width = 425 ит. 272 PREZANT mantle edge when extended. Protrusion of the glands in L. gouldii or L. pugetensis gives the mantle edge a tentaculated appearance (Fig. 4). Eversion of the glands probably oc- curs via hydrostatic pressure as numerous hemocoelic spaces pervade the mantle tis- sues of lyonsiids. The arenophilic glands of Lyonsia are vari- able in size and number (Fig. 5). They range from an overall length of 75 um and diameter of 22 um in L. floridana, to 970 ¡um long and 90 um wide in L. pugetensis. Lyonsia flori- dana, a small lyonsiid rarely exceeding 15 mm in length, has only a few arenophilic glands situated along its mantle edge. Lyon- sía pugetensis, on the other hand, reaches overall lengths exceeding 50 mm and may posess well over 100 pairs of arenophilic glands densely packed along its mantle edge (Fig. 5). The glands in many species of Lyon- sia are evenly distributed along the mantle edge and are often found in alignment with fine, radial striations of the periostracum. Sand frequently adheres along these perio- stracal striae as a reflection of this one-to-one relationship of gland to external shell orna- mentation. Arenophilic glands are often con- centrated along the siphonal margin and over 40 glands may surround these small atria in adult L. californica (Fig. 6). Most museum specimens of Lyonsia that were examined (including L. arenosa, L. flabellata, L. nesiotes and L. norvegica) plus the specimens used in this study, retained at least some adherent particles, especially ventrally and along the siphonal border (Fig. 7). The arenophilic mantle glands of Ento- desma do not open into the periostracal groove and the surrounding epithelial sheath FIG. 7. Lyonsia pugetensis. Shells with peripheral sand cover. Shell length = 28 mm. is not a modification or extension of the mid- die mantle fold. Instead, the glands, which are still an elaboration of the outer mantle fold epithelium, open distal to the periostracal groove, and the surrounding sheath is an ex- tension of the outer mantle fold (Fig. 1). The glandular secretion obviousiy cannot be im- mediately emplaced above the periostracum as it is in Lyonsia. The arenophilic gland secretion in Entodesma, which also is pro- duced as a thin, cylindrical thread, pierces the periostracum intermittently (Figs. 1, 8), and emerges as small threads, webs or tufts, radi- ally aligned along the periostracum. In sec- tion, it is readily seen that the periostracum is distinctly interrupted at the point of glandular secretion. This interruption produces a pore through which the arenophilic gland secretion emerges. The pore is not preformed but is a result of contact with the gland secretion. The arenophilic glands of Entodesma are never as abundant as they are in Lyonsia. А specimen of Е. beana 15 mm long, for ex- ample, may have only five pairs of glands; three along the siphons and two along the pedal gape. There is some indication that the glands may quantitatively regress with growth or age in some members of the genus. Thus, older or larger specimens tend to have fewer arenophilic glands than smaller, juvenile specimens. A specimen of E. chilensis 8 mm long has 11 pairs of glands while a specimen 25 mm long has only 2 pairs that are situated along the pedal gape. Similar evidence for this regression, or at least for quantitative variability, has been found in a small growth series of E. saxicola (Table 2). Aside from gross differences in location and sheath composition, the arenophilic glands of Lyonsia and Entodesma are quite similar. They are club-shaped organs, cylindrical in cross-section, and surrounded by a secretory TABLE 2. Examination of mantle edge of a small growth series of Entodesma saxicola indicating a possible quantitative regression of arenophilic glands with overall size (length) of the clam. Size (length in mm) Total number of glands 29 40 36 36 36 35 44 23 50 25 58 13 63 23 LYONSIID MANTLE GLANDS 273 sheath (Figs. 9, 10). A thick secretion emerges from the glandular system, initially confined between the central gland and the surrounding sheath, and eventually emerges, through an invagination of the outer mantle fold, above the periostracum. The glands of Entodesma usually, though not always, occur in pairs along the mantle edge and, in many species, are most abun- dant along the pedal gape. This concentration of glands was especially evident in some of the larger species of Entodesma (E. saxicola). In some species, such as E. chilensis, the glands are also aligned with radial striations of the periostracum. In a specimen of E. chi- lensis 8 mm long, radial periostracal striations are situated 0.706 mm apart at the ventral shell edge. Arenophilic glands of this speci- men average 0.727 mm apart along the mantle edge. Sand adheres along the radial striations of E. fretalis as well and a specimen 16 mm long has about 14 pairs of glands. Thin retractor-like muscles are present in most species of Entodesma that were ex- amined but the protrusion of glands was never evident. Species of Entodesma are usually thicker- shelled and larger than Lyonsia. The areno- philic glands, however, do not reflect this dif- ference in overall size. They range from 92 шт long and 30 um in diameter in Е. beana to 540 um long and 32 um in diameter in Е. saxicola. The latter species is one of the largest species of Entodesma while E. beana is relatively small for this genus. The largest lyonsiid species do tend to have the largest FIG. 8. Entodesma saxicola. Arenophilic gland secretion (arrow) penetrating periostracum, longitudinal section. Hollande's fluid, Pantin modification. P, periostracum. Horizontal field width = 95 ит. FIG. 9. Entodesma saxicola. Arenophilic gland (section primarily through type 2 cells) with thick, darkly staining, surrounding secretory layer, frontal section. Zenker's fluid, Pantin modification. Horizontal field width = 65 ит. FIGURE 10. Entodesma chilensis. Arenophilic gland, longitudinal section. Space surrounding gland is fixation artifact. Gland secretion at arrow. Zenker's fluid, Azan. Horizontal field width = 80 um. 274 PREZANT mantle glands, but much variation in size of the glands exists in medium sized species. The third marine genus of lyonsiid, Mytili- meria, found embedded within compound as- cidians, lacks arenophilic radial mantle glands in specimens 6 mm long and larger. Speci- mens smaller than this have not been ex- amined. General histology The arenophilic mantle gland system, in both Lyonsia and Entodesma, is composed of three primary secretory cell types (Figs. 11- 13). The club-shaped central gland basically consists of two cell types while the third pri- mary secretory cell constitutes parts of the surrounding sheath. In cross-section the cen- tral arenophilic glands are C-shaped (Figs. 13, 14), the open end being the region of merger between the central gland and the surrounding sheath (Fig. 13). The two pri- mary, central cells often form a pseudostrati- fied epithelium when viewed in mid-longitudi- nal section (Fig. 11). Cell type 1, roughly ovoid or triangular in shape, stain basophil- ically, have small round, central nuclei with single nucleoli, and are packed with deeply staining granules. These cells either taper towards the broad surface periphery of the gland (Fig. 11) and thus appear bluntly tri- angular, or are ovoidal or goblet-shaped in section and do not make contact with the lat- ter surface. These cells sometimes show a dense accumulation of large vacuoles that fill most of the intracellular area. A dense array of fine granules is more common. Both condi- tions are especially evident in L. californica and probably represent different stages of secretion maturity. Type 1 cells are near the open end of the C-shaped central gland. At this merger zone the central gland consists of a series of bluntly ovoid cells that are similar to type 1 cells but may stain even more baso- philically. These cells usually have a central or basal ovoid nucleus with a single nucleo- lus, and a high concentration of deeply stain- ing secretory granules. These cells, also categorized as type 1 cells, form the basic juncture cell between the central gland and the sheath cells in both Lyonsia and Ento- desma. The central gland is always dominated by flask-shaped or tall triangular cells which taper proximally within the organ (Figs. 11- 13, 15). These cells are termed cell type 2. The cells are deeply basophilic, have small round or ovoid nuclei with single nucleoli and FIG. 11. Lyonsia californica. Diagrammatic repre- sentation of longitudinal section through portion of arenophilic gland. Cell type 1 and 2 are clearly shown. Spindle-shaped swelling of surrounding sheath forms outer wall and source of gland sec- retion (stippled area outside cells). Horizontal dia- gram width = 25 um. FIG. 12. Entodesma beana. Diagrammatic cross- section through arenophilic gland. Horizontal dia- gram width = 32 um. FIG. 13. Entodesma chilensis. Diagrammatic cross- section through arenophilic gland. Horizontal dia- gram width = 30 um. LYONSIID MANTLE GLANDS 275 often alternate along a longitudinal medial plane, with the shorter, less regular type 1 cells forming the pseudostratification. Type 2 cells rarely show large secretory granules within the cytoplasm in histological sections, but often have fine granules concentrated along their external border. Maximum width of type 2 cells range from 3-7 ит depending upon species and exact location in the gland. Merger of the central gland and the surrounding sheath occurs along the cross-sectional open end of the C, which is directed away from the inner epitheli- um of the outer fold. The juncture cells are usually of equal width to the widest of the other central gland cells. The sheath, derived from the middle fold in Lyonsia and the outer fold in Entodesma, 1$ confluent with the juncture cells. This thin sur- rounding epithelium is composed of flat squa- mous cells that irregularly form bulbous or spindle-shaped cells (Fig. 13) that usually contain a circular nucleus with a single nucle- olus and a dense, basophilic cytoplasm. The latter is often perfused with deeply staining secretory granules. The spindle-shaped cells of the sheath are especially evident at the re- gion of merger or confluence between the central gland and the sheath. The encircling sheath was previously reported as being a re- taining layer in L. hyalina (Prezant, 1979a). The thick secretion surrounding the central gland and medial to the sheath, however, 15 the compound result of joint secretory activi- ties from both tissues. The secretion 15 bi- layered as a result of bi-directional secretory activities. The thick, inner layer of secretion is produced by type 2 cells which dominate the central gland while the thinner outer secretory layer is produced by the sheath cells and the ovoid or bluntly triangular type 1 cells of the central gland. Cellular distinctions are readily evident in species of Lyonsia which possess numerous, large glands. Arenophilic glands of L. puge- tensis show very tall and thin type 2 cells that measure 45 ит by 4 ит, and possess a cen- tral ovoid nucleus that fills most of the width of the cell and 6 ит of its length. Type 1 and 2 cells of L. pugetensis sporadically alternate along the longitudinal-medial plane of the central gland. Type 1 cells of this species are deeply basophilic, range in height from 10-20 рт, and have small central nuclei. These cells do not taper into an extended, narrow neck that reaches the broad gland surface. Contrary to this, the two primary cell types of the central gland of L. californica alternate along the medial-longitudinal plane of the central gland and both cells extend the entire 18 um height so that the broad apical surface of type 2 cells lie next to the thin, tapered neck of type 1 cells. This, in effect, produces a pseudostratified columnar epithelium. This type of central gland is found in L. floridana, L. gouldii, and L. hyalina. While glands of members of the genus Lyonsia open just behind the periostracal groove cells, arenophilic glands of Ento- desma may open some distance distal to the typically small periostracal groove (Figs. 1, 16-17). In Entodesma chilensis the areno- philic glands open about 400-500 um away from the periostracal groove. The glands of this species show the usual type 1 cells which are roughly triangular and have thin exten- sions leading to the broad surface of the cen- tral gland, are 14 um tall and 5 um wide at their inner edge and have small, round nuclei. Type 2 cells of E. chilensis are flask-shaped with a broad peripheral surface and a central ovoid nucleus with a single nucleolus (Fig. 138): Cells of the central gland show some plas- ticity in overall sizes, as seen in Entodesma patagonica. In this species the cells of the central gland along most of its length are about 15 ¡um tall. At the internal border of the central gland, the cells reach heights of 25 ит. Variation in size of type 1 cells especially in width is common. This is likely a result of various maturation stages of secretions in these mucocyte-like cells. Ultrastructure and histochemistry The ultrastructure and histochemistry of the central gland and sheath cells confirm their active secretory nature. For the most part, each species showed similar ultrastructure (Figs. 18-19) and histochemical reactions. Results of the histochemical tests are shown in Table 3. Briefly, type 1 cells stain beta metachromatically with toluidine blue (Fig. 20) and give strong positive reactions for: alco- holic periodic acid Schiff stain (with and with- out diastase); alcian blue at pH 1.0 and 2.5 (Fig. 20), alcian blue with critical electrolyte levels ranging from 0.2 to 0.8 M magnesium chloride (Fig. 22), and give trace positive re- actions for alcian blue at a 1.0 M electrolyte level, ninhydrin Schiff, and bromphenol blue. These reactions indicate a weakly acidic mu- copolysaccharide with a small protein com- 276 PREZANT ponent. Similar results were reported for Lyonsia hyalina (Prezant, 1979a). These cells have a prominent Golgi complex. In L. cali- fornica the Golgi complex has a deeply con- cave face and up to 14 cisternae which con- tain an electron dense material (Fig. 23). The Golgi complex, variously located but often prominent basally or just next to the nucleus with the mature face directed towards the nucleus, pinches off numerous small vesicles which condense into large vacuoles just next to the Golgi complex (Fig. 24). The Golgi ap- paratus of type 1 cells in Entodesma saxicola are typically composed of fewer and straighter cisternae than in L. californica. Type 1 cells also have a dense population of rough endo- plasmic reticula, a smaller amount of smooth endoplasmic reticula, numerous small round or ovoid mitochondria, and an electron dense, grainy cytoplasm. Previously, type 2 cells were thought to be supportive cells (Prezant, 1979a); current research shows they are secretory as well. Type 2 cells are histochemically distinct from type 1 cells. Type 2 cells stain positively for alcoholic periodic acid Schiff (with and without diastase) (Fig. 21), ninhydrin Schiff, brom- phenol blue, dihydroxy-dinaphthyl-disulfide, mercury orange, and the latter two stains with thioglycerol. These reactions indicate a highly proteinaceous secretory component with di- sulfide groups and a minor carbohydrate com- ponent. The secretion from type 2 cells is probably a glycoprotein. The external border of type 2 cells bears a dense array of micro- villi. The microvilli of Lyonsia californica and L. floridana (Fig. 25) are similar to those of L. hyalina (Prezant, 1979a) in that they are short, squat and irregular. The microvilli of type 2 cells of Entodesma saxicola are dis- crete units which reach lengths of 0.6 ит (Fig. 26). They are uniformly distributed along the exposed cell surface and are of relatively equal dimensions. Electron dense secretory granules in type 2 cells of Lyonsia californica indicative of exocytosis (Fig. 27). Type 2 cells are often packed with large, osmophilic sec- retory vacuoles and granules, especially pre- valent in E. saxicola (Fig. 28). Secretory granules in type 2 cells of Lyonsia californica and L. floridana are always smaller than those in Entodesma. Type 2 cells also possess a small Golgi complex, numerous small mito- chondria located especially above the nucle- us, numerous free ribosomes and a less elec- tron dense cytoplasm than that present in cell TABLE 3. Summary of the histochemical reactions carried out on the arenophilic mantle gland system of Lyonsia californica. Similar results were found for L. floridana and Entodesma saxicola. Abbreviations: —, no reaction; tr, trace reaction; +, positive reaction; ++, very strong positive reaction; ab, alcian blue positive; pas, periodic acid Schiff positive; CEC, critical electrolyte concentration; DDD, dihydroxy- dinaphthyl-disulfide. Central gland cells Secretory layer Histochemical Sheath test Type 1 Type 2 cells Inner Outer Toluidine blue (metachromasia) beta alpha beta beta beta PAS (alc.) ++ + + tr + PAS/diastase ++ - + tr + Alcian blue, pH 1.0 + - > - + Alcian blue, pH 2.5 + — + - + Alcian blue, pH 5.7 CEC 0.1 M + - + - + CEC 0.2 М eo - - — + СЕС 0.4 М ++ - + - + CEC 0.5 M ++ = + - + CEC 0.6 M ee - - - + CEC 0.8 M + - + - + CEC 1.0 M tr — tr - tr Alcian blue, pH 2.5/PAS ab pas ab tr pas ab Ninhydrin Schiff tr ++ tr ++ + Bromphenol blue tr ++ tr ++ tr Mercury orange — ++ ~ . - Mercury orange/thioglycerol ~ ++ - + = DDD — tr ~ = = DDD/thioglycerol = tr — + — LYONSIID MANTLE GLANDS PTT. FIG. 14. Lyonsia californica. Closely set arenophilic glands, cross-section. Gland secretion forms dark layer around each central organ. Zenker's fluid, Pantin modification. Horizontal field width = 45 ит. FIG. 15. Entodesma saxicola. Type 2 cell domination of arenophilic gland, frontal section. Thick secretory layer surrounds central gland. Zenker's fluid, Pantin modification. Horizontal field width = 35 um. FIG. 16. Entodesma chilensis. Arenophilic gland (open arrow), longitudinal section. Debris adhering to periostracum is evident (closed arrow). P, periostracum. Zenker's fluid, Pantin modification. Horizontal field width = 750 ит. FIG. 17. Entodesma fretalis. Arenophilic gland opening (open arrow) ventral to periostracal groove cells, periostracal groove at arrow, cross-section of mantle edge. Formalin, ethanol, basic fuchsin and fast green. Horizontal field width = 265 um. FIG. 18. Lyonsia californica. Transmission electron micrograph of arenophilic gland, frontal section. 1, cell type 1; 2, cell type 2; $, sheath cell; arrow, gland secretion. Horizontal field width = 50 ит. FIG. 19. Lyonsia californica. Transmission electron micrograph of arenophilic gland, longitudinal section. Type 1 cells show large secretory vesicles. 1, cell type 1; 2, cell type 2; s, sheath cell; arrow, gland secretion. Horizontal field width = 25 ит. 278 PREZANT FIG. 20. Lyonsia gouldii. Arenophilic gland, frontal section showing metachromasia of cell type 1. Formalin, ethanol, toluidine blue. Horizontal field width = 85 ит. FIG. 21. Lyonsia californica. Arenophilic gland, frontal section. Medial cells type 1 react positively to alcian blue, laterally placed cells type 2 react positively to PAS; secretion barely reactive. Rossman's fluid, alcian blue pH 2.5 and PAS. Arrow, gland secretion. Horizontal field width = 80 ит. FIG. 22. Lyonsia californica. Arenophilic gland, frontal section. Cells type 1 and sheath cells give strong positive reaction to alcian blue at critical electrolyte level of 0.5 M. Rossman's fluid. Horizontal field width = 135 um. type 1. The nucleus of type 2 cells has densely staining, widely dispersed chromatin material. The commanding aspect of type 2 cells is the very dense concentration of rough endoplasmic reticulum (Fig. 29). The rough endoplasmic reticulum of the cells in L. cali- fornica usually occur as tightly packed parallel cisternae. The reticulum 1$ dispersed through- out the cytoplasm of type 2 cells but the paral- lel arrangement of cisternae is especially pre- valent next to the nucleus in L. californica. Round, vesicular-like sections of the rough endoplasmic reticulum are dispersed through- out the cytoplasm of type 2 cells of E. saxi- cola. The Golgi complex of type 2 cells in L. californica has a concave face but is com- posed of only 5 or 6 cisternae. An electron dense material fills the Golgi cisternae and small, electron dense vesicles are evidently pinched off. The Golgi complex of type 2 cells usually occurs supranuclearly. The Golgi complex of type 2 cells of E. saxicola (Fig. 30) is less prominent than in L. californica. In the former, the Golgi apparatus appears as a small system of straight lamellae that pinch off electron lucent vesicles that gradually in- crease in density. Some of the large secretory granules in the type 2 cells of E. saxicola pos- sess microlamellar inclusions that form a fingerprint or paracrystalline pattern within the electron dense structure (Fig. 31). In cell type 2 of E. saxicola, a large, multivesicular body was found just below the cell surface and above the previously described secretory granule (Fig. 30) in one series of sections. The secretory granule was typically sur- rounded by rough endoplasmic reticulum and was often capped by an electron lucent vacuole. In all observed stages, the granule or vacuole is surrounded by a somewhat less LYONSIID MANTLE GLANDS 279 FIG. 23. Lyonsia californica. Transmission electron micrograph of type 1 cell Golgi complex. Note small membrane enclosed granules pinched off from ends of sacs and large vesicles above Golgi. Horizontal field width = 2.5 ит. FIG. 24. Lyonsia californica. Transmission electron micrograph of basal regions of type 1 cells. д, Golgi complex; m, mitrochondrion; п, nucleus; у, vesicle. Horizontal field width = 7.5 um. FIG. 25. Lyonsia floridana. Transmission electron micrograph of arenophilic gland secretion. Arrow indi- cates division in secretory layers. Horizontal field width = 10.5 ит. FIG. 26. Entodesma saxicola. Transmis- sion electron micrograph of outer region of type 1 cell. Secretory layer is obviously biphasic outside microvilli. Horizontal field width = 5 ит. FIG. 27. Entodesma saxicola. Transmission electron micrograph. Exocytotic release of two secretory granules from cell type 2 of arenophilic gland. Note homogeneous electron density of secretory granules and inner secretory layer. Horizontal field width = 1.7 ит. electron dense cytoplasm that in turn 1$ sur- rounded by a single membrane (Fig. 30). Dur- ing some phases of this granule development the secretory material condenses into an electron dense substance with internal micro- lamellae This type of system is seen only rarely in type 2 cells and may be indicative of lysosomal activity. The cells of the central gland are joined by numerous desmosomes (Fig. 32). These areas of cell attachment are usually most noticeable near the gland periphery. The in- tercellular space is approximately 220 A. The sheath cells for the most part are very thin, but irregularly, and always at the juncture with the central gland, become spindle- shaped (Figs. 28, 33-35). These spindle- shaped cells show histochemical similarities to type 1 cells and, based on these properties, may secrete the same substance. The spindle-shaped sheath cells have a grainy cytoplasm filled with a high concentration of rough endoplasmic reticulum, numerous small round or ovoid mitochondria and, in Lyonsia californica, at least some regions are filled with a dense array of microfibrils that run normal to the cell surface (Fig. 36). These microfibrils appear to be composed of elec- tron dense granules which are closely aligned and abut upon the external secretory layer. The secretory surface of the sheath cells of Entodesma saxicola bears an irregular, short microvillar border (Fig. 28). In E. saxicola the sheath cells show numerous membrane bound secretory granules with concentric fibril lamellations as seen rarely in type 2 cells of this species (Figs. 37-38). The fine fibrils, that compose the lamellae within the secretory granule circumscribe the large, mature va- cuole and seem to form a bounding, multi- PREZANT У, iy N u ena RR HR i: 4$: Ра." FIG. 28. Entodesma saxicola. Transmission electron micrograph showing electron dense secretory granules of type 2 cells, note secretory granule emerging from sheath cell (arrow). Horizontal field width = 15.5 um. layered coat around the central mass of he- terogeneous, grainy material. The individual electron dense fibrils are less than 55 A wide. These sheath cell granules are less homo- geneous than similar granules of type 2 cells. Granules of this sort, emerging from the sheath cells, have been found among the microvilli of these cells. The secretion produced by the arenophilic gland system is composed of two distinct layers; an electron dense inner layer pro- duced by type 2 cells, and a less electron dense outer layer produced by type 1 cells and sheath cells (Figs. 25-27, 39). Type 1 cells secrete along the entire concave portion (in cross-section) of the central gland, and its secretion merges with the sheath cell se- cretion and is deposited above the thicker, denser, cell type 2 secretion. The inner gly- coproteinaceous layer is homogeneous in Entodesma saxicola and somewhat more grainy in Lyonsia californica and L. floridana. The outer secretory layer is always thinner and in Е. saxicola shows a dense array of microlamellae, irregularly present within the grainy matrix (Figs. 40-41). These fibrils are similar to the microfibrils present within the sheath cell secretory granules. Fibrils in the secretion appear almost membranous ог similar to thin, elongated microtubules. These usually, but not consistently, run normal to the gland surface along the outer secretion and LYONSIID MANTLE GLANDS 281 FIG. 29. Lyonsia californica. Transmission electron micrograph of type 2 cells (upper region of micrograph) and portion of type 1 cells. Note dense array of rough endoplasmic reticula near nucleus of type 2 cell and large Golgi complex of type 1 cells. д, Golgi; г, rough endoplasmic reticula. Horizontal field width = 7.7 um. FIG. 30. Entodesma saxicola. Transmission electron micrograph illustrating micro-lamellar inclusions of secretory granules (open arrow) of type 2 cells. b, multivesicular body. Horizontal field width = 3.8 ит. FIG. 31. Entodesma saxicola. Transmission electron micrograph of secretory granule with micro-lamellar inclu- sion in type 2 cell. Horizontal field width = 1.1 ит. FIG. 32. Entodesma saxicola. Transmission electron micrograph of central gland and sheath cells. Note microvillar border on either side of secretory layer. d, desmosome; $, sheath cell; arrow, glandular secretion. Horizontal field width = 9 um. exhibit a periodicity of about 130 A. The fibrils of the secretory granules in the sheath cell are more tightly packed together than those found extracellularly. The dual nature of the secretion is obvious in most specimens examined ultrastructurally. The inner secretory layer (i.e., the glyco- protein layer) exceeds 2 um in thickness in Entodesma saxicola while the outer (i.e., mucopolysaccharide) layer is usually less than 0.8 um thick. The outer layer in Lyonsia californica is less than 0.01 ¡um thick while the inner glycoprotein layer is about 0.2 ¡um thick. The outer layer of L. floridana averages less than 0.5 ¡um in thickness while the inner layer is just over 2 um thick. In the latter species the two layers are difficult to dis- tinguish since both have approximately equal electron density (Fig. 25), but a fine line se- parates the two layers. This is reminiscent of the vague separation of secretory layers in L. hyalina (Prezant, 1979a). A large nerve trunk runs along the region of merger between the central gland and the sheath in Lyonsia californica (Fig. 34). This trunk measures about 8.6 ит wide by 3.2 ит tall. A similar neural bundle, showing numer- ous axons with synaptic vesicles, runs longi- tudinally along the sheath cells of Entodesma saxicola. Details of the potential nervous innervation of the arenophilic glands were not obtained. 282 PREZANT FIG. 33. Entodesma saxicola. Transmission electron micrograph of sheath cell. s, sheath cell; arrow, glandular secretion. Horizontal field width = 16.7 ит. FIG. 34. Lyonsia californica. Transmission electron micrograph showing merger of sheath cells and central gland at ovoid border cell. Nerve trunk runs along this region of тегдег. С, central gland; $, sheath cell; t, nerve trunk. Horizontal field width = 17.4 ит. FIG. 35. Lyonsia californica. Transmission electron micrograph of sheath cell and periphery of central gland. g, Golgi; m, mitochondrion; г, rough endoplasmic reticular; s, sheath cell. Horizontal field width = 10.5 um. Sediment replacement Five weeks after the shells of selected indi- viduals had been cleared of adhering sedi- ment and the animals had been replaced in a fine, sand substratum, both Lyonsia floridana and L. californica attached new sand grains along the ventral periphery of their shells. Reattachment of sediment reflected new periostracal growth; sand adhered only to the newly grown region. Entodesma saxicola under similar conditions, emplaced only a few LYONSIID MANTLE GLANDS 283 FIG. 36. Lyonsia californica. Transmission electron micrograph of granular fibrils surrounding portions of outer secretory layer in sheath cells. Horizontal field width = 2 ит. FIG. 37. Entodesma saxicola. Transmis- sion electron micrograph of secretory granules with micro-lamellar inclusions among microvilli of sheath cells. Horizontal field width = 1.6 ит. FIG. 38. Entodesma saxicola. Transmission electron micrograph of secretory granules with micro-lamellar inclusion within sheath cell. Horizontal field width = 1.2 ит. FIG. 39. Lyonsia californica. Transmission electron micrograph of biphasic secretion. Note fibrils in surrounding sheath. Horizontal field width = 1.7 ит. FIG. 40. Entodesma saxicola. Transmission electron micrograph of biphasic secretion. Note electron dense inner secretory layer and less dense, lamellated outer layer. Hori- zontal field width = 2.5 ит. FIG. 41. Entodesma saxicola. Transmission electron micrograph showing micro-lamellations of outer secretory layer. Horizontal field width = 1.1 um. sand grains mainly along the siphonal and pedal gape regions of the shell. Specimens of Mytilimeria nuttalli from various size classes (the smallest being 6 mm long at start of the five week experiment) grew close to 3 mm by the end of this period. No specimen showed any indication of sediment adhesion during or after the experimental time. DISCUSSION Adhesion of sediments to the shell is not unique to the Lyonsiidae. Samarangia quad- rangularis, a venerid bivalve, for example, deposits discretely sculptured patterns of sand over its shell surface (Clench, 1942). Many gastropods as well, particularly among the xenophorids, are able to decorate their shells with foreign materials. Even land snails of the family Sagdidae cement debris to their shell, probably for protection (Clench, 1942). These molluscs, however, incorporate extra- neous matter into the calcareous shell (Xe- nophora) (Morton, 1958), while the other mol- luscs may secrete an adhesive mucus from “typical,” unicellular pallial mucocytes. Dis- crete, multicellular organs, producing an ex- ternal adhesive involved in cementing foreign matter to the periostracum, are known only in lyonsiids and verticordiids. The Lyonsiidae have undergone a high de- gree of adaptive radiation with regard to their habitats, as discussed earlier. п each case, the periostracum and associated components play an intricate role in the maintenance of the bivalve within its appropriate microhabitat. The presence or absence of arenophilic glands correlates well with the life styles of the three genera. Lyonsia species are attached to the substratum by only a few, weak byssal threads and are often exposed to shifting sediments and tides. All species of Lyonsia 284 PREZANT possess arenophilic mantle glands that are usually evenly distributed along the entire mantle edge and are particularly concentrated along the siphons. The glands secrete into the periostracal groove and over the perio- stracum and the sand, which adheres to this surface, adds surface area, provides a pro- tective coat, and adds weight to the shell. The attached sand may play several roles, includ- ing stabilizing the bivalve within the sub- stratum, acting as a protective cover, and pro- viding some camouflage. The latter two functions are of special import along the pos- terior shell surfaces of Lyonsia. Here, where glands may be most abundant, attached particles cover extensions of the periostracum that flare out beyond the calcareous shell edge. The siphonal region of shell cannot be fully adducted but upon withdrawal of the si- phons, the periostracal flaps fold inward and attached sand grains form an effective oper- culum over the siphonal даре. This is parti- cularly important in most species of Lyonsia. These lyonsiids are typically only partially buried within the substratum, and their si- phonal regions would otherwise be exposed to potential predation or silting were it not for the protective sandcover. Species of Entodesma are generally thicker-shelled than Lyonsia and also produce numerous, thick, pliable byssi. These fea- tures, in conjunction with nestling habits, us- ually leaves adult bivalves of this genus well secured and protected. Juvenile individuals, on the other hand, are thinner-shelled and may not fit as securely within a given crevice. Smaller Entodesma species are also gener- ally thinner shelled than larger members. Juveniles and adults of typically smaller species of Entodesma, may possess more arenophilic mantle glands than larger, thicker shelled specimens. Some evidence indicates a quantitative regression of these glands in some species of Entodesma. Very large specimens of Entodesma rarely have many foreign particles adhering to their shell. Specimens of Mytilimeria, greater than 6.0 mm in length, lack arenophilic mantle glands. The loss of these organs in this genus 1$ understandable. M. nuttalli larvae settle on colonies of Eudistoma psammion or Distaplia occidentalis and the bivalve seems to inhibit growth of the tunicate at the point of settle- ment (Yonge, 1952). Thus, the ascidian col- ony grows around the mollusc. The bivalve shell adheres within the tunicate by a “sticky periostracum” (Yonge 4 Thompson, 1976). The role of stabilization, protection and cam- ouflage have been “taken over” by the host. The bivalve has in turn evolved cryptic colora- tion of the siphons to match the tunicate test, a physically adhesive periostracum (Prezant, 1980a), and a method of maintaining an open respiratory and feeding passage through the ascidian. The fine structure of lyonsiid arenophilic glands suggests their secretory roles and products. The elaborate Golgi apparatus found in type 1 cells of arenophilic mantle glands and the very dense concentrations of rough endoplasmic reticulum in type 2 cells correspond well to previous findings reporting similar secretions. Secretory granules origi- nating from the Golgi complex often signify a high proportion of carbohydrate in inverte- brate mucoid secretions (Peterson & Leblond, 1964). The large Golgi complex of the cere- bral organ of the nemertean Lineus ruber, in conjunction with a high concentration of rough endoplasmic reticula, may indicate a high pro- tein component in the mucus secreted by this organ (Storch & Welsch, 1972). The muco- cytes of the cephalic tentacles of the proso- branch Pomatias elegans have a dense population of rough endoplasmic reticulum which may give rise directly to mucoid se- cretory granules (Storch & Welsch, 1972). The arenophilic gland system of lyonsiids produces two secretory components, a gly- coprotein and a weakly acidic mucopoly- saccharide, which form a tight unit secretion. The two components probably have separate functions. The thin, outer mucopolysac- charide layer may act as a lubricant to aid in the passage of the more viscous glycoprotein through the glandular inpocketing. Possibly, the mucopolysaccharide component also acts as the initial adhesive for primary particle at- tachment prior to “tanning” or hardening of the glycoprotein layer which may be re- sponsible for more permanent adhesion. In the case of Entodesma, at least one of the secretory components must contain some solvent that permits penetration through the periostracum. This organic solvent is most likely an enzyme, probably a protease. The secretion has not yet been tested for en- zymes. The glycoprotein layer is the more likely candidate for the possession of the potential protease in Entodesma. If this is the case, the thin mucopolysaccharide layer may act as a protective shield preventing extraneous per- fusion of the enzyme throughout the mantle LYONSIID MANTLE GLANDS 285 epithelium. The viscous nature of the glyco- protein may also help localize the secretion. A similar situation may exist in the oyster-drill, Urosalpinx стегеа. In this muricid, the accessory boring organ may contain an en- zyme that aids in the dissolution of the perio- stracum and organic matrix of bivalve shells (Nylen et al., 1969; Carriker, 1969, 1978; Carriker 8 Williams, 1978). Secretion from the accessory boring organ contains a neutral mucopolysaccharide or mucoprotein. The viscid nature of these substances allows close application of the secretion to the bore- hole and helps prevent the secretion from running out of the drill site (Carriker et al., 1978). Zottoli 8 Carriker (1974) also found an external protease released by several species of tube-dwelling polychaetes. This enzyme may be mixed with mucus produced by the worm's epidermis and emplaced on the inner walls of the tube. lt may be activated upon exposure to seawater and may prevent foul- ing of the worm tubes. The accessory boring organ of the prosobranch Polinices lewisi consists of two distinct epithelial regions; one producing a mucopolysaccharide and the other a proteinaceous substance (Bernard & Bagshaw, 1969). It is possible that the muco- polysaccharide component acts as a lubricant or seal during boring (Bernard 8 Bagshaw, 1969) while the proteinaceous component contains the active solvents. This may also, as mentioned, be the case in Entodesma, but of course, not for interspecific predation or boring. Mucocytes within several different taxa contain secretory granules that have dense arrays of microtubule-like structures (Storch 8 Welsch, 1972). The exact nature of these often paracrystalline components in most cases 15 not well known, but they are probably protein aggregates (Welsch & Storch, 1976). While crystalline or microfibrillar inclusions are well known intracellular components occurring within the nucleus, endoplasmic reticulum, mitochondria, Golgi apparatus, secretory granules, or in the cytoplasm of numerous cell types (Welsch & Storch, 1976: 33), the maintenance of fibrillar uniformity following secretion is not. The ventral gland cells involved in production of the secretion of the outer periostracal layer of the shell of Lit- torina littorea possess Golgi granules that contain a material with a distinct periodicity of 300 A (Bevelander & Nakahara, 1970). The outer periostracum of this gastropod also has a regular periodicity of about 300 A, although Beverlander and Nakahara (1970) suggest that ventral gland secretion undergoes a dis- persion and reaggregation at discharge. Periodic delineations in the outer secretory layer from the arenophilic glands of Ento- desma saxicola exhibit a regularity of about 130 A while granules of the sheath cell have an internal periodicity of less than this. It is likely that reaggregation is also taking place in the arenophilic gland system as there is no evidence of maintenance of periodicity in the secretion during release. The micro-lamellar or para-crystalline com- ponents found within the outer secretory layer produced by the arenophilic glands of Ento- desma saxicola may offer structural support of this secretion during periostracal penetra- tion. In species of Entodesma, the proteolytic portion of the secretion may be short-lived and only active prior to exposure to seawater. Thus, a mechanism probably exists whereby there is a periodic secretion, under nervous control, from arenophilic glands of Entodesma during a period of no, or minimal, periostracal growth. The secretion chemically penetrates the periostracum and produces a small ad- hesive tuft, thread, or web on top of the newly tanned outer organic shell layer. The exact location, or time involved in periostracal tan- ning in Entodesma is not known, but it is likely that the periostracum is tanned prior to pene- tration by the arenophilic gland secretion. Waite & Wilbur (1976) hypothesized that periostracal polymerization must occur shortly after secretion to avoid environmental dis- solution. Reflection of the periostracum over the shell edge would put adhesive com- ponents in direct contact with the surrounding environment. The secretion, upon exposure to seawater, probably loses its proteolytic powers but remains adhesive for a short time after. Following this, glands stop secreting and the periostracum, under appropriate con- ditions, resumes growth. This cycle is repeti- tive and periodic. Radially situated rows of fine short threads, or tufts regularly dot the periostracal striations of E. saxicola, E. fretalis and E. chilensis. Within an individual, the presence of dis- crete secretory organs producing an organic solvent that dissolves its own periostracum appears unique within the bivalves and limited to Entodesma. Among gastropods, some spiny muricids have the ability to dissolve parts of their own shell that would otherwise interfere with normal shell growth (Carriker, 1972). Among arthropods, an insect moulting 286 fluid able to dissolve inner proteinaceous cuti- cular layers during ecdysis is well known (Wigglesworth, 1972). A search for arenophilic radial mantle glands in other members of the Anomalo- desmata other than Lyonsiidae and Verti- cordiidae, has thus far proved negative. No comparable organs have been found in the species of Periploma, Thracia, Cochlodesma, Pandora and Laternula that were examined. lt is phylogenetically significant that thus far within this subclass only two families have been found to possess these specialized organs. Allen & Turner (1974) found mantle glands in some Verticordiidae that are quite similar to those of Lyonsia. These glands are similarly composed of a central gland with a surrounding epithelial sheath and empty into the periostracal groove. It is likely, based upon the possession of these organs and a wide variety of other morphological similari- ties offered by Allen & Turner (1974), that Ver- ticordiidae and Lyonsia have a close ancestry. К is unlikely that such specialized glands arose separately in two such similar bivalve families. The Verticordiidae may have evolved from a Lyonsia-like ancestor. A line- age of this sort, described by Allen & Turner (1974), involves an evolutionary descent from Lyonsia to Policordia to Lyonsiella. In this series the dorsal hood of the stomach is lost, the gastric shield is expanded posteriorly, and the stomach walls are muscularized (Allen & Turner, 1974). Evolution within the Lyonsiidae has, to some extent, followed a lineage leading to more sedentary habits (Yonge, 1952) starting with a free-living Lyonsia ancestor. The fourth pallial aperture, common to all lyonsiids, may be a symplesiomorph of a deep-burrowing ancestor (Ansell, 1967). The fused mantle lobes of lyonsiids in conjunction with the fourth aperture are found together only in Mactridae and Solenidae, both of which possess numer- ous deep-burrowing members (Pelseneer, 1906). The Lower Carboniferous bivalve, Wilkingia, is considered the first deep- burrowing member of the Anomalodesmata (Runnegar, 1974). These elongated bivalves had deep pallial sinuses and may have evolved from genera such as Cuneamya and Pholadella of the Lower Paleozoic (Runnegar, 1974). According to Runnegar (1974) the Pholadomyidae were well es- tablished by the Middle Paleozoic, and Ansell (1967) feels that most recent Anomalodes- mata have “secondarily returned to a shallow PREZANT burrowing habit.” Thus the lineage leading to many extant Anomalodesmata 15 from deeper burrowers (still retained in the Pholado- myidae) to shallow dwellers, and evolution within the Lyonsiidae is towards more seden- tary habits (Yonge, 1952). This dual progres- sion corresponds with the development of the arenophilic glands. If we visualize some ancestral Anomalodes- mata as deep burrowers with long siphons, it seems evident that the role of shell elabora- tions (i.e., spines, deep ribs, sediment coat, etc.) would not be of the same importance as in near surface dwellers. Deeper in the sub- stratum there is less shifting of sediments and the stabilizing elaborations are not “needed.” If we raise the bivalve to a microhabitat closer to the sediment-water interface, the higher energy environment would dictate the “need” for some method of control to maintain the bivalve's stability and protect thin shelled, ex- posed members from predation. This is seen many times among near surface dwelling bi- valves (i.e., Pitar, Pinna, Trachycardium, Cerastoderma, etc.) in the production of spines and ribs while deeply burrowing species are often quite smooth-shelled (i.e., Tellina, Ensis, Tagelus, etc.). Members of the Lyonsiidae lack well-developed spines or deep ribs and are fairly smooth shelled except for low, radial striations and small spinules (Prezant, 1979b). Many members of the Ver- ticordiidae also have only very small spines radially ornamenting their shells (Allen & Turner, 1974). Both the lyonsiids and verti- cordiids are either shallow burrowers or epi- faunistic. The mucoid secretions from the arenophilic glands essentially allow the shell to incorporate extraneous particles and thus added surface area. A similar condition exists in Laternula truncata. This bivalve produces small prefabricated spinules over its shell to aid in stabilization (Aller, 1974). The arenophilic radial mantle glands offer a unique tool to help decipher the course of evolution within the Lyonsiidae. Three primary distinctions exist between mantle glands of Lyonsia and Entodesma: 1) location relative to the periostracal groove, 2) origin of sheath cells, and 3) penetration of the periostracum by the glandular secretion in Entodesma (Fig. 1). It is not difficult to envisage the evolu- tionary development of the two gland types. Mucocytes are common along the mantle edge of most bivalves. Glands in Lyonsia are elaborations of the outer mantle fold sur- rounded by a glandular sheath from the LYONSIID MANTLE GLANDS 287 middle fold. Both the central gland and the sheath in Entodesma are elaborations of the outer fold. The latter is evident not only in histological preparations but also by the oc- currence of microvilli along the sheath cells. Microvilli commonly occur on the outer mantle epithelium. The differentially placed arenophilic radial mantle glands in the lyonsiids suggest a major change in the currently accepted intrafamilial lineage. The direct Lyonsia to Entodesma to Mytilimeria lineage predicted by Yonge (1952) is unlikely to have occurred since it is difficult to envisage an evolutionary transition from mantle glands of Lyonsia to those of Entodesma. For this to have taken place would have necessitated a movement of gland cells through periostracal groove cells to take up a position distal to the groove or, a regression of proximal gland cells in Lyonsia and a redevelopment of the gland distal to the periostracum secreting cells. It is more likely that arenophilic glands evolved under similar conditions in both genera from a series of mucocytes surrounding the periostracal groove. Thus, from the ancestral mantle edge, three developments can be postulated: 1) mucocytes proximal to the groove de- veloped as arenophilic glands (as found in Lyonsia), 2) mucocytes distal to the groove evolved into arenophilic glands (as found in Entodesma), or 3) the mantle edge remained unchanged, as in Муштепа. Each of these developments, of course, would involve cor- responding chemical evolution of appropriate mucoid substances and enzymes. The third development is unlikely on the basis of total morphology and habits of Mytilimeria. Mytili- тепа is probably descended through the Entodesma stock since there is a good pro- gression evident in extant species. Ento- desma are slightly heteromyarian while Mytili- тепа are heteromyarian; there 15 an increas- ing degree of sedentariness in the two genera; many shallow water species of Ento- desma found in rocky intertidal zones posess a well-developed, granular, homogeneous layer in their shell ultrastructure and Mytili- тепа retains a thin homogeneous layer (Pre- zant, 1980a); and there 15 an increasing body angle in the two genera (Yonge, 1952). Both also possess a muscularized mantle edge, lack photoreceptors in their siphons, have a high concentration of mucocytes adorning their inner mantle margin, have a small cy- lindrical foot and a small, circular pedal gape with a raised internal rim, and have increasing diameters in their siphonal lumens, ге- spectively. Lyonsia on the other hand, are isomyarian or very slightly heteromyarian; free-living; lack a homogeneous shell layer; have a poorly muscularized mantle edge; have numerous, small photoreceptor organs in a dense band along their exhalent siphon; have a low-lying elongated pedal aperture without a raised internal rim; and have narrow siphonal lumens. At least some species of Lyonsia as well as Mytilimeria possess а modified type of ring nacre (Prezant, 1980a). These generic distinctions are outlined and discussed in more detail elsewhere (Prezant, 1980a, b). Based upon these distinctions, | hypothe- size that Lyonsia and Entodesma diverged early in their evolution from a common Lyon- sia-like ancestor that originally lacked аге- nophilic glands (Fig. 42). Mytilimeria evolved from intertidal Entodesma stock, perhaps al- ready associated with (nestled among) com- pound ascidians, and secondarily lost are- nophilic mantle glands sometime after assum- ing a sessile life-style within compound as- cidians. In this microhabitat, the endo- symbionts developed a physically adhesive periostracum (Prezant, 1980a) and gradually lost, not only the arenophilic glands, but an active byssal attachment and thick shell. Ring nacre, presently known only from thin shelled genera of lyonsiids, is probably a primitive trait that has been lost in the thicker shelled members of Entodesma and most species of Lyonsia. Following the development of а thick, LYONSIA MYTILIMERIA ENTODESMA LYONSIID ANCESTOR FIG. 42. Hypothetical phylogenetic tree of lyonsiid evolution. The initial divergence signifies the split in arenophilic gland formation (a parallel event pro- ducing similar apomorphs). The secondary diver- gence (resulting in Mytilimeria and Entodesma) signifies dichotomies in retension or loss of areno- philic glands, and morphological distinctions pre- viously outlined. 288 PREZANT strong shell and byssus, and movement into stable quarters (i.e., nestling habits) where shells may be molded to conform with the out- line of their shelters, some species of Ento- desma may gradually lose their arenophilic glands. The strong shell and anchorage obvi- ate the “need” for a protective or stabilizing extraneous coat. The same reasoning 1$ ap- plied to Mytilimeria nuttalli where the host as- cidian insures protection and stabilization, roles previously assumed by an extraneous sand coat still typical of the thin-shelled, more exposed Lyonsia. The unique intergeneric adaptations of the shell (Prezant, 1980a) and the mantle edge in lyonsiid bivalves reflects their diverse accommodations to a wide array of microhabitats, as well as their phylogenetic plasticity. These same adaptations have un- doubtedly insured the survival of this intrigu- ing molluscan group. ACKNOWLEDGMENTS | gratefully acknowledge the generous donations and loans of specimens that made this work possible. Thanks to: Dr. J. A. Allen, University Marine Biological Station, Millport, Isle of Cumbrae, Scotland; Dr. F. Bernard, Pacific Biological Station, Nanaimo, British Columbia; Dr. K. Boss, Museum of Compara- tive Zoology, Harvard University, Cambridge, Massachusetts; Drs. D. D. Chivers and W. Lee, California Academy of Sciences, San Francisco, California; Dr. @. М. Davis, Academy of Natural Sciences of Philadelphia, Philadelphia, Pennsylvania; R. Dillon, Uni- versity of Pennsylvania, Philadelphia, Pennsylvania; R. Fay, Pacific Bio-Marine Labs., Venice, California; Dr. R. Fernald, Friday Harbor Laboratories, University of Washington, Friday Harbor, Washington; C. Gallardo, Instituto de Zoología, Universidad Austral de Chile, Valdivia, Chile; Dr. J. H. McLean, Los Angeles County Museum of Natural History, Los Angeles, California; Dr. J. Rosewater, National Museum of Natural History, Smithsonian Institution, Washington, D.C.; and Dr. R. Virnstein, Harbor Branch Foundation, Inc., Fort Pierce, Florida. | am also most grateful for the continuous aid and direction given by Dr. M. R. Carriker during this study and valuable discussions with Dr. G. Davis. Thanks are also extended to Dr. N. W. Riser for help in some of the histological interpretations. Drs. R. T. Abbott, M. R. Carriker, F. C. Daiber, R. E. Hillman and N. W. Riser kindly reviewed the manuscript and offered many helpful suggestions. The col- laborative support given by my family is great- ly appreciated. Thanks also to P. Savage for typing the manuscript and P. Palinski for help with some of the micrographs. LITERATURE CITED ALLEN, J. A., 1958, Observations on Cochlodesma praetenue (Pulteney) (Eulamellibranchia). Journal of the Marine Biological Association of the United Kingdom, 37: 97-112. ALLEN, J. A. & TURNER, J. F., 1974, On the func- tional morphology of the family Verticordiidae (Bivalvia) with descriptions of new species from the abyssal Atlantic. Philosophical Transactions of the Royal Society of London, ser. В, 268: 401- 536. ALLER, R. C., 1974, Prefabrication of shell orna- mentation in the bivalve Laternula. Lethaia, 7: 43-56. ANSELL, А. D., 1967, Burrowing in Lyonsia nor- vegica (Gmelin) (Bivalvia: Lyonsiidae). Proceed- ings of the Malacological Society of London, 37: 387-393. BARKA, T. & ANDERSON, P. J., 1965, Histo- chemistry: Theory, practice and bibliography. Harper and Row, New York, 660 p. BERNARD, F. R. & BAGSHAW, J. W., 1969, His- tology and fine structure of the accessory boring organ of Polinices lewisi (Gastropoda, Proso- branchiata). Journal of the Fisheries Research Board of Canada, 26: 1451-1457. BEVELANDER, G. & NAKAHARA, H., 1970, An electron microscope study of the formation and structure of the periostracum of a gastropod, Lit- torina littorea. Calcified Tissue Research, 5: 1- 12. CARRIKER, М. R., 1969, Excavation of boreholes by the gastropod Urosalpinx: an analysis by light and scanning electron microscopy. American Zoologist, 9: 917-933. CARRIKER, M. R., 1972, Observations on removal of spines by Muricid gastropods during shell growth. Veliger, 15: 69-74. CARRIKER, M. R., 1978, Ultrastructural analysis of dissolution of shell of the bivalve Mytilus edulis by the accessory boring organ of the gastropod Urosalpinx cinerea. Marine Biology, 48: 105- 134. CARRIKER, M. R., WILLIAMS, L. G. & VAN ZANDT, D., 1978, Preliminary characterization of the secretion of the accessory boring organ of the shell-penetrating muricid gastropod Uro- salpinx cinerea. Malacologia, 17: 125-142. CARRIKER, M. R. & WILLIAMS, L. G., 1978, The chemical mechanism of shell dissolution by pre- datory boring gastropods: A review and a hypo- thesis. Malacologia, 17: 143-156. CLENCH, W. J., 1942, A remarkable development of pseudosculpture on a bivalve. Nautilus, 55: 73-74. DAKIN, W. J., 1928, The eyes of Pecten, Spondy- LYONSIID MANTLE GLANDS 289 lus, Amusium and allied lamellibranchs with a short discussion of their evolution. Proceedings of the Royal Society of London, ser. B, 103: 355— 365. FRETTER, V. 8 GRAHAM, A., 1954, Observations on the opisthobranch mollusc Acteon tornatilis (L.). Journal of the Marine Biological Association of the United Kingdom, 33: 565-583. GILMOUR, Т. H. J., 1967, The defensive adapta- tions of Lima hians. Journal of the Marine Bio- logical Association of the United Kingdom, 47: 209-221. HILLMAN, В. E., 1964, The functional morphology of the fourth fold of the mantle of the northern quahog, Mercenaria mercenaria. Journal of the Elisha Mitchell Scientific Society, 80: 8-12. HILLMAN, R. E., 1968, Histochemistry of muco- substances in the mantle of the clam Mercenaria mercenaria. |. A glycosaminoglycan in the first marginal fold. Transactions of the American Microscopical Society, 88: 420425. HILLMAN, R. E., 1969, Histochemistry of muco- substances in the mantle of the clam Mercenaria mercenaria. ||. Mucosubstances in the second marginal fold. Transactions of the American Microscopical Society, 88: 420—425. HILLMAN, В. E. & SHUSTER, С. N., Jr., 1960, Observations on the mantle of the northern qua- hog Mercenaria mercenaria. Proceedings of the National Shellfisheries Association, 51: 15-22. HILLMAN, R. E. 8 SHUSTER, C. N., Jr., 1966, A comment on the origin of the fourth fold in the mantle of the quahog Mercenaria mercenaria. Chesapeake Science, 7: 112-113. HODGKIN, N. M., 1962, Limestone boring by the mytilid Lithophaga. Veliger, 4: 123-129. HUMASON, С. L., 1972, Animal Tissue Tech- niques. Freeman, San Francisco, ed. 3, 641 p. HYMAN, L., 1967, The Invertebrates, Vol. VI. Mol- lusca |. McGraw-Hill, New York, 602 р. MORTON, J. B., 1958, The adaptations and rela- tionships of the Xenophoridae (Mesogastro- poda). Proceedings of the Malacological So- ciety of London, 33: 89-100. MUSCATINE, L., 1967, Glycerol excretion by sym- biotic algae from corals and Tridacna and its control by the host. Science, 156: 516-519. NYLEN, М. U., PROVENZA, V., & CARRIKER, M. R., 1969, Fine structure of the accessory boring organ of the gastropod, Urosalpinx. American Zoologist, 9: 935-965. PEARSE, A. G. E., 1968, Histochemistry, Theo- retical and Applied. Ed. 3, Vol. 1. Williams and Wilkins, Baltimore, Maryland, 759 p. PELSENEER, P., 1906, Part 5, Mollusca. In LANKESTER, E. В. (ed.), A Treatise on Zoology. PETERSON, M. 4 LEBLOND, C. P., 1964, Syn- thesis of complex carbohydrates in the Golgi region as shown by radioautography after т- jection of labeled glucose. Journal of Cell Bio- logy, 21: 143-154. PREZANT, R. S., 1979a, The structure and func- tion of the radial mantle glands of Lyonsia hya- lina (Bivalvia: Anomalodesmata). Journal of Zoology 187: 505-516. PREZANT, R. S., 1979b, Shell spinules of the bi- valve Lyonsia hyalina. Nautilus, 93: 93-95. PREZANT, R. S., 1980a, Comparative shell ultra- structure of lyonsiid bivalves. Veliger. PREZANT, В. S., 1980b, Taxonomic re-evaluation of the bivalve family Lyonsiidae. Nautilus. RONKIN, R. R., 1952, Cytological studies on mucus formation and secretion in Busycon. Bio- logical Bulletin, 102: 252-260. RONKIN, В. В. & RONKIN, E., 1951, Cytochemical studies on mucous epithelial cells. Biological Bulletin, 101: 228. RUNNEGAR, B., 1974, Evolutionary history of the bivalve subclass Anomalodesmata. Journal of Paleontology, 48: 904-939. SCOTT, J. E., DORING, J. 8 QUINTARELLI, G., 1964, Differential staining of acid glycosamino- glycans by alcian blue in salt solutions. Bio- chemical Journal, 90: 4-5. STASEK, С. В. 8 MCWILLIAMS, W. R., 1973, The comparative morphology and evolution of the molluscan mantle edge. Veliger, 16: 1-19. STORCH, V. & WELSCH, U., 1972, The ийга- structure of epidermal mucus cells in marine in- vertebrates (Nemertini, Polychaeta, Proso- branchia, Opisthobranchia). Marine Biology, 13: 167-175. THOMPSON, S. W., 1966, Selected Histochemical and Histopathological Methods. Thomas, Springfield, Illinois, 1639 p. THOMPSON, T. E., 1960, Defensive acid-secretion in marine gastropods. Journal of the Marine Bio- logical Association of the United Kingdom, 39: 115-122. WAITE, J. H. & WILBUR, K. M., 1976, Phenol oxi- dase in the periostracum of the marine bivalve Modiolus demissus. Journal of Experimental Zoology, 195: 359-367. WELSCH, U. 8 STORCH, V., 1976, Comparative animal cytology and histology. University of Washington Press, 343 p. WIGGLESWORTH, V. B., 1972, The Principles of Insect Physiology. Ed. 7. Chapman and Hall, London, 827 p. YONGE, C. M., 1952, Structure and adaptation in Entodesma saxicola (Baird) and Mytilimeria nut- talli Conrad, with a discussion on evolution within the family Lyonsiidae (Eulamellibranchia). Uni- versity of California Publications in Zoology, 55: 439-450. YONGE, C. M. & MORTON, B. 1979, Ligament and lithodesma in the Pandoracea and the Poro- myacea with a discussion on evolutionary history in the Anomalodesmata (Mollusca: Bivalvia). Journal of Zoology, 191: 263-292. УОМСЕ, С. M. & THOMPSON, Т. E., 1976, Living marine molluscs. William Collins, Glasgow, 288 р. ZOTTOLI, В. А. & CARRIKER, М. R., 1974, Burrow morphology, tube formation, and microarchi- tecture of shell dissolution by the spionid poly- chaete Polydora websteri. Marine Biology, 27: 307-316. ote en ci dé ; CRE vi mors има Е Wir: vor uber ФР ori wre Bos Ar" Are у 5 y». д al 1 Анис" Ai A past PO a NCAA т APN AY Mt: res AS EA MN г MO № PUR: CEE TU De mnt? _ An р. DA == . il ta f ten Tee e я T i u 1 171 ЗУ 1% В. 4 + Y ve ET TT bd > a ree 07m TATTOO yh дробь > ИЕ VF " (Ve tid Aa id "ADA Pr, ete ~*! A PAL AA E С Ч A Herr lapas Ию de у ТУ Ma EN > + N Ze | В O Fr Tr rein MINA su rl PL p ARA Cre N 2% г 2 (Ds № 2 ie р й # ns à m u EN 3 У de Us, ES qn TIFF EM a Au A DEA MU С № ad A <4 sige ВИ MALACOLOGIA, 1981, 20(2): 291-305 RHODOPETALINAE, A NEW SUBFAMILY OF ACMAEIDAE FROM THE BOREAL PACIFIC: ANATOMY AND SYSTEMATICS David R. Lindberg Center for Coastal Marine Studies, University of California, Santa Cruz, California 95064, U.S.A. and Department of Invertebrate Zoology, California Academy of Sciences, San Francisco, California 94118, U.S.A. ABSTRACT Doubt concerning the familial assignment of the patellacean limpet Rhodopetala rosea (Dall, 1872) has existed since the species was described. The shell morphology and structure are patellid features, while the radular teeth configuration is distinctly acmaeid. The anatomy of A. rosea is basically acmaeid, but there are several significant differences and patellid features. The gill is located in the nuchal cavity and is rudimentary. lts structure is analogous to the individual lappets that form the secondary gill found in the mantle groove of patellid limpets and some acmaeids. The gill lacks filaments, a septum, distinct ciliated bands, and skeletal support and arises from the mantle skirt rather than the posterior wall of the nuchal cavity. Correspond- ingly, the vessels of the auricle are modified to receive blood from the haemocoelic spaces in the nuchal cavity roof. In addition, the circumpallial vessel connects with the haemocoelic spaces rather than directly with the auricle. The anterior portions of the right and left excretory organs extend above the nuchal cavity within the mantle skirt. Structures of the digestive system— looping of the alimentary tract and radular sac, radular dentition—are acmaeid features, while the position of the gonad is like that seen in patellid limpets. Modifications of the respiratory and circulatory systems may be adaptations associated with the brooding behavior of the species. The unique shell structure of this acmaeid limpet and the anatomical characters warrant a new subfamiliar category within the family Acmaeidae. Rhodopetalinae subfam. nov. is distinguished from other acmaeid subfamilies by a helcioni- form rather than conical shell, an interior central area with a silvery metallic lustre rather than a porcelaneous central area, and the presence of a rudimentary gill without typical ctenidial struc- tures. The combination of the patellid shell structure and acmaeid-like anatomy suggest that the Rhodopetalinae is an ancestral intermediate group between the acmaeid and cellanid limpets. INTRODUCTION The superfamily Patellacea Rafinesque, 1815 includes three Recent families: Acmaeidae Forbes, 1850; Lepetidae Dall, 1869, and Patellidae Rafinesque, 1815 (Knight et al., 1960). Members of these sub- families have a docoglossate radula, subcen- tral to anterior shell apex, and a horseshoe- shaped myostracum (muscle scar). Families are distinguished by anatomical criteria. Acmaeid limpets are the only patellaceans with a ctenidium, in addition, some species also have a secondary gill (branchial cordon). In the Lepetidae gills are completely lacking, respiratory exchange taking place in the lining of the mantle groove (Powell, 1973). п mem- bers of the Patellidae only a secondary gill is present. Radular teeth configurations can also be used to distinguish families, however, dentition is more useful for distinguishing genera within families. MacClintock (1967) introduced a new char- acter into patellacean systematics: shell struc- ture. MacClintock found 17 different types of crystal structure and layering in Recent and fossil patellacean shells, 10 in the Patellidae, 8 in the Acmaeidae, and 1 in the Lepetidae. Thus, the gill, radular, and shell structure characters of each of the 3 families, taken in aggregate, clearly delineate and define them, despite some shared characters such as sec- ondary gills and shell structure (Table 1). While preparing a revision ofthe Acmaeidae | took under study the familial assignment of the small boreal patellacean, Rhodopetala rosea (Dall, 1872). Doubt concerning the fa- milial assignment of this species has existed since the species was described. In the origi- nal description R. rosea was questionably as- (291) 292 LINDBERG TABLE 1. Characters of Recent Patellacean Limpets A AA A A AAA AAA A A AAA A IEEE Shell structure Radula Taxon group по.! (M-L-R-L-M)2 Gill Family Patellidae Genus Patella 6, 8 3-3-0-3-3 complete secondary gill 6, 8, 9, 10 3-3-1-3-3 complete secondary gill Genus Helcion 6, 7 3-3-0-3-3 incomplete secondary gill 6 3-3-1-3-3 incomplete secondary gill Genus Cellana 21314 3-2-0-2-3 incomplete secondary gill Genus Nacella 11 3-2-0-2-3 complete secondary gill Family Acmaeidae Genus Acmaea 15 0-3-0-3-0 ctenidium Genus Pectinodonta 15 0-3-0-3-03 ctenidium Genus Tectura 1 0-3-0-3-04 ctenidium Genus Rhodopetala 12 0-3-0-3-0 rudimentary gill Genus Notoacmea 14,25 0-3-0-3-05 ctenidium Genus Problacmaea 2 0-3-0-3-0 ctenidium Genus Collisella 116 1-3-0-3-1 ctenidium Genus Lottia 1 1-3-0-3-1 ctenidium & secondary gill Genus Scurria 3 1-3-0-3-1 ctenidium & secondary gill Genus Patelloida 2 2-3-0-3-2 ctenidium Family Lepetidae Genus Lepeta 15 2-2-0-2-2 absent Genus Cryptobranchia 15 2-2-0-2-2 absent A р о НЗ nn 2 1After MacClintock (1967). 2M = # of marginal teeth, L = # of lateral teeth, В = # of rachidian teeth. 3Multicuspid lateral teeth. 4Basal plates simple. SBasal plates complex. MacClintock’s (1967) shell structure group no. 17 restricted to Eocene patellaceans. signed to the genus Nacella Schumacher, 1817 in the family Patellidae. This familial as- signment was followed by Pilsbry (1891), who placed the species in the genus Patella, sub- genus Helcion Montfort, 1890 because of the submarginal apex. Dall (1921) changed his original familial assignment and established Rhodopetala, by monotypy and without ex- planation, as a section of the Acmaeidae. Oldroyd (1927), Keen (1937), and Burch (1946) all followed Dall and considered Я. rosea to be an acmaeid. Keen (1960) transferred Rhodopetala back to the family Patellidae, placing it as a sub- genus of Helcion. McLean (1966) also con- sidered R. rosea to be a patellid, and treated it as a subgenus of Ansates Sowerby, 1839, re- garded by Keen (1960) and McLean (1966) as the prior name for Patina Gray, 1847. MacClintock (1967), after studying the shell structure, assigned Rhodopetala as a sub- genus of the patellid genus Cellana H. Adams, 1869. All of these workers utilized characters found in the shell because whole animals were unknown. Golikov & Kussakin (1972) published on the first known whole specimens of R. rosea, indicating its ovoviviparity and the distinct acmaeid configuration of the radular teeth. They placed it in the family Tecturidae Gray, 1847 [= Acmaeidae]. Powell (1973), in a monograph of the Patellidae and Christiaens (1976), in a revision of the Acmaeidae, have also considered Я. rosea to be an acmaeid. However, there has remained the paradox of the shell structure being patellid (MacClintock, 1967) and the radula being acmaeid (Golikov & Kussakin, 1972). The absence of special- ized respiratory structures (Golikov, personal communication, 1978), a lepetid character, further obscures the familial position of these limpets. In the present paper aspects of the anat- omy of Я. rosea are described and illustrated for the first time. The findings of this study have caused me to reconsider my earlier fa- milial assignment (Lindberg, 1977) and | now consider R. rosea to belong to the family Acmaeidae. Because much of the anatomy of R. rosea differs so little from previously stud- RHODOPETALINAE SUBFAM. NOV. 293 ied acmaeids, only specific characters, signifi- cant anatomical differences, and diagnostic familial characters are presented and dis- cussed. However, several of the anatomical differences are significant enough to warrant subfamilial rank and | therefore propose a new subfamily within the Acmaeidae. MATERIALS AND METHODS Specimens of Я. rosea were collected by С. E. O'Clair from intertidal areas on Amchitka Island, Aleutian Islands, Alaska in 1971, 1972, and 1974 (Table 2). The limpets were fixed in 10% formalin and then placed in 70% isopropyl alcohol. A single specimen was de- hydrated, cleared, and embedded in paraffin. Serial, transverse sections were cut on a microtome at 10 um and stained with haemo- toxylin and eosin. The organ systems were reconstructed from the sections by mapping the dimensions and positions of structures at intervals of 50 um or less. Four additional specimens were dissected to corroborate the reconstruc- tions. The tissue sections and radula prepara- tions are deposited in the Natural History Museum of Los Angeles County. Unless otherwise stated, organs and struc- tures are illustrated as viewed in the dorsal aspect with the anterior towards the top of the page. ANATOMY Shell The shell (Fig. 1) is small (less than 10 mm long), and of medium height; the apex overhangs the anterior margin. The anterior slope is concave and the posterior and lateral slopes convex. The aperture is ovoid and the sides subparallel. Exterior sculpture consists of concentric growth lines and obsolete radial ribs. Shell color ranges from pink to red, but the apex typically is eroded to white. The in- terior margin of the shell also ranges from pink to red. The intermediate area 1$ red, but changes to white in wet preserved speci- mens. The myostracum is horseshoe-shaped and opens broadly anteriorly. A fine pallial line connects the anterior portions of the myo- stracum. The central area, in both dry and wet preserved specimens, is silvery white. Rhodopetala rosea belongs to MacClin- tock’s (1967) shell structure group no. 12 (Fig. 2). The exterior of the shell and interior margin TABLE 2. Material examind. Specimen Size no. (mm) Sex Depository Remarks 1 3.6 fe} 1 sectioned 2 4.8 2 1 3 5.5 wind: 1 dissected 4 7.6 2 1 dissected 5 8.5 3 1 dissected 6 4.7 © 2 7 4.8 2 2 8 5.4 ind. 2 9 5.8 3 2 10 723 3 2 11 8.6 о 2 dissected ind. = indeterminate; 1 = Natural History Museum of Los Angeles County 471-252; 2 = National Museum of Natural Science, Ottawa #1976-30. FIG. 1. Rhodopetala rosea (Dall, 1872) (Natural History Museum of Los Angeles County #71-252). consist of a complex prismatic layer. Two layers are present in the intermediate area. Nearest to the interior margin the shell struc- ture is foliated. This is followed by a radial crossed-lamellar layer that extends to the myostracum. Interior of the myostracum the central area is composed of a complex crossed-lamellar layer. Altogether there are 5 layers including the myostracum. External anatomy Removed from the shell and viewed in the ventral aspect (Fig. 3), the foot is small and subcircular, covering approximately 60% of 294 LINDBERG Interior OOOO OY (| ххх XY ОХ ER ХХ A) o ‘\ XX) ХАК SS >< o, 2525 <) <) Schematic < Y Position Structure RR complex crossed-lamellar X И M myostracum radial crossed-lamellar foliated complex prismatic FIG. 2. Shell structure of Rhodopetala rosea. dorsal outer lip mantle inner lips nuchal cavity mouth | lappet anal cephalic papilla =p _ tentacle mantle à groove mantle edge foot 1 тт FIG. 3. Ventral view of Rhodopetala rosea removed from shell. the aperture posteriorly. The head is oval with a single pair of cephalic tentacles arising dorsolaterally. The mouth appears as a sagittal slit and is surrounded by a broad outer lip. Laterally the outer lip is drawn out into large oral lappets. The anal papilla is visible below the right cephalic tentacle. The mantle margin is thickened around the perimeter of the aperture, especially along the anterior nuchal cavity cephalic head 7 tentacle pericardial sac foot mantle edge FIG. 4. Dorsal view of Rhodopetala rosea removed from shell. portion of the nuchal cavity. No pallial tenta- cles are visible. Viewed in the dorsal aspect (Fig. 4), the large nuchal cavity is visible at the anterior of the limpet. A small Нар of tissue, the gill, is visible through the mantle in the middle of the nuchal cavity. To the left of the gill is the peri- cardial sac. The visceral mass is surrounded laterally and posteriorly by the shell muscle. This horseshoe-shaped muscle is made up of nondiscrete muscle bundles and opens broadly anteriorly. In the center of the visceral RHODOPETALINAE SUBFAM. NOV. 295 mass is the kidney-shaped stomach. A lobate digestive gland 15 to the right of the stomach. Between the digestive gland and the shell muscle, the dorsal surface of the right excre- tory organ can be seen. The gonad lies to the left of the stomach and against the left shell muscle. The intestine, arising from under the posterior end of the stomach, curves around the left side of the stomach and extends di- agonally across the viscera, terminating at the anal papilla, which is adjacent to the right of the gill and anterior of the shell muscle. Internal anatomy Nuchal cavity—The nuchal cavity 1$ large and is 40% of the animal's total length. The anterior end of the cavity is narrowed, con- forming to the submarginal apex, and the thickened mantle margin. Between the dorsal and ventral mantle epithelia is a large, contin- uous haemocoelic space crossed by numer- ous tissue strands (Fig. 5a). The ciliated pos- terior portion of the roof of the nuchal cavity has large concentrations of cilia on the left side. А corresponding concentration of cilia occurs on the left side of the head immediate- ly below this region. The single gill lies in the midline of the nuchal cavity. The gill arises from the ventral mantle epithelium, not from the posterior wall of the nuchal cavity (Fig. 6). To the right, at the base of the gill lie the left excretory organ, rectum, and right excretory organ. The anteri- or portions of the excretory organs extend anteriorly into the mantle skirt, and the ex- cretory pores open ventrally. The anal papilla is elongated and directed ventrally so that it dorsal mantle epithelium tissue -strands “+ ventral mantle epithelium haemocoelic yv ‘space 50 pm opens into the mantle groove directly in front of the shell muscle (Fig. 3). The osphradia are situated on the nape of the neck far back in the cavity. The left osphradium is larger than the right. A hypobranchial gland is not present. Gill—The single gill (Figs. 6, 7) arises in the posterior portion of the nuchal cavity from the ventral mantle epithelium at the midline of the limpet’s body. It is flat and triangular with the left side slightly longer giving the apex a hooked appearance. The gill is small, 0.62 mm wide at the base and 0.68 mm long in a specimen 7.6 mm long. The gill lacks fila- ments, a laterally compressed axis (Septum), distinct ciliated bands, and skeletal support. It closely resembles the individual lappets that form the secondary gills found in patellid lim- pets and some acmaeids. Along the edge of the gill runs a marginal vessel that connects on the right side to the haemocoelic space of the roof of the nuchal cavity and on the left side to the auricle of the heart. A haemocoelic space in the central portion of the gill opens on both sides into the marginal vessel. The outer surface of the gill is folded and ciliated, with longer cilia on its ventral surface. Digestive system—The looping of the ali- mentary tract is simple (Fig. 8). The esopha- gus lies largely to the left and rotates counter- clockwise approximately 135 degrees based on the position of the dorsal folds. The poste- rior portion of the esophagus is slightly ex- panded. Directly behind the dilation it turns to the left and anteriorly broadens into a large stomach. The anterior end of the stomach turns to the right and downward. A constric- tion marks the begining of the intestine, which _haemocoelic space “ent on elium * ventral mantle epithelium 50 ym FIG. 5. Transverse section through the mantle skirt. (a) Rhodopetala rosea, (b) Tectura rubella. 296 LINDBERG pericardial sac auricle foot haemocoelic space efferent marginal vessel 0.5mm FIG.7. Gill of Rhodopetala rosea. proceeds anteriorly for a short distance and then turns to the left and upward. It passes over the esophagus and then turns posteriorly again alongside the posterior portion of the esophagus. Crossing again under the stom- ach in a broad loop to the left, the intestine turns anterodorsally along the left side of the stomach and finally diagonally crosses the left excretory organ gill nuchal cavity rectum №. $ к) Y cephalic tentacle A AP 0.5mm — intestine posterior esophagus stomach digestive gland ducts intestine 0.5mm FIG. 8. Alimentary tract of Rhodopetala rosea. visceral mass towards the right posterior por- tion of the nuchal cavity, terminating in the rectum and anal papilla. The intestine has only two loops, the first counterclockwise in the anterior portion of the visceral mass, and the second clockwise in the posterior portion of the visceral mass. The looping of the radula is mostly in an RHODOPETALINAE SUBFAM. NOV. 297 elastic membrane radular caecum radula —> 1 mm FIG. 9. Radular sac of Rhodopetala rosea. oblique sagittal plane between the folds of the digestive gland (Fig. 9). Behind the head the radular sac extends diagonally along the right side of the esophagus. Immediately posterior to the cross nerve between the right and left pedal nerve cords the radula makes a U- shaped turn upward and runs anteriorly for a short distance. № then makes another U- shaped turn and proceeds posteriorly. After a third upward U-shaped turn the radula ex- tends anteriorly again. In the vicinity of the anterior constriction of the stomach it turns downward proceeding almost to the dorsal surface of the foot where it forms a tight loop and proceeds anteriorly terminating in the radular caecum. The radula has approximately 40 rows of mature lateral teeth and 20 rows with imma- ture teeth. Each row bears three pairs of lat- eral teeth (Fig. 10a). The first pair of lateral teeth is closely set at the anterior edge of the 2nd 3 rd 60 um ribbon. The medial and lateral edges of the first lateral teeth are convex. The second pair of lateral teeth are posterior and slightly lat- eral to the first pair; the medial edges are con- vex and the lateral edges concave. The third pair of lateral teeth are slightly posterior and lateral to the second pair; the medial edges are convex and the lateral edges are straight. Marginal teeth are lacking. Radular rows consist of two ventral plates each with three lateral tooth plates (Fig. 10b). The first lateral plates are large and kidney- shaped. They extend beyond the anterior edges of the ventral plates. The second later- al plates are posterior to the first lateral plates and have straight posterior edges. The second lateral plates are separated from the third lateral plates by partial sutures. The lobate third lateral plates have lobes that extend to the margins of the ventral plates. The ventral plates are rectangular with an anterior proc- ess and posterior notch. The anterior process is rectangular and the medial edges of the processes continue under the first lateral plates forming a strong anterior suture. The jaw of R. rosea (Fig. 11) is thickened medially and there are two rounded lateral ex- tensions. The lateral edges of the extensions and the dorsal regions immediately adjacent to the medial area also are thickened. Circulatory system—The pericardial sac lies to the left of the visceral mass against the shell muscle and behind the nuchal cavity (Figs. 4, 6). It contains a thin-walled auricle and a muscular ventricle and aortic bulb (Fig. 12). Both auricle and ventricle are attached to the right side of the pericardial sac. Blood enters the auricle from the haemo- coelic space in the roof of the nuchal cavity anterior suture 1st lateral anterior process 2 nd lateral posterior plate notch b FIG. 10. Radular row of Rhodopetala rosea. (a) dentition, (b) lateral plate morphology. 298 LINDBERG medial area > ЧР Nos lateral extension 0.5mm FIG. 11. Jaw of Rhodopetala rosea. mantle skirt posterior aorta FIG. 12. Heart of Rhodopetala rosea. (Fig. 6) and from the left and right circumpal- lial vessels. A small opening in the right lobe of the pericardial sac connects with the right and probably left excretory organs via the renopericardial canal. Excretory organs—The left and right ex- cretory organs are at the right side of the visceral mass (Fig. 13). The left excretory organ is oblong and, except for its posterior- most portion is enclosed within the mantle skirt (Fig. 6). The left excretory pore opens ventrally and is surrounded by a thickened lip. | could not locate the opening from the reno- pericardial canal into the left excretory organ. The right excretory organ also extends for a excretory pore left excretory organ posterior wall of nuchal cavity right excretory organ SA A Al сие" A Pat hig EG a: 0.5mm FIG. 13. Excretory organs of Rhodopetala rosea. D : yo 50m a slo right excretory FIG. 14. Transverse section through the right ex- cretory organ of Rhodopetala rosea. short distance above the nuchal cavity, but most of it is inside of the visceral mass. The right excretory pore is ventral and surrounded by a thickened lip (Fig. 14). Immediately be- hind the nuchal cavity is a large left lobe into which the renopericardial canal opens; a RHODOPETALINAE SUBFAM. NOV. papilla and subanal lobe are absent. The right excretory organ narrows posteriorly and ex- tends ventrally under the digestive gland. Further posterior, the right excretory organ narrows and continues along the posterior portion of the visceral mass and up the left side along the shell muscle. Reproductive system—The single gonad lies on the left side of the visceral mass (Fig. 4), immediately extending behind the peri- cardial sac to the posterior shell muscle. To its left lies the shell muscle and on the right the stomach and digestive gland. | could not find a connection with the right excretory organ, but | do not doubt that it exists. Rhodopetala rosea broods its young in the nuchal cavity (Golikov 8 Kussakin, 1972) and appears to be gonochoric unlike the hermaph- roditic brooding acmaeids, Problacmaea sybaritica (Dall, 1871), P. moskalevi Golikov & Kussakin, 1972, and Tectura rubella (Fabricius, 1780). All 11 specimens including gravid individuals collected during the breed- ing season, which lasts from at least May to September, comprised separate sexes. There were no size differences suggestive of pro- tandric hermaphroditism (Table 2). DISTRIBUTION USSR: Kuril Islands, Onekotan Island (49°25'N, 154°45'E) to Paramushir Island 709, 300 NE 170°Е 299 (50%25'N, 155°50'E), and ALASKA: Aleutian Islands; Rat Islands, Kiska Island (51°57’М, 178°30’E) to Afognak Island (58°21’М, 152° 30'W). Records (Fig. 15): (Western Pa- cific) USSR: Kuril Islands; Onekotan Island (Golikov personal communication, 1978), Paramushir Island (Golikov 8 Kussakin, 1972). (Eastern Pacific) ALASKA: Aleutian Islands; Rat Islands, Kiska Island (U.S. Na- tional Museum of Natural History #30789), Amchitka Island (Natural History Museum of Los Angeles County 471-252, National Mu- seum of Natural Sciences, Ottawa # 1976-30); Adak Island (San Diego Museum of Natural History #11610); Shumagin Islands, Simeonof Island [Type-locality] (U.S. National Museum of Natural History #213813, 30790, 635464): Sitkalidak Island and Afognak Island (Eyerdam, 1946). ECOLOGY Little is known of the ecology of Rhodope- tala rosea. The limpets on which Golikov & Kussakih (1972) reported were collected from rocks and stones (Golikov personal commu- nication, 1978). O'Clair (1977) reported В. rosea on the holdfasts and fronds of Laminaria yezoensis Miyabe, 1902 at Amchitka Island. Specimens were also col- lected from rock dominated by the coralline algae Clathromorphum spp. and Thalassio- O We + Bering Sea \ ` : AN dé 4% „a Aleutian 'S* =>. e == =" — no? ss ung 77 xurit 15° a 170% 1000 KM FIG. 15. Distribution of Rhodopetala rosea O = collection records, & = Attu Island. 300 LINDBERG phylum clathrus (Gmelin) Postels 8 Ruprecht, 1840 (O'Clair personal communication, 1979). O'Clairs specimens were collected from exposed and semiprotected intertidal areas between +2.0 feet (+0.6 m) and —2.0 feet (—0.6 т) (datum = mean lower low water). Asiatic specimens were found from the low intertidal to a depth of 10 m (Golikov personal communication, 1978). Gut contents suggest that R. rosea feeds on both coralline algae and the cortical cells of laminarian algae. The shape of the radular teeth and their configuration suggests a diet of coralline algae, but bear little resemblance to those species that feed on laminarian algae (e.g. Collisella instabilis (Gould, 1846)). How- ever, another species, Collisella ochracea (Dall, 1872), also feeds on both coralline and laminarian algae and has a radular morphol- ogy and configuration suggestive of only a coralline diet (Lindberg, 1979). DISCUSSION The anatomy of Rhodopetala rosea is basi- cally acmaeid, with several significant differ- ences that set it apart from other members of the family. These are seen in the shell, respi- ratory system, circulatory system, and in ti > position of the excretory organs. The shell structure of R. rosea is found in 15 species of the patellid genus Cellana (Table 5 of MacClintock, 1967) and is charac- terized by a complex crossed-lamellar layer inside of the myostracum. In all other acmaeids, except Collisella scabra (Gould, 1846) and C. edmitchelli (Lipps, 1966), the layer inside the myostracum is radial crossed- lamellar. In C. scabra and C. edmitchelli the inner layer 15 modified foliate (Lindberg, 1978). The common shell structure of R. rosea and the Cellana species strongly suggests a phyletic relationship between the two. The gill of Я. rosea differs markedly in structure and orientation from those of other previously studied acmaeids. In his classic account of the pallial organs of aspidobranch gastropods, Yonge (1947: 466) described the ctenidium of the Lottiidae [= Acmaeidae] as having “the usual structure with alternating filaments identical on the two sides.” The acmaeid ctenidium arises from the left poste- rior wall of the nuchal cavity and extends to the right across the cavity. Yonge's (1947) description of the structure and orientation of the acmaeid ctenidium cor- roborates earlier studies by Willcox (1898, 1906), Fisher (1904), and Theim (1917b), and has been confirmed again in the Brazilian acmaeids by Righi (1966). The gill of Я. rosea arises from the nuchal cavity roof, not the left posterior wall of the nuchal cavity as reported in other acmaeids. The gill encloses a haemocoelic space through which the blood flows between mar- ginal vessels (Fig. 7). Ctenidial filaments and distinct bands of cilia are absent. The only ciliary concentration is a group of long cilia on the left ventral surface. Similar outpocketings of the ventral mantle epithelium in the nuchal cavity of the lepetid limpet /othia coppingeri Smith, 1882 have been reported by Moskalev (1977). The gill of R. rosea appears analogous to the individual lappets that form the secondary gill of patellid and some acmaeid limpets. Each lappet has a marginal vessel that con- nects with a central haemocoelic space (Gib- son, 1887; Davis & Fleure, 1903; Nuwayhid & Davies, 1978). Structures of the digestive system of RA. rosea are acmaeid-like. Whereas the patellid intestine typically has numerous, complex loops (Davis & Fleure, 1903; Fleure, 1904; Graham, 1932; Graham & Fretter, 1947), Я. rosea has a simply looped intestine similar to eastern Pacific acmaeids (Walker, 1968). The esophagus of A. rosea is rotated approxi- mately 135 degrees. Fleure (1904) reported that the maximum rotation of the acmaeid esophagus is 250 degrees, while in the Patel- lidae it is 270 to 330 degrees. While | cannot account for the limited rotation in Я. rosea, it is far below the 270 degree minimum given for the Patellidae. The radula loops between the lobes of the digestive gland in a pattern simi- lar to that of other eastern Pacific acmaeids (Walker, 1968), and not like that of Cellana species with which АЯ. rosea shares shell structure (Theim, 1917a). The radular denti- tion and basal plate morphology are distinctly acmaeid. Although the radular teeth bear a superficial resemblance to those of Acmaea s.s. Eschscholtz, 1833, the complexity of the basal plates more closely resembles those found in members of the genera Collisella Dall, 1871 and Notoacmea Iredale, 1915. The heart of R. rosea has both acmaeid and patellid features. In other acmaeids the efferent blood vessel from the ctenidium con- nects directly to the auricle (Willcox, 1898, 1906; Fisher, 1904; Fleure, 1904; Theim, 1917b; Righi, 1966; Kingston, 1968), and the RHODOPETALINAE SUBFAM. NOV. 301 ctenidial circumpallial vessel vessel aorta posterior senta scale unknown FIG. 16. Heart of acmaeid limpet. Redrawn from Fleure (1904). circumpallial * vessel ne y mantle skirt Se vessels aortic bulb * posterior aorta scale unknown FIG. 17. Heart of patellid limpet. Redrawn from Fleure (1904). circumpallial vessel enters the auricle im- mediately to the left of the ctenidial vessel (Fig. 16). The patellid heart (Fig. 17) is identi- cal to the acmaeid heart but patellids lack a ctenidium and there is no ctenidial vessel. In- stead, blood enters the auricle from the por- ous roof of the nuchal cavity (Fleure, 1904). As in acmaeids, blood also enters the auricle from the circumpallial vessel. Rhodopetala rosea, like other acmaeids, has a connection between the gill and auricle, but it is not a distinct vessel. Instead, blood from the gill is collected in the haemocoelic spaces that com- municate with other haemocoelic spaces of the nuchal cavity roof. Here, blood passed through the gill is mixed with blood passed only through the nuchal cavity roof before it enters the auricle. There is not a distinct con- nection between the auricle and the circum- pallial vessel. Instead, blood from the circum- pallial vessels enters the porous roof of the nuchal cavity and proceeds to the heart. In this respect vessels of the auricle differ from both patellids and acmaeids (cf. Figs. 12, 16, 17): The small size and rudimentary state of the gill of R. rosea, combined with the modifica- tions of the auricle, suggest that the nuchal cavity roof is an important site for respiration. This same surface is thought to be a respira- tory surface in the Patellidae (Fleure, 1904) and in other prosobranchs (Hyman, 1967: 205). Typically т the Acmaeidae, the inner sur- face of the mantle margin 1$ a secondary sur- face for respiration (Fisher, 1904; Kingston, 1968). Located here is the circumpallial sinus from which highly branched vessels arise and anastomose. Blood passes through this sinus before it is recollected and returned to the auricle through the circumpallial vessels. The tiny circumpallial sinus of R. rosea does not appear to be developed for respiration. Gas exchange on the mantle margin would be in- significant to gas exchange in the nuchal cav- ity roof through which all blood returning to the auricle passes. It is not clear if the roof of the nuchal cavity is a respiratory surface in other acmaeids. Circulatory patterns, as demon- strated with injection of dye in live and pre- served specimens, have not implicated the roof of the nuchal cavity for respiration in other eastern Pacific species (Fisher, 1904; Kingston, 1968). Moreover, there are struc- tural differences in the nuchal cavity roof of different species. Righi (1966: fig. 8) illustrat- ed a large haemocoelic space in the roof of the nuchal cavity of Collisella subrugosa (Orbigny, 1841). In contrast, Tectura rubella, a brooding circumarctic species, has a very weakly developed space (Fig. 5b). The excretory organs of other acmaeids and patellids lie behind the posterior wall of the nuchal cavity, the left excretory organ 1$ much smaller than the right, and both open into the nuchal cavity through anterior excre- tory pores (Lankester, 1867; Gibson, 1887; Willcox, 1898, 1906; Davis 8 Fleure, 1903; Fisher, 1904; Theim, 1917b; Righi, 1966; Walker, 1968). In R. rosea the excretory organs are not totally behind the posterior wall of the nuchal cavity. Except for the posterior- most portion, the left excretory organ 1$ within the mantle skirt above the nuchal cavity (Fig. 6). A corresponding portion of the larger right excretory organ also extends over the nuchal 302 LINDBERG cavity within the mantle skirt, but the bulk re- mains in the visceral mass. In addition, the excretory pores open ventrally into the cavity, not anteriorly as in other species. Otherwise, the location and size of the excretory organs are as in other acmaeids; the subanal lobe of the right excretory organ, reported in the Patellidae by Lankester (1867) is not present in R. rosea. The reproductive systems of patellid and acmaeid limpets are very similar. In both fami- lies the gonad originates as a flat tubular structure appressed to the dorsal surface of the foot on the left side of the visceral mass. When gravid, the acmaeid gonad lies ventral to the visceral mass (Fisher, 1904; Righi, 1966; Walker, 1968). In the patellids, the gravid gonad lies ventral to the visceral mass, but also extends up the left side along the shell muscle (Gibson, 1887; Davis & Fleure, 1903; Branch, 1974). The presence of the gonad of R. rosea on the left side of the visceral mass appears to be a patellid feature. The haemocoelic space in the nuchal cavity roof, the modifications of the circulatory sys- tem, and the position of the anterior portions of the excretory organs may be modifications of the nuchal cavity associated with brooding. The presence of young in the nuchal cavity would undoubtedly hamper water circulation and gas exchange along a typical ctenidium; thus modifications of the nuchal cavjty roof for gas exchange may be an adaptation to in- crease the respiratory surface area. Corre- spondingly, circulatory patterns would be modified to ensure that oxygenated blood from the nuchal cavity roof would flow directly to the heart. Because of the rich blood supply of the excretory organs, the placement of the anterior portions in the mantle skirt may also serve to increase the respiratory surface (Fretter, personal communication, 1979). Initially | thought that the disjunct distribu- tion of R. rosea was an artifact of incomplete collecting in the Aleutian Islands. However, thorough searches of intertidal and subtidal localities on Attu Island (Fig. 15) over a 5 week period failed to procure a single speci- men of Я. rosea. Therefore, the disjunct dis- tribution of R. rosea may, in part, be real. In summary, the anatomy of Я. rosea indi- cates that this species is an acmaeid limpet. The distinguishing characters are a gill in the nuchal cavity, rotation of the esophagus, loop- ing of the intestine and radula, radular denti- tion, and the proportions of the right excretory organ. Major modifications have occurred in the respiratory and circulatory systems, not- ably the reduction of the ctenidium to a ves- tigial state, use of the nuchal cavity roof as a respiratory surface, and corresponding changes in vessels associated with the auri- cle of the heart. These deviations from typical acmaeid anatomy, combined with the unique shell structure warrant a separate subfamilial category within the Acmaeidae, and | there- fore propose the following new taxon. SYSTEMATICS MOLLUSCA Gastropoda Cuvier, 1797 Archaeogastropoda Thiele, 1925 Patellacea Rafinesque, 1815 Family Acmaeidae Forbes, 1850 Acmaeidae Forbes, 1850: 76. Tecturidae Gray, 1847: 158. Lottiidae Habe, 1944: 171. Diagnosis Shell conical or cap-shaped, apex posi- tioned between middie and anterior of shell; myostracum horseshoe-shaped, open anteri- orly. Radula docoglossate; marginal teeth one or two pairs or absent, lateral teeth three pairs, rachidian tooth lacking. Nuchal cavity with single gill; secondary gill in mantle groove present in some genera. Triassic to Recent. Rhodopetalinae subfam. nov. Type-genus Rhodopetala Dall, 1921: 171. Diagnosis Shell helcioniform, apex positioned at ante- rior quarter of shell, submarginal. Marginal teeth lacking, lateral teeth approximately equal in size and shape. Gill rudimentary, situ- ated medially at back of nuchal cavity; sec- ondary gill absent. Shell structure of 5 layers, outer surface and interior margin complex prismatic, followed by foliate, radial crossed- lamallar, myostracum, and complex crossed- lamellar layers. Recent. RHODOPETALINAE SUBFAM. NOV. 303 Remarks The subfamily Rhodopetalinae is distin- guished from other acmaeid subfamilies by a helcioniform rather than conical shell, an in- terior central area with a silvery metallic lustre rather than a porcelaneous central area found in other subfamilies, and the presence of a rudimentary gill that lacks typical ctenidial structures. Only the monotypic genus Rhodopetala 1$ assignable to this subfamily. Fossil members of Rhodopetalinae are not known, but could be recognized by a helcioniform shell belong- ing to MacClintock's (1967) shell structure group no. 12. All limpets with this shell struc- ture have been thought to belong to the Patel- lidae, but | believe that the group no. 12 shell structure and the acmaeid-like anatomy sug- gest that the Rhodopetalinae is an ancestral intermediate group between acmaeid and cellanid limpets. ACKNOWLEDGEMENTS | acknowledge Dr. J. H. McLean (Natural History Museum of Los Angeles County) for providing specimens of Rhodopetala rosea for study. | am grateful to Drs. McLean, V. Fretter (University of Reading, England), and J. S. and V. B. Pearse (University of Cali- fornia, Santa Cruz) for providing valuable criti- cism and comments on the manuscript. Dr. C. E. O'Clair (Auke Bay Fisheries Laboratory, Alaska) provided additional specimens and habitat notes from Amchitka Island, Alaska, and Dr. A. N. Golikov (Academy of Sciences, Leningrad) provided habitat and locality data for the Asiatic specimens of R. rosea. Speci- mens of Tectura rubella, for comparison, were kindly provided by Dr. J.-A. Sneli (Uni- versitetet i Trondheim, Norway). This study was made possible by the histological talents of Dr. D. E. Dunn (California Academy of Sciences), Ms J.-C. Ramsaran (Natural His- tory Museum of Los Angeles County), and especially Dr. A. T. Newberry (University of California, Santa Cruz). The anatomical draw- ings were prepared by Ms J. A. Christmann (University of California, Santa Cruz). LITERATURE CITED ADAMS, H., 1869, Descriptions of a new genus and fourteen new species of marine shells. Proceed- ings of the Zoological Society of London, 1869: 272-275, pl. 19. BRANCH, С. M., 1974, The ecology of Patella Linnaeus from the Cape Peninsula, South Africa. 2. Reproductive cycles. Transactions of the Royal Society of South Africa, 41: 111-160. BURCH, J. Q., 1946, Family Acmaeidae. Minutes of the Conchological Club of Southern Cali- fornia, 57: 5-16. CHRISTIAENS, J., 1976, De familie Acmaeidae. Gloria Maris, 15(3): 49-64. CUVIER, G., 1797, Tableau élémentaire de l'his- toire naturelle des animaux. Paris, 710 p. DALL, W. H., 1869, Materials for a monograph of the family Lepetidae. American Journal of Conchology, 5: 140-150, pl. 15. DALL, W. H., 1871, On the limpets; with special reference to the species of the west coast of America, and to a more natural classification of the group. American Journal of Conchology, 6: 227-282. DALL, W. H., 1872, Preliminary descriptions of new species of mollusks from the northwest coast of America. Proceedings of the California Acad- emy of Sciences, 4: 270-271, pl. 1. DALL, W. H., 1921, Summary of the marine shell- bearing mollusks of the northwest coast of Amer- ica, from San Diego, California, to the Polar Sea, mostly contained in the collection of the United States National Museum, with illustrations of hitherto unfigured types. United States National Museum Bulletin, 112: 1-217. DAVIS, J. В. A. & FLEURE, H. J., 1903, Patella (The Common Limpet). Liverpool Marine Bio- logical Committee Memoir, 10: 1-76. ESCHSCHOLTZ, F., 1833, т: ВАТНКЕ, Н., Zoologischer Atlas, enthaltend Abbildungen und Beschreibungen neuer Thierarten, während des Flottcapitains von Kotzebue zweiter Reise um die welt...in... 1823-26 beobachtet von...F. Eschscholtz. Part 5, Berlin, 28 р. EYERDAM, W. J., 1946, In: BURCH, J. Q., Minutes of the Conchological Club of Southern Cali- fornia, 57: 15. FISHER, W. K., 1904, The anatomy of Lottia gigantea Gray. Zoologische Jahrbucher, 20: 1- 66. FLEURE, H. J., 1904, On the evolution of topo- graphical relations among the Docoglossa. Transactions of the Linnean Society of London, 9: 269-290. FORBES, E., 1850, On the genera of British Patellacea. Reports of the British Association for the Advancement of Science, 1849, Part 2: 75- 76. GIBSON, R. J. H., 1887, Anatomy and physiology of Patella vulgata. Part |. Anatomy. Transactions of the Royal Society of Edinburgh, 32: 601-638. GOLIKOV, A. N. & KUSSAKIN, O., 1972, Sur la biologie de la reproduction de patelles de la famille Tecturidae (Gastropoda: Docoglossa) et sur la position systématique de ses subdivisions. Malacologia, 11: 287-294. 304 LINDBERG GRAHAM, A., 1932, On the structure and function of the alimentary canal of the limpet. Transac- tions of the Royal Society of Edinburgh, 57: 287- 308. GRAHAM, A. & FRETTER, V., 1947, The life his- tory of Patina pellucida (L.). Journal of the Marine Biological Association of the United Kingdom, 26: 590-601. GRAY, J. Е., 1847, A list of the genera of Recent Mollusca, their synonyma and types. Proceed- ings of the Zoological Society of London, 1847: 129-219. HABE, T., 1944, On the Japanese Lottiidae (= Acmaeidae). Japanese Journal of Malacology (Venus), 13: 171-187. HYMAN, L. H., 1967, The Invertebrates, Volume VI, Mollusca 1. McGraw-Hill, New York, 792 р. IREDALE, T., 1915, A commentary on Suter's Man- ual of the New Zealand Mollusca. Transactions and Proceedings of the New Zealand Institute, 1914, 47: 417-497. KEEN, A. M., 1937, An Abridged Check List and Bibliography of West North American Marine Mollusca. Stanford University Press, Stanford, California, 87 p. KEEN, A. M., 1960, Family Patellidae Rafinesque, 1815. In: MOORE, В. C., (ed.), Treatise on In- vertebrate Paleontology, Part I, Mollusca 1, Geological Society of America 4 University of Kansas Press, Lawrence, Kansas, p. 234-235. KINGSTON, R. S., 1968, Anatomical and oxygen electrode studies of respiratory surfaces and respiration in Acmaea. Veliger, 11 (Supplement): 73-78. KNIGHT, J. B., COX, L. R., KEEN, A. M., BAT- ТЕМ, В. L., YOCHELSON, E. L., 4 ROBERT- SON, R., 1960, Archaeogastropoda. In: MOORE, R. C., (ed.), Treatise on Invertebrate Paleontology, Part |, Mollusca 1, Geological So- ciety of America & University of Kansas Press, Lawrence, Kansas, p. 169-310. LANKESTER, E. R., 1867, On some undescribed points in the anatomy of the limpet (Patella vulgata). Annals and Magazine of Natural His- tory, 20: 334-337. LINDBERG, D. R., 1977, The enigmatic Rhodo- petala. Western Society of Malacologists Annual Report, 10: 7. LINDBERG, D. R., 1978, On the taxonomic affini- ties of Collisella edmitchelli (Lipps) (Gastropoda: Acmaeidae), a late Pleistocene limpet from San Nicolas Island, California. Bulletin of the South- ет California Academy of Sciences, 77: 65-70. LINDBERG, D. R., 1979, Variations in the limpet Collisella ochracea and the northeastern Pacific distribution of Notoacmea testudinalis (Acmaeidae). The Nautilus, 93: 50-56. MACCLINTOCK, C., 1967, Shell structure of patelloid and bellerophontoid gastropods (Mol- lusca). Peabody Museum of Natural History Yale University Bulletin, 22: 1-140. MCLEAN, J. H., 1966, West American Proso- branch Gastropoda: Superfamilies Patellacea, Pleurotomariacea, and Fissurellacea. Ph.D. Dissertation, Stanford University, Stanford, Cali- fornia, 255 p. MONTFORT, P.D. DE, 1808-1810, Conchyliologie systématique, et classification méthodique des coquilles ..., Paris, 2 vols. MOSKALEV, L. I., 1977, To the revision Lepetidae (Gastropoda, Prosobranchia) of world ocean. Trudy Instituta Okeanologii, Akademiya Nauk SSSA, 108: 52-78 [in Russian]. NUWAYHID, M. A. & DAVIES, P. S., 1978, Gill structure in the common limpet Patella vulgata. Journal of the Marine Biological Association of the United Kingdom, 58: 817-823. O'CLAIR, С. E., 1977, Marine Invertebrates in Rocky Intertidal Communities. Chapter 18 /n: The Environment of Amchitka Island, Alaska. MERRILL, M. L. & FULLER, R. G., (eds.). Tech- nical Information Center, U.S. Energy Research and Development Administration, Springfield, Virginia, 682 p. OLDROYD, I. S., 1927, The Marine Shells of the West Coast of North America, Volume 2, Part 3. Stanford University Press, Stanford, California, 339 p. PILSBRY, H. A., 1891, Acmaeidae, Lepetidae, Patellidae, Titiscaniidae. Manual of Concholegy: Structure and Systematics, Volume 13, Phila- delphia, Pennsylvania, 195 p. POWELL, A. W. B., 1973, The patellid limpets of the World (Patellidae). Indo-Pacific Mollusca, 3(15): 75-206. RAFINESQUE, C. S., 1815, Analyse de /a Nature, ou Tableau de l'univers et des corps organisés. Palermo, 224 p. RIGHI, G., 1966, On the Brazilian species in the Acmaea subrugosa complex (Gastropoda: Prosobranchia: Patellacea). Malacologia, 4: 269-295. SCHUMACHER, С. F., 1817, Essai d'un Nouveau Systeme des Habitations des vers Testacés avec ХХИ planches. Copenhagen, 287 р. SOWERBY, С. В. (Il), 1839, A Conchological Manual. London, 134 p. THEIM, H., 1917a, Beitráge zur Anatomie und Phylogenie der Docoglossen; |, zur Anatomie von Helcioniscus ardosiaeus Hombron et Jacquinot unter Bezugnahme auf die Bearbeit- ung von Erich Schuster in den Zoolog. Jahrb., Supplement XIII, Bd., IV, 1913. Jenaische Zeitschrift Naturwissenschaft, 54: 333—404. THEIM, H., 1917b, Beiträge zur Anatomie und Phylogenie der Docoglossen; Il, Die Anatomie und Phylogenie der Monobranchen (Akmäiden und Scurriiden nach der Sammlung Plates). Jenaische Zeitschrift Naturwissenschaft, 54: 405-630. THIELE, J., 1925-1926, Mollusca = Weichtiere: Handbuch der Zoologie, begründet И. Kükenthal, Volume 5, Berlin, Leipzig, 275 p. WALKER, C. G., 1968, Studies on the jaw, diges- RHODOPETALINAE SUBFAM. NOV. 305 tive system, and coelomic derivatives in repre- rial—external anatomy. American Naturalist, 40: sentatives of the genus Acmaea. Veliger, 11 171-187. (Supplement): 88-97. YONGE, С. M., 1947, The pallial organs in the WILLCOX, М. A., 1898, Zur Anatomie von Acmaea fragilis Chemnitz. Jenaische Zeitschrift Natur- wissenschaft, 32: 411-456. WILLCOX, M. A., 1906, Anatomy of Acmaea testudinalis Müller, Part 1. Introductory mate- aspidobranch Gastropoda and their evolution throughout the Mollusca. Philosophical Transac- tions of the Royal Society of London, Series B. Biological Science, 232(591): 443-518. В rated MIA. PA? ON, Sa ‹ Mur i DO VW teow a ED > DORE. 21h 6 a ма zu ias Y Y pe ine ias de Ten LS wos aro Piña А (ra INN à > ID e Pi ар A Рог o aa ram À MR, Y MO Де u ur. | Ayan, Var AS Pen т’ а AS >, = У Le wed DO № иг E ER ANNA ts $ | we 7 Tes NE at = u wale Sy 831 + А À “a, _ A te an Lee #9 y ws; 2 a Os u à à era Piti = WI "WWW 14 7 o о Write An ALE ren и >. à tros - FA a o 1 en; > à 4 Alas у 5 a y TE ei р | Sid MALACOLOGIA, 1981, 20(2): 307-347 А RELICT NEOGENE CAENOGASTROPOD FAUNA FROM NORTHERN SOUTH AMERICA Edward J. Petuch! Rosenstiel School of Marine and Atmospheric Science, 4600 Rickenbacker Causeway, Miami, Florida 33149, U.S.A. ABSTRACT A previously unknown relict Neogene gastropod fauna has been found to exist in shallow water near upwelling systems along northern Colombia and Venezuela. Forty-five relict caenogastropods of the families Turritellidae, Calyptraeidae, Naticidae, Cypraeidae, Cassidae, Ficidae, Muricidae, Columbellidae, Buccinidae, Fasciolariidae, Olividae, Mitridae, Volutomitri- dae, Volutidae, Marginellidae, Cancellariidae, Conidae, Terebridae, and Turridae, are rede- scribed and their Recent and fossil distributions outlined. Also reported from the upwelling areas are the paciphilic genera Agladrillia, Aphera, and Truncaria, and the relict Neogene genera Conomitra, Panamurex s.s., and Paraborsonia. Based on the unusual ecological conditions in which the relicts have been found to be living and on their paleozoogeography, the hydrographic and substrate parameters of the Neogene southern Caribbean are conjectured. A possible ecological barrier to tropical molluscivorous predation is proposed as an explanation for the existence of several thin-shelled, Shallow water relict gastropods. INTRODUCTION To both the systematist and the zoogeo- grapher, the Recent shallow water inner shelf (0-100 m) molluscan fauna of northern South America is an enigma. Although still poorly known and never completely surveyed, this area contains a remarkable fauna with an un- usually high degree of endemism. Not only are many of the mollusks of the northern South American coastline restricted to that area, they also show few affinities with mollusks from the surrounding Caribbean region. The existence of this anomalous faunal pocket in the center of the Caribbean Molluscan Prov- ince has generated many unanswered ques- tions about the origin of the tropical western Atlantic molluscan fauna and its relationship to other faunas around the world. | first became interested in the northern South American mollusks during a series of collecting trips starting in 1974 and ending in 1979. During this time, | had several oppor- tunities to work on Colombian and Venezu- elan commercial shrimp boats and was able to sample a large number of previously unex- plored offshore areas. Since the shrimp fleets are restricted to their own national territorial waters, the sample areas were limited to the Colombian and Venezuelan continental shelves. Within this sampling area, however, a particularly interesting molluscan assem- blage was encountered. As the shrimpers moved up the coast of Colombia toward the Guajira Peninsula, large numbers of unusual mollusks were taken with increasing regular- ity. The numbers of species seen in the net hauls reached maximum level in the area of the Golfo de Venezuela, Peninsula de Paraguana, and Golfo de Triste, Venezuela. A rich molluscan fauna was encountered along this coast, containing a large proportion of previously unknown living species. Upon re- sumption of my research work at the Univer- sity of Miami in 1976, | had the opportunity to work with the R/V Pillsbury expedition collec- tions taken during cruises along the Colom- bian and Venezuelan coasts in 1966 and 1968. These extensive collections supported my findings and added several more anomal- ous mollusks to a rapidly growing species list. The unexpectedly large number of species of mollusks of all classes from the area around the Golfo de Venezuela made it nec- essary to restrict the scope of my studies. The paper presented here deals only with the higher prosobranch gastropod families con- taining members with average adult shell lengths of over 15 mm. Even with this restric- tion, the caenogastropod families covered in, 1Present address: Department of Zoology, University of Maryland, College Park, Maryland 20742, U.S.A. (307) 308 РЕТУСН this study are still among the best indicators of the distribution of shallow water benthic as- semblages. As highly specialized animals, they are often tightly restricted in their habitat preferences and, for this reason, are a power- ful tool for determining zoogeographic pat- terns. These families are also well represent- ed in the fossil record of the Caribbean region, and are excellent organisms for tracing the evolutionary trends in the western Atlantic. When | began to analyze the material from my field work and that of the R/V Pillsbury collections, it became apparent that a con- siderable part of the Colombian and Venezu- elan caenogastropods could not be assigned to any known Recent fauna. A review of the literature of the Caribbean fossil mollusks showed that these living gastropods were, in fact, previously unknown Miocene and Plio- cene relicts. From the diversity of this relict fauna, it became evident that Plio-Pleistocene glacial sea level fluctuations and temperature changes failed to destroy completely all Neo- gene shallow water molluscan assemblages in the Caribbean region. As pointed out by myself (1976: 323-325) and Vermeij (1978: 231-237), the shallow water area along northern South America has acted as a refug- ium. The full extent and size of the fauna of this refugium, however, was previously un- known and greatly underestimated. The fol- lowing description of an extant Neogene mol- luscan assemblage, together with a review of the literature on the Recent and fossil caeno- gastropods of northern South America, sheds more light on the spatial and temporal hetero- geneity of the Recent tropical western Atlantic molluscan fauna. MATERIALS AND METHODS Collections studied The specimens of the relict caenogastro- pods that form the framework of this study are part of four collections. The first of these com- prised several hundred lots taken by myself while shore collecting and working on shrimp boats along the Colombian and Venezuelan coasts from 1974 to 1979. This collection was divided and deposited at the National Muse- um of Natural History, Smithsonian Institution, and in the Invertebrate Museum collection of the Rosenstiel School of Marine and Atmos- pheric Science, University of Miami. The second was that of the R/V Elliott Pills- bury, taken as part of two oceanographic sur- veys of the southern Caribbean in 1966 and 1968 and is housed in the Invertebrate Muse- um of the Rosenstiel School. lt represents several thousand lots of wet and dry, sorted and unsorted material. The third collection used was that of the Universidad Simón Bolívar Marine Laboratory at Puerto Cabello, Venezuela, and was ex- amined by me in March, 1979. This largely un- sorted collection contains several hundred lots of caenogastropods, with most of the specimens coming from the shrimp trawlers that fish the adjacent Golfo de Triste. Dr. P. Penchaszadeh, Director of the laboratory, kindly donated a large number of specimens and these are now housed at the Smithsonian Institution along with my original study collec- tion. Considering the small size of the collec- tion, it contained a disproportionately large number of relict species. These specimens indicate the large size of the poorly-known Golfo de Triste relict pocket. In March, 1979, | studied the extensive col- lection of fossil and Recent material of Dr. and Mrs. J. Gibson-Smith, of Caracas, Venezuela. Within this collection were housed several thousand lots of caenogastropods, including virtually complete collections of the gastropods of the Cabo Blanco, Mare, and Cantaure for- mations. The Gibson-Smiths kindly donated a number of important specimens which now reside in the Smithsonian Institution mollusk collection. Access to both this collection and that of the Universidad Simón Bolívar Marine Laboratory enabled me to compare Recent Golfo de Triste material with well-preserved fossils. Many of my initial species determina- tions came from these comparisons. Collecting was limited to a depth range of 0-100 m. These depths incorporate, in an oceanographic sense, a shallow water range and encompass the range of my shore col- lecting, shrimp boat work, and the pertinent R/V Pillsbury stations. The R/V Pillsbury col- lections were taken with 10 and 40 foot otter trawls following one-half to one hour hauls. The stations are given in Appendix 1. The commercial shrimp boats working along northern South America use two 10 foot otter trawls much like those of the R/V Pillsbury, but usually haul them for three to six hours, running at about 1-2 knots. The collected material is then sorted on deck. Shore collecting involved either beach col- lecting, as was done along the Colombian coast near Cartagena and Riohacha, or by RELICT CAENOGASTROPODS 309 wading in shallow bays, such as Bahía Amuay on the Peninsula de Paraguaná, Venezuela (Petuch, 1976). This limited means of collect- ing was a result of the extremely turbid water conditions along this coast which, in turn, pre- vents skin or SCUBA diving. Methods of analysis For positive identification, the relicts were compared to fossil specimens either in the literature or in the Gibson-Smith collection. Particularly useful were the works of Jung (1965, 1969), Woodring (1959, 1964, 1970), and Weisbord (1962), all of which clearly illus- trate and describe the fossil holotypes of many of the extant species. Pflug's 1969 study was the most valuable comparison work because of its illustrations of the lectotypes of Sowerby's Santo Domingo fossils. The relicts collected along the Colombian and Venezu- elan coasts are illustrated here to permit com- parison with the specimens, fossil and living, illustrated in other works. Abbreviations used in this study include: UMML—Invertebrate Museum Collection, Rosenstiel School of Marine and Atmospheric Science, University of Miami. USNM—Collection of the Division of Mol- lusks, National Museum of Natural History, Smithsonian Institution. P-000—R/V John Elliott Pillsbury station number. LITERATURE ON THE SYSTEMATICS OF THE NORTHERN SOUTH AMERICAN CAENOGASTROPODS Very little work has been done on the sys- tematics of the Recent shallow water caeno- gastropod fauna of the Colombian and Vene- zuelan coasts. Dautzenberg (1900) was the first to publish on the molluscan fauna of this area, using material from the cruise on the yacht “Chazalie.” Clench, in the Johnsonia series of monographs starting in 1941, pub- lished scattered records and single species descriptions for many caenogastropod fami- lies; particulary noteworthy were the works on the Conidae (1942,.1953), Muricidae (1945) (with Pérez Farfante) and (1959), Thaididae (1947), and Cassidae (1944). Most of his materi- al came from private collectors who had visited the area over the early decades of this cen- tury. Much of this material is deposited in the Museum of Comparative Zoology, Harvard University, and in the National Museum of Natural History and was Clench's primary source of southern Caribbean records during the war years of 1941-1945. No compendium or species checklist was undertaken for the Colombian and Venezu- elan coastlines until 1962 when Rehder pub- lished the first survey of the area in his study of the mollusks of Los Roques and La Orchila Islands off the coast of Venezuela. In large monographic revisions of muricid genera, Виз (1964), E. Vokes (1967a,b, 1968, 1974, 1975) and Gertman (1969) included new species from Colombian and Venezuelan waters. The Cassidae of northeast Venezuela was outlined by Flores (1966) while Work (1969) continued Rehder's pioneer study by producing the first extensive checklist of the Los Roques molluscan fauna. This not only included a species list, as did Rehder's, but also covered aspects of the ecology and zoo- geography of many of the species. The German biologists, Kaufmann & Gôtting (1970), working from the Instituto Colombo- Alemán at Santa Marta, Colombia, compiled the first species list of shallow water gastro- pods from the Colombian coast. This list 1$ invaluable because of its excellent illustra- tions. The 1970's saw an end to the incognita status of coastal northern South American waters. Bayer (1971) redescribed and illus- trated several shallow water Colombian and Venezuelan muricids (pp. 151-169). The families Thaididae and Muricidae of the shal- low waters of Venezuela were outlined in de- tail by Gonzalez 8 Flores (1972). In 1973, Flores (1973a, b) published two papers on the ecology and systematics of the family Littor- inidae in Venezuela that clarified for the first time the morphologically conservative Lit- torina ziczac species complex. Also in 1973, J. Gibson-Smith outlined the living and fossil Voluta from the Venezuelan coast and de- scribed two new fossil species. The mol- luscan faunas of Isla Margarita, Isla Cubagua, and Isla Coche, Venezuela were surveyed by Princz (1973). His was the first work to corre- late species assemblages with shallow water biotopes. J. Gibson-Smith & W. Gibson-Smith (1974) reviewed the Recent and fossil Venezuelan columbellids of the genus Strombina and described the second known living Atlantic species. Although outside the 310 PETUCH scope of Venezuelan and Colombian waters, the species list published by Altena (1975) on the gastropods of Surinam is of particular interest to Caribbean workers. It not only il- lustrates the first known living Caribbean Fusiturricula but also describes and illustrates many gastropods from neighboring Venezuela and Colombia. The first compendium of Venezuelan mollusks, both fossil and living, was published by Tello in 1975. The caeno- gastropod section alone covers 132 pages and contains detailed species records and literature citations. The genus Voluta т Colombia was outlined in detail by von Cosel (1976), who also worked from the Instituto Colombo-Alemán, Santa Marta. He deline- ated a possible zoogeographic barrier near the mouth of the Rio Magdalena. Similarly, Vink (1977) described the Conidae of the Conus cedonulli species complex of the Southern Caribbean and described a new species endemic to the Santa Manta area. Based on the distributions of the various cone species, several zoogeographic barriers were also suggested. In 1979, | reviewed the relict cypraeid genus Siphocypraea and described a new Colombian-Venezuelan species. Fol- lowing the examples of my predecessors, | incorporated into this study aspects of the ecology and zoogeography of the living sipho- cypraeas. LITERATURE ON THE PALEONTOLOGY OF THE NORTHERN SOUTH AMERICAN CAENOGASTROPODS In 1850, G. B. Sowerby | undertook the first review of the fossil gastropods of the Carib- bean area. This landmark work described new species collected by Colonel Heneken in Santo Domingo. The tremendous diversity of the preglacial Caribbean region was further documented by Gabb (1860 and 1881) who described many new species from Costa Rica. In 1873 and 1875, he added more new taxa to Sowerby's original Santo Domingo species list. By far the most prolific of his con- temporaries, R. J. L. Guppy published over forty works on Caribbean fossil mollusks be- tween the years 1863 and 1911. Most note- worthy are his studies of the Jamaican fossils (1866 and 1873) and those of Trinidad (1909 and 1911). Maury (1912, 1917, and 1925) also described large numbers of new fossil gastropods from Santo Domingo and Trinidad. These initial studies laid the groundwork for a rich literature, making the Caribbean Terti- ary fossil gastropods one of the best-studied molluscan faunas in the New World. Follow- ing the works of Gabb, Olsson (1922 and 1942) further expanded on the fossils of Costa Rica and described many new Pliocene species. The apogee of early twentieth cen- tury paleontological studies was reached in Woodring's 1928 work on the Pliocene mol- lusks of the Bowden formation of Jamaica. Woodring’s techniques and field experience culminated in his incomparable group of works on the Gatun formation of Panamá (1957, 1959, 1964, and 1970). Containing new species descriptions, faunal correlations, and paleoecological data, these five volumes together are the single most important con- tribution to the evolutionary history of the Caribbean mollusks. Although containing no new species descriptions, Pflug's 1961 study on the Santo Domingo caenogastropods 1$ extremely important because of the illustra- tions of many unfigured lectotypes of Sower- by's species. Jung (1969), on the other hand, added to the faunal diversity of the Pliocene of Trinidad by describing many new caeno- gastropods. His works on the fossils of the poorly-known Peninsula de Paraguana, Venezuela (1965), and Carriacou (1971) have facilitated the correlation of fossil as- semblages from other areas of the Caribbean. The unusual Pliocene gastropod fauna of Ecuador was described in detail by Marks (1951) and Olsson (1964), as was that of the Isthmus of Tehuantepec, Mexico by Perrilliat- Montoya (1963). These works were particu- lary useful in the formation of my concept of provinciality in the preglacial Caribbean and in establishing provincial boundaries. Weis- bord (1962) produced another large and im- portant work on the upper Pliocene-Pleisto- cene gastropod faunas of northern Vene- zuela. This study described some of the youngest formations in the southern Carib- bean and made it possible to correlate those faunas with the older assemblages outlined by Woodring and Olsson. Besides the research on entire molluscan faunas, there are a number of important pale- ontological studies on specific southern Caribbean gastropod genera and families. Of particular importance to my work were the publications of Schilder (1939) and Ingram (1939, 1947a, b) on the Cypraeidae, Olsson (1965), J. Gibson-Smith (1976), and S. Hoerle 8 E. Vokes (1978) on the Volutidae, E. Vokes (1967a, b, 1968, 1970, 1974 and 1975) on the ВЕНСТ CAENOGASTROPODS 311 Muricidae, J. Gibson-Smith & W. Gibson- Smith (1974) on the genus Strombina, and Hodson (1926) on the Turritellidae. DISCUSSION The Colombian-Venezuelan Neogene relict pocket represents the oldest known intact shallow water molluscan fauna in the western Atlantic. The Gulfs of Venezuela and Triste regions contain extant elements of several well-documented Neogene Caribbean fossil formations. Of the forty-five additional extant species reported in the following systematic section and shown here in Table 1, nineteen have also been found in the Bowden forma- tion of Jamaica, fifteen in the Gurabo forma- tion of Santo Domingo, seventeen in the Gatun formation of Panamá, four from the Grand Bay formation of Carriacou, and twelve from the Mare and Cabo Blanco formations of Venezuela. The widespread Neogene genera Conomitra, Panamurex s.s., and Parabor- sonia, previously thought to have been extinct since preglacial times, have been found to be living components of the relict assemblage. The paciphilic genera Ag/adrillia, Aphera, and Truncaria are also present in the relict pocket and represent the first records of these taxa in the Recent Atlantic. The archaic genera Fusiturricula, Siphocypraea, Strombina, and Subcancilla, previously thought to be repre- sented in the Caribbean by one or only a few species, have been found to be more diverse. These diversity trends, along with the pres- ence of relict and paciphilic genera, empha- size the anachronistic aspect of the fauna and its closeness to the Recent Panamic Mollus- can Province. Table 1. Additional relict species from northern South America. CAPA AR A A AAA Turritellidae 1. Turritella paraguanensis F. Hodson, 1926 Calyptraeidae 2. Crucibulum marense Weisbord, 1962 3. Crucibulum springvaleense Rutsch, 1942 Naticidae 4. Natica stenopa Woodring, 1957 Cypraeidae 5. Siphocypraea henekeni (Sowerby, 1850) Cassidae 6. Morum dominguense (Sowerby, 1850) 7. Sconsia laevigata (Sowerby, 1850) TABLE 1 (Continued) Ficidae 8. Ficus pilsbryi (В. Smith, 1970) Muricidae 9. Panamurex gatunensis (Brown 4 Pilsbry, 1911) Columbellidae 10. Strombina caboblanquensis Weisbord, 1962 11. Strombina sp. Buccinidae 12. Antillophos elegans (Guppy, 1866) 13. Truncaria sp. Fasciolariidae 14. Latirus anapetes Woodring, 1964 15. Fusinus caboblanquensis Weisbord, 1962 16. Fusinus marensis Weisbord, 1962 Olividae 17. Ancilla venezuelana Weisbord, 1962 18. Ancilla sp. 19. Oliva schepmani Weisbord, 1962 Mitridae 20. Subcancilla illacidata (Woodring, 1928) 21. Subcancilla rhadina (Woodring, 1928) 22. Subcancilla venezuelana (F. Hodson, 1931) Volutomitridae 23. Conomitra caribbeana Weisbord, 1929 24. Conomitra lehneri Jung, 1971 25. Conomitra sp. Volutidae 26. Lyria cf. limata S. Hoerle and E. Vokes, 1978 Marginellidae 27. Persicula hodsoni Weisbord, 1962 Cancellariidae 28. Agatrix epomis (Woodring, 1928) 29. Aphera islacolonis (Maury, 1917) Conidae 30. Conus consobrinus Sowerby, 1850 31. Conus planiliratus Sowerby, 1850 32. Conus symmetricus Sowerby, 1850 Terebridae 33. Strioterebrum bowdenensis (Woodring, 1928) 34. Strioterebrum gatunensis kugleri (Rutsch, 1934) 35. Strioterebrum ischna (Woodring, 1928) 36. Strioterebrum quadrispiralis (Weisbord, 1962) 37. Strioterebrum trispiralis (Weisbord, 1962) Turridae 38. Polystira barretti (Guppy, 1866) 39. Agladrillia lassula Jung, 1969 40. Hindsiclava consors (Sowerby, 1850) 41. Fusiturricula acra (Woodring, 1970) 42. Fusiturricula humerosa (Gabb, 1873) 43. Fusiturricula ¡ole Woodring, 1928 44. Fusiturricula jaquensis (Sowerby, 1850) 45. Paraborsonia varicosa (Sowerby, 1850) AA TEA AA A A 312 PETUCH In Appendix 2, | give a list of known large caenogastropod species from the Golfo de Triste and Golfo de Venezuela areas, this being gleaned from the pertinent literature and personal communications and observa- tions. Of the 97 known species listed in this appendix, 19 species, Siphocypraea don- moorei, S. mus, Calotrophon velero, Phyl- lonotus margaritensis, Murex dommoorei, Mazatlania aciculata, Strombina pumilio, Fusinus closter, Ancilla glabrata, A. tanker- уе Oliva oblonga, Olivella perplexa, Persicula tessellata, Prunum glans, P. pulchrum, Conus optabilis, C. undatus, Clath- odrillia gibbosa, and Hindsiclava chazaliei, are restricted to northern South America, and in particular, the Colombian-Venezuelan coasts. When the 45 additional relicts given in the systematic section are added to this spe- cies list, it can be seen that roughly 46% of the large caenogastropods are Neogene relicts confined to the soft bottom community of the upwelling areas. This percentage would be even larger if the allopatric intertidal com- ponents, such as the littorinids and thaidids, were removed. The survival of the relict pocket into the Recent as an intact entity is probably a func- tion of the substrate and hydrographic condi- tions that are unique to northern South Amer- ica. As shown by Meyer (1977: 45—71), this area is influenced by four continuous wind- driven upwelling systems (Fig. 130) and con- tains cold, atypical Caribbean water. The substrate of the Gulf of Venezuela area 1$ also atypical for the continental Caribbean, con- taining large amounts of coarse quartz sands (Schuchert, 1935: 652-653) and turbid water conditions due to erosion of the Pleistocene fluvially deposited coastline (Petuch, 1979: 221). The Peninsulas of the Guajira and Paraguaná, together the center of distribution of the relict fauna, are desert areas with little or no river input and few mangrove environ- ments. Nutrient enrichment of coastal waters of this area, therefore, derives from the up- wellings and not from terrigenous input. The cold, turbid water and non-reef environ- ment of the northern South American coast most probably acts as a physiological and ecological barrier to typical tropical Caribbean molluscivorous predators. This possible pre- dation barrier may explain why such extreme- ly thin-shelled relict species as Turritella paraguanensis, Natica stenopa, Sipho- cypraea mus, Siphocypraea henekeni, Ficus pilsbryi, Ancilla venezuelana, Oliva schep- mani, Agatrix epomis and Strioterebrum ischna, can exist in abundance in shallow water within this region (Petuch, 1976: 322- 325) but are not found elsewhere in the Carib- bean area. Extrapolating backwards through time, the atypical Caribbean environment of the Recent northern South American coast probably ap- proximates what had been the typical shallow water marine environment of much of the Neogene southern Caribbean. This hypothe- sis seems defensible when considering that the fauna of the relict assemblage had once been widespread throughout the southern Caribbean, but is now found only in the geo- graphically small area of the upwelling sys- tems. Where tropical waters and carbonate substrate environments now predominate in the southern Caribbean, the Neogene fauna is extinct and replaced by a widespread post- Pleistocene-Holocene fauna. On the other hand, where the environment comprises tur- bid, nutrient-rich water with temperatures be- low 25°C and a substrate of silicoclastic sedi- ments, a typical Neogene southern Caribbean gastropod assemblage is still extant (Vermeij, 1978: 231-236). The northern South American Neogene environment and its accompanying relict fauna is geographically well-defined. In the west, the carbonate and coral reef environ- ment of the San Blas Islands, Panamá, acts as an effective barrier to the Neogene shallow water relicts. In the east, the extensive brack- ish water estuary at the mouth of the Orinoco River also acts as an effective barrier, limiting most of the relict forms. Only Fusiturricula jaquensis and possibly Sconsia laevigata (as S. nephele) have been reported from outside these geographical and physiological bound- aries. Offshore islands, such as the Nether- lands Antilles, Las Aves, and Los Roques, all have extensive coral reef growth, lack contin- uous upwelling systems, and contain a more typical Caribbean-West Indian gastropod fauna. No Neogene relicts have yet been re- ported from any of these island groups or their associated carbonate environments. Philosophically, the relict pocket poses an interesting question; how can the ancestors of a group of animals still exist and retain an intact community structure while being sur- rounded, contemporaneously, by communi- ties made up of their descendants? This can be clearly seen by the following species pairs—one being the supposed direct ances- tor now found only in the communities com- RELICT CAENOGASTROPODS 313 prising the relict pocket, and the other being the wide-ranging Caribbean descendant; Morum dominguense-Morum dennisoni, Sconsia laevigata-Sconsia striata, Latirus anapetes-Latirus angulatus, Oliva schep- mani-Oliva reticularis, Lyria limata-Lyria beaui, Conus consobrinus-Conus cedonulli, Conus planiliratus-Conus atractus, and Hindsiclava consors-Hindsiclava alesidota. These mollusks occur contemporaneously, although allopatrically, within the same faunal province. In summary, evolution has been greatly slowed in the gastropod fauna found along northern Colombia and Venezuela. Con- versely, faunas in other areas of the Carib- bean have undergone extremely rapid speci- ation since the beginning of the Pleistocene. The concept of a heterochronous Caribbean Province implies several modes of provincial differentiation and development. The homo- chronous sister Panamic Province probably represents an area of more evenly distributed evolutionary pressures, with a nearly homo- geneous fauna within its boundaries. The Caribbean, on the other hand, represents an area that was exposed to differential specia- tion and extinction pressures. This has mani- fested itself along the northern South Ameri- can coast as an intact but spatially reduced Neogene relict pocket surronded by a large cortex of post-Pleistocene-derived species. SYSTEMATIC SECTION Descriptions of Neogene relict Caenogastropods As briefly outlined in the literature section, the first detailed analysis and discussion of the Gatunian relict pocket was done by J. Gibson-Smith & W. Gibson-Smith in 1979. They recognized two Pliocene archaeo- gastropods, Тедша puntagordana Weisbord, 1962 and Parviturbo venezuelensis Weis- bord, 1962 as extant along the Venezuelan coast. In their discussion of Weisbord's Plio- cene caenogastropod taxa (1979: 24, 26), the Gibson-Smiths recognized two more extant species, Fasciolaria hollisteri Weisbord, 1962 (Е. tulipa Linnaeus, 1758 variety?) and Oliva schepmani Weisbord, 1962. They synony- mized three others as forms of Recent spe- cies; Fusinus caboblanquensis Weisbord, 1962 as a form of F. closter (Philippi, 1851), Ancilla venezuelana Weisbord, 1962 as A. tankervillei (Swainson, 1825), and Persicula hodsoni Weisbord, 1962 as P. interrupto- lineata (Megerle von Mühlfeld, 1816). | recog- nize the last three taxa of Weisbord as valid relict species. These, along with Oliva schepmani, Crucibulum marense, Strombina caboblanquensis, Fusinus marensis, Strio- terebrum quadrispiralis, and $. trispiralis, bring to a total of twelve the number of fossil gastropods described by Weisbord that can be shown to be extant. Besides the five relict caenogastropods discussed by the Gibson-Smiths, two other relicts have been reported in the recent litera- ture. These are Agatrix epomis (Woodring, 1928), shown to be extant along Venezuela by Petit (1976), and Fusiturricula jaquensis (Sowerby, 1850), recognized as still living off Surinam by Abbott (1974). This last relict was described as a new species of Knefastia (a genus not found in the western Atlantic) by Princz (1980: 70-72). A number of other well- known northern South American caeno- gastropods, all described from Recent speci- mens, were also shown by Weisbord (1962) to have existed in the Venezuelan Pliocene assemblages, and as such, constitute relict species. These include Siphocypraea donmoorei Petuch, 1979 (as S. henekeni variety), Мигех donmoorei Bullis, 1960 (as М. recurvirostris), Calotrophon velero (E. Vokes, 1970) (as “Latirus” recticanalis), Ancilla tankervillei (Swainson, 1825), Mazatlania aciculata (Lamarck, 1822), Conus puncticu- latus Hwass, 1792 (as C. jaspideus caboblanquensis), and Clathodrillia gibbosa (Born, 1778). As shown by myself (1979), Bayer (1971), E. Vokes (1970), and Vermeij (1978: 231-235), these species are restricted to the northern coast of South America and fall within the ranges of the other known relict species. Adding to the preliminary works of Weis- bord and the Gibson-Smiths, | herein rede- scribe several more previously unknown Neogene relict species. Family Turritellidae Genus Turritella Lamarck, 1799 1. Turritella paraguanensis Hodson, 1926 Figs. 1-2 Turritella variegata paraguanensis Hodson, 1926: 31,12 11195:.2, 7. PETUCH yan 5 AA Ud Fr A ll tr LE ви +% | = = = > > 4 y Y ee 2 FIGS. 1-13. 1-2. Turritella paraguanensis Hodson: USNM 784451, L = 81 тт. 3—4. Turritella variegata (Linnaeus): P-712, L = 78 mm. 5-6. Crucibulum marense Weisbord: USNM 784452, L = 16 mm. ба. Crucibulum marense attached to Polystira barretti shell. 7-9. Crucibulum springvaleense Rutsch: P-712, L = 18 mm. 10-11. Майса stenopa Woodring: USNM 784570, L = 29 mm. 12-13. Siphocypraea mus (Linnaeus): UMML 8278, L = 31 mm. RELICT CAENOGASTROPODS 315 Turritella maiquetiana Weisbord, 1962: 146- 1507 plo 11; figs: 1-16. Material examined—5 specimens, lengths 75-85 mm, exposed at low tide, Amuay Bay, Paraguaná Peninsula, Estado Falcón, Vene- zuela, 21 March 1979, UMML 8276; 7 speci- mens, lengths 80-96 mm, same locality and date, USNM 784451. Major citations—Redescribed in detail, with diagnosis, by Weisbord, 1962 (as Т. maique- tiana n.sp.). Additions to original description—Shell surface smooth, slightly shiny, with a silky tex- ture; color variable, usually gray or bluish- gray with numerous fine revolving lines of black dots; some specimens deep blue-black; base color overlaid with intermittent vertical flammules of dark gray or black; dark speci- mens frequently with white or pale gray verti- cal flammules alternating with dark flam- mules. Remarks— Although originally described as a subspecies of the Recent Т. variegata (Linnaeus, 1758), Т. paraguanensis is also a well-known member of Plio-Pleistocene as- semblages throughout Venezuela (Weisbord, 1962: 146-147). Weisbord described (as 7. maiquetiana) fossil specimens of Т. para- guanensis from the Pliocene beds of the Playa Grande and Lower Mare formations. This description is important because of the emphasis on the form and structure of the early whorls and their use in separating 7. paraguanensis from other known species. Turritella paraguanensis belongs to a long and species-rich lineage that originated with the lower Miocene Venezuelan 7. berjadinen- 5/5 (Hodson, 1926, pl. 20, fig. 5) and the Мю- cene Colombian 7. cartagenensis Pilsbry and Brown, 1917. In the Recent fauna, 7. para- guanensis most closely resembles 7. banksi Reeve, 1849 and Т. gonostoma Valenci- ennes, 1832 from the Panamic Province. Tur- ritella variegata (Figs. 3—4) is the only Atlantic species that shows any relationship to 7. paraguanensis. The Venezuelan relict differs from 7. variegata in being larger, by having a blue-gray ground color (like Т. gonostoma), and by lacking the heavy spiral sculpturing of T. variegata. The shell outlines of the two spe- cies also differ consistently; that of Т. varie- gata is straight-sided while Т. paraguanensis is distinctly turreted due to the characteristic- ally deeply impressed suture. Fossil distribution—Playa Grande and Cabo Blanco formations, Venezuela. Recent distribution—At present, known only from the Golfo de Venezuela, in shallow bays. Family Calyptraeidae Genus Crucibulum Schumacher, 1817 2. Crucibulum (Dispotaea) marense Weisbord, 1962 Figs. 56a Crucibulum (Dispotaea) marense Weisbord, 1962: 218, pl. 20, figs. 10, 11. Material examined—Length 16 mm, width 15 mm, height 12 mm, attached to a living Polystira barretti trawled from 35 m depth off Cabo La Vela, Peninsula de Guajira, Colom- bia (12°10’М, 72°15’W), December, 1974, USNM 784452. Additions to original description—Shell waxy, smooth; color uniform pale yellow- orange; internal cup translucent white. Remarks—Crucibulum marense differs from most known Atlantic calyptraeids, both fossil and living, in having only a small portion of the margin of the internal cup adherent to the shell interior. In this respect, C. marense most closely resembles the Recent C. per- sonatum Keen, 1958 from the Panamic Prov- ince. Crucibulum waltonense Gardner, 1947 from the Alum Bluff series of Florida and С. ecuadorense Olsson, 1932 from the Progreso formation of Ecuador are similar to C. marense in having the internal cup adherent to the shell interior; but their attachments are larger, often including over half of the margin of the internal cup. The single living specimen of C. marense was found attached to the posterior columel- lar area, near the anal slit, of another relict species, Polystira barretti (Guppy, 1886). As can be seen in Fig. 5, the shell margin of C. marense corresponds perfectly to the body whorl of the Polystira. Interestingly enough, the unusual horizontally-arranged primary sculpture of C. marense also corresponds to the raised spiral cords on the body whorl of the turrid, and represents a xenomorphic growth pattern. Although the holotype of C. marense lacks the raised horizontal ribs of the Recent Colombian specimen, the structure of the protoconch, the shape and form of the internal cup, and the dichotomous fine sur- face are identical in both specimens. The presence of raised horizontal ribs appears to 316 РЕТУСН be а response to living on the heavily sculp- tured substrate of the turrid shell. E. Vokes (in litt.) has suggested that this and the following species may actually be forms of C. planum Schumacher, a poorly- known and “lost” species. Until the system- atics of this complex group is better known and since this shell is well-illustrated and de- scribed in Weisbord's paper, | prefer to use the taxon C. marense for the Venezuelan fos- sil and living species. Fossil distribution—Mare formation, Vene- zuela. Recent distribution—Known only from 35 m depth off the Peninsula de Guajira, Colombia. 3. Crucibulum (Dispotaea) springvaleense Rutsch, 1942 Figs. 7-9 Crucibulum springvaleense Rutsch, 1942: 138, pl. 4, fig. 8. Woodring, 1957: 83-84, pl. 19, figs. 8-10. Material examined—Lengths 18 mm and 17mm, P-712 (11%08'N, 63°18’W), 25m depth. Major citations—Redescribed in detail, with diagnosis, by Woodring, 1957. Additions to original description—Shell color pale yellow-tan to yellow-white with radiating patches of reddish-brown dots and flammules; color pattern readily visible on both surfaces; internal cup translucent white, one-third adherent to shell interior; external surface with radiating raised ridges. Remarks— The specimen illustrated here is most probably a juvenile, as it is indistinguish- able from the early stages of the fossil illus- trated by Woodring (1957, pl. 19, figs. 8, 9). Crucibulum springvaleense is similar to C. marense but can be separated by having a larger internal cup attachment and a different color pattern. By having attached internal cups, both C. marense and C. springvaleense differ from the other northern South American congener, C. auricula (Gmelin, 1791). Fossil distribution—Springvale formation, Trinidad; Turbera formation, Colombia; Gatun formation, Costa Rica and Panama. Recent distribution—Known only from off the Venezuelan coast, at depths of around 25 т: Family Naticidae Genus Natica Scopoli, 1777 Subgenus Naticarius Duméril, 1805 4. Natica stenopa Woodring, 1957 Figs. 10-11 Natica (Naticarius) stenopa Woodring, 1957: 85-86, pl. 20, figs. 4-6. Material examined—Lengths 29 mm and 20 mm, trawled by commercial shrimp boats from 35 m depth in Golfo de Triste, Venezuela (11°42’N, 69°22.5'W), March, 1979, USNM 784570. Additions to original description—Shell pale cream-gray becoming gray on spire; margin of suture white; interior of aperture white with wide pale brown band. Remarks—Although somewhat larger than the fossil specimens illustrated by Woodring, the two Recent specimens agree quite closely with the illustrations of Woodring's types. The specimen shown here has the characteristic spire sculpture consisting of short, retractive axial grooves that extend from the suture onto the shoulder of later whorls. Fossil distribution—Gatun Panamá. Recent distribution—35 т depth in Golfo de Triste, Venezuela. formation, Family Cypraeidae Genus Siphocypraea Heilprin, 1886 The Siphocypraea henekeni species complex The Tertiary South American and West Indian Siphocypraea species complex has been one of the more controversial groups in Caribbean paleontological studies. As abundant indicator organisms, members of the genus are well-documented in the litera- ture. During the last several decades, two dif- ferent approaches have been taken by fossil cypraeid workers, leading to two schools of thought on speciation within the complex and treatment of specific morphological char- acters. One group, principally Maury (1925) Schilder (1939), and Ingram (1939, 1940, 1947a, b), described several new species of what appeared to be a close-knit complex centered around the widespread Sipho- cypraea henekeni (Sowerby, 1850). These taxa were erected on what some workers RELICT CAENOGASTROPODS 317 consider trivial, although consistent charac- ters. The second group, including Woodring (1959), Pflug (1961), and Weisbord (1962), has argued that the species of Maury, Schilder, Ingram and others are to be con- sidered intraspecific variants, or at best popu- lation variants, of a morphologically plastic S. henekeni. Until 1979, only a single living species of the S. henekeni lineage, S. mus (Linnaeus, 1758), was known to exist. In that year, | de- scribed a second living species, S. don- moorei, from off the coasts of Colombia and Venezuela. It is now recognized that S. don- moorei is also present in the fossil record, as demonstrated by the specimen illustrated by Weisbord (1962, pl. 22, figs. 5, 6, as “$. henekeni’). In March of 1979, five specimens of another species of Siphocypraea were taken from off the Venezuelan coast. The anatomy of this Siphocypraea differed from the anatomies of the other two living species. The shell morphology of this third species compared very closely with specimens of $. henekeni illustrated by Woodring (1959, pl. 32, figs. 1, 4, 6, 9). It is now apparent that there are, in fact, three living species of Siphocypraea, with the five Venezuelan shells being representative of an extant popu- lation of the previously ubiquitous $. henekeni. When the shell morphologies of S. mus, S. donmoorei, and S. henekeni are compared, no striking differences can be observed; this morphological conservatism is the principal reason why the shells of the three species had been confused for so long. When the liv- ing animals are compared, the three species can be easily separated. Siphocypraea henekeni has a uniformly bright orange-red animal, with long, unbranched mantle papillae (Fig. 129). The animal of S. donmoorei is white with only a few scattered patches of gray, while the mantle papillae are elongate and dendritic (Fig. 128). The gray and black- mottled S. mus has very reduced, wart-like papillae, these often being white or pale orange (Fig. 127). Considering that the three extant Siphocypraea species have similar-looking shells but different animals, one has to view the supposed variants of S. henekeni in a dif- ferent light. In all probability, the consistent morphologies seen in many of the fossil forms, such as the thickened dorsal callus of S. lacrimula (Maury, 1925) and the elongate and flaring beaks of S. projecta (Ingram, 1947), represent full species characteristics and not merely variations within a single gene pool. Based on the biological evidence seen in the living species, | prefer to treat the taxa of Maury, Ingram, Schilder, etc., as full spe- cies and not as synonyms of S. henekeni. The anatomy and ecology of the well-known S. mus was outlined by Petuch (1979: 217-220). Since S. donmoorei is also a relict, it is in- cluded here to facilitate a comparison with the other two living species of Siphocypraea. Siphocypraea donmoorei Petuch, 1979 Figs. 18-21 Cypraea (Muracypraea) henekeni, Weisbord, 1962 (non C. henekeni Sowerby, 1850), 236-238, pl. 22, figs. 5, 6. Siphocypraea donmoorei Petuch, 1979: 216- 225, pl. 1, figs. О-1. Material examined—Holotype, USNM 770731, length 64 mm, 37 m depth off Cabo La Vela, Colombia (12%10'N, 72°15'W); length 55 mm, 39 т depth off Cartagena, Colombia (10°22’N, 75°47'W), UMML 8162; length 59 mm, 40 m depth in Golfo de Urabá, Colombia (8°38’N, 77°2'W), UMML 8161; length 55 mm, 30m depth off Cartagena, Colombia, USNM 770732; length 60 mm, 37 m depth off Cabo La Vela, Colombia, USNM 784453. Major citations—Detailed description of fossil specimens given by Weisbord, 1962 (as C. henekeni variety); living specimens de- scribed by Petuch, 1979. Remarks—Morphologically, the shells of $. donmoorei and $. henekeni appear similar. The animals of the two species are, however, very different (as illustrated here). Upon closer examination, the shells of the two spe- cies do show some consistent differences, and these can be used in separating mixed lots of shells without the animals. The shell of $. donmoorei is larger than that of $. henekeni (67-75 mm average length as op- posed to the 40-50 mm average length of 5. henekeni), darker in color, and has а uni- formly narrow, arcuate aperture. The labial and columellar dentition of $. donmoorei are coarser and more developed, but less numer- ous than the dentition of $. henekeni. | had previously stated (Petuch, 1979: 224) that the cypraeid illustrated by Weisbord (1962, pl. 22, figs. 5, 6) probably represented РЕТУСН FIGS. 14-23. 14-15. Siphocypraea mus (Linnaeus): USNM 784454, L = 50 тт. 16-17. Siphocypraea henekeni (Sowerby): USNM 784455, L = 48 mm. 18-19. Siphocypraea donmoorei Petuch: USNM 770732, L = 55 mm. 20-21. Siphocypraea donmoorei Petuch (holotype): USNM 770731, Е = 60 mm. 22-23. Siphocypraea henekeni (Sowerby): USNM 784455, L = 46 mm. ВЕНСТ CAENOGASTROPODS 319 the ancestor of both $. mus and $. don- moorei. п March, 1979, | had the opportunity to study several similar specimens, taken from the same fossil deposits of Weisbord's, in the Gibson-Smith collection, Caracas, Vene- zuela. These fossil specimens are identicai to living specimens of $. donmoorei taken from off the Guajira Peninsula of Colombia. Geo- logically, $. donmoorei is now known to range from the middle Pliocene to the Recent. The closely-related S. mus most probably repre- sents a Pleistocene offshoot of the wide-rang- ing S. donmoorei; having become adapted to the Thalassia-based ecosystem of the Golfo de Venezuela region. Fossil distribution—Mare Grande formations, Venezuela. Recent distribution—Golfo de Urabá, along the Colombian coast and Peninsula de Guajira, into the Golfo de Venezuela off the Peninsula de Paraguana, Venezuela, in depth ranging from 15-50 m. and Playa 5. Siphocypraea henekeni (Sowerby, 1850) Figs. 16-17, 22-23 Cypraea henekeri Sowerby, 1850: 45, pl. 9, fig. 3. Cypraea henekeni Gabb, 1873: 235. Emen- dation for С. henekeri Sowerby. Woodring, 1959: 194-196, pl. 31, figs. 6-10, pl. 32, figs. 1, 4, 6, 9. Pflug, 1961: 30-32. Petuch, 1979: 216-217, table 1. With discussion of Gabb's emendation. Material examined—Lengths 48 mm and 46 mm, from 20 m depth in Golfo de Triste, Venezuela (11°42'N, 69°40'W), March, 1979, USNM 784455; length 37 mm, from 25m depth off Punto Fijo, Golfo de Venezuela, Venezuela (11°52’, 70°22'W), March, 1979, USNM 784456; lengths 42 mm and 38 mm, from 20m depth in Golfo de Venezuela, UMML 8277; preserved animals in mollusk collection of INTECMAR, Universidad Simón Bolívar, Caracas, Venezuela. Major citations—Redescribed in detail, with synonymies, by Woodring, 1959 and Pflug, 1961. Additions to original description—Color of dorsum pale bluish-white to light blue with numerous pale tan spots; spots distinctly separate and not coalescing; some speci- mens with pale brown bar-shaped markings along sides and dark brown or black patch on posterior and near apex; base of shell flat- tened, varying in color from dark tan to dark brown; pale dorsum color and dark base color meeting along flattened lateral margin; columellar and labial teeth dark chocolate brown; interior of aperture pale tan to yellow; dorsum of adult specimens with deeply in- cised central sulcus running from apex to anterior tip; beaks of adult specimens charac- teristically flaring, ear-like, greatly extended; living animal uniformly orange-red with elon- gate, simple mantle papillae. Remarks—A comparison between the shell morphologies of $. donmoorei and $. henekeni is outlined in the previous species description. Siphocypraea henekeni differs from S. mus not only in coloration of the living animals and the structure of the mantle papillae but also in shell morphology. Siphocypraea mus has a darkly colored, in- flated shell with weak labial dentition and poorly developed (and often absent) columel- lar dentition. The oddly-flattened, pale- colored shell of S. henekeni is in direct con- trast, with the columellar dentition well devel- oped and the labial dentition coarser and more numerous. Of the living Siphocypraea, the enlarged, flange-like beaks are unique to $. henekeni. Fossil distribution—Bowden formation, Jamaica; Gurabo formation, Santo Domingo; Springvale formation, Trinidad; Cantaure and Punta Gavilan formations, Venezuela; Gatun formation, Panama; Esmeralda formation, Ecuador; Gatun formation, Costa Rica. Recent distribution—From Golfo de Vene- zuela to Golfo de Triste, Venezuela, at depths of 20-30 m. Family Cassidae Genus Morum Röding, 1798 6. Morum (Oniscidia) dominguense (Sower- by, 1850) Figs. 24-25 Oniscia dominguensis Sowerby, 1850: 47, pl. 10, fig. 3. Morum dominguense, Pilsbry, 1922: 363. Pflug, 1961: 37-38, pl. 7, figs. 9-12, pl. 8, figs. 1,2, 9-7: Morum dennisoni Bayer, 1971 (non Reeve, 1843): 140, fig. 16, lower two figures. Material examined—Length 35 mm, P-772 (12°20:27М, 71°55.1'W), 11) medepth: Major citations—Redescribed in detail by Pflug, 1961, with illustration of lectotype (pl. 8, figs. 6, 7). PETUCH FIGS. 24-34. 24-25. Morum dominguense (Sowerby): P-772, L = 35 тт. 26-28. Sconsia laevigata (Sowerby): USNM 784457, L = 71 mm. 29-31. Ficus pilsbryi (В. Smith): USNM 784458, L = 76 mm. 32. Ficus pilsbryi, same specimen, showing detail of body whorl sculpture. 33-34. Panamurex gatunensis (Brown 8 Pilsbry): USNM 784571, L = 26 mm. RELICT CAENOGASTROPODS 321 Additions to original descriptions—Shell color pinkish-white with scattered patches of bright pink; body whorl with two reddish- brown bands; spire with intermittent large reddish-brown patches; parietal shield sal- mon-pink with lavender border and white pustules; outer lip salmon-pink with numerous lavender-purple radiating bars; teeth of outer Ир salmon-pink; interior of aperture white; protoconch and early whorls white. Remarks—Bayer (1971, fig. 16, lower two figures) illustrated this species but considered it a variety of Morum dennisoni (Reeve). Morum dominguense differs from that spe- cies, however, in having a broader, tabulate spire, by having a smaller and less developed parietal shield, and by having larger and more developed labial teeth. The two species can also be differentiated by the colors of the parietal shields; М. dennisoni has a bright red-orange shield, while that of M. doming- uense is a rich salmon-pink with lavender shadings. Although the single known Recent speci- men closely resembles the fossil illustrated by Pflug (pl. 8, figs. 1, 2) in both shape and sculp- turing, it differs in having more varices per whorl (12 in the fossil, 16 in the Recent speci- men). Otherwise, the fossil and living forms are virtually indistinguishable. Fossil distribution—Gurabo formation, Santo Domingo; Bowden formation, Jamaica; Gatun formation of Costa Rica and Panama. Recent distribution—Known only from off the Peninsula de Guajira, Colombia, 11 т depth. Genus Sconsia Gray, 1847 7. Sconsia laevigata (Sowerby, 1850) Figs. 26-28 Cassidaria laevigata Sowerby, 1850: 47, pl. LO Во. 2. Sconsia laevigata, Maury, 1917: 275, pl. 19, fig. 2. Woodring, 1928: 309-310. Woodring, 1959: 201-202. Pflug, 1961: 36-37, pl. 7, figs. 1-8. Sconsia striata Bayer, 1971 (non Lamarck, 1816): fig. 14, lower two figures. Material examined—Length 71 mm, trawled by commercial shrimp boat, from 35 m depth off Cartagena, Colombia, Decem- ber, 1976, USNM 784457; length 43 mm, P-353 (8 13:2 N, M6:9-051:W), 0 25°m depth; lengths 49 тт and 51 mm, P-361 (8°52'N, 76°37'W), 37m depth; length 62 тт, P-362 (8°57'N, 76°34’W), 60m depth; length 24mm, Р-367 (931'N, 75°50'W), 36 т depth; length 52 mm, Р-756 (11°33.1'N, 69°12.6’W), 25 т depth; length 32 mm, P-760 (12°15.4'N, 69°57.5’W), 62 т depth. Major citations—Redescribed by Pflug, 1961, with illustrations of the lectotype and а representative series (pl. 7, figs. 1-8). Additions to original description—Shell color white with variable amount of red-brown mottling, usually in form of vertical flammules in zebra-like pattern; color pattern may vary from pure white to having 4-6 rows of brown checkers or alternating vertical flammules composed of coalesced checkered bands (as in Bayer, 1971: fig. 14, lower two figures); spire white, often with band of brown flam- mules; interior of aperture white; outer lip well developed, thickened; porcelaneous white. Remarks—Bayer (1971, fig. 14, lower two specimens) illustrated a Recent specimen of Sconsia laevigata (P-353) but misidentified it as a variant of S. striata. Sconsia laevigata is a common species in offshore Colombian and Venezuelan waters and is morphologically consistent; a large series from all along this area shows little variation in form and color pattern. This consistency is also seen in the fossil record. The Recent specimen illustrated here is identical to the fossil illustrated by Pflug (1961, pl. 7, figs. 1-4, 6). The characteristic cancel- late sculpture on the outer lip of the fossil, a feature well-illustrated by Pflug (fig. 6), is the same as that of the Recent specimen (Fig. 28). The morphologically very similar S. nephele Bayer, 1971 from 18m depth off Grenada may possibly be only a color morph of S. laevigata. This is substantiated by the fact that several of the specimens collected (P-363, P-760, P-362) had patterns com- posed of bands of alternating light and dark blocks which, in turn, were overlaid by a pat- tern of vertical flammules. These specimens act as intergrades between the zebra color morph illustrated here and the checkered S. nephele. E. Vokes (in litt.) has stated that a fossil specimen of $. laevigata examined by her under ultraviolet light revealed a color pat- tern like that of S. nephele or like that of the P-760 specimen of S. /aevigata. Fossil distribution—Bowden formation, Jamaica; Gurabo formation, Santo Domingo; 322 PETUCH Gatun formation, Costa Rica and Panama; Esmeraldas formation, Ecuador. Recent distribution—From Golfo de Urabá, Colombia, along Colombian coast, into the Golfo de Venezuela, and to at least Isla Margarita, Venezuela, at depths of 15-16 m. Family Ficidae Genus Ficus Róding, 1798 8. Ficus pilsbryi (В. Smith, 1907) Figs. 29-32 Pyrula pilsbryi B. Smith, 1907: 213-214, fig. 1. Maury, 1917: 277. Ficus pilsbryi, Woodring, 1928: 313-314, pl. 20; fig: 9, pl. 24 165. 1, 2. Material examined—Lengths 82mm, 76 тт, and 71 тт, trawled by commercial shrimp boats from 15 m depth off Punto Fijo, Golfo de Venezuela, March, 1979, USNM 784458; length 29mm, P-709 (11°24.7°N, 62°40.5'W), 46m depth; length 74mm, Р-712 (11°08.0’М, 63°18.0’W), 25 т depth; lengths 73mm and 54mm, P-767 (12°16.1’N, 71%03.3'W), 25 m depth; length 31 mm, Р-772 (12°20.2'N, 71°55.1’W), 11m depth. Major citations—Redescribed, with diag- nosis, by Woodring, 1928. Additions to original description—Shell sculpture cancellate, consisting of strong pri- mary spiral threads and one weak secondary thread between each pair of primary spiral threads; spiral sculpture intersected by strong axial threads equal to primary spiral threads (Fig. 32); color dark brownish-tan with scat- tered wide vertical bands of darker brown; strong axial threads with alternating white and brown dashes; weak secondary axial threads uniformly dark tan; spire white with white cal- lus over last whorl; protoconch white; interior of aperture tan turning white towards Пр. Remarks—This is the first shallow water Ficus reported from the Recent southern Car- ibbean. Although F. pilsbryi bears some re- semblance to the Carolinian F. communis (Say), the relict's brown color and character- istic cancellate sculpture make it readily sepa- rable from the Carolinian species. The sculpture of the fossil specimen illus- trated by Woodring (1928, pl. 20, fig. 9) is identical to that of the Recent Venezuelan specimens. Fossil distribution—Bowden formation, Jamaica; Gurabo formation, Santo Domingo. Recent distribution—Peninsula de Guajira, Colombia, into the Golfo de Venezuela, and to the Golfo de Triste, Venezuela, at depths of around 10-35 m. Family Muricidae Subfamily Muricinae Rafinesque, 1815 Genus Panamurex Woodring, 1959 9. Panamurex gatunensis (Brown & Pilsbry, 1911) Figs. 33-34 Murex (Phyllonotus) gatunensis Brown & Pilsbry, 1911: 354, pl. 26, fig. 2. Paziella (Panamurex) gatunensis, Woodring, 1959: 217-218, pl. 35, figs. 6, 7, 10. Material examined—Length 26 mm, trawled by commercial shrimp boats from 35 m depth in Golfo de Triste, Venezuela, November, 1977, USNM 784571. Major citations—Redescribed in detail, with diagnosis and illustrations, by Woodring, 1959: Additions to original description—Shell pure white, covered with a white chalky in- tritacalx. Remarks—The single Recent specimen of P. gatunensis 1$ very similar to the specimens figured by Woodring both in size and shell sculpture. Although also placed in Panamurex by E. Vokes (1971: 114), the sympatric Co- lombian-Venezuelan Calotrophon velero (Vokes, 1970) was shown not to belong to Woodring's genus by Radwin 8 D'Attilio (1976: 31-32). Panamurex gatunensis, there- fore, is the first and only known living member of this once widespread group of muricids. E. Vokes (in litt.) has suggested that my specimen represents a new species that 1$ very close to P. gatunensis. The living Venezuelan specimen shown here, however, is identical in every way to a Gatun formation fossil specimen in the Vermeij collection at the University of Maryland. Since my living specimen 15 identical to the Panamanian fos- sil, in size, aperture shape, and having varical flanges, and since the morphological variation of the populations of P. gatunensis, both living and fossil, is not known, | prefer to retain the taxon P. gatunensis for the single living Venezuelan specimen. RELICT CAENOGASTROPODS 323 Fossil distribution—Gatun formation, Panama; Tuberá formation, Colombia. Recent distribution—35 m depth in Golfo de Triste, Venezuela. Family Columbellidae Genus Strombina Mörch, 1852 10. Strombina caboblanquensis Weisbord, 1962 Figs. 35-36 Strombina caboblanquensis Weisbord, 1962: 323-327, pl. 28, figs. 25-30, pl. 29, figs. 1-4. Gibson-Smith & Gibson-Smith, 1974: 24. Material examined—7 specimens, lengths 15-20 mm, P-712 (11°08’N, 63°18’W), 25 m depth; 5 specimens, lengths 13-16 mm, Р-721 (11%06.5'N, 64°22.5'W), 26 m depth; lengths 19 mm and 20 mm, trawled by com- mercial shrimp boat, 20 m depth in Golfo de Triste, Venezuela (11°42’N, 69°22.5’W), March 1979, USNM 784459. Major citations— Expanded diagnosis given by J. Gibson-Smith 8 W. Gibson-Smith, 1974. Additions to original description—Shell shiny, waxy; color bright yellow-ochre with scattered irregular white flammules; outer lip pure white; aperture and labial callus pure white; columellar callus yellow. Remarks—As outlined by Weisbord (1962: 326) and Gibson-Smith & Gibson-Smith (1974: 53-54), S. caboblanquensis can be easily separated from its congener, S. pumilio (Reeve, 1859) (Figs. 37-38). The Recent Car- ibbean S. pumilio differs from S. cabo- blanquensis in being a broader, less turreted shell, with a lower spire and with the outer lip more developed and flaring. Strombina caboblanquensis has more and finer varices than does S. pumilio, a character that shows up well in Figs. 35 and 36. Although S. pumilio also has a diamond criss-cross pattern of lines on the body whorl, these are wider apart and are not as developed nor as deeply in- cised as those of S. caboblanquensis. Strombina pumilio was previously thought to have been the only living Atlantic Strom- bina. In 1974, however, Gibson-Smith 4 Gibson-Smith described the second known living species, S. francesae, from Los Roques Islands, Venezuela. Strombina cabo- blanquensis, although described as a fossil, has now been found to be extant along the Venezuelan coast, and represents the third living Atlantic Strombina. А possible fourth Atlantic species is presented here in the fol- lowing description. Fossil distribution—Mare formation and Maiquetia member of the Playa Grande for- mation, Venezuela. Recent distribution—From the Peninsula de Paraguaná to the Golfo de Triste region, Venezuela, at depths of 15—40 т. 11. Strombina sp. Figs. 39—40 Material examined—Length 15 mm, trawled by commercial shrimp boats from 35 m depth in Golfo de Triste, Venezuela, March, 1979, USNM 784572. Description—Shell with 6 whorls, smooth, waxy; body whorl broad, laterally flattened; spire roughly Y total shell length; body whorl with 10 evenly-spaced sharp-edged axial ribs; spire whorls with 10-12 ribs; color white with large yellow patch on dorsum and near aper- ture; spire and protoconch yellow; outer lip and labial dentition white; diamond criss- cross sculpture reduced, absent on axial ribs. Remarks—This small Strombina most probably represents an undescribed species, sympatric with both S. pumilio and S. cabo- blanquensis. | have included it here in order to show that the Strombina fauna of northern South America is actualy larger than was originally assumed. Of the four known Atlantic species of Strombina, the new species 15 the smallest. It differs from the two mainland spe- cies in being squatter, having evenly-spaced smooth axial ribs over the entire dorsum, and in lacking the prominent dorsal hump seen in both $. pumilio and $. caboblanquensis. Family Buccinidae Genus Antillophos Woodring, 1928 12. Antillophos elegans (Guppy, 1866) Figs. 43-44 Phos elegans Guppy, 1866: 290, pl. 16, fig. 13. Maury, 1917: 250-251, pl. 40, fig. 10. Olsson, 1942: 88. Tritiaria (Antillophos) elegans, Woodring, 1928: 262, pl. 16, fig. 1. Antillophos elegans, Pflug, 1961: 46—47, pl. 11, figs. 1-3, 9, 10, 14, 16-19. PETUCH FIGS. 3548. 35-36. Strombina caboblanquensis Weisbord: USNM 784459, L = 20 mm. 37-38. Strombina pumilio (Reeve): Р-712, L = 16 mm. 39-40. Strombina sp.: USNM 784572, L = 15 mm. 41-42. Antillophos сапае! (d'Orbigny): P-712, L = 27 mm. 43-44. Antillophos elegans (Guppy): USNM 784460, L = 20 mm. 45—46. Truncaria sp.: P-734, fragment L = 22 mm. 47-48. Truncaria sp.: P-734, fragment L = 20 тт. RELICT CAENOGASTROPODS 325 Material examined—Lengths 20 mm and 15 mm, trawled by commercial shrimp boats, from 35 m depth, Golfo de Triste, Venezuela, March, 1979, USNM 784460. Major citations—Holotype and representa- tive series illustrated by Pflug, 1961 (pl. 11, figs. 9, 10). Additions to original description—Shell color pale tan with three wide reddish-brown bands, one on shoulder, one around mid- body, one around base; body whorl near lip pure white; interior of aperture white; proto- conch and early whorls pale tan. Remarks—The two specimens from the Golfo de Triste firmly establish this well known and distinctive Pliocene fossil as a component of the Recent Venezuelan fauna. Antillophos elegans can be separated from the South American variety of the ubiquitous Caribbean А. candei (d'Orbigny) (Figs. 41-42) by its smalier size, by having strong, beaded axial costae instead of a cancellate sculpture, and by having a color pattern of three solid red- brown bands. The Recent specimen illustrat- ed is indistinguishable from the fossils Шиз- trated by Pflug (1961, pl. 11, figs. 1, 3, 14, 16), in shape, sculpturing, and size. Antillophos elegans is sympatric with A. candei in the Golfo de Triste. Fossil distribution—Gurabo formation, Santo Domingo; Bowden formation, Jamaica; Springvale formation, Trinidad; Punta Gavilan formation, Venezuela; Limon and Gatun for- mations, Costa Rica. Recent distribution—Known only from the Golfo de Triste, Venezuela, 35 m depth. Genus Truncaria Adams & Reeve, 1848 13. Truncaria sp. Figs. 45-48 Material examined—2 fragments, lengths 22 тт and 20 тт, P-734, 65 т depth off Cabo Cordera, Venezuela. Description—Shell fragments smooth, pale cream colored, with fine brown revolving hair- lines; sculptured with fine raised spiral striae. Remarks—The two Cabo Cordera frag- ments constitute the first record of the paci- philic genus Truncaria in the Atlantic Ocean. Although these fragments represent a new Caribbean species, a formal description will have to wait until better material is collected. Enough of the body whorl is intact, however, to show that the new species is close to its Panamic cognate species, Truncaria brun- neopicta (Dall, 1896). Both species have sculpturing of spiral striae and coloration of brown hairlines. Family Fasciolariidae Subfamily Fasciolariinae Genus Latirus Montfort, 1810 14. Latirus (Polygona) anapetes Woodring, 1964 Figs. 49-50 Latirus (Polygona) anapetes Woodring, 1964: 274, pl. 47, fig. 12. Jung, 1965: 539, pl. 73, el Material examined—Length 55 mm, P-708 (11°24.7'N, 62°40.5'W), 70 т depth; length 40 тт, Р-736 (10%57'N, 65°52’W), 100 m depth; lengths 48 тт and 41 mm, Р-718 (11°22.5'N, 64°08.6'W), 60 m depth. Additions to original description—Shell color pale yellow-orange, darker on early whorls; interior of aperture white. Remarks—The Recent specimen illus- trated here is very similar to the fossil illustrat- ed by Woodring (1964, pl. 47, fig. 12), especi- ally so in that both have a widely-flaring umbil- ical region and a strongly constricted suture. Fossil distribution—Chagres Sandstone, Gatun formation, Panamá. Recent distribution—Off the Venezuelan coast, from the Peninsula de Paria to near Isla Margarita, 60-100 m depth. Subfamily Fusininae Swainson, 1840 Genus Fusinus Rafinesque, 1815 15. Fusinus caboblanquensis Weisbord, 1962 Figs. 51-52 Fusinus closter caboblanquensis Weisbord, 1962: 364-368, pl. 32, figs. 13, 14, pl. 33, figs. 1, 2. J. Gibson-Smith & W. Gibson- Smith, 1979: 26. Material examined—Length 160 mm, trawled by commercial shrimp boats from 25 m depth in Golfo de Venezuela, off Punto Fijo, Peninsula de Paraguana, Venezuela, May, 1976, USNM 784462; length 111 mm, same locality and date, UMML 8280. Additions to original description—Shell color pale tan, becoming darker on siphonal canal; siphonal canal tipped with dark brown; РЕТУСН FIGS. 49-60. 49-50. Latirus anapetes Woodring: Р-708, L = 55 mm. 51-52. Fusinus caboblanquensis Weisbord: USNM 784462, L = 160 mm. 53-54. Fusinus marensis Weisbord: 784463, L = 112 mm. 55-56. Ancilla venezuelana Weisbord: P-722, L = 30 mm. 57-58. Ancilla sp.: USNM 784573, L = 35 mm. 59-60. Oliva schepmani Weisbord: USNM 784464, L = 39 mm. RELICT CAENOGASTROPODS 327 base color overlaid with scattered brown ver- tical lines and spots; prominent raised shoulder keel white on knobs, dark brown in depressions; interior of aperture and columella white; outer lip edged with dark brown dashes; protoconch and early whorls brown. Remarks—Although described as a fossil subspecies of the Recent F. closter (Philippi, 1951), F. caboblanquensis can be separated from that species on the basis of having a well developed and darkly colored shoulder keel. Together, the sharp-angled keeled shoulder and the regularly spaced axial folds give the shell a decidedly knobbed appearance. The more slender, white F. closter from Isla Mar- garita and the Lesser Antilles not only lacks the shoulder keel of F. caboblanquensis but also does not seem to reach the shell length of the relict species. The two species are sympatric along part of the Venezuelan coast. Fossil distribution—Mare and Playa Grande formations, Venezuela. Recent distribution—Known from near Isla Margarita to the Golfo de Venezuela, 25 m depth. 16. Fusinus marensis Weisbord, 1962 Figs. 53-54 Fusinus marensis Weisbord, 1962: 262-264, 032,105. 11, 12: Material examined—Length 112mm, trawled by commercial shrimp boat, 20 m depth, in the Golfo de Venezuela off the Peninsula de Paraguana, Venezuela, March, 1979, USNM 784463; lengths 53 mm and 50 mm, P-758 (11°42.4'N, 69°40'W), 16 т depth; lengths 36mm and 27 mm, P-761 (11°52'N, 70°22'W), 35 т depth. Additions to original description—Shell color tan, overlaid with numerous thin, dark brown vertical flammules; siphonal canal dark brown; protoconch and early whorls brown; columella and interior of aperture white; mar- gin of lip and siphonal canal edged with pur- ple. Remarks—The holotype of F. marensis 1$ a slender juvenile specimen only 50 mm in length. The specimen illustrated here is over twice that length and shows the bulbous, tab- ulate later whorls characteristic of this spe- cies. The early whorls and spiral sculpturing of the Recent specimens are otherwise indis- tinguishable from those of Weisbord's holo- type. Fossil distribution—Mare formation, Venezuela. Recent distribution—Off the Venezuelan coast, from the Golfo de Venezuela to the Golfo de Triste, 15-35 m depth. Family Olividae Genus Ancilla Lamarck, 1799 17. Ancilla venezuelana Weisbord, 1962 Figs. 55-56 Ancilla (Eburna) venezuelana Weisbord, 1962: 393-395, pl. 36, figs. 5, 6. Gibson- Smith 8 Gibson-Smith, 1979: 26. Material examined—Lengths 30 mm and 13 mm, P-722 (11°04’М, 64°44’W), 91m depth; 2 fragments, lengths 21 тт and 20mm, Р-722; length 30mm, Р-734 (11°01.8’М, 65°34.2'W), 64 т depth. Additions to original description—Shell color bright red-orange; spire and upper one- half to one-third of body whorl glazed over with yellow-orange enamel; enamel of sub- sutural area darker orange; fascicular area bi- partite, dark orange on anterior half, paler orange on posterior half; protoconch and glazed early whorls pale yellow-orange; in- terior of aperture orange; operculum thin, corneous, yellow-brown. Remarks— Although considered a synonym of the Recent A. tankervillei (Swainson, 1825) by Gibson-Smith & Gibson-Smith (1979: 26), A. venezuelana is a valid species. This relict has a more slender, fusiform shell, contrast- ing with the turreted outline of A. tankervillei. Ancilla venezuelana can also be distin- guished from A. tankervillei by its smaller size, deep orange-red callus color and form and extent of the subsutural callus. The callus does not extend as far onto the body whorl as that seen in A. tankervillei. The extent of the subsutural nacre of A. venezuelana varies with individuals. One specimen (P-734) had a callus arrangement similar to the holotype il- lustrated by Weisbord (1962, pl. 36, figs. 5, 6), while the specimen here illustrated had a less extensive area of callus. This variability of nacre production is not seen in A. tankervillei. Fossil distribution—Mare formation, Venezuela. Recent distribution—Off the Venezuelan coast from the Golfo de Triste to Isla Margar- ita, in depths of 60-100 m. 328 PETUCH 18. Ancilla sp. Figs. 57-58 Material examined—Length 35 mm, trawled by commercial shrimp boats from 35 m depth in Golfo de Triste, Venezuela, March, 1979, USNM 784573. Description—Shell shiny, elongate; spire turreted, 2 of total shell length; color pale yel- low-orange. Remarks— Although similar to the sympatric Ancilla venezuelana, the single specimen shown here appears to represent a new spe- cies. | have included this new species with the other relicts in order to emphasize the un- usual nature of the northern South American olivid fauna. Here, and nowhere else in the Atlantic Ocean, are five species of Ancilla all living in close proximity; Ancilla venezuelana, along with the new species, being restricted to the Golfo de Triste; A. balteata (Swainson, 1825) being endemic to neighboring Aruba; and A. glabrata (Linnaeus, 1758) and A. tankervillei being widespread along the entire coast. In the Golfo de Triste, the new species, A. venezuelana, A. glabrata, and A. tanker- уе! are all sympatric and the last three men- tioned species are often brought up together in the same net haul. Genus Oliva Hwass, 1789 19. Oliva schepmani Weisbord, 1962 Figs. 59-60 Oliva schepmani Weisbord, 1962: 370-374, pl. 33, figs. 5-13. Gibson-Smith & Gibson- Smith, 1979: 24. Material examined—Length 39 mm, trawled by commercial shrimp boats, from 30m depth, in the Golfo de Triste, Venezuela, March, 1979, USNM 784464; lengths 42 mm and 41 mm, same location, depth, and date, UMML 8281. Additions to original description—Shell color dark greenish-gray mottled with pale brown; base color overlaid with numerous fine pale green triangular markings; body whorl with two faint bands of brown vertical lines, one anterior to mid-body line, one posterior; suture edged with band of alternating black and yellow dashes; spire with thin, pale purple callus; protoconch purplish-brown; interior of aperture pale purple becoming grayish-yellow toward edge of lip; columella lavender-purple; posterior edge of columella with deep purple stain; anterior tip of fasciole with large, deep purple spot; interior of siphonal canal with thin, pale purple callus; protoconch purplish- brown; interior of aperture pale purple becom- ing grayish-yellow toward edge of lip; colum- ella lavender-purple; posterior edge of columella with deep purple stain; anterior tip of fasciole with large, deep purple spot; in- terior of siphonal carıal with purple stain; ani- mal color pale yellow-orange to cream with numerous dark brown little flecks. Remarks—Among the fossil Olividae, O. schepmani most closely resembles O. couvana Maury, 1925, from the Springvale formation of Trinidad. Oliva couvana, how- ever, is a more slender shell, and differs from O. schepmani in having a higher spire. Of the Recent Olividae, the relict most closely re- sembles O. julietta Duclos, 1833 from the Panamic Province. That species has a larger shell and a completely different color pattern, with a yellow and green base color with prominent scattered black spots and white triangles. The Recent O. fulgurator Lamarck, 1810, endemic to Aruba, is similar to O. schepmani but differs in being more inflated, by having a white base color patterned with orange-red flammules and triangles, and by having a much larger protoconch in propor- tion to the shell size. Fossil distribution—Mare, Playa Grande, and Abisinia formations, Venezuela. Recent distribution—Known only from the Golfo de Triste, Venezuela. Family Mitridae Genus Subcancilla Olsson & Harbison, 1953 20. Subcancilla illacidata (Woodring, 1928) Figs. 63-65 Mitra (Tiara) henekeni illacidata Woodring, 1928; 243, pl. 14, fig. 13. Material examined—Three specimens, lengths 15mm, 17mm, 18mm, trawled by commercial shrimp boats from 35 m depth in Golfo de Triste, Venezuela, March, 1979, USNM 784574. Additions to original descriptions—Shell color very pale yellow, some specimens with pale brown vertical flammules. Remarks—Subcancilla illacidata resem- bles members of the widespread Mio- Pliocene S. dariensis (Brown & Pilsbry, 1911) complex but differs in having only three RELICT CAENOGASTROPODS FIGS. 61-76. 61-62. Subcancilla rhadina (Woodring): Р-749, L = 21 тт. 63-64. Subcancilla illacidata (Woodring): USNM 784574, L = 18 mm. 65. Subcancilla illacidata (Woodring): USNM 784574, L = 15 mm. 66. Subcancilla venezuelana (F. Hodson): USNM 784575, L = 24 mm. 67-68. Conomitra caribbeana Weisbord: Р-722, L = 12 mm. 69-70. Сопотйга lehneri Jung: UNSM 784576, L = 15 mm. 71. Сопотйга lehneri Jung: USNM 784576, L = 14 mm. 72-74. Conomitra sp.: USNM 784577, L = 10 mm. 75-76. Lyria cf. limata S. Hoerle & E. Vokes: P-758, fragment, L = 72 mm. 330 columellar plications instead of the four seen in the $. dariensis complex. In the Recent fauna, S. funiculata from the Panamic Prov- ince shows a close relationship to this relict species. Fossil Jamaica. Recent distribution—Known only from the Golfo de Triste, Venezuela, 59 m depth. distribution—Bowden formation, 21. Subcancilla rhadina (Woodring, 1928) Figs. 61-62 Mitra rhadina Woodring, 1928: 243-244, pl. 14, fig. 14. Material examined—Three specimens, lengths 20 mm, 21 mm, and 29 mm, P-749 (10°37'N, 67°57.9'W), 59 т depth. Additions to original description—Shell color pure white with raised light brown spiral cords. Remarks—Subcancilla rhadina resembles the previous species but differs in being a more slender, high-spired shell, and in having more numerous and brown-colored spiral cords. Fossil Jamaica. Recent distribution—35 m depth in Golfo de Triste, Venezuela. distribution—Bowden formation, 22. Subcancilla venezuelana (F. Hodson, 1931) Fig. 66 Mitra dariensis venezuelana F. Hodson in F. Hodson & H. K. Hodson, 1931: 42, pl. 20, figs. 6, 7. Material examined—Two specimens, both lengths 24 mm, trawled by commercial shrimp boats from 35 m depth т Golfo de Triste, March, 1979, USNM 784575. Additions to original description—Shell color pure white. Remarks—The two Recent specimens agree closely with Hodson's figured speci- men. As pointed out by Woodring (1964: 284), S. venezuelana is more closely related to the Miocene $. longa (Gabb, 1873) than to such members of the S. henekeni complex as S. dariensis or S. colombiana (Weisbord, 1929). The major differentiating characteristics that separate S. venezuelana from the S. henekeni complex are the more numerous spiral cords, higher spire, and attenuated body form. PETUCH Fossil distribution—Mio-Pliocene beds of Falcon State, Venezuela. Recent distribution—35 m depth in Golfo de Triste, Venezuela. Family Volutomitridae Genus Conomitra Conrad, 1865 This genus was thought to have been ex- tinct since the upper Miocene (Gardner, 1937: 420), and the following three species consti- tute the first records of the genus from the Recent Caribbean fauna. The Recent species assigned to this genus by Dall (1889: “Conomitra” blakeana, “Сопотйга” laevior, and “Conomitra” intermedia have now been shown to be members of the genus Micro- voluta (Abbott, 1974: 240-241). 23. Conomitra caribbeana Weisbord, 1929 Figs. 67-68 Conomitra caribbeana Weisbord, 1929: 48, pl. 6, figs. 14, 15. Material examined—Two specimens, lengths 12 mm and 14 тт, P-722 (11°0.4’N, 64°44'W), 91 m depth. Additions to original description—Shell color tan, with two narrow white bands, one above mid-body line, one below mid-body line; interior of aperture tan with white band; ptotoconch large, glassy, tan in color. Remarks—Conomitra caribbeana differs from the following species by having fewer axial ribs per whorl and by lacking the vertical flammules characteristic of C. lehneri. Fossil distribution—Tuberá formation, Colombia. Recent distribution—Off Isla Margarita, Venezuela, 91 т depth. 24. Conomitra lehneri Jung, 1971 Figs. 69-71 Conomitra lehneri Jung, 1971: 200-201, pl. 14, figs. 12-16. Material examined—Eleven specimens, lengths 12 mm to 18 mm, trawled by com- mercial shrimp boats from 35 т depth in Golfo de Triste, Venezuela, March, 1979, USNM 784576. Additions to original description—Shell color white with numerous axial flammules and zigzags of tan; some specimens also en- circled with two tan bands. RELICT CAENOGASTROPODS 331 Remarks—The Recent specimens shown here closely resemble the fossil type series illustrated by Jung, having the same general body form and numerous thin axial ribs. Fossil distribution—Grand Bay formation, Carriacou, Grenadines, Lesser Antilles. Recent distribution—Golfo de Triste, Venezuela, 35 m depth. 25. Сопотйга sp. Figs. 72-74 Material examined—Length 10 mm, trawled by commercial shrimp boat in 35 m depth off Cabo La Vela, Peninsula de Guajira, Colom- bia, December, 1974, USNM 784577. Shell description—Shiny, with 5 whorls; body with numerous fine axial ribs intersected with numerous spiral ribs, giving shell pustu- lose appearance; protoconch large, bulbous, composed of 2 whorls; columella with four plications; color pale tan with 2 bands of arrow-shaped dark brown flammules; shoul- der with intermittent dark brown blotches; be- tween dark shoulder blotches are small white patches; tan base color overlaid with pattern of spiral bands of small brown dots; proto- conch brown; interior of aperture tan with two bands of dark brown. Remarks—This small Conomitra is quite unlike its sympatric congeners and represents an undescribed species. The brown and white color markings readily separate the new spe- cies from both С. caribbeana and С. lehneri. Family Volutidae Subfamily Lyriinae Pilsbry & Olsson, 1954 Genus Lyria Gray, 1847 26. Lyria cf. limata S. Hoerle & E. Vokes, 1978 Figs. 75-76 Lyria limata S. Hoerle 8 E. Vokes, 1978: 111, pl. 1, figs. 4a, 4b, 5a, 5b. Material examined—Fragment, length 72 тт, P-758 (11°42.4'N, 69°40’W), 16m depth. Additions to original description—Shell color pale yellowish-tan with three broad light brown bands, one on shoulder, one around mid-body, one at base in siphonal region; three bands darker when crossing axial costae; shoulder and mid-body bands over- laid by three dark brown continuous spiral stripes; basal band overlaid by five stripes; pale tan areas between bands with the three dark brown spiral stripes; dark spiral stripes continue as sharp barbs on margin of outer lip; inner side of lip pale orange; columella pale yellow. Remarks—Unfortunately, this giant Lyria is only known from the Recent as a single frag- ment, roughly one-third of the body whorl, a small portion of the columellar region, and a small section of the preceding whorl. The height and form of the missing spire can only be guessed at. Judging from the general shell contours of other species of Lyria, it would appear that a complete specimen of the Golfo de Triste Lyria would probably exceed 100 mm in length. Because of the incomplete condition, the specimen is referred with some reservation to Hoerle & Vokes’ taxon. The ultra-violet light photographs of the holotype of L. limata illustrated by Hoerle & Vokes (1978, pl. 1, figs. 5a, 5b) show a color pattern identical to that of the Venezuelan fragment. The contours of the ощег lip and the arrangement and coloring of the axial costae of both the fossil holotype and the Recent fragment are also identical. The main differ- ence between the fossil and Recent specimen is one of size; the entire holotype is only 38.8 mm in length while the fragment alone is 72 mm in length. Of all the Recent northern South American relicts, L. limata is the only species to have been originally described from the northern Caribbean region (Chipola formation of Flor- ida). lts disappearance from Florida Pliocene assemblages, its absence from Mio-Pliocene assemblages in the Gatunian region and its reappearance in the Recent along a small stretch of Venezuelan coastline is problemat- ical. Fossil Florida. Recent distribution—Known only from the Golfo de Triste, Venezuela. distribution—Chipola formation, Family Marginellidae Genus Persicula Schumacher, 1817 27. Persicula (Rabicea) hodsoni Weisbord, 1962 Figs. 77-78 Persicula (Rabicea) hodsoni Weisbord, 1962: 412-413, pl. 38, figs. 5-8. Persicula interruptolineata J. Gibson-Smith & W. Gibson-Smith, 1979 (non Megerle von Mühlfeld, 1816): 26. PETUCH FIGS. 77-92. 77-78. Persicula hodsoni Weisbord: USNM 784466, L = 14 mm. 79-80. Agatrix epomis (Woodring): P-750, L = 16 mm. 81. Aphera islacolonis (Maury): USNM 784467, L = 10 mm, shell sculpture enhanced by coating with magnesium oxide. 82-83. Persicula interruptolineata (Megerle von Mühlfeld): USNM 784465, L = 12 mm. 84. Persicula interruptolineata: USNM 784465, L = 11 mm. 85-86. Aphera islacolonis (Maury): USNM 784467, L = 10 mm. 87-88. Conus consobrinus Sowerby: P-708, L = 28 mm. 89-90. Conus consobrinus Sowerby: P-734, L = 33 mm. 91-92. Conus consobrinus Sowerby: P-734, L = 42 mm. RELICT CAENOGASTROPODS 333 Material examined—Length 14 mm, оп beach, Adicora, Peninsula de Paraguaná, Venezuela, April, 1975, USNM 784466. Additions to original description—Shell color cream-white overlaid with 15 dark red- brown stripes; callused outer lip and columel- lar region white; interior of aperture white; body whorl with large rectangular dark brown patch on dorsum, slightly posterior to midline. Remarks—Although considered a synonym of P. interruptolineata (Megerle von Mühlfeld, 1816) by Gibson-Smith & Gibson-Smith (1979: 26), P. hodsoni is a valid species. Both marginellids occur sympatrically along the coast of the Peninsula de Paraguaná, al- though P. interruptolineata has a more exten- sive rangethroughoutthe southern Caribbean. Persicula hodsoni can be separated from P. interruptolineata by its larger size and by its color pattern; fifteen red-brown stripes, broken a few times by vertical white bars, as opposed to the numerous rows of brown dots seen in P. interruptolineata (Figs. 82-84). This striped pattern is similar to that of P. bandera Coan 8 Roth, 1965 from the Panamic Province. Fossil distribution—Mare and Abisinia for- mations, Venezuela. Recent distribution—Along the Peninsula de Paraguaná, Venezuela. Family Cancellariidae Genus Agatrix Petit, 1967 28. Agatrix epomis (Woodring, 1928) Figs. 79-80 Tribia epomis Woodring, 1928: 223, pl. 12, fig. 10. Agatrix epomis, Petit, 1976: 38, pl. 1, fig. 3. Material examined—Length 11 mm, P-717 (11°21'N, 64°10°W), 64 т depth; 6 speci- mens, Отт —Р-718;. (11225N, 64°8.6'W), 60 т depth; lengths 7 тт and 17 mm, Р-721 (11°6.5'N, 64°22.5°W), 26 т depth; lengths 12 тт and 16 mm, P-750 (10°36.1'N, 68°12.2’W), 24 m depth. Major citations—Living specimens recog- nized, described, and illustrated by Petit, 1976. Remarks—After the discovery of living Fusiturricula jaquensis (Sowerby) (described by Altena, 1975), A. epomis was the second known living example of a supposedly-extinct Gatunian species. The R/V Pillsbury speci- mens show that this relict is a relatively com- mon member of the offshore Venezuelan mol- luscan assemblages. Fossil distribution—Bowden Jamaica. Recent distribution—Along the Colombian and Venezuelan coasts, in depths of 24- 64 m. formation, Genus Aphera H. & A. Adams, 1854 29. Aphera islacolonis (Maury, 1917) Figs. 81, 85-86 Cancellaria islacolonis Maury, 1917: 65, pl. 10, figs. 12, 12a, 12b. Cancellaria (Aphera) islacolonis, Olsson, 1922: 86, pl. 6, fig. 12. Cancellaria ellipsis Pilsbry, 1922: 333-334, DI: 22.1458: 9: Aphera islacolonis, Woodring, 1970: 344, pl. SONGS ae: Material examined—Length 10 mm, trawled by commercial shrimp boats, from 35m depth, in Golfo de Triste, Venezuela, March, 1979, USNM 784467. Additions to original description—Shell color white; dorsum with single large light tan patch; interior of aperture white; Fig. 81 shows characteristic cancellate sculpture, enhanced by coating of magnesium oxide. Remarks—The discovery of a living Aphera in the Atlantic demotes the genus from the rank of paciphile (Vermeij, 1978: 232, table 8.2)—in having a single Panamic species, A. tessellata (Sowerby, 1832), and a Caribbean species, A. islacolonis. The Recent specimen is very close to the fossils illustrated by Maury (1917, pl. 10, figs. 12a, 12b) and Olsson (1922, pl. 6, figs. 1, 2), but differs in having finer sculpture than the fossil illustrated by Woodring (1970, pl. 56, figs. 1, 2). This muta- bility of sculpture patterns is probably repre- sentative of ecophenotypic variation and not full species rank (Woodring, 1970: 344). The Golfo de Triste Aphera is identical in sculpture to A. ellipsis (Pilsbry, 1922: pl. 22, figs. 8, 9), which Pilsbry himself (p. 334) said may be only the juvenile of A. islacolonis. Fossil distribution—Cercado and Gurabo formations, Santo Domingo; Gatun formation, Costa Rica and Panama. Recent distribution—Known only from the Golfo de Triste, Venezuela, 35 m depth. 334 PETUCH Family Conidae Genus Conus Linnaeus, 1758 30. Conus consobrinus Sowerby, 1850 Figs. 87-92 Conus consobrinus Sowerby, 1850: 45. Woodring, 1928: 214-215, pl. 11, figs. 6, 7. Pflug, 1961: 62, pl. 17, figs. 1-10. Material examined—Lengths 42 mm and 33 mm, P-734 (11°1.8'N, 65°40.5’W), 65 т depth; length 28mm, Р-708 (11°24.7'N, 62°40.5'W), 70m depth; length 34 mm, P-773 (12°17'N, 72°15'W), 62 т depth. Major citations—Redescribed, with diag- nosis, by Woodring, 1928; lectotype and type series illustrated by Pflug, 1961, pl. 17, figs. 1-10. Additions to original description—Shell color pale salmon-pink with prominent wide orange band just posterior to anterior tip; body whorl with thin orange bands, varying from one to four, posterior to mid-body line; base color overlaid with numerous faint spiral rows of tiny pale orange dots; spire salmon-pink with scattered dark orange flammules; interior of aperture pale orange; juvenile specimens heavily pustulose, strongly coronated, and biconic; adult specimens smoother, more elongated, with shoulder of last whorl non- coronate. Remarks—This once-widespread Gatunian indicator species is now restricted to deeper water off the Colombian and Venezuelan coasts. Conus consobrinus is so distinctive that it cannot be confused with any other living western Atlantic cone. This relict is related to the Recent C. cedonulli-mappa-aurantius species complex of the Lesser Antilles and shallow water areas along the Colombian and Venezuelan coasts. Conus consobrinus can be separated from members of this complex by its high, heavily coronated spire and by the lack of the elaborate color patterns character- istic of the C. cedonulli group. The small specimen from P-708 (Figs. 87— 88) is similar to the fossil illustrated by Wood- ring (1928, pl. 11, fig. 7), while the large spec- imen from P-734 (Figs. 89-90) is virtually identical to the fossil illustrated by Pflug (1961, 0:17, 19:6): Fossil distribution—Agueguexquite forma- tion, Mexico; Gurabo formation, Santo Domingo; Bowden formation, Jamaica; Grand Bay formation, Carriacou; Gatun formation, Costa Rica and Panamá. Recent distribution—30-80 т depth off Venezuela and northern Colombia. 31. Conus planiliratus Sowerby, 1850 Figs. 93-95 Conus planiliratus Sowerby, 1850: 44. Olsson, 1922: 50, pl. 3, figs. 10, 13. Wood- ring, 1928: 210-212, pl. 10, figs. 7-9, pl. 11, figs <2. Material examined—Lengths 27 mm and 21mm, 25 m depth in Golfo de Triste, Vene- zuela, trawled by commercial shrimpers, March, 1979, USNM 784469. Major citations—Redescribed in detail by Woodring, 1928, with discussion of possible species complex. Additions to original descriptions—Shell color white to salmon-pink, with two bands of yellow maculations around mid-body; spire with scattered small, brown, crescent-shaped flammules; aperture white, periostracum thin, smooth, translucent yellow. Remarks—The four known Recent speci- mens are indistinguishable from the fossil specimens illustrated by Olsson (1922, pl. 3, fig. 10) and Woodring (1928, pl. 10, figs. 7, 9). The only Recent cone that bears any resem- blance to C. planiliratus is C. stimpsoni Dall, 1902, from deep water off Florida, Georgia, the Carolinas, and in the Gulf of Mexico. Conus planiliratus differs from C. stimpsoni by being a consistently more slender shell, by having two bands of yellow maculations around the mid-body, by having a heavily sculptured spire, and by having numerous in- cised spiral sulci on the body whorl. Woodring (1970: 346) was correct in his prediction that C. planiliratus could still be living in the Atlan- tic. Fossil distribution—Bowden formation, Jamaica; Gurabo formation, Santo Domingo; Gatun and Limón formations, Costa Rica. Recent distribution—In the Golfo de Triste, Venezuela, 35 m depth. 32. Conus symmetricus Sowerby, 1850 Figs. 96-97 Conus symmetricus Sowerby, 1850: 44, pl. 9, fig. 1. Maury, 1917: 200, pl. 7, fig. 7. Wood- ring, 1928: 204. Pflug, 1961: 63-64, pl. 18, figs. 1-11. Woodring, 1970: 35-354. Material examined—Length 37 mm, trawled by commercial shrimp boats, 35 m depth, in ВЕНСТ CAENOGASTROPODS FIGS. 93-112. 93-94. Conus planiliratus Sowerby: USNM 784469, L = 21 mm. 95. Conus planiliratus Sowerby: USNM 784469, L = 27 mm. 96-97. Conus symmetricus Sowerby: USNM 784470, L = 37 mm. 98. Strioterebrum bowdenensis (Woodring): USNM 784578, L = 20 mm. 99-100. Strioterebrum gatunense kugleri (Rutsch): USNM 784471, L = 33 mm. 101-102. Strioterebrum ischna (Woodring): USNM 784579, L = 11 mm. 103-104. Strioterebrum quadrispiralis (Weisbord): USNM 784472, L = 13 mm. 105- 106. Strioterebrum trispiralis (Weisbord): USNM 784473, L = 14 тт. 107-108. Polystira barretti (Guppy): USNM 784477, Е = 63 mm. 109-110. Agladrillia lassula Jung: USNM 784474, L = 25 mm. 111-112. Hindsiclava consors (Sowerby): USNM 784476, L = 38 mm. 336 PETUCH Golfo de Triste, Venezuela, March, 1970, USNM 784470. Major citations—Lectotype and representa- tive series illustrated by Pflug, 1961, pl. 18, figs. 4, 8, 11. Additions to original description—Body whorl sculptured with 18 prominent, raised, pustulated spiral cords; spire sculpture with six incised spiral sulci; shell color pure white with small scattered pale orange-brown flam- mules; spire white with regularly-spaced, in- termittent brown flammules; protoconch and early whorls pale orange; aperture white. Remarks—The wide-shouldered and flat- spired aspects of C. symmetricus are unlike those of any other living cone. The Recent specimen shown here easily fits into the series illustrated by Pflug (1961, pl. 18, figs. 1-11), especially so in having spiral rows of raised pustules on the body whorl and in hav- ing a characteristically sculptured spire like that of Pflug's fig. 5 and as seen here in Fig. 96. One of the specimens of Pflug's series (fig. 6) is nearly identical to the Recent Vene- zuelan specimen. Fossil distribution—Bowden formation, Jamaica; Gurabo formation, Santo Domingo; Gatun formation, Costa Rica and Panama. Family Terebridae Genus Strioterebrum Sacco, 1891 33. Strioterebrum bowdenensis (Woodring, 1928) Fig. 98 Terebra (Strioterebrum) bowdenensis Wood- ring, 1928: 138-139, pl. 3, figs. 3-8. Material examined—Two specimens, lengths 20 mm and 22 mm, trawled by com- mercial shrimp boats from 35m depth in Golfo de Triste, Venezuela, December, 1978, USNM 784578. Additions to original description—Shell pure white. Remarks—The two Recent specimens are indistinguishable from the type-series Шиз- trated by Woodring, both in size and sculptur- ing. Fossil Jamaica. Recent distribution—In Golfo de Triste, Venezuela, 35 m depth. distribution—Bowden formation, 34. Strioterebrum gatunensis kugleri (Rutsch, 1934) Figs. 99-100 Terebra (Strioterebrum) gatunensis kugleri Rutsch, 1934: 106-108, pl. 8, fig. 18, pl. 9, figs. 12, 13. Weisbord, 1962: 428-430, pl. 40, figs. 12, 13, pl. 45, figs. 24, 25. Material examined—Lengths 33 mm and 32 mm, on beach, Crespo, Cartagena, Co- lombia, after storm, December, 1974, USNM 784471; length 24 mm, same locality and data, UMML 8282. Major citations—Redescribed in detail, with diagnosis, by Weisbord, 1962. Additions to original description—Shell color deep gray-brown with alternating flam- mules of dark brown; beaded junctions of axial cords and spiral ridges light tan; lower part of body whorl with white band; base of shell dark chocolate brown; subsutural collar white with alternating brown patches corre- sponding to brown flammules on body whorl; interior of aperture dark brown. Remarks—Besides lack of color, the fossil specimen of $. gatunensis kugleri illustrated by Weisbord (1962, pl. 40, figs. 12, 13) is almost identical to the Recent specimen illustrated here. The color pattern of brown flammules and checkers and the distinctive sculpturing of raised beads readily separates S. gatunen- sis kugleri from any other known Recent At- lantic Strioterebrum. The relict species 15 closest to $. spiriferum (Dall, 1895) from the Gurabo formation, Santo Domingo, and also $. glaucum (Hinds, 1844) from the Panamic Province. Strioterebrum gatunensis kugleri is well-represented in the fossil record of Venezuela. Fossil distribution—Punta Gavilán, Mare, and Cabo Blanco formations, Venezuela. Recent distribution—Known only from the Colombian coast near Cartagena but proba- bly occurs elsewhere along the Colombian and Venezuelan coasts. 35. Strioterebrum ischna (Woodring, 1928) Figs. 101-102 Terebra (Strioterebrum) ischna Woodring, 1928: 142, pl. 3, fig. 18, pl. 4, fig. 1. Material examined—Five specimens, lengths 6-11 mm, on beach, Adicora, Penin- RELICT CAENOGASTROPODS 337 sula de Paraguaná, Estado Falcón, Venezu- ela, December, 1974, USNM 784579. Addition to original description—Shell color uniformly pale tan. Remarks—The Recent specimens are identical to the fossil type-specimens illustrat- ed by Woodring. Fossil distribution—Bowden Jamaica. Recent distribution—North end of Peninsu- la de Paraguana, Venezuela. formation, 36. Strioterebrum quadrispiralis (Weisbord, 1962) Figs. 103-104 Terebra (Strioterebrum) quadrispiralis Weis- Бога, 1962: 431—432, pl. 41, figs. 14. Material examined—Three specimens, lengths 11-13 mm, on beach, Adicora, Реп- insula de Paraguaná, Venezuela, April, 1975, USNM 784472; length 11 mm, on beach, Punta Mangle, Isla Margarita, Venezuela, 1977, UMML 8282 (from Gibson-Smith collec- tion). Additions to original description—Shell color pale rose-white with darker band along suture; base of shell dark reddish-brown; in- terior of aperture white, dark reddish-brown in siphonal region; protoconch white. Remarks—Along with the following spe- cies, this small terebrid resembles no other living Atlantic species. Strioterebrum quad- rispiralis and S. trispiralis represent the last of a long lineage of small, beaded terebrids cen- tered around the Middle Miocene $5. eleutheria (Woodring, 1928) and $. midiensis (Olsson, 1922). This and the following species may be popu- lation variants of an undescribed Bowden species (Woodring, 1928: pl. 3, figs. 13, 14). As such, they would represent true relict spe- cies. If they are distinct species that have long been endemic to the Venezuelan coast, how- ever, they may only represent old, unchanged species inhabiting their original range and would not be considered true relicts. In either case, the existence of these two terebrids re- inforces the archaic nature of the relict pocket. Fossil distribution—Mare formation, Vene- zuela. Recent distribution—From the Peninsula de Paraguaná to Isla Margarita, Venezuela, in shallow water. 37. Strioterebrum trispiralis (Weisbord, 1962) Figs. 105-106 Terebra (Strioterebrum) trispiralis Weisbord, 1962: 430-431, pl. 40, figs. 14, 15. Material examined—Lengths 14 mm and 13 тт on beach, Adicora, Peninsula de Paraguaná, Venezuela, April, 1975, USNM 784473; 3 specimens, lengths 11-14 mm, on beach Punta Mangle, Isla Margarita, Venezu- ela, 1977, UMML 8283 (from Gibson-Smith collection). Additions to original description—Shell color gray-brown, darker along suture; base dark purple-brown; interior of aperture white, purple in siphonal region. Remarks—Strioterebrum trispiralis is closely related to the preceding species, and pending anatomical studies, may prove to be conspecific. The main difference between the two species is seen in the structure and form of the varices and varical nodes. In S. tri- spiralis, the varices are complete, forming costae that are intersected by two spiral sulci, giving the shell the characteristic tripartite form. In S. quadrispiralis, the varices are in- tersected by three sulci, giving the effect of four rows of raised beads. Fossil distribution—Mare formation, Vene- zuela. Recent distribution—From the Peninsula de Paraguana to Isla Margarita, Venezuela, in shallow water. Family Turridae Subfamily Turrinae Swainson, 1875 Genus Polystira Woodring, 1928 38. Polystira barretti (Guppy, 1866) Figs. 107-108 Pleurotoma barretti Guppy, 1866: 290, pl. 17, fig. 6. Turris albida barretti, Maury, 1917: 214, pl. 8, fig: 5: Polystira barretti, Woodring, 1928: 146, pl. 4, fig. 6. Pflug, 1961: 70-71, pl. 20, figs. 1, 4. Material examined—Length 63 mm, trawled by commercial shrimp boat, 35 т depth, in Golfo de Triste, Venezuela, March, 1979, USNM 784477; length 72mm, trawled by 338 PETUCH commercial shrimp boat, 35m depth, off Cabo La Vela, Peninsula de Guajira, Colom- bia, December, 1974 (with Crucibulum marense attached), USNM 784478; lengths 65 mm and 61 тт, P-712 (11°8’N, 63°18’W), 25 m depth. Major citations—Redescribed in detail by Woodring, 1928; holotype illustrated by Pflug, 1961, pl. 20, figs. 1, 4). Addition to original description—Shell pure white; periostracum thin, gray-green. Remarks—The well-developed shoulder carina, seen in both fossil and Recent speci- mens, sets P. barretti aside from all other known Atlantic Polystira species. The Recent specimen illustrated here is similar to the illus- tration of the holotype in Pflug (1961). Fossil distribution—Gurabo formation, Santo Domingo; Bowden formation, Jamaica. Recent distribution—Off the Colombian and Venezuelan coasts, 20-40 m depth. Subfamily Clavinae Powell, 1942 Genus Agladrillia Woodring, 1928 39. Agladrillia lassula Jung, 1969 Figs. 109-110 Agladrillia lassula Jung, 1969: 550-551, pl. 59, figs. 1-3. Material examined—Length 25 mm, trawled by commercial shrimp boat, from 35 m depth, in Golfo de Triste, Venezuela, March, 1979, USNM 784474. Additions to original description—Shell color pale tan with white axial costae; outer lip white; interior of aperture tan; early whorls pinkish-tan. Remarks—The Recent specimen of A. lassula from the Golfo de Triste 1$ very close to the fossil illustrated by Jung (1969, pl. 59, figs. 2, 3). The conspicuous lateral hump, which marks the termination of the axial costae and the beginning of the smooth dor- sum, is a specific character seen in both the fossil and Recent specimens. There are no known Recent species of Agladrillia that bear any resemblance to this rather aberrant turrid, and this 1$ the first known Atlantic species of the formerly paciphilic genus. Fossil distribution—Melajo Clay Member, Springvale formation, Trinidad. Recent distribution—Known only from the Golfo de Triste, Venezuela. Genus Hindsiclava Hertlein & Strong, 1955 40. Hindsiclava consors (Sowerby, 1850) Figs. 111-112 Pleurotoma consors Sowerby, 1850: 50. Turris (Crassispira) consors, Rutsch, 1934: 99, pl. 8, figs. 13-16. Crassispira consors, Pflug, 1961: 67, pl. 19, figs. 4, 7, 10. Jung, 1965: 565, pl. 76, figs. 14, 15 Crassispira (Hindsiclava) consors consors, Woodring, 1970: 378-380, pl. 58, figs. 1, 22. Material examined—Length 38 mm, trawled by commercial shrimp boat, 35 m depth, in Golfo de Triste, Venezuela, March, 1979, USNM 784476. Major citations—Lectotype illustrated by Pflug, 1961; detailed redescription and diag- nosis by Woodring, 1970. Additions to original description—Shell color pale yellow; subsutural band white. Remarks—The Recent specimen illustrat- ed here 15 nearly identical to the fossil lecto- type illustrated by Pflug (1961, pl. 19, figs. 4, 10). The only other southern Caribbean Hindsiclava species that could be confused with H. consors is H. chazaliei (Dautzenberg, 1900) (Fig. 123) from off Surinam, Venezuela, and Colombia. The sympatric H. chazaliei dif- fers from H. consors by having a lower spire and raised axial costae, and being dark brown in color. Fossil distribution—Gurabo formation, Santo Domingo; Bowden formation, Jamaica; Springvale formation, Trinidad; Punta Gavilan formation, Venezuela; Limon formation, Costa Rica; Gatun formation, Costa Rica and Panama. Recent distribution—Known only from the Golfo de Triste, Venezuela, 35 m depth. Subfamily Turriculinae Powell, 1942 Genus Fusiturricula Woodring, 1928 41. Fusiturricula acra (Woodring, 1970) Figs. 113-114 Pleurofusia acra Woodring, 1970: 367-368, РЁ! 57, tige: Material examined—Length 34 mm, P-727 (10°20’М, 65°2’W), 64 т depth. RELICT CAENOGASTROPODS FIGS. 113-126. 113-114. Fusiturricula acra (Woodring): P-727, L humerosa (Gabb): P-737, L = 19 mm. 117-118. Fusiturricula ¡ole Woodring: USNM 784580, L = 17 mm. 119-120. Fusiturricula jaquensis (Sowerby): USNM 784475, L = 46 mm. 121-122. Paraborsonia varicosa = 34mm. 115-116. Fusiturricula (Sowerby): USNM 784581, L = 16 mm. 123. Hindsiclava chazaliei (Dautzenberg): P-712,L = 33 тт. 124-126. Paraborsonia varicosa (Sowerby): USNM 784581, L = 15 mm. 340 РЕТУСН Additions to original description—Shell color tan with numerous thin brown vertical flammules; subsutural nodes white, separat- ed by brown patches. Remarks—Fusiturricula acra can be sepa- rated from the other three living northern South American Fusiturricula species by hav- ing white subsutural shoulder nodes and noded spiral threads on the lower part of the body whorl. This last character was used by Woodring (1970: 368) to separate F. acra from its fossil congeners. Fossil distribution—Gatun Panamá. Recent distribution—Near Isla Margarita, Venezuela, 64 m depth. formation, 42. Fusiturricula humerosa (Gabb, 1873) Figs. 115-116 Turris (Surcula) humerosa Gabb, 1873: 208. Surcula humerosa, Pilsbry, 1922: 315-316, pl. 17, figs. 4, 5. Material examined—Two specimens, lengths 15 mm and 19 mm, Р-737 (10°44’N, 66°7'W), 65 т depth. Major citations—Redescribed and illustrat- ed with diagnosis by Pilsbry, 1922. Addition to original description—Shell color pale tan with alternating lavender purple vertical flammules; raised spiral cords рае yellow; siphonal canal orange; interior of aperture tan; protoconch and early whorls orange. Remarks—Although similar in size to the following species, Fusiturricula humerosa dif- fers from F. iole in being a more colorful shell with a pattern of vertical purple flammules and by having a large, angled axial swelling on the dorsum of the last whorl. The specimens il- lustrated by Pilsbry closely resemble the specimen shown here. Fossil distribution—Gurabo Santo Domingo. Recent distribution—Off Cabo Cordera, Venezuela, 65 m depth. formation, 43. Fusiturricula ¡ole Woodring, 1928 Figs. 117-118 Fusiturricula ¡ole Woodring, 1928: 167, pl. 6, fig. 4. Material examined—Two specimens, lengths 17 mm and 20 mm, trawled by com- mercial shrimp boats from 35m depth in Golfo de Triste, Venezuela, March, 1979, USNM 784580. Addition to original description—Shell pure white. Remarks—Fusiturricula ¡ole can be sepa- rated from both F. acra and F. humerosa by its smaller size, more angled shoulder, sharp shoulder coronations, and pure white color. The Recent specimen illustrated here is very close to the fossil holotype illustrated by Woodring. Fossil distribution—Bowden formation, Jamaica. Recent distribution—Golfo de Triste, Venezuela, 35 m depth. 44. Fusiturricula jaquensis (Sowerby, 1850) Figs. 119-120 Pleurotoma jaquensis Sowerby, 1850: 51. Knefastia jaquensis, Woodring, 1928: 167. Fusiturricula jaquensis, Jung, 1965: 568, pl. 77, fig. 5. Abbott, 1974: 264, no. 2918. Altena, 1975: 62-63, pl. 4, figs. 8, 9. Knefastia paulettae Princz, 1980: 71, fig. 1. Material examined—Length 46 тт, trawled by commercial shrimp boats, 35 m depth, in Golfo de Triste, Venezuela, March, 1979, USNM 784475. Major citations—Living specimens de- scribed by Altena, 1975 and Princz, 1980 (as Knefastia paulettae). Remarks—Fusiturricula jaquensis was the first supposedly extinct Gatunian species to be recognized as part of the Recent mol- luscan fauna of northern South America. The specimen illustrated here is nearly identical to those illustrated by Altena (living) and Jung (fossil). Fossil distribution—Bowden formation, Jamaica; Gurabo formation, Santo Domingo; Cantaure and Punta Gavilan formations, Venezuela. Recent distribution—Surinam to Golfo de Triste, Venezuela, at depths of 35-100 m. Subfamily Borsoninae Bellardi, 1875 Genus Paraborsonia Pilsbry, 1922 Like Panamurex and Conomitra, this endemic American genus was presumed to have died out at the end of the Miocene (Woodring, 1970: 373). The discovery of this distinctive relict genus further reinforces the archaic nature of the upwelling faunal pocket. ВЕНСТ CAENOGASTROPODS 341 45. Paraborsonia varicosa (Sowerby, 1850) Figs. 121-122, 124-126 Mitra varicosa Sowerby, 1850: 46. Cordiera varicosa, Gabb, 1873: 270. Borsonia (Paraborsonia) varicosa, Pilsbry, 1922: 325-326, pl. 17, figs. 19-21. Materials examined—Five specimens, lengths 14 mm to 16 mm, trawled from 35 m depth in Golfo de Triste by commercial shrimp boats, March, 1979, USNM 784581. Major citations—Diagnosis and illustra- tions, especially of protoconch, by Pilsbry, 1922. Additions to original description—Shell color pale yellow with darker yellow crescent- shaped flammules along shoulder; anterior tip of columella dark yellow; interior of aperture white. Remarks—The Recent specimens of this relict genus closely resemble the specimen of Paraborsonia varicosa illustrated by Pilsbry. Paraborsonia cantaurana Jung, 1965 from the Cantaure formation of Venezuela 1$ simi- lar but differs in having a much higher spire, approaching the genus Scobinella. Parabor- sonia laeta Jung, 1971 from the Grand Bay formation of Carriacou is also similar to the relict but has a much more developed shoulder carina. Fossil distribution—Bowden formation, Jamaica; Grand Bay formation, Carriacou; Gurabo formation, Santo Domingo. Recent distribution—Golfo de Triste, Venezuela, 35 m depth. ACKNOWLEDGEMENTS To my former major professor, Dr. Gilbert L. Voss of the Rosenstiel School of Marine and Atmospheric Sciences, University of Miami, | extend my most heartfelt thanks for the input into and support of my research. For critical review of the manuscript, | would like to thank the following: Dr. Donald R. Moore, Dr. Lowell P. Thomas, Dr. Peter L. Lutz, and Mr. Robert C. Work, also of the Rosenstiel School; and Dr. Richard S. Houbrick, Department of Mol- lusks, National Museum of Natural History, Smithsonian Institution. To Dr. Geerat J. Vermeij, Department of Zoology, University of Maryland, | give special thanks for the sharing of insights into the evolutionary history of the western Atlantic and for helping to crystallize my concept of Caribbean zoogeograpy. À auf IN УИ O Dg SHA Е FIGS. 127-129. Living animals of the extant South- ern Caribbean Siphocypraea spp., drawn from life: 127. Siphocypraea mus (Linnaeus). 128. Siphocypraea donmoorei Petuch. 129. Sipho- cypraea henekeni (Sowerby). My deepest gratitude is extended to Dr. Jack and Mrs. Winifred Gibson-Smith, Caracas, Venezuela, who not only nursed me through a bout of Dengue Fever during a re- search trip to Venezuela in 1979, but also generously shared their extensive knowledge of the fossil record of Venezuelan mollusks and graciously donated much valuable study material. Thanks also go to Dr. Pablo 342 Caribbean PETUCH Sea FIG. 130. Map of Southern Caribbean, showing distribution of upwelling system along northern Colombian and Venezuelan coasts. Actual upwelling areas denoted by heavy stipling; areas influenced by upwellings denoted by light stipling. After Meyer (1977, fig. 1). Penchaszadeh, Universidad Simón Bolivar, Caracas, Venezuela, for his donation of re- search material. For loan of seemingly unob- tainable but much-needed literature, | extend my sincerest thanks to Dr. William K. Emer- son, Department of Invertebrates, American Museum of Natural History, New York. Special thanks also go to Mr. Gonzalo Cruzat of Miami, Florida, Mr. Pat 1220, Uni- versity of Maryland, and Mrs. Sally D. Kaicher of St. Petersburg, Florida, for the excellent photographs. Finally, my deepest appreciation goes to Mr. M. G. Harasewych, College of Marine Studies, University of Delaware, for his help in collecting specimens of the relicts in Venezu- ela in 1979, and for sharing my enthusiasm for the unexplored southern Caribbean biota. Most of this work was done at the Rosen- stiel School of Marine and Atmospheric Sci- ences, University of Miami, as partial fulfill- ment of the degree of Doctor of Philosophy in Marine Biology. LITERATURE CITED ABBOTT, R. T., 1974, American seashells. Ed. 2. Van Nostrand Reinhold, New York, p. 264, no. 2918. ALTENA, С. O. VAN R., 1975, The marine Mollusca of Suriname (Dutch Guiana), Holocene and Re- cent, Part Ill, Gastropoda and Cephalopoda. Zoologische Verhandelingen, Leiden, 139: 104 p., 11 pl. BAYER, Е. M., 1971, New and unusual mollusks collected by the R/V JOHN ELLIOTT PILLS- BURY and R/V GERDA in the tropical western Atlantic. Bulletin of Marine Science, 21: 111- 236. BROWN, A. P. & PILSBRY, Н. A., 1911, Fauna of the Gatun Formation, Isthmus of Panama. Proceedings of the Academy of Natural Sci- ences of Philadlephia, 64: 336-373, pl. 22-29. BULLIS, H. R., Jr., 1964, Muricidae (Gastropoda) from the northeast coast of South America, with descriptions of four new species. Tulane Studies in Zoology, 11: 99-107, 1 pl. CLENCH, W. J., 1942, The genus Conus in the western Atlantic. Johnsonia, 1(6): 1—0, 15 pl. CLENCH, W. J., 1944, The genera Casmaria, Galeodea, Phalium, and Cassis in the western Atlantic. Johnsonia, 1(16): 1-16, 8 pl. CLENCH, W. J., 1947, The genera Purpura and Thais in the western Atlantic. Johnsonia, 2: 61- 92, 8 pl. CLENCH, W. J., 1953, The genus Conus in the western Atlantic. Johnsonia, 2: 363-376, 5 pl. CLENCH, W. J., 1959, The genus Murex in the western Atlantic. Johnsonia, 3: 331-334, 2 pl. CLENCH, W. J. & PEREZ FARFANTE, I. 1945, The genus Murex in the western Atlantic. Johnsonia, 1(17): 1-58, 29 pl. COSEL, В. VON 1976, Contribución al сопо- cimiento del genero Voluta Linné, 1758 (Proso- RELICT CAENOGASTROPODS 343 branchia) en la costa del Caribe de Colombia. Instituto Colombo-Alemán Investigaciones Cientificas, 8: 83-104, 5 pl. DALL, W. H., 1889, A preliminary catalogue of the shell-bearing marine mollusks and brachiopods of the southeastern coast of the United States, with illustrations of many of the species. Bulletin of the United States National Museum, 37: 221 p., 74 pl. DAUTZENBERG, P., 1900, Croisieres du yacht “Chazalie” dans l'Atlantique. Mollusques. Mémoire de la Société Zoologique de France, 13: 145-265. FLORES, C., 1966, La familia Cassidae (Mollusca: Mesogastropoda) en las costas nororientales de Venezuela. Boletín del Instituto de Oceano- graphia de la Universidad del Oriente, 5(1-2): 7-37. FLORES, C., 1973a, La familia Littorinidae (Mol- lusca: Mesogastropoda) en las aguas costeras de Venezuela. Boletín del Instituto de Oceano- graphia de la Universidad del Oriente, 12(1): 3- 223" pl: FLORES, C., 1973b, Notas sobre la distribucion horizontal y vertical de los Littorinidae (Mollusca: Gastropoda) en las aguas costeras de Venezu- ela. Boletín del Instituto Oceanographia de la Universidad del Oriente, 12(1): 67-74. GABB, W., 1860, Description of some new Tertiary fossils from Chiriqui, Central America. Proceed- ings of the Academy of Natural Sciences of Philadelphia, 12: 567-569. САВВ, W., 1873, Description of some new genera of Mollusca. Proceedings of the Academy of Natural Sciences of Philadelphia, 24: 270-274, pi. 9-11. GABB, W., 1875, Notes on West Indian fossils. Geological Magazine, 2: 544-545. GABB, W., 1881, Descriptions of Caribbean Mio- cene fossils. Journal of the Academy of Natural Sciences of Philadlephia, 8: 337-348. GARDNER, J., 1937, The Molluscan fauna of the Alum Bluff Group of Florida. Part VI. Pteropoda, Opisthobranchia, and Ctenobranchia (in part). United States Department of the Interior Geo- logical Survey Professional Paper 142-F: 251- 435, pl. 37-48. GERTMAN, R. L., 1969, Cenozoic Typhinae (Mol- lusca: Gastropoda) of the western Atlantic re- gion. Tulane Studies Geology and Paleontology, 7: 143-191, 8 pl. GIBSON-SMITH, J., 1973, The genus Voluta (Mol- lusca: Gastropoda) in Venezuela, with descrip- tion of two new species. Geos, 20: 65-73, 3 pl. GIBSON-SMITH, J., 1976, A new taxon in the genus Charonia (Mollusca: Gastropoda) from the Pliocene of Venezuela. Asociación Venezo- lano de Geología, Minas, y Petroleo, Boletín In- formativo, 19(1): 1-17, 1 pl. GIBSON-SMITH, J. 8 GIBSON-SMITH, W., 1974, The genus Strombina (Mollusca: Gastropoda) in Venezuela, with description of a new Recent and some fossil species. Asociacion Venezolano de Geología, Minas, y Petroleo, Boletín Informativo, 17(4-6): 49-70, 4 pl. GIBSON-SMITH, J. & GIBSON-SMITH, W., 1979, The genus Arcinella (Mollusca: Bivalvia) in Venezuela and some associated faunas. Geos, 24: 11-32. GONZALEZ, А. В. & FLORES, C., 1972, Nota sobre los generos Thais Róding, Purpura Bruguière, y Murex Linnaeus (Neogastropoda: Muricidae) en las aguas costeras de Venezuela. Boletín del Instituto de Oceanographia de la Universidad del Oriente, 11(2): 67-82, 11 figs. GUPPY, R. J. L., 1866, On the Tertiary Mollusca of Jamaica. Quarterly Journal of the Geological Society of London, 22: 281-295, pl. 16-18. GUPPY, В. J. L., 1873, On some new Tertiary fos- sils from Jamaica. Proceedings of the Scientific Association of Trinidad, 2: 72-88. Reprint in Bulletins of American Paleontology, 8(35): 204— 220. GUPPY, R. J. L., 1909, On a collection of fossils from Springvale near Couva, Trinidad. Agricul- tural Society of Trinidad and Tobago Scientific Paper, 440: 1-55. Reprint in Bulletins of Ameri- can Paleontology, 8(35): 292-305. GUPPY, В. J. L., 1911, Fossils from Springvale near Couva, Trinidad, Second report. Agricul- tural Society of Trinidad and Tobago Paper, 454: 1-15. Reprint in Bulletins of American Paleontology, 8(35): 306-314. HODSON, F., 1926, Venezuelan and Caribbean Turritellas; with a list of Venezuelan type strati- graphic localities. Bulletins of American Paleon- tology, 11(45): 31, pl. 21, figs. 2, 7. HODSON, F. & HODSON, H. K., 1931, Some Venezuelan mollusks, Part 1. Bulletins of Amer- ican Paleontology, 16: 1-94, 24 pl. HOERLE, 5. Е & VOKES, Е. H., 1978, A review of the volutid genera Lyria and Falsilyria (Mollusca: Gastropoda) in the Tertiary of the western Atlan- tic. Tulane Studies in Geology and Paleontol- ogy, 14: 105-130. INGRAM, W. M., 1939, New fossil Cypraeidae from the Miocene of the Dominican Republic. Bul- letins of American Paleontology, 24: 327-340, pl. 22. INGRAM, W. M., 1940, Two new Cypraeas from Costa Rica. Journal of Paleontology, 14: 505- 506. INGRAM, W. M., 1947a, New fossil Cypraeidae from Venezuela and Colombia. Bulletins of American Paleontology, 31: 1-12, 2 pl. INGRAM, W. M., 1947b, New fossil Cypraeidae from the Miocene of Florida and Colombia. Proceedings of the California Academy of Sci- ences, ser. 4, 26: 125-133, pl. 2, figs. 1-2. JUNG, P., 1965, Miocene Mollusca from the Para- guaná Peninsula, Venezuela. Bulletins of Amer- ican Paleontology, 49: 384-652, pl. 50-79. JUNG, P., 1969, Miocene and Pliocene mollusks from Trinidad. Bulletins of American Paleontol- ogy, 55: 289-657, pl. 13-60. JUNG, P., 1971, Fossil mollusks from Carriacou, 344 West Indies. Bulletins of American Paleontol- ogy, 61: 1-262, 21 pl. KAUFMANN, В. & GOTTING, K.-J., 1970, Proso- branchia aus dem Litoral der Karibischen Kúste Kolumbiens. Helgoländer wissenschaftliche Meeresuntersuchungen, 21: 333-398, 148 figs. MARKS, J. G., 1951, Miocene stratigraphy and paleontology of southwestern Ecuador. Bulletins of American Paleontology, 33: 142, pl. 9, figs. 10,11 MAURY, С. J., 1912. A contribution to the paleon- tology of Trinidad. Proceedings of the Academy of Natural Sciences of Philadelphia, 64: 132- 134. MAURY, C. J., 1917, Santo Domingo type sections and fossils. Bulletins of American Paleontology 5(29): part 1, Mollusca: 165-415, pl. 29-65. MAURY, С. J., 1925, A further contribution to the paleontology of Trinidad. (Miocene Horizons). Bulletins of American Paleontology 10: 159- 402. MEYER, G. B., 1977, Algas marinas bénticas indi- cadoras de un area afectada por aguas de surgencia frente a la costa Caribe de Colombia. Anales del Instituto Marino de Punta Betin 9:45- ИИ Ща OLSSON, А. A., 1922, The Мюсепе of northern Costa Rica with notes on its general stratigraphic relations Part 1. Bulletins of American Paleontol- оду, (6): 42-55, 3 pl. OLSSON, A. A., 1942, Tertiary and Quaternary fossils from the Burica Peninsula of Panamá and Costa Rica. Bulletins of American Paleontology, 27: 1-106, 12 pl. OLSSON, A. A., 1964, Neogene mollusks from northwestern Ecuador. Paleontological Re- search Institute, Ithaca, New York, р. 1-256, 38 pl. OLSSON, A. A., 1965, A review of the genus Voluta and the description of a new species. Bulletins of American Paleontology, 49: 655-671, pl. 80-83. PERRILLIAT-MONTOYA, M. C., 1963, Moluscos de la Formación Agueguexquite (Mioceno Medio) del Istmo de Tehuantepec, Mexico. Paleontología Mexicana, 14: 1-45, 6 pl. PETIT, R. E., 1976, Notes on Cancellariidae (Mollusca: Сазгорода)—!. Tulane Studies т Geology and Paleontology, 12: 33-43, 2 pl. PETUCH, E. J., 1976, An unusual molluscan as- semblage from Venezuela. Veliger, 18: 322-325, 1 pl. PETUCH, E. J., 1979, A new species of Sipho- cypraea (Gastropoda: Cypraeidae) from north- ern South America with notes on the genus in the Caribbean. Bulletin of Marine Science, 29: 216- 225. PFLUG, H. D., 1961, Mollusken aus dem Tertiar von Santo Domingo. Acta Humboldtiana. Series Geologica et Palaeontologica No. 1. Steiner Verlag, Wiesbaden, p. 1-107, 26 pl. PILSBRY, H. A., 1922, Revision of W. М. Gabb's Tertiary Mollusca of Santo Domingo. Proceed- ings of the Academy of Natural Sciences of Philadelphia, 73: 305-435, pl. 16—47. PETUCH PRINCZ, D., 1973, Moluscos gastrópodos y pelecipodos del Estado Nueva Esparta, Vene- zuela. Memoria de la Sociedad de Ciencias Naturales La Salle, 33: 169-222, 4 pl. PRINCZ, D., 1980, Los moluscos marinos del Golfo de Venezuela. Fundación La Salle de Ciencias Naturales Contribución, 78: 51-75. RADWIN, С. E. 8 D'ATTILIO, A., 1976, Murex shells of the world; an illustrated guide to the Muricidae. Stanford University Press, Stanford, California, xi + 284 p. REHDER, H. A., 1962, Contribución al conocimi- ento de los moluscos marinos del Archipelago de los Roques y La Orchila. Memoria de la Sociedad de Ciencias Naturales La Salle, 22: 116-138. RUTSCH, R., 1934, Die Gastropoden aus dem Neogen der Punta Gavilán in Nord-Venezuela. Mémoire de la Société Paleontologique Suisse, 54-55: 1-169, 9 pl. RUTSCH, R., 1942, Die mollusken der Springvale- Schichten (Obermiocaen) von Trinidad (British West-Indien). Naturforschenden Gesellschaft Basel Verhandlunger, 54: 96-182, pl. 3-9. SCHILDER, F. A., 1939, Cypraeacea aus dem Tertiär von Trinidad, Venezuela, und den Antillen. Schweizer Paleontologische Gesellschaft Abhandlungen, 62: 1-35. SCHUCHERT, C., 1935, Historical Geology of the Antillean-Caribbean Region. Wiley, New York, р. 230-231, 252-253, 420-455, 570-591. SMITH, B., 1907, A contribution to the morphology of Pyrula. Proceedings of the Academy of Natural Sciences of Philadelphia, 59: 208-219, pl. 17, figs. 1-2. SOWERBY, С. B., 1, 1850, Description of new spe- cies of fossil shells found by J. S. Heniker [sic], Esq. Quarterly Journal of the Geological Society of London, 6: 44-53, pl. 9-10. TELLO, J., 1975, Catálogo de la fauna Venezolana, vol. 8, Mollusca. Sociedad Venezolana Ciencias Naturales, p. 33-163. VERMEIJ, G., 1978, Biogeography and adapta- tion. Harvard, Cambridge, Massachusetts, р. 1- 5, 216-237. VINK, D. L., 1977, The Conus cedonulli complex. Zoologische Mededeelingen, Leiden, 51: 79- 93, 4 pl. VOKES, E. H., 1967a, Observations on Murex messorius and Murex tryoni, with the description of two new species of Murex. Tulane Studies in Geology, 5: 81-90, 4 pl. VOKES, E. H., 1967b, Cenozoic Muricidae of the western Atlantic region. Рай Ill—Chicoreus (Phyllonotus). Tulane Studies in Geology, 5: 133-166, 6 pl. VOKES, E. H., 1968, Cenozoic Muricidae of the western Atlantic region. Part IV—Hexaplex and Murexiella. Tulane Studies in Geology, 6: 85- 126, 8 pl. VOKES, Е. H., 1970, Cenozoic Muricidae of the western Atlantic region. Part V—Pterynotus and Ротепа. Tulane Studies т Geology and Paleontology, 8: 1-50, 7 pl. RELICT CAENOGASTROPODS 345 VOKES, E. H., 1971, Catalogue of the genus Murex Linné (Mollusca: Gastropoda): Muricinae, Ocenebrinae. Bulletins of American Paleontol- ogy, 61: 5-141. VOKES, E. H., 1974, Notes on Chicoreus (Mol- lusca: Gastropoda) from the Cenozoic of the western Atlantic region, with the description of a new species. Tulane Studies in Geology and Paleontology, 11: 81-95, 3 pl. VOKES, E. H., 1975, Cenozoic Muricidae of the western Atlantic region. Part Vi—Aspella and Dermomurex. Tulane Studies in Geology and Paleontology, 11: 121-162, 6 pl. WEISBORD, N. E., 1929, Miocene Mollusca of northern Colombia. Bulletins of American Paleontology, 14: 1-57, 9 pl. WEISBORD, N. E., 1929, Miocene Mollusca of northern Colombia. Bulletins of American Paleontology, 42: 1-672, 48 pl. WOODRING, W. P., 1928, Miocene mollusks from Bowden, Jamaica. Carnegie Institution of Wash- ington Publication 385: 1-564, 40 pl. WOODRING, W. Р., 1957, Geology and раеоп- tology of Canal Zone and adjoining parts of Panama. Description of Tertiary mollusks (Gastropods: Trochidae to Turritellidae). [United States] Geological Survey Professional Paper, 306-A: 42-62, 82-84, pl. 19, figs. 8-10. WOODRING, W. P., 1959, Geology and paleontol- ogy of Canal Zone and adjoining parts of Panama. Description of Tertiary mollusks (Gastropods: Vermetidae to Thaididae). [United States] Geological Survey Professional Paper, 306-B: 193-202, pl. 31, figs. 6-10, pl. 32, figs. 1, Я, 6; 9. WOODRING, W. Р., 1964, Geology and paleontol- ogy of Canal Zone and adjoining parts of Panama. Description of Tertiary mollusks (Gastropods: Columbellidae to Volutidae). [United States] Geological Survey Professional Paper 306-C: 255-256, pl. 39, figs. 6-8; 271, 274-275, pl. 47, fig. 12; 287-288. WOODRING, W. P., 1970, Geology and paleontol- ogy of Canal Zone and adjoining parts of Panama. Description of Tertiary mollusks (Gastropods: Eulimidae, Marginellidae to Helminthoglyptidae). [United States] Geological Survey Professional Paper, 306-D: 344, 345-359, 373, 378, 379; pl. 56, figs. 1, 2, 8, 7, 9; pl. 57, figs. 13, 14; pl. 58, figs. 1, 22. WORK, R. C., 1969, Systematics, ecology, and distribution of the mollusks of Los Roques, Venezuela. Bulletin of Marine Sciene, 19: 615- ТА APPENDIX 1. R/V John Elliott Pillsbury station data, arranged by area. Date Depth (all July) 1966 1. Golfo de Urabá, Colombia Р-353 (8°13.2'N, 76°50.1'W) 30 m Yi P-361 (851.9'N, 76°37.2'W) 40 m 12 Р-362 (8°57.5'N, 76°33.6'W) 72m 12 2. Golfo de Morrosquillo, Colombia Р-367 (9°31.3’N, 75°49.5’W) 40 m 13 3. Off Peninsula de Guajira, Colombia Р-766 (12°14.3'N, 70°40.0'W) 64m 28 Р-767 (12°16.1'N, 71°03.3'W) 25m 28 Р-768 (12°33.4'N, 71°10.8’W) 65 m 28 P-772 (12°20.2'N, 71%55.1'W) 11m 29 P-773 (12°17.0'N, 72°15.0'М/) 62m 29 4. Golfo de Venezuela, Venezuela P-759 (12°09.0'N, 69°57.5'W) 36 т 27 Р-760 (12°15.4'N, 69°57.5'W) 62 m 27 P-761 (11°52.0'N, 70°22.0'W) 35 m 27 5. Golfo de Triste, Venezuela P-749 (10°37.0'N, 67°57.9'W) 59 m 25 Р-750 (10°36.1'N, 68°12.6’W) 24 т 25 Р-756 (11°33.1’N, 69°12.6°W) 30 m 27 P-758 (11°42.2'N, 69°40.0’W) 16m 27 6. Off Cabo Cordera, Venezuela P-734 (11°01.8'N, 65°34.2'W) 65 m 22 Р-736 (10°57.0'N, 65°52.0’W) 100 m 22 P-737 (10°44.0'N, 66°07.0'W) 65 m 22 7. Off Isla Margarita, Venezuela P-716 (11°29.0'N, 63°51.0'W Р-717 (11°21.0’М, 64°10.0'W Р-718 (11°22.5'N, 64°08.0’W 60 m 20 Р-721 (11°06.5’N, 64°22.5'W 26m 21 ( ( ( ) 63m 20 ) | P-722 (11°04.0’М, 64°44.0'W) 91m 21 ) ) ) 64 m 20 P-723 (10°43.5'N, 64°16.0'W 65 m 21 P-727 (10°20.0'N, 65°02.2’W 64m 21 P-728 (10°22.5'N, 65°23.0'W 86 m 21 8. Off Peninsula de Paria, Venezuela Р-708 (11°24.7'N, 62°40.5'W) 70 т 19 Р-709 (11°08.8'N, 62°46.1'W) 46m 19 P-712 (11°08.0'N, 63°18.0’W) 25m 19 346 PETUCH APPENDIX 2. List of known living caenogastropods Cymatiidae from shallow water in the Golfo de Venezuela and Cymatium aquatile Reeve, T, 5 Golfo de Triste, Venezuela. T = collected in Golfo Сутайит krebsi Mörch, V, 5 de Triste; V = collected in Golfo de Venezuela; 1 = Cymatium parthenopeum (von Salis), V, 5 reported by Princz, 1980; 2 = records from R/V Cymatium pileare (Linnaeus), T, 5 Pillsbury expedition material; 3 = reported by Distorsio clathrata (Lamarck), T, V, 1, 5 Vermeij (personal communication); 4 = reported by Distorsio macgintyi Emerson & Puffer, V, 5 Gonzalez € Flores, 1972; 5 = personal observa- ое tions; 6 = reported by Flores, 1973, a, b. к Tonna даа (Linnaeus), V, 1, 5 Littorinidae Bursidae Littorina angulifera (Lamarck), T, 6 Bursa bufo (Hwass), T, V, 2, 5 Littorina cf. angustior Mórch, V, 3, 5 Muricidae Littorina flava King 8 Broderip, T, 5, 6 Calotrophon velero (E. Vokes), T, V, 2, 5 Littorina lineata Orbigny, T, 6 Littorina lineolata Orbigny, T, 6 Littorina meleagris (Potiez & Michaud), T, V, 1, 5, 6 Littorina nebulosa (Lamarck), T, V, 1, 5, 6 Littorina tessellata Philippi, T, V, 1 Chicoreus brevifrons (Lamarck), V, 1, 3, 4, 5 Dermomurex pauperculus (C. B. Adams), V, 3 Phyllonotus margaritensis (Abbott), V, 1, 3, 5 Murex аоптооге! Bullis, T, V, 2, 4, 5 Murex messorius Sowerby, T, V, 1, 2, 4, 5 Littorina ziczac (Gmelin), T, V, 1, 3, 5, 6 Thaididae Nodilittorina tuberculata (Menke), T, V, 5, 6 Purpura patula (Linnaeus), T, 4, 5 Tectarius muricatus (Linnaeus), T, V, 3, 5, 6 Thais coronatum (Lamarck), V, 5 к e Thais deltoidea (Lamarck), T, 4, 5 Architectonicidae = Thais haemastoma floridana Conrad, T, V, 1, 4 Architectonica nobilis (Róding), T, V, 1, 2, 5 Thais rustica (Lamarck), T, 4, 5 Turritellidae Thais trinitatensis (Guppy), V, 5 Turritella exoleta (Linnaeus), T, 2 Turritella variegata (Linnaeus), T, V, 1, 2, 5 Columbellidae Cerithiidae Columbella cf. mercatoria (Linnaeus), T, V, 1, 3, Cerithium atratum (Born), T, V, 3, 5 5 Cerithium eburneum Hwass, T, V, 3, 5 Mazatlania aciculata (Lamarck), V, 3, 5 Nitidella laevigata (Linnaeus), V, 3, 5 Dianazjeae Strombina pumilio (Reeve), T, V, 5 Planaxis nucleus (Hwass), T, V, 1, 3, 5 Buccinidae Antillophos candei Eh № №295 Pisania lauta (Reeve), Т, М, ee Pisania auritula (Link), T, u er Nassariidae Nassarius vibex (Say), V, 3, 5 Pallacera guadalupensis (Petit), V, 3, 5 Epitoniidae Epitonium albidum (Orbigny), V, 1 Epitonium lamellosum (Lamarck), T, V, 2, 5 Calyptraeidae Calyptraea centralis (Conrad), V, he Crepidula cymbaeformis Conrad, V, 3 Crepidula plana Say, V, 3 Crucibulum auricula (Gmelin) V, 1 Melongenidae Stombidae Te melongena (Linnaeus), T, V, 1, 2, 3, Strombus gigas Linnaeus, T, V, 1, 5 Strombus pugilis Linnaeus, T, V, 1, 2, 5 Fasciolariidae Strombus raninus Gmelin, T, 5 Latirus angulatus (Róding), T, V, 1, 2, 5 Naticidaó Latirus infundibulum (Gmelin), Ve Va 255 Polinices hepaticus (Röding), Т, V, 2, 5 Leucozonía nassa (Gmelin), V, 3, 5 Polinices lacteus (Guilding) т V > 5 Fasciolaria cf. tulipa (Linnaeus), V, 1, 2, 3, 5 AN Tr ae Fusinus closter Philippi, V, 1, 2, 5 Cypraeidae Turbinellidae Cypraea cinerea Gmelin, T, V, 2, 5 Vasum muricatum (Born), V, 1, 3, 5 Cypraea spurca acicularis Gmelin, T, V, 2, 5 Olividae Cypraea zebra Linnaeus, T, V, 2, 5 3.5 Siphocypraea donmoorei Petuch, V, 5 Ancilla glabrata (Linnaeus), T, V, 1, 2, Siphocypraea mus (Linnaeus), V, 1, 3, 5 V 5 1 Ancilla tankervillei (Swainson), V, 2, Oliva oblonga Marrat, V, 5 Cassidae Oliva scripta Lamarck, V, 1, 2, 5 Cassis madagascariensis E V, Olivella perplexa Olsson, V, 1, 5 4 Phalium granulatum (Born), T, V, 1, 2, 5 Olivella verreauxi (Duclos), Т, V, 1, 3, 5 RELICT CAENOGASTROPODS 347 Volutidae Conus daucus Hwass, V, 5 Voluta musica Linnaeus, V, 3, 5 Conus mappa Lightfoot, T, V, 5 Marginellidae Conus optabilis A. Adams, V, 2, 5 ; ; р SJ Conus puncticulatus Hwass, T, V, 1, 3, 5 ag das po (Megerle von Mühl Conus spurius Gmelin, T, М, 1, 2, 5 Persicula tessellata (Lamarck), V, 2, 5 Conus undatus Kiener, V, 2, 5 Prunum glans (Menke), V, 2, 5 Terebridae Prunum marginatum (Born), T, V, 1, 2, 5 Hastula salleana (Deshayes), T, V, 1, 3, 5 Prunum prunum (Gmelin), V, 5 Paraterebra taurina (Solander), V, 1, 5 Prunum pulchrum (Gray), V, 2, 5 Turidae Cancellariidae Clathodrillia gibbosa (Born), V, 5 Cancellaria reticulata (Linnaeus), T, V, 2, 5 Hindsiclava chazaliei (Dautzenberg), V, 2, 5 Conidae det barretti (Guppy) (as “P. albida”), T, V, Conus centurio Born, T, V, 2, 5 Prem Pass ¿IA Cas ita = we er IH wre oi hs + Le A <> MALACOLOGIA, 1981, 20(2): 349-357 LARVAL DEVELOPMENT, SETTLEMENT AND METAMORPHOSIS OF THE TROPICAL GASTROPOD TROCHUS NILOTICUS Gerald A. Heslinga Department of Zoology, University of Hawaii, Honolulu, Hawaii 96822, U.S.A. ABSTRACT Larvae of the archaeogastropod Trochus niloticus Linnaeus were reared through meta- morphosis in the laboratory at Palau, Western Caroline Islands (8°N, 135”E). Swimming trocho- phore larvae hatched 12 hr after fertilization at 27-30°C. Settlement and metamorphosis were induced by the red coralline alga Porolithon sp. and by a primary algal film on culture dishes. п the presence of these substrates, larval settlement began at 50-60 hr, and metamorphosis, the loss of the velar cilia, was completed as early as 3 days and as late as 8 days after fertilization. In the absence of live algal inducers, settlement occurred in 3-10 days and metamorphosis oc- curred in 4-21 days. Larvae were lecithotrophic and showed no evidence of ingestion of phyto- plankton prior to settlement. Unlike the free-swimming veligers of many tropical mesogastropods, Trochus niloticus larvae are characterized by lecithotrophy, rapid development, and limited ability to prolong the plank- tonic stage in the absence of a suitable settling substrate. The poor dispersal potential of this species corroborates and may explain its restricted native range in the Indo-West Pacific. It is suggested (1) that free-swimming archaeogastropod larvae are phylogenetically constrained to lecithotrophy and short planktonic lives, and (2) that high ambient temperatures act to accelerate development and restrict larval dispersal in tropical representatives of this group. INTRODUCTION A planktonic larval stage in the life history of benthic marine invertebrates enhances op- portunities for dispersal, colonization of new habitats, and genetic exchange (Scheltema, 1971a; Strathmann, 1974). The duration of the larval stage directly influences dispersal potential, and is thus of interest from ecologi- cal, biogeographical and evolutionary per- spectives. The present paper is an experi- mental investigation of the larval planktonic period of Trochus niloticus Linnaeus, a con- spicuous and economically important ar- chaeogastropod found on Indo-West Pacific coral reefs. This is the first such study to in- volve any of the numerous tropical represent- atives of the family Trochidae. The free-swimming veligers of tropical mesogastropods are predominantly plankto- trophic, often with long pelagic stages and great capacities for dispersal (Thorson, 1946, 1961; Robertson, 1964; Scheltema, 1966, 1968, 1971a, 1971b; Struhsaker & Costlow, 1968; Crisp, 1974; Brownell, 1977). Pechenik (1980) maintains that this trend 1$ a result of strong selection for larval longevity in tropical seas. In temperate seas, free-swimming ar- chaeogastropod larvae are lecithotrophic with short planktonic lives (Crofts, 1937, 1955; Fretter & Graham, 1962; Anderson, 1965; Desai, 1966; Underwood, 1972; Manly, 1976). It is thus relevant to ask whether the numerous and ecologically important tropical archaeogastropods, including the limpets, abalones, turbans and top shells, conform to the reproductive trends set by tropical meso- gastropods, or whether they are phylogeneti- cally constrained to lecithotrophy, rapid de- velopment, and poor dispersal ability. Investi- gation of the larval biology of a number of tropical archaeogastropods is desirable be- cause it will permit a comparative approach to the problem of how phylogeny, mode of de- velopment, temperature and latitude interact to determine larval life spans. Although top shells of the family Trochidae have been frequent subjects of morphoge- netic studies, there has been no experimental consideration of how environmental variables influence the timing of metamoprhosis. Poor laboratory survival of trochid larvae has in some cases precluded this kind of analysis (Underwood, 1972). Moorhouse (1932) ob- served spawning of Trochus niloticus in Australia but was not successful in rearing the larvae. Attempts to produce viable T. niloticus larvae by fertilization of excised gametes (349) 350 HESLINGA (Rao, 1937), injection of КС! (Smith, personal communication) or exposure to H202 (Hill- mann, personal communication) have also been unsatisfactory. Under appropriate la- boratory conditions, however, T. niloticus spawn epidemically on a predictable monthly schedule (Heslinga 4 Hillmann, in press), making this species ideal for investigations of larval biology. Here | describe early develop- ment and the effects of algal substrates on metamorphosis. METHODS Trochus niloticus collected on the outer reefs of Paiau were held in flowing seawater tanks (27-30°С) at the Micronesian Maricul- ture Demonstration Center. Following epi- demic spawning in the tanks, fertilized eggs were dipped out in 1 liter plastic beakers. After the eggs settled the supernatant water was decanted, the eggs were rinsed twice with Millipore filtered seawater (0.45 um роге size) and adjusted to a density of about 1 egg/10 ml. Beakers were covered with plastic wrap and kept indoors at 27-30°С. Aeration was not provided. Larvae were transferred by pipet to experimental petri dishes, or placed on depression slides for microscopic exami- nation, measurement by ocular micrometer, or photomicrography. Settlement experiments were conducted in 50 ml plastic petri dishes with covers. Each dish was filled with 40 ml seawater and stocked with 20 veliger larvae. Filtered sea- water was used in all cultures except those calling for a raw (unfiltered) seawater regime. A stated sample size of 6, for example, refers to 6 replicate petri dish cultures containing a total of 120 larvae. Five culture regimes were tested: 1)Primary algal film cultures Prior to use in experiments, petri dishes were kept for 2 weeks in unfiltered running seawater, until covered by a thin film of bac- teria, diatoms, flagellates and fine filamentous green algae (hereafter called a primary algal film). Treatment A was initiated day 2, N=6; treatment B was initiated day 6, N=3; treat- ment C was initiated day 10, N=3. 2) Filtered seawater cultures Sterile petri dishes with filtered seawater and no added substrate. Treatment D was in- itiated day 2; N=9. 3) Raw seawater cultures Same as filtered seawater cultures, except raw (unfiltered) seawater was used. Treat- ment G was initiated day 2; N=3. 4) Сога!те red alga (Porolithon) cultures Fragments of live Porolithon were scraped from the shell of an adult Trochus niloticus collected on the reef margin. A sample was preserved, examined histologically, and iden- tified to genus (Gordon et al., 1976). Live frag- ments were rinsed repeatedly in filtered sea- water, then added to culture dishes such that the total live algal surface area was = 1 ст?/ dish. Treatment E was initiated day 2; N=3. Treatment F was initiated day 6; N=3. 5) Control substrate for coralline algae cultures Fragments of dead, sunbleached coralline algae were collected from dried T. niloticus shells and boiled in fresh water. Fragments were then rinsed in filtered seawater and added to petri dishes in quantities similar to #4. Treatment H was initiated day 2; N=3. Dishes were kept indoors at 27-30°C, away from direct sunlight. Water in each dish was changed at 2 day intervals by pipetting off 35 ml culture water and replacing it with an equal volume of filtered seawater or raw sea- water, as required. Cultures were examined twice daily (AM and PM) under a dissecting microscope, and all dead or metamorphosed larvae were removed. Metamorphosis was considered complete after total loss of the velar cilia. Squash preparations of newly metamorphosed juveniles were examined for evidence of food ingestion. Cultures were terminated when all larvae had either meta- morphosed or died. RESULTS Development Eggs released by female Trochus niloticus were dark green and approximately 185 ит in diameter (Fig. 1A). They were surrounded by a pitted jelly layer measuring 475-500 ит across. Fertilization occurred immediately and within a few minutes the vitelline mem- brane rose to a diameter of about 225 um. Cleavage began after 30 minutes and fol- lowed the typical spiral pattern (e.g. Robert, 1902). Gastrulation by epiboly occurred after 5-6 hr. By 8 hours post-fertilization, trocho- DEVELOPMENT OF TROCHUS NILOTICUS 351 22 бо: se 0.1 mm SS! С 10 a 12 D F FIG. 1. Embryonic and larval Trochus niloticus. A, fertilized egg, time (t) = 3 minutes after fertilization; В, gastrula just before hatching, t = 11.5 hr; C, pretorsional trochophore with embryonic shell growth beginning, { = 14 hr; D, veliger with complete larval shell, after 90° torsion, t = 20 hr; E, veliger with reduced velum, t = 65 hr; Е, metamorphosed juvenile, t = 73 hr. All figures drawn to same scale. 1, vitelline membrane; 2, jelly layer; 3, embryonic shell; 4, foot rudiment; 5, velum; 6, foot; 7, larval retractor muscle attachment site; 8, mantle margin; 9, propodium; 10, eye; 11, cephalic tentacles; 12, operculum. blast cells were ciliated (Fig. 1B), and rotation of the gastrula within the vitelline membrane began after about 9 hours. Rotation was at first feeble and intermittent, but became increas- ingly vigorous until trochophores hatched at 12 hr. Newly hatched trochophores swam near the top few centimeters of water in 1 liter beakers, alternately rising to the surface and dropping briefly through the water column. The embryonic shell appeared near the vegetal pole within an hour after hatching, and spread rapidly to the anterior (Fig. 1C). The first phase of torsion occurred between 14 and 18 hr, rotating the shell and visceral hump through 90* with respect to the head and foot. By 20 hr the larval shell was complete and measured approximately 290 ит in diameter (Fig. 1D). At this point the velum was a simple circular ridge with a single band of prominent cilia. Full retraction into the larval shell was first observed at 24 hr. Larvae continued to swim near the top of the water column, but began to show negative phototaxis, swim- ming to the opposite side of the beaker in re- sponse to light. In the presence of live algal substrate, veligers began settling and creeping at 50-60 hr. Newly settled larvae probed the substrate with the tip of the propodium, frequently flex- ing the anterior edge of the foot and passing the propodium over the mouth. By this time the velum had developed a mid-ventral cleft, resulting in 2 distinct lobes. The lobes were gradually resorbed, revealing 2 cephalic ten- tacle rudiments and 2 black eye rudiments near the center of the velum. The velar lobes became progressively reduced in size and their cilia were sloughed off until only a few remained beating. The last cilia were cast from the vestigial velum at about 70 hr (Fig. 1F). The larval shell had not increased in size since its completion at 20 hr. Induction of metamorphosis Mean survival through metamorphosis in all petri dish cultures was 91% (range: 75- 100%), with 8 of 33 cultures exhibiting 100% survival. Mortalities were not clearly associ- ated with any particular stage of development. In filtered seawater cultures with no algal sub- strate, metamorphosis occurred on days 4- 352 HESLINGA 100 algal film О Lu (7) о 80 E О. X О 60 = < A B r uU 40 = > 20 filtered sw 10 15 20 AGE (DAYS) FIG. 2. Metamorphosis of Trochus niloticus larvae in response to a primary algal film on culture dishes. Treatments were initiated on day 2 (A), day 6 (B), and day 10 (C). Treatment Dis filtered seawater control with no algal film. Points are means +1 S.E. А, N = 6 replicate cultures, each containing 20 larvae; В, М = 3; С, М = 3; 0, М = 9. 21, and was completed by 50% of the larval population by days 16-17 (Fig. 2D). Among larvae that had not metamorphosed by day 8, the velum became conspicuously reduced in size and the digestive gland became increas- ingly transparent as the yolk supply dimin- ished. By day 10, 100% settlement had oc- curred among unmetamorphosed larvae, and from day 10-21 creeping was the mode of locomotion. During this creeping period the velar cilia continued to beat sporadically but no further swimming was observed. Larvae exposed to a primary algal film be- ginning on the 2nd, 6th or 10th day meta- morphosed earlier than those reared with no algal substrate (Fig. 2). When algal film ex- posure was initiated on day 2, 14% of the larval population metamorphosed by day 3, 50% metamorphosed by days 4-5, and 93% metamorphosed by day 8 (Fig. 2A). When algal film exposure was initiated on day 6, 32% of the larval population metamorphosed by day 7, 50% metamorphosed by days 8-9, and 100% metamorphosed by day 12 (Fig. 2B). The most rapid response to algal film was shown by the oldest larvae. When this treatment was initiated on the morning of day 10, 91% of the larval population metamor- phosed within 10 hours (Fig. 2C). In contrast, mean percent metamorphosis in filtered sea- water cultures with no algal substrate re- mained below 10% through day 10; in 3 of 9 replicates no metamorphosis occurred until day 11 (Fig. 2D). These results suggest that the algal films induced metamorphosis in Trochus niloticus larvae and that the intensity of response was age dependent. Fragments of live Porolithon introduced on day 2 or day 6 caused metamorphic re- sponses similar to those caused by a primary DEVELOPMENT OF TROCHUS NILOTICUS 353 100 coralline algae METAMORPHOSED % 10 15 20 AGE (DAYS) FIG. 3. Metamorphosis of Trochus niloticus larvae in response to living fragments of the coralline red alga Porolithon sp., introduced on day 2 (E) and day 6 (F). Treatment © is raw unfiltered seawater with no added substrate. Treatment H is filtered seawater control with dead, sterile fragments of coralline algae. Points are means +1 S.E. N = 3 for all treatments. algal film (Figs. 3E, 3F). In both cases where live Porolithon was introduced, 20-30% of the larval population metamorphosed within 24 hr, while the remaining individuals metamor- phosed during the next 2-4 days. Моп- quantitative behavioral observations indicated that larvae crept and metamorphosed prefer- entially on the living. algal fragments, and re- mained on or under the fragments after metamorphosing. No evidence of food inges- tion prior to settlement was found among larvae exposed to a primary algal film or to Porolithon fragments. Larvae reared in dishes containing boiled coralline algae fragments showed no ap- parent attraction to the fragments after settling. No metamorphosis occurred in these cultures until day 8, after which the response curve resembled that of larvae reared in filtered sea- water with no algal substrate (Fig. 3H). Metamorphosis of larvae reared in raw sea- water began on day 3 and was completed by day 19 (Fig. 3G). Despite an abundance of small flagellates and diatoms in the culture water, close examination of live larvae and squash preparations showed no indication that planktonic feeding had occurred prior to settlement. DISCUSSION The family Trochidae includes species that deposit benthic egg masses and others that broadcast their spawn. Trochus niloticus be- longs to the second category, and like previ- ously described archaeogastropods (Fretter & Graham, 1962), hatches in the trochophore stage. Gohar & Eisawy (1963) and Eisawy (1970) reported without supporting evidence 354 HESLINGA that T. niloticus hatches as a free swimming veliger. The gross features of development in Trochus niloticus appear similar to those of 7. dentatus, described in detail by Езаму (1970). T. niloticus eggs are similar in size to those of Т. dentatus, but differ in that they are surrounded by а conspicuous pitted jelly layer. In the temperate water trochids Mono- donta lineata and Gibbula cineraria, the jelly layer disperses within minutes and may act as an early block to polyspermy (Desai, 1966; Underwood, 1972). This is clearly not the function of the layer in Т. niloticus since it does not inhibit fertilization and since it per- sists until hatching. The jelly layer increases the surface area of T. niloticus eggs by at least 500%, and in nature it would slow the sinking rate of embryos and thus reduce the risk of predation by benthic planktivores. The minimum time required for Trochus niloticus larvae to metamorphose was about 3 days; in the presence of algal inducers all larvae shed the velum by day 8. Although some individuals in filtered seawater cultures retained the velar cilia until day 21, most of the larval population settled during the first week, and no swimming at all occurred after day 10. In the absence of a suitable settling substrate, velar resorption was delayed by no more than a few days. By day 10 the reduced velum is probably no longer functional as a swimming organ, even though it may retain some cilia. Anderson (1962) suggested that the velar cilia act to stabilize larval creeping in the limpet Cellana tramoserica. This could explain why archaeogastropod larvae often retain the velar cilia for some time after settle- ment. To the extent that this is true, the time of total velar cilia loss must give an overestimate of the duration of the planktonic period. Induction of metamorphosis by organic extracts or substrates, including algal films, has been documented in a variety of marine invertebrate larvae (see Crisp, 1974, 1976; Chia € Rice, 1978). Morse et al. (1979) identi- fied gamma-amino butyric acid and related compounds, isolated from coralline red algae, as inducers of metamorphosis in temperate water archaeogastropods of the genus Haliotis. lt seems likely that a similar induc- tion-response interaction occurs between Porolithon and the larvae of Trochus niloticus. Porolithon and other crustose coralline algae are suspected or known to induce settlement and metamorphosis in several asteroids as well (Henderson 4 Lucas, 1971; Yamaguchi, 1973, 1974; Lucas 4 Jones, 1976; Barker, 1977). Morse et al. (in press) argue that such interactions are mutually beneficial and co- evolved; larval invertebrates are induced to metamorphose on a suitable microhabitat which may also be a food source, while the inducing algae are grazed free of potentially harmful fouling epiphytes. The sensory mech- anisms by which larvae perceive inducer sub- stances are not known. The dramatic response of Trochus niloticus larvae to live coralline algae and to a primary algal film on culture dishes indicates that this prosobranch is not narrowly restricted in sub- strate preference, as are some prey-specific or host-specific opisthobranch juveniles (Hadfield, 1977; Perron & Turner, 1977). If T. niloticus larvae in nature settle preferentially on red coralline algae or on algal film covered substrates, the potential space for recruitment would appear to be broad. Porolithon is a dominant alga on Indo-West Pacific coral reef margins, and it is also common on inner and outer reef flats (Gordon et al., 1976). T. niloticus juveniles are most often found near outer reef flat rubble and cobble zones (Moorhouse, 1932; McGowan, 1956; Gail 8 Devambez, 1958; Smith, 1979) where fila- mentous and encrusting algae are common and live coral cover 1$ low. Larval settlement probably occurs here. Trochus niloticus is indigenous only to the Indo-Malaysian area, Melanesia, and Yap and Palau in Micronesia (Hedley, 1917; Rao, 1937; Vermeij, 1978), but because of its eco- nomic importance the species has been intro- duced to several other island groups to the east, where it is now common. Successful 7. niloticus introductions have been carried out in the Marianas (see Smith, 1979), the Caro- line, Marshall and Cook Islands (Gail 8 Devambez, 1958), French Polynesia (Doumenge, 1972) and Hawaii (Kay, 1979). The rather restricted indigenous range may be attributable in part to the short larval life span. Before the planktonic larva of 7. niloticus had Бееп identified, McGowan (1958: 18) proposed that the planktonic term is “not of very long duration, otherwise Trochus niloticus would have bridged the gap naturally between Palau and Ngulu [=300 km], or Yap and Ulithi [=200 km], as have other species of Micronesian Trochus.” The native range of 7. niloticus appears to have been limited by poor larval dispersal rather than by habitat unsuitability on other tropical Pacific islands. DEVELOPMENT OF TROCHUS NILOTICUS 355 The larval life span of Trochus niloticus 15 significantly shorter than that of other broad- cast spawning archaeogastropods from higher latitudes. T. niloticus metamorphosed within 3 days at 27-30°C, while the temperate trochids Gibbula cineraria (Underwood, 1972) and Monodonta lineata (Desai, 1966) required a minimum of 9 days at 12° and 7 days at 15-20°C, respectively. Three temper- ate acmaeid limpet species reared by Ander- son (1965) metamorphosed in 7-10 days at 20°, while several temperate haliotids (Leighton, 1972, 1974) and patellids (Dodd, 1957; Anderson, 1962) metamorphosed after 1-2 weeks when reared at local ambient temperatures. There is no evidence that temperate ar- chaeogastropod larvae are capable of sus- pension feeding while swimming (Strath- mann, 1978) though Crofts (1937) and Morse et al. (in press) suggested that some benthic feeding may occur between settlement and metamorphosis. If planktotrophy exists in ex- tant archaeogastropods it should be evident in tropical species. The data presented here, however, suggest that tropical archaeo- gastropods are similar to their temperate counterparts in being phylogenetically con- strained to lecithotrophy and short larval lives. Moreover, unless rate-compensating me- chanisms exist, high tropical temperatures should accelerate development in lecitho- trophic larvae and further reduce the duration of the swimming phase. These considerations may account for the limited native range of Trochus niloticus, and may in part explain the restricted tropical distributions of several ar- chaeogastropod families previously noted by Kay (1967) and Vermeij (1972). ACKNOWLEDGEMENTS | am grateful to Drs. W. М. Hamner and К. Chave for advice and encouragement during the study. | thank the staff of the Micronesian Mariculture Demonstration Center, especially A. Hillmann, N. Idechong and R. Braley, for technical assistance. Drs. E. A. Kay, J. E. Bardach, M. Hadfield and F. Perron offered critical comments on the manuscript. This work was supported by grants from the Re- source Systems Institute, East-West Center (Honolulu), the Hawaiian Malacological So- ciety, the Society of Sigma Xi, and the Marine Resoures Division, U.S. Trust Territory of the Pacific Islands. REFERENCES CITED ANDERSON, D. T., 1962, The reproduction and early life histories of the gastropods Bembicium auratum (Quoy and Gaimard) (Fam. Littorinidae), Cellana tramoserica (Sower.) (Fam. Patellidae) and Ме/апейа melanotragus (Smith) (Fam. Neritidae). Proceedings of the Linnaean Society of New South Wales, 87: 62-82. ANDERSON, D. T., 1965, The reproduction and early life histories of the gastropods Notoacmaea petterdi (Ten.-Woods), Chiazacmaea flammea (Quoy and Gaimard) and Patelloida alticostata (Angas). Proceedings of the Linnaean Society of New South Wales, 90: 106-114. BARKER, M. F., 1977, Observations on the settle- ment of the brachiolaria larvae of Stichaster australis (Verrill) and Coscinasterias calamaria (Gray) (Echinodermata: Asteroidea) in the labo- ratory and on the shore. Journal of Experimental Marine Biology and Ecology, 30: 95-108. BROWNELL, W. N., 1977, Reproduction, labora- tory culture, and growth of Strombus gigas, S. costatus and S. pugilis in Los Roques, Venezu- ela. Bulletin of Marine Science, 27: 668-680. CHIA, Е. 5. & RICE, М. Е. (eds.), 1978, Settlement and Metamorphosis of Marine Invertebrate Larvae. Elsevier, New York, 290 p. CRISP, D. J., 1974, Factors influencing the settle- ment of marine invertebrate larvae. In: GRANT, Р. Т. 8 MACKIE, А. М. (ed.), Chemoreception in Marine Organisms. Academic Press, New York, 290 p. CRISP, D. J., 1976, Settlement responses in ma- rine organisms. In: NEWELL, R. C. (ed.), Adap- tations to Environment: Essays on the Physiol- оду of Marine Animals. Butterworth's, London, р. 83-124. CROFTS, D. R., 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. R., 1955, Muscle morphogenesis in primitive gastropods and its relation to torsion. Proceedings of the Zoological Society of Lon- don, 37: 1-17. DESAI, В. N., 1966, The biology of Monodonta lineata (da Costa). Proceedings of the Malaco- logical Society of London, 37: 1-17. DODD, J. M., 1957, Artificial fertilization, larval de- velopment and metamorphosis in Patella vulgata L. and Patella coerula L. Pubblicazioni della Stazione Zoologica di Napoli, 29: 172-186. DOUMENGE, F., 1972, Commercial exploitation of Trochus (French Polynesia). South Pacific Com- mission Fisheries Newsletter, 3: 4, 32. EISAWY, A. M., 1970, The spawning and develop- ment of Trochus (Tectus) dentatus Forskal. Bul- letin of the Institute of Oceanography and Fish- eries, Cairo, 1: 379-393. FRETTER, V. & GRAHAM, A., 1962, British Proso- branch Molluscs. Ray Society, London, 755 p. 356 HESLINGA GAIL, В. & DEVAMBEZ, L., 1958, Selected anno- tated bibliograpy of Trochus. South Pacific Commission Technical Paper, 111: 1-20. GOHAR, H. A. F. 8 EISAWY, A. M., 1963, The egg masses and development of Trochus (Infundi- bulops) erythraeus Brocchi. Publications of the Marine Biological Station, Al-Ghardaga, Cairo, 12: 191-203. GORDON, С. D., MASAKI, T. 8 AKIOKA, H., 1976, Floristic and distributional account of the com- mon crustose coralline algae of Guam. Micro- nesica, 12: 247-277. HADFIELD, М. G., 1977, Chemical interactions in larval settling of a marine gastropod. In: FAULKNER, D. J. & FENICAL, W. H. (ed.), Marine Natural Products Chemistry. Plenum, New York, p. 403-413. HEDLEY, C., 1917, The economics of Trochus niloticus. Australian Zoologist, 1: 69-73. HENDERSON, J. A. & LUCAS, J. S., 1971, Larval development and metamorphosis of Acanthaster planci (Asteroidea). Nature, 232: 655. HESLINGA, G. A. & HILLMANN, A., 1981, Hatch- ery culture of the commercial top snail Trochus niloticus in Palau, Caroline Islands. Aquaculture, 22: 3543. KAY, E. A., 1967, The composition and relation- ships of marine molluscan fauna of the Hawaiian Islands. Venus, 25: 94-104. KAY, E. A., 1979, Hawaiian Marine Shells. Bishop Museum, Honolulu, 653 p. LEIGHTON, D. L., 1972, Laboratory observations on the early growth of the abalone, Haliotis sorenseni, and the effect of temperature on larva! development and settling success. Fishery Bulletin, 70: 373-381. LEIGHTON, D. L., 1974, The influence of tempera- ture on larval and juvenile growth in three spe- cies of Southern California abalones. Fishery Bulletin, 72: 1137-1145. LUCAS, J. $. 4 JONES, М. S., 1976, Hybrid crown- of-thorns starfish (Acanthaster planci x A. brevispinus) reared to metamorphosis in the laboratory. Nature, 198: 610-611. MANLY, R., 1976, The larval development of Тисойа pullus (L.). Journal of Molluscan Studies, 42: 361-369. MCGOWAN, J. A., 1956, The current status of the Trochus industry in Micronesia. Report to the High Commissioner, U.S. Trust Territory of the Pacific Islands, Saipan, 11 p. MCGOWAN, J. A., 1958, The Trochus fishery of the Trust Territory of the Pacific Islands. Report to the High Commissioner, U.S. Trust Territory of the Pacific Islands, Saipan, 46 p. MOORHOUSE, F. W., 1932, Notes on Trochus niloticus. Scientific Reports of the Great Barrier Reef Expedition, 1928-1929, 3: 315-324. MORSE, D. E., HOOKER, N., DUNCAN, H. & GENSEN, L., 1979, Y-Aminobutyric acid, a neurotransmitter, induces planktonic abalone larvae to settle and begin metamorphosis. Sci- ence, 204: 407-410. MORSE, D. E., TEGNER, M., DUNCAN, H., HOOKER, N., TREVELYAN, G. & CAMERON, A., in press, Induction of settling and meta- morphosis of planktonic molluscan (Haliotis) larvae. In: MULLER-SCHWARZ, D., (ed.), Chemical Signaling in Vertebrates and Aquatic Animals. Plenum, New York. PECHENIK, J. A. 1980, Growth and energy balance during the larval lives of three prosobranch gastropods. Journal of Experimental Marine Biology and Ecology, 44: 1-28. PERRON, Е. Е. & TURNER, В. D., 1977, Develop- ment, metamorphosis, and natural history of the nudibranch Doridella obscura Verrill (Corambi- dae: Opisthobranchia). Journal of Experimental Marine Biology and Ecology, 27: 171-185. RAO, H. S., 1937, On the habitat and habits of Trochus niloticus Linn. in the Andaman Seas. Records of the Indian Museum, Calcutta, 39: 47-82. ROBERT, A., 1902, Recherches sur le develop- pement des troques. Archives de Zoologie experimentale et générale, 10: 269-538. ROBERTSON, R., 1964, Dispersal and wastage of larval Philippia krebsii (Gastropoda: Architectoni- cidae) in the North Atlantic. Proceedings of the Academy of Natural Sciences of Philadelphia, 116: 1-27. SCHELTEMA, R. S., 1966, Evidence for trans- Atlantic transport of gastropod larvae belonging to the genus Cymatium. Deep Sea Research, 13: 83-95. SCHELTEMA, В. S., 1968, Dispersal of larvae by ocean currents and its importance to the zoo- geograpy of shoal water tropical species. Nature, 217: 1159-1162. SCHELTEMA, В. S., 1971a, Larval dispersal as a means of genetic exchange between geographi- cally separated populations of shallow-water benthic marine gastropods. Biological Bulletin, 140: 284-322. SCHELTEMA, R. S., 1971b, The dispersal of the larvae of shoal-water benthic invertebrate spe- cies over long distances by ocean currents. In: CRISP, D. J. (ed.), Fourth European Marine Bi- ology Symposium. Cambridge University, Cam- bridge, p. 7-28. SMITH, B. D., 1979, Growth rate, abundance and distribution of the topshell Trochus niloticus on Guam. M.S. Thesis, University of Guam, 24 p. STRATHMANN, В. R., 1974, The spread of sibling larvae of sedentary marine invertebrates. Amer- ican Naturalist, 108: 29-44. STRATHMANN, R. R., 1978, The evolution and loss of feeding larval stages of marine inverte- brates. Evolution, 32: 894-906. STRUHSAKER, J. W. 8 COSTLOW, J. D., 1968, Larval development of Littorina pica (Proso- branchia: Mesogastropoda), reared in the labo- ratory. Proceedings of the Malacological So- ciety of London, 38: 153-160. THORSON, G., 1946, Reproduction and larval de- velopment of Danish marine bottom inverte- DEVELOPMENT OF TROCHUS NILOTICUS 357 brates. Meddelelser fra Kommissionen for Danmarks Fiskeri- og Havundersdgelser, ser. Plankton, 4, 523 p. THORSON, G., 1961, Reproduction and larval ecology of marine bottom invertebrates as re- lated to larval transport by ocean currents. In: SEARS, M. (ed.), Oceanography. American As- sociation for the Advancement of Science, Washington, D.C., р. 455-474. UNDERWOOD, A. J., 1972, Spawning, larval de- velopment and settlement behavior of Gibbula cineraria (Gastropoda: Prosobranchia) with a reappraisal of torsion in gastropods. Marine Bi- ology, 17: 341-349. VERMEIJ, С. J., 1972, Endemism and environ- ment: some shore molluscs of the tropical Atlan- tic. American Naturalist, 106: 89-101. VERMEIJ, С. J., 1978, Biogeography and Adapta- tion. Harvard University, Cambridge, Massachu- setts, 332 p. YAMAGUCHI, M., 1973, Early life histories of coral reef asteroids, with special reference to Acanthaster planci (L.). In: JONES, А. O. 4 ENDEAN, R. (eds.), Biology and Geology of Coral Reefs, Vol. 2, Biology I. Academic Press, New York, p. 367-387. YAMAGUCHI, M., 1974, Larval life span of the coral reef asteroid Gomophia egyptiaca Gray. Micronesica 10: 57-64. ITA ES = vo da A ime | ar Hire, Ligue вам о su oe i ah par gti! À 1168 | i LI nn E SIMI ny i ) qe 1 бо, Sud AA er | i pris he sent vet CRU Ад Ñ i ct prea a MACRO Oy tiger ete Win: tet Ae AN m M Pre > uy as a JE NE a A: Mine D LATE) A UE HR vi ee | ht $ pri м An a Fr idR | | | e Dal Ae 7” Mi de ta di. hr У ar It le vie à ton" \ у ми АГ 1 € Li in 4 | ia y Ку Canina МИЯ a Mi j a na И eut th tir el \ 1 : = | ' | | à eue, LA вы arte ¿e } ÿ % que en h | À Head l'a ¿e AO y у o ‘ ES ‚И } WE air; | | ayy 0 Aa ALO lia = is Ad р CE PM “re > | qe AMERICAN MALACOLOGICAL UNION SYMPOSIUM: FEEDING MECHANISMS OF PREDATORY MOLLUSCS Organized by Alan J. Kohn President: Clyde F. E. Roper Louisville, Kentucky 24 July, 1980 MALACOLOGIA, 1981, 20(2): 359 FEEDING MECHANISMS OF PREDATORY MOLLUSCS: INTRODUCTION TO THE SYMPOSIUM Alan J. Kohn Department of Zoology, University of Washington, Seattle, Washington 98195, U.S.A. The most important problem an animal faces in nature is obtaining a sufficient quan- tity of the right kind of food. An animal that fails in this endeavor soon learns the meaning of the term ‘energy crisis.‘ The animal that solves this basic problem can at least face the other trials and tribulations of its world on a full stomach. The goal of this Symposium was to address recent advances in understanding of how predatory molluscs detect, obtain, and ingest their prey. The Symposium focused on the feeding mechanism itself, but the presenta- tions and discussions ranged widely over many aspects of feeding: how predators be- come aware that prey 1$ nearby, the structure and functioning of mechanisms for overcom- ing prey organisms, and how their feeding biology аНес{ how predatory molluscs inter- act with other organisms in their communities. The Symposium was held 24 July 1980, at the annual meeting of the American Malacologi- cal Union in Louisville, Kentucky. How important is predation in the Mollusca? IS it primitive in the phylum? Although the Symposium participants did not attempt to answer these questions, they provided a heuristic framework for the discussions. If the molluscs evolved from a platyhelminth-like ancestor as many zoologists believe, the answer to the second question is probably yes. As to the first question, all 7 classes of molluscs contain predatory representatives, and 4 are probably almost exclusively preda- tory. These are the Monoplacophora, Aplaco- phora, Scaphopoda, and Cephalopoda. Many gastropods are predatory, although this is probably a derived characteristic in this class. In the remaining two classes, Bivalvia and Polyplacophora, carnivory is uncommon and undoubtedly derived. The class Gastropoda contains the largest number and greatest diversity of predatory molluscs. Nevertheless, it is unfortunate that only one other class could be represented in the Symposium papers. In the Introduction to an earlier A.M.U. symposium, George Davis (Malacologia, 1978, 17: 163) noted the рге- dominance of studies of terrestrial gastropods as a source of “the most exciting and encom- passing work in molluscan evolution over the last 15 years.” The papers that follow contrib- ute importantly to understanding the feeding biology of predatory molluscs. They also demonstrate that evolutionarily less well known taxa of molluscs have equally exciting possibilities and that examination of distinct phyletic lineages can provide independent tests of theories in evolutionary biology. (359) E A | | nace N I № Ах a aa ast + у ры Ve) rs N PAR AN Kur а Le | 5 | | ne “i can м u + À ay y м et bi = и В fe) PANNE rot MALACOLOGIA, 1981, 20(2): 361-383 COMPARATIVE MORPHOLOGY OF THE RADULAE AND ALIMENTARY TRACTS IN THE APLACOPHORA! Amelie H. Scheltema Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543, U.S.A. ABSTRACT The alimentary tract was studied in one genus of Neomeniomorpha and in five genera of Chaetodermomorpha. The cuticular oral shield of the Chaetodermomorpha is part of the foregut cuticle. A dorsal ciliated tract or typhlosole, an unarticulated radula on a radular membrane, an odontophore with bolsters within the haemocoele, and paired tubular salivary glands are con- servative molluscan characters. It is not certain whether an undivided stomach-digestive gland (Neomeniomorpha) or separate stomach and digestive diverticulum (Chaetodermomorpha) rep- resents the primitive midgut in the Aplacophora. The molluscan style may primitively have been formed throughout the stomach and anterior intestine (Scutopus). A style sac with protostyle and a gastric shield have evolved together independently in one family of carnivorous Aplacophora (Chaetodermatidae). The genera studied here exhibit an evolution of the radula from rows of distichous teeth firmly affixed to a divided or fused radular membrane to (1) a gastropod-like articulated radula and (2) a highly specialized pincers-like radula. The odontophore has evolved from a structure scarcely protruded into the buccal cavity to one with the tip lying free, surrounded by deep buccal pouches and sublingual cavity. A carnivorous diet is related both to a primitive radula (Gymnomenia) and to the specialized radulae of Prochaetoderma and the Chaetodermatidae (Chaetoderma and Falcidens). Evolution to a gastropod-like radula combined with jaws which hold the mouth open in Prochaetoderma has made possible a diet which 15 independent of particle size. A broad food source may be one reason that some species of Prochaetoderma are numerically dominant members of the fauna in the deep sea, where food may be limiting. INTRODUCTION The alimentary tract of the Aplacophora, excepting the radula, has generally received less attention than other organ systems. The radula itself has usually been described from histologic preparations; isolated radulae with complete radular membranes have been fig- ured for only a few species in the subclass Chaetodermomorpha (= Caudofoveata) (Kowalevsky, 1901; Scheltema, 1972, 1976; Ivanov, 1979) and for only two species of Epimenia in the subclass Neomeniomorpha (= Solenogastres sensu Salvini-Plawen) (Baba, 1939, 1940). Gut morphologies have usually been described as part of species de- scriptions; no integrated overview exists for the class as a whole outside of literature re- views in the standard invertebrate treatises. The literature on feeding and digestion in the Neomeniomorpha was reviewed by Salvini- Plawen (1967b), who has recently proposed evolutionary sequences in the digestive sys- tem in the mollusks (1980). The only develop- mental studies on the alimentary tract are for Epimenia verrucosa (Baba, 1938) and Neomenia carinata (Thompson, 1960). This paper examines the morphologies of the radula and fore- and mid-guts of certain aplacophoran families and relates these mor- phologies to the feeding type and the ecologic importance of these families in the deep-sea benthos. Possible phylogenetic relationships of the aplacophoran radula and gut morphol- ogies are proposed and the bearing of these relationships to understanding molluscan evolution is discussed. Primary consideration is given to the Chaetodermomorpha, but one primitive neomeniomorph is examined (Fig. 1). 1Contribution No. 4655 from the Woods Hole Oceanographic Institution. (361) 362 SCHELTEMA FIG. 1. Alimentary tract of Aplacophora, semi-schematic. A: Gymnomenia n. sp., anterior two-thirds only; B: Scutopus robustus; С: Limifossor talpoideus; D: Prochaetoderma sp. y, posterior end not shown; Е: Falcidens caudatus, posterior end not shown: F: Chaetoderma nitidulum, anterior half only. A, F from sagittal sections; B-E from cleared specimens. Gonad not indicated in B, E, or F. Scales in mm. 1, foregut; 2, dorsal caecum; 3, atrium; 4, midgut of undifferentiated stomach and digestive gland; 5, odontophore and radula; 6, buccal cavity; 7, stomach; 8, digestive diverticulum; 9, intestine; 10, style sac; 11, jaws; 12, gonad. MATERIALS AND METHODS Thirteen species in four families and six genera were examined, two species by histo- logic sections only, three by isolated radula preparations only, and eight by both histologic and radula preparations. All unnamed spe- cies or species identified by letter only will be formally described elsewhere. Subclass Neomeniomorpha Fam. Wireniidae (regarded as primitive on basis of spicule shape, thin integument, and lack of ventral foregut glands, Salvini-Plawen, 1978). (1) Gymnomenia n. sp. 620 m, off Walvis Bay, Namibia, Africa (23°00'S, 12°58’E). 4 specimens (cross and sagittal sections, 2 radula preparations). Subclass Chaetodermomorpha Fam. Limifossoridae (regarded as primitive on basis of vestige of ventral foot furrow in Scutopus, Salvini-Plawen, 1972a). (2) Scutopus megaradulatus Salvini- Plawen, 1972. 650m, off Cape Hatteras, North Carolina, U.S.A. (34°14.8’N, 75° 46.7'W); 2 specimens (cross sections and radula preparation). (3) Scutopus robustus Salvini-Plawen, 1970. 660 m, Bay of Biscay (48°56’N, 11° 02'W); 4 specimens (sagittal sections, 2 gut dissections, and 1 radula preparation). (4) Limifossor talpoideus Heath, 1904. 508—572 т, Alaska; 3 specimens (type mate- rial) (cross and sagittal sections, whole mount). (5) Limifossor n. sp. 188-195 m, off east Florida, U.S.A. (27°25'N, 79°53'W); 3 speci- mens (cross sections, 2 radula preparations). ALIMENTARY TRACTS IN APLACOPHORA 363 (6) Limifossor ?fratula Heath, 1911. Loca- tion unknown. 1 specimen (sagittal sections; Heath material). Fam. Prochaetodermatidae (7) Prochaetoderma sp. y. (a) 805-811 т, S of Woods Hole, Massachusetts, U.S.A. (39°51.3’М, 70°54.3’W); 6 specimens (radula preparations). (b) 1330-1470 m, S of Woods Hole (39°46.5'N, 70°43.3'W); 21 specimens (9 cross and sagittal sections, 12 radula prep- arations). (c) 1546-1559 m, off Walvis Bay, Namibia, Africa (23%05'S, 12°31’E); 1 speci- men (radula preparation). (8) Prochaetoderma sp. c. (a) 1330- 1470 m, S of Woods Hole (39°46.5'N, 70° 43.3'W); 4 specimens (radula preparations). (6) 2178m, S of Woods Hole (39°38.5’N, 70%36.5'W); 2 specimens (radula prepara- tions). (с) 2091 m, off Scotland (57°59.7’М, 10°39.8’W); 1 specimen (radula preparation). (9) Prochaetoderma sp. p. 1624-1796 m, off Dakar, West Africa (10°30.0’М, 17° 51.5'W); 6 specimens (2 cross and sagittal sections, 4 radula preparations). Fam. Chaetodermatidae (10) Falcidens n. sp. 650 m, off Cape Hat- teras (34°14.8’N, 75°46.7’W); 1 specimen (radula preparation). (11) Falcidens caudatus (Heath, 1911). 1102 т, $ of Woods Hole (39°48.7'N, 70° 40.8'W) and 1330-1470 m, $ of Woods Hole (39°46.5’N, 70°43.3’W); 5 specimens (sagittal and cross sections). (12) Chaetoderma nitidulum Loven, 1844 (= C. canadense Nierstrasz, 1902; Schel- tema, 1973). 74 m, St. Margaret's Bay, Nova Scotia (44°33’01”М, 65°58’09”М/). 4 speci- mens (sagittal and cross sections) and nu- merous radula preparations. (13) Chaetoderma abidjanense Scheltema, 1976. 80m, off Ivory Coast, West Africa (5°02.5'N, 3°47'W); 1 specimen (radula, re- drawn from Scheltema, 1976). Most specimens were fixed as part of an entire washed sample in 10% buffered forma- lin and changed for preservation to 70 or 80% ethyl alcohol within 24hr. Chaetoderma nitidulum was fixed in Bouin's for histologic sections; all others were refixed in HgCl, and acetic acid before sectioning. Stains em- ployed were Delafields’ haematoxylin, with eosin, Gray’s double contrast, or Ponceau S as counter-stains. Radulae were isolated by dissecting out the buccal mass and treating with 5% sodium hypochlorite (household bleach) to remove the tissue. The isolated radulae were washed in distilled water and examined in glycerin using a Zeiss interfer- ence contrast microscope. Drawings were made with the aid of a camera lucida. One radula of a Prochaetoderma species was ex- amined with a scanning electron microscope. COMPARATIVE MORPHOLOGY OF ALIMENTARY TRACTS Mouth The external tissue surrounding the mouth in Aplacophora is usually supplied with mu- cous cells and nerve strands and the mouth is closed by a sphincter muscle. In some Neo- meniomorpha there is a peri-oral fold; in Gymnomenia this fold bears numerous cuti- cular processes, which are extensions of the peri-oral cuticle and presumably receive tac- tile stimuli (Fig. 2M, N). The cuticle of the peri- oral fold is a continuation of the foregut cuti- cle, and both are supplied by large mucous glands or masses of mucous cells (ducts were not clearly seen). The Chaetodermomorpha all have a cuticu- larized oral shield, divided or undivided and more or less surrounding the mouth opening (Fig. 2D, E, J, L). The cuticle of the oral shield is not part of the epidermal, integumental cuti- cle (Hoffman, 1949) (Fig. 2A), but is a thick- ened continuation of the cuticle of the oral tube and buccal cavity in Scutopus, Limi- fossor, and Prochaetoderma (Fig. 2B, С, К). Nierstrasz (1903) noted the same condition in Metachaetoderma challengeri, and Schwabl (1961) considered the oral tube epithelium to be a continuation of the oral shield epithelium in Falcidens hartmani. In Chaetoderma nitidulum and Falcidens caudatus the cuticle of the shield joins that of the oral tube; the latter continues for only a short distance be- fore grading into very dense, long cilia (Fig. 2H, 1), which in turn shorten and continue into the buccal cavity (see also Schwabl, 1961). The epithelial cells underlying the oral shield vary in detail among genera, but certain generalizations seem to hold. There is an abrupt change in epithelial cell type between the cells of the oral shield and those of the oral tube (Fig. 2C, F, arrow); however, the 364 SCHELTEMA cuticle itself appears to be homogeneous ex- cept for a thickened outermost layer of the oral shield and a zone of fibrils running be- tween the epithelial cells of the oral shield and the cuticle (Fig. 2A, 7; C). The epithelial cells of the oral shield contain vacuoles and secretory granules (Fig. 2А, F). In Scutopus the cuticle is pierced by channels and by scattered pyriform mucous cells which are not grouped into lobes (Fig. 2A). In Chaetoderma = z Y Y FIG. 2. Mouth of Aplacophora. A-D: Scutopus megaradulatus; E-G: Prochaetoderma sp. y; mee (be Chaetoderma nitidulum: J: Limifossor n.sp.; К: Limifossor talpoideus, one half of section; М, N: Gymnomenia n. sp.; B, G, H, K, cross-section through oral shield and oral tube, cuticle of gut black, cuticle of integument stippled; D, E, J, L, M, external views of mouth; A, detail of oral shield and integument; C, F, histologic detail of change from oral shield to oral tube (arrow); |, histologic detail of oral tube; N, tactile extensions of peri-oral cuticle. Scales in mm. 1, epidermal cuticle; 2, oral shield cuticle; 3, outer thickened layer of oral shield cuticle; 4, cuticle of oral tube; 5, cilia of oral tube; 6, channel; 7, zone of fibrils; 8, mucous cell; 9, vacuole; 10, epidermal cell; 11, muscle of body wall; 12, epithelial cell of oral tube; 13, muscles; 14, precerebral ganglion; 15, cuticular peri-oral fold; 16, cilia of pedal pit. ALIMENTARY TRACTS IN APLACOPHORA 365 nitidulum the mucous cells form lobes which open at the lateral edges of the oral shield (Hoffman, 1949, and confirmed here). The oral shield seems to serve both in loco- motion and as a sensory organ; it is highly innervated by several precerebral ganglia (Hoffman, 1949; Salvini-Plawen, 1972a) (Fig. 2В). One specimen of Scutopus megaradulatus shows that although the thickest part of the oral shield bends away ventrally from the mouth, it is continuous with and surrounds the mouth opening (Fig. 2D), an observation that does not agree with the original decription (Salvini-Plawen, 1972b). Buccal Cavity Gymnomenia n. sp. As in many Neomenio- morpha, the foregut appears to be suctorial and a buccal cavity as such 1$ not distinct from the rest of the foregut (Fig. 1A). Two sphinc- ters and numerous circular muscles surround the foregut, in addition to the anterior sphinc- ter that closes the mouth. The radula lies be- tween the two posterior sphincters; the poste- riormost one defines the juncture of fore- and midgut. Masses of goblet cells surround the foregut, but there is no ventral pair of salivary glands (Fig. 4F). Between the mouth and the first sphincter the secretory cells are baso- philic; between the first and second sphinc- ters they stain orange (Orange Il counter- stain). Scutopus (S. megaradulatus). The dorsal half to two-thirds of the buccal cavity is lined by tall goblet cells bearing a thick striated cuticle (Fig. 3A). The goblet cells secrete large yellow granules and empty through the cuticle; they occur in all stages of vacuoliza- tion (Fig. 31). A pair of simple tubular salivary glands 150 ит in length lies ventral to the buccal cavity; they empty near their posterior ends laterally into the buccal cavity at the level of the anterior end of the radula. The tip of the buccal mass does not lie free in the buccal cavity; thus there is no sublingual cavity, and the odontophore remains within the main space of the haemocoele (Figs. 1B, 3A). Limifossor (L. talpoideus). The large odon- tophore tip lies free in the cuticle-lined buccal cavity, and there is a spacious sublingual cavity (Figs. 1C, 3B). Tall goblet cells with large yellow granules similar to those in Scutopus line the buccal cavity laterally and dorsally; ventrally the goblet cells are scat- tered. A pair of tubular salivary glands emp- ties dorsally into the buccal cavity near the radula tip; they originate anteriorly and rather far ventrally (arrow, Fig. 3B). Prochaetoderma (spp. y, р). The spacious buccal cavity is lined by a thick cuticle (Fig. 3D, E). The epithelium is formed of medium- high columnar to cuboidal cells filled with fine granules; some have a single large yellow secretory body with or without connection to a vacuole (Fig. 3K). The anterior part of the buccal cavity is dominated by a pair of cuticu- lar jaws which hold the mouth open during feeding (Kowalevsky, 1901; unpublished data) (Figs. 1D, 3D). The jaws are abutted by the epithelium of the buccal cavity; laterally they lie directly against basement membrane (Fig. 3H). Thus, they are not part of the buccal cavity cuticle as reported by Schwabl (1961) and are not homologous to gastropod jaws but are unique structures among the mol- luscs. The bases of the jaws lie wholly within the haemocoele (Fig. 3F; cf. Fig. 7B, C). At the point where their bases join the long an- terior ends, the jaws pierce through the buccal cavity wall (Figs. 1D, 3E, G). The cuticle of the jaws is perhaps secreted at the point where the jaws are abutted by the epithelium of the buccal cavity, as indicated by a change in epithelial cell type and by the direction of the striations in the jaws (Fig. 3G, H). Although the major part of the jaws lies within the buc- cal cavity, they appear to have originated as part of the odontophore mass in the haemocoele (see below under Radula). The tip of the odontophore lies free in the buccal cavity, and the lateral buccal pouches are deep (Fig. ЗЕ). А pair of salivary glands with compound tubules opens dorsally near the beginning of the short esophagus. Chaetoderma (C. nitidulum). The epitheli- um of the spacious buccal cavity is formed by tall, brush-bordered columnar cells containing fine granules; there are also scattered goblet cells with large yellow secretions (Fig. 3C, J). Two pairs of salivary glands with compound tubules open laterally and dorsally into the buccal cavity, one pair at the level of the tip of the radula, the other just anterior to the esophagus (as reported by Wirén, 1892). The odontophore lies free in the buccal cavity for one-half or more of its length. Falcidens (F. caudatus). The buccal cavity is similar to that of Chaetoderma, but is less capacious. There are perhaps also two pairs of salivary glands; however, ducts were not clearly seen for the dorsalmost pair. SCHELTEMA 366 a SUB AY D ALIMENTARY TRACTS IN APLACOPHORA 367 Radula The aplacophoran radula has been shown throughout the literature to be very diverse in form, and far more plastic than the gastropod radula. However, there are certain structures common both to aplacophorans and other mollusks. All isolated radulae that | have studied, ex- cept those of the Chaetodermatidae, have a discrete radular membrane with attached distichous rows of teeth issuing from a radular sac, which is a diverticulum of the buccal cavi- ty known to secrete the radular membrane and teeth in gastropods (Fretter 4 Graham, 1962) and appearing to do so in Aplacophora. There is no evidence for a primitive so-called “basal membrane” that is part of the foregut cuticle and different in some way from a true radular membrane (Boettger, 1956; Salvini- Plawen, 1972a). The radular membrane 1$ supported by an odontophore which lies in the haemocoele. There are one or more pairs of bolsters formed of connective tissue and muscle, or of chondroid tissue, or perhaps of collagen and muscle; in one case there 1$ cuticularization. Protractor and retractor muscles run between the odontophore mass and the body wall, and presumably all aplacophoran radulae can be protracted to, or through, the mouth. In gastropods, muscles that run between a subradular membrane and the bolsters move the radula itself (Graham, 1973); in Aplacophora a subradular mem- brane is usually, but not always, lacking. The radula musculature has been described for only a few aplacophoran species and will not be described here except for a few particular cases. А bending plane may be either present or lacking; if present, there is no fixed position along the odontophore from genus to genus where it is situated. Gymnomenia n.sp. The tiny radula of Gymnomenia was overlooked in the original description of the genus (Odhner, 1921) (Fig. 4); it is considered to be secondarily reduced by Salvini-Plawen (1978). There are about 28 rows of hooked distichous teeth, each with two median denticles in various stages of be- ing tanned. None of the teeth show wear. Each tooth is attached to the radular mem- brane for one-half its length (Fig. 4B, С). In interference contrast, the radular membrane was seen to be continuous (a) between the teeth of each row as a slight ridge (Fig. 4B, C), which in turn runs down the length of the radula; (b) along and slightly below the base of the teeth lengthwise along the radula (Fig. 4A); and (c) lengthwise along the radula at the level of the denticle in the middle of each tooth (Fig. 4D). Thus, the radular membrane 1$ a continuous sheet which appears to be fused medially; it bears two longitudinal rows of well-affixed teeth. Teeth attached so firmly to the radular membrane can have only limited movements. The orientation of the radula is similar to that described for Сепйосота (Salvini- Plawen, 1967a). The fore-end of the radula 1$ positioned dorsoventrally, where it lies in a blind sheath (Fig. 4E, F). About two-thirds of the distance towards the newest formed teeth in the radular sac there is a bending plane, over which the teeth open into the foregut. The short dorsal radular sac is perhaps bifid as in other Wireniidae, as indicated by the medial ridge of the radular membrane, but further histologic material is needed for sub- stantiation. The base of the fore-end of the radula lies against the connective tissue (? and commis- sure), defining the pedal sinus (Fig. 4F); di- rectly beneath this, in the sinus, are about seven calcareous statoliths, each produced by a statocyst (Fig. 4G). In Genitoconia, Salvini-Plawen (1967a) described a pedal commissure sac with vesicles which he con- sidered perhaps to be a balancing organ (“ein statisches Organ”). The odontophore protractors and retractors have been described for Genitoconia (Salvini- Plawen, 1967a), but the exact manner in which they operate the radula is not clear. The function of the enclosed fore-end of the radula ——— FIG. 3. Buccal cavity in Chaetodermomorpha. A, |: Scutopus megaradulatus; В: Limifossor talpoideus, arrow indicates level of blind end of salivary gland; С, J: Chaetoderma nitidulum; D-H, К: Prochaetoderma sp. y. Scales in mm. 1, bolster; 2, salivary gland; 3, opening of salivary gland into buccal cavity; 4, radular tooth (diagrammatic); 5, radular membrane; 6, subradular membrane; 7, jaw; 8, ventral approximator of bolsters; 9, ventral approximator of jaw; 10, tensor between bolster and base of jaw; 11, lumen of odonto- phore; 12, buccal pouch; 13, cuticle of buccal cavity; 14, dorsal cuticular membrane between distal end of jaws; 15, basement membrane; 16, deeply staining portion of jaw cuticle; 17, goblet cell; 18, large yellow granule; 19, sublingual cavity; 20, esophagus. 368 SCHELTEMA A sian) 20 Vi | Е FIG. 4. Radula and statoliths of Gymnomenia п. sp. A-D: radular teeth; A, lateral view; В, С, anterior view; D, median view at level of middle denticle indicated by arrow on A; E: radula, lateral view, dorsal at top, anterior to right, most recently formed teeth to left, teeth between arrows exposed in pharynx; F: cross-section showing exposed teeth in foregut and sheath around fore-end of radula resting against statocyst. G: two statoliths and a statocyst cell filled with amorphous substance. Scales in mm, А-О at same scale. 1, radular membrane; 2, radula; 3, foregut; 4, paired dorsal caecum of midgut; 5 goblet cells; 6, statocyst; 7, pedal sinus; 8, pedal pit. is not known; it may act as a supporting rod- like structure. The proximity of the radula to the statocysts of the pedal sinus may or may not indicate a direct relationship between them. The exposed teeth perhaps are able weakly to tear at soft tissue as it is sucked into the foregut, and very probably serve to move food backwards toward the midgut. Scutopus (S. robustus, S. megaradulatus). The radula is formed of seven or more pairs of teeth in a straight, nearly anteroposterior posi- tion, with the distal ends of the teeth lying anteriorly to the proximal ends (Figs. 5A, B, 11D). Thus, the older of any two rows of teeth lies beneath and anterior to the younger, and the odontoblasts lie on the dorsal side of the radular sac. There is no bending plane. The teeth of $. megaradulatus and $. robustus are thick and massive, with pointed tips and many large median denticles which curve ventrally and posteriorly (Fig. 5D). From histo- logic sections and whole preparations the radular membrane was seen to be formed of two longitudinal bands connected only be- tween each pair of teeth (Fig. 5C). Each band extends laterally along the side of each tooth, but these extensions are not connected (Fig. 5C, D). Thus, the teeth are free to slide past each other, perhaps moved by the ventral tensor (Fig. 5A); they can also be closed by a large dorsal approximator muscle running be- tween the anterior pair of bolsters (Fig. 5A, B). These rather limited possibilities for move- ment combined with the absence of a bending plane and of a sublingual cavity suggest a simple shovelling or pulling in of food. The radula probably cannot be protruded very far beyond the end of the radular sac, and the teeth do not show any wear. Limifossor (Limifossor п. зр., L. talpoid- eus). The radula with its massive odonto- phore was well described by Heath (1905), ALIMENTARY TRACTS IN APLACOPHORA 369 FIG. 5. Radula of Scutopus. A, B, D: S. robustus; C: S. megaradulatus. A, lateral view of buccal mass, anterior to left, teeth beyond radular sheath black; B, dorsal view, anterior end at top; C, diagrammatic representation of radular membrane; D, pair of radular teeth. Scales in mm, A and B at same scale. 1, midgut epithelium; 2, ventral tensor muscle; 3, protractors; 4, dorsal approximator of bolsters; 5, radular sac; 6, anterior bolster; 7, posterior bolster; 8, anterior limit of radular sac; 9, radular membrane connecting pair of teeth; 10, radular membrane attached to base of teeth; 11, lateral extension of radular membrane; 12, odontoblasts of radula. who illustrated the musculature and watched the radular movements of living animals. A few observations may be added to his. As Heath noted, the radular membrane is a continuous sheet only at the posterior end of the radula; farther anteriorly it splits along the midline (arrow, Fig. 6B) and continues as two bands. The radular membrane extends for a short distance up the lateral side of each tooth, and these lateral extensions are in con- nection along the radula (Fig. 6A, C). The teeth are massive, with long lateral hooks and shorter median hooks turned posteriorly (Fig. 6C, D); the bases have a thickened ridge posteriorly. The teeth are set close to each other along the radula; each tooth thus ap- pears to act as a fulcrum for the next posterior one (Fig. 6C, D). There is a large mass of muscle fibers, the tooth adductor, that Heath (1905) found to be responsible for moving the opposed teeth toward each other (Fig. 6B); there is a dorsal approximator of the bolsters, present as in Scutopus, which is also prob- ably important in bringing the two rows of teeth together (Fig. 6B). None of the teeth show wear. 370 SCHELTEMA FIG. 6. Radula of Limifossor n. sp. A: lateral view of radula, dorsal at top, anterior to left; B: dorsal view of radula and odontophore, anterior at top; arrow indicates point where radular membrane splits into two bands; С, D: oblique and dorsal views of teeth in natural position. Scales in mm, А-В and C-D at same scales. 1, radular membrane; 2, subradular membrane?; 3, odontophore mass; 4, anterior limit of radular sac; 5, attachment of dorsal approximator of bolsters; 6, area of large tooth adductor muscle. A certain amount of rotation of the teeth is possible, as shown in Figure 6B. These movements are made possible by (a) the median split in the radular membrane, which frees the two longitudinal rows anteriorly, (b) the use of each tooth as a fulcrum by the next posterior one, (c) the presence of a rudimen- tary bending plane, (d) the possible existence of a subradular membrane for tensor inser- tion, and (e) a deep sublingual cavity, which frees the entire buccal mass from the haemo- coele (Fig. 3B). Heath reported that the odontophore swept past the teeth when the mouth was open. Certainly the radular teeth are able to go through a more complicated set of movements than can those of Scutopus or Сутпотета. Less certain is whether the radula is used for tearing or simply is an im- proved form of rake. Prochaetoderma (spp. у, р, с). Kowalevsky (1901) figured the isolated distichous radula of Prochaetoderma raduliferum; some details can be added. Most noticeable are the two large cuticular jaws that nearly fill the space in the head (Fig. 7A-C). Anteriorly they are connected by a membrane (Figs. 3D, 7C); posteriorly within the haemocoele a large bundle of muscle fibers runs between their bases, and a small fiber runs between each base and the chon- АНМЕМТААУ TRACTS IN APLACOPHORA 371 ¿OS FIG. 7. Radula of Prochaetoderma. A-F: Prochaetoderma sp. y; G: Prochaetoderma sp. р. A, lateral view of jaws, radula, and odontophore, anterior to left; B, same as A, tissue removed; C, dorsal view of B, anterior at top; D, base of tooth, with lateral tooth-like extension and radular membrane in darker stippling; E: oblique view of central plate, or tooth, in relation to proximal ends of teeth; F: row of central plates, anterior at top; G: single radular tooth with membranous denticulate medial brush and lateral membranous wing, attachment to radular membrane between arrows. Scales in mm, А-С and D-F at same scales. 1, cuticular jaw; 2, radular sac; 3, tensor muscles between jaws; 4, chondroid bolster; 5, toothlike lateral extension of radular mem- brane; 6, subradular membrane; 7, case of jaw which lies within haemocoele; 8, radular membrane; 9, membranous wing of radular tooth; 10, dorsal cuticular membrane. 372 SCHELTEMA droid bolster immediately dorsal to it (Fig. 3F). The jaws serve to hold the mouth open, and the radula is protruded between them. The musculature which protracts and retracts the jaws has not been described. The radular membrane is a continuous sheet to which only the tips of the proximal ends of the teeth are attached (Fig. 7D, G; observation substantiated by scanning elec- tron microscopy). Laterally the membrane 1$ drawn out into a tooth-like extension beside each tooth; the extension is not attached to the tooth but appears to support it in some manner (Figs. 7B, D, 11A). A bending plane lies at the anterior end of the odontophore (Fig. 7B). Uniquely in the Aplacophora, a central plate, or tooth, lies between the bases of each pair of teeth (Fig. 7E, F). The four to six pairs of anterior teeth are crossed and used in feeding (Figs. 7C, 11C); the anterior- most pair are worn (Fig. 11B). The posterior teeth, which probably remain within the radu- lar sac, seem to function as a backstop for food particles carried between the membran- ous median brush-like extensions of the teeth (Figs. 7G; 11C). There is a subradular mem- brane, distinct from the radular membrane (Figs: SE; 7B): The radula of Prochaetoderma appears to reduce the size of food material by rasping before ingestion on the following evidence: (a) the mouth can be held open by the jaws prob- ably independent of radular protrusion; (b) the teeth can articulate, for they are free of the radular membrane laterally and there are median supportive teeth; (c) the chondroid tissue of the bolsters provides a stiff structure to work beneath the protruded radula; (а) there is a bending plane at the anterior end of the odontophore over which the teeth can be articulated; (e) the anterior teeth are worn. Kowalevsky (1901) described the protracted, anterior crossed teeth in living animals as pro- jecting through a wide-open mouth and con- stantly in motion as if to seize something. Falcidens and Chaetoderma (several spe- cies). The very specialized radulae of the Chaetodermatidae (Fig. 8) have already been described in detail from isolated preparations (Scheltema, 1972). Paired denticles or lateral projections attached to the end of a cone- shaped rod presumably act as grasping pin- cers; tensors run between them and the bolsters (Schwabl, 1961; Ivanov, 1979) (Fig. 8C). There are three published interpretations of the cone-shaped structure: it represents a fused radula (Scheltema, 1972); it is a greatly thickened basal membrane (Salvini-Plawen, 1972a); it is one of three teeth of a mono- segmental radula (Ivanov, 1979). The cone lies within an epithelial sheath, perhaps the radular sac (Fig. 8C), and is secreted at its thick, ventral end. The identity of growth lines in the cone of Chaetoderma with those in Prochaetoderma jaws is highly improbable (Salvini-Plawen 8 Nopp, 1974), for the jaws appear to be a part of the odontophore mass FIG. 8. Radulae of Chaetoderma and Falcidens. A: Falcidens n.sp.; В: С. abidjanense (from Scheltema, 1976); C: C. nitidulum, histologic sec- tion. Scales in mm. 1, cuticle surrounding buccal mass (subradular membrane?) 2, cone-shaped tooth; 3, lateral projection; 4, tensors of lateral pro- jection; 5, bolster; 6, epithelium surrounding cone- shaped tooth (radular sac?). ALIMENTARY TRACTS IN APLACOPHORA 373 within the haemocoele and are not underlain by epithelium (cf. Figs. ЗО, E, F, 8C). Esophagus The esophagus 15 defined as that рай of the foregut forming a tube above the radula and connecting the buccal cavity and stomach. lts epithelium 15 differentiated from the epitheli- um of both the buccal cavity and the stomach. Gymnomenia п. sp. An esophagus is lack- ing in Gymnomenia but not in all Neomenio- morpha, although Odhner (1921) considered the posterior pharynx of G. pellucida to be an esophagus. Scutopus (S. megaradulatus). The buccal cavity opens dorsally into a short, wide esophagus formed of low cuboidal epithelium with a brush border; the cells are filled with fine, yellow granules. A short distance poste- riorly the lateral walls acquire folds, and the ventral wall thickens. The folds merge dorsal- ly and become ciliated; posteriorly they coa- lesce into a typhlosole that continues into the stomach. Limifossor (L. talpoideus, L. ?fratula). As described by Heath (1905), the esophagus is a long, ciliated, narrow tube with several longitudinal folds; the cells are filled with granules. Dorsally the ciliated epithelium is continued into the stomach. Prochaetoderma (sp. y, p). The esophagus is extremely short and bears no cilia; how- ever, Schwabl (1963) reported а ciliated esophagus in P. californicum. Falcidens (Е. caudatus). The esophagus 1$ discernible from the buccal cavity only by its long, slender goblet cells which lie between the buccal cavity and stomach. Chaetoderma (C. nitidulum). Cells with a brush border line a muscular esophagus. At the entrance to the stomach there are very long, slender goblet cells but no cilia; these coalesce dorsally and become the ciliated typhlosole in the stomach. Midgut Neomeniomorpha. The stomach 1$ a single wide tube interrupted laterally at regular inter- vals by the dorsoventral musculature (Fig. 1A). In most Neomeniomorpha there is an antero-dorsal paired or unpaired caecum (Fig. 4F). The cell types of the midgut are not described here. A dorsal ciliated tract or fold runs the length of the midgut and leads into a short, posteriorly placed, ciliated intestine (Pruvot, 1891; Salvini-Plawen, 1978); it was not seen in Gymnomenia, however. Chaetodermomorpha. All members of the Chaetodermomorpha investigated here have a stomach, a sac-like ventral digestive diver- ticulum that opens into the posteror end of the stomach, and a long ciliated intestine that fol- lows a bend in the posterior stomach. Nier- strasz (1903) reported that Metachaetoderma challengeri lacked a separate midgut gland in the one incomplete specimen he examined, but this observation needs to be repeated. Except in Prochaetoderma there is either a dorsal ciliated typhlosole or a groove that runs down the stomach to the ciliated intestine. The epithelial cells lining the stomach are homogeneous and contain granules; cell shape varies among genera and species. The cells of the digestive gland are unique among mollusks: the dorsal wall is lined by a band of cells packed with coarse yellow granules (lacking in Prochaetoderma) (Fig. 10H); laterally and ventrally are cells which secrete large basophilic spheres (Fig. 10G). A mucoid or proteinaceous rod 15 present in all genera except Prochaetoderma, but its position in the gut varies. Scutopus (S. megaradulatus, S. robustus). The stomach is long and divided by septa which do not run its entire length. These do not appear to be the same as the outpouch- ings that Salvini-Plawen (1972a, fig. 16) Шиз- trated for S. ventrolineatus related to dorso- ventral musculature. The granular cells of the stomach epithelium have a striated or brush border; anteriorly they are low and cuboidal, but farther posteriorly they become high and club-shaped (Fig. 10A, B). A strip of the stomach epithelium passes into the digestive diverticulum and continues there as a dorsal band of granular cells with greatly coarsened granules (Figs. ЭА-С, 10B, H, 11D). A dorsal ciliated typhlosole (Fig. 10A, C) runs from the esophagus to the intestine; there is a second ciliated typhlosole arising at the base of the stomach that also runs to the intestine. A patch of ciliated cuboidal cells with densely staining borders opposes the bend that joins stomach and intestine (Fig. 9A, |). In three specimens out of a sample of 19 $. robustus, the stomach epithelium was nearly colorless owing to the lack of cell granules. Dissection of two of these colorless speci- mens revealed that the stomach contained several solid, proteinaceous (stained by rose Bengal), acellular, parallel rods which were presumably formed by secretions from the 374 SCHELTEMA АНМЕМТААУ TRACTS IN АРЕАСОРНОВА 375 septate stomach (Fig. 11Е). Crystals of about 40 um adhered to the outsides of the rods; these crystals became more densely packed posteriorly. Both rods and crystals passed into the anterior intestine; farther posteriorly the crystals, but not the rods, formed part of a fecal mass. The crystals may be organic, as they dissolved in dilute НС! (but not NH¿OH) and broke down into an amorphous yellow mass when subjected to pressure by squeez- ing them beneath a glass coverslip. In S. megaradulatus sections, the stomach was empty and the stomach cells were packed with granules; there was only a short mucoid rod at the anterior end of the intestine. Fecal material in Scutopus is formed into a long spindle-shaped mass along a straight intestine (Fig. 1B). Limifossor (L. talpoideus, L. ?fratula). The epithelium of the very short (L. talpoideus, Fig. 1C) or very long (L. ?fratula) stomach are formed of tall (former) or short (latter) cuboidal granular cells with a striated or brush border. A dorsal typhlosole runs from the esophagus to the intestine. At the posterior bend between the stomach and intestine in L. talpoideus the cells are thickly ciliated (?brush border) and have a thick amorphous border resembling cuticle (Fig. 9D, J). Within the anterior ciliated intestine (interpreted originally as a style sac, Scheltema, 1978) is a mucoid rod (Fig. 9D, E). The digestive diverticulum is long. Fecal material is formed into oblong masses along a straight intestine. Prochaetoderma (spp. y, p). The stomach is lined by low cuboidal cells probably with a cuticular border; cilia are lacking. The short digestive diverticulum lacks a dorsal band of granular cells and is formed only of secretory cells which are modified from the type found in other chaetoderms; the cell granules are eosinophilic and there are few basophilic, spherical secretions. There is no mucoid rod. The short anterior section of the intestine may — бе bent ог straight. Fecal material is formed into discrete spherical masses strung out along a long, convoluted intestine (Fig. 1D). Chaetoderma (C. nitidulum). The stomach and digestive diverticulum are long (Fig. 1F). A dorsal typhlosole starts just posterior to the esophagus and runs the length of the stom- ach and into a style sac (Fig. 9N); only the medial cells of the typhlosole are ciliated anteriorly (Fig. 10E). The stomach epithelium is formed of low cuboidal cells with a striated or brush border; the granular cells of the dorsal band in the digestive diverticulum are very tall with a striated border and were not seen to be in connection with the stomach epithelium. At the base of the stomach there is a thick, hooklike cuticular gastric shield un- derlain by tall columnar cells which are gran- ule-filled distally and striated basally; fibrils run between the cuticle and cell walls (Fig. 9L, N). The ciliated style sac runs between the stomach and the intestine transversely to the long axis of the body; it contains a mucoid rod in some specimens (Fig. 9M) (see also Scheltema, 1978, fig. 1В). The rod appears to rotate against the gastric shield, inasmuch as food material between the rod and the shield occurs in spiral swirls. The style sac is formed of granular cells with dense, short cilia; a broad ridge borders a groove with longer cilia which continues into the intestine (Fig. 9F-H, M). Fecal material is formed into oblong masses; the intestine is straight. Falcidens (F. caudatus). The stomach is short and bilobed (Fig. 1Е); its epithelium is formed of low cuboidal cells with yellow granules and a cuticular border (Fig. 10D). A strip of these cells continues, without a cuti- cular border, into the digestive diverticulum where it becomes the dorsal band of granular cells of that organ. The digestive diverticulum extends broadly to where the body narrows into a “tail.” A dorsal ciliated groove, rather than typhlo- FIG. 9. Posterior stomach, anterior intestine, and opening of digestive diverticulum in Chaetodermomorpha. AC, |: Scutopus megaradulatus; D, E, J: Limifossor talpoideus; F-H, L-N: Chaetoderma nitidulum; К: Falcidens caudatus. А-С, cross-section from anterior to posterior, viewed anteriorly; О, E, nearly adjacent sagittal sections, anterior to right; F-H, M, oblique sections through style sac of two specimens, one showing groove (enlarged in G) running into intestine (H) and one with protostyle (М); I-L, morphocline of cells from ciliated to cuticularized at bend between stomach and intestine (| enlarged from A, 8; J from D, 8; К from М, 7). М, oblique view anterior to M at junction of stomach and style sac, showing gastric shield. Scales in mm, A-C, DE, I-L, M-N, at same scales. 1, dorsal band of granular cells; 2, secretion cells with basophilic spheres; 3, intestine; 4, stomach; 5, bend between stomach and intestine; 6, style sac; 7, gastric shield; 8, specialized cells at bend between stomach and intestine; 9, style sac ridge; 10, style sac groove; 11, dorsal typhlosole; 12, mucoid rod (protostyle); 13, bolus entering intestine; 14, gonad. 376 SCHELTEMA FIG. 10. Cells of alimentary tract in Chaetodermo- morpha. A-C, H: Scutopus megaradulatus; D: Falcidens caudatus; E-G: Chaetoderma nitidulum. A-F, granular stomach cells and dorsal ciliated typhlosole or groove (B, C posterior to A); G, diges- tive gland secretory cell with basophilic sphere; H, cell from dorsal band of granular cells of digestive gland. sole (Fig. 10D), starts about half way down the length of the stomach and leads to a hook- like gastric shield (Scheltema, 1978, fig. 1C), and thence continues into a style sac. A sec- ond, ventral ciliated band starts at about the level of the gastric shield and joins the dorsal typhlosole to form a style sac with a mucoid rod and ciliated ridge bordering a groove. The rod appears to rotate against the gastric shield. Schwabl (1961) described and figured schematically the gastric shield and style sac for F. hartmani without considering them as such, although referring to the style sac as “caecum-like”; a mucoid rod is not mentioned. The cells underlying the gastric shield are cuboidal with large granules distally (Fig. 9K). The style sac is transverse to the body axis (Fig. ТЕ). The intestine is convoluted and filled with spherical fecal masses. Diet The diet of the species under discussion 1$ based on stomach contents. Not available to me at this time of writing is Salvini-Plawen's work (in press) on diet (see Literature Cited). Сутпотет п. sp. As in most Neomenio- morpha, there are many unexploded nemato- cysts within the cells of the midgut; Gymno- тета is therefore considered to feed оп Cnidaria. Scutopus. The diet is not known; fecal material contains organic (?) crystals and perhaps sediment particles. The radula morphology suggests that the diet is particle- size dependent, probably detritus. Limifossor. The diet is not known. Fecal material contains very small bits of unidenti- fied frustules, spicules and other hard parts of organic origin. Although Limifossor has usual- ly been considered a carnivore (Heath, 1905; Salvini-Plawen, 1975), it seems quite as likely from radula morphology that it is a detritivore and possibly particle-size dependent. Prochaetoderma. The diet seems to be a wide variety of both prey and organic debris. The stomach of several specimens hold Foraminifera with sand tests (?Saccorhiza), crustacean parts, radular teeth of smaller Prochaetoderma, and bits of unidentified organic remains; much of the food material still contains stained cytoplasm. There are very few sand grains. Chaetoderma and Falcidens. The Chaetodermatidae are considered to be se- lective carnivores, taking in entire Foramini- fera, “worms,” small snails and other uniden- tified organisms which are found in the stom- ach with stained cytoplasm. Ivanov (1979) has figured the action of feeding. C. nitidulum can be a contaminant in laboratory cultures of living Foraminifera upon which they will feed (В. Christensen, personal communication). It is not known whether members of the Chaetodermatidae also feed on organic debris. There are few sand grains in the gut. ALIMENTARY TRACTS IN APLACOPHORA 377 FIG. 11. Radula of Prochaetoderma sp. y (А-С) and alimentary tract of Scutopus robustus (D, E). A: lateral tooth-like extensions of radular membrane; B: worn anterior pair of denticles; C: food material caught in crossed pairs of anterior teeth, held against posterior 6-7 pairs touching at distal tips; note darkened, tanned distal tips; D: entire preserved specimen in transmitted light; E: proteinaceous rods with adhered crystals dissected from stomach and intestine of a specimen without dark granules evident in D. Scales in mm. 1, radula; 2, darkly pigmented stomach; 3, band of granular cells passing from stomach to dorsal wall of digestive gland; 4, rods from anterior stomach; 5, rods and crystals at base of stomach and entrance into intestine; 6, fecal material. 378 SCHELTEMA DISCUSSION Phylogenetic Considerations: Intraclass The Neomeniomorpha and Chaetodermo- morpha are considered by me to be sub- classes belonging to the class Aplacophora (Scheltema, 1978) and their great specializa- tion of acquiring a worm shape to have evolved as a single event before evolution of the molluscan shell (i.e., a shared derived character state). The chaetoderm oral shield was thought by Hoffman (1949) to be homol- ogous to the outer wall of the ventral foot fur- row of the neomeniomorphs (homology is not with the foot sole, Scheltema, 1978). This homology is not substantiated by the observa- tion that the chaetoderm oral shield is formed from cuticularized gut epithelium that has come to lie externally like lips (Fig. 2). More- over, the mucous cells of the oral shield in the primitive species Scutopus megaradulatus are diffuse and do not occur in lobes as re- quired by this homology (Hoffman, 1949). Therefore the separation of the Chaeto- dermomorpha (Caudofoveata) from all other mollusks as the most primitive molluscan class on the basis of this homology is not up- held (see Salvini-Plawen, 1972a, paragraph 16). In considering which character states of the alimentary tract may be primitive and which may be derived among the Aplacophora (Table 1), the following assumptions are made: (a) the Aplacophora are the sole living representatives of the primitive pre-placo- phorous mollusks and geologically very old (see Stasek, 1972; Salvini-Plawen, 1972a; Scheltema, 1978); (b) the least differentiated character state is usually the most primitive, unless there is some evidence for loss of structure; (c) a character state shared by most or all members is usually primitive, unless some evidence points to the contrary; (d) the radula capable of the least amount of manipu- lation is most primitive. From the table certain relationships are clear (numbers below refer to character num- ber in the table). The primitive character states held in common between the two sub- classes are: cuticularization of the foregut (1); paired tubular salivary glands (lying ventrally both in Scutopus megaradulatus and in most Neomeniomorpha; perhaps secondarily lack- TABLE 1. Primitive and derived character states of the aplacophoran alimentary tract (C = Chaetoderma, F = Falcidens, G = Gymnomenia, L = Limifossor, P = Prochaetoderma, S = Scutopus). Derived Character Primitive (a, b, independently derived) Oral shield (Neomeniomorpha not considered) Entire Divided SEC E Buccal cavity 1. Cuticle Present Absent CASA ENG 2. Goblet cells Dominant Scattered Grom PREG 3. Tubular salivary glands One pair (a) Two pairs (most Neomeniomorpha) (F?), С SAP (b)! Lacking G 4. Buccal sublingual pouch Absent, or пеапу so Present Gis LPH EAC Radula 5. Radular membrane Divided or partially so, or Entire line of fusion PE? 162) (CAS 6. Subradular membrane Absent Present а, $ (2) Р.Е? ©?) АНМЕМТААУ TRACTS IN APLACOPHORA TABLE 1 (Continued). 379 e | | a, Character Primitive Derived (a, b, independently derived) nn 7. Dentition 8. Relationship of teeth to radular membrane 9. Bolster tissue 10. Cuticular structure derived from odontophore 11. Dorsal approximator of bolsters Esophagus 12. Length 13. Ciliation Midgut 14. Ciliated dorsal band, groove, or typhlosole 15. Digestive diverticulum Chaetodermomorpha only: 16. Dorsal granule cells, digestive diverticulum 17. Lining of stomach 18. Gastric shield 19. Style sac 20. Mucoid or protein rod(s) Feeding, diet 21. Feeding type 22. Particle size Distichous, without central plate GASAE Not articulated GYSAEL Connective tissue, muscle (GASTE CG Absent GSM С Present (primitive?) SA Long, short ЕС Ciliated Sul Present (nearly all Neo- meniomorpha) SEG Absent (primitive?) G (and all other Neo- meniomorpha) Present SALAENGC Not cuticular SIC Absent SALER Absent SAP Present throughout stom- ach and anterior intestine 5 Detritivore-omnivore SEL Dependent S, (L?) F, C (a) Distichous, with central plate P (b) Reduced FG (a) Articulated P (b) Reduced EXC Chondroid-like Р Present P Absent (derived?) СИР Е © (a) Extremely short PAR (b) Absent (derived?) G Not ciliated PREG Absent (G?), P Present (derived?) S, L, P, F, C (and all other Chaetodermomorpha) Absent P Cuticular (P2)E Present EXC Present EG (a) Present, restricted location ЕЕ, © (b) Absent P Selective carnivore С.Е. С Independent (a) Suctorial G (b) Rasping P 1Considered primitive by Salvini-Plawen (1978). 380 SCHELTEMA ing in Gymnomenia) (3); a distichous radula lacking articulation (7, 8); a divided or fused radular membrane and lack of a subradular membrane (5, 6); and a ciliated dorsal band, groove, or typhlosole that runs the length of the midgut to a ciliated intestine (14). The foregut goblet cells (2) may not be homolo- gous (cf. Figs. ЗА, В, 4F). The greatest difference between the two subclasses lies in the presence or absence of a digestive diverticulum (15). The undivided midgut of the Neomeniomorpha has been interpreted as primitive on the basis of (a) the lack of digestive adaptations (digestive gland, protostyle, gastric shield) for microphagous feeding (Salvini-Plawen, 1980) and (b) the presence of regular outpouchings caused by serially arranged dorso-ventral musculature (Boettger, 1956; Salvini-Plawen, 1969). (These outpouchings were first considered to be primitively lacking in Genitoconia, а тет- ber of the Wireniidae which includes Gymno- menia [Salvini-Plawen, 1967a], but later the lack of lateral pouches was considered to be secondarily derived [Salvini-Plawen, 1978]). Most neomeniomorphs have a very special- ized cnidarian diet and thus the undivided midgut may be a specialized or reduced state, and not a primitive one. The single digestive diverticulum of the chaetoderms appears to have developed as a lobe from the stomach; it retains the evidence of its origin in the dorsal band of granular cells which can be traced forward to the stomach epithelium. Among the Chaetodermomorpha there are two lines of evolutionary change from the least differentiated and therefore presumed primitive state found in Scutopus. One direc- tion has been toward increased elaboration of the stomach into a posterior style sac, restric- tion of the protostyle to this sac, and in- creased cuticularızation at the base of the stomach to form a gastric shield; morpho- clines of these character states exist from Scutopus through Limifossor to Falcidens and Chaetoderma (Fig. 9). The other direction has been toward reduction as found in Prochaetoderma, with a single type of diges- tive cell in a shortened digestive diverticulum, no dorsal ciliated typhlosole, and no proto- style. The gastric shield ıs not correlated with general cuticularization of the stomach epi- thelium (Table 1: 17, 18). A convoluted intes- tine is found independently in the two genera that have long, thin ‘tails,’ Prochaetoderma and Falcidens (Fig. 1D, E). The aplacophoran radula has evolved to- wards freeing the teeth from their primitively broad attachment to the radular membrane and toward development of a sublingual pouch (4, 8). The result has been increased ability to manipulate or break down the food source. In Gymnomenia the radula appears to be one of the most primitive among the Aplaco- phora (Fig. 4), but much work remains to be done on the diverse radular types found in other Neomeniomorpha (Nierstrasz, 1905; Salvini-Plawen, 19676, 1978). Among cnidarian feeders with a suctorial foregut, reduction and specialization could be expect- ed; nevertheless, a primitive type of radula occurs in carnivores in the Aplacophora. The radulae among chaetoderm genera dif- fer greatly in morphology and cannot readily be derived from a primitive type or from each other except in terms of function. Primitively, teeth are affixed to the radular membrane and the odontophore 1$ scarcely free in the buccal cavity; only a sliding motion combined with closing opposed teeth 1$ possible (Scutopus, Fig. 5). More complicated movements can occur in Limifossor with a split radular mem- brane and a relatively enormous odontophore (Figs. 1C, 6B; Heath, 1905). A rasping gastropod-like radula has evolved only т Prochaetoderma (Fig. 7). The reduced, highly modified radula of Falcidens and Chaeto- derma 1$ probably capable of precise move- ment in prey capture (Ivanov, 1979). The two most highly evolved radulae occur in the two groups which are carnivorous or carnivorous- omnivorous and which also have the most modified midguts: Prochaetoderma with the most complex radula and most reduced mid- gut and the Chaetodermatidae (Falcidens and Chaetoderma) with the most modified radula and most complex midgut. There does not appear to be a morphocline in radula type in the Chaetodermomorpha (see Salvini- Plawen, 1975). Phylogenetic Considerations: Interclass The style sac and gastric shield are shown by the Aplacophora to have evolved more than once in the Mollusca. In the Aplaco- phora, a protostyle has evolved before a style sac, and a style sac and gastric shield occur only in a carnivorous family (Chaetodermati- dae). A radula capable of rasping seems to re- quire a single radular membrane, a subradu- lar membrane, a bending plane, firm bolsters, ALIMENTARY TRACTS IN APLACOPHORA 381 and some way for the teeth to articulate on the radular membrane. There also must be some way to keep the mouth open during rasping. In gastropods the mouth opens as part of radula protraction (Graham, 1973), but in Prochaetoderma, which uses its head for locomotion (burrowing) as well as for feeding, unique jaws have evolved which can keep the mouth open during rasping. The significance of rasping as a feeding mechanism is that feeding is not particle-size dependent (Table 1: 22); large pieces of food can be broken down and manipulated before ingestion, whether the food be a large algal mat on a hard surface, prey, or large pieces of detritus. The ability to manipulate food before inges- tion may be one of the reasons for the great success of the gastropods. К is not possible on the evidence presented here to determine the structure of the archi- molluscan alimentary tract. Certainly it had a nonarticulated radula with protractors, ге- tractors, and bolsters, paired tubular salivary glands, a cuticular foregut, and a dorsal ciliat- ed tract running down the midgut. If the Neo- meniomorpha have retained a primitive mid- gut even though they have become food specialists, then a digestive gland must have been derived more than once in the mollusks. On the other hand, if the neomeniomorph midgut is reduced, then a protostyle without a style sac or gastric shield and a digestive di- verticulum could have been primitively pres- ent in the mollusks, a condition that would lead more directly to parallel evolution in the molluscan midgut of a style sac and gastric shield. The two studies on aplacophoran gut development for two Neomeniomorpha did not have this question in mind (Baba, 1938; Thompson, 1960), but Baba observed that the intestine arises from endoderm and that the midgut epithelium when it first forms is thick- est laterally and ventrally, as it is in the chaetoderm digestive diverticulum. The evidence from entire, isolated radulae of Aplacophora indicates for the mollusks an original state of distichous rows of teeth on a divided radular membrane. The evidence for an original single basal membrane with rows of broad monoserial teeth rests on recon- structions from histologic sections of the neomeniomorph radula of Dondersia (Nierstrasz, 1905), on histologic sections of Simrothiella (Salvini-Plawen, 1972a), and on ontogenetic studies on chitons (Minichev & Sirenko, 1974). Kerth (1979) has shown, on the other hand, that distichous teeth on a sin- gle membrane develop ontogenetically in the pulmonates. The questions of whether the archimol- luscan radula was a single or paired structure and whether or not the midgut had a digestive diverticulum is left open for further observa- tions on isolated aplacophoran radulae, com- parative histologic studies and studies on development. Ecological Considerations Although Aplacophora are ubiquitous in the deep sea from the edge of the continental shelf to the deepest abysses and hadal depths, they seldom are numerically an im- portant constituent of the macrofauna. The two chaetoderm genera described here which are the most primitive also have the fewest known species: Scutopus (4) and Limifossor (4). The number may be doubled, at most, from existing collections not yet described. The carnivorous species belonging to Falcidens and Chaetoderma are far more numerous, although their total numbers in any one sample are never great (unpublished data). Species of Prochaetoderma are numerous (unpublished data) and can be the numerical- ly dominant macrofaunal animals in quantita- tive samples. Prochaetoderma sp. y was the dominant species in a total of twenty-five 35- cm? tubular cores taken at one 1,760-m sta- tion off Woods Hole (Grassle, 1977), although dominance was not high (6.0%; discrepancy from Grassle’s data due to recent recognition of a sibling species). The next four most numerous species were polychaete worms (5.1%, 4.4%, 4.2%, and 3.7%) (nematodes, ostracods, and copepods excluded). Actual density of Prochaetoderma y was 309 m-2. In a Va-m? spade box core taken in the same area, this species was the fourth most numer- ous species with a density of 192 т-2. In other quantitative samples in the same area, Prochaetoderma y ranged in numbers up to 237 m-2, and in grab samples taken between 1141 and 2148 m depths it ranged up to 400 m-2. In a Va-m2 spade box core taken in the re- markably productive Aleutian Trench off Alaska at a depth of 7298m, another Prochaetoderma was one of the dominant species at a density of 124 m-2 (Jumars & Hessler, 1976). The numerical success of some species of Prochaetoderma may be attributable in part 382 SCHELTEMA to their efficient gastropod-like rasping radula, which has made a wide size range of food sources available to them in an environment where food 15 probably a limiting factor. CONCLUSION The Aplacophora exhibit a wide variation in morphology of the alimentary tract. Compara- tive studies of these morphologies give insight into evolutionary events and function among the Mollusca and lead to a greater under- standing of feeding in the deep sea. ACKNOWLEDGMENTS Material was made available to me by Howard L. Sanders (Woods Hole Oceano- graphic Institution), Centre Océanologique de Bretagne, Paul S. Mikkelson (Harbor Branch Foundation, Fon Pierce, Florida), and P. Leloeuff and A. Intes (Centre de Recherches Océanographiques, Abidjan, Cóte d'lvoire). Arthur G. Humes, June F. Harrigan, and Isabelle P. Williams kindly read the manu- script critically. To all of these are due my gratitude. To Rudolf S. Scheltema go especial thanks for many a provocative discussion about clarity in presenting morphology. The materials from Woods Hole Oceanographic Institution were collected and sorted under National Science Foundation grants GB 6027X, GA 31105, and GA 36554. LITERATURE CITED BABA, K., 1938, The later development of a soleno- gastre, Epimenia verrucosa (Nierstrasz). Journal of the Department of Agriculture, Kyushu Uni- versity, 6: 21—40. BABA, K., 1939, Preliminary note on Epimenia verrucosa (Nierst.). Venus, 9: 35—42 (in Japan- ese). BABA, K., 1940, Epimenia ohshimai, a new soleno- gastre species from Amakusa, Japan. Venus, 10: 91-96 (in Japanese). BOETTGER, С. R., 1956, Вейгаде zur Systematik der Urmollusken (Amphineura). Zoologischer Anzeiger, Suppl. 19: 223-256. FRETTER, V. 8 GRAHAM, A., 1962, British proso- branch molluscs. Ray Society, London, 755 p. GRAHAM, A., 1973, The anatomical basis of func- tion in the buccal mass of prosobranch and amphineuran molluscs. Journal of the Zoological Society of London, 169: 317-348. GRASSLE, J. F., 1977, Slow recolonisation of deep sea sediment. Nature, 265: 618-619. HEATH, H., 1905, The morphology of a soleno- gastre. Zoologische Jahrbücher, Abteilung für Anatomie und Ontogenie der Tiere, 21: 703- 734, pl. 42-43. HOFFMAN, S., 1949, Studien über das Integument der Solenogastren nebst Bemerkungen über die Verwandtschaft zwischen den Solenogastren und Placophoren. Zoologiska Bidrag fran Uppsala, 27: 293-427. IVANOV, D. L., 1979, Structure and functional morphology of the radular system in Chaeto- derma (Mollusca, Caudofoveata). Zoologi- scheskii Zhurnal, 58: 1302-1306 (in Russian). JUMARS, P. A. & HESSLER, R. R., 1976, Hadal community structure: implications from the Aleu- tian Trench. Journal of Marine Research, 34: 547-560. KERTH, K., 1979, Phylogenetische Aspekte der Radulamorphogenese von Gastropoden. Mala- cologia, 19: 103-108. KOWALEVSKY, A., 1901, Sur le genre Chaeto- derma. Archives de Zoologie experimentale et general, ser. 3, 9: 261-284, pl. 10-12. MINICHEV, A. S. & SIRENKO, B. J., 1974, Devel- opment and evolution of the radula in Poly- placophora. Zoologicheskii Zhurnal, 53: 1133- 1139 (in Russian). NIERSTRASZ, H. F., 1903, Neue Solenogastren. Zoologische Jahrbücher, Abteilung für Anato- mie und Ontogenie der Thiere, 18: 359-386, pl. 35-36. NIERSTRASZ, H. F., 1905, Kruppomenia minima und die Radula der Solenogastren. Zoologische Jahrbücher, Abteilung für Anatomie und Ontogenie der Thiere, 21: 665-702, pl. 39-41. ODHNER, N. H., 1921, Norwegian Solenogastres. Bergens Museums Aarbok 1918-1919, Nr. 3, 1- 86, pl. 1-6. PRUVOT, G., 1891, Sur l'organisation de quelques néoméniens des cótes de France. Archives de Zoologie expérimentale et general, ser. 2,'9: 699-806, pl. 25-31. SALVINI-PLAWEN, L. V., 1967a, Neue scandi- navische Aplacophora (Mollusca, Aculifera). Sarsia, 27: 1-63. x SALVINI-PLAWEN, L. V., 1967b, Uber die Beziehungen zwischen den Merkmalen von Standort, Nahrung und Verdauungstrakt bei Solenogastres (Aculifera, Aplacophora). Zeitschrift für Morphologie und Okologie der Tiere, 59: 318-340. SALVINI-PLAWEN, L. V., 1969, Solenogastres und Caudofoveata (Mollusca, Aculifera): Organisa- tion und phylogenetische Bedeutung. Mala- cologia, 9: 191-216. SALVINI-PLAWEN, L. V., 1972a, Zur Morphologie und Phylogenie der Mollusken: Die Beziehungen der Caudofoveata und der Solenogastres als Aculifera, als Mollusca und als Spiralia. Zeit- schrift für wissenschaftliche Zoologie, 184: 305— 394. ALIMENTARY TRACTS IN APLACOPHORA 383 SALVINI-PLAWEN, |. V., 1972b, Die Caudo- foveata des Mittelmeeres und das Genus Scutopus (Mollusca, Aculifera). In Fifth Euro- pean Marine Biology Symposium, Battaglia, B. (ed.) Piccin, Padua, p. 27-51. SALVINI-PLAWEN, L. V., 1975, Mollusca Caudo- foveata. Marine Invertebrates of Scandinavia 4, Universitetsforlaget, Oslo, 55 p. SALVINI-PLAWEN, L. V., 1978, Antarktische und subantarktische Solenogastres (eine Mono- graphie: 1898-1974). Zoologica (Stuttgart), 44: 1-315. SALVINI-PLAWEN, L. V., 1980, Molluscan digestive system in evolution. Haliotis, 10(2): 127 [abstract]. SALVINI-PLAWEN, L. V., in press? Diets for aplaco- phoran and monoplacophoran molluscs. Hand- book of Nutrition and Food, CPC Press. SALVINI-PLAWEN, L. V. & NOPP, H., 1974, Chitin bei Caudofoveata (Mollusca) und die Ableitung ihres Radulaapparates. Zeitschrift für Morph- ologie der Tiere, 77: 77-86. SCHELTEMA, A. H., 1972, The radula of the Chaetodermatidae (Mollusca, Aplacophora). Zeitschrift für Morphologie der Tiere, 72: 361- 370. SCHELTEMA, A. H., 1973, Heart, pericardium, coelomoduct openings, and juvenile gonad in Chaetoderma nitidulum and Falcidens caudatus (Mollusca, Aplacophora). Zeitschrift für Morph- ologie der Tiere, 76: 97-107. SCHELTEMA, A. H., 1976, Two new species of Chaetoderma from off West Africa (Aplaco- phora, Chaetodermatidae). Journal of Molluscan Studies, 42: 223-234. SCHELTEMA, A. H., 1978, Position of the class Aplacophora in the phylum Mollusca. Malaco- logia, 17: 99-109. SCHWABL, M., 1961, Crystallophrisson (= Chaetoderma) hartmani nov. spec., eine neue Aplacophore aus dem Ostpazifik. Zoologischer Anzeiger, 166: 258-277. SCHWABL, M., 1963, Solenogaster mollusks from southern California. Pacific Science, 17: 261-281. STASEK, C. R., 1972, The molluscan framework. In Chemical Zoology, МИ. Mollusca, FLORKIN, М. & SCHEER, В. T. (eds.), р. 144. THOMPSON, T. E., 1960, The development of Neomenia carinata Tullberg (Mollusca Aplaco- phora). Proceedings of the Royal Society of London, ser. B, 153: 263-278. WIREN, A., 1892, Studien über die Solenogastres. |. Monographie des Chaetoderma nitidulum Lovén. Konglige Svenska Vetenskaps-Acade- miens Handlingar, 24(12): 66 p., 7 pl. "O ay LA.) | мч + y ag “nr | им Ar a т qe a Му. у e № ie ti alu к à à A IP ds pda Sd | | y tu Ag me mL A ah ee |: Cen A ne gi ane y vel ет м, ak 1a AA | 7 ral. ро ri | Lu ae ie um LRO x phate pin dr и ot etwas 1 nite Ai Ir, i + oo ras an FA fs see ¿y aren ААРОН eae? TA A Ts Lot | mn on dE: IÓ oir de? AA | ра 1 щи a ¿a en ne Е de er udn ri AT iv Ve SEHR иг A? { 1 ++ Гы ton hé т fea Зы Ava L я pa DEN) ne та dant PAPA ne РИ ЕТ” vin \ ot dot! lu th Mi ‘al у Sem Я | «AY, A © EN a Ta i Mir uli aes. т т à A | IE d | eternal А: dust! Rio qed AA ATA ый on ии. sa? ee | yang Lice PI? EL NEA ani N ero N ; “Y и rem nm } y ^ AA asl Exif. à j 4 NE, | gl | VAR { ash D NULS WIN | : SU : a= | в EF = 2 5 . MALACOLOGIA, 1981, 20(2): 385—402 MORPHOLOGICAL AND BEHAVIOURAL ASPECTS OF FEEDING IN THE CASSIDAE (TONNACEA, MESOGASTROPODA) Roger М. Hughes and Helen Р. 1. Hughes Department of Zoology, University College of North Wales, Gwynedd, LL57 2UW, United Kingdom ABSTRACT Anatomy of the feeding apparatus, properties of the proboscis gland secretion, diet, and feeding behaviour of the Cassidae are reviewed from published data. New data are presented on the alimentary morphology of Cassis tuberosa and on the pursuit, attack, prey penetration and feeding methods of C. tuberosa and Cypraecassis testiculus. Tonnacean proboscis gland (PG) secretion is compared with accessory boring organ (ABO) secretion of the Naticidae and Muricidae. The Cassidae are mainly nocturnal predators that feed specifically on echinoids; diets prob- ably reflect the availabilities of specific echinoids in the habitat. The basic design of the ali- mentary system is similar to that of other tonnacean families. Two large proboscis glands deliver a secretion rich in sulfuric acid (pH < 1) via long ducts that pass through the nerve ring, along the proboscis to the buccal cavity. Prey are gripped by the foot and penetration of the test is achieved within about 10 min by the combined action of sulfuric acid and the radula. Scanning electron micrographs reveal severe etching, but no radular marks, on the cut edges of the test. Prey do not appear to be anaesthetized during attack. Consumption of internal tissue takes about 1-2 hr, but the feeding time can be more than doubled when all external tissue and spines are eaten. There appears to be no size selection of prey. ABO secretion of naticids and muricids, which drill bivalves, and tonnacean PG secretion both dissolve minerals with the aid of an inorganic acid and probably a chelating agent. PG secretion, however, is much more acidic and is produced in far greater quantities than ABO secretion, and may lack the ability of the latter to attack the organic matrix of calcareous skeletal material prior to mineral dissolution. INTRODUCTION The beautiful shells of the Cassidae (helmet shells) have attracted attention al- most to the total exclusion of the living ani- mals. Yet these mesogastropods, which have a world-wide distribution in tropical to warm temperate sand and reef habitats, are in- triguing not only because of their shells, but also because they feed almost exclusively on echinoids. In spite of this rather bizarre diet for a gastropod, accounts of the feeding biology of cassids were, until recently, confined to brief notes scattered in various journals. In the present paper, we have synthesized informa- tion from the literature and present new data on the feeding methods of the Cassidae. Where relevant, we have also included pub- lished information on the Cymatiidae, Bursidae and Tonnidae that together with the Cassidae and Ficidae comprise the super- family Tonnacea. Our review deals sequenti- ally with the anatomy of the feeding ap- paratus, properties of the associated glands and their secretions, mode of attack, penetra- tion and consumption of prey. ANATOMY The general plan of the alimentary system is similar throughout the Tonnacea, useful recent accounts being those of Day (1969) for Argobuccinum, and Houbrick «€ Fretter (1969) for Cymatium and Bursa. Reynell's (1905) anatomical description of Galeodea (= Cassidaria) rugosa (L.) suffers from poorly reproduced diagrams. The description of the cassid alimentary system given here is based on new data for Cassis tuberosa (L.) (Fig. 1a, b). Proboscis and buccal apparatus The mouth and buccal mass lie at the tip of the proboscis (Fig. 1, P) which, when fully ex- (385) 386 HUGHES AND HUGHES tended, is 1 to 1.5 times the length of the shell, and when retracted lies within the pro- boscis sheath whose entrance forms a false mouth or rhynchodeum (RH). Retractor mus- cles running longitudinally in the proboscis wall become free in the proximal region of the proboscis, forming strap-like connections (PRM) with the walls of the cephalic haemo- coel (CH). Retraction of the proboscis there- fore involves both shortening, due to contrac- tion of muscles in the proboscis wall, and in- version of the proximal region due to contrac- tion of the free retractor muscles. The Cymatiidae lack free retractor muscles, and retraction occurs solely by contraction of the proboscis wall (Day, 1969; Houbrick 8 Fretter, 1969). The tonnacean proboscis is therefore rather different from the pleurembolic type of most neogastropods, where retractor muscles lying freely within the proboscis cavity are in- serted along the length of the buccal mass (Day, 1969). The bulk of the retracted pro- boscis and the huge proboscis glands are compensated by reductions in the pallial organs in the anterior half of the pallial cavity. The hypobranchial gland becomes thin to- wards its anterior end and the anus 1$ posi- — FIG. 1. a. Cassis tuberosa. The visceral mass, mantle and dorsal wall of the cephalic haemocoel have been removed to reveal the feeding apparatus and associated glands. A aorta, AO anterior esoph- agus, BG buccal gland, CH cephalic haemocoel, CHW cephalic haemocoel wall, CM columella mus- cle, F foot, J jaw, M mantle, NR nerve ring, OG esophageal gland, OP operculum, PEGO pedal gland opening, PG proboscis gland, PGD proboscis gland duct, PL proboscis lumen, PO posterior esophagus, PRM proboscis retractor muscle, PSL proboscis sheath lumen, RA radula, RH rhyncho- deum, SG salivary gland. b. Cassis tuberosa re- moved intact from its shell. DG digestive gland, E TEN ) 5 eye, HG hypobranchial gland, К kidney, P pro- N In DG boscis, PS proboscis sheath, S stomach; other > symbols as for Fig. 1a. b 2 ст ae dd rin ti MN e ВНЗ = — FIG. 2. a. Tripneustes ventricosus drilled by Cassis flammea, which had swallowed half the spines; b. 7. ventricosus drilled by Cassis tuberosa. An area of spines has been cleared and a hole made in the test by cutting out a disc, which can be seen still in place. This urchin was removed from the C. tuberosa 10 min after the initial attack; c. Edge of hole in (b), showing undercutting caused by erosion by sulfuric acid; d. edge of disc showing severe etching by sulfuric acid; white areas are unetched surface material; e. higher magnification of spine boss in (d), showing severe etching; f. jaw of Cassis flammea; 9. radula of Cassis tuberosa; В. central (rachidian) teeth showing wear on median cusps; i. marginal teeth; ci are scanning electron micrographs. 387 FEEDING IN CASSIDAE e* men ten er Ae e 388 HUGHES AND HUGHES tioned much further back than in other meso- gastropod superfamilies. The buccal mass, anchored within the tip of the proboscis by fine radiating muscles, con- tains a pair of horny jaws (J), the radula (RA) and associated odontophore, a region of glandular tissue (BG) behind the odonto- phore, and the openings of the paired pro- boscis gland ducts. Each jaw forms a bluntly pointed plate comprised of rows of contiguous rods, the distal ends of which form diamond- shaped subunits, giving the outer surface of the jaw a 'snake-skin' appearance. Instead of having a sharp saw-toothed edge as т Cymatium or Charonia, the jaw of Cassis has a blunt, rounded edge suitable for gripping spines and strands of flesh (Fig. 2f). The radula has 7 teeth per transverse row, comprised of a central (rachidian) tooth on either side of which are a lateral and two marginal teeth (Fig. 2g). The central and lat- eral teeth are heavily cusped (Fig. 2h), where- as the marginal teeth are more delicately armed with 3 to 4 cusps arranged to form a claw (Fig. 2i). Ontogenetically, the marginal teeth become functional before the central and lateral teeth, and are used to grip shreds of flesh by intermeshing as the radula rolls backwards over the bending plane of the odontophore. More distally along the radula, the central and lateral teeth are used to rasp the calcareous tests of urchins, as a conse- quence of which their cusps become worn (Fig. 2h) and the delicate marginal teeth be- come torn away. The buccal glandular tissue (BG) may be homologous to the “partly paired buccal gland” described in Argobuccinum argus (Gmelin) by Day (1969), to the “glandular patches” described in Bursa granularis (Röding) by Houbrick & Fretter (1969) and to the “blindsack” of Tonna galea (L.) (Weber, 1927). These glandular structures are of un- known function. Also opening into the buccal cavity are the paired proboscis gland ducts that deliver a secretion rich in sulfuric acid used to etch the calcium carbonate tests of urchins (Fig. 2 c-e). Alimentary canal On leaving the buccal mass, the anterior esophagus (Fig. 1a, AO) forms a narrow tube with dorsal and ventral typhlosoles, and runs along the proboscis, finally dropping sharply downwards through the nerve ring (NR). On emerging from the nerve ring, the esophagus immediately bends dorsalwards, whereupon its upper side is modified into the septate esophageal gland (OG), which runs alongside the aorta (A) beneath the proboscis glands (PG), almost for the remaining length of the cephalic haemocoel. The repeated transverse folds of the esophageal gland are richly sup- plied with secretory cells that in Argobuccin- um argus are of three types, one secreting mucus and the others secreting unidentified digestive enzymes (Day, 1969). The posterior esophagus (PO), which is a simple ciliated tube with thick muscular walls, leaves the esophageal gland and opens into the U-shaped stomach (Fig. 2b, S). The large digestive gland (DG) is connected to each limb of the stomach by separate ducts. The short intestine emerges, without sharp dis- tinction, from the stomach and runs forward through the kidney sac (K) to the muscular rectum and anus. Proboscis glands The proboscis gland ducts (Fig. 1a, PGD) run from the buccal mass along either side of the esophagus, through the nerve ring, to the relatively massive glands lying in the cephalic haemocoel. This anatomical feature is unique to the Tonnacea. Each proboscis gland con- sists of 2 to 3 lobes, the right gland being considerably smaller and placed more cen- trally and further forward than the left gland beneath it (Fig. 1a). The proboscis glands are covered by an intricate network of muscles, the contraction of which forces the secretion through the proboscis gland ducts into the buccal cavity. The glands are anchored by fine muscle strands connecting to the body wall and foot. The proboscis gland's histology has been described by Núske (1973) for Galeodea (= Cassidaria) echinophora (Lamarck) and by Day (1969) for Argobuccinum argus. The fol- lowing account is synthesized from both sources. Each gland consists of numerous tubules draining into collection ducts (Fig. 3a). The tubule walls are made of a single layer of large cells that produce the acid secretion (Fig. 3b). Núske (1973) recognized three phases of cellular secretion. During secretion formation, large vacuoles are formed in the apical cell region by confluence of smaller vacuoles appearing to originate from the Golgi apparatus. The cells increase in size and attain a smooth boundary at the lateral and basal sides where they were previously FEEDING IN CASSIDAE 389 1mm FIG. 3. a. dissected tubules of proboscis gland of Argobuccinum argus (after Day, 1969); b. transverse and longitudinal sections of PG tubules of Galeodea echinophora (after Núske, 1973). L tubule lumen, М nucleus, V secretion vacuole. thrown into interdigitating folds. These folds may represent a plasma membrane reservoir for the enlargement of the cells during secre- tion ‘ormation. When full, the cells are each cccupied by one or a few large secretion vacuoles. The nucleus, cytoplasm, organelles and abundant small vacuoles are confined to the basal layer (Fig. 3b). After liberation of the apical secretion vacuole, the cytoplasm, con- taining numerous small vacuoles, becomes evenly spread throughout the entire cell. Meanwhile the cells shrink and become ex- tensively folded along their lateral and basal sides. New secretion vacuoles appear within a few hours after the proboscis glands have been emptied (Núske, 1973). Proboscis gland secretion The chemical properties of the proboscis gland secretion have been studied for Galeodea echinophora by Fange & Lidman (1976) and for the cymatiid Argobuccinum argus by Day (1969), these recent analyses agreeing closely with the much earlier study of Panceri (1869) on Tonna galea and G. echinophora. The proboscis gland secretion of G. echinophora is hypertonic to seawater, with a pH of about 0.13. Hydrogen and sulfate ions are the predominant components, while sodium and chloride concentrations are lower than in the blood, and potassium and mag- nesium concentrations equal those in sea- water. Only minute traces of organic material are present, and in this respect the Cassidae may differ from the Cymatiidae and Bursidae that are known to secrete toxins that anaes- thetize the prey (Asano & Itoh, 1960; Day, 1969; Houbrick & Fretter, 1969). The secre- tion of the cymatiid, A. argus, has a pH of 1.1, and consists largely of 0.4-0.5 М sulfuric acid. Day (1969) established that the secretion contains no enzymes involved with the dis- solution of calcium carbonate, but that 33% of the calcium carbonate (bivalve shell) dis- solved by experimentally administered secre- tion was attacked by some component other than sulfuric acid, probably a chelating agent. How the proboscis gland cells store the strongly acid secretion without cellular dam- age, and the nature of the physical-chemical pathways which produce such high concen- trations of hydrogen and sulfate ions, remain unknown. (aduapina |1ецие]$шпэлэ) рюшцэа Apnıs juasaid sopeqieg pıoßueyeds paıyyuspıun Buimounq (youewe7) 1/61 ‘seuBny $ sayßnH sopeqieg ‘шпыепбе SNSODLJUAA salsnaudu] spioulyoa 1/61 “sayBny 9 sayßnH sopeqieg ‘wnuenbe (1) sazunon¡ едашоицэз ¡eunejida (7) Pawwey $15$е) (y9/euue7) 6961 '“Jaysay) Áyeso] payioadsun 'ueaqqueo взоэщиал ешоэи/ 6961 ‘1э4$эц5 (инэшо) splouly98 “9561 'SJO0ON AlP90] Paıyıoadsun ‘sAay ерио|- sipuesB snssugoibejq Buimounq (youewe) smebanen sploulyoe (y9Jeu1e7) 1661 “uBwÁ7 ерио|- (sa3snaudoxo/=) Snuiy28}A7 ¡eunejida $/5иаиеэ$ебереш $1552) (youewe7) 6961 HOM epio¡4 ‘шпиепбе взоощиал ешоэи/ (уцешо) 6961 ‘JOM epio¡Y4 ‘шпиепбе 5/риелб snssu1qoibeld 0 6961 ‘JOM epio¡4 ‘шпиепбе ‘ds елои/ 2 6961 ‘OM epio¡4 ‘wnuenbe ‘ds snauoulyo3 (6) (84537) = 6961 ‘HIOM epio¡4 ‘шпиепбе ejesouadsambunb ejay a (Ae19) Z 6961 HOM ерис|- ‘шпиепбе snssaldapqns /э15еэ4/!9 E 1961 ‘18/80 seweyeg ‘IUILUIg (1) snaoesos Ja3seadÁ/o u youewe] spioulyoa 5 8/61 ‘лэ!эре9 5рие!$| UIBIIA шпаеэодиеэ sninpisseg BuimoJinq =) 6961 HOM epio¡4 ‘wnuenbe ‘ds eıoeguw = 6961 ‘HIOM epuoj4 ‘wnuenbe 1/61 ‘seuBny $ seuBnH sopeqieg 'wunienbe (*1) лэлипэп/ eN9wouly9F 0/61 ‘JepAus 3 18pAUS seweyeg “ug 6961 ‘HIOM epuo|4 ‘шпиепбе 8/61 ‘лэуерею (1ddijiyd) 2961 'Japaolyos spueisj иблл шпае|цие ешэреа 1561 ‘чешАл epuo|4 (youewe) зтебаиел 2896! ‘uoqqy Aq pajonb uvosuaqoy seweyeg ‘iuimig ‘wnuenbe (Sajsnaudoxoy =) зпицээИ7 Арп}$ juasaid sopeqieg 1/61 ‘seubny 9 seubny sopeqieg ‘wnuenbe 6961 HOM epio¡4 “unuenbe (yoJeue7) spioulyoa 9761 ‘2100N shey ерио|- зпзоощиал $а]5паиам1 ¡eunejida (7) eso/aqn] Sıssen = Auounvy Ayıle90 saioeds A914 adh} Adıd 10}ep91d с Е А AAA ES ‘эер!5$е) ay, jo Áeld рэрлоээн ‘| 3718991 nn een SSS 8961 'noqqy Aq payo 2961 ‘I21qe9) 3 чозлэцаэеи (эоцарма |ецие]зшполэ) eg961 'noqqy Аа pajonb “ajzuay 391 7/61 ‘Aeq 6961 ‘eq (aduapina |ецие}$шпэлэ) 8/61 ‘1э\э}ре9 9561 ‘2100W иоцеэипшшоо ¡euosiad 'suoog 28961 'noqqy Aq pajonb ‘Aids 8461 'ABUIUBA (ээцэрма |ецие}$шполэ) 8/61 ‘10Ае1 1161 “a¡puaH 4461 “a¡puay 2161 ‘лэриэн 6961 440M FEEDING IN CASSIDAE 6961 ‘HIOM Apnıs juasaid Apnıs juasaid 2161 ‘лэриэн 2161 ‘18JPUSH 8961 'UOSISYJIN Арп}$ juasaud 1161 “Ja¡puaH 1661 'uBuwÁ7] 5/61 ‘зрипшрз 9 spunwp3 (uodaı pajenuejsqnsun) 6961 ‘чеэриз 5161 “2/61 'uBapuy eIe1sny uooße| Y0}8MIUZ Boy ymos Bly ymos spueisj UIBJIA skay epuo|s ooıxayy “eyinAes еэщу 1583 eIUPZUE] ‘рие|$| IMIZEW sanıpjew ‘IIOIY NPPY eweued eweued eweued epio¡4 ‘wnuenbe epio¡4 ‘wnuenbe sopeqieg ‘шпиепбе sopeqieg 'wnuenbe eweued eweued epuoj4 ‘wnuenbe sopeqieg ‘wnuenbe eweued po] еиеце ‘шпиепбе е|едзп\у ‘J90H Jaleg е|езп\у ‘joey JalJeg (элцепээ4$ aq Леш) рэшоэ4двип zısseßy “y sapio -18]SEIWOY SNSSUQOUIYY (anıyeinoads aq Aew) paijioadsun (aysa7) sme/opadsiq SN2SIPOUIY27 youewe шплеэодиеэ SNINPISSET (94537) ejyesopadsainbuinb ej an zısseßy “y NUNIQUEA едэшоицэ3 sulyojn Bas „рэша$ yoys,, paiyioadsun (yo/ewe7) snjoajid (sa] -snaudoxo, =) sajsnauduy zısseßy JOJOD/UN snssug ayse7 snwo]soj9Á9 SnauOUIYyOF (youewe 7) saplojnquy suepiong зрюшцээ paioadsun jo „Азэмел apim, ('Ча!!Ча) une] ue euwapelg (1981487) SNSODLJUBA $9]5пэиаи1 zisseBy SIDUIA едэшоицэу (7) 4zajunon) едлэашоицэ3 (yo/ewe7) зтебаиел (5$э15пэиаохо| =) snuiyoajA7 1949$041 Y J8IINN теле] 1э}5еэ/0 (1) 1oue/d sajseyjueoy (94597) Wnsojas ewapeig ёзэмела plouIy98 Buimouing сзэмела piouIyoa Buimounq spioulyoa Buimouing ploulyde ¡eunejida зрюццоэ ¡eunejida spioulyoa Buimouing SplouIy98 euneyide pioiayse &pioiayse pIouIy98 ¡euneyida (youewe) wnsouesbiwas шпуеца (\эибем $ ueqnuos) un]e9/nsiq шпуеца (yo/euwe7) (wnomejhoz =) шпе!де| шпуеца (41089) ште/пиелб шпуеца (Aquamos) 2/29/8209 SISSPISeIdÁAN) (7) Byns sissegaeidÁo (7) sn¡nonse] sisseoaeldÁo (snIA0UOJE) esouids ‘7 =) (U!9UIE)) е]е//э$$а} sisseg (`]) eanuloo $1552) 392 HUGHES AND HUGHES Salivary glands Partially embedded in the anterior lobe of each proboscis gland is a much smaller acinar salivary gland (Fig. 2a, SG). It has commonly been assumed that the salivary glands empty into the proboscis gland ducts, but Day (1969) found that the salivary glands of Argobuccinum argus empty directly into the esophagus immediately before the esoph- ageal gland. Nüske (1973) described the salivary gland of Galeodea echinophora as consisting of mucus-secreting cells and “canaliculi” cells. The latter are characterized by intercellular canaliculi densely filled with microvilli and cilia, and Núske (1973) sug- gested that they may produce the chloride component of the secretion emptying into the buccal cavity. This interpretation would seem to be erroneous if, as in A. argus, the salivary glands of cassids empty by separate ducts into the esophagus. Moreover, Fänge & Lidman (1976) found chloride to be present only in low concentrations in the secretion ejected from the proboscis of G. echinophora. Day (1969) claims that the salivary glands of Argobuccinum argus have a high amylase activity. THE DIET Records of prey taken by cassids are sum- marized in Table 1, from which it is evident that with very few exceptions, the Cassidae feed exclusively on echinoids. The exceptions are as follows. Morum oniscus (L.), not listed in Table 1, failed to eat any kind of echino- derm when an exhaustive series was offered in the laboratory (Work, 1969). Cassis cornuta (L.) was reported by a shell collector to eat the crown of thorns starfish Acanthaster planci (L.) (cited in Endean, 1969), and рге- sumably this was the origin of Profant's (1970) claim that C. cornuta eats A. planci. Endean (1973), however, thought that the ob- server may have mistaken the urchin Diadema setosum (Leske) (known prey of C. cornuta) for A. planci. Cassis tessellata (Gmelin) (= C. spinosa Gronovius) “rasped away” at the astreroid Oreaster clavatus Muller & Troschel in captivity (Edmunds & Edmunds, 1973), but the success of this feed- ing attempt was not reported and the incident may have been induced by artificial conditions in the aquarium. Phalium labiatum (Lamarck) and Phalium semigranosum (Lamarck) are both reputed to feed on bivalves. The mol- luscan diet of Phalium is unsubstantiated and the reports may be speculative, since bivalves are common in the sandy areas populated by those cassids. Phalium spp. might be capable of drilling bivalve shells since the secretion of the cymatiid Argobuccinum argus etches Macoma sp. shells (Day, 1969). Neverthe- less, bivalves have not been recorded in the diet of the better known species Phalium granulatum (Born). A wide range of echinoids is consumed by the Cassidae, and diets seem to reflect the common species of urchin present in the habi- tat. Only stout-spined urchins such as Eucidaris tribuloides (Lamarck) tend to be avoided, although this species has been re- corded in the natural diet of Cypraecassis testiculus (Hendler, 1977) and is readily taken by the cymatiid Charonia variegata Lamarck (McPherson, 1968; Work, 1969). Cassids liv- ing on sea-grass beds or near reefs feed on epifaunal echinoids, whereas populations of the same species living on soft substrata feed on burrowing echinoids. Phalium spp. always inhabit sandy substrata where they feed on burrowing echinoids, notably clypeastroids. Possibly, individuals become conditioned to specific echinoid prey. In the present study, Cassis tuberosa fed readily in the laboratory on Tripneustes ventricosus (Lamarck) and Echinometra lucunter (L.), but ignored Diadema antillarum (Philippi). However, Schroeder (1962), Snyder & Snyder (1970) and Gladfelter (1978) recorded C. tuberosa feeding on D. antillarum in the field. Of four C. testiculus (L.) that we collected from a single locality in Barbados, one fed rapaciously on D. antillarum, whereas the other three readily attacked Т. ventricosus and E. lucunter but would not attack D. antillarum. However, they occasionally fed communally on D. antillarum that had been overcome by the first C. testicu- lus. Perhaps the technique of attacking the active, long-spined D. antillarum has to be learned. Diets of newly settled or newly hatched cassids are unknown. They may feed on very small juvenile echinoids, but it is possible that they subsist on other, more readily available, invertebrate taxa. On three occasions in the laboratory, Cassis tuberosa were seen to stop hunting and to twist the head back along the side of the shell aperture, extending the proboscis to reach the highest parts of the shell (Fig. 4d). The proboscis plucked at the shell surface as if to remove the algae growing there. The pur- FEEDING IN CASSIDAE 393 FIG. 4. a. Cassis tuberosa arching over Tripneustes ventricosus in initial phase of attack; b. urchin grasped by front part of foot; c. a groove between lobes of foot supports proboscis during penetration and feeding; d. extension of proboscis to 'clean' shell surface. pose of this behaviour remains a mystery, as removal of the algae would detract from the camouflage they provide; the possibility that the algae supply some essential nutrients, not available from the prey, seems unlikely. FEEDING METHODS AND BEHAVIOUR The following account of feeding methods is based largely on observations of 4 individ- uals of Cassis tuberosa, 1 of C. flammea (L.) and 4 of Cypraecassis testiculus collected in Barbados during April 1980. The animals were kept at about 29°С in a shallow concrete tank (3.95 m long and 0.76 m wide) filled to a depth of 18 cm with running seawater. The C. testiculus were partitioned off in an area of the tank and provided with a 4 cm layer of sand. Observations were made throughout the night using a dimmed torch. 394 HUGHES AND HUGHES Hunting behaviour Cassis spp. and Cypraecassis testiculus seldom moved during daylight, remaining partially (Cassis spp.) and wholely (С. testi- culus) buried in sand when it was available. Animals began feeding at various times throughout the night, but generally ceased feeding at daybreak. The single specimen of Cassis flammea was exceptional because, during the day, it often remained feeding on urchins that it had attacked the previous night. In contrast to these observations, Boone (personal communication) saw Cypraecassis coarctata (Sowerby) feeding on urchins dur- ing the day in the field. The typically nocturnal feeding habits of the Cassidae are paralleled by those cymatiids that feed on mobile prey (Houbrick 4 Fretter, 1969; Laxton, 1971). Diurnal feeding apparently occurs only in sluggish cymatiids that 'graze' on sedentary invertebrates (Laxton, 1971). While hunting, Cassis tuberosa moves steadily at a speed of about 0.3 cm per second, turning only gradu- ally during its trajectory. The much smaller Cypraecassis testiculus moves at a speed of about 0.5 cm per second, with frequent erratic turns. Method of attack Cassis spp. and Cypraecassis testiculus have quite different attack methods. С. tuberosa detects prey by olfaction, and when it approaches to within a few cm of an urchin, the siphon bends forward and the tentacles become fully extended. Just before tentacular contact is made, the front edge of the foot 1$ raised slightly. As soon as the tentacles touch the urchin, the front half of the foot is raised in a high arch so that the shell is inclined at an angle of about 30°. С. tuberosa continues to move forwards on the hind portion of its foot, at the same time extending its head over the urchin (Fig. 4a). During this maneuver, which usually takes less than 10 sec, no contact 15 made with the urchin except for very brief deli- cate touches by the tentacles. This is impor- tant, because most epifaunal urchins can move faster than C. tuberosa and would often escape if alarmed before they were covered by the predator. Such escapes were observed frequently in the laboratory, the urchins ap- parently alarmed by the scent of the ap- proaching С. tuberosa. Snyder & Snyder (1970) however, found that Diadema antil- larum were unresponsive to C. tuberosa placed upcurrent in the field, but reacted violently when touched by the predator. Schroeder (1962) maintained that in the field, pursuits of fleeing D. antillarum Бу С. tuberosa were usually successful. When the urchin is adequately covered, Cassis tuberosa drops down on the prey, grasping it with the bilobed front portion of the foot (Fig. 4b). The lobes of the foot provide a secure hold on the urchin and the groove be- tween them supports the proboscis during penetration and feeding. The proboscis 1$ also supported during feeding by folds in the front edge of the foot in Cypraecassis testi- culus (Fig. 5b, c), Charonia spp. (Laxton, 1971) and Cymatium nicobaricum (Röding) (Houbrick & Fretter, 1969). Schroeder's (1962) description of the attack behaviour of C. tuberosa feeding in the field on Diadema antillarum agrees with ours of C. tuberosa feeding on Tripneustes ventricosus and Echinometra lucunter, except that when feed- ing on D. antillarum, C. tuberosa repeatedly inserts its proboscis among the spines while arching over the prey. D. antillarum is a par- ticularly agile urchin, and it is possible that C. tuberosa administers a toxin among the spines to hinder the urchin's retreat. lt is un- certain, however, whether cassids are able to secrete toxins. Our single specimen of Cassis flammea did not arch over its prey, but tried to mount urchins directly, with the result that many of them escaped. When successful, C. flammea gripped the side, or at most, only half the up- per surface of the urchin with the front of its foot. This specimen of C. flammea was col- lected on a fine sand substratum where the only potential echinoid prey were spatangoids and clypeastroids. Although we could not in- duce the С. flammea to eat Mellita quin- quiesperforata (Leske) or large Meoma ventricosa (Lamarck) in the laboratory, it is likely that it had been feeding on burrowing echinoids in the field; these prey may require a different attack method than epifaunal echinoids. Moore (1956) described how, when feeding on the large burrowing urchin Plagiobrissus grandis (Gmelin), Cassis madagascariensis (Lamarck) burrows down- wards at a steep angle through the sand and clasps the echinoid in the front part of its foot. Moore (1956) also found Phalium granulatum feeding on M. quinquiesperforata; the pre- dators were perched on top of their prey and had penetrated the test near the centre of the aboral surface. Kier (personal communica- FEEDING IN CASSIDAE 395 b 2cm © 2 ст FIG. 5. а. Cypraecassis testiculus shooting out its proboscis in initial attack on Diadema antillarum; b. spines are gripped by front part of foot; с. Cypraecassis mounts urchin and commences feeding. tion) recorded, by time lapse photography, the feeding of P. granulatum on the burrowing echinoid Cassidulus cariboearum Lamarck. After moving about the aquarium for several hours of the night, P. granulatum stopped and burrowed partially into the sediment. After 8 minutes a dead denuded C. cariboearum ap- peared at the surface and P. granulatum re- sumed locomotion. This represents an out- standingly rapid attack procedure that clearly deserves further study. The film also revealed that C. cariboearum sensed the approach of Р. granulatum and responded by surfacing and moving away as fast as possible. Cypraecassis testiculus 15 able to detect echinoids by olfaction from at least 30 cm downstream in a weak current. Once detect- ed, the prey are approached swiftly and di- rectly. On making first tentacular contact with Tripneustes ventricosus and Echinometra lucunter, C. testiculus extends its proboscis and applies it repeatedly to the test among the spines, penetration commencing within 1-2 min. Meanwhile, the spines are grasped by the forepart of the foot, and by the time pene- tration is well underway, C. testiculus has the urchin securely enveloped. Urchins under at- tack wave their tube feet and spines wildly, and С. testiculus is frequently dragged 30- 50 cm before the prey becomes incapable of further locomotion. Diadema antillarum elicits a rather different attack behaviour. When 1-2 shell lengths away from the urchin, Cypraecassis testi- culus quickly shoots out its proboscis and ap- plies it repeatedly to the periproct (Fig. 5a). The urchin invariably reacts by violently wav- ing its spines and tube feet, retreating hastily. Often, C. testiculus is able to gain on the retreating D. antillarum and to grasp its spines with the front of the foot, whereupon the at- tack always succeeds (Fig. 5b, c). Sometimes 396 HUGHES AND HUGHES D. antillarum escapes before С. testiculus is able to grasp the spines. In the confines of the laboratory aquarium, a chase ensues that may involve several unsuccessful attacks, but usually ends in victory. The swift application of the fully extended proboscis when С. testi- culus attacks D. antillarum is reminiscent of the probing action of the proboscis of Cassis tuberosa when it also attacks this species, again suggesting that a toxin is being deliv- ered to hinder the retreat of such a mobile prey. Secretion of toxins The reputation of the Cassidae for secret- ing toxins derives both from inference, since certain other members of the Tonnacea are known to produce toxins, and from the experi- ments of Cornman (1963), who found that the secretion delivered from the proboscis of Cassis tuberosa, when diluted to 1/1000, would incapacitate the spines and tube feet of Diadema antillarum immersed in the solution. We examined the mobility of spines and tube feet of numerous urchins under attack from Cassis spp. and Cypraecassis testiculus, but could find no evidence for action of a toxin. Spines and pedicellariae remained active throughout the attacks, and remained active for some time even after all internal tissue had been eaten. The tube feet remained active and retained their ability to attach themselves to the substratum for at least 10 min after the initial attack, by which time penetration of the test was usually achieved. After this time, there was a tendency for the tube feet on certain ambulacra to curl up and lose their response to touch, while those on other ambulacra remained active. Internal examina- tion of the tests showed that the ampullae had been stripped from those ambulacra with in- active tube feet. К is still possible, however, that cassids produce a toxin that interferes with the co- ordination of the spines and tube feet, thus impeding locomotion without inhibiting their activity. Such a subtle effect would be hard to detect experimentally. By contrast, cymatiids and bursids such as Argobuccinum argus (Day, 1969), Cymatium nicobaricum and Bursa spp. (Houbrick 4 Fretter, 1969) pro- duce toxins that have very marked anaes- thetic effects on their prey. Moreover, the toxin produced by Fusitriton (= Argobuc- cinum) oregonensis (Redfield) has been identified as tetramethylammonium tetramine (Asano 4 Itoh, 1960). Cymatiid toxin appears to be secreted by the proboscis gland, as Laxton (cited by Endean, 1972) found that extract only from the posterior segment of the “salivary” gland of Charonia rubicunda (Perry) would paralyse starfish. Fánge & Lidman (1976), however, found only minute traces of organic material in the secretion of Galeodea echinophora. The ability of cassids to produce toxins thus remain open to ques- tion. Role of mucus While handling its prey, Cassis tuberosa secretes a thick layer of stiff mucus from the transverse slit along the front edge of the foot (Fig. 1a, b, PEGO). The prey's spines be- come flattened down, all facing away from the area of penetration, under the layer of mucus by gradual pressure from the predator's foot. In this way, the spines do not damage the snail. Pedicellariae become detached and trapped in the mucus, thereby rendered harm- less. Copious quantities of mucus are also secreted by Cypraecassis testiculus, as a re- sult of which the snail is able to climb onto its prey without damage from spines and pedi- cellariae. Mucus is an important agent in the attack method of other tonnaceans. Cymati- um nicobaricum secretes а thick, resilient curtain of mucus over the aperture of its gastropod prey, thus forming a seal around the proboscis which 15 inserted into the lumen of the prey's shell (Houbrick 4 Fretter, 1969). PENETRATION OF THE PREY Position of penetration Cassids usually penetrate their prey through the test, but sometimes entry is gained through the membranous peristome, or in the case of Diadema antillarum, through the periproct. The diameter of the hole made in the test reflects the size of the predator, being about 9mm for adult Cassis mada- gascariensis, 5-6 тт for С. tuberosa, 3- 4 тт for С. flammea, 2-3 тт for Cyprae- cassis testiculus and 1-3mm for Phalium granulatum. Penetration may occur anywhere on the test, but some regions are penetrated more frequently than others. Hughes Hughes (1971) found that Cassis tuberosa feeding on Tripneustes ventricosus and Echinometra lucunter penetrated about 50% of the urchins through the side, about 13% FEEDING IN CASSIDAE 397 through the top and about 13% through the base of the test. The rest were entered through the peristomeal membrane. When feeding on the burrowing echinoid Cassidulus cariboearum, C. tuberosa usually penetrates the relatively spine-free ventromedial region of the test, but this site-preference is lost when very small individuals are attacked (Gladfelter, 1978). Foster (1947) recorded a C. tuberosa penetrating the burrowing urchin Clypeaster rosaceus (L.) near the anal re- gion, while Moore (1956) described a speci- men of the large burrowing urchin, Plagio- brissus grandis, as having been penetrated through the anterolateral edge by Cassis madagascariensis. In the present study, C. flammea feeding on T. ventricosus and Е. lucunter penetrated 6 urchins through the side, 3 through the top and 1 through the base of the test. C. testiculus, also feeding on T. ventricosus and Е. lucunter, penetrated 15 urchins through the periostomeal membrane, 1 through the base of the test, 5 through the side and 7 through the top. Five D. antillarum were penetrated by C. testiculus through the periproct and 1 through the peristomeal mem- brane, but none was penetrated through the test. Because of its membranous periproct, D. antillarum is particularly vulnerable to attack in this region. Moore (1956) found 3 P. granulatum to have penetrated the суре- astroid Mellita quinquiesperforata near the centre of the aboral surface. Beu et al. (1972) observed that New Zealand Tertiary Spatangoida presumed to have been eaten Бу cassids or cymatiids, were drilled mainly near the anterior end of the test, and re- marked that a predator holding the prey in an anterior position would prevent the prey escaping and would be able to push inside the backwardly directed spines more easily. The bias of Cypraecassis testiculus to penetrate Tripneustes ventricosus and Echinometra lucunter through the perio- stomeal membrane contrasts with the tend- ency of Cassis tuberosa and C. flammea to penetrate these urchins through the side of the test. In all cases, the proboscis can easily be extended to all inner parts of the prey and the position of entry would seem to be cor- related either with the way the prey is gripped during feeding (Cassis spp. feeding on 7. ventricosus and E. lucunter) or with the ease of penetration (С. tuberosa feeding on Cassidulus cariboearum, C. testiculus feed- ing on 7. ventricosus, E. lucunter and Diadema antillarum). Mechanics of penetration Before cassids penetrate the prey test, an area slightly larger in diameter than that of the proboscis is cleared of spines that are swal- lowed by Cassis spp. but, except for very small ones, discarded by Cypraecassis testi- culus. Cassis tuberosa and Cypraecassis testiculus complete this phase in 4-5 min. A circular groove is then cut in the test, again taking about 5 minutes. The disc of test there- by cut out is usually pushed inwards, but sometimes is displaced outside the test. Cut- ting is achieved by the combined action of sulfuric acid and radula. Scanning electron micrographs reveal severely etched surfaces but no signs of radular scrape marks (Fig. 2c- e). The median cusps of the central teeth, however, show considerable wear (Fig. 2h), and the buccal mass of C. testiculus was seen to be in perpetual rhythmic action during penetration. The radula is thus undoubtedly used in the drilling process and probably removes the calcium sulfate produced during etching, thereby exposing new layers of calci- um carbonate to the sulfuric acid and maxi- mizing the rate of erosion. Day (1969) found that etching was impeded by the precipitate of calcium sulfate formed when drops of sulfuric acid were placed on bivalve shells. The radula must also be instrumental in rasping through the peristomeal and periproct membranes that are more resistant to the action of sulfuric acid than the test. The etched surface of the test is confined to within a few ит of the edge of the hole (Fig. 2c), suggesting that the ‘lips’ of the proboscis form a seal around the area penetrated, thereby preventing leakage and dilution of the sulfuric acid. Considerable amounts of sul- furic acid may continue to be secreted after penetration because the plates of the test be- come loosened and the test is easily crushed. The delicate tests of Diadema antillarum are often crushed by the weight of Cypraecassis testiculus during feeding. Collapse occurs merely because the tests have been weak- ened by erosion and not, as suggested by Schroeder (1962), because the predator has crushed them with its foot to render the inner parts more accessible. CONSUMPTION OF PREY Having penetrated the test, Cassis tuber- osa, C. flammea and Cypraecassis testiculus 398 HUGHES AND HUGHES consume all the internal tissue, leaving only the gut contents, the peristomeal membrane, and occasionally some of the ampullae. When the internal tissue has been eaten, C. tuber- osa and C. flammea often commence to eat the tube feet, pedicellariae and spines, the proportion eaten varying widely and depend- ing on the appetite. The spines are voided intact in the feces, stacked in parallel in bun- dles and intermingled with thin black filaments up to 30 cm long. С. testiculus swallows only the very smallest spines but eats the tissue and musices at the bases of the larger spines that become detached and drop to the sub- stratum. The long mobile spines of Diadema antillarum have very large basal muscles and these must comprise a relatively high propor- tion of the diet when this species 15 eaten. Surprisingly, we found no significant cor- relation between the time taken to penetrate and consume an urchin (handling time) and the size of the prey (Fig. 6). Much of the varia- tion in handling time was due to the proportion of spines eaten. Hughes & Hughes (1971) found that of 295 Tripneustes ventricosus and 320 5 240 160 Handling Time тт 00 Sa 0 Or Ze A Echinometra lucunter eaten by Cassis tuberosa, about 30% had <9% spines eaten, 50% had 10-90% spines eaten, and 20% had 90-100% spines eaten. The mean handling time for C. tuberosa feeding on T. ventricosus and E. lucunter in the present study was 79 min, S.E. = 6.4 min, n = 38. The single Cassis flammea always consumed all the spines and took an average of 5.5 hr to finish each meal, whereas C. tuberosa seldom took more than 2.5 hr, even when all spines were eaten. In spite of its small size, Cypraecassis testiculus consumed T. ventricosus and E. lucunter at the same rate as C. tuberosa, the mean handling time being 79 min, S.E. = 9 ИТ, = 25: Dietary value of prey and size selection In order to estimate the dietary value of the internal and external tissue of Tripneustes ventricosus, Echinometra lucunter and Diadema antillarum, we dried samples of in- tact urchins and those preyed upon by Cassis spp. and Cypraecassis testiculus to constant AAN SAO Urchin Test Diameter cm FIG. 6. Handling time (time taken to penetrate test and finish meal plotted against diameter (excluding spines) of prey; Vv Cassis tuberosa feeding on Tripneustes ventricosus; O Cypraecassis testiculus feeding on Echinometra lucunter; O С. testiculus feeding on Т. ventricosus; & C. testiculus feeding on Diadema antillarum. FEEDING IN CASSIDAE 399 ES N И. 619 Edible Organic Matter д O — — © Ro со о Dune Pé eras Go TO Urchin Test Diam. cm FIG. 7. Ash free dry weight of total edible organic matter as a function of diameter of prey. E Echinometra lucunter, D Diadema antillarum, T Tripneustes ventricosus. Regression equations for the data after log, transformation of both variables are: E y = 2.8x — 3.5, T y = 3.0x — 4.5, D y = 2.9x — 3.6. Intercept values for internal edible organic matter are E — 4.0, T — 5.0, D — 4.5. weight at 60°С, heated them in a muffle fur- nace for 3 hr at 500°С and reweighed them. Only urchins that had either all or none of the spines eaten were chosen. We estimated that the external tissue accounted for about 31% of the total ash free dry weight of the meal when all spines were eaten. The total ash free dry weight of edible material (internal plus ex- ternal) increased approximately as the cube of the diameter (1.е. volume) of the test (Fig. 7). The nutritional value of an urchin will, of course, differ greatly according to whether its gonads are fully пре, as in our material, or spent. Since the time taken to penetrate urchins 1$ only a small proportion of the total handling time, the dietary value of urchins would seem to increase monotonically with size. We should expect, therefore, that larger urchins would be preferred to smaller ones, as long as they can be caught easily. There is no evi- dence, however, for the size selection of prey by cassids. The most substantial data are for Cassis tuberosa feeding оп Cassidulus cariboearum (Gladfelter, 1978). In this case, the size frequency of live urchins in the prey population was similar to that of tests cast up on the shore after being drilled by С. tuberosa. At least in shallow water, cassids are usual- ly nocturnal predators that retreat into the sand at daybreak to avoid predation. During the hours of darkness there 15 time to eat only a few urchins and, if the encounter rate with prey is low or uncertain, it may pay to attack the first urchin encountered, irrespective of size. This is a ‘time-minimizing’ as opposed to an ‘епегду maximizing' feeding behaviour (Hughes, 1980). In support of this interpre- tation, we found that on completing a meal, Cassis tuberosa and Cypraecassis testiculus would usually retreat into the sand until the following evening. Size selection of prey, sug- gesting an energy maximizing feeding be- haviour has, however, been recorded in Cymatium nicobaricum, which prefers larger to smaller gastropods (Houbrick 8 Fretter, 1969). It is perhaps significant that this preda- tor will continue feeding during the day, albeit less frequently, than at night. DISCUSSION The ability to drill through the calcareous plates and shells of echinoderms and mol- lusks has evolved independently in the Топпасеа, Naticidae and Muricidae. In the Naticidae and Muricidae drilling is achieved by alternating application of the accessory boring organ (ABO) and the radula to the ex- cavation. The ABO, located in the propodium, secretes a mucoid, slightly acidic, hypertonic fluid rich in hydrogen, chloride and sodium ions. The ABO secretion also contains car- bonic anhydrase and probably other enzymes and chelating agents (Carriker 8 Williams, 1978). Etching of the prey shell begins by the preferential dissolution of the organic matrix, probably by proteolytic enzymes, which has the effect of increasing the surface area of mineral crystals exposed to solubilization and facilitating the removal of shell material by the radula (Carriker, 1978). Minerals are dis- solved by the hydrochloric acid and probably also by a chelating agent and the action of carbonic anhydrase. Calcium ions freed from the shell in the borehole enter the microvilli of the ABO and pass into the foot of the snail, the transmembrane flux of calcium probably being aided by carbonic anhydrase and adenosine triphosphate (Carriker & Williams, 1978). The chemical etching process of the tonnacean proboscis gland (PG) secretion 400 resembles that of the naticid and muricid ABO secretion in that mineral dissolution involves an inorganic acid and probably a chelating agent (Day, 1969). PG secretion, however, 15 produced in much larger quantities than ABO secretion (large Galeodea echinophora can eject up to 1 ml of secretion when irritated [Fánge & Lidman, 1976], compared with the few ul secreted by the ABO of Urosalpinx cinerea follyensis Baker during excavation [Carriker et al., 1978]) and is more acidic (pH 0.13 in G. echinophora [Fánge & Lidman, 1976], pH 1.1 in Argobuccinum argus [Day, 1969], compared with ABO secretion of pH 3.8—4.0 when in contact with seawater in U. cinerea [Carriker et al., 1978]). PG secretion differs further from ABO secretion in having a high concentration of sulfate and a relatively low concentration of chloride ions, a much low- er concentration of organic matter and an ap- parent lack of enzymes (Fánge & Lidman, 1976; Day, 1969). The in vitro calcium car- bonate solubilizing properties of A. argus PG secretion are unimpaired by boiling the secre- tion (Day, 1969), whereas ABO secretion of U. cinerea loses its etching properties after heat- ing to 80°C (Carriker & Williams, 1978). Cassids penetrate echinoid test at a speed of about 0.1 mm per minute in contrast to the muricid Urosalpinx cinerea that penetrates oyster shell at a speed of about 0.3-0.5 mm per day (Carriker & Williams, 1978). Most of this difference in drilling speed is probably at- tributable to the porous structure of echinoid test (Fig. 2c-e) compared with the denser material of bivalve shell. This comparison, however, is complicated by the fact that cas- sids cut out a disc, whereas naticids and muricids Боге a hole. The drilling speed of tonnaceans, if they could be induced to pene- trate bivalve shells, would be of great interest. Day (1969) found that PG secretion of Argobuccinum argus had produced a shallow depression 2-3 hr after being placed on the valve of Macoma sp., but of course the calci- um sulfate accumulating over the etched sur- face would have progressively retarded ero- sion. Day (1969) concluded that the mineral fraction was dissolved faster than the organic fraction of the shell because remnants of the organic matrix atthe edges ofthe eroded area were visible under the microscope. If this in- terpretation is correct, the action oftonnacean PG secretion would contrast with that of naticid and muricid ABO secretion, which at- tacks the organic matrix prior to the dissolu- tion of minerals. HUGHES AND HUGHES The drilling of molluscan shells by ton- naceans has never been recorded with cer- tainty. Molluscivorous cymatiids penetrate their prey through the shell apertures. Cymatium nicobaricum inserts its proboscis into the mantle cavity of its gastropod prey (Houbrick & Fretter, 1969) and Monoplex australasiae Perry pushes its proboscis down through the substratum and in between the valves of its bivalve prey (Laxton, 1971). Sulfuric acid, although adequate for corroding the porous plates and spicules of echino- derms, may not be as effective as ABO se- cretion of naticids and muricids for etching the more solid material of molluscan shells. It is debatable whether the secretion of sul- furic acid by tonnaceans evolved first as a defensive or as an offensive device. Sulfuric acid would seem, a priori, not to be a particu- larly effective agent for attacking prey lacking calcareous armory, yet well developed sul- furic acid-secreting proboscis glands are pos- sessed by all tonnaceans, even though many of them in the Cymatiidae and Bursidae feed on soft bodied prey such as polychaetes, sipunculans, ascidians and sponges (Hou- brick € Fretter, 1969; Laxton, 1971; Taylor, 1978). Sulfuric acid may, however, be effec- tive in overcoming soft-bodied prey. When feeding оп the sabellariid polychaete, Gunnarea capensis (Schmarda), Argobuc- cinum argus inserts its proboscis into the worm's tube and ejects a secretion into the crown of flattened setae that protect the head of the prey. The setae become loosened and can be dislodged by the predator's radula. Further experiments are needed, of course, to determine whether enzymes are secreted in addition to the sulfuric acid. Echinoderms predominate in the diets of tonnaceans; the Cassidae specialize on echinoids; the Tonnidae probably feed largely on holothurians (Bakus, 1973; Grange, 1974; Taylor et al., 1980), although Weber (1927) claimed that Tonna galea accepted echinoids; the Ficidae feed “оп sea urchins and other echinoderms” (Abbott, 19686), but actual data on ficid diets are lacking in the literature; the Bursidae feed on ophiuroids, echinoids and crinoids in addition to non-echinoderm taxa (Taylor, 1978), and members of the genus Charonia within the Cymatiidae feed on echinoids, asteroids and holothurians (Kisch, 1952; McPherson, 1968; Work, 1969; Laxton, 1971; Percharde, 1972; Endean, 1973; Thomassin, 1976). Cymatiid phylogeny is not sufficiently well known for firm conclu- FEEDING IN CASSIDAE 401 sions to be drawn on the origin of echinoderm diets. Phylogenetically more ‘primitive’ recent cymatiids, however, feed on a variety of phyla and it is possible that a specialized diet on echinoderms is not a primitive characteristic. The phylogenetic relationship between the Cassidae and Cymatiidae is obscure, pre- cluding meaningful speculation on the evolu- tion of the cassid feeding method (Taylor, personal communication). Sohl (1969) pub- lished a photograph of the fossil burrowing echinoid Hamea alta (Arnolde 8 Clarke) with a hole in the test that was clearly made by a cassid, and this late Eocene fossil appears to be the earliest known record of cassid feeding activities. ACKNOWLEDGEMENTS We thank Finn Sander for providing facili- ties at the Bellairs Research Institute of McGill University, Barbados. Special thanks go to George Godson, whose unstinted help and knowledge of the habitats of Cassis tuberosa saved us from the brink of disaster due to the lack of animals, to Tom Tomasik for collecting for us the splendid specimen of Cassis flam- mea, and to Norman Runham and Andrew Davis for their skillful use of the scanning electron microscope. Constance Boone, Gene Everson and Porter Kier kindly supplied their unpublished data on cassid diets. Our warm thanks are extended to Melbourne Carriker and Robert Robertson for construc- tively reviewing our manuscript. Financial support enabling us to deliver the paper at the AMU symposium, 1980, was provided by the Royal Society and the American Malacologi- cal Union. Finally, we thank Alan Kohn for in- citing us to take another look at cassids. REFERENCES CITED ABBOTT, В. Т., 1968a, Helmet shells of the world (Cassidae): 1. Indo-Pacific Mollusca, 2: 7-201. АВВОТТ, В. Т., 1968b, Seashells of North America. A guide to field identification. Golden Press, New York, 280 p. ASANO, М. & ITOH, M., 1960, Salivary poison of a marine gastropod, Neptunea arthritica Bernardi, and the seasonal variation in its toxic- ity. Annals of the New York Academy of Sci- ence, 90: 674-688. BAKUS, С. J., 1973, The biology and ecology of tropical holothurians. In: JONES, O. A. 4 ENDEAN, R. (eds.), Biology and geology of coral reefs, 2 Biology 1: 325-367. Academic Press, London. BEU, A. G., HENDERSON, В. A. 8 NELSON, С. S., 1972, Notes on the taphonomy and paleoecol- ogy of New Zealand Tertiary Spatangoida. New Zealand Journal of Geology and Geophysics, 15: 275-286. CARRIKER, M. R., 1978, Ultrastructural analysis of dissolution of shell of the bivalve Mytilus edulis by the accessory boring organ of the gastropod Urosalpinx cinerea. Marine Biology, 48: 105- 134. CARRIKER, М. В. & WILLIAMS, L. G., 1978, The chemical mechanism of shell dissolution by predatory boring gastropods: a review and an hypothesis. Malacologia, 17: 143-156. CARRIKER, М. R., WILLIAMS, L. G. & VAN ZANDT, D., 1978, Preliminary characterisation of the secretion of the accessory boring organ of the shell-penetrating muricid gastropod Uro- salpinx cinerea. Malacologia, 17: 125-142. CHESHER, R. H., 1969, Contributions to the biol- ogy of Meoma ventricosa (Echinoidea: Spatan- goida). Bulletin of Marine Science, 19: 72-110. CORNMAN, I., 1963, Toxic properties of the saliva of Cassis. Nature, 200: 88-89. DAY, J. H., 1969, Feeding of the cymatiid gastro- pod, Argobuccinum argus, in relation to the structure of the proboscis and secretions of the proboscis gland. American Zoologist, 9: 909- 916. DAY, J. H., 1974, А guide to marine life on South African shores, Balkema, Cape Town and Rot- terdam, iii + 300 р. EDMUNDS, J. & EDMUNDS, M., 1973, Preliminary report on the Mollusca of the benthic communi- ties off Tema, Ghana. Proceedings of the Fourth European Malacological Congress; Malaco- logia, 14: 371-376. ENDEAN, R., 1969, Report on investigations made into aspects of the current Acanthaster planci (crown of thorns) infestation of certain reefs of the Great Barrier Reef. Fisheries Branch, Queensland Department of Primary Industries, Brisbane. ЕМОЕАМ, R., 1972, Aspects of molluscan pharma- cology. In: FLORKIN, М. & SCHEER, В. T., (eds.) Chemical Zoology, 7: Mollusca, p. 421— 466. Academic Press, New York, London. ENDEAN, R., 1973, Population explosions of Acanthaster planci and associated destruction of hermatypic corals in the Indo-West Pacific re- gion. In: JONES, O. A. & ENDEAN, В. (eds.), Biology and geology of coral reefs, 2: Biology 1: _389-438. Academic Press, London. РАМСЕ, В. & LIDMAN, U., 1976, Secretion of sul- furic acid in Cassidaria echinophora Lamarck (Mollusca: Mesogastropoda, marine carnivorous snail). Comparative Biochemistry and Physi- ology, 53A: 101-103. FOSTER, R. W., 1947, Cassis tuberosa L. feeding on an echinoid (Clypeaster rosaceus L.). Nauti- lus, 61: 35-36. GLADFELTER, W. В., 1978, General ecology of the cassiculoid urchin Cassidulus caribbearum [sic]. Marine Biology, 47: 149-160. 402 HUGHES AND HUGHES GRANGE, K. R., 1974, The feeding of Tonna ceverisina. Poiriera, 7: 57-61. HENDLER, G., 1977, The differential effects of seasonal stress and predation on the stability of reef-flat echinoid populations. Proceedings Third International Coral Reef Symposium, 1 Biology: 217-223. HOUBRICK, J. В. & FRETTER, V., 1969, Some aspects of the functional anatomy and biology of Cymatium and Bursa. Proceedings of the Mala- cological Society of London, 38: 415—429. HUGHES, R. N., 1980, Optimal foraging theory in the marine context. Oceanography and Marine Biology: an Annual Review, 18: 423—481. HUGHES, В. М. & HUGHES, H. P. 1., 1971, A study of the gastropod Cassis tuberosa (L.) preying upon sea urchins. Journal of experimental marine Biology and Ecology, 7: 215-315. KISCH, В. S., 1952, Further note on Charonia lampas (L.). Journal of Conchology, 23: 266. LAXTON, J. H., 1971, Feeding in some Australas- ian Cymatiidae (Gastropoda: Prosobranchia). Zoological Journal of the Linnaean Society, 50: 1-9. LYMAN, F. B., 1937, Food of Cassis madagascar- iensis. Nautilus, 51: 34. MCPHERSON, В. F., 1968, Contributions to the biology of Eucidaris tribuloides (Lamarck). Bul- letin of Marine Science, 18: 400-443. MOORE, О. R., 1956, Observations of predation on echinoderms by three species of Cassidae. Nautilus, 69: 73-76. NUSKE, H., 1973, Cytologische Untersuchungen an der Sauredruse der Meeresschnecke Cassidaria echinophora. Cytobiologie, 7: 164— 180. PANCERI, P., 1869, Gli organi e la secrezione dellacido solforico nei gasteropodi. Atti dell'Academia delle scienze fisiche e mate- matiche, Napoli, 4: 1-56. PERCHARDE, P. L., 1972, Observations on the gastropod Charonia variegata in Trinidad and Tobago. Nautilus, 85: 84-92. PROFANT, R. J., 1970, Reef wreckers. Tabulata, January 1, 1970: 3-5. (Santa Barbara Malaco- logical Society, Santa Barbara). REYNELL, A., 1905, Some account of {Пе anatomy of Cassidaria rugosa (Linn.). Proceedings of the Malacological Society of London, 6: 292-299. SCHROEDER, R. E., 1962, Urchin killer. Sea Fron- tiers, 8: 156-160. SNYDER, N. & SNYDER, H., 1970, Alarm re- sponse of Diadema antillarum. Science, 168: 276-278. SOHL, N. F., 1969, The fossil record of shell boring by snails. American Zoologist, 9: 725-734. TAYLOR, J. D., 1978, Habitats and diet of preda- tory gastropods at Addu Atoll, Maldives. Journal of experimental marine Biology and Ecology, 31: 83-103. TAYLOR, J. D., MORRIS, М. J. 4 TAYLOR, С. N., 1980, Food specialization and the evolution of predatory prosobranch gastropods. Palaeon- tology, 23: 375—409. THOMASSIN, В. A., 1976, Feeding behaviour of the felt-, sponge-, and coral-feeder sea stars, mainly Culcita schmideliana. Helgoländer wissenschaftliche Meeresuntersuchungen, 28: 51-65. WEBER, H., 1927, Der Darm von Dolium galea L., eine vergleichende anatomische Untersuchung unter besonderer Berucksichtigung der Triton- ium-arten. Zeitschrift für wissenschaftliche Biologie, Abt. A, Zeitschrift für Morphologie und Okologie der Tiere, 8: 663-804. WORK, R. C., 1969, Systematics, ecology and dis- tribution of the mollusks of Los Roques, Venezu- ela. Bulletin of Marine Science, 19: 614-711. YANINEK, J. S., 1978, A comparative survey of reef-associated gastropods at Maziwi Island, Tanzania. Journal of the East African Natural History Society and National Museum, 31: 1-16. MALACOLOGIA, 1981, 20(2): 403-422 SHELL PENETRATION AND FEEDING BY NATICACEAN AND MURICACEAN PREDATORY GASTROPODS: A SYNTHESIS Melbourne R. Carriker College of Marine Studies, University of Delaware, Lewes, Delaware 19958, U.S.A. ABSTRACT Predatory gastropod shell borers occur among the Capulidae, Naticacae, Tonnacea, Muricacea, and Vayssiereidae. With the exception of boring nudibranchs, all known gastropod borers are shelled. This synthesis 15 concerned primarily with naticacean and muricacean borers that excavate smooth, round, beveled holes. They occur in every coastal region of the world that has been examined, and identify prey chemoreceptively. The shell penetrating mechanism includes at least an accessory boring organ (ABO) and radula. The ABO is located in three separate anatomical regions in different groups of borers: in muricaceans, in the sole of the foot anterior to the ventral pedal gland or atop the ventral pedal gland; in naticaceans, under the tip of the proboscis. Studies of the ABO of several species of naticacean and muricacean snails reveal a common ultrastructural form. An acid (possibly HCI) and unidentified chelating agents and enzymes in a hypertonic mucoid secretion released by the ABO are hypothesized to dissolve shell during hole boring. All 33 species of naticacean and muricacean snails examined possess an ABO and are shell borers; the ABO does not appear to have evolved in other shell penetrating molluscs. The role of tubular salivary glands (missing in some muricids and naticids), hypo- branchial glands, and anterior pedal mucous glands in shell penetration is uncertain. Borers release paralytic substances from the hypobranchial gland, and possibly also from other glands associated with the proboscis. Gastropods known to bore holes in prey shell date from the Jurassic and perhaps the late Triassic, some two hundred million years ago. Progress is being made in the control of commercially important species of muricaceans, but not of naticaceans. INTRODUCTION A notable characteristic of many molluscs is their capacity to secrete a protective calcare- ous exoskeleton. Another is ironically the ability of some of these same molluscs to bore or burrow into the shells of other inverte- brates. A voluminous literature has described the structure and development of molluscan shell (Grégoire, 1972; Wilbur, 1972; Watabe 8 Wilbur, 1976), but much less is known about processes of shell breakdown by in- vertebrates and lower plants (Carriker, Smith 8 Wilce, 1969). Molluscan calcibiocavites (Carriker 4 Smith, 1969) have been reported in three classes: the Bivalvia (burrowers), Gastropoda (borers), and Cephalopoda (borers). Among the Gastropoda, shell pene- trating snails occur in the mesogastropod family Capulidae, mesogastropod superfami- lies Naticacea and Tonnacea, neogastropod superfamily Muricacea, and the nudibranch family Vayssiereidae (Carriker, Smith 8 Wilce, 1969). With the exception of boring nudibranchs, all known shell penetrating gastropods possess a shell. (403) This review is concerned primarily with the biology of shell penetration and feeding by predatory gastropods in the superfamilies Naticacea and Muricacea. DISTRIBUTION AND TYPES OF BORING MECHANISMS Every coastal region of the world that has been examined supports populations of bor- ing gastropods (see representative examples in Table 1). Most species are subtropical or tropical, the number increasing toward the equator (Taylor et al., 1980). It is likely that further zoogeographical investigations will locate them off the shores of most land masses (Sohl, 1969). As suggested by the presence of bore holes in prey shells, boring snails range in depth from intertidal zones to at least 2,700 m (Carriker, 1961), and their numbers decrease into deeper water (Taylor et al., 1980). Clarke (1962) lists several species of Naticidae and Muricidae that occur in abyssal regions of the oceans, but it is not known whether they are borers, or whether 404 CARRIKER TABLE 1. Species, source of specimens, and comparative anatomy of accessory boring organ (ABO), ventral pedal gland (VPG), and tubular salivary glands of muricacean and naticacean boring gastropods. S, ABO in anterior midventral sole of foot; aVPG, anterior to ventral pedal gland; р, ABO on anterior ventral tip of proboscis; relative size of tubular salivary glands: O, absent; 1-5, small to large. Nomenclature of North American species based on Abbott (1974). Accessory boring Size organ, location Ventral tubular pedal salivary Species Source Male Female gland gland Muricacea: Bedeva hanleyi Port Jackson, Australia $ aVPG Reduced 1 Eupleura caudata North Carolina to Massa- chusetts, U.S.A. $ aVPG Large 3 E. caudata etterae Virginia, U.S.A. S aVPG Large 4 E. sulcidentata Florida, U.S.A. $ avPG Large 4 Murex brevifrons Puerto Rico S aVPG Reduced 3 M. cellulosus Florida, U.S.A. S aVPG Absent 3 M. florifer Florida, U.S.A. $ aVPG Large 2 M. fulvescens North Carolina, U.S.A. S aVPG Reduced 1 M. pomum Florida, U.S.A. 5 aVPG Absent 0 Muricopsis ostrearum Florida, U.S.A. S aVPG Absent 4 Nucella emarginata Washington, U.S.A. $ avPG Large 5 N. lamellosa Washington, U.S.A. 5 aVPG Large 5 N. lapillus England; Massachusetts, U.S.A. $ aVPG Large 4 Ocenebra erinacea England $ aVPG Large 5 O. inornata (= japonica) Japan; Washington, U.S.A. S aVPG Large 4 Pterorytis foliata Washington, U.S.A. S aVPG Large 3 Purpura clavigera Japan S atop VPG Large 4 Rapana thomasiana Japan S atop VPG Large 5 Thais haemastoma Bimini, Bahamas $ аюр VPG Large 5 T. haemastoma floridana North Carolina, U.S.A. 5 аюр VPG Large 5 T. haemastoma canaliculata West coast Florida, U.S.A. 5 аюр VPG Large 5 T. deltoidea Bimini, Bahamas 5 aVPG Absent 5 Urosalpinx cinerea England; Florida to Massa- chusetts, U.S.A. $ aVPG Large 4 U. cinerea follyensis Virginia, U.S.A. 5 aVPG Large 4 U. perrugata Florida, U.S.A. $ aVPG Large 3 U. tampaensis Florida, U.S.A. 5 aVPG Large 3 Naticacea: Lunatia heros Massachusetts, U.S.A. p p Absent 0 L. lewisi Washington, U.S.A. р р Absent 0 L. triseriata Massachusetts, U.S.A. р р Absent 0 Natica severa Korea р р Absent 0 Neverita didyma Korea р р Absent 0 Polinices duplicatus Florida, North Carolina, Massa- chusetts, U.S.A. р р Absent 0 Sinum perspectivum Florida, North Carolina, U.S.A. р р Absent 0 members of other families are also shell penetrants. A study of gastropod boreholes from the deep sea could clarify some of these questions. Naticacean and muricacean boreholes typically possess smooth walls, beveled outer edges, decreasing diameters with depth, and are generally circular and perpendicular to the shell surface. The typical naticid borehole is a truncated spherical paraboloid; muricid holes, on the other hand, although also variously countersunk, are considerably more varied in vertical section than naticid holes (Carriker & Yochelson, 1968). Nudibranchs (Okadaia elegans; Young, 1969) also excavate smooth, round, beveled holes, while capulids (Capulus danieli; Orr, 1962), and cephalo- pods (Octopus vulgaris; Arnold & Arnold, NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 405 1969; Nixon, 1979) excavate asymmetric, sometimes jagged boreholes. Identificatior. of shell-penetrating molluscs on the basis of their boreholes is thus difficult, except possi- bly for naticids. Although the anatomy of the shell-penetrat- ing mechanism differs among different spe- cies, all 33 naticacean and muricacean spe- cies and subspecies that | have examined possess an accessory boring organ (ABO) and excavate boreholes in the shell of their prey (Table 1) (Carriker, 1961). In all murica- cean males the ABO is located in the mid- anterior ventral part of the foot (Fig. 1). In most muricacean females the organ occurs in the mid-anterior ventral part of the foot but anterior to the ventral pedal gland (the egg capsule gland of some authors) (Fig. 2) when the gland is present. In a small number of muricean females the ABO lies atop, and is continuous with, the ventral pedal gland, so that during its eversion the ABO passes through the cavity of the gland (Fig. 3). In all naticaceans examined, the ABO is located on the anterior ventral lip of the proboscis (Fig. 4). In seven of the muricacean species ex- amined, the ventral pedal gland was absent, or present only as a shallow depression, at the time of dissection, but the ABO was fully formed (Table 1). In these species the ventral pedal gland develops and is functional at the time of oviposition. Fretter & Graham (1962) reviewed hole- boring by the muricids Nucella lapillus, Ocenebra erinacea, and Urosalpinx cinerea, and by the naticids Natica nitida and N. catena. Radwin & Wells (1968) observed bor- ing in the laboratory by Murex pomum, M. fulvescens, M. florifer, M. cellulosus, Muri- copsis ostrearum, Urosalpinx perrugata, and U. tampaensis, and Hemingway (1973, 1975a, b) discussed boring by the muricid Acanthina spirata. Observations on boring by these species corroborate those for similar species listed in Table 1. Tubular salivary glands (accessory salivary glands of some authors) occur in most Muri- cacea (Fretter & Graham, 1962). All the muri- caceans listed in Table 1 possess obvious tubular salivary glands except Murex pomum in which none was found. The size of the glands relative to the height of each snail varies markedly, being rather small in Bedeva hanleyi, Murex florifer arenarius, and M. fulvescens, and largest in the genera Rapana and Thais. No tubular salivary glands were — 1777 FIG. 1. Drawing of sagittal section of anterior part of foot of a male muricacean, Rapana thomasiana, through the accessory boring organ, ABO. S, ABO sinus containing arteries (A), nerves (N), and mus- cles (M) passing to back of ABO. V, ABO vestibule through which ABO is extended to borehole. P, pro- podium, 7, transverse furrow. — ¡mM FIG. 2. Drawing of sagittal section of anterior part of foot of a female muricacean, Urosalpinx cinerea follyensis, through the accessory boring organ, ABO, and ventral pedal gland, VPG. S, ABO sinus. V, ABO vestibule. P, propodium, 7, transverse fur- row. N, nerve. A, artery, M, muscle. found in species of Naticacea. Hemingway (1973, 1975a, b) reported that the tubular salivary glands of Acanthina spirata are simi- lar to those of Urosalpinx cinerea. The vari- able size of tubular salivary glands in most muricaceans, and their absence in natica- ceans and one species of Muricidae, cast 406 CARRIKER — / mm FIG. 3. Drawing of sagittal section of anterior part of foot of a female muricacean, Rapana thomasiana, through the accessory boring organ, ABO, and ventral pedal gland, VPG. The ABO is located atop the ventral pedal gland and in eversion passes through the lumen о the gland. $, ABO sinus, N, nerve. A, artery. M, muscle, V, ABO vestibule. P, propodium. 7, transverse furrow. ABO FIG. 4. Drawing of left side of proboscis of a naticacean. Polinices duplicatus, opened laterally to illustrate relationship of accessory boring organ, ABO, to buccal mass, BM, and to proboscidial hemocoel, PH. RM, retractor muscle. M, mouth. A, radular sac. E, esophagus. ORM, odontophoral retractor muscle. PRM, proboscidial retractor muscle. doubt on the direct functional role of these RESPONSE TO PREY glands in the shell boring process. The curious position of the ABO on top of Muricaceans feed on a wide variety of bi- the ventral pedal gland in species of Purpura, Rapana, and Thais suggests a close affinity of these taxa. Likewise, the absence of this anatomical arrangement in Nucella lapillus supports the contention that the species N. lapillus does not belong in the genus Thais (Abbott, 1974). valves, barnacles, gastropods, small crabs, encrusting bryozoans, and carrion of fish (though they generally select live over dead prey), and may on occasion become canni- balistic (Carriker, 1955; Hanks, 1957; Chew & Eisler, 1958; Fretter & Graham, 1962; Largen, 1967; Radwin & Wells, 1968; NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 407 Morgan, 1972; Menge, 1974; Pratt, 1974a; Bayne 8 Scullard, 1978; Barnett, 1979). Naticaceans, on the other hand, are more restricted in their diet and feed primarily on live bivalves (Hanks, 1952, 1953, 1960; Fretter & Graham, 1962; Franz, 1977; Edwards & Huebner, 1977; Wiltse, 1980). Prey utilization curves of a number of species of small boring gastropods are skewed toward large prey size, and those of large predators are skewed toward small prey size (Sassaman, 1974; but see also Taylor et al., 1980). Boring gastro- pods feed on the flesh of prey through bore- holes excavated by them in the shell of prey, through unbored slits between valves when these are present (as in some bivalves and barnacles), or on gaping prey recently killed by other predators. A curious exception to this is the “commensal” muricid Genkaimurex varicosa that bores a hole in the shells of scal- lops and 1$ thought to “suck juices” from them (Matsukuma, 1977). Response of boring gastropods to prey has been studied primarily in muricids. Under ex- perimental conditions in the laboratory and in the field, these snails can identify preferred live prey some distance away (Carriker, 1955; Wood, 1968; Carriker & Van Zandt, 1972a; Morgan, 1972; Pratt, 1974a). However, all in- dividuals in a population may not respond at the same time. In the laboratory, for example, only 50 to 80% of a population of Urosalpinx cinerea will respond to recently introduced live prey (Carriker, 1957). Nor is preference for prey genetically fixed; existence of prey and predator in similar intertidal zones and relative abundance of prey account for prey selection (Wood, 1968; Pratt, 1974a). U. cinerea can be ingestively conditioned in the laboratory, tending to prefer effluents from a given prey species after it has ingested living tissues of that species (Wood, 1968). Starved U. cinerea are repelled by effluent from starved oyster drills and attracted to the efflu- ent of satiated oyster drills. These responses probably increase foraging efficiency by di- recting snails away from unproductive areas and toward their prey (Pratt, 1974a, 1976). Not all potential prey are attacked by muricids. When, for example, Urosalpinx cinerea 15 confined with a variety of species of bivalves, all are bored except Anomia simplex (Pratt, 1974a; Carriker, Van Zandt & Grant, 1978). Since these snails can bore through empty valves of A. simplex in labora- tory experiments (Carriker, Van Zandt 4 Grant, 1978), it 15 likely that they are sup- pressed by a chemical associated with living A. simplex. Young boring gastropods, recently emerged from egg capsules (Urosalpinx стегеа; Carriker, 1957) and egg collars (Natica диаШепапа; Berg, 1976; Polinices duplicatus; Wiltse, 1980), also are attracted to young prey, bore holes in them and feed on the soft tissues. To what extent and how soon after initiation of feeding young snails become ingestively conditioned is uncertain. The mat- ter requires investigation. Most prey are incapable of defending them- selves against boring gastropods. A striking exception to this is Crepidula fornicata that frequently jabs at an approaching borer with the radula, or dislodges the predator from its valve by pressing the predator against an ob- stacle (Pratt, 1974b). Apparently passive “retribution” on the part of prey occurs occa- sionally. There is a report of an oyster that apparently closed its valves on the proboscis of an Eupleura caudata that was inserted through a hole bored in the margin of the shell, and held the snail until it died. Shell material was then deposited around the predator's shell, permanently affixing it to the oyster's right valve (Burrell, 1975)! Another example is that of Urosalpinx cinerea which in the laboratory can be immobilized by byssi of Mytilus edulis at temperatures at which bi- valves are active but snails have gone into hibernation. This probably does not occur to any extent in the field, as snails move toward the bottom away from mussels as the temper- ature of the seawater drops approximately below 15°С (personal observations). Boring gastropods possess (a) chemo- receptive mechanism(s) for detecting prey and approaching them from a distance. Snails respond to chemical substances characteris- tic of the effluents of prey species they have eaten (Wood, 1968; Carriker 8 Van Zanat, 1972b; Morgan, 1972; Pratt, 1974a). Although attractiveness of prey is often marked and responded to by a large proportion of a preda- tor population, the chemical stimulus that guides predators to prey has been identified only as one or more of the metabolic products of prey (Carriker 8 Van Zandt, 1972b). Experimentation on carnivorous meso- gastropods and neogastropods (Kohn, 1961; Crisp, 1973; Newell 8 Brown, 1977) suggests that the osphradium plays a primary role in distance chemoreception. The function of the 408 CARRIKER mantle edge, tentacles, and propodium in sensing chemical cues is probably also impor- tant and needs further investigation. Primary recognition of the immediate pres- ence of prey by muricaceans appears to de- pend on identification of a chemical cue in the exhalant water of prey. Snails creep over the bottom toward their prey, locating them most rapidly when the prey are on the upstream side of tidal currents (Carriker, 1955). Wheth- er snails respond to the same chemicals from prey at a distance and close to prey, is un- clear. When approaching actively pumping prey, Urosalpinx cinerea, for example, often raises the anterior part of the foot, stands on the posterior tip of the foot, propodium and tentacles fully extended, and swings the pro- podium back and forth in a pattern suggestive of searching. Whether distance or close- range attractant(s), or both, in exhalant sea- water is reinforced by a further stimulus as- sociated with the prey is uncertain. Reinforce- ment might come from valvular movements of the prey, chemical attractant adsorbed to the shell, topography of the prey shell, chemicals in the organic matrix of shell, or even unknown cues from the animal within the shell (Carriker & Van Zandt, 1972b). Pratt (1974a) reported that epibiota on the shell of prey did not play an important role in oyster drills' attacks on Crepidula fornicata. In laboratory experi- ments, Carriker & Van Zandt (1972b) noted that something on the surface of oyster valves, possibly microorganisms, enriched by effluent from pumping oysters, attracted snails to the oysters, but did not stimulate them to bore the shell. The problem needs clarification. Very little information is available on the behavior of prey recognition by naticaceans (Kohn, 1961; Fretter & Graham, 1962; Carri- ker & Yochelson, 1968). The burrowing habit of these snails makes them difficult subjects for this kind of research. The ability of boring gastropods to detect prey is influenced by environmental factors. For example, response to prey by Urosalpinx cinerea declines as temperature of the sea- water drops in the fall from 15 to 7°C, depend- ing on the latitude and other environmental factors (Carriker, 1954; Carriker & Van Zandt, 1972a). A salinity of 12.5 %oo is near the lower limit for location of prey by both U. cinerea and Eupleura caudata (Manzi, 1970). Feeding activities of Thais haemastoma stop at tempeatures of 10°C and below (Gunther, 1979). Such naticaceans as Polinices dupli- catus in temperate zones cease to identify prey at about 5°C and a salinity of 6 %oo, whereas Lunatia heros, a species found generally in deeper water than P. duplicatus, continues its activities at temperatures as low as 2°C but to a salinity of only 10 2/00 (Hanks, 1952, 1953; Edwards & Huebner, 1977; Carriker, unpublished observations). PENETRATION OF PREY Selection of Borehole Site Muricaceans Little is known about borehole site selection by boring gastropods. Urosalpinx cinerea, after crawling onto an epifaunal bivalve, for example, undertakes a series of exploratory activities leading to selection of the penetra- tion site. Exploration can range from a few minutes to half an hour. During the search the proboscis is extended intermittently to the shell surface, and, its tip undulating with mi- nute wave-like movements, is passed slowly over the substratum, stopping now and then to rasp at small, live, sessile organisms (Carriker & Van Zandt, 1972a). What determines the specific site for boring is unclear. Nor is it known whether individuals express a consistent preference for a particu- lar part of the shell surface of successive prey, or whether an environmental cue plays a part in selection. Urosalpinx cinerea (Carri- ker & Van Zandt, 1972b) and Nucella lapillus (Morgan, 1972) appear to excavate boreholes randomly on prey valves, though U. cinerea locates its holes primarily away from the edge of the valves, reflecting avoidance of valve edges probably because of valvular motion. Breaks in valves away from valve edges, or along valve edges when valves are held shut by rubber bands, are quickly located and used as penetration sites in lieu of boring through solid shell. It appears that metabolites from active living, non-wounded prey not only trig- ger the initial attack on prey, but also deter- mine penetration sites when seepage occurs through tiny holes between valve edges. Thus, tightly closed living oysters are not penetrated, nor are empty valves bored even in the presence of attractant from pumping oysters nearby (Carriker & Van Zandt, 1972a). In contrast to U. cinerea, Acanthina spirata bores holes most commonly at the margin of the prey valves (Hemingway, 1973), NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 409 and Dicathais aegrota, away from the margin of the univalve of the limpet (Black, 1978). Naticaceans A series of behavioral patterns involving prey capture and prey manipulation, present upon metamorphosis of the snails, deter- mines the position of the borehole in this group (Berg, 1976). These gastropods, char- acterized by an exceptionally large, flat foot that facilitates their movements within the sediment and with which they tightly grip their prey, crawl through clean to slightly muddy sand both above and below the sediment- water interface. When infaunal prey are lo- cated, probably chemoreceptively, snails bur- row rapidly to their level, and generally bore into the shell below the benthic surface. In the process of prey capture, these naticids secrete copious quantities of mucus. In the laboratory Lunatia nitida covers its prey with mucus to help hold the prey closed and prevent it from escaping (Richter, 1962). In some cases, after coating its prey, L. nitida tows the bivalve behind it by a rope of mucus, the prey held closed by the mucus sheet until the snail is ready to bore into it. L. heros, like- wise in the laboratory, sometimes places a bivalve in a pocket formed by underfolding of the posterior part of the foot, and сатез the prey there until ready to consume it (personal observation). Positions for boring seem to be related to the manner in which prey are grasped, and holes are thus usually limited to a small area of prey valves, commonly on one valve more frequently than the other. Position of bore- holes appears to vary with the species of predator and prey (Boettger, 1930; Ziegel- теег, 1954; Fretter & Graham, 1962; Carri- ker 8 Yochelson, 1968; Taylor et al., 1980). Berg (1976) found that after metamorphosis young Natica gualtieriana bored their first prey by a single hole in a stereotyped position. As these snails matured and gained experi- ence at boring, there was no change in the angular distribution of the boreholes in each whorl, but whorl preference changed. Shell Penetration Muricaceans All muricids that have been studied closely employ a similar chemical-mechanical mechanism for penetration of prey valves though the manner of penetration may vary (Carriker & Van Zandt, 1972b; Morgan, 1972; Gunter, 1979). For example, once Urosalpinx cinerea has commenced excavation of a borehole, it continues until penetration has been completed. Only dislodgment of the snail by exterior forces or precipitous environ- mental changes are apt to terminate boring; and even then, many snails, if remaining close by the borehole, will return to the hole. U. стегеа can penetrate the shell of its prey in the absence of the live animal, pro- vided boring has been initiated on live whole prey. Thus, boreholes once started can be completed without stimulation of any kind from live prey (Carriker 8 Van Zanat, 1972a; Carriker, Van Zandt & Grant, 1978). On the other hand a young Thais haemastoma bores holes on a valve until it reaches a height of 5 cm; at larger sizes it penetrates at valve edges apparently relaxing prey by a paralytic substance, and in one-third of the oysters consumed, boring no hole (McGraw 4 Gunter, 1972; Krutak, 1977; Gunter, 1979). Initial identification of a boring site by Urosalpinx cinerea is made by the propodium and by the proboscis tip. In early stages of exploration, the snail frequently extends and passes its proboscis over the spot, and occa- sionally the mouth opens and the buccal cavi- ty enlarges in what appears to be a “tasting” reaction, Anterior propodial ridges are used only partially, and sometimes not at all, in supporting the proboscis during search for a penetrating site (Carriker, 1943; Carriker 8 Van Zanat, 1972a). After the boring site is selected, the snail positions itself on the shell surface with the pore of the retracted ABO located over the prospective boring site. Thereafter the poste- rior part of the foot remains firmly attached to the shell in the same position. The anterior part of the propodium is then retracted deeply, and the lateral propodial ridges are over- folded, forming a fleshy tube over the bore- hole site down which the proboscis is ex- tended. Rasping is limited principally to the bottom of the incomplete borehole. The odontophore can rotate on its long axis in- dependent of rotation of the proboscis by at least 180°; thus, by swinging to the left and then to the right in two half turns, the odonto- phore covers the circumference of the bore- hole. Rasping over the surface of the incom- plete borehole by the radula is uniformly firm, and the pattern of rasping appears random (Carriker & Van Zanat, 1972a). 410 CARRIKER After the brief rasping period, the proboscis is infolded into the cephalic hemocoel. Simul- taneously the mid-anterior part of the pro- podium, already at the posterior edge of the borehole where it surrounded the proboscis, is extended into the borehole. The propodium then presses the transverse furrow (Fretter & Graham, 1962) closely against the shell, slides it forward across the surface of the incom- plete borehole and back onto the surface of the shell to assume a normally extended position and a tight contact between the epi- thelium of the snail's foot and the prey's shell. In this maneuver the propodium voids sea- water from the incomplete borehole prior to entrance of the ABO. The propodium is fol- lowed immediately by the ABO which slides gently into position, and presses closely against the shell surface. Once in position, the organ continues to pulsate gently. During its stay in the borehole, the organ secretes solu- bilizing fluid that removes a thin layer of shell at the bottom and obliterates most of the marks of the previous rasping period. After the period of shell dissolution, the ABO 1$ withdrawn from the borehole. Simultaneously, the propodial tube is formed, the proboscis 15 extended into the borehole to resume rasping, and a new penetration cycle commences. As soon as the borehole is completed and the break into the extrapallial space of the bivalve is large enough to admit the proboscis, the snail presses the proboscis against the flesh and starts feeding (Carriker & Van Zanat, 1972a). The boring behavior of Eupleura caudata, as observed in oyster models (Carri- ker & Van Zandt, 1972a), is identical to that of Urosalpinx cinerea. The boring behavior of Nucella lapillus is said to be similar to that of U. стегеа (Morgan, 1972). Using a motion picture camera taking single exposures every 1.5 minutes, Morgan showed that in the period of 73.3 hours required to Боге, М. lapillus, like U. cinerea, moved its position on the prey only slightly. Naticaceans Because these snails wrap prey in the foot during boring and bore primarily when buried in the sand (Fretter 8 Graham, 1962), it is difficult to study their shell-penetration proc- ess. Ziegelmeier's (1954) account of Lunatia nitida 1$ the most detailed. The bivalve is held by the propodium, which overfolds much as does that of muricids, to form a fleshy tube down which the long proboscis is extended from the cephalic hemocoel to the surface of the prey shell. During penetration the pro- boscis is rotated a 90” quadrant at a time so that rasping is done systematically sector by sector from the center of the incomplete bore- hole to the periphery. The center of the hole, where the least radular rasping occurs, thus results in a boss characteristic of incomplete naticacean boreholes. After the rasping of a quadrant, the proboscis is raised from the incomplete borehole and the ABO, located under the ventral lip, is placed in the hole. (Ziegelmeier was not able to see the change in position.) As in muricids, the ABO solubilizes the surface layer of shell in the borehole, and the weakened shell is rasped free by the radula during the next round of mechanical boring. In Urosalpinx cinerea the process of hole boring is easily observed in an oyster model (Carriker & Van Zandt, 1972a); по apparatus has yet been devised to permit viewing of the process in naticaceans. None- theless, from the information available, and from general observations on feeding by Lunatia heros, L. triseriata, and Polinices duplicatus in the laboratory (Carriker, per- sonal observations), it appears that the mechanism of shell penetration in muri- caceans and naticaceans is similar (see also Fretter 8 Graham, 1962). Proboscis and Radula Proboscis A long proboscis evolved in prosobranchs that feed on food not immediately accessible to them (Fretter 8 Graham, 1962; Graham, 1973). In boring prosobranchs, the length of the proboscis is about as long as the height of the shell. This is a distinct advantage because predators can not only bore a hole in the shell of prey, but can also extend the proboscis deep into prey to feed safely within a wide radius of soft tissues until the valves of prey gape. When valves open, nearby predators, especially small crabs, join in feeding. In view of the predatory success of both groups of snails, the muricacean pleurembolic and the naticacean acrembolic types of proboscides appear to be equally effective (Carriker, 1943). After the muricacean proboscis 1$ amputated accidently by being pinched be- tween valves of prey, by small crabs feeding alongside the proboscis in gaping prey, or by experimental procedures т the laboratory, it regenerates rapidly to its former size and NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 411 function (Urosalpinx cinerea, Eupleura caudata: Carriker, Person, Libbin & Van Zandt, 1972; Thais haemastoma: Gunter, 1968). Loss of this important organ is thus not fatal, as the snail possesses enough meta- bolic reserves to survive until a new proboscis has formed. In the absence of the proboscis the snail is unable to bore, even though the ABO 15 present. The regenerative capacity of the proboscis of naticaceans has not been tested, but it is likely that it, too, can reform in the event of accidental proboscisectomy. Radula Although radulae of muricaceans (rachi- glossan, formula 1 + R + 1) and naticaceans (taenioglossan, formula 2 + 1 + R +1 + 2) differ in organization, they are both long, slender structures limited to a few teeth in each transverse row. The narrow radula 1$ admirably adapted to hole boring, the central rachidian tooth in each row bearing the brunt of rasping over the surface of boreholes and the marginal teeth serving synchronously with rachidian teeth in tearing flesh from prey (Car- riker, Schaadt & Peters, 1974; Krutak, 1977). The radula of boring gastropods has been a favorite subject for light (25 species of muri- cids: Wu, 1965b) and scanning electron microscopy (Urosalpinx cinerea: Carriker € Van Zandt, 1972a, Carriker, Schaadt & Peters, 1974; Nucella lapillus: Runham, 1969; several species of Acanthina and Eupleura triquetra: Hemingway, 1975а, b; Thais haemastoma: Krutak, 1977). Scanning microscopy shows admirably the successive locking of each tooth over its neighbor, spreading the impact against the shell surface over several rachidian teeth as the radula slides over the tip of the odontophore against the borehole. Independent forward movement of the radula over odontophoral cartilages as the radula scrapes forward against the bore- hole adds efficiency to the shell-rasping process and spreads the wear of cusp tips over several rows of rachidian teeth (Carriker 8 Van Zandt, 1972a; Carriker, Schaadt 4 Peters, 1974). Hole boring wears the teeth down to their base. Gradual replacement of the radula by formation of new teeth in the radular sac insures that a supply of sharp teeth 15 available for each successive round of shell-boring (Isarankura & Runham, 1968). Hardness of radular teeth is known only for muricids (Carriker & Van Zandt, 1972b); naticid teeth have not been tested. The mar- ginal teeth of Urosalpinx cinerea are about twice as hard as rachidian teeth, and the latter are about the same hardness as oyster shell. Thus, without the aid of the solubilizer secret- ed by the accessory boring organ, the radula would make little progress into the shell. Cal- cium is a major chemical element of the teeth of U. cinerea and strontium and silicon are present as major to trace constituents (Car- riker & Van Zandt, 1972a). Abrasion of radu- lar teeth during boring wears cusps smoothly. No sharpening occurs as it does in teeth of the grazer, Patella vulgata, in which the lead- ing edge of each tooth is backed by a softer region that insures self-sharpening of this edge during wear (Runham, Thornton, Shaw 8 Wayte, 1969). Unworn teeth of boring gastropods are ex- ceedingly sharp and could readily shred the lining of the buccal cavity during boring and feeding. This is generally avoided by a protec- tive, flexible, cuticularized buccal armature that lines the buccal cavity and prevents damage to buccal tissues. Even so, light abrasion still occurs on the more elevated parts of the buccal lining, but this lining is augmented further by secretion from the buc- cal epithelia (Carriker, Schaadt & Peters, 1974). As demonstrated in Urosalpinx cinerea, gastropod borers swallow fragments of shell rasped from the borehole during penetration of prey (Carriker, 1977). Depending on their orientation relative to the surface of the in- complete borehole, shell units (prisms, lamel- lae) are broken off, coated with secretion from the ABO, and further pelleted by mucus on their passage down the alimentary canal to be voided as feces. The envelope of mucoid ma- terial undoubtedly reduces or prevents lacera- tion of the epithelium of the alimentary canal. Naticaceans also swallow shell fragments scraped from the borehole (Ziegelmeier, 1954; Fretter & Graham, 1962). These also pass down the esophagus and appear out- side the anus as white fecal pellets. Since most shell excavated from boreholes appears to be discharged through the anus, it is ques- tionable that minerals in shell fragments are used metabolically by snails to any extent. The matter should be investigated by tagging shell of prey with radioactive calcium. Accessory Boring Organ The ABO is an essential component of the shell penetrating mechanism of boring gastro- 412 CARRIKER FIG. 5. Light micrograph of histological sagittal section of accessory boring organ of Urosalpinx cinerea follyensis extended from foot. S, secretory epithelium. C, connective tissue in center of organ supporting retractor muscles, capillaries, and nerve fibers. Organ 1 mm in diameter. pods (Fig. 5). When this organ is removed from, for example, Игоза/ртх стегеа, by ex- perimental excision, the snail recovers, but is unable to bore, even though the proboscis is present and functional. The organ regener- ates relatively rapidly, and the muricid soon resumes boring (Carriker & Van Zandt, 1972a). The effect of removal of the ABO on the shell penetrating capacity of a naticacean has not been determined, but is likely similar to that observed in muricids. А! species of boring muricacean and naticacean gastropods that have been stud- ied to date possess an АВО, but these consti- tute only a small sample of the large number of species of boring gastropods that exist in the world oceans. Many more species need to be examined before we can generalize on the universality of a shell solubilizing gland in bor- ing gastropods. Detailed structural studies carried out so far on the ABO of two species of muricids (Urosalpinx cinerea: Nylen, Provenza & Carriker, 1969; and Nucella lapillus: Chétail, Binot & Bensalem, 1968; Derer, 1975; Webb 8 Saleuddin, 1977) and one species of naticid (Polinices lewisi: Bernard & Bagshaw, 1969) show that the histology and fine structure of the secretory disc of the organ is similar in the three species. The organ of the naticid differs from that of the muricid organ in possessing a peripheral zone of subdermal mucocytes around the central disc. The peduncle that supports the disc is long and cylindrical in muricids, to accommodate the position of the gland deep within the foot, and short in nati- cids, in which the organ is attached to the lower lip of the proboscis (Webb & Saleuddin, 1977). The secretory disc of the muricid and nati- cid ABO is composed of a single layer of tall columnar epithelial cells arranged in groups. A brush border of unusually long, densely packed microvilli covers the surface of the disc. The center of the organ consists of con- nective tissue that supports muscles, capil- laries, and nerve fiber bundles passing to the base of the secretory epithelium. Dense populations of mitochondria are present near the surface of the epithelium, more abundant in secreting (ABO's of actively boring snails) than in resting (non-boring snails) secreting cells. Dense membrane-bound secretory granules, multivesicular bodies, and single NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 413 vesicles are also conspicuous in the cells (Nylen, Provenza 8 Carriker, 1968, 1969; Chétail, Вто! & Bensalem, 1968; Bernard & Bagshaw, 1969; Derer, 1975; Webb & Saleuddin, 1977). Whereas the mechanical phase of shell penetration by the radula is well understood (Carriker, Schaadt & Peters, 1974), knowl- edge of the chemical phase 1$ still in a hy- pothetical state. Earliest physiological ге- search on the ABO of muricids disclosed, a) a pH ranging from 3.8 to 4.1 in the released secretion of a normally functioning gland (Carriker, Van Zandt & Charlton, 1967), b) active aerobic metabolism in the secretory cells (Person, Smarsh, Lipson & Carriker, 1967), and, c) substantial amounts of car- bonic anhydrase in the organ (Carriker, Person, Smarsh, Lipson & Chauncey, 1968; Chetail, Binot & Bensalem, 1968). Subse- quent research on the chemical phase of penetration was summarized by Carriker & Williams (1978). They hypothesized that a combination of enzymes, an inorganic acid (possibly HCl), and possibly chelating agents is employed in a hypertonic secretion to facili- tate dissolution of shell and intracellular transport of calcium during the chemical phase of shell penetration. Secretion granules and vesicles in the secretory epithelium of the ABO and in the released secretion, organic matter in the secretion, and inactivation by heat and papain of the etching capacity of ex- cised ABO’s suggest the presence of en- zymes. Hydrogen, chloride, and sodium ion concentrations demonstrate the hypertonic and acidic nature of the released secretion. An unidentified chelating agent and a muco- protein appear to be present in the secretory epithelium; the latter perhaps is the chelator. In a study of lysosomal enzymes, acid phos- phatase, and carbonic anhydrase in the ABO of Nucella lapillus, Webb & Saleuddin (1977) concluded that there is minimal involvement of extracellular enzymes in the boring proc- ess. They postulated, instead, that hydrogen ions, derived from hydration of metabolic car- bon dioxide, are released by the secretory epithelium for dissolution of calcium carbon- ate of the shell. This supports the earlier find- ings by Carriker, Van Zandt & Carlton (1967) on the pH of the secretion in Urosalpinx cinerea. However, the findings of Webb & Saleuddin (1977) on extracellular enzymes are at variance with those of Carriker and Chauncey (1973) who reported that released secretion collected from live U. cinerea was positive for carbonic anhydrase. The similarity determined by scanning electron microscopy of ultrastructural patterns of dissolution etched in the shell of Mytilus edulis by the secretion of the ABO and those produced artificially by НС! and ethylene- diaminotetra-acetic acid, suggest that these chemicals, or similar ones are constituents of the secretion of the ABO. Lactic and succinic acids and a chitinase-like enzyme were also suggested as possible components. How- ever, alteration of shell fracture surfaces by experimental application of these and other chemical agents was not sufficiently compar- able to that etched by the secretion of the ABO to support this suggestion. A marked variation in the rate of dissolution of different ultra-structural parts of the mineral components of shell occurs in shell surfaces when they are etched by the secretion of the ABO (Carriker, 1978). As differential dissolu- tion could result, in part, from variation in the composition of trace and minor mineral con- stituents of shell, Carriker, Van Zandt & Grant (1978) tested the capacity of Urosalpinx cinerea to penetrate several kinds of non- molluscan minerals commonly present in trace or minor amounts in bivalve shell. The rate of penetration of these minerals de- creased in the following order: bivalve (mainly CaCO,) shell, strontianite (SrCO,), anhydrite (CaSO,), witherite (BaCO3), and magnesite (MgCO3), lending support to the original hy- pothesis of differential dissolution (Carriker, 1978). A variety of biogenically formed calcareous minerals was also tested, and all of these, except the radula of U. cinerea, were penetrated. The relative resistance of radular teeth to dissolution by the secretion is not unexpected, since the radula is exposed to the secretion for a relatively long time during penetration. Clearly, much more research must be carried out on the chemical phase of penetration before the mechanism can be fully understood. The anatomical location and structure of organs involved in hole boring by other mol- luscan penetrants such as cymatiid meso- gastropods (Day, 1969), vayssiereid nudi- branchs (Young, 1969), and octopuses (Nixon, in press) are significantly different from those of the accessory boring organ in muricaceans and naticaceans, yet the shell of their prey is penetrated effectively. Study of the chemical mechanism of shell excavation by these predators should provide a deeper understanding of shell penetration by muricaceans and naticaceans than is now available. 414 CARRIKER Tubular Salivary Glands Two kidney-shaped, muscular, tubular sali- vary glands (also known as accessory sali- vary glands) discharge through a common duct into the ventral lip of the mouth of most muricaceans (Table 1; Graham, 1941; Car- riker, 1943; Fretter 8 Graham, 1962; Carriker, 1977). These glands are distinctive morpho- logical features of the Muricidae (Ponder, 1973). Four separate functions have been hypothesized for the glands: Lubrication. Discharge of the secretion of the glands into the path of the functioning odontophore suggests a source (in addition to that from the salivary glands) of lubricant for the radula during the boring process (Fretter 8 Graham, 1962). This suggestion is sup- ported by the fact that the spongy layer about the mouth and opening of the tubular salivary gland duct in living Urosalpinx cinerea stain a deep purple-red color with methylene blue. (The only other external structures in the snail giving a similar staining reaction are the ven- tral and lateral surfaces of the foot that se- crete copious quantities of mucus.) (Carriker, 1943). Furthermore, extracts of the glands of Nucella lapillus and Ocenebra erinacea have a pH of about 6.0, application of the glands or their extracts to the polished inner surface of mollusc shell leaves no etched mark and no proteolytic or amylolytic enzymes appear to be present (Graham, 1941). In contrast the secretion of the ABO when applied to pol- ished shell does etch (Carriker & Van Zanat, 1964). Hole boring. That the glands could also be involved in shell penetration may be deduced from their position in the distal end of the pro- boscis, but there 15 little else to support this conjecture. These glands are present in most muricaceans (Table 1) in which they vary in relative size, and are absent in naticaceans and apparently also in nonboring gastropods. Conceivably the unexplained role of muri- cacean tubular salivary glands could be equivalent to that of the mucocytes that sur- round the naticacean ABO (Carriker, 1977), but there is no information on this. What struc- ture replaces the tubular salivary glands in muricaceans that lack them has not been de- termined. Paralysis. The histological resemblance be- tween tubular salivary glands and the poison gland of toxoglossans (Graham, 1941; Fretter 8 Graham, 1962) suggested to Graham (1941) and to Mantoja (1971) that tubular sali- vary glands could produce some toxic sub- stance. Graham (1941), however, found that their extract has no effect on the heart of Cardium sp., and noted that many prey of bor- ing gastropods are sedentary and do not have to be paralyzed before consumption. A further point that might have a bearing on the prob- lem is that salivary glands of stenoglossans (Ocenebra aciculata, for example) lack alkaline phosphatase in their cells, while both the tubular salivary glands and the gland of Leiblein are rich in this enzyme (Franc, 1952; Fretter & Graham, 1962). Because of their intrinsic biological interest, and their possible involvement in shell pene- tration, tubular salivary glands of muricaceans deserve further attention. Extracorporeal Enzymes From experiments on attraction of hermit crabs to simulated gastropod predation sites, Rittschof (1980, in press) suggested that gastropod predators (such as the fasciolariids Pleuroploca gigantea and Fasciolaria tulipa) release a protease while feeding. Peptides released from gastropod prey while predators consume prey flesh serve as cues that en- hance the attractiveness of prey several times over that of prey flesh alone. Rittschof sup- ported his hypothesis by addition of trypsin to prey flesh which in the absence of a predator made the flesh as attractive to hermit crabs as was prey flesh being actively consumed by a gastropod predator. This finding has significant implications for the study of shell penetration. Boring gastro- pods possess salivary glands, buccal glands, and in the case of most muricaceans, tubular salivary glands that empty directly into the buccal cavity. Mansour-Bek (1934) reported the presence of proteolytic enzymes including a trypsin-like protease in the saliva (presum- ably from the salivary glands) of Murex anguliferus. Enzymes discharged into the buccal cavity around odontophore and radula could easily trickle into the borehole during the rasping phase of shell penetration, and if a constituent of the secretion were a conchio- linase-type enzyme, attack the organic com- ponents of the shell (Carriker, 1969; Travis & Gonsalves, 1969). If the enzymes aid in shell penetration, they should be demonstrable during boring but prior to feeding. Preliminary attempts by Carriker (1978) to identify enzymes that hydrolyze the intercrystalline organic matrix of shell were inconclusive and NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 415 should be repeated. Until now, we have hy- pothesized that shell solubilizing enzymes, if present, are secreted by the ABO (Carriker 8 Williams, 1978). Identification of hydrolytic enzymes in the buccal region of boring gas- tropods, and testing of these enzymes on shell preparations should thus provide addi- tional information on the chemical phase of shell penetration. Anterior Pedal Mucous Gland This gland is a collection of clusters of sub- epithelial secretory cells arranged in nests in the anterior part of the foot. The gland dis- charges into a sagittal canal that empties into the transverse furrow between the propodium and the podium (Fretter & Graham, 1962). The propodium sweeps across the bottom of the incomplete borehole during boring, and the furrow, in an anatomical position to wipe secretion over the surface of the hole, 1$ car- ried along. Most of the cells of the gland stain so as to suggest that their secretion contains mucoprotein. These constituents, if present, could function as chelating agents in solubili- zation (Carriker & Williams, 1978). The pH of the secretion in the furrow ranges from 7.0 to 7.8 (Carriker, Williams & Van Zanat, 1978). However, shell etched by the secretion from the ABO in the absence of furrow secretion, revealed the normal pattern of dissolution found in boreholes (Carriker, 1978). The role of the secretion in shell penetration is thus uncertain; at the least the secretion could serve as a lubricant and as a sealant to hold the ABO secretion within the bore hole. Study of the gland needs to be undertaken before the chemical mechanism of shell penetration by boring gastropods can be fully understood. PARALYSIS OF PREY That some muricacean gastropods syn- thesize biotoxins to quiet or kill their prey has been suspected for some time (Gunter, 1968). For example, while most boring gas- tropods bore a hole large enough to admit the proboscis, adult Thais haemastoma Боге comparatively small holes at the valve mar- gins that do not admit the proboscis. This fact, together with behavioral observations, sug- gested to McGraw 8 Gunter (1972) and Gunter (1968, 1979) that T. haemastoma in- jects a paralytic substance into prey that causes them to gape and die. Paralytic agents, elaborated in the hypo- branchial gland (Whittaker 8 Michelson, 1954: Whittaker, 1960; Endean, 1972; Hem- ingway, 1978), have been identified as pharmacologically active esters of choline: urocanylcholine (in Murex trunculus, M. fulvescens, Ocenebra erinacea, Nucella lapillus, and Urosalpinx cinerea), and senecioylcholine (т Thais floridana). Acrylylcholine is present in the nonboring snail Buccinum undatum, but по choline esters occur in Busycon canaliculatum or in several species of taenioglossans (Whittaker, 1960). The salivary glands of nonboring spe- cies of buccinids and cymatiids contain tetramine in addition to choline esters. The hypobranchial gland secretes mucus contain- ing both Tyrian purple and the choline ester that is probably carried to prey by ciliary cur- rents on the surface of the mantle and pro- podium (Whittaker, 1960; Hemingway, 1978). Urocanylcholine has marked hypertensive as well as a neuromuscular blocking action. Senecioylcholine resembles urocanylcholine but is somewhat less active as a blocking agent, acrylylcholine has only an extremely brief and feeble blocking action (Whittaker, 1960). The first two biotoxins are present in shell boring gastropods and the third in a nonboring gastropod. A paralytic substance with a high acetyl- choline equivalency is also present in the com- bined salivary and tubular salivary gland complex (as well as in the hypobranchial gland) of the muricid, Acanthina spirata (Hemingway, 1973, 1978). As analyses were performed on the combined glands, it is not clear whether one or both of the glands re- lease the biotoxin. Graham's (1941) report, that extract of tubular salivary glands has no effect when injected into the heart of a bi- valve, suggests that the acetylcholine is pro- duced by the salivary glands. The matter re- quires verification. Hemingway (1978) noted that different choline esters in the hypobranchial glands of predatory gastropods may be as numerous as the species of muricaceans (see also Whit- taker, 1960). The apparent specificity of choline esters from these glands led Feare (1971) to make the provocative suggestion that choline esters released by them could also be involved in species recognition or mating behavior! It is understandable that a predator, like Buccinum undatum, which at- tacks bivalve prey without boring through the shell, would be aided in attacking prey by pro- 416 CARRIKER ducing acrylylcholine, but not why shell- penetrating muricaceans, which prey on bi- valves that are generally sedentary (Graham, 1941), release urocanylcholine that has a strong blocking action. No reports are available on whether glands in the proboscis or mantle cavity of nati- caceans emit paralytic chemicals. Since these gastropods bind prey in large quantities of mucus during capture and manipulation prior to boring, the mucus itself, secreted presumably by pedal surfaces, could contain paralytic substances. These interesting pos- sibilities call for attention. EVOLUTION The greatest known concentration of muri- cacean and naticacean borers occurs in shal- low water around continents in tropical lati- tudes (Carriker, 1961; Taylor et al., 1980). Since no boring gastropods have been dis- covered in freshwater (Carriker & Smith, 1969), and relatively few borers have been reported from the deep-sea (Carriker, 1961; Taylor et al., 1980), it is likely the shell boring habit in prosobranchs evolved in relatively shallow, tropical, marine waters (see also Clarke, 1962). Gastropods presently known to bore holes in shells of prey date back to the Jurassic and perhaps as early as the Late Triassic, some 200 million years ago (Carriker & Yochelson, 1968; Sohl, 1969; Ponder, 1973; Krutak, 1977; Taylor et al., 1980). Evolution of the shell-penetrating mechanism in muricaceans and naticaceans could have taken place in three major morphological steps in this order in geologic time: a) development of the radula (Firby & Durham, 1974; Krutak, 1977; Taylor et al., 1980), b) elongation of the head to form a proboscis (Graham, 1973), and c) formation of the accessory boring organ (Carriker, 1943; Fretter, 1941, 1946). The mechanism for se- cretion of paralytic substances could have evolved after the appearance of the radula (Taylor et al., 1980) and could have pre- adapted snails to become predators of non- shelled prey. Appearance of the ABO in two separate anatomical locations among muricaceans (in front of the ventral pedal gland, and atop the ventral pedal gland) and in an entirely dif- ferent region in naticaceans—under the pro- boscis tip (Carriker, 1961)—is an enigma. Dif- ference in the position of the organ in the two superfamilies might be attributed to the strik- ingly different epifaunal and infaunal boring behaviors, respectively, of the two groups. However, the general position of the anterior central part of the foot of the predator on its prey, the placement of the organ in the bore- hole, and alternation of radula and organ in the borehole during penetration are similar in the two superfamiles and within the muri- caceans. A comparative embryological study of the development of the ABO in representa- tive muricaceans and naticaceans is urgently needed to determine whether the organ de- velops anew in its respective anatomical spot in different groups, or is formed in one place and migrates to its definitive position in the adult. In any event, the development of such similar organs as the ABO on different parts of the body is one of the most interesting paral- lels in molluscan morphology (Bernard & Bagshaw, 1969). К is curious that the ABO seems to have evolved only in muricaceans and nati- caceans, and not in other predatory molluscs. Whether all species of these two distantly related superfamilies possess an accessory boring organ has not been determined. Too few species have been examined to permit a generalization. There is, for example, an omnivorous muricid, Drupa ricina, pedal anatomy unknown, that feeds on sponges, holothurians, and carrion, and is not thought to be a typical predator of hard-shelled mol- luscs (Wu, 1965a). lts tubular salivary glands are fully developed. It will be important to determine whether this snail possesses a fully developed, or a vestigial, ABO, or none at all. Shell dissolution in muricaceans 15 not limited to shell boring. The mantle edge of spiny muricids, for example, dissolves spines at their base as the body whorl is deposited from one varix to the next, to eliminate block- age of the aperture (Carriker, 1972). The broad temporal, spatial, and systematic distri- bution of calcibiocavites, the capacity for dis- solution of shell by many invertebrates in noncalcibiocavitic activities, and the promi- nence of osteoclastic activity in the verte- brates, suggest that calcibiocavitation may be a latent and fundamental characteristic of organisms, expressing itself especially in epithelia, that has appeared from time to time without regard to systematic or morphological position (Carriker & Smith, 1969). Evolution of the proboscis and the ABO opened to boring gastropods a broad spec- trum of prey not otherwise easily available, NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 417 and undoubtedly has helped account for the historical longevity and ubiquity of the group (Carriker 8 Van Zandt, 1972a). In the event of loss of either the proboscis or the ABO, through pinching or amputation during pene- tration of prey, relatively rapid functional regeneration of both organs occurs (Carriker & Van Zandt, 19726; Carriker, Person, Libbin & Van Zandt, 1972)—a unique safe- guard insuring full replacement of the me- chanism and survival of the organism through geologic time. CONTROL During the last 50 years shellfish growers and shellfish biologists have devoted consid- erable time and effort in attempts to control muricacean predators. Examples of better known predators include Eupleura caudata, Ocenebra inornata (= japonica), Thais haemastoma, and Urosalpinx cinerea in the United States; Ocenebra erinacea and Urosalpinx cinerea in Great Britain; Ocenebra тотаа, Нарапа thomasiana, Thais bronni, and T. tumulosa in Japan; and Bedeva hanleyi and Morula marginalba in Australia. There are many other species in other regions of the world. Efforts to control muricacean borers (also called drills) by physical methods have met only with partial, and then only temporary, success. Hand picking, forks, concrete pillars, oyster dredges, deck screens, drill dredges, drill box traps, and drill trapping of Urosalpinx cinerea have all been tried more or less in- tensively. A more mechanical, less labor in- tensive method employing a hydraulic suction dredge has been used with some success in the Long Island Sound area (Carriker, 1955). Loose material on the bottom is drawn onto a screened conveyer belt that allows oysters and shell to pass back overboard into the water. Fine materials, including oyster borers, collect in bins under the screen and are later discharged in shallow water to kill the borers by suffocation. The suction dredge is limited to dredging in intermediate depths of water, and on relatively firm bottoms. Invention of a more economical method of disposing of the snails than currently used would significantly reduce the cost of this method of control (Carriker, 1955; Hancock, 1959, 1969). At- tempts to trap Thais haemastoma on oyster beds have been unsuccessful because no baits more attractive than the surrounding oysters and mussels have been found (Gunter, 1979). Efforts to eradicate muricacean predators and their young by desiccation, flaming, fresh and brine waters, magnesium sulfate, copper sulfate, mercuric chloride, formalin, rotenone and chlorinated benzene, and other chemi- cals have been ineffective on a commercial scale, or effective, but too harmful to other organisms and the environment to be em- ployed (Carriker, 1955; Castagna, Haven & Whitcomb, 1969). Copper barriers have also been suggested by Glude (1956) and Huguenin (1977), but these, like other metals, would contaminate the environment, and their application would be labor intensive and cost- ly. The use of freshwater curtains, created by release of fine streams of fresh water, to con- trol muricacean borers has not been at- tempted, but merits consideration (D. Ritt- schof, personal communication). Naticaceans (moon snails), serious preda- tors of infaunal bivalves, decimate popula- tions of such commercially important species as Mya arenaria and Mercenaria mercenaria in estuaries and embayments and Spisula solidissima on the continental shelf (Franz, 1977). Abortive attempts have been made to control them by manual collecting in the inter- tidal zone (for example, Lunatia heros, Medcof & Thurber, 1958). As with similar at- tempts at control of muricaceans, this method has serious limitations, primarily because these predators occur subtidally as well and soon replace those removed from the inter- tidal zone. The response of boring gastropods to at- tractive chemical signals from prey, or from female snails during mating, or repulsion of them by unattractive biochemical cues from other organisms, provide the basis for pos- sible ecological control. Attractive or unattrac- tive chemical signals, if they can be identified and synthesized, could possibly be used as bait in trapping, or as dispersive or repelling agents. A great advantage of such signals is that they are biodegradable, and would not contaminate the environment. Ideally they might be species specific. CONCLUSIONS Interest in organisms that penetrate hard calcareous substrata dates back many cen- turies. Aristotle, some 2,300 years ago, is credited for recognizing that predatory marine 418 CARRIKER gastropods have the capacity to bore holes through shells of prey (Jensen, 1951). Since then advances in the knowledge of shell penetration by boring gastropods has been rapid (Carriker 8 Smith, 1969; present re- view). In spite of this progress, however, sev- eral important aspects of the biology of shell penetration require further study; these are summarized in this section. Information on the zoogeographical distri- bution of boring gastropods is limited, not only in shallow coastal areas but more so in the deep-sea (Clarke, 1962), and is difficult to ob- tain. Bore holes in prey shell indicate the presence of borers in the geographic vicinity, but provide no clues on the specific identity of the borers. Identification of shell penetrants can be determined by holding snails in aquar- ia with potential prey, and observing whether hole boring takes place. This procedure gen- erally works well with snails from shallow water, but could be difficult with gastropods from the deep-sea even in pressurized aquar- ia. À more practical approach would be to ex- amine suspected shell penetrants for the presence of the ABO by anatomical and his- tological techniques. All naticacean and muricacean gastropods studied so far possess an ABO and are shell borers, Whether all species of these super- families are borers needs to be determined by examination of a wide spectrum of species of these groups, as well as non-naticacean- muricacean predators, from widely different regions of the oceans. The ABO is known to occur in three dif- ferent anatomical positions in different spe- cies of boring gastropods. However, the num- ber of species that has been examined 1$ small, and it is possible that the ABO could occur in other than the described anatomical locations. The ABO appears to be proportion- ately larger in young individuals than in adult ones (for example, Thais haemastoma; Gunter, 1968, 1979). This condition is not characteristic of most gastropod boring spe- cies, and could be interpreted as suggesting that this species has evolved toward a lesser use of the ABO in adults than in the young. On the other hand, species of borers could exist in which the ABO is an incipient organ, and the snails could be evolving either toward or away from the boring habit. A species worth exploring in this regard is Drupa ricina (Wu, 1965a). The study of transitional stages of the ABO, as well as the embryological de- velopment of the ABO in different anatomical positions, would be of considerable evolution- ary interest. From an ecological and behavioral point of view it is of interest to know whether the chemical attractant associated with each prey species is a single, or a combination of differ- ent molecules, and whether attractants are species specific. This fundamental informa- tion is prerequisite to the formulation of a bait for control of these predatory snails. Although substantial progress has been made in the study of the behavior of shell penetration by boring gastropods and of the gross and fine structure of the ABO, we know rather little about the chemical aspects of shell penetration. Study of the chemistry of the ABO secretion is difficult because the ABO is a relatively small organ, and amounts of released secretion are very small. The presence of a mild acid in the secretion has been verified with pH electrodes, but the com- position of the acid, suspected of being HCl, is uncertain. Preliminary observations suggest that an unidentified enzyme(s) and chela- tor(s) may be components of the active shell solubilizing secretion. This needs confirma- tion. The ABO is probably the principal organ involved in the chemical phase of shell pene- tration. However, close association of duct openings of the salivary glands, buccal glands and tubular salivary glands with the buccal cavity and mouth, and of the anterior pedal mucous gland with the anterior part of the foot, suggests that these glands could play at least a part in the mechanism of shell penetra- tion. Their potential role cannot be discounted until more is known about their functions. Some boring gastropods appear to be able to quiet or kill their prey by applying a paralytic substance to them through the borehole. Suspected sources of paralytic agents are the hypobranchial gland, salivary glands, and tubular salivary glands. Whether salivary glands can secrete both paralytic and shell solubilizing substances is questionable, but worth exploring. The source of these bio- toxins, the method of injection into prey, and the physiological effect on prey also need in- vestigation. Shell swallowed by boring gastropods ap- parently passes through the alimentary canal and is voided relatively unchanged in feces. There 15 the possibility, however, that some nutrients could be extracted from the organic and inorganic components of shell fragments in the stomach of the snail and absorbed. The NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 419 metabolic fate of absorbed nutrients, if any, could be tested with radioactive tracers. Costly depredations by boring gastropods of commercial bivalve populations in all parts of the world confer a high priority on these snails as subjects for the investigations pro- posed in this synthesis. Especially important would be a search for components of the shell penetrating mechanism that might be blocked in order to control the predators. The results of such a study would benefit not only the shellfish industry but would also contribute new knowledge on the biology of predation by these ubiquitous, refractory—and very inter- esting—marine snails. ACKNOWLEDGMENTS My special thanks are extended to F. Bernard, B. Brown, T. Evans, J. Gordon, A. Kohn, R. Prezant, D. Rittschof, and L. Wood who kindly reviewed the manuscript and of- fered many valuable suggestions. Preparation of the review was supported in part by the Delaware Sea Grant College Pro- gram. College of Marine Studies Contribution No. 155: LITERATURE CITED ABBOTT, R. T., 1974, American Seashells. The Marine Mollusca of the Atlantic and Pacific Coasts of North America. Ed. 2. Van Nostrand Reinhold, New York, 663 p. ARNOLD, J. М. & ARNOLD, К. O., 1969, Some aspects of hole-boring predation by Octopus vulgaris. American Zoologist, 9: 991-996. BARNETT, B. E., 1979, A laboratory study of pre- dation by the dog-whelk Nucella lapillus on the barnacle Elminius modestus and Balanus balanoides. Journal of the Marine Biological As- sociation of the United Kingdom, 59: 299-306. BAYNE, В. |. & SCULLARD, C., 1978, Rates of feeding by Thais (Nucella) lapillus (L.) Journal of Experimental Marine Biology and Ecology, 32: 113-129. BERG, С. J., уг., 1976, Ontogeny of predatory be- havior in marine snails (Prosobranchia: Nati- cidae). Nautilus, 90: 1-4. BERNARD, Е. В. & BAGSHAW, J. W., 1969, His- tology and fine structure of the accessory boring organ of Polinices lewisi (Gastropoda, Proso- branchiata). Journal of the Fisheries Research Board of Canada, 26: 1451-1457. BLACK, R., 1978, Tactics of whelks preying on limpets. Marine Biology, 46: 157-162. BOETTGER, C. R., 1930, Die Lage der Bohrstelle beim Angriff der Raubschnecken aus der Familie Naticidae. Zeitschrift für wissenschaftliche Zoologie, 126: 453-463. BURRELL, V. G., Jr., 1975, One oyster's solution to the drill problem. Veliger, 17: 334. CARRIKER, M. R., 1943, On the structure and function of the proboscis in the common oyster drill, Urosalpinx cinerea Say. Journal of Mor- phology, 73: 441-506. CARRIKER, M. R., 1954, Seasonal vertical move- ments of oyster drills (Urosalpinx cinerea). Proceedings of the National Shellfisheries As- sociation, 45: 190-198. CARRIKER, М. R., 1955, Critical review of biology and control of oyster drills, Urosalpinx and Eupleura. Fish and Wildlife Service, Special Scientific Report: Fisheries No. 148: 1-150. CARRIKER, M. R., 1957, Preliminary study of be- havior of newly hatched oyster drills, Urosalpinx cinerea (Say). Journal of the Elisha Mitchell Sci- entific Society, 73: 328-351. CARRIKER, M. R., 1961, Comparative functional morphology of boring mechanisms in gastro- pods. American Zoologist, 1: 263-266. CARRIKER, М. R., 1969, Excavation of boreholes by the gastropod, Urosalpinx: an analysis by light and scanning electron microscopy. Ameri- can Zoologist, 9: 917-933. CARRIKER, M. R., 1972, Observations on removal of spines by muricid gastropods during shell growth. Veliger, 15: 69-74. CARRIKER, M. R., 1977, Ultrastructural evidence that gastropods swallow shell rasped during hole boring. Biological Bulletin, 152: 325-336. CARRIKER, M. R., 1978, Ultrastructural analysis of dissolution of shell of the bivalve Mytilus edulis by the accessory boring organ of the gastropod Urosalpinx cinerea. Marine Biology, 48: 105— 134. CARRIKER, M. R. & CHAUNCEY, H. H., 1973, Ef- fect of carbonic anhydrase inhibition on shell penetration by the muricid gastropod Urosalpinx cinerea. Malacologia, 12: 247-263. CARRIKER, M. R., PERSON, P., LIBBIN, R. & VAN ZANDT, D., 1972, Regeneration of the proboscis of muricid gastropods after amputation, with emphasis on the radula and cartilages. Biologi- cal Bulletin, 143: 317-331. CARRIKER, M. R., PERSON, P., SMARSH, A., LIPSON, S. & CHAUNCEY, H. H., 1968, Role of carbonic anhydrase т decalcification by Urosalpinx cinerea follyensis (oyster drill). Inter- national Association for Dental Research, 46th General Meeting, and North American Division of IADR, March, 1968, Program and Abstracts of Papers, p. 188, Abstract No. 604. CARRIKER, M. R., SCHAADT, J. G. & PETERS, V., 1974, Analysis by slow-motion picture pho- topography and scanning electron microscopy of radular function in Urosalpinx cinerea follyensis (Muricidae, Gastropoda) during shell penetra- tion. Marine Biology, 25: 63-76. 420 CARRIKER CARRIKER, M. R. 8 SMITH, E. H., 1969, Com- parative calcibiocavitology: summary and con- clusions. American Zoologist, 9: 1011-1020. CARRIKER, М. R., SMITH, E. H. 8 WILCE, В. T., 1969, Penetration of calcium carbonate sub- strates by lower plants and invertebrates, an international multidisciplinary symposium. Amer- ican Zoologist, 9: 629-1020. CARRIKER, М. В. 8 VAN ZANDT, D., 1964, Use of polished mollusk shell for testing demineraliza- tion activity of accessory boring organ of muricid boring gastropods. Biological Bulletin, 127: 365. CARRIKER, M. R. 4 VAN ZANDT, D., 1972a, Regeneration of the accessory boring organ of muricid gastropods after excision. Transactions of the American Microscopical Society, 91: 455— 466. CARRIKER, M. R. & VAN ZANDT, D., 1972b, Pred- datory behavior of a shell-boring muricid gastro- pod. In: WINN, H. E. & OLLA, В. L. (eds.), Behavior of Marine Animals: Current Perspec- tives in Research, Vol. 1, Invertebrates, Plenum, New York, p. 157-244. CARRIKER, M. R., VAN ZANDT, D. 4 CHARLTON, G., 1967, Gastropod Urosalpinx: pH of acces- sory boring organ while boring. Science, 158: 920-922. CARRIKER, М. R., VAN ZANDT, D. & GRANT, T. J., 1978, Penetration of molluscan and non- molluscan minerals by the boring gastropod Urosalpinx cinerea. Biological Bulletin, 155: 511-526. CARRIKER, М. В. & WILLIAMS, L. G., 1978, The chemical mechanism of shell dissolution by predatory boring gastropods; a review and an hypothesis. Malacologia, 17: 143-156. CARRIKER, М. R., WILLIAMS, |. G. & VAN ZANDT, D., 1978, Preliminary characterization of the secretion of the accessory boring organ of the shell-penetrating muricid gastropod Urosal- pinx стегеа. Malacologia, 17: 125-142. CARRIKER, М. В. 4 YOCHELSON, E. L., 1968, Recent gastropod boreholes and Ordovician cylindrical borings. Contributions to Paleontol- ogy, [United States] Geological Survey Profes- sional Paper, 593B: 1-26. САЗТАСМА, M., HAVEN, О. $. & WHITCOMB, J. B., 1969, Treatment of shell cultch with poly- stream to increase the yield of seed oysters, Crassostrea virginica. Proceedings of the Na- tional Shellfisheries Association, 59: 84-90. CHETAIL, M., BINOT, D. 4 BENSALEM, M., 1968, Organe de perforation de Purpura lapillus (L.) (Muricidae): histochimie et histoenzymologie. Cahiers de Biologie Marine, 9: 13-22. CHEW, K. K. 8 EISLER, R., 1958, A preliminary study of the feeding habits of the Japanese oyster drill, Ocinebra japonica. Journal of the Fisheries Research Board of Canada, 15: 529- 535: CLARKE, А. H., Jr., 1962, On the composition, zoogeography, origin and age of the deep-sea mollusk fauna. Deep-Sea Research, 9: 291- 306. CRISP, M., 1973, Fine structure of some proso- branch osphradia. Marine Biology, 22: 231-240. DAY, J. A., 1969, Feeding of the cymatiid gastro- pod, Argobuccinum argus, in relation to the structure of the proboscis and secretions of the proboscis gland. American Zoologist, 9: 909— 919. DERER, M., 1975, L’organe de perforation de Thais lapillus L. (Gastéropodes, Prosobranches), etude en microscopie optique et électronique. Archives d’Anatomie Microscopique et de Morphologie Experimentale, 64: 1-26. EDWARDS. D. С. & HUEBNER, J. D., 1977, Feed- ing and growth rates of Polinices duplicatus preying on Mya arenaria at Barnstable Harbor, Massachusetts. Ecology, 58: 1218-1236. ENDEAN, R., 1972, Aspects of molluscan pharma- cology. In: FLORKIN, М. 8 SCHEER, В. T. (eds.), Chemical Zoology, Vol. 7, Mollusca. Academic Press, New York, р. 421-466. FEARE, C. J., 1971, The adaptive significance of aggregation behaviour in the dogwhelk Nucella lapillus (L.). Oecologia, 7: 117-126. FIRBY, J. В. & DURHAM, J. W., 1974, Molluscan radula from earliest Cambrian. Journal of Paleontology, 48: 1109-1119. FRANC, A., 1952, Notes écologiques et anatom- iques sur Tritonalia (Ocinebrina) aciculata (Lk.) (Mollusque Prosobranche). Bulletin du Labora- toire Maritime de Dinard, 36: 31-34. FRANZ, D. R., 1977, Size and age-specific preda- tion by Lunatia heros (Say, 1822) on the surf clam Spisula solidissima (Dillwyn, 1817) off western Long Island, New York. Veliger, 20: 144-150. FRETTER, V., 1941, The genital ducts of some British stenoglossan prosobranchs. Journal of the Marine Biological Association of the United Kingdom, 25: 173-211. FRETTER, V., 1946, The pedal sucker and anal gland of some British Stenoglossa. Proceedings of the Malacological Society of London, 27: 126-130. FRETTER, V. 8 GRAHAM, A., 1962, British Proso- branch Molluscs, their Functional Anatomy and Ecology. Ray Society, London, 755 p. GLUDE, J. B., 1956, Copper, a possible barrier to oyster drills. Proceedings of the National Shell- fisheries Association, 47: 73-82. GRAHAM, A., 1941, The oesophagus of the stenoglossan prosobranchs. Proceedings of the Royal Society of Edinburgh, section B (Biology), 61: 1-23. GRAHAM, A., 1973, The anatomical basis of func- tion in the buccal mass of prosobranch and amphineuran molluscs. Journal of Zoology, 169: 317-348. GREGOIRE, C., 1972, Structure of the molluscan shell. In: FLORKIN, M. & SCHEER, B. T. (eds.), Chemical Zoology, Vol. 7, Mollusca, Academic Press, New York, р. 45-102. GUNTER, G., 1968, Some factors concerning the drilling apparatus and the feeding and predation of prosobranchiate gastropods especially on NATICACEAN AND MURICACEAN SHELL PENETRATION AND FEEDING 421 other molluscs. Proceedings of Symposium on Mollusca. Marine Biological Association, India, Bangalore Press, p. 370-378. GUNTER, G., 1979, Studies of the southern oyster borer, Thais haemastoma. Gulf Research Re- ports, 6: 249-260. HANCOCK, D. A., 1959, The biology and control of the American Whelk tingle Urosalpinx cinerea (Say) on English oyster beds. Great Britain Min- istry of Agriculture, Fisheries and Food, Fishery Investigations, ser. 2, 22: 1-66. HANCOCK, D. A., 1969, Oyster pests and their control. Great Britain Ministry of Agriculture, Fisheries and Food, Laboratory Leaflet, new ser. no. 19, 30 p. HANKS, J. E., 1952, The effect of changes in water temperature and salinity on the feeding habits of the boring snails, Polinices heros and Polinices duplicata. Fifth Report on Investigations of the Shellfisheries of Massachusetts, Massachusetts Division of Marine Fisheries, p. 33-37. HANKS, J. E., 1953, A comparative study of the feeding habits of the boring snails, Polínices heros (Say) and Polinices duplicata (Say) as re- lated to water temperature and salinity. Master of Science Thesis, University of New Hampshire, ЭЙ р: HANKS, J. E., 1957, The rate о feeding о the com- mon oyster drill, Urosalpinx cinerea (Say) at con- trolled water temperature. Biological Bulletin, 112: 330-335. HANKS, J. E., 1960, The early life history of the New England clam drills, P. duplicatus (Say), P. heros (Say), and P. triseriata (Say) (Naticidae; Gastropoda). Doctoral Dissertation, University of New Hampshire, 122 p. HEMINGWAY, G. T., 1973, Feeding in Acanthina spirata (Prosobranchia: Neogastropoda). Master of Science Thesis, California State University, San Diego, 85 p. HEMINGWAY, G. T., 1975a, Comparacion de la morfologia funcional de alimentacion en cuatro especies de perforadores marinos (Neogastro- poda Muricacea). Ciencias Marinas, 2: 1-5. HEMINGWAY, G. T., 1975b, Functional morphol- ogy of feeding in the predatory whelk Acanthina spirata (Gastropoda Prosobranchia). Bulletin of the American Malacological Union, 41: 64-65. HEMINGWAY, G. T., 1978, Evidence for a paralytic venom in the intertidal snail Acanthina spirata (Neogastropoda: Thaisidae). Comparative Bio- chemistry and Physiology, 60C: 79-81. HUGUENIN, J. E., 1977, The reluctance of the oyster drill (Urosalpinx cinerea) to cross metallic copper. Proceedings of the National Shellfisher- ies Association, 67: 80-84. ISARANKURA, К. 8 RUNHAM, М. W., 1968, Stud- ies on the replacement of the gastropod radula. Malacologia, 7: 71-91. JENSEN, A. S., 1951, Do the Naticidae (Gastro- poda Prosobranchia) drill by chemical or me- chanical means? Videnskabelige Meddelelser fra Dansk naturhistorisk Forening, 113: 251- 261. КОНМ, A. J., 1961, Chemoreception in gastropod molluscs. American Zoologist, 1: 291-308. KRUTAK, P. R., 1977, Gastropod radulae: their potential in the fossil record. Transactions of the Gulf Coast Association of Geological Societies, 27: 314-322. LARGEN, М. J., 1967, The diet of the dog-whelk, Nucella lapillus (Gastropoda: Prosobranchia). Journal of Zoology, 151: 123-127. MANSOUR-BEK, J. J., 1934, Uber die proteoly- tischen Enzyme von Murex anguliferus Lamk. Zeitschrift für vergleichende Physiologie, 20: 343-369. MANZI, J. J., 1970, Combined effect of salinity and temperature on the feeding, reproductive, and survival rates of Eupleura caudata (Say) and Urosalpinx cinerea (Say) (Prosobranchia: Muricidae). Biological Bulletin, 138: 35-46. MARTOJA, M., 1971, Données histologiques sur les glandes salivares et oesophagiennes de Thais lapillus (L.) (= Nucella lapillus, Proso- branche Néogastropode). Archives de Zoologie Expérimentale et Générale, 112: 249-291. MATSUKUMA, A., 1977, Notes on Genkaimurex varicosa (Kuroda, 1953) (Prosobranchia: Neogastropoda). Venus, Japanese Journal of Malacology, 36: 81-88. MCGRAW, К. А. & GUNTER, G., 1972, Observa- tions on killing of the Virginia oyster by the Gulf oyster borer, Thais haemastoma, with evidence for a paralytic secretion. Proceedings of the Na- tional Shellfisheries Association, 62: 95-97. MEDCOF, J. С. & THURBER, L. W., 1958, Trial control of the greater clam drill (Lunatia heros) by manual collection. Journal of the Fisheries Re- search Board of Canada, 15: 1355-1369. MENGE, J. L., 1974, Prey selection and foraging period of the predacious rocky intertidal snail Acanthina punctulata. Oecologia, 34: 1-16. MORGAN, P. R., 1972, Nucella lapillus (L.) as a predator of edible cockles. Journal of Experi- mental Marine Biology and Ecology, 8: 45-52. NEWELL, P. F. 8 BROWN, A. C., 1977, The fine structure of the osphradium of Bullia digitalis Meuschen (Gastropoda, Prosobranchia). Mala- cologia, 16: 197-205. NIXON, M., 1979, Hole boring in shells by Octopus vulgaris in the Mediterranean. Malacologia, 18: 431-444. NIXON, M., in press, The salivary papilla of Octo- pus as an accessory radula for drilling shells. Journal of Zoology, 190: NYLEN, М. U., PROVENAA, D. V. & CARRIKER, М. В., 1968, Fine structure of the accessory bor- ing organ of the gastropod Urosalpinx. American Zoologist, 8: 801-802. NYLEN, M. U., PROVENAA, D. V. 8 CARRIKER, М. R., 1969, Fine structure of the accessory bor- ing organ of the gastropod Urosalpinx. American Zoologist, 9: 935-965. ORR, V., 1962, The drilling habit of Capulus дате! (Crosse) (Mollusca: Gastropoda). Veliger, 5: 63— 67. PERSON, P. A., SMARSH, A., LIPSON, S. J. 4 422 CARRIKER CARRIKER, M. R., 1967, Enzymes of the acces- sory boring organ of the muricid gastropod Urosalpinx cinerea follyensis, 1. Aerobic and related oxidative systems. Biological Bulletin, 133: 401410. PONDER, W. F., 1973, The origin and evolution of the Neogastropoda. Malacologia, 12: 295-338. PRATT, D. M., 1974a, Attraction to prey and stimu- lus to attack in the predatory gastropod Urosal- pinx стегеа. Marine Biology, 27: 37-45. PRATT, D. М., 19746, Behavioral defenses of Crepidula fornicata against attack by Urosalpinx cinerea. Marine Biology. 27: 47-49. PRATT, D. M., 1976, Intraspecific signalling of hunting success or failure in Urosalpinx cinerea. Journal of Experimental Marine Biology and Ecology, 21: 7-9. RADWIN, G. E. & WELLS, H. W., 1968, Сотрага- tive radular morphology and feeding habits of muricid gastropods from the Gulf of Mexico. Bulletin of Marine Science, 18: 72-85. RICHTER, G. 1962, Beobachtungen zum Beutefang der marinen Bohrschnecke Lunatia nitida. Natur und Museum, 92: 186-192. RITTSCHOF, D., 1980, Chemical attraction of hermit crabs and other attendants to simulated gastropod predation sites. Journal of Chemical Ecology, 6: 103-118. RITTSCHOF, D. in press, Enzymatic production of small molecules attracting hermit crabs to simu- lated gastropod predation sites. Journal of Chemical Ecology, 6: RUNHAM, N. W., 1969, The use of the scanning electron microscope in the study of the gastro- pod radula: the radulae of Agriolimax reticulatus and Nucella lapillus. Malacologia, 9: 179-185. RUNHAM, М. W., THORNTON, P. В., SHAW, D. A. 8 WAYTE, R. C., 1969, The mineralization and hardness of the radular teeth of the limpet Patella vulgata L. Zeitschrift fur Zellforschung, 99: 608-626. SASSAMAN, C., 1974, Resource utilization func- tion of boring gastropods. American Zoologist, 14: 1287. SOHL, N. F., 1969, The fossil record of shell boring by snails. American Zoologist, 9: 725-734. TAYLOR, J. D., MORRIS, N. J. 8 TAYLOR, C. N. 1980, Food specialization and the evolution of predatory prosobranch gastropods. Palaeontol- оду, 23: 375—409. TRAVIS, D. С. & GONSALVES, M., 1969, Com- parative ultrastructure and organization of the prismatic region of two bivalves and its possible relation to the chemical mechanism of boring. American Zoologist, 9: 635—661. WATABE, N. & WILBUR, К. M., 1976, The Mechanisms of Mineralization in the Inverte- brates and Plants. University of South Carolina Press, Columbia, South Carolina, 461 p. WEBB, В. $. & SALEUDDIN, А. S. M., 1977, Role of enzymes in the mechanism of shell penetra- tion by the muricid gastropod Thais lapillus. Canadian Journal of Zoology, 55: 1846-1857. WHITTAKER, V. P., 1960, Pharmacologically ac- tive choline esters in marine gastropods. Annals of the New York Academy of Science, 90: 695- 705. WHITTAKER, V. P. 8 MICHAELSON, I. A., 1954, Studies on urocanylcholine. Biological Bulletin, 107: 304. WILBUR, K. M., 1972, Shell formation in mollusks. in FLORKIN, М. & SCHEER, В. Т. (ed.), Chemi- cal Zoology, vol. 7, Mollusca, Academic Press, New York, p.103-145. WILTSE, W. 1., 1980, Predation by juvenile Polinices duplicatus (Say) on Gemma gemma (Totten). Journal of Experimental Marine Biology and Ecology, 42: 187-199. WOOD, L., 1968, Physiological and ecological as- pects of prey selection by the marine gastropod Urosalpinx cinerea (Prosobranchia: Muricidae). Malacologia, 6: 267-320. WU, S.-K., 1965a, Comparative functional studies of the digestive system of the muricid gastropods Drupa ricina and Morula granulata. Malaco- logia, 3: 211-233. WU, S.-K., 1965b, Studies of the radulae of Taiwan muricid gastropods. Bulletin of the Institute of Zoology, Academia Sinica, 4: 95-106. YOUNG, О. K., 1969, Okadaia elegans, a tube-bor- ing nudibranch mollusc from the central and west Pacific. American Zoologist, 9: 903-907. ZIEGELMEIER, E., 1954. Beobachtungen über den Nahrungserwerb bei der Naticidae Lunatia nitida Donovan (Gastropoda Prosobranchia). Helgo- länder wissenschaftliche Meeresunter- suchungen, 5: 1-33. MALACOLOGIA, 1981, 20(2): 423-438 FUNCTIONAL MORPHOLOGY AND EVOLUTION OF THE TOXOGLOSSAN RADULA Ronald L. Shimek and Alan J. Kohn Friday Harbor Laboratories and Department of Zoology, University of Washington, Seattle, Washington 98195, U.S.A. ABSTRACT Gastropods of the superfamily Toxoglossa have a chemically aided rapid-strike feeding method that immobilizes large prey animals; these are then swallowed whole. This report char- acterizes the major structural grades and functional types of the radula component of this mechanism and proposes a hypothesis of the evolutionary relationships of toxoglossan radulas. These probably evolved during the Cretaceous from the taenioglossan radula of a mesogastro- pod ancestor. Most of their major features are hypothesized to be simplifications from the taenioglossan condition: reduced number of rows of teeth, loss of one marginal tooth on each side, loss of the central and lateral teeth, and reduction of the radular membrane. The remaining marginal teeth have evolved in the directions of larger size and increased complexity; the most derived types function as hypodermic needles, injecting a virulent venom into the prey. Several lines of evidence support the proposed polarity of radula character states. A cladistic analysis of the subfamilies of Turridae using these data suggests that all major features of the toxoglossan radula evolved within the family Turridae; the radula has been less profoundly altered within the derivative families Conidae and Terebridae. The toxoglossan radula was an anatomical break- through that probably facilitated the rapid adaptive radiation of this superfamily in the Tertiary. INTRODUCTION The toxoglossan gastropods have a rapid- strike, chemically aided feeding method that immobilizes and holds large prey animals that are then swallowed whole. Polychaete an- nelids are the usual prey organisms, but some members of the families Terebridae and Conidae eat enteropneusts and echiurans, some Turridae and Conidae eat other gastro- pods and sipunculans, and some conids are the only gastropods known to overpower and consume vertebrates (Kohn, 1959; Pearce, 1966; Miller, 1970; Kohn 8 Nybak- ken, 1975; Shimek, 1975, 1977; Bouchet 8 Waren, 1980; Nybakken, personal communi- cation; Maes, personal communication). The toxoglossan radular apparatus is a functionally innovative anatomical break- through that probably led to the occupation of new adaptive zones by this taxon as well as its remarkable adaptive radiation. The rates of diversification of the families of Toxoglossa are among the highest in the Mollusca (Stanley & Newman, 1980), and at the pres- ent time it is probably the largest prosobranch superfamily in number of species. In this report, we first describe the function- ing of the radula in Turridae in which feeding has been directly observed, and we suggest the probable modus operandi of turrid radula types whose functioning has not yet been ob- served. We then hypothesize the evolutionary derivation of radula types within the Turridae and discuss the relationships of turrid radulas to those in other toxoglossan families. Finally we examine the degree to which a phenetic classification of turrid subfamilies based on radula characters is congruent with a classifi- cation based on shell morphometric charac- ters. We employ the turrid subfamily classifi- cation of McLean (1971; McLean in Keen, 1971). FUNCTIONAL MORPHOLOGY OF TURRID RADULA TYPES Methods: Radula Preparation Radulas selected for scanning electron mi- croscopy were obtained by removing the radu- lar sac and placing it in 5% sodium hypochlo- rite. The radula was agitated periodically to aid the dissolution of tissues. When the tissue sur- rounding the radula was almost gone, the radula was transferred to distilled water and cleaned further by agitation. The radula was (423) 424 SHIMEK AND KOHN then dehydrated in a series of alcohol rinses until it was in 70% ethanol. If a radular mem- brane was present, the preparation was com- pletely dehydrated, dried in a critical point drier, and mounted on double-stick tape in preparation for plating. If there was no radular membrane, the individual teeth were further dehydrated to absolute ethanol, transferred to a 1:1 mixture of petroleum ether and absolute ethanol, allowed to air dry, and mounted indi- vidually on double-stick tape on an observa- tion stub. The stubs were plated with gold- palladium in a vacuum evaporator and ex- amined in a JEOL JSM-35 scanning electron microscope. Results The feeding apparatus of the Toxoglossa consists of a venom apparatus, a specialized and unique proboscis (Robinson, 1960; Smith, 1967; Sheridan et al., 1973), and a highly derived radula. In all of the turrids con- sidered here, a venom apparatus opens into the buccal cavity slightly posterior to the opening of the small and relatively immobile odontophore. The lack of acrembolic and pleurembolic proboscides preclude radular use as a boring or rasping organ. Instead the specialized intraembolic and/or polyembolic proboscis (Smith, 1967) assists in prey seiz- ure. In those forms with detached hypodermic teeth, the proboscis functions to hold the tooth during envenomation. In all other forms, the proboscis probably functions to assist prey laceration by holding the prey in place. We have categorized turrid radula types in six groups, discussed below. The Slicing-Rasping Radula The slicing-rasping radula, apparently the most basic type in the Turridae, is found in the subfamilies Pseudomelatominae and Clavinae; however, the design is different in the two groups. The Clavinae contains sever- al functional radular types, probably reflecting the variety of design possible with four or five radular teeth per row. For reasons given be- low, we consider this the most primitive con- dition in the extant Turridae. The slicing-rasp- ing clavinate radula has five teeth per row (Fig. 1), a probably non-functional, reduced central tooth, a pair of curved, comb-like, in- terdigitating lateral teeth, and a pair of flat- tened, dagger-like marginal teeth. Lateral teeth do not occur in the radulas of any other turrid subfamily. FIG. 1. Calliclava albolaqueata (Carpenter, 1865) subfamily Clavinae: slicing-rasping radula. Dorsal view of radula, bending plane to the bottom. С = central tooth. L = lateral tooth. M = marginal tooth. Scale bar = 20 ит. Note interdigitation of posterior lateral teeth. The lateral teeth are erected passing over the bending plane during radular protraction and extension. They probably lacerate the prey and may tear off fragments which are conveyed to the esophagus during radular retraction, assisted by the interdigitation of the cusps of the lateral teeth. The marginal teeth, erected as they cross the bending plane, would slice the prey during the action stroke. The marginal teeth are longer than the lateral teeth and would cut deeper into the prey, per- haps allowing for deeper venom penetration. If the venom contains lytic enzymes, further prey fragmentation would be facilitated. The Pseudomelatominae also have a slic- ing-rasping radula, but of a fundamentally dif- ferent design (Fig. 2A). Lateral teeth are ab- sent. The central tooth is massive and uni- cuspid, functioning as a slicing tooth and a brace for the large scythe-like marginal teeth. The marginal teeth lacerate the prey, and they may tear off and convey fragments of the prey to the esophagus within the “basket” of teeth created as the radular ribbon passes over the bending plane (Fig. 2B). The Slicing Radula Three radular designs that are primarily slicers form the next functional grade. The TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 425 FIG. 2. Tiariturris libya (Dall, 1919) subfamily Pseudomelatominae: slicing-rasping radula. А. Dorsal view of the radula, bending plane to the top. B. Bending plane of the radula, antero-lateral view. C = central tooth. M = marginal tooth. Scale bars = 20 um. first type, found only in the Clavinae (e.g. in Agladrillia, not illustrated) is a modification of the clavinate slicing-rasping radula described above. The small central tooth has been lost. The lateral teeth are curved and comb-like; however, their cusps do not interdigitate on the return stroke. The flattened marginal teeth have a lateral strengthening rib. Except that the lateral teeth do not interdigitate, and con- sequently cannot convey materials as effici- ently into the esophagus, this radula functions much as the clavinate slicing-rasping radula. The second slicing radular design is found in a large group of species in the subfamilies Turrinae (Fig. 3), Turriculinae (Figs. 3,4), Crassispirinae (Fig. 5), and Strictispirinae, and is characterized by flattened rear-pointing marginal teeth. Originally these were probably shaped like flattened knife-like blades, but there are trends throughout these subfamilies toward the development of reinforcing ribs or struts. When the ribs are fully developed, the resulting tooth is of the so-called “duplex” type (Powell, 1942, 1966; Morrison, 1966; Maes, 1971; McLean, 1971). Examined with the light microscope, the struts often appear to be separate from the main blade of the tooth, and have been termed “accessory limbs.” The accessory limb has been hypothesized to be the evolutionary remnant of the missing taenioglossan marginal tooth (Maes, 1971). Fig. 4 shows the de novo origin of the acces- sory limb from a fold in the developing mar- ginal tooth. The marginal teeth in this type of FIG. 3. Aforia circinata (Dall, 1919) subfamily Tur- riculinae: slicing radula. A. Dorso-lateral view of the radula, bending plane to the top. Ce = central tooth. M = marginal tooth. Scale bar = 25 um. B. Bending plane of the radula in the buccal cavity, frontal view. Note orientation of teeth as they cross the bending plane. Scale bar = 100 um. C. Poly- stira picta (Reeve, 1843) subfamily Turrinae: slicing radula. One half radula. Note the accessory limbs of the teeth. Scale bar = 5 um. 426 SHIMEK AND KOHN FIG. 4. Knefastia dalli (Bartsch, 1944) subfamily Turriculinae: slicing radula. A. Frontal view of radula. M = marginal tooth. r = radular ribbon. Note strengthening strut or accessory limb. Scale bar = 50 um. В. Dorsal aspect of area shown in square in A, anterior to top. Note that there is no suture line. Scale = 5 ит. С. Dorsal view of immature radular teeth of the same radula; sequence from younger (top left) to older teeth. Arrows indicate area of tooth shown in B. Teeth at this stage are flexible. Note that the fold indicated by the arrows in С de- velops into the accessory limb of the more mature teeth in A and B. Scale bar = 25 um. FIG. 5. Crassispira rudis (Sowerby, 1834) sub- family Crassispirinae: slicing radula. A. Dorsal view, one half radula. Arrow indicates strengthening strut or accessory limb. B. Immature teeth of the same radula. Note accessory limbs developing from a fold in the tooth material, younger teeth to the top. Scale bars = 20 um. radula, erected as they cross the bending plane, function as slicing blades during the effective stroke. The substantial stress on these long, relatively thin, erect teeth could be countered in several ways. The side of the tooth opposite the cutting edge can be thick- ened or supported by a strengthening fold, or both lateral edges can be thickened with the cutting function relegated to the tooth tip. The central tooth, when present, may or may not retain a small slicing cusp, but it probably functions primarily to brace the marginal teeth during their slicing stroke. The central tooth is lost in many species within this grade, pre- TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 427 sumably in response to simplifying selective pressures acting upon the radula as the in- ternal bracing of the marginal teeth becomes more efficient. Additionally, this radular struc- tural grade is probably polyphyletic; some lineages may have developed from clavinate ancestors lacking the central tooth. The final development of the purely slicing radula, seen in some of the Crassispirinae, 15 the development of the supporting rib to give the marginal tooth a prominent free lateral brace (Fig. 6). As with the previous radular type, the erected marginal teeth function as slicing agents, but with the development of NA 4 dl À 7 eet FIG. 6. Hindsiclava militaris (Reeve, 1843) sub- family Crassispirinae: slicing radula. A. “Wishbone” or flying buttress marginal tooth, lateral view. Arrow indicates the strengthening strut. B. Dorsal aspect of radula. C. Immature teeth of the same radula as in A, arrows indicate developing strut from a fold in the tooth material. All scale bars = 10 um. the “flying buttress” the slicing portion of the tooth is thinner and presumably more effec- tive. A bracing central tooth is absent. The Slicing-Stabbing Radula A further modification of the basic slicing radula is found in some Clavinae. Here the lateral teeth are curved and comb-like with short, non-interdigitating cusps, and when the teeth are erected as they cross the bending plane they function to lacerate the prey. There is no central tooth. The marginal teeth, unlike other clavinate marginal teeth, however, are barbed and rolled to form a central channel (Fig. 7). In addition to forming an efficient lacerating organ as they cross the bending plane, the rolled tooth could facilitate en- venomation of the prey. The length and rear- ward orientation of the marginal teeth would also help force the prey farther into the esophagus with each return stroke of the radula. The Stabbing Radula The radula of the Zonulispirinae is a deriva- tion of the slicing-stabbing radula seen in the Clavinae (Fig. 8). The lateral teeth have been lost, but the rolled, hollow, often barbed, mar- ginal teeth remain attached to a strong radular membrane. These teeth probably function to pierce the prey, introducing venom on the radular action stroke. On the recovery stroke, they probably fold rearward, forcing prey into the esophagus. As the edges of the marginal overlap less than % of the total tooth length, it is probably not an efficient hypodermic in- jector for thick-skinned prey. In all of the above turrid radulas with a func- tional radular membrane, the rearward mov- ing radular teeth on the return stroke of the radula probably tend to act as a ratchet, forc- ing prey into the esophagus. Indeed, due to the small size of the radula, the laceration function of the radula might be minimal, and the major functions of radular action may be envenomation through the wounds created, and the assistance of swallowing. The Hypodermic Radula with Reduced Membrane Rolled hollow marginal teeth, overlapped for 34 of their length or more, are found in the Clathurellinae and Daphnellinae (Fig. 9). Sim- 428 SHIMEK AND KOHN FIG. 7. Imaclava unimaculata (Sowerby, 1834) subfamily Clavinae: slicing-stabbing radula. A. Anterior view of broken radular ribbon. L = lateral tooth. M = marginal tooth. Note coiled nature of the broken marginal teeth. Scale bar = 100 ит. В. Dorsal view of the bending plane of the radula. т = radular membrane. Note orientation shift of the marginal teeth as they cross the bending plane. Scale bar = 100 um. С. Isolated marginal tooth, note central channel and barb. Scale bar = 50 um. ilar to, and presumably derived from the stab- bing radular type described above, here the radular membrane is reduced and appears too weak to hold the tooth during active radu- lar movement. Instead, the teeth are probably sloughed off the end of the membrane, charged with venom, and used individually as hypodermic needles, held either in the true mouth or the tip of the extensible proboscis. The Hypodermic Radula with Vestigal Membrane The radular membrane is further reduced in the final two radular types. As in the above two types, only marginal teeth remain. In the first of these, found only in some Mangeliinae, the teeth are rolled on themselves to form an open channel (Fig. 10). This is the “hilted dagger” tooth of Powell (1942, 1966). The teeth are sloughed off the end of the radular membrane, stored in the short arm of the radular sac, and used individually as hypo- dermic venom injectors. These teeth often have a prominent spur at the base which pre- sumably aids the proboscis in gripping the tooth. Venom flows into the prey through the channel, which may be closed dorsally by the proboscis or forced shut by the pressure of the gripping proboscis. The second type of radular design with a vestigial membrane 15 also found in the Man- geliinae (Fig. 11), but additionally in the Mitrolumninae, the Borsoniinae (Fig. 12), and the Conidae. Here the marginal teeth are tubular, but the degree of overlap varies con- siderably, particularly within the Mangeliinae. In Oenopota it ranges at least from 25% to 85% of the length of the tooth (Fig. 11); it is less than 50% in most species. п the other TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 429 FIG. 8. Compsodrillia duplicata (Sowerby, 1834) subfamily Zonulispirinae: stabbing radula. Dorsal view of the radula, membrane 1$ torn. Arrow indi- cates barb on the marginal tooth. Note amount of overlap of edges of teeth. Scale bar = 25 ит. FIG. 9. Gymnobela sp. Subfamily Daphnellinae: hypodermic radula with reduced radular mem- brane. Mature radular teeth sloughed from mem- brane are on the left. Arrow indicates immature teeth still fastened to the radular membrane on the right. The teeth lack barbs, which are found on the teeth of some other species of Gymnobela. Scale bar = 10 um. FIG. 10. Kurtziella plumbea (Hinds, 1843) sub- family Mangeliinae: hypodermic radula with vestig- ial radular membrane. А, С. Immature teeth, lateral and dorso-lateral views, respectively. Note channel or groove in tooth, and the lack of a radular mem- brane. s = spur. B. Mature teeth, seen from the opposite lateral aspect from those in A and C. Note hole in the tip of the tooth and the strong elongate base of the tooth. Scale bars = 10 um. subfamilies the overlap region exceeds 75% of the tooth length. Teeth of this type are sometimes barbed, but they lack the large basal spur of the first type. The teeth are indi- vidually held in the proboscis and stabbed into the prey (Fig. 13) introducing venom. Barbs on the teeth often hold the prey during inges- tion. ORIGIN OF THE TOXOGLOSSAN FEEDING MECHANISM The prey-capturing component of the toxo- glossan feeding mechanism probably evolved during the Mesozoic from a mesogastropodan precursor in one of two ways: 430 SHIMEK AND KOHN о ше о q 5 de _ = 2 À E 2 . “+ E . ; ee FIG. 11. Oenopota spp. Subfamily Mangeliinae: Hypodermic radula with vestigial membrane. A. O. simplex (Middendorff, 1849). B. O. fidicula (Gould, 1849). C. O. turricula (Montagu, 1803). Arrow indicates spur. Note degree of overlap of edges of tooth and cantilevered strut supporting tip of tooth. Lower tooth is rotated 90° to the left of upper one. D, E. O. rugulata (Troschel, 1866). Lateral and apical views. b = apical barb. c = central channel. F. O. tabulata (Carpenter, 1864). Note apical barb, basal spur, amount of overlap, and surface texture. All scale bars = 50 um. TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 431 FIG. 12. Subfamily Borsoniinae: hypodermic radula with vestigial radular membrane. A. Ophiodermella inermis (Hinds, 1843). Compare with Fig. 11 A,B. Note difference in blade (bl) orientation, basal struc- ture, and the presence of a remnant of the connec- tion to the radular membrane. Compare also with Fig. 9. Scale bar = 50 um. B-D. Suavodrillia kennicottii (Dall, 1871). В. Lateral view, bl = blade. С. Fractured tooth. Note coiled structure. D. As in С showing flared base of tooth. Scale bars: B = 50 ит. С and D = 10 um. FIG. 13. Ophiodermella inermis (Hinds, 1843) A. Attacking prey organism, Owenia fusiformis delle Chiaje, 1844. Tooth is held in the tip of the intra- embolic proboscis (p), and forced into the prey. B. Ingestion of the prey by the snail. Scale bars = 1 cm. 432 SHIMEK AND KOHN 1, from a taenioglossan (2-1-R-1-2) ances- tor, or 2, from a “reduced rhipidoglossan” (1-1-В- 1-1) ancestor (Ponder, 1973). According to hypothesis 1, loss of one marginal tooth on each side is an initial apomorphy of the Toxo- glossa. Although Maes (1971) proposed that the single marginal tooth evolved within the Toxoglossa by fusion of the two marginals on each side, our evidence cited above (Fig. 4) makes this unlikely. According to hypothesis 2, reduction of the central tooth is the most likely initial apomorphy of the Toxoglossa. Both hypotheses consider the major evolu- tionary trends to involve 1) structural simplifi- cation by reduction in number of teeth per row, 2) enlargement and elaboration of the remaining pair of marginal teeth, and 3) re- duction of the radular membrane. The differ- ence between hypotheses 1 and 2 would dis- appear if the “reduced rhipidoglossan” an- cestor of hypothesis 2 were phylogenetically intermediate between the Taenioglossa and Toxoglossa. However, if it is considered as ancestral to both taxa, then the Taenioglossa have added a second marginal tooth on each side (Ponder, 1973). The following lines of evidence tend to support hypothesis 1 and the polarity of the three major evolutionary trends listed above: 1. Commonality of primitive state. Accord- ing to the commonality principle, “a character state widespread within a group is likely to be primitive for that group” (Eldredge, 1979). The taenioglossan radula formula occurs in all families belonging to the 12 of 13 meso- gastropodan superfamilies in which a radula is present (Thiele, 1929; Fretter 4 Graham, 1962; Morton, 1979). 2. Limited distribution of derived state. Derived states are expected to have more limited distributions, resulting in the patterns of special resemblance (Eldredge, 1979). If the taenioglossan formula is considered prim- itive, derived states are found in the three neogastropodan superfamilies (Ponder, 1973). In addition, the limited distribution of some similar radula patterns within meso- gastropodan families (e.g. Lamellariidae; Behrens, 1980) provides additional support- ing evidence. 3. Stratigraphic distribution of taxa. The argument for polarity of primitive and derived states is strengthened if stratigraphic ranges conform. While the evidence is strongest for taxa whose stratigraphic ranges do not over- lap, this strength is not available, because radula formulas can be known only from ex- tant representatives of the taxa in question. The Mesogastropoda are known from Ordo- vician time, were widespread throughout the Paleozoic, and at least three extant super- families have Paleozoic representatives (Cox, 1960). The Neogastropoda probably arose during the Mesozoic (from the superfam- Пу Subulitacea in Ponders (1973) view); the Toxoglossa as well as other neogastro- podan superfamilies are known from the early Cretaceous (Ponder, 1973; Sohl, 1977). CLADISTIC ANALYSIS OF THE TOXOGLOSSAN RADULA In view of the preceding arguments, we conclude that the taenioglossan radula con- tains the maximum number of four primitive character states (Table 1) and thus serves as the ‘ground plan' (Wagner, 1969) for the cladistic divergence analysis that follows. The cladogram (Fig. 14) was constructed by the Camin-Sokal monothetic method (Sneath 4 Sokal, 1973: 336f). The operational taxonomic units (OTUs) are the Taenioglossa and 13 taxa of Turridae at the subfamily and genus levels. For this analysis we consider the series of taxa represented by these OTUs monophyletic from the first branch point. The classification and radula data are based on McLean (1971), modified by original scanning electron microscopy studies by the first auth- or; the latter are the basis for subdividing 3 subfamilies. The cladogram shows the trans- formations relating observed traits to shared, derived character states. Postulated evolu- tionary changes are indicated by crossbars numbered to agree with the transitions indi- cated in Table 1. The alignment of taxa in the cladogram 1$ generally consistent with the functional group- ings described in the previous section. The main differences are the position of Imaclava, whose slicing-stabbing radula combines the intermediate character state of channeled marginal teeth with the primitive retention of lateral teeth, and of Kurtziella, whose hypo- dermic radula combines channeled marginal teeth with the most derived states of the other three characters. Available evidence of stratigraphic distribu- tion of turrid subfamilies is also consistent with the cladistic relationships shown in Fig. 14. All 30 species of Turridae of the well studied Lower Oligocene Keasey Formation TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 433 Functional | Slicing- Slicing Le д ù type rasping Je - Slicing 2 А Slicing {sat — Hypodermic Hi N e : Y — Ws 2 $ 9 < 2% Ss 8 $ 5 ш 9 9 Sis N Е à = < è 5° N $ S 5 2 > AS) DS Q x= € AN о = S о ET WwW 3 > Ww % u 2 N Q au < x WW a © < Ww < + os € at 2 г 5 2 .. 2 < = w Е S IZ = Ww = = = я IS} > = ax = ul 4 2 2 < = На = a x a ч = = г m) Le ++ == а EZ ^ a = u 4 = WwW wa on = a о = a — 2 3 Y < gu 2 ЕТ 7) D =] 5 u =) zZ = ры = ш 7 aw) uy 2 O о la] = = ES о 3 5 © E E D > > > EN = = D = E & F x eae age 3 == 3 & 5 2 = Я ES © o o FO n 3 o N = о = ao 3f 3f 4b 3b 4b 3c 4b Зе 1b 3b 40 3c 4a 3d 3b 1b 2 2 la За TAENIOGLOSSA FIG. 14. Cladogram of radula characteristics of the Taenioglossa, subfamilies and certain genera of Turri- dae, and Conidae, constructed by the monothetic method. The cladogram relates the evolutionary transfor- mations, designated as in Table 1, that relate observed traits to shared, derived characters. Evolutionary steps representing different states of the same character are grouped closely; otherwise the position of steps on the branches has no significance. TABLE 1. Cladistic analysis of toxoglossan radula characters: polarity of character states. Evolutionary transformations are numbered in parentheses. Character States Primitive Derived 1. CENTRAL TOOTH Large ———————> Small ———————=> Absent (1a) (1b) 2. LATERAL TOOTH Present 2 Absent (3c) 1 Wishbone 3. MARGINAL TEETH 2 Present—1 Solid—1 аа (За) (3b) (3d 1 Hollow, 1 Hollow, 1 Hollow, overlapped (So) overlapped Gi overlapped <%a length, > % length, >34 length not barbed barbed 4. RADULAR MEMBRANE Strong ——————-> Weak > Vestigial (4a) (4b) 434 SHIMEK AND KOHN in Oregon belong to the subfamilies Turrinae, Clavinae, and Turriculinae. These three sub- families comprise only 35% of the Recent tur- rid species in similar Eastern Pacific habitats, where the Daphnellinae (42%) and Bor- зопипае (17%) are also major components (Hickman, 1976). Figure 14 suggests that all major features of toxoglossan radula evolution, shown in Table 1, occurred within the family Turridae. The most derived turrid radulas are of the hypodermic type, consisting only of long, hol- low, barbed marginal teeth with a vestigial membrane. These characterize the subfamily Borsoniinae, which share all of these char- acter states with the Conidae. More profound evolutionary transformations of the radula oc- curred within the Turridae than between the Taenioglossa and Turridae or between the Turridae and Conidae. Radula morphology varies considerably in the third major family of Toxoglossa, the Terebridae, but this family has as yet been the subject of even less com- parative study than the Turridae and Conidae, and we can shed little light on phylogenetic patterns. Rudman (1969) established a new family for the genus Pervicacia, in which each radular row consists only of a pair of solid, flattened, pointed marginal teeth supported by a strong radular ribbon, thus resembling the radula of the turrid subfamily Strictispirinae. Troschel (1866) had described a similar radula but with more complex teeth in Myurella, generally considered a subgenus of Terebra (Thiele, 1929). In the terebrid genus Hastula and in some species of Terebra s.s., the radula teeth share all the derived char- acter states indicated above in the Bor- soniinae and Conidae (Troschel, 1866; Miller, 1980; Mills, 1977). Finally, some members of at least two toxoglossan families completely lack a radula. This is the case in Cenodagreutes (Turridae: Smith, 1967) and in a number of species of Terebra (Rudman, 1969; Troschel, 1866). No Conus without a radula is known, but the radular apparatus is extremely small and like- ly vestigial т С. leopardus, which engulfs prey organisms without stinging them (Kohn, 1959). RELATION OF TOXOGLOSSAN RADULA CHARACTERS AND SHELL FORM If one classifies the Toxoglossa according to cladistic relationships of radula character- istics, is this classification congruent with one based on shells? If this were so, both char- acter sets could be used to develop a hy- pothesis of phylogenetic relationships. Classi- fications of gastropods are of course based predominantly on shell form. Shell characters are more accessible and more easily studied than radulas, and shells are durable and available in the fossil record. A classification based primarily on radula characters has the major drawback of being unable to deal with the fossil record (Hickman, 1976). The hallmark of the turrid shell is the anal sinus, but at the present time we have no in- dependent assessment of polarity of trends in shell form within this family. In the very pre- liminary analysis that follows, we therefore present only a phenetic classification derived from objective expression of shell shape using 5 basic parameters of the coiled shell applicable to turrids: shape of the generating curve (S), rate of translation (T), rate of spire translation (ST), rate of whorl expansion (W), and relative whorl height (RWH) (Raup, 1966; Kohn 4 Riggs, 1975). We measured these for one species in each genus listed separately in Fig. 14 and averaged values for three species representing different genera in each sub- family. Measurements are based on illustra- tions in Keen (1971) selected for suitability for morphometric analysis. For the one genus not illustrated by Keen, we used shells of Oenopota elegans from Friday Harbor, Washington. The phenogram (Fig. 15A) was constructed by the unweighted pair-group method using arithmetic averages (UPGMA) of Euclidean distances (Sneath 8 Sokal, 1973). Two parameters, T and S, account for most of the groupings of OTUs: The Turrinae and Hindsiclava have very high values of T; the Pseudomelatominae, Turriculinae, other Cras- sispirinae, Borsoniinae and Zonulispirinae have rather high T and S values; the Mi- trolumninae has very high S and rather high T values; and the Clavinae, Strictispirinae, Daphnellinae, Clathurellinae, and Man- geliinae have lower values of T. In order to avoid introducing a variable in addition to general aperture shape, we omitted from con- sideration species in which high values of S were due to anterior elongation of the aper- ture into a shell siphon. The distinct separa- tion of the Turrinae in the phenogram (Fig. 15A) is consistent with a classification based primarily on position of the anal sinus (Hick- man, 1976). Unfortunately, reliance on pub- TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 435 TURRINAE CRASSISPIRINAE : #/ndsic/ava PSEUDOMELATOMINAE TURRICULINAE ZONULISPIRINAE BORSONIINAE CLAVINAE: Ca//iclava CLAVINAE: /mac/ava MANGELIINAE: Oenopota STRICTISPIRINAE DAPHNELLINAE CLATHURELLINAE MANGELIINAE : Aurtze//a MITROLUMNINAE A AN GG AA ee 3:0 2.5 2.0 IS 1,0 0,5 O Taxonomic distance PSEUDOMELATOMINAE CLAVINAE: Ca//iclava TURRINAE, TURRICULINAE, ETC. CRASSISPIRINAE: Hinds/c/ava CLAVINAE: /mac/ava STRICTISPIRINAE ZONULISPIRINAE MANGELIINAE: Aurtzie/la MANGELIINAE: Oenopo'a CLATHURELLINAE DAPHNELLINAE MITROLUMNINAE BORSONIINAE NA LOL EE EL зте 5 47. зе Шо Taxonomic distance FIG. 15. Phenograms of subfamilies and certain genera of Turridae based on UPGMA cluster analysis of taxonomic distance measures: A, 4 multistate radula characters; cophenetic correlation coefficient = 0.74; В, 5 quantitative shell geometry characters; cophenetic correlation coefficient = 0.88. In В, taxa have been rotated about axes to maximize conformance with A. 436 SHIMEK AND KOHN lished ventral views precluded our use of this and the other quantitative shell characters presented by Hickman (1976). In order to facilitate direct comparison with the phenogram based on shell characters (Fig. 15A), the cladistic relationship of turrid radula characters was converted into a phenogram of the same character states (Fig. 15B) by measuring taxonomic distance as the number of differences between OTUs in Fig. 14. The correlation coefficient between the taxonomic distance matrices of shell and radula characteristics is only 0.02. The pre- liminary analysis thus does not indicate con- gruence between the classification based on the phenetic similarities of the two character sets. DISCUSSION The Toxoglossa first appear in the fossil record in the Lower Cretaceous; they prob- ably arose from a mesogastropod ancestor. The toxoglossan radular apparatus probably evolved from the taenioglossan radula. In ap- plying the cladistic method to phylogeny of the Toxoglossa, we identified apomorphic char- acter states shared by the derived subfamilies of the Turridae, as well as by the Terebridae and Conidae. The maximum number of radular teeth per row in the Toxoglossa is 5: one pair of mar- ginal teeth, a pair of lateral teeth, and a cen- tral tooth. This condition is found only in some genera of the subfamily Clavinae of the family Turridae. The most derived radular character states are: 1) loss of the central tooth, 2) loss of the lateral teeth, 3) elongation and rolling of the marginal tooth into a sharp-pointed hollow tube bearing barbs, and 4) a vestigial radular membrane. Three lines of evidence, based on commonality of the primitive states, limited distribution of the derived states, and corre- spondence with the stratigraphic distribution of taxa, support this interpretation of the polar- ity of character states, from the primitive con- dition shown in Table 1. If this interpretation is correct, the main trends of toxoglossan radu- lar evolution occurred within the Turridae and involved primarily ‘streamlining’ (Regal, 1977) or simplification: reductions in the num- ber of teeth per row, number of rows of teeth, and importance of the radular membrane. Concomitantly the main evolutionary elabora- tion involved the marginal teeth and included increased size and complexity. The set of shared, derived radula character states we employ thus includes states with both rather low and high information content (Hecht & Edwards, 1976). The central and lateral teeth and the radular membrane pro- vide less information because the derived states result from the reduction and loss of more complex structures. The marginal teeth provide the most important type of informa- tion, because they are “shared and derived character states which are unique and inno- vative in structure” (Hecht 8 Edwards, 1976). In the most primitive turrid genera (sub- family Clavinae), the radula probably retains a rasping function, but the marginal teeth prob- ably slice or lacerate the prey. Loss of the lateral and central teeth led to the slicing and stabbing radulas employing only the one elaborated marginal tooth on each side of the radular row. The elongate, hypodermic mar- ginal tooth is regarded as the most derived toxoglossan radula type. The major selective pressure in the direction of such large, hollow, pointed and barbed marginal teeth appears to have been selection for increased efficiency of envenomation of prey. As the primitive tur- rid radula was probably not within an extensi- ble proboscis, selection toward specialization for rasping or boring probably would not have been effective. The shape of the rolled marginals var- ies considerably and it is likely that this structural grade has been arrived at through at least two different pathways. The first pathway from the pre-adapted rolled marginal teeth of the slicing-stabbing radula of the Clavinae to the true toxoglossan hypodermic radula of the Borsoniinae and Conidae would involve several intermediate structural grades. Three major evolutionary events characterize this pathway: first, development of the rolled marginal teeth of the slicing-stab- bing radula with a strong radular membrane; second, loss of the lateral teeth with retention of the strong radular membrane; third, gradual reduction of the radular membrane to a ves- tigial ligament attached to the base of the tooth. The presence in living animals of many, if not all, intermediate forms provides sup- portive evidence for this sequence of events. The second pathway would involve the groov- ing, rolling, and subsequent overlapping of a blade-like, cutting marginal tooth. This path- way is characterized by the development of grooved marginals probably in the slicing radula lineage of the Turrinae-Turriculinae complex. The complete loss of the radular TOXOGLOSSAN RADULA MORPHOLOGY AND EVOLUTION 437 membrane and the sequence from grooved to barely overlapped to completely rolled mar- ginal teeth is found in the Mangeliinae. Cladistic analysis of radular features is con- sistent with the argument that the true toxo- glossan condition evolved at least twice by indicating that the events leading to the toxo- glossate condition could have occurred т several different turrid subfamilies. Because there has been no prior attempt to assess the relationships of the subfamilies of Turridae, we compared phenograms based on radular and shell morphometry characters. Although there was no correlation, analyses of additional taxa and character sets might lead to a higher degree of congruence. While this study must be considered an ini- tial and preliminary approach to the evolution of the toxoglossan feeding mechanism and to the phylogeny of this superfamily, the objec- tive methods employed introduce testable procedures to the study of evolutionary rela- tionships of gastropod taxa, an area long dominated by more subjective judgments. ACKNOWLEDGMENTS We thank E. М. Kozloff, Acting Director, for providing research facilities at the Friday Harbor Laboratories, T. E. Schroeder for much assistance with scanning electron mi- croscopy, and P. A. Raymore and M. A. Rex for providing specimens used in this study. We are grateful to В. $. Houbrick and J. H. McLean for critically reviewing the manu- script. This research was partially supported by NSF Grant DEB 77-24430. LITERATURE CITED BEHRENS, D. W., 1980, The Lamellariidae of the North Eastern Pacific. Veliger, 22: 323-339. BOUCHET, P. & WAREN, A., 1980, Revision of the North-East Atlantic bathyal and abyssal Tur- ridae (Mollusca, Gastropoda). Journal of Mol- luscan Studies Supplement 8: 119 p. COX, L. R., 1960, Gastropoda: General character- istics of Gastropoda. In MOORE, R. C. (ed.), Treatise on Invertebrate Paleontology, Part 1, Mollusca 1. Geological Society of America and University of Kansas Press, Lawrence, р. 185- 1169. ELDREDGE, N., 1979, Cladism and common sense. In CRACRAFT, J. 8 ELDREDGE, N. (eds.), Phylogenetic analysis and paleontology. Columbia University Press, New York, р. 165- 198. FRETTER, V. 8 GRAHAM, A., 1962, British proso- branch molluscs. Ray Society, London, 755 p. HECHT, M. K. 8 EDWARDS, J. L., 1976, The de- termination of parallel or monophyletic relation- ships: the proteid salamanders. American Naturalist, 110: 653-677. HICKMAN, C. S., 1976, Bathyal gastropods of the family Turridae in the early Oligocene Keasey Formation in Oregon, with a review of some deep-water genera in the Paleogene of the Eastern Pacific. Bulletins of American Paleon- tology, 70: 119 p. KEEN, A. M., 1971, Sea shells of tropical West America. Stanford University Press, Stanford, California, 1064 p. KOHN, А. J., 1959, The ecology of Conus in Hawaii. Ecological Monographs, 29: 47-90. KOHN, A. J. 8 NYBAKKEN, J. W., 1975, Ecology of Conus on eastern Indian Ocean fringing reefs: Diversity of species and resource utilization. Marine Biology, 29: 211-234. KOHN, А. J. & RIGGS, А. C., 1975, Morphometry of the Conus shell. Systematic Zoology, 24: 346- 359: MAES, V., 1971, Evolution of the toxoglossate radula and methods of envenomation. American Malacological Union Annual Reports for 1970, p. 69-72. MCLEAN, J. H., 1971, A revised classification of the family Turridae, with the proposal of new sub- families, genera, and subgenera from the east- ern Pacific. Veliger, 14: 114-130. MILLER, B. A., 1970, Studies on the biology of Indo-Pacific Terebridae. Ph.D. dissertation, Uni- versity of New Hampshire, 213 p. MILLER, В. A., 1980, The biology of Hastula т- constans (Hinds, 1844) and a discussion of life history similarities among other Hastulas of similar proboscis type. Pacific Science, 33: 289- 306. MILLS, Р. M., 1977, Radular tooth structure in three species of Terebridae. Veliger, 19: 259-265. MORRISON, J. P. E., 1966, On the families of Tur- ridae. American Malacological Union Annual Reports for 1965, p. 1-2. MORTON, J. E., 1979, Molluscs. Hutchinson, London, 264 p. PEARCE, J. B., 1966, On Lora trevelliana (Turton) (Gastropoda: Turridae). Ophelia, 3: 81-91. PONDER, W. F., 1973, The origin and evolution of the Neogastropoda. Malacologia, 12: 295-338. POWELL, A. W. B., 1942, The New Zealand Re- cent and fossil Mollusca of the family Turridae, with general notes on turrid nomenclature and systematics. Bulletin of the Auckland Institute and Museum, no. 2, 188 p. POWELL, A. W. B., 1966, The molluscan families Speightiidae and Turridae. Bulletin of the Auck- land Institute and Museum, no. 5, 184 p. RAUP, D. M., 1966, Geometric analysis of shell coiling: general problems. Journal of Paleontol- ogy, 40: 1178-1190. 438 SHIMEK AND KOHN REGAL, P. J., 1977, Evolutionary loss of useless features: is it molecular noise suppression? American Naturalist, 111: 123-133. ROBINSON, E., 1960, Observation on the toxo- glossan gastropod Мапдейа brachystoma (Philippi). Proceedings of the Zoological Society of London, 135: 319-338. RUDMAN, W. B., 1969, Observation on Pervicacia tristis (Deshayes, 1859) and a comparison with other toxoglossan gastropods. Veliger, 12: 53-64. SHERIDAN, R., VAN MOL, J.-J. 4 BOUILLON, J., 1973, Etude morphologique du tube digestif de quelques Turridae (Mollusca-Gastropoda- Prosobranchia-Toxoglossa) de la région de Roscoff. Cahiers de Biologie Marine, 14: 159- 188. SHIMEK, В. L., 1975, The morphology of the buc- cal apparatus of Oenopota levidensis (Gastro- poda, Turridae). Zeitschrift fur Morphologie der Tiere, 80: 59-96. SHIMEK, R. L., 1977, Resource utilization and natural history of some northeastern Pacific Tur- ridae. Ph.D. Dissertation, University of Washing- ton, 232 p. SMITH, E. H., 1967, The proboscis and oesopha- gus of some British turrids. Transaction of the Royal Society of Edinburgh, 67: 1-22. SNEATH, Р.Н. A. & SOKAL, В. R., 1973, Numeri- cal taxonomy. Freeman, San Francisco, 573 p. SOHL, N., 1977, Utility of gastropods in biostrati- graphy. п KAUFFMAN, Е. С. 8 HAZEL, J. E. (eds.), Concepts and methods of biostrati- graphy. Dowden, Hutchinson 4 Ross, Strouds- burg, Penna., p. 519-539. STANLEY, S. М. & NEWMAN, W. A., 1980, Com- petitive exclusion in evolutionary time: the case of the acorn barnacles. Paleobiology, 6: 173- 183. THIELE, J., 1929, Handbuch der systematischen Weichtierkunde. Gustav Fischer, Jena, 1: 1-376. TROSCHEL, F. H., 1866, Das Gebiss der Schnecken zur Begrúndung einer natürlichen Classification, vol. 2. Nicolai, Berlin. WAGNER, W. H., Jr., 1969, The construction of a classification. In Systematic biology, Pub. 1692, National Academy of Sciences, Washington, D.C., p. 104-128. MALACOLOGIA, 1981, 20(2): 439-449 FEEDING MECHANISMS OF WEST AMERICAN NUDIBRANCHS FEEDING ON BRYOZOA, CNIDARIA AND ASCIDIACEA, WITH SPECIAL RESPECT TO THE RADULA James Nybakken and Gary McDonald Moss Landing Marine Laboratories, P.O. Box 223, Moss Landing, CA 95039, U.S.A. ABSTRACT Employing information available in the literature and observations made over the last several years by personnel at the Moss Landing Marine Laboratories, correlations were sought between radula morphology of West American nudibranchs and type of prey consumed. The study was restricted to nudibranchs feeding upon Bryozoa, Tunicata and Cnidaria. Initial results indicated that one type of radula morphology characterized nudibranchs feeding on tunicates. Bryozoan and hydroid predators have more varied radulas. Closer examination of the latter revealed that in some cases radula morphology is correlated with the way in which the nudibranch feeds upon the prey. Different species have different ingestive mechanisms which are reflected in their radula types. Certain specialist nudibranchs were found to have unique radula types. A compli- cating factor in the analysis was the discovery that the radula type for certain species is different between juvenile and adult and often reflects different prey preferences. Suggestions are made for future work and limited predictions advanced for correlation of certain radula morphological types and special prey categories. INTRODUCTION Since the early 1960s, several workers have dealt extensively with the food of nudi- branch mollusks (Swennen, 1961; Miller, 1961, 1962; Thompson, 1964; McBeth, 1971; McDonald & Nybakken, 1978). The resulting body of literature emphasizes the identifica- tion and enumeration of the food items. How- ever, there has been a lack of discussion of the relationships between actual feeding mechanisms and prey type. The only major correlation that has been documented 15 that of Bloom (1976) who was able to show a cor- relation between dorid radula morphology and certain broad aspects of sponge skeletal morphology. This is, however, a “loose” cor- relation involving mainly low encrusting sponges which behave as a relatively mo- notonous or homogeneous “grazing surface” in contrast to the array of higher invertebrate morphologies and textures for which nudi- branchs have evolved strategies. Those nudibranchs known to consume ascidians, bryozoans and cnidarians include species from not only the Doridacea, but also the Dendronotacea and Aeolidacea (McDonald & Nybakken, 1978). Considered together, these taxa have a much more di- verse array of radula types. In addition, the three prey taxa are themselves more highly varied in structure than the sponges. lt would therefore seem likely that study of the radula types of the non-sponge consuming nudi- branchs and the morphology of their prey would offer a chance to see if specific correla- tions can be established between radula morphology and prey type. If such broad cor- relation can be discerned, they may suggest certain mechanisms of feeding or, alterna- tively, special feeding strategies from which we might be able to make predictions con- cerning food and mechanisms of feeding for those nudibranch species for which food 15 currently unknown. The absence of correla- tions might suggest distinct methods of deal- ing with the same prey item which, in turn, would help to re-establish a correlation, or alternatively, that radular form is constrained predominantly by other factors unrelated to prey (see Hickman, 1980). Our purpose 1$ to test the hypothesis that there is a correlation between food type and radula morphology among non-sponge- eating nudibranchs. The data base for testing the hypothesis is McDonald & Nybakken (1978), plus additional observation of prey species by personnel at Moss Landing Marine Laboratories and known radula morphologies (MacFarland, 1966; McDonald, 1977). We (439) 440 МУВАККЕМ AND MCDONALD have restricted our correlation of radula morphology, food and feeding mechanisms to West American nudibranchs. METHODS AND MATERIALS The West American species of nudibranchs known to feed upon Ascidiacea, Bryozoa and Cnidaria were derived primarily from the sum- mary paper of McDonald 8 Nybakken (1978). Additional information was provided by Cooper (1979) and unpublished observations by the authors and other personnel at Moss Landing Marine Laboratories accumulated over the last several years. Radular anatomy for all species is taken from descriptions and il- lustrations in McDonald (1977) or MacFarland (1966). For each radula, the general shape of the various teeth was noted and sketched, and the number of teeth per row and the number of rows of teeth noted. General size of the animals was also noted to see if radula morphology changes with age. Morphological attributes were noted for nudibranch prey items as follows: For Bryozoa, each prey species was noted as being fleshy or calcified, with or without calci- fied front, with or without avicularia and encrusting or erect. Cnidaria consumed by Pacific coast nudibranchs fall into two classes, Hydrozoa and Anthozoa. Hydrozoan prey were ranked as to thecate or athecate and whether erect or encrusting. Anthozoans were classified as to whether they were Octocorallia or Hexacorallia and each further subdivided into the respective orders. The radulas of nudibranchs feeding on unique prey items were further analyzed to identify any specialized features that might corre- spond to the anatomy or other features of their prey. When correlations seemed unclear or ab- sent, or when several different radula types were associated with the same prey type, we undertook further analyses of the various mechanisms of feeding to discern if the differ- ent radula types could be correlated with a certain way of attacking or consuming prey. Finally, where data existed, we looked at the radula morphology of juvenile and adults of the species to see if there are changes in the radula and/or the prey consumed. RESULTS Food data and radula morphology were as- sembled for forty-four nudibranch species. The species, their average sizes, their dis- tribution among the higher taxa and the cate- gory of prey consumed are given in Table 1. Those species that are reported to feed upon Bryozoa and Ascidiacea are compared with features of the prey in Table 2. In addition, certain features of the radula are presented. Tables 3 and 4 present a similar breakdown for the Cnidarian predators. For outline draw- ings of the radulas of all species, see McDonald (1977) and MacFarland (1966). DISCUSSION Are there any obvious aspects of radula morphology that can be correlated with feed- ing On a particular type of prey item? Within the limits of the current small sample size, we believe that two correlations stand out. The first is seen in Table 2. Here, we find that only the two species of Acanthodoris and Onchidoris muricata, all members of the fam- ily Onchidorididae, feed upon fleshy Bryozoa or Tunicata. We have combined in the “fleshy” category both the Bryozoa of the order Ctenostomata, as well as the Ascidiacea, since both groups are grossly similar morpho- logically, with zooids living surrounded by a soft matrix and lacking opercula to close the opening to the outside. The radula morphol- оду of these three species is similar. They all have relatively few teeth per row, usually from eight to sixteen, of which the marginal teeth are small and the single lateral on each side is enormously enlarged. The rachidian tooth 1$ absent. Figure 1 illustrates the half row of the radula teeth of these three species, plus the radula teeth of the other species of Acantho- doris and Onchidoris found in California. The other four species of Acanthodoris are all sim- ilar in radula morphology, strongly suggesting to us that they should feed primarily on ctenostome Bryozoa or ascidians. The other two species of Onchidoris, O. bilamellata and O. hystricina, have a very dif- ferent radula morphology from O. muricata. In all three species, the lateral tooth 15 the larg- est in each row, but its shape is different (Fig. NUDIBRANCH FEEDING AND RADULAE 441 TABLE 1. Nudibranch species investigated together with features of their size and main prey category. Suborder Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Doridacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Dendronotacea Arminacea Arminacea Arminacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Aeolidacea Family Corambidae Corambidae Okeniidae Okeniidae Onchidorididae Onchidorididae Onchidorididae Onchidorididae Triophidae Triophidae Polyceridae Polyceridae Polyceridae Polyceridae Tritoniidae Tritoniidae Tritoniidae Dendronotidae Dendronotidae Dendronotidae Dendronotidae Dendronotidae Dotidae Dotidae Arminidae Dironidae Zephyrinidae Coryphellidae Coryphellidae Coryphellidae Eubranchidae Eubranchidae Eubranchidae Cuthonidae Cuthonidae Cuthonidae Fionidae Facelinidae Facelinidae Aeolidiidae Aeolidiidae Aeolidiidae Spurillidae Spurillidae Species Corambe pacifica Doridella steinbergae Ancula pacifica Hopkinsia rosacea Acanthodoris nanaimoensis Acanthodoris pilosa Onchidoris bilamellata Onchidoris muricata Triopha catalinae Triopha maculata Laila cockerelli Polycera atra Polycera hedgpethi Polycera zosterae Tritonia festiva Tritonia diomedia Tochuina tetraquetra Dendronotus albus Dendronotus diversicolor Dendronotus frondosus Dendronotus iris Dendronotus subramosus Doto amyra Doto kya Armina californica Dirona picta Antiopella barbarensis Coryphella trilineata Coryphella cooperi Coryphella ¡odinea Cumanotus beaumonti Eubranchus olivaceus Eubranchus rustyus Precuthona divae Tenellia adspersa Cuthona columbiana Fiona pinnata Phidiana crassicornis Phidiana hiltoni Aeolidia papillosa Aeolidiella takanosimensis Cerberilla mosslandica Spurilla oliviae Spurilla chromosoma Avg. size of adult nudibranch 5 mm 5 mm 10 mm 20 mm 30 mm 25 mm 15 mm 5 mm 40 mm 15 mm 15 mm 12 тт 15 тт 10 mm 20 тт 150 тт 120 тт 25 тт 40 тт 25 тт 60 тт 25 тт 8 тт 8 тт 30 mm 25 тт 20 тт 20 тт 20 тт 30 тт 8 тт 8 mm 8 тт 15 тт 5 тт 8 тт 20 mm 25 тт 40 тт 40 тт 25 тт 7 тт 20 тт 20 тт Category of prey consumed Bryozoa Bryozoa Entoprocta Bryozoa Ascidiacea Ascidiacea Cirripedia Bryozoa Bryozoa Bryozoa Bryozoa Bryozoa Bryozoa Bryozoa Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Bryozoa, Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cirripedia, Cnidaria Cnidaria, Ascidiacea Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria Cnidaria 442 NYBAKKEN AND MCDONALD TABLE 2. Radula characteristics of nudibranchs consuming Bryozoa and Ascidiacea, together with morpho- logical features of the prey. € AAA A A AAA AAA A AAA AA AA AAA AAA AAA A gg EEE a aa Radula Bryozoa or Ascidiacea Teeth Broad Fleshy Calci- Uncal- With Without Nudibranch per or or cified Encrust- avicu- амси- species row narrow soft front front ing Erect laria laria Corambe pacifica 10-14 = narrow X X X Doridella steinbergae 10-12 narrow X X X Ancula pacifica! 4 narrow Hopkinsia rosacea 4 narrow x x Acanthodoris pilosa 8-10 narrow x Acanthodoris nanaimoensis 10-16 narrow x Onchidoris muricata 12 narrow x x x x x Triopha catalinae 40-82 broad X X X Triopha maculata 20-39 broad x x x x Laila cockerelli 25-33 broad X X X Polycera atra 8-12 narrow X X X X Polycera hedgpethi 10-12 narrow X X X X Polycera zosterae 14-16 narrow X X X X Antiopella barbarensis 29—45 broad X X X TFeeding on phylum Entoprocta. 1). Furthermore, the number of teeth per row is reduced in both species to five, and a rachidian tooth is present. This suggests a dif- ferent prey type, which indeed is the case. Onchidoris bilamellata is a specialist on barnacles (Miller, 1961; Swennen, 1961), while O. hystricina preys upon the bryozoan Tubulipora sp. (McDonald, 1977). The uni- formity of radula anatomy among the fleshy bryozoan and tunicate feeders suggests that where this radula type is found, the species should feed in similar fashion upon similar prey. At this time, the question of how this particular morphology is related to the actual consumption of the prey is unknown. A second strong correlation can be ob- served between radula morphology and anemones as prey items. Table 4 lists five species of aeolid nudibranch feeding on actiniarian anemones, and one species of dendronotacean nudibranch, Dendronotus iris, feeding on ceriantharian anemones. The basic radula morphology of the five species feeding on actiniarian anemones, Aeolidia papillosa, Aeolidiella takanosimenis, Cerber- illa mosslandica, Spurilla oliviae and 5. chromosoma, is again very similar, as shown in Fig. 2. The radula in each case is uniseriate (one tooth per row). Each tooth is very broad and has the entire anterior border covered with small denticles. The actual form of the denticled border varies among the species from broadly concave to slightly convex, but the entire outline makes these radulas quite distinct. We predict that other eolids with this same radula morphology will also be found to feed upon actiniarians. А somewhat special case concerns Dendronotus iris, the only other Pacific coast nudibranch known to consume anemones. NUDIBRANCH FEEDING AND RADULAE 443 TABLE 3. Radula characteristics of nudibranchs consuming hydrozoan prey, together with morphological features of the prey. Radula Hydrozoa Nudibranch Teeth Broad or species Uniseriate Triseriate per row narrow Thecate Athecate Chondrophore Dendronotus albus 13-19 narrow X Dendronotus diversicolor 13-19 narrow X Dendronotus frondosus 13-23 narrow X X Dendronotus subramosus 5-15 narrow X Doto amyra X 1 narrow X Doto kya X 1 narrow X Dirona picta 5 narrow X Coryphella trilineata X 3 narrow X Coryphella iodinea X 3 narrow X Coryphella cooperi X 3 narrow X Cumanotus beaumonti X 8 narrow x Eubranchus olivaceus x 3 narrow x Eubranchus rustyus x 3 narrow X X Precuthona divae X 1 narrow X Tenellia adspersa X 1 narrow X X Cuthona columbiana X 1 narrow 2 X Fiona pinnata X 1 narrow X Phidiana crassicornis X 1 narrow X Phidiana hiltoni X 1 narrow X Aeolidia papillosa X 1 narrow Aeolidiella takanosimensis X 1 narrow Cerberilla mosslandica X 1 narrow Spurilla oliviae X 1 narrow Spurilla chromosoma X 1 narrow Antiopella barbarensis 2745 narrow X 44 NYBAKKEN AND MCDONALD = "WD 1 > А 90995 | > DA? LE PR FIG. 1. Radulas of the Onchidorididae. A, Acanthodoris brunnea; В, Acanthodoris hudsoni; С, Acanthodoris lutea; D, Acanthodoris nanaimoensis; E, Acanthodoris pilosa; Е, Acanthodoris rhodoceras; G, Onchidoris bilamellata; H, Onchidoris hystricina; |, Onchidoris muricata. The bars represent 50 um. Each radula is а single half row with marginals to the left. со FN EN Pt C F FIG. 2. Radulas of anemone predators. A, Aeolidia papillosa; В, Aeolidiella takanosimensis; С, Spuril- la chromosoma; D, Dendronotus iris; E, Cerberilla mosslandica; F, Spurilla oliviae. The bars repre- sent 50 um. The radula of D. iris is a half row, rachidian to the left, laterals and marginals to the right. Dendronotus iris, as an adult, is a specialist on the ceriantharian anemone Pachycerianthus fimbriatus. In contrast to its co-occurring con- geners, it has a radula with 23-24 teeth per row; none of its Pacific coast relatives have more than 23 teeth per row. The shape of the teeth of D. iris is similar to those of its hydroid consuming congeners, except for the ab- sence of serrations on the marginal and lateral teeth. lt is not likely that one could predict the food source here from radula anatomy alone. We believe this is partially due to the different mechanisms of dealing with the prey. Dendronotus iris is reported by Wobber (1970) to climb up on the tubes of P. fimbri- atus and selectively attack the tentacles of the anemone. The aeolid anemone feeders gen- erally feed on whole anemones (Waters, 1973). Here is a case where further study of the actual feeding process between these two groups might elucidate the significance of the different radula morphology. Three dendronotacean nudibranchs, Tochuina tetraquetra, Tritonia diomedea and T. festiva, plus the arminacean Armina cali- fornica, are known to consume octocorallians (Thompson, 1971; Wicksten & DeMartini, NUDIBRANCH FEEDING AND RADULAE 445 TABLE 4. Radula characteristics of nudibranchs consuming anthozoan prey and morphological features ofthe prey. 1 Anthozoa Broad —— Nudibranch Teeth or Actini- Cerian- Stolo- Pennatu- Gorgo- Alcyo- species per row патом ara tharia nifera lacea nacea nacea Other food Tritonia diomedia 117-193 broad x Tritonia festiva 36-72 narrow x x X Tochuina tetraquetra 328-625 broad X X Dendronotus frondosus 13-23 narrow Ascidiacea Dendronotus iris 23—43 narrow Armina californica 81-163 broad X Dirona picta 5 narrow Bryozoa Coryphella iodinea 3 narrow Ascidiacea Phidiana crassicornis 1 narrow X Ascidiacea Aeolidia papillosa 1 narrow X Aeolidiella takanosimensis 1 narrow X Cerberilla mosslandica 1 narrow X Spurilla oliviae 1 narrow X Spurilla chromosoma 1 narrow X Antiopella barbarensis 29—45 narrow Bryozoa AAA RE EP a RER Ета 1973). In the case of the three dendrono- taceans, the radulas are very similar in the form of the individual teeth (Fig. 3). The lateral teeth are strong and the marginals are curved. Armina californica has much different teeth, in that they are bifid at the tips and less massive. Both А. californica and the dendro- notaceans, however, share the characteristic of having many teeth per row (Table 4) and therefore, of having a broad radula. The major exception is 7. festiva, the smallest of the group (Table 1). Tritonia festiva is the only species to feed on the octocoral order Stoloni- fera, specifically a species of the genus Clavularia. Whereas other octocorals (Penna- tulacea, Gorgonacea, Alcyonacea) are rather large, massive colonies, usually much larger than the nudibranch, Clavularia sp. is a very tiny form, often smaller than T. festiva. The individual polyps of Clavularia sp. arise from a single, narrow stolon attached to the rock, and the unusually narrow radula of 7. festiva seems to correlate with the morphology of the prey. A larger, more massive radula would not be as effective. Of those nudibranchs for which we have data, two specialists occur, Ancula pacifica and Hopkinsia rosacea. The former feeds on members of the small phylum Entoprocta and the latter on the encrusting bryozoan Eury- stomella bilabiata (McBeth, 1971). Both spe- cies have only four teeth per row, but the morphology of the teeth of the two species 15 very different (Fig. 4). We know little about the feeding behavior of A. pacifica, but for H. roseaca we can note that it is the only one of the bryozoan feeders which feeds exclusively on a bryozoan that has a calcified front. Pre- sumably, this 15 facilitated by the strong lateral teeth with the hooks on the inner edge. We 446 NYBAKKEN AND MCDONALD MG N: BA (are: | [AGS ı D FIG. 3. Radulas of octocoral predators. A, Tochuina tetraquetra; В, Tritonia diomedia; С, Ттота festiva; D, Armina californica. The bars represent 50 um. Each radula represents a half row with rachidian to the right and laterals and marginals to the left. think that the combination of massive lateral teeth, no rachidian and small marginals, is an adaptation to attack Bryozoa with calcified fronts, more specifically, to break the indi- vidual zooecia, rather than to graze up large numbers of polypides at one time, such as seen in the other bryozoan feeders (Table 2). This idea is reinforced by noting the narrow- ness of this radula as compared to that of the other bryozoan feeders (Table 2). Among the remaining nudibranchs consid- ered here, it is difficult to find any obvious radula morphology associated with a particu- lar type of prey. We therefore asked ourselves whether there might be correlations if we looked more closely at just how a series of nudibranch predators handle their food. For this, it was necessary to find studies in which the actual mechanism of feeding was de- scribed for nudibranchs feeding on a single class of prey items. We were fortunate in find- = Q 005 ©0657, MS (SO, FIG. 4. Specialists and bryozoan predators. A, Hopkinsia rosacea; B, Ancula pacifica; C, Triopha catalinae; D, Triopha maculata. The bars represent 50 um. Each radula is a half row with rachidian teeth to the right in Triopha and laterals and margin- als to the left. In Hopkinsia and Ancula, the larger tooth 15 a lateral, the smaller a marginal. ing two such studies (Waters, 1966; Cooper, 1979) that included data on specific feeding strategies of aeolid nudibranchs feeding on hydroids. As may be seen in Table 3, many aeolid nudibranchs feed upon hydroids. These aeolids fall into two classes with respect to the radulas, either uniseriate (one tooth per row) or triseriate (three teeth per row). The hydroids upon which they feed may be either thecate (polyps covered by a hard perisarc) or athecate (polyps naked). Our initial grouping in Table 3 revealed no strong correlation be- tween radula type and hydroid type. Aeolids with both uniseriate and triseriate radulas were reported to feed on both thecate and athecate hydroids. Cooper (1979) studied the nudibranchs as- sociated with the athecate hydroid Tubularia crocea and carefully noted those which fed on it and where. Waters (1966) made similar careful observations on the aeolids feeding on the thecate hydroid Obelia commissuralis. Tables 5 and 6 list the species of nudi- branchs found with each of these prey spe- cies and а summary of where the nudibranchs were found feeding on the hydroid, or if they were found not to be feeding or feeding only on other hydroids. NUDIBRANCH FEEDING AND RADULAE 447 TABLE 5. Tubularia-associated nudibranchs (from Cooper, 1979). Numbers in parentheses indicate whether radula is uniseriate (1) or triseriate (3). Species Hydranth Coryphella cooperi X (3) Coryphella trilineata X (3) Coryphella sp. X (3) Cuthona albocrusta Cuthona columbiana Tenellia adspersa Phidiana crassicornis X (1) Feeding on Not feeding or feeds on Gonophores Stolon other hydroids X X X X (1) X (1) X (1) Cuthona fulgens X Dendronotus iris Obelia Dendronotus frondosus Obelia Doto amyra Bougainvillia Eubranchus rustyus Obelia TABLE 6. Nudibranchs associated with Obelia commissuralis (from Waters, 1966). Feeding on Radula Species Hydranth Stolon Triseriate Uniseriate Eubranchus olivaceus X X Cuthona concinnia X X Coryphella fusca X X Phidiana crassicornis X X Once this breakdown is done, some cor- relations are suggested when one considers an additional fact concerning thecate and athecate hydroids. That fact is that in both hydroid divisions, the main stalks and stems are covered with a tough chitinous perisarc and the only difference between the two is that, in thecate hydroids, the perisarc extends up around the base of the hydranth, but does not cover it. Examination of Tables 5 and 6 reveals that, with one exception, those eolids with triseriate radulas fed upon the naked hydranths or gonophores, whereas those nudibranchs with uniseriate radulas fed upon the stolons that are covered with a perisarc. lt seems likely that the uniseriate radula is better adapted for drilling holes than is a radula with three teeth per row. This observation is supported by the fact that sacoglossan opisthobranchs all have uniseriate radulas and all are known to use these to pierce the tough cell walls of certain algal cells (Thompson, 1976). The single ex- ception in our study is Phidiana crassicornis, which is known to be a generalist feeding on many different groups. Our analysis also points out the danger of listing as a food source the hydroid upon which a nudibranch is found. As Table 5 shows, Cooper (1979) found five species of nudibranch associated with Т. crocea that feed only upon other hydroids. Unfortunately, we do not have more de- tailed information for the exact mechanism of feeding on hydroids by other eolids. We there- fore cautiously predict from this small sample size that future studies should indicate that, in general, hydroid-feeding nudibranchs with uniseriate radulas should feed by piercing the stolons, stems or hydrothecae of hydroids and triseriate species should feed on the polyps without piercing the perisarc. The only group of bryozoan feeders for which detailed data exist with respect to the mechanism of feeding are the two species of Triopha, T. catalinae and T. maculata. These two species have somewhat different radulas, but they are not markedly divergent (Fig. 4). One significant difference is that there are more teeth per row in 7. catalinae and hence the radula is more broad (Table 2). As Nybakken 8 Eastman (1977) have shown, 448 NYBAKKEN AND MCDONALD these two species differ in their prey prefer- ence, T. catalinae feeding on erect Bryozoa and Т. maculata mainly on encrusting forms. The teeth of 7. catalinae are more “hooked,” which corresponds to its mechanism of feed- ing, which is to rip off whole pieces of the erect bryozoans. In contrast 7. maculata has a narrower, less robust radula and less “hook” to the teeth and feeds by scooping out the polypides from the bryozoans. If our observations hold true in future stud- ies of bryozoan feeders, we would expect a broad radula and strongly hooked teeth to be associated with species feeding on erect Bryozoa, whereby the colony pieces are actually ingested. Conversely, we might ex- pect that narrower radulas (fewer teeth per row) and those with less “hook” to them, would be features of species that feed on en- crusting (perhaps also erect) bryozoans in which polypides are grazed out and the re- mainder of the colony left intact. Although the results discussed above sug- gest that, at least in some cases, it is possible to relate radula morphology to food and/or feeding mechanism, this is not always the case. One reason for the difficulty may lie in the fact that both radula morphology and prey species consumed may change during ontogeny. Two examples are documented from work at Moss Landing Marine Labora- tories. Nybakken 8 Eastman (1977) found that the radula anatomy of Triopha maculata was different between juveniles and adults. The juveniles had but one marginal tooth, whereas adults had from four to eight margin- al teeth. Correlated with that difference was a change in diet, the juveniles feeding primarily on encrusting bryozoans, while the adults fed on both encrusting and arborescent bryo- zoans. Similarly, Cooper (1979) reported Dendronotus iris juveniles to be found associ- ated with, and consuming, Obelia, whereas the adults are seemingly specialist predators on the anemone Pachycerianthus fimbriatus. Unfortunately, in the latter case, no radula preparations were made of the juveniles. Similar changes in diet with age have been reported for D. frondosus by Thompson (1964). These latter two observations suggest two things; first, that more specific correlations between certain radula morphological types and prey reported in the literature may be ob- scured by not knowing if the animal was a juvenile or an adult and/or by not checking to see if the radula was the same in both juvenile and adult; secondly, the fact that, in those cases investigated where the radula morphol- оду does undergo change, there 1$ also a cor- responding change in the major food. Both ob- servations support the contention that radula morphology is indeed correlated with type of prey and/or mechanism of feeding. Finally, it is not clear to what extent convergence in feeding behavior has directed evolution of radula morphology and what specific radular characters are most directly affected. It may thus be that future, more careful, studies of nudibranchs, particularly the changes in both food and radulas between juveniles and adults, may well elucidate some of these cor- relations. CONCLUSIONS 1. Aeolid nudibranchs that feed on anemones tend to have uniseriate radulas with broad teeth that are heavily serrated on the anterior border. 2. Nudibranchs that feed either on fleshy ctenostome bryozoans or ascidians have similar radula morphologies with each half row dominated by a very large, massive —1-shaped lateral tooth. 3. Aeolid nudibranchs that feed on hy- droids have uniseriate or triseriate radulas. Those with unseriate radulas prey upon hy- droids by puncturing the perisarc, usually somewhere along the stolon or stem, and eat- ing out the coenosarc. Those with triseriate radulas tend to feed directly upon the polyps. 4. Most bryozoan feeders prey upon bryozoans which lack calcified fronts (Anasca). 5. Nudibranchs that feed on members of the orders Pennatulacea, Gorgonacea and Alcyonacea have very broad radulas. Those that feed on the order Stolonifera have nar- row radulas. 6. Some specialists, such as Hopkinsia rosacea and Ancula pacifica, have unique radulas that may be related to the specific prey item. 6. Evidence exists for a few nudibranchs that radula morphology and food preference change with ontogeny. ACKNOWLEDGEMENTS The authors thank Dr. Carole Hickman and Mr. David Lindberg for reviewing the manu- script, John Cooper for providing some of the much needed data, and Ms. Lynn McMasters for drawing the figures. NUDIBRANCH FEEDING AND RADULAE 449 LITERATURE CITED BLOOM, S. A., 1976, Morphological correlations between dorid nudibranch predators and sponge prey. Veliger, 18: 289-301. COOPER, J., 1979, Ecological aspects of Tubu- laria crocea (Agassiz, 1862) and its nudibanch predators in Elkhorn Slough, California. MS thesis, California State University, Hayward, 107 p. HICKMAN, C., 1970, Gastropod radulae and the assessment of form in evolutionary paleontol- ogy. Paleobiology, 6: 276-294. MACFARLAND, F. M., 1966, Studies of opistho- branchiate mollusks of the Pacific coast of North America. Memoirs of the California Academy of Sciences, 6: xvi + 546 p., 72 pl. MCBETH, J. W., 1971, Studies on the food of nudi- branchs. Veliger, 14: 158-161. MCDONALD, G. R., 1977, A review of the nudi- branchs of the California coast. MS thesis, Cali- fornia State University, Hayward, 373 p. MCDONALD, G. R. & NYBAKKEN, J. W., 1978, Additional notes on the food of some California nudibranchs with a summary of known food habits of California species. Veliger, 21: 110- 119. MILLER, M. C., 1961, Distribution and food of the nudibranchiate Mollusca of the south of the Isle of Man. Journal of Animal Ecology, 30: 95-116. MILLER, M. C., 1962, Annual cycles of some Manx nudibranchs with a discussion of the problem of migration. Journal of Animal Ecology, 31: 545- 569. NYBAKKEN, J. W. & EASTMAN, J., 1977, Food preference, food availability and resource parti- tioning in Triopha maculata and Triopha car- penteri (Opisthobranchia: Nudibranchia). Veliger, 19: 279-289. SWENNEN, C., 1961, Data on distribution, repro- duction and ecology of the nudibranchiate Mol- lusca occurring in the Netherlands. Netherlands Journal of Sea Research, 1: 191-240. THOMPSON, T. E., 1964, Grazing and the life cycles of British nudibranchs, р. 275-287. In: CRISP, D. J. (ed.), Grazing in Terrestrial and Marine Environments. Blackwell, Oxford, England. THOMPSON, T. E., 1971, Tritoniidae from the North American Pacific coast (Mollusca: Opisthobranchia). Veliger, 13: 333-338. THOMPSON, T. E., 1976, Biology of opistho- branch molluscs. Vol. 1. Ray Society, vol. 151, 207 р. WATERS, V., 1966, Feeding ecology and other aspects of the natural history of the nudibranch, Eubranchus olivaceus. MS thesis, University of Washington, 88 p. WATERS, V., 1973, Food preferences of the nudi- branch Aeolidia papillosa and the effect of the defenses of the prey on predation. Veliger, 15: 174-192. WICKSTEN, М. К. 8 DEMARTINI, J., 1973, Obser- vations of the feeding habits of Tochuina tetra- quetra (Pallas) (Gastropoda, Tritoniidae). Veliger, 15: 195. М/ОВВЕН, D. R., 1970, A report on the feeding of Dendronotus iris on the anthozoan Cerianthus sp. from Monterey Bay, California. Veliger, 12: 383-387. : р rm. а en TN re ra ое oe Е | | e y к ne Ae, | net hi ane - J A, у Ki TR м e ATAR a = DV PA Y PRAT fi "Vay arr. * > Que В Paru: ee DE A ГЛ ! > F a ns a as, ‚ae yd мега ANT Ay Es 21 e “$. REN | de ea u Lv) hu i я Mr A : + | A } = 7 г an "i= o яв | } ’ meer ue aio ad A ‘ и T0 : tes lade aati Г à Le e te ^^ 7 ja mdr. kA mo Al ote Fa La нм! eN Rt. a Ale m Wi un by $5284 rd г AZ NGS. He а tune" dun dd) & end IN $ ‚а АИ и d'un mio e Res ANR RATER eb el | ve ul AS у a hae sae nee KAA AT фтп мех MÈRE MEE Are e ie = HE IEA AAA pre р Ace 1 ay? Suede #4 a IN 4 ¿AN N и 102 у Le y lee ait io We E у Pe | FEA In > Sr в i À, © К le oe ii a. Se ‘ A В ES (UT, LA СИ . pa г QU 2 i= ven М , utero = es i A : 5 4 #] ore: u ¡A P hem a sera a а E me MALACOLOGIA, 1981, 20(2): 451-469 THE ROLE OF CARNIVOROUS GASTROPODS IN THE TROPHIC ANALYSIS OF A FOSSIL COMMUNITY Robert J. Stanton, Jr.,? Eric N. Powell? and Penelope С. Nelson? ABSTRACT Trophic structure is an essential attribute in the description and analysis of a community. It is particularly useful in the study of fossil communities because it is relatively independent of geologic time. That is, pathways of energy flow in a community and efficiencies from level to level in the ecological pyramid have probably changed relatively little through time compared to the changes in the species composition of the community. Reconstruction of the trophic structure of a fossil community is severely hampered by the bias introduced by nonpreservation of the soft-bodied organisms. Among the preserved component, gastropods are the dominant carnivores, particularly in Mesozoic and Cenozoic communities. Consequently, they are the key element in any attempt to reconstruct the trophic structure of the original community. Numerical abundance of individuals is the simplest means of estimating the relative im- portance of the different gastropod taxa within the community. Differences in size frequency distributions and in survivorship curves for the taxa present in the community indicate, however, that age and size attributes of the population as well as number of individuals must be con- sidered in a detailed analysis of trophic structure. A trophic analysis should be based on mea- sures of energy flow through the community rather than on numerical abundances. Calculations of produced biomass (growth) and cumulative biomass (maintenance) incorporate these attri- butes and provide estimates of energy flow in the community. They also incorporate a time dimension, which is necessary in paleoecologic interpretation of a time-averaged fossil as- semblage but which may not be a consideration in an ecological study. These biomass mea- sures for the carnivorous gastropods can be used to estimate prey production. In the Eocene Stone City Formation, the fossil assemblage consists of more than 100 species, but biomass calculations indicate that far less than half of the original community is preserved. The non-preserved рай consists primarily of crustaceans at the carnivore level and of deposit feeding soft-bodied organisms at the primary consumer level. INTRODUCTION A major trend in paleoecology in recent years has been to analyze the fossil record from an increasingly biological point of view. One aspect of this line of investigation has been to consider fossil assemblages as relict representations of original communities of organisms. The comparison has been, and is largely still in taxonomic terms. The fossil communities are compared with one another and with modern communities by similarities in taxa present. If the age differences be- tween the fossil assemblages 15 not great, the taxonomic level of comparison may be the genus or even the species; with greater age differences the comparison may be at the order or family level. In the development of community paleo- ecology, however, structural attributes have become increasingly important. Measures of diversity have been particularly useful, and the paleoecologist has avidly attempted to employ the concepts, indices and interpreta- tions of diversity that have been so plentiful in the ecological literature during the past fifteen years. In more recent years, trophic structural characteristics have been increasingly ex- amined as a means of comparing fossil com- munities (Walker, 1972; Walker 8 Bambach, 1974; Stanton & Dodd, 1976; Hoffman et al., 1978; Scott, 1978; Bosence, 1979). The ultimate objective of community paleo- ecology would be to reconstruct from the fos- sil assemblage the original community from which it had been derived. This is clearly a difficult if not impossible task because the original community is generally very poorly 1Department of Geology, Texas A&M University, College Station, TX 77843, U.S.A. Department of Oceanography, Texas A&M University, College Station, TX 77843, U.S.A. 1504 E 37th St., Tulsa, OK 74105, U.S.A. (451) 452 STANTON, POWELL AND NELSON preserved. Reconstruction of the original community is very important, however, for several reasons. First, it is the only way to recognize and describe the non-preserved component of the fossil record. Thus it leads to improved knowledge of the fossil record. Nonpreservable organisms in Recent com- munities comprise 50-75% of the total num- ber of organisms present (Lawrence, 1968; Stanton, 1976; Schopf, 1978). As in modern communities, the non-preservable com- ponent of ancient communities probably in- cluded a number of numerical dominants (Sanders, 1958; Holland et al., 1973; Maurer, 1977) as well as species that may have strongly influenced community composition through processes such as sediment rework- ing (Myers, 1977 a, b), sediment stabilization (Young & Rhoads, 1971; Wilson, 1979), and predation (Virnstein, 1977, 1979). Second, the interpretation of the fossil community leads to improved paleoenvironmental recon- struction because the paleoecologist is able to deal with a greater proportion of the original biota, because trophic and other structural characteristics can be directly interpreted en- vironmentally, and because the processes that have led to some parts of the original community being preserved and others des- troyed can be recognized and are sympto- matic, themselves, of environmental con- ditions. Third, by attempting to reconstruct the original community, the paleoecologist is in a position to ask questions about the biological aspects of the ancient ecosystem such as the evolution of communities, niches, and organ- ism interactions. The role of gastropods in ancient communities is of major importance because they are the dominant carnivore pre- served in many fossil assemblages. Community paleoecology requires three types of data. The first is a comprehensive knowledge of the fossil record. Although bio- stratigraphy and many other paleontologic disciplines can function with only the domi- nants in the fauna, paleocommunity analysis depends upon as full a representation of the taxa in the assemblage as possible, as well as information about the relative abundances and population dynamics of the taxa. The second type of data is that derived from a thorough taphonomic analysis of the fossil assemblage. That is, an analysis of the fossils and the sediments in which they occur in order to identify and evaluate the processes that resulted in the fossil assemblage as a more or less biased representation of the original community. The third type of data consists of well-documented modern com- munities and detailed life histories of compar- able modern species that can serve as analo- gues for the fossil communities and taxa. Trophic analysis is a major component of paleocommunity reconstruction. It provides insight into the functional relationships within the community by examining one of the т- portant parameters binding the organisms to- gether into the community—their trophic in- teractions. Because the preservable and non- preservable components in a fossil com- munity must have been tied together in the trophic web, the nature and importance of the soft-bodied fauna in the original community can be inferred through trophic analysis of the fossil assemblage. These inferences can be drawn from either preserved prey which show evidence of predation, such as mollusc shells with crab nips on them, or preserved preda- tors that fed on non-preserved prey. In most fossil assemblages, the predators that are preserved are primarily gastropods. In this study, for example, almost all the gastropods, which amount to about half of the assem- blage, are carnivores. The objectives of this paper are to describe an Eocene assemblage and the original com- munity as far as it can be reconstructed from the fossil assemblage and to analyze the role played by the predatory gastropods in the community. GEOLOGIC SETTING The Stone City Formation (Claiborne Group, Middle Eocene) consists of interbedded fos- siliferous bioturbated glauconite pellet sand- stone and laminated lignitic mudstone at its type locality on the Brazos River in south- eastern Texas (Fig. 1). The Formation was deposited during the initial transgressive phase of a major Eocene depositional cycle in the northwest Gulf of Mexico. The upward in- crease in the proportion of glauconite sand- stone to lignitic mudstone reflects progres- sively more marine conditions during depo- sition. The depositional setting was probably similar to that of the lakes and sounds north- east of the present Mississippi Delta. The fauna of the Main Glauconite bed, near the top of the Stone City Formation, is the basis of this study. The bed consists predominantly of ovoidal glauconitized fecal pellets 0.5 mm long and 0.2 mm across in a matrix consisting CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 453 (SOUTH) ЕКА CADDEL CROCKETT CLAIBORNE WECHES RECKLAW WILCOX TRANSGRESSIVE MARINE CLAIBORNE OUTCROP Tas = STONE CITY, : À. j я 130 km , 7 = e SAN ANTONIO. ^^) (NORTH) LZ QUEEN REGRESSIVE DELTAIC GROGKED MES MAIN GLAUCONITE BED STONE CITY CITY CARRIZO om || РЕ! SPARTA GLAUCONITE ARENITE, ABUNDANT INVERTEBRATE FOSSILS, BIOTURBATION te CLAYSTONE, LAMINATED, ABUNDANT PLANT F== FOSSILS, WITH INTERBEDDED SS LENSES, [== SOME X-BEDDED, WITH INVERTEBRATE FRAGMENTS FIG. 1. Location of the study site and stratigraphic setting of the Main Glauconite bed, Stone City Formation. A, Outline map of Texas; location of the Stone City study site within the Claiborne outcrop belt. B, Location of Stone City Formation within the Middle Eocene transgressive-regressive stratigraphy; lithology of the Stone City Formation. of equal proportions of clay and fine to very fine, well sorted, angular quartz sand. The fecal pellets and sediment texture suggest that the sediments were thoroughly mixed by bioturbation during deposition. More compre- hensive and detailed descriptions of the Formation can be found in Stenzel, Krause & Twining (1957), Stanton & Warme (1971), and Stanton (1979). FOSSIL ASSEMBLAGE Macrofossils greater than 1 mm across form the primary data for this study. The body fossils are scattered uniformly throughout the bed and are generally in excellent condition with original shell material and fine morpho- logic details preserved. Microfossils have been studied in only general terms by us (see, 454 STANTON, POWELL AND NELSON TABLE 1. Taxonomic and trophic characteristics of the fossil assemblage. Col. 1: Number of genera and species within the taxon; Col. 2: Percent of individuals in macrofossil assemblage belonging to taxon; Col. 3: Trophic level of taxon in ecologic pyramid—1: Primary producer; 2: primary consumer; 3-5: low to high levels of carnivory; Col. 4: Abundance of bored individuals in taxon—A: greater than 20% of individuals; C: 10-20%; Р: 5-10%; В: less than 5% of individuals are bored; Col. 5: Abundance of crustacean-chipped specimens in taxon—percentage ranges as in Column 4; Col. 6: Feeding information for the taxon. 1 2 3 4 5 6 G-S % Е Вог Пр: Feeding habits Coelenterata OS carnivorous Scleractinia 2-2 CPS — — Alcyonaria 1-1 OF 53 — — Bryozoa 6-7 AL 2 — — phytoplankton Scaphopoda 2-2 1293 С С carnivorous on forams, small bivalves Gastropoda 49.3 Fissurellidae 1-1 <0.1 23 — В detritus and carnivorous оп small sponges Turbinidae 1-1 OR? — R benthic diatoms, filamentous algae, eaten by Fasciolariidae Vitrinellidae 2-2 ONO R P Architectonicidae 14 04 4 R С carnivorous оп anemones Turritellidae 26 AO A A suspended detritus and phytoplankton Caecidae 1-1 922 —- С interstitial diatoms Epitoniidae 2-3 QA 04 — С carnivorous and parasitic оп coelenterates Eulimidae 2-3 O4 — — parasitic on echinoderms (starfish, holothuri- ans), polychaetes Pyramidellidae 1-1 0511342 P — parasitic/carnivorous on polychaetes, bivalves Naticidae 3-3 16.1 34 С А carnivorous оп bivalves, gastropods, scapho- pods Ficidae 1-2 O7? — P 2 Cymatiidae 1-1 0.2 34 Р P — сагпмогои$ on mollusks Muricidae 1-1 Оо — — _ Carnivorous on bivalves, barnacles, gastro- pods Pyramimitridae 1-1 ONE? — 2 e Buccinidae 2-5 3.03 С С scavengers; carnivorous on bivalves, crusta- ceans, worms Nassariidae 1-1 06 2 А А nonselective deposit feeder: diatoms, detritus; scavenger Fasciolariidae 2-2 2.6 3—4 P P carnivorous on gastropods, bivalves, poly- chaetes, barnacles Volutidae 1-2 AS R Р carnivorous оп bivalves Olividae 1-1 OSs #3 R С carnivorous оп small mollusks, forams Marginellidae 1-1 0.4 4? — P carnivorous on ? Mitridae 1-1 <0.1 4? — R carnivorous on crustaceans, sipunculid worms Cancellariidae 2-3 2.003 Р С carnivorous on soft-bodied interstitial micro- organisms Conidae 1-1 <0.1 4 — B carnivorous on herbivorous polychaetes, fish, gastropods Terebridae 2-2 215 4 Р Р carnivorous on worms, enteropneusts Turridae 12-14 129 34 D C carnivorous on annelids, nemerteans Acteocinidae 1-2 29 34 R С carnivorous on other opisthobranchs, forams Mathildidae 2-3 OZ? — С 2 Ringiculidae 1-1 OS R — Carnivorous on polychaetes, forams Bivalvia 36.6 feed from suspension or sediment surface. Dietary preferences generally not known, probably largely microflora and detritus but nonselective, including bacteria, microfauna. Nuculidae 1-1 2a 2 С Р deposit feeder Nuculanidae 14 3:0) 962 — С deposit feeder Arcidae 1-1 021802 — R suspension feeder Noetiidae 1-2 2.62 R R suspension feeder Ostreidae 22 ileal 32 Р Р suspension feeder Anomiidae 1-1 RG В С suspension feeder CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 455 TABLE 1 (Cont.) 1 2 3 4 G-S % TIE Вог Carditidae 1-1 ри D В Diplodontidae 1-1 0512 —- Semelidae 1-1 DS? R Tellinidae 1-2 O22 — Mactridae 1-1 0.1 2 R Veneridae 1-1 Е eae = Corbulidae 3-3 2383 72 © Cephalopoda (Aturia, Belosepia) 0.1 35 — Echinodermata (heart urchin) —0:1 2 = Arthropoda (crustacean) <0.1 34 — Foraminifera OA 3 — Chordata Elasmobranchii 3-3 <0.1 5 — Congridae 1-1 <0.1 35 — Beryciformis 1-2 08 35 — Serranidae 1-1 <0.1 35 — Scianidae 3-3 08 35 — Ophididae 3-3 08 35 — Soleidae 1-1 <0.1 45 — 5 6 Chip. Feeding habits С suspension feeder С suspension feeder R suspension feeder P deposit feeder Р suspension feeder R suspension feeder С suspension feeder carnivorous on crustaceans, fish, mollusks — non-selective deposit feeder — _ Carnivorous, scavenger — diatoms, bacteria — carnivorous on fish, cephalopods — carnivorous on bottom-living fish, crusta- ceans, cephalopods — carnivorous on benthic crustaceans — carnivorous оп benthic crustaceans and poly- chaetes as juvenile, on small fish as adult — carnivorous on benthic polychaetes and crus- taceans, mollusks?; planktonic crustaceans, fish and squid — carnivorous оп benthic crustaceans (shrimp, crabs, stomatopods), juvenile fish, poly- chaetes — carnivorous on benthic crustaceans RE IE EE SE however, Greenfield, 1957). The sample соп- sists of 6,616 individuals representing 96 genera and 120 species. Table 1 lists the fauna at the taxonomic level for which in- formation about trophic characteristics could be found. Before the original community character- istics can be derived from the fossil assem- blage, the post-mortem processes that may have modified the community must be con- sidered. Non-preservation of soft-bodied in- dividuals was probably the principal modifying process. Solution of the calcareous shells was probably not important, based on the fact that small and fragile shells and fine sculptural detail are all excellently preserved. The im- portance of physical processes that might have mixed, winnowed, abraded, and broken shells was minor, based on analysis of 1) shell condition, sorting and orientation, 2) sedi- mentary structures, and 3) right-left ratios of bivalves. In general, the assemblage 1$ an in-place residue of the original community, with modi- fications resulting primarily from non-preser- vation of soft-bodied organisms and biological processes, such as predation. For a more de- tailed description of the fauna, construction of Table 1, laboratory procedures, and tapho- nomic analysis, see Stanton & Nelson (1980). Richness or species diversity is high but equitability is low in the assemblage (Fig. 2). Size-frequency diagrams and survivorship curves have been calculated for some of the most common organisms (Figs. 3-11). In these diagrams, gastropod size is shell length from apex to abapical point. Bivalve size 15 anterior-posterior length. Size values have been converted to age values using the general logarithmic relationship between the two (Levinton 4 Bambach, 1969). The maxi- mum age value for each population is based on the largest specimen, assuming that the largest shell found represents the oldest indi- 456 STANTON, POWELL AND NELSON Number of Species 1 JU 5 10 15 20 % of Individuals FIG. 2. Dominance histogram. Bar height repre- sents the percent of species that account for a given percentage of individuals in the sample ana- lyzed. For example, one species accounted for more than 17% of the individuals whereas 90% of the species each accounts for less than 1% of the individuals. EZI FREQUENCY OF BORED INDIVIDUALS мы | | + | SIZE RANGE PER BAR UNIT 0 5mm 20| (2.5 - 2.9 mm, etc.) 366 30% OF INDIVIDUALS BORED [O FREQUENCY OF INDIVIDUALS Frequency (%) — 20 40 60 60 100 % Max Age FIG. 3. Size-frequency distribution and survivorship curve for Polinices aratus (naticid gastropod). 20r N = 105 | 17% OF INDIVIDUALS BORED 10+ 2 4 6 8 10 12 14 16 18 Frequency (%) Size (mm) % Surviving 1 =L 20 40 60 80 100 FIG. 4. Size-frequency distribution and survivorship curve for Latirus moorei (fasciolariid gastropod). N = 61 7% OF INDIVIDUALS BORED Frequency (%) % Surviving 1 1 AA — = LI E 20 40 50 80 100 % Мах Аде FIG. 5. Size-frequency distribution and survivorship curve for Bonellitia parilis (cancellariid gastropod). CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 457 20 N = 56 30F = 0% OF INDIVIDUALS BORED 3 2 E 10 720 f N - 35 2 5 6% OF INDIVIDUALS BORED 2 $ С 0 SEE 8 sob 2 4 6 8 10 [ra Size (mm) 0 2 4 6 8 10 12 14 Size (mm) % Surviving % Surviving у ao 40 0 a 100 % Max. Age nu | E | FIG. 8. Size-frequency distribution and survivorship FIG. 6. Size-frequency distribution and survivorship curve for Виссйтоп sagenum (buccinid gastro- curve for Michela trabeatoides (turrid gastropod). pod). m о == 71 20 N - 70 = N = = 23% OF INDIVIDUALS BORED = 10% OF INDIVIDUALS BORED > 2 10 > $ 5 10+ o 2 E dle E H Fre le {ra 4 6 8 10 12 14 16 18 1 E E] Size (mm) 0 2 4 6 8 10 Size (mm) 100 ii ni 100 SOF 10F o c > | ё 5 > n o с = 2 = 5 n 1h | SE | 1 1 L 1 (aS ES ee 1 20 40 60 80 100 1 SR Y Se NN! % Max. Age 20 40 60 80 100 % Max. Age FIG. 7. Size-frequency distribution and survivorship curve for Hesperiturris nodocarinatus (turrid gas- FIG. 9. Size-frequency distribution and survivorship tropod). curve for Retusa kellogii (retusid gastropod). 458 STANTON, POWELL AND NELSON 20 N = 731 33% OF INDIVIDUALS BORED Frequency (%) FIG. 10. Size-frequency distribution and survivor- ship curve for Notocorbula texana (corbulid bi- valve). vidual present in the original community. It is also assumed that the death assemblage pro- vides an accurate picture of the population dynamics of the original community (Thayer, 1977). Most of the fossils are small, with only Conus (Conidae) and a small percentage of the Venericardia (Carditidae) and Katherinella (Veneridae) exceeding 2.5 cm maximum length. Thus, differences in maximum size be- tween species are not large, but differences in size-frequency distributions and survivorship curves are significant. The food preferences of the fossils are listed in Table 1, col. 6. These are based on our own observations and on the literature for the living representatives of the fossil taxa (Fretter & Graham, 1962; Graham, 1955; and others referenced herein), and on direct evi- dence of predation in the fossil assemblage. Dietary information derived from living organ- isms is presented at the family level for two reasons. First, it is difficult to justify applying more specific modern feeding characteristics to Eocene fossils. Second, it is difficult to be more specific for a living organism because it may feed at several levels of the ecologic py- N = 245 18% OF INDIVIDUALS BORED Frequency ( th 1 1 = a 1 L 1 20 40 60 80 100 % Max. Age 1 | + FIG. 11. Size-frequency distribution and survivor- ship curve for Vokesula smithvillensis petropolitana (corbulid bivalve). ramid at the same time; it may feed at dif- ferent levels through time, shifting its pre- ferences with age and changing food avail- ability; or it may have very narrow food pre- ferences that differ from locality to locality. In addition, it is difficult to be certain that an organism 1$ actually utilizing what it eats. For example, some amphipods consume plant detritus but derive nutritive value only from the microorganisms attached to it (Fenchel, 1970), and some molluscs may generally feed on de- tritus in the same way, as described, for ex- ample, by Newell (1965) for the gastropod Hydrobia and the bivalve Macoma. Food preferences determined from direct evidence of predation are based on borings caused by gastropods and shells chipped by crustaceans. About 15% of the gastropods, bivalves, and scaphopods are bored. The taxonomic distribution and abundance of bored and chipped shells are indicated in Table 1, cols. 4 and 5, and Figs. 3-11. Based on their shape (Fretter 4 Graham, 1962), the borings are predominantly caused by naticid gastropods. The borings of the three most abundant species, Polinices aratus (13.4%), CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 459 “Natica” (Naticarius) semilunata (2.6%), and other organisms such as larger crustaceans, Sinum bilix (0.1%), cannot be distinguished, birds, or rays (Schäffer, 1972; Trewin 8 however. Muricids and cymatiids also bore Welsh, 1976). Positive evidence of the pre- (Owen, 1966), but are very rare in the assem- sence of any of these organisms 15 otherwise blage. None of the borings can be referred to lacking. either of these groups. Rare borings similar to These dietary data are used to construct those made by the living Octopus are present the ecologic pyramid (Table 1, col. 3) and the and imply, although other evidence 1$ lacking, trophic web for the Eocene community (Fig. that Octopus was a member of the original 12): biocoenosis. Taxa with similar trophic position are Mollusc shells chipped by crustaceans are grouped together in the trophic web. Detailed abundant (Table 1, col. 5). Many prey had information, such as the predation by naticids survived attack, for scars subsequently and crustaceans on each molluscan taxon, patched during further growth are common. could be included in the web by separating None of the chipped shells in the fossil as- out each individual genus or species. This in- semblage can be explained by gastropod pre- formation, however, would make the trophic dation, which produces a distinctive scar very web so complex that it would be unintelligible, different from that of crustaceans (Carriker, and is available in Table 1. More importantly, 1951). Many larger fossil fragments may these detailed interrelationships that could be have been the result of bioclastic breakage by included in the trophic web are speculative for TROPHIC LEVELS: > A Se aa à PRIMARY PRIMARY CARNIVORES PRODUCER| CONSUMER LOW MIDDLE HIGH | 2 3 4 5 Е FORAM SCAPHOPOD FISH & < = Micro- Г] СЕРН =) y E Bar -—-?0 - > O OF ou erbivores CRUST rae a == = =— == > —0 | O Bi px SHARK zu Ка | O GAST 5) IN = H. UR | 7 o © | | NATICID | GAST. | | ZW > 7 Oz о ВУ у | о COEL. OTHER Da | U 2% | >ш LGAst] | CARNIV ) / ета | wo Ни] GAST ‘ F juil y Zooplankton = _ Bao / = аб nu Bee Pycnogonid “oi E а da 2“ Nudibranch / 2 BRY Regular Echinoid , / A 7 Opisthobranch ] / / Enteropneust / / BIV Bivalve Sponge 4 / BRY Bryozoan Holothurian E 7 CEPH Cephalopod Starfish / COEL. Coelenterate Crinoid р CRUST Crustacean Barnacle AS 2, GAST Gastropod Polychaete & | H.UR Heart Urchin other worms FIG. 12. Trophic web of the Main Glauconite bed community. Box sizes are proportional to numbers of individuals at each position. Solid lines and capital lettering indicate components present in the fossil assemblage and feeding relationships documented in the fossil assemblage or based on modern relation- ships. Dashed lines and lower case lettering indicate inferred components and relationships in the original assemblage based on modern trophic data involving components not preserved. FOSSIL CALCULATED ASSEMBLAGE 10% 20% efficiency CR TH 8 49 SN | 4 BREST => 24.6 46 46 D у . à 3 | а = 34 4 123 33! СЯ ко o _|Assemblage~ 4 NEN! 2 2 + + I 5 ¿e = 4| 246 a © Calculated 20% 10% + It + | N 4 —= Individuals FIG. 13. Proportions of individuals at each trophic level in the ecologic pyramid based on data from Fig. 12, and those expected assuming a 10-20% transfer efficiency from one level to another with numbers in level four fixed (see text and Stanton & Nelson, 1980 for explanation). the Eocene community because they are based on modern feeding information, which is itself incomplete. In such a detailed Eocene web based on modern data, there would be much less than meets the eye. Because the different organisms in the Original community are not equally likely to be preserved, the sizes of the boxes in Fig. 12 indicate original abundances as modified by different degrees of preservation. Several lines of evidence indicate that the effect of differential preservation may be considerable: 1) deposit feeding primary consumers are re- latively rare in the fossil assemblage (the sus- pension feeder:deposit feeder ratio is 15.7: 1), but the abundance of fecal pellets in the sedi- ment, their nutritive value (Frankenberg & Smith, 1967), and the high degree of biotur- bation suggest that the original abundance of deposit-feeding organisms must have been much greater. 2) The proportions of indi- viduals in the fossil assemblage at different trophic levels (Fig. 13) do not agree with those expected in modern communities, with ef- ficiency from level-to-level of 10% to 20% (Odum, 1971). Using level 4 as a reference point, carnivores are too numerous relative to primary consumers. MEASURES OF TROPHIC IMPORTANCE The trophic web (Fig. 12) is effectively the State-of-the-art for trophic analysis today. It provides the basis both for description of the taphonomic processes that have formed the fossil assemblage from the original com- munity, and for description of the structure and evolution of ancient communities. It is im- precise, however, because it is based on STANTON, POWELL AND NELSON numerical abundance data. Ideally, a trophic analysis should be based on estimates of the total energy flow transmitted from one trophic level to another in the community. The im- portance of any prey-predator interaction must be judged by the percent of the total energy flow for which that interaction is re- sponsible (e.g. Levine, 1980). A trophic web constructed from a fossil assemblage is in- herently imprecise because of the imper- fections of the fossil record. Working within that framework, however, the web can be im- proved by refining the numerical data to be more representative of the energy flow in- volved. In the following sections we will pro- pose three approaches by which the nu- merical data can be improved in order to es- timate energy flow. Relative age distribution The boxes in Fig. 12, representing relative numerical abundances of individuals at differ- ent positions within the trophic web, are prob- ably a poor estimate of trophic structure be- Cause size and age distribution, as well as numbers, determine the food requirements and thus the energy flow for each species. The adult component of the population of a carnivorous gastropod is probably much more important than the juvenile component be- cause, with increased size, the potential number, size and diversity of prey taken in- crease. Consequently, the number of indi- viduals attaining adulthood should be a better estimate of relative importance of a species than the simple number of individuals. Many molluscan life strategies include the pro- duction of a large number of offspring and thus the recruitment of a large number of ju- veniles into the community (e.g. Thorson, 1966; Fotheringham, 1971). For a predatory species, however, subsequent high juvenile mortality would limit its role in the trophic web because the many juveniles would have con- sumed substantially fewer total prey than the adults, and with high juvenile mortality, the prey demand by few surviving adults would be low. In this case, numbers will over-estimate trophic importance. The effect of age structure on food require- ments of a species can be estimated from survivorship curves, such as those in Figs. 3-11 for several common species from the Main Glauconite bed. The upward concavity of the survivorship curve reflects the intensity of juvenile mortality. It is greatest, for example, CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 461 for Michela and Polinices (Figs. 6 and 3) and is least for Bonellitia and Retusa (Figs. 5 and 9). A curve with high slope at low and high values of age and with a lower slope at an intermediate age indicates high juvenile mortality but then low mortality until old age. Buccitriton and Latirus provide examples of this pattern (Figs. 8 and 4). The survivorship curve is determined by age-dependent causes of mortality. These may be рага- meters of the external physical environment or may be age- or size-dependent predator or prey interactions such as naticid predation, indicated by the cross-hatched bars in the size-frequency distribution plots. This preda- tion pattern affects not only the survivorship of the prey, but also the predator in terms of prey availability. The potential usefulness of sur- vivorship curves in paleoecology is demon- strated by recent studies by Hoffman (1976a, 1976b). Differences in the survivorship curves for the otherwise similar corbulid species of Notocorbula and Vokesula indicate that this form of analysis brings out subtle niche dif- ferences in the community. Species with relatively low juvenile morta- lity, and thus with a larger proportion of the population attaining adulthood, should be re- latively important in terms of biomass and energy flow in the community. The carni- vorous gastropods can be ranked according to this criterion visually from the survivorship curves. In order to rank them quantitatively, the percent and number surviving to 25% and 50% of maximum age have been calculated (Table 2, cols. В and С). It is evident that al- though Polinices is very abundant, most indi- viduals are very small, with only a few percent exceeding 25% of maximum age. In contrast, by this criterion Retusa and Bonellitia appear to be the dominant carnivorous gastropods in the community. Produced Biomass Analysis of survivorship curves takes into account the age and size distribution of a population in determining the species’ relative Significance in the trophic web of the com- munity. That relative significance can be ex- pressed quantitatively by the total biomass of each predator species, because predator bio- mass should be proportional to the biomass of prey consumed. That is, the number of prey required to support a predator population is determined by the amount of energy input re- quired to produce and maintain the predator population. Thus, biomass values reflect the contributions of the different species to energy flow through the community. One way to estimate biomass values is to estimate pro- duction for each predator species from the number and size distribution of individuals in the population. Production estimates are given in Table 2, col. E for seven common predatory gastropods at Stone City. Each number is the total biovolume of individuals of a given species, less the biovolume present at minimum size (assumed to be recruitment size). The biovolumes are based on the ex- ternal gastropod volume so they give over- estimations of true biovolumes, but only the relative numbers are important here to in- dicate the relative importance of the different species. The numbers are, in effect, mea- sures of secondary production of the predator species in the community (equivalent to G of Calow, 1977; Odum, 1971), as estimated from growth of new biomass. Conversion co- efficients between biovolume and biomass are not available from the literature, but are being determined for present day snails to re- fine this procedure. In the meantime, bio- volume provides estimates of relative pro- duction among the seven gastropods. When the values are compared to the numerical abundance values (Table 2, col. A), it is evident that the importance of Polinices is significantly reduced and that of Latirus, the turrids, and Retusa is increased, as was al- ready predicted from analysis of the survivor- ship curves. The locations of maximum size and of size at 50% maximum age for Latirus moorei and Polinices aratus are plotted on a single logarithmic growth curve in Fig. 14. The Size (cm) \ N Latirus moorei ; e maximum size и А size at 50% max. age Polinices aratus о maximum size asize at 50% max. age Age FIG. 14. Comparison of population dynamics for two species as reflected by the location of the maxi- mum size and the size at 50% maximum age on the logarithmic growth curve. STANTON, POWELL AND NELSON 462 PE NI EN EE D A A A D A A el 9 901 iv 922 (рипа209)— вив)//оодаа $15иэ/илцуииз eINnsayoA ez Е 871 el Olt (pinq109)—euexa] EINGIO9OJON Sal 60 gr bl’ € 82 bz 22 LLL (риюэпа)—шпиэбе$ voy ong 60 80 ве rach 02 81 Op ve ELL (piue¡¡9oueo)—syed еци//эиоя 792 02 9'bZ yl G ge GL LL LE (рылп)—5теииеоорои sin edsey 9211 0% 6'be 95! с 8 05 02 2S (риле|0!$е})—иэл00ш $п.цет 0'91 vz рат gu rd a 9 Le LLL (puun)—sapıojesgen вэцэ!и/ 0'€ Sie ez Ach Ob La es! 08 161 (plulooajoe)—Boyay esnjay ov vy Le 19° So 85 er 188 (proneu)—sn]e/e seo/u!jOd (эшпол (эшпол abe ‘хеш (wo) эбе ‘хеш abe ‘хеш оэбе ‘хеш ee ‘хеш sjenplA -01q Se) -oIq Se) %06 je 26e sabe ‘хеш %0G 0} Bul %06 0} Bu 9sz oy би! %55 0} би! -1ри! jo ‘ou ssewoilg ssewolq элцеэнН %OG JE 9ZIS -AIAINS ‘OU -AIAINS % -AININS ‘OÙ -AIAINS % [8101 элцетшпЭ рээпрола 3 3 а à a у SRE ‘эзэшлер se yipim шпилхеш pue jybiay se цзбиа| [eus цим ‘seuoo se spodo4]seb лэщо ¡je pue лэринАэ e se esnjay бицеэд Aq pajejno¡eo ‘эшполо!а шод pa]ewIse si ssewoig q ‘109 ш ‚abe wnwixew 2,06 Je abe элце|эн,, jo чоцеиа!ахеа 10} y | '6!—4хэ1 999 ‘эбе!ашеззе |15$0} ay] ul sarads jo ajdwes aA1yejuasaldas e 10; SSEWOIG элцепшпо pue 'sseuwoiq paonpold ‘abe шпиихеш %0G pue %Gz Je вер abe pue azis 'seouepunqe ¡eouauny ‘г 3719v1L CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 463 Polinices are small at maximum age and at 50% maximum age, with few surviving to adulthood; thus production 1$ lower than might be expected from the number present. Latirus and the turrids, though fewer in number, are larger at maximum age and at 50% maximum age; thus more growth has occurred relative to numbers present. Retusa provides an in- teresting example of a species in which size at 50% maximum age 1$ even less than in Polinices, but 21% of the individuals exceed this age as compared to only 0.5% of Poli- nices. Cumulative Biomass The bulk of the assimilated energy of an organism will be used for growth or main- tenance. Produced biomass measures energy used for growth but does not take into account the energy cost of maintenance of biomass (i.e., respiration, excretion, etc.). As a gastropod gets older, the growth efficiency declines as more and more energy is ex- pended simply in biomass maintenance (Calow, 1977; Wilbur & Owen, 1964). There- fore, estimates of energy flow based on growth alone err in favor of the smaller or ju- venile individuals, although not as badly as do the estimates based on abundance. An analysis that includes maintenance must consider the combination of age, growth, and survivorship in order to estimate cumula- tive prey consumption by the carnivorous gastropod. This can be done by summing the biovolume (as an estimate of biomass) at each age unit by the number of age units pre- sent (the specimen's age). Thus, a gastropod living and growing for 3 age units might have a biomass of 1 gram at age 1, 2 at age 2 and 3 at age 3. The number used to calculate pro- duced biomass (secondary production of the gastropod as measured by growth) would be 3 grams, but the number of grams of predator maintained by the prey population over three years is really 6. Ecologists rarely consider ecological efficiencies summed over an or- ganism's lifespan. Paleoecologists, however, must work with long-term summed data. A specimen represents a single unit of biomass to an ecologist, and ecological inferences, even when time units are included, are nearly always as snapshots, for a point or narrow interval in time. To a paleoecologist, a speci- men represents an individual that lived and grew and was part of a community during a span of time. Thus, the value 6, the cumula- tive biomass, estimates maintenance require- ments such as energy lost by respiration and excretion during the animal's lifespan, where- as produced biomass represents the standing crop. Values of cumulative biomass, esti- mates from cumulative volume, are listed in Table 2, col. F. These values should be con- sidered in relative magnitude of values and importance of species within a column, and their differences in ranking between columns. Comparison of estimates These three approaches give very different, and we believe better, estimates of the roles of the carnivorous gastropods than do abun- dance values. They depend on several as- sumptions associated with the survivorship curves: (1) Conversion of size to age is based on a general logarithmic relationship between the two. Only rarely can age be determined directly from shell form or microstructure. (2) Comparison of different species is based on the assumption that the age-growth rela- tionship 15 similar for each (as in Fig. 14). This assumption 15 only approximate (Hallam, 1967), and maximum age and size data for Recent taxa are sorely needed to calibrate the age-growth relationships (e.g. Comfort, 1957; Green, 1979). (3) Survivorship curves and age determinations are based on length mea- surements, but biomass would be more pre- cise. The use of length measurements intro- duces errors to the extent that the length/ biomass ratio changes as shell form, and par- ticularly rates of whorl expansion and whorl translation, change with size. (4) Minimum age is assumed to be size 0, but for some organisms recruitment size may be sub- stantial. (5) The largest specimen collected is assumed to be representative of the maxi- mum size (and maximum age) of the species. Data from the literature, though sparse, sup- port this in all cases except Michela and Buc- citriton (see Palmer, 1937; Harris, 1937; Gardner, 1945, for size data). In these two taxa, our largest specimens are approxi- mately 50% actual maximum sizes as re- ported in the literature. The percent maximum age, as used here, refers only to percent maximum age at Stone City. (6) Size differ- ences due to sexual dimorphism are not pre- sent. (7) The specimens in the fossil assem- blage reflect the actual size-frequency distri- butions present in the living community. Re- cent data for living gastropods are insufficient to test these assumptions. We believe that 464 STANTON, POWELL AND NELSON only assumptions 2 and 3 may be sufficiently questionable to introduce significant error into the analysis. Numerical abundance (Table 2, col. A) considers only the number of individuals. In this ranking, Polinices is clearly the dominant gastropod in the community. Survivorship (Table 2, col. В and С) considers the relative abundance during adulthood rather than total abundance. This suggests that Polinices is much less important and that Retusa, Bonelli- tía, and Latirus are relatively more important in the trophic web relative to estimates based on numerical abundance alone. Size fre- quency-survivorship data are imperfect mea- sures of energy flow, however. The longer- lived and larger organisms, Latirus and the two turrids in particular, maintain a larger bio- mass than the others, and, all else being equal, can be expected to consume many more prey over their lifespans than the other gastropods. A species' contribution to total energy flow in the community and, therefore, its importance in the trophic web can be best understood by using estimates of produced biomass (Table 2, col. E) and cumulative bio- mass (Table 2, col. F), which are estimates of the energy required for growth and main- tenance by the animals over their lifespans. The size and relative age at 50% maximum age (Table 2, col. D; Fig. 14) of these gastro- pods indicate that substantial differences are to be expected in the energy required for growth and maintenance. Estimates of pro- duced biomass indicate that Polinices, Retusa, the turrids, and Latirus account for the bulk of the secondary production by pre- dators. Polinices, though most abundant, is responsible for less secondary production than Latirus and no more than about twice that of Retusa and the turrids. If energy used for maintenance of biomass is also con- sidered, using cumulative biomass, Polinices is relatively insignificant, whereas the turrids and Latirus are extremely important for energy flow in the community. The produced biomass and cumulative bio- mass data are not directly comparable, un- fortunately, because neither is in energy units. Thus it is not possible to combine the two and arrive at the best estimate of total energy flow for these species. If biovolume to biomass conversion coefficients were available and if metabolic data for modern analogous species were known, both estimates could be con- verted into energy terms. Produced biomass values could be converted using a biomass- to-energy conversion for typical protoplasm (Odum, 1971). Cumulative biomass would re- quire the use of the relationship М = КМ with k being obtained from respiration values for recent analogues (see Prosser, 1973), M be- ing metabolic rate expressed in Kcal/time unit, where the time unit would be the indi- vidual’s life span, and W being the cumulative biomass maintained by the individual over its lifespan. Food input (i.e., grams of prey con- sumed) could then be estimated because energetic costs of growth and maintenance could be derived directly from the fossil as- semblage, and estimates of assimilation and growth efficiencies could be obtained from recent analogues (e.g. Calow, 1977). Meta- bolic rate has been calculated for extinct mammals using this procedure (Martin, 1980). A rough test of the quality of the estimates of produced and cumulative biomass can be made using the naticids, because in this case both prey and predator can be counted. Mol- luscan prey were consumed primarily by crabs and gastropods. Crabs accounted for roughly 20% of the molluscs preyed upon (Stanton & Nelson, 1980). Naticid borings oc- cur in about 15% of all individuals (bivalves, scaphopods, and gastropods) at Stone City. Thus, of the molluscs escaping crab preda- tion, about 20% succumbed to naticid pre- dators. If all borings were successful (i.e., pre- dation efficiency = 100%), then as much as 20% of the molluscan biomass was naticid prey. Even the largest naticids are small, how- ever. The size distributions of borings in the gastropods (Figs. 3-11) indicate that the larger individuals effectively escaped naticid predation. This is particularly clear for the three largest gastropods, Latirus, Hesperitur- ris, and Michela, in which boring incidence in individuals of size greater than 0.6-1.0 cm is low. Because their prey were smaller than average, it is safe to assume that naticids consumed less than 20% of the molluscan biomass produced at Stone City. If we assume that most molluscan mortality was caused by predation and that Polinices, Latirus, and Buccitriton comprised the bulk of the molluscan predators, the percent of mol- luscan prey each consumed should roughly equal the percent that each comprised of the total produced biomass and cumulative bio- mass of molluscan predators at Stone City. For Polinices, the percentages are 36% ana 3% for produced biomass and cumulative bio- mass, respectively. If the two biomass esti- CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 465 mates were combined, the actual percentage would be somewhere in between. Consider- ing the roughness of the estimates and the assumptions involved, this agrees reasonably well with the estimate of somewhat less than 20% of molluscan biomass consumed by the naticids. Thus the estimates of energy flow presented here appear to be sufficiently pre- cise to be useful in reconstructing the trophic web of this community. INTERPRETATION None of the primary producers (level 1) of the original community is preserved in the fossil assemblage. The preserved primary consumers (level 2) are largely suspension- feeding bivalves and bryozoans. The exact composition of their diet is unknown but is largely phytoplankton for living members of these taxa. The role of detritus is unclear for it may have served as food directly, or as food for the bacteria and fungi that are used as food at this level. Pollen, spores, and diatoms may have been present but probably would have comprised only a minor part of the origi- nal primary food source. None of the primary consumers in the community fed on benthic macrophytes, and organisms that might have been epizoans on such plants are absent from the fauna. Thus, larger plants were probably uncommon or absent. The fact that the important preserved pri- mary consumers are suspension-feeding bi- valves is no more surprising than the fact that the important preserved predators are gas- tropods. The dominant preservable organ- isms in most present-day communities are molluscs. Of these, the bivalves are largely suspension feeders and the higher proso- branch gastropods are largely carnivores. Most other hard-bodied organisms are either suspension feeders or carnivores (e.g., bryo- zoans, brachiopods, corals, etc.). In contrast, most deposit feeders are soft-bodied and not preserved. In many communities today, suspension- feeding bivalves and carnivorous gastropods are much less numerous than soft-bodied organisms such as polychaetes and amphi- pods, and so are not numerical dominants (e.g., Frankenberg, 1971; Boesch, 1973; Maurer, 1977). Thus, almost all fossil assem- blages will consist of suspension feeders and carnivores regardless of the species com- position and of the trophic structure of the original community, and a trophic web based solely on numerical abundances of preserved organisms will present a poor picture ofthe original community. Trophic analysis in pale- ontology must attempt to identify differences in the original communities from subtle differ- ences within the biased records of the fossil assemblages. The prediction of prey-item abundance, based on the estimates of energy flow and population dynamics of the preda- tors, is one potentially valuable tool. In the Stone City assemblage, Polinices appears to have been a dominant predator based on its numerical abundance. This was unlikely, however, because the Polinices indi- viduals were cannibalistic and were pre- dominantly very small, with many fewer of them reaching 50% maximum age (0.5%) than other gastropods in the community. The high juvenile mortality of Polínices may have been caused by recruitment into a sub- optimal environment, because naticids today are typically found in sandier, higher-energy habitats than that of the Stone City Formation (Hunter 8 Grant, 1966; Kinner et al., 1974; Franz, 1976). Survivorship to 50% maximum age was much greater for the gastropods Retusa and Bonellitia, but their small size and short lifespan limited their contribution to energy flow in the community. Survivorship curves scaled to percent of maximum age may be misleading because maximum age may be significantly different for the various species. For example, few indi- viduals of either Hesperiturris or Polinices lived to 50% maximum age, but if growth rates of the two were similar, the maximum age of Hesperiturris would have been much greater than that of Polinices, and individuals of Hes- periturris would have existed in the com- munity over a longer period of time than indi- viduals of Polinices. Thus, the high juvenile mortality suggested by the two survivorship curves for turrids probably occurred over a real time span significantly greater than the juven- Не mortality of Polinices. Consequently, Michela and Hesperiturris, and to a lesser ex- tent Retusa, were important predators. Few of the turrids survived to 50% maximum age, but the size at this age is so large that the pro- duced biomass and cumulative biomass are large. Moreover, the total number of turrid genera and species at Stone City is large. Al- though a large percent of individuals of Bonel- Ша and Retusa survived to 50% maximum age, the individuals were small and probably had a short lifespan (about one year for 466 STANTON, POWELL AND NELSON Retusa; Smith, 1967). Thus, they probably had a limited role in the energy flow within the community. Prey items for Retusa probably included other molluscs among the preserved component as well as some soft-bodied ani- mals (Smith, 1967). Prey items for the turrids were primarily soft-bodied (Fretter 4 Graham, 1962) and therefore not preserved. The fasciolariid, Latirus, appears to have been the dominant predator. Present day fasciolariids prey primarily on molluscs (Wells, 1958), and if Latirus did likewise, it must have been one of the primary predators on the pre- served component, the molluscs, at Stone City, including many other carnivorous gas- tropods. Judging from the abundance of crab- nipped shells, crabs were also important pre- dators of molluscs (see also Virnstein, 1977). and the Polinices-molluscan prey link was third in importance. The ecologic pyramid based on the pre- served component (Fig. 13) indicates that a large рай of the original Stone City com- munity was not preserved. The amount, how- ever, is difficult to determine. A comparison of produced biomass and cumulative biomass for Latirus and Polinices, which fed on pre- served prey, and the two turrids measured which fed on non-preserved prey, suggests that at least % to Уз of all prey biomass was not preserved. If the other species of turrids (turrids comprise Ya of all Stone City gastro- pods) have similar size-frequency distri- butions, the estimated percentage of soft- bodied prey increases to at least one half. It is important to recognize that trophic an- alysis of fossil assemblages is inherently bi- ased toward the preservable components. Predator-prey links among preservable or- ganisms are preserved, and some predator- prey links between preserved and non-pre- served organisms can be reconstructed. However, the importance of predator-prey links between non-preservable organisms can almost never be estimated, although they are of great importance in modern com- munities. Thus the estimate that one-half of the prey biomass is non-preserved, based on information retained in the fossil record, must be a minimum. In all likelihood, the Stone City community was dominated by non-preserv- able organisms both numerically and in bio- mass. There are several other pieces of evi- dence which support this contention: 1. Crab predation was important. Crabs feed on both the preservable and non-preservable com- ponent today (Virnstein, 1977) and probably also did in the Eocene. 2. The sediment is pelleted and bioturbated. Those responsible were most probably soft-bodied deposit feeders. 3. Present day soft-bottom com- munities are rarely dominated by gastropods as is the assemblage in the Stone City forma- tion—rarely does any carnivorous gastropod make up 1% or more of the total individuals present (for example, Frankenberg, 1971; Maurer, 1977; Sanders, 1958; Sanders et al., 1962; and many others). This is not unex- pected since a carnivorous trophic position re- quires that the animal be relatively rare. Be- cause the gastropods are even more common than the bivalves at Stone City, it is reason- able to expect that the entire preserved com- ponent was a small fraction of the living com- munity. The bivalves are at a lower trophic position than the gastropods and so might be expected to be much more common. That they are not suggests that they, too, were not numerical dominants. 4. Dominant predators feeding on molluscs are limited to the fascio- lariid, naticids, and the few buccinids. Based on numerical abundance, 45% of the gastro- pods fed on soft-bodied prey and 47% on molluscs, but 70% of the latter are naticids. If molluscs were really a major component, mol- lusc-feeding predators would be expected to be more abundant. The Stone City community was predomi- nantly composed of soft-bodied organisms; preservable organisms probably made up less than one-half of the biomass; the trophic web contained primarily soft-bodied prey and both preserved and non-preserved predators, with the well preserved molluscan predator- prey links playing a subordinate role in the entire community. These conclusions are consistent with recent studies on the preserv- ability of organisms in Recent communities. DISCUSSION Trophic analysis can be a powerful tool in reconstructing the original community. It is more useful than trace fossils in determining the non-preservable component of the biota because bioturbation is caused by processes operating at a different time scale than the accumulation of body fossils. For example, in heavily bioturbated areas, the sediment may be completely reworked once or more within the lifespan of the preservable organisms present (e.g., Powell, 1977, and references therein). Thus, the number of times the sedi- CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 467 ment is reworked 1$ not preserved, and esti- mation of the relative standing crop of deposit feeders is not possible. Trophic analysis, however, uses the preservable component to estimate the non-preservable, and because they are bound together trophically, their in- teractions are likely to be within the same time frame. Thus, estimation of relative com- position should be possible, yielding a much more complete reconstruction of the com- munity. Trophic analysis, even at this primitive level, can be a useful tool in identifying and estima- ting the role of the non-preservable com- ponent in the fossil community. The basic as- sumptions, however, as described earlier, need to be evaluated in recent communities, even though the trophic analysis of a fossil community can never be exactly comparable to that of a Recent community. The ecologist can deal with the community in space and time as the faunal composition varies over periods of months to a few years (e.g., Bu- chanan et al., 1974; Davis & VanBlaricom, 1978; Poore € Rainer, 1979). In effect, he can integrate a number of separate snapshots of the community over time. On the other hand, the paleoecologist can never recognize or quantify short-term fluctations of this type in a time-averaged fossil assemblage (Peterson, 1977). Preservation acts as a low-pass filter, passing through only the long-term changes and filtering out the short-term community changes in species composition and in abun- dance and size of individuals, which con- tinuously modify the trophic web. Within the “noise” of short-term fluctuations are also dis- turbance-related phenomena which may have had profound effects on species composition and trophic structure. The fossil record pre- serves the integrated average, upon which paleoecological trophic theory must be based. Ecological theory rarely, if ever, addresses phenomena on this time scale. The analysis of the Stone City fossil as- semblage suggests that the paleoecologist's ability to reconstruct a community trophically, based on the assumption that the fossil as- semblage can be treated as a snapshot, is relatively far advanced. Our ability to interpret the data contained in the analysis of energy flow phenomena—survivorship curves and size frequency distributions, for example—is inhibited primarily by the paucity of data avail- able for modern communities. These data can be acquired, however, so much progress can be expected in the future. On the other hand, it is the assumption that the community can indeed be treated as a snapshot that needs careful consideration. Here paleoecologic theory (and ecological theory) is not suffi- ciently far advanced. It is here that work must be done in recent communities to determine how to take the short-term noisy ecologic re- cord and adapt it to the long-term picture needed for paleoecologic trophic analysis. Multiple stable points (Gray, 1977), distur- bances (Woodin, 1978), and periodic in- vasions of explosive opportunistic species (Levinton, 1970; Grassle & Grassle, 1974) all frequently occur in modern communities. The time scales, however, are within the time- determined noise level of paleoecology. In some cases phenomena such as oppor- tunistic incursions may be recognized, and the naticids at Stone City may well represent an example. It is probably not generally possible to quantify and subtract out these phenomena, however, so community recon- struction will always yield a “filtered” com- munity, and paleocological trophic theory must be based on the long-term averaged picture. ACKNOWLEDGMENTS We thank Sayed El-Sayed and Stefan Gartner for their suggestions, which improved the manuscript. REFERENCES CITED BOESCH, D., 1973, Classification and community structure of macrobenthos in the Hampton Roads area, Virginia. Marine Biology, 21: 226- 244. BOSENCE, D., 1979, Live and dead faunas from coralline algal gravels, Co. Galway. Palae- ontology, 22: 449-478. BUCHANAN, J., KINGSTON, P. 8 SHAEDER, M. 1974, Long-term population trends of the benthic macrofauna in the offshore mud of the North- umberland coast. Journal of the Marine Bio- logical Association of the United Kingdom, 54: 785-795. CALOW, P., 1977, Conversion efficiencies in he- terotrophic organisms. Biological Reviews, 52: 385—409. CARRIKER, М. R., 1951, Observations оп the penetration of tightly closing bivalves by Busycon and other predators. Ecology, 32: 73-83. COMFORT, A., 1957, The duration of life in mol- luscs. Proceedings of the Malacological Society of London, 32: 219-241. 468 STANTON, POWELL AND NELSON DAVIS, М. & VANBLARICOM, С. R., 1978, Spatial and temporal heterogeneity in a sand bottom epifaunal community of invertebrates in shallow water. Limnology and Oceanography, 23: 417- 427. FENCHEL, T., 1970, Studies on the decomposition of organic detritus derived from the turtle grass Thalassia testudinum. Limnology and Oceano- graphy, 15: 14-20. FOTHERINGHAM, N., 1971, Life history patterns of the littoral gastropods Shaskyus festivus (Hinds) and Ocenebra poulsoni Carpenter (Ргозо- branchia: Muricidae). Ecology, 52: 742-757. FRANKENBERG, D., 1971, The dynamics of benthic communities off Georgia, U.S.A. Tha- lassia Jugoslavica, 7: 49-55. FRANKENBERG, D. 8 SMITH, K. L., 1967, Copro- phagy in marine animals. Limnology and Oceanography, 12: 443-450. FRANZ, D., 1976, Benthic molluscan assemblages in relation to sediment gradients in northeastern Long Island Sound, Connecticut. Malacologia, 15: 377-399. FRETTER, V. & GRAHAM, A., 1962, British Pro- sobranch Molluscs. Ray Society, London, 755 p. GARDNER, J., 1945, Mollusca of the Tertiary for- mations of northeastern Mexico. Geological Society of America Memoir, 11: 1-332. GRAHAM, A., 1955, Molluscan diets. Proceedings of the Malacological Society of London, 31: 144— 159. GRASSLE, J. & GRASSLE, J., 1974, Opportunistic life histories and genetic systems in marine benthic polychaetes. Journal of Marine Re- search, 32: 253-284. GRAY, J. S., 1977, The stability of benthic eco- systems. Helgolánder wissenschaftliche Meeresuntersuchungen, 30: 427-444. GREEN, R., 1979, Matrix population models ap- plied to living populations and death assem- blages. American Journal of Science, 279: 481- 487. GREENFIELD, L. L., 1957, Stratigraphy and micro- paleontology of the Cook Mountain Formation, western Fayette County, Texas. Unpubl. М.А. thesis, University of Texas, 71 p. HALLAM, A., 1967, The interpretation of size-fre- quency distributions in molluscan death assem- blages. Palaeontology, 10: 25-42. HARRIS, G., 1937, Turrid illustrations, mainly Clai- bornian. Paleontographica Americana, 2(7): 1-96. HOFFMAN, A., 1976a, Mortality patterns of some gastropods from the Badenian (Miocene) Koryt- nica Clays, Poland. Neues Jahrbuch für Geo- logie und Paläontologie; Abhandlungen, 152: 293-306. HOFFMAN, A., 1976b, Mortality patterns of some bivalves from the Badenian (Miocene) Korytnica Clays, Poland.Neues Jahrbuch für Geologie und Paläontologie; Monatshefte, 1976(6): 337-349. HOFFMAN, A., PISERA, A. 8 STUDENCKI, W., 1978, Reconstruction of а Miocene kelp- associated macrobenthic ecosystem. Acta Geo- logica Polonica, 28: 377-387. HOLLAND, J., MACIOLEK, М. & OPPENHEIMER, C., 1973, Galveston Bay benthic community structure as an indicator of water quality. Contri- butions in Marine Science, 17: 169-188. HUNTER, W. & GRANT, D., 1966, Estimates of population density and dispersal in the naticid gastropod, Polinices duplicatus, with a dis- cussion of computational methods. Biological Bulletin, 131: 292-307. KINNER, P., MAURER, D. 8 LEATHEM, W., 1974, Benthic invertebrates in Delaware Bay: animal- sediment associations of the dominant species. Internationale Revue der gesamten Hydrobio- logie, 59: 685-701. LAWRENCE, D. R., 1968, Taphonomy and in- formation losses in fossil communities. Geo- logical Society of America Bulletin, 79: 1315-1330. LEVINE, S., 1980, Several measures of trophic structure applicable to complex food webs. Journal of Theoretical Biology, 83: 195-207. LEVINTON, J., 1970, The paleoecological signifi- cance of opportunistic species. Lethaia, 3: 69-78. LEVINTON, J. & BAMBACH, R., 1969, Some eco- logical aspects of bivalve mortality patterns. American Journal of Science, 268: 97-112. MARTIN, R. A., 1980, Body mass and basal meta- bolism of extinct mammals. Comparative Bio- chemistry and Physiology, 66A: 307-314. MAURER, D., 1977, Estuarine benthic inverte- brates of Indian River and Rehoboth Bays, Dela- ware. Internationale Revue der gesamten Hy- drobiologie, 62: 591-629. MYERS, A., 1977a, Sediment processing in a marine subtidal sandy bottom community: 1. Physical aspects. Journal of Marine Research, 35: 609-632. MYERS, А., 19776, Sediment processing in a marine subtidal sandy bottom community: Il. Bio- logical consequences. Journal of Marine Re- search, 35: 633-647. NEWELL, R., 1965, The role of detritus in the nutri- tion of two marine deposit feeders, the proso- branch Hydrobia ulvae and the bivalve Macoma balthica. Proceedings of the Zoological Society of London, 144: 25-45. ODUM, E., 1971, Fundamentals of Ecology, ed. 3, Saunders, Philadelphia, 574 p. OWEN, G., 1966, Feeding. In WILBUR, К. M. 4 YONGE, С. М. (eds.), Physiology of Mollusca Il, Academic Press, New York, 1-51. PALMER, K. V., 1937, The Claibornian Scapho- poda, Gastropoda, and dibranchiate Cephalo- poda of the southern United States. Bulletins of American Paleontology, 7: 1-548. PETERSON, C., 1977, The paleoecological signifi- cance of undetected short-term temporal vari- ability. Journal of Paleontology, 51: 976-981. POORE, G. & RAINER, S., 1979, A three year study of benthos of muddy environments in Port CARNIVOROUS GASTROPODS IN TROPHIC ANALYSIS 469 Phillip Bay, Victoria. Estuarine and Coastal Marine Science, 9: 477—497. POWELL, E. N., 1977, Particle size selection and sediment reworking in a funnel feeder, Lepto- synapta tenuis (Holothuroidea, Synaptidae). /n- ternationale Revue der gesamten Hydrobiologie, 62: 385—408. PROSSER, С. L., 1973, Comparative Animal Phy- siology, ed. 3, Saunders, Philadelphia, 966 p. SANDERS, H. L., 1958, Benthic studies in Buz- zards Bay. |. Animal-sediment relationships. Limnology and Oceanography, 3: 245-248. SANDERS, H. L., GOUDSMIT, E. M., MILLS, E. L. 8 HAMPSON, С. E., 1962, A study of the inter- tidal fauna of Barnstable Harbor, Massachusetts. Limnology and Oceanography, 7: 63-79. SCHAFER, W., 1972, Ecology and Paleoecology Of Marine Environments. University of Chicago, Chicago, Illinois, 568 р. SCHOPF, T., 1978, Fossilization potential of an in- tertidal fauna. Friday Harbor, Washington. Paleobiology, 4: 261-270. SCOTT, R. W., 1978, Approaches to trophic ana- lysis of paleo-communities. Lethaia, 11: 1-14. SMITH, S., 1967, The ecology and life history of Retusa obtusa (Montagu) (Gastropoda, Opis- thobranchia). Canadian Journal of Zoology, 45: 397-405. STANTON, R. J., Jr., 1976, The relationship of fos- sil communities to the original communities of living organisms. In SCOTT, R. W. 8 WEST, В. R. (eds.), Structure and Classification of Paleo- communities, Dowden, Hutchinson and Ross, Stroudsburg, Pa., 107-142. STANTON, R. J., Jr., 1979, The Stone City Forma- tion. In KERSEY, D. G. (ed.), Claiborne sedi- ments of the Brazos Valley, Southeast Texas. Houston Geological Society Guidebook, 73-77. STANTON, R. J., Jr. & DODD, J. R., 1976, The application of trophic structure of fossil com- munities in paleoenvironmental reconstruction. Lethaia, 9: 327-342. STANTON, В. J., Jr. & NELSON, Р. C., 1980, Re- construction of the trophic web in paleontology: Community structure in the Stone City Formation (Middle Eocene, Texas). Journal of Pale- ontology, 54: 118-135. STANTON, В. J., Jr. 8 WARME, J. E., 1971, Stop 1: Stone City Bluff. In PERKINS, B. F. (ed.), Trace Fossils, a Field Guide to selected locali- ties in Pennsylvanian, Permian, Cretaceous and Tertiary rocks of Texas and related Papers. School of Geoscience, Louisiana State Univer- sity Miscellaneous Publication, 71-1, р. 3-10. STENZEL, H. B., KRAUSE, Е. К. 8 TWINING, J. T., 1957, Pelecypods from the type locality of the Stone City beds (Middle Eocene) of Texas. Uni- versity of Texas Publication, 5704: 1-237. THAYER, C., 1977, Recruitment, growth, and mortality of a living articulate brachiopod with im- plications for the interpretation of survivorship curves. Paleobiology, 3: 98-109. THORSON, G., 1966, Some factors influencing the recruitment and establishment of marine bene- thic communities. Netherlands Journal of Sea Research, 3: 267-293. TREWIN, М. H. & WELSH, W., 1976, Formation and composition of a graded estuarine shell bed. Palaeogeography, Palaeoclimatology, Palaeo- ecology, 19: 219-230. VIRNSTEIN, R. W., 1977, The importance of pre- dation by crabs and fishes on benthic infauna in Chesapeake Bay. Ecology, 58: 1199-1217. VIRNSTEIN, В. W., 1979, Predation on estuarine infauna: response patterns of component species. Estuaries, 2: 69-86. WALKER, K. R., 1972, Trophic analysis: a method for studying the function of ancient communities. Journal of Paleontology, 46: 82-93. WALKER, K. 8 BAMBACH, R., 1974, Feeding by benthic invertebrates: classification and termino- logy for paleoecological analysis. Lethaia, 7: 67- 78. WELLS, H., 1958, Predation of pelecypods and gastropods by Fasciolaria hunteria (Perry). Bul- letin of Marine Science of the Gulf and Carib- bean, 8: 152-166. WILBUR, K. M. & OWEN, G., 1964, Growth. In WILBUR, K. M. & YONGE, C. M. (eds.), Physio- logy of Mollusca |, Academic Press, New York, p. 211-242. WILSON, W. H., Jr., 1979, Community structure and species diversity of the sedimentary reefs constructed by Petaloproctus socialis (Poly- chaeta: Maldanidae). Journal of Marine Re- search, 37: 623-664. WOODIN, S. A., 1978, Refuges, disturbance, and community structure: a marine soft-bottom ex- ample. Ecology, 59: 274-284. YOUNG, D. К. & RHOADS, D. C., 1971, Animal sediment relations in Cape Cod Bay, Massa- chusetts. |: a transect study. Marine Biology, 11: 242-254. u ne erre Sd eye ^ : L e Als мА дала | fan err EP EL LR AA PARU де м nr Le N A 5 Vs À y) Srey Г. » фра QUE. RT do o Pen A e ee $ FA NT ERW ITEM or is мы лоза ет al ur EI III ET na IEA ad oe et vil [LEE Y Pu MAA y we В у LE >: eg Fae ys! > й UTA р. АУ NAS WAI MAGA UA NU A ETC Moe re Pep 7 Li % РА Ap: e LA ENT 7 ares A No A LE à 4 De ha м и pis dy eros A DAN MAL D ur mí, € fer rw fi fis = 0 a Y | Вы { ¡e Cr EN № : № | i PRADERA IAB LR in sat NON M ne MT р in eas MATE y AS E E A AS on Vw Ol Ts À a RATE" ia NEE pot wf м 4 . et | Г’ u > y = a "ия i A ni JN o + ) a de п en \ furet TA: h — ИУ y a rev \ nn | w 3 url 1M i = k ve | . Arey nr a | ; ag Maa vere ee My ne КЕ dB ER ur INDEX TO SCIENTIFIC NAMES IN VOLUME 20, NOS. 1-2 An asterisk (*) denotes a new taxon abidjanense, Chaetoderma, 362, 372 Acanthaster, 391, 392 Acanthina, 405, 408, 411, 415 Acanthodoris, 440-442, 444 achaeus, Anomia, 147 acicularis, Cypraea, 346 aciculata, Mazatlania, 312, 313, 346 aciculata, Ocenebra, 414 Acmaea, 292, 300 Acmaeidae, 291-305 acra, Fusiturricula, 311, 338-340 acra, Pleurofusia, 338 Acteocinidae, 454, 462 Actiniaria, 442, 444 Actinonaias, 219, 229, 234 aculeata, Anomia, 149 Adapedonta, 224 adspersa, Tenellia, 441, 443, 447 Adulomya, 164, 185 aegrota, Dicathais, 409 Aeolidacea, 439, 441, 442, 444, 446, 448 Aeolidia, 99-116, 441-445 Aeolidiella, 112, 201, 202, 441-445 Aeolidiidae, 441 Aeolidoidea, 112 affinis, Cardita, 163 Aforia, 425 Agatrix, 311-313, 332, 333 Agladrillia, 307, 311, 335, 338, 425 akanudaensis, Calyptogena, 185 Akebiconcha, 162, 163, 185, 186, 188, 189 Alasmidonta, 219, 234, 237, 243, 245, 251 Alasmidontinae, 234, 249, 250, 253 alata, Proptera, 219, 221, 226 albanyensis, Goniobasis, 64, 65, 67, 68, 70, 71, 76 albida, Turris, 337 albidum, Epitonium, 346 albocrusta, Cuthona, 447 albolaqueata, Calliclava, 424 albus, Dendronotus, 441, 443 Alcyonacea, 445, 448 Alcyonaria, 454 alderi, Aeolidiella, 201, 202 alesidota, Hindsiclava, 313 allognoto, Proserpina, 25 alsidii, Goniotrichum, 135 alta, Hamea, 401 Amblema, 219, 226, 227, 229, 233, 234, 239-241 Amblemidae, 218, 234, 236, 238, 246, 249, 250, 253 Ambleminae, 217, 231-239, 241-246, 249-253 Amblemini, 217, 232, 234, 236-242, 246, 247, 250- 253 amyra, Doto, 441, 443, 447 Anadara, 55 anapetes, Latirus, 311, 313, 325, 326 anceps, Helisoma, 262, 263 апсеу!, Brazzaea, 246 Ancilla, 311-313, 326-328, 346 Ancula, 441, 442, 445, 446, 448 angulata, Gonidea, 219, 252 angulata, Vesicomya, 187 angulatus, Latirus, 313, 346 angulifera, Littorina, 346 anguliferus, Murex, 414 angustior, Littorina, 346 Anisomyaria, 224 Anodonta, 206, 219, 221, 224-226, 233-238, 243- 245, 251, 252, 263 Anodontinae, 217, 231, 233-239, 242-246, 249, 250-253 Anomalodesmata, 267-289 Anomia, 143-150, 407 Anomiacea, 143, 144, 147-149 “Anomiadae,” 143 Anomiidae, 143-151, 454 Ansates, 292 Anthozoa, 440, 442, 444, 445, 448 antillarum, Diadema, 390, 392, 394-399 Antillophos, 311, 323-325, 346 Antiopella, 441-443, 445 anus, Distorsio, 157 Aphera, 307, 311, 332, 333 apiculata, Quadrula, 219, 241, 250 Aplacophora, 359, 361-383 aquatile, Cymatium, 346 Araphidineae, 136 aratus, Polinices, 456, 458, 461, 462 Arbacia, 390 Archaeogastropoda, 1-33, 302 *Archecharax, 17, 18, *20-24 Archidoris, 114 Architectonica, 346 Architectonicidae, 346, 454 Archivesica, 186, 187 Archosargus, 135 Arcidae, 454 Arctica, 163, 183 Arcticidae, 162, 163, 189 arcula, Alasmidonta, 243 arenaria, Mya, 224, 225, 417 arenarius, Murex, 405 arenosa, Lyonsia, 272 Argobuccinum, 385, 388, 389, 392, 396, 400 argus, Argobuccinum, 388, 389, 392, 396, 400 Armina, 441, 444-446 Arminacea, 440, 444 Arminidae, 441 Arthopoda, 455 Ascidiacea, 439-449 asianus,Chicoreus, 158, 160 Assimineidae, 137 Astarte, 182 Asteroidea, 459 athearni, Goniobasis, 64-68, 70-72, 75, 76 atlantica, Callocardia, 162 atra, Polycera, 441—442 atractus, Conus, 313 atratum, Cerithium, 121, 138, 346 Ашпа, 455 aurantia, Trinchesia, 114 (471) 472 MALACOLOGIA aurantius, Conus, 334 auricula, Crucibulum, 316, 346 auricularia, Lymnaea, 263 auritula, Pisania, 346 australasiae, Monoplex, 400 australis, Anomia, 143 australis, Patro, 143-146, 149, 150 Australorbis, 263 axicornis, Chicoreus, 159 Azotobacter, 263 Bacillariophyceae, 136 balteata, Ancilla, 328 bandera, Persicula, 333 banksi, Turritella, 315 barbarensis, Antiopella, 441-443, 445 barretti, Pleurotoma, 337 barretti, Polystira, 311, 314, 315, 335, 337, 338, 347 barretti, Turris, 337 Basommatophora, 256 Batillaria, 124, 127, 137, 138 beana, Entodesma, 268, 272-274 beaui, Lyria, 313 beaumonti, Cumanotus, 441, 443 Bedeva, 404, 405, 417 belcheri, Forreria, 158 Belosepia, 455 berendti, Proserpinella, 23, 24 berjadinensis, Turritella, 315 Beryciformis, 455 bidentata, Despoenella, 26, 29 bidentata, Proserpina, 26, 28, 29 bilabiata, Eurystomella, 445 bilamellata, Onchidoris, 440-442, 444 bilix, Sinum, 259 bimaculata, Venericardia, 35, 41, 42, 52, 54 Biomphalaria, 202 Biplex, 153, 158 Biraphidineae, 136 bisperforatus, Echinodiscus, 391 bisulcatum, Phalium, 391 Bithyniidae, 137 Bittium, 127, 135, 137-139 Bivalvia, 39, 48, 205-252, 267-289, 359, 403, 454,. 455, 458, 459, 464—466 blakeana, Conomitra, 330 blandiana, Cyane, 20, 21 blandianus, Archecharax, 21-23 bonairensis, Pagurus, 135 Bonellitia, 345, 461, 462, 464, 465 Вогзота, 341 Borsoninae, 340, 428, 431, 433-436 Bougainvillia, 447 bowdenensis, Strioterebrum, 311, 335, 336 bowdenensis, Terebra, 336 Brazzaea, 246 brevifrons, Chicoreus, 346 brevifrons, Murex, 404 Brissus, 391 bronni, Thais, 417 brunnea, Acanthodoris, 444 brunneopicta, Truncaria, 325 Bryozoa, 439—449, 454, 459, 465 Buccinidae, 307, 311, 323, 346, 454, 462, 466 Buccinum, 206, 415 Buccitriton, 457, 461-464 buckleyi, Elliptio, 219, 221, 226, 227, 229, 250 buckleyi, Popenaias, 234, 236 bufo, Bursa, 346 Bullata, 241 Bursa, 153, 346, 385, 388, 396 Bursidae, 153, 385, 389, 400 Busycon, 415 caboblanquensis, Conus, 313 caboblanquensis, Fusinus, 311, 313, 325, 327 caboblanquensis, Strombina, 311, 313, 323, 324 cactoides, Caulerpa, 36 Caecidae, 454 Caenogastropoda, 307-347 californica, Armina, 441, 444—446 californica, Cerithidea, 137 californica, Lyonsia, 268-272, 274-283 californicum, Prochaetoderma, 373 Calliclava, 424, 432, 435 Callinectes, 135 Callocardia, 162, 187 Callogonia, 182, 187 Callothrix, 135 Calotrophon, 312, 313, 322, 346 Calybium, 25 Calyptogena, 161-194 Calyptraea, 346 calyptraeformis, Sigapatella, 41, 43, 57 Calyptraeidae, 307, 311, 315, 346 Camplyodiscus, 136 canadense, Chaetoderma, 363 canaliculata, Thais, 404 canaliculatum, Busycon, 415 Cancellaria, 333, 347 Cancellariidae, 307, 311, 333, 347, 454, 462 candei, Antillophos, 323, 325, 346 candidus, Modulus, 117, 139 cantaurana, Paraborsonia, 341 capensis, Gunnarea, 400 Capulidae, 403 Capulus, 404 carchedonius, Modulus, 118, 140 Cardiacea, 233 Cardiidae, 233 Cardita, 163 Carditidae, 162, 163, 183, 455, 458 Cardium, 208, 414 caribaearum, Cassidulus, 390, 391, 395, 397, 399 caribbea, Vesicyoma, 187 caribbeana, Conomitra, 311, 329-331 carinata, Actinonaias, 219, 229 carinata, Neomenia, 361 Carocolla, 17 cartagenensis, Turritella, 315 Carunculina, 219, 229, 234, 239 Cassidae, 153, 307, 309, 311, 319, 346, 385-402 Cassidaria, 321, 385, 388 Cassidulus, 390, 391, 395, 397, 399 catalinae, Triopha, 441, 442, 446—448 cataracta, Anodonta, 219, 221, 224, 226 cataracta, Pyganodon, 234 INDEX, VOL. 20 473 catena, Natica, 405 caudata, Eupleura, 158, 404, 407, 408, 410, 411, 417 caudatus, Falcidens, 362, 363, 365, 373, 375, 376 Caudofoveata, 361, 378 Caulerpa, 36 cedonulli, Conus, 310, 313, 333, 334 Cellana, 292, 300, 354 cellulosus, Murex, 404, 405 Cenodagreutes, 434 Centrales, 136 centralis, Calyptraea, 346 Cepaea, 201 Cephalopoda, 35, 39, 48, 359, 403, 455, 459 cepio, Pododesmus, 143, 147, 148 Cerastoderma, 286 “Ceratodiscinae,” 13 Ceratodiscus, 13 Ceratostoma, 158 Cerberilla, 441-445 Ceres, 1-3, 4, 6-8, 12, 13, 16, 17 *Ceresidae, 1, 12, *13, 15, 16, 20 Ceriantharia, 442, 444 Cerion, 83 Cerithiacea, 117, 142 Cerithidea, 137 Cerithiidae, 117, 136-138, 346 Cerithiinae, 137 Cerithiopsis, 138, 139 Cerithium, 121, 124, 127, 131, 135, 137-139, 346 Chaetoderma, 361-365, 367, 372, 373, 375, 376, 378, 380, 381 Chaetodermatidae, 361, 363, 367, 372, 376, 380 Chaetodermomorpha, 361-363, 367, 373, 375, 376, 378, 380 challengeri, Metachaetoderma, 363, 373 Charonia, 388, 392, 394, 396, 400 chazaliei, Hindsiclava, 312, 338, 339, 347 Chersodespoena, 19 Chicoreus, 158, 159, 456 chilensis, Entodesma, 268, 272-275, 285 chilensis, Pyura, 268 chitanii, Adulomya, 164, 185 chitanii, Akebiconcha, 185 chitanii, Calyptogena, 185 Chlorella, 99-102, 106-112, 114, 115 Chlorophyta, 135 Chordata, 455 chromosoma, Spurilla, 441-445 chuni, Vesicomya, 187 cineraria, Gibbula, 354, 355 cinerea, Cypraea, 346 cinerea, Urosalpinx, 285, 400, 404, 405, 407-415, 417 cinnamomea, Chersodespoena, 19 cinnamomea, Despoena, 19 cinnamomea, Linidiella, 20 circinata, Aforia, 425 Cirripedia, 441, 459 Cladophora, 135 claibornensis, Lampsilis, 219 clathrata, Distorsio, 346 Clathrodrillia, 312, 313, 347 Clathromorphum, 299 clathrus, Thalassiophylum, 300 Clathurellinae, 433-435 clavatus, Oreaster, 391, 392 clavigera, Purpura, 404 Clavinae, 338, 424, 428, 432-436 Clavularia, 445 clenchi, Goniobasis, 65 closter, Fusinus, 312, 313, 325, 346 Clypeaster, 390, 397 Clypeomorus, 137, 138 Cnidaria, 376, 439-449 coarctata, Cypraecassis, 391, 394 Cochlodesma, 269, 286 cockerelli, Laila, 441, 442 Coelenterata, 454, 459 Collisella, 292, 300, 301 colombiana, Subcancilla, 330 Columbella, 346 Columbellidae, 307, 311, 323, 346 columbiana, Cuthona, 441, 443, 447 commissuralis, Obelia, 446, 447 complanata, Elliptio, 219, 224, 250 complanata, Lasmigona, 243 Compsodrillia, 429 concentrica, Pronucula, 35, 41, 43, 51-54 concinnia, Cuthona, 447 Congridae, 455 Conidae, 307, 309, 311, 334, 423, 428, 432-434, 436, 454, 458 Conomitra, 307, 311, 329-331, 340 consobrinus, Conus, 311, 313, 332-334 consors, Crassispira, 338 consors, Hindsiclava, 311, 335, 338 consors, Pleurotoma, 338 consors, Turris, 338 Conus, 310-313, 332-334, 347, 434, 458 cooperi, Coryphella, 441, 443, 447 coppingeri, lothia, 300 Corambe, 441, 442 Corambidae, 441 Corbicula, 205-216 Corbulidae, 455, 461, 462 cordatum, Pleurobema, 219 Cordiera, 341 corneum, Sphaerium, 208, 209 corneus, Planorbarius, 263 cornuta, Cassis, 391, 392 coronatum, Thais, 346 corpulenta, Anodonta, 251 Coryphella, 113, 441, 443, 445, 447 Coryphellidae, 441 Coscinodiscineae, 136 costata, Lasmigona, 219, 226, 239 cousini, Archecharax, 21, 23 cousini, Proserpinella, 23 couvana, Oliva, 328 crassicornis, Hermissenda, 99-116 crassicornis, Phidiana, 441, 443, 445, 447 crassidens, Elliptio, 219, 250 Crassispira, 338, 426, 433 Crassispirinae, 425, 426, 433-435 Crepidula, 135, 346, 407, 408 474 MALACOLOGIA Crinoidea, 459 crocea, Tubularia, 446, 447 Crucibulum, 311, 313-316, 346 crucicula, Mastogloia, 136 Crustacea, 459, 464, 466 crustacea, Callothrix, 135 Cryptobranchia, 292 cryptus, Phascolion, 135 Ctenostomata, 440 Cucullaea, 187 Cumanotus, 441, 443 Cumberlandia, 219, 221, 225, 226, 234, 237 Cumberlandiinae, 234, 237, 250 Cuneamya, 286 curvicostata, Goniobasis, 64, 65, 67, 68, 70, 71, 76 Cuthona, 113, 441, 443, 447 Cuthonidae, 441 Cyane, 18, 20, 21, 23 Cyanophyta, 135 Cyclobranchia, 10 Cyclocardium, 54 Cyclonaias, 217, 219, 221, 225-227, 229, 234, 240, 241, 251 Cyclopecten, 35, 41, 48, 55 cyclostomus, Echinoneus, 391 cylindrica, Orthonymus, 240, 241 cylindrica, Quadrula, 219, 221, 225, 226, 229, 240, 241 Cymatiidae, 153, 346, 385, 386, 389, 400, 401, 454, 459 Cymatium, 153, 156, 157, 160, 346, 385, 388, 394, 396, 399, 400 cymbaeformis, Crepidula, 346 Cypraea, 317, 319, 346 Cypraecassis, 385, 391, 399 Cypraeidae, 307, 310, 311, 316, 346 Cyprina, 183 Cyprinidae, 162, 163 Cyprogenia, 242, 244 Cyrtonaias, 241, 244 cytaeum, Anomia, 147 dalli, Knefastia, 426 danieli, Capulus, 404 Daphnellinae, 429, 433—435 dariensis, Mitra, 330 dariensis, Subcancilla, 328, 330 daucus, Conus, 347 deltoidalis, Tellina, 35, 41, 43, 51, 52 deltoidea, Thais, 346, 404 delumbis, Villosa, 219, 221, 226, 250 demissa, Guekensia, 224, 225 Dendraster, 188 Dendronotacea, 439, 441, 442, 444, 445 Dendronotidae, 441 Dendronotus, 441—445, 447, 448 dennisoni, Morum, 313, 319, 320 Dentalium, 267 dentatus, Trochus, 354 depressa, Despoenella, 25, 29 depressa, Odontostoma, 25, 26 depressa, Proserpina, 2, 8, 10, 25, 26, 28, 29 derbyi, Proserpina, 17 derbyi, Staffola, 17-19 Dermomurex, 346 Despoena, 19, 25 Despoenella, 2, 10, 24-30 Diadema, 390-392, 394-399 Dialidae, 136 Dicathais, 409 dickinsoni, Goniobasis, 64, 65, 67, 68, 70, 71, 74- 78 didyma, Neverita, 404 diemenensis, Notocallista, 35, 41, 44, 52, 54 Digenae, 249, 251 Dimorphoptychia, 16 Dimorphoptychinae, 16 Dimya, 149 Dimyacea, 149 Dimyidae, 149 diomedia, Tritonia, 441, 444-446 Diotocardia, 10, 12 Diplodontidae, 455 Dipodomys, 78 Dirona, 441, 443, 445 Dironidae, 441 Dispotaea, 315, 316 Distaplia, 284 Distorsio, 153, 157, 346 divae, Precuthona, 441, 443 diversicolor, Dendronotus, 441, 443 dombeyanus, Plectomerus, 219, 227, 229 dominguense, Morum, 311, 313-320 donaciformis, Myselia, 35, 41, 44, 51, 52, 54, 55 Dondersia, 381 donmoorei, Мигех, 312, 313, 346 donmoorei, Siphocypraea, 312, 313, 317-319, 346 Doridacea, 439, 441 Doridella, 100, 441, 442 Dotidae, 441 Doto, 441, 443, 447 Dromus, 242, 244 Drosophila, 72-74, 76, 77, 84 Drupa, 416, 418 Dunaliella, 100-102, 106-112, 114 duplicata, Compsodrillia, 429 duplicatus, Polinices, 404, 406-408, 410 ebena, Fusconaia, 219, 227, 229, 250 Eburna, 327 eburneum, Cerithium, 346 Echinodermata, 455, 459 Echinodiscus, 391 Echinometra, 390-392, 394-399 Echinoneus, 390, 391 echinophora, Cassidaria, 388 echinophora, Galeodea, 388, 389, 392, 396, 399, 400 echo, Cymatium, 156 Ectenagena, 162-165, 179, 185-187, 189 ecuadorense, Crucibulum, 315 edmitchelli, Collisella, 300 edule, Cardium, 208 edulis, Mytilus, 54, 206, 213, 214, 407, 413 Elasmobranchii, 455 elegans, Antillophos, 311, 323-325 elegans, Haminoea, 135 elegans, Oenopota, 434 INDEX, VOL. 20 elegans, Okadaia, 404 elegans, Phos, 323 elegans, Tritiaria, 323 eleutheria, Strioterebrum, 337 ellipsis, Aphera, 333 ellipsis, Cancellaria, 333 Elliptio, 217-219, 221, 224-227, 229, 234, 236- 241, 250 Elliptionidae, 234, 249, 253 Elliptioninae, 234, 249 Elliptionini, 234, 240, 251 Elliptoideus, 251 elodes, Lymnaea, 210, 202 elongata, Akebiconcha kawamurai, 186 elongata, Calyptogena, 161-165, 177, 179-181, 185-190 elongata, Ectenagena, 164, 179, 186 elyros, Patro, 143 emarginata, Nucella, 404 Enigmonia, 143, 144, 147-149 Ensis, 54, 206, 286 Enteromorpha, 135 Enteropneusta, 459 Entodesma, 267-289 Entoprocta, 441, 442, 445 eolina, Carocolla, 17 eolina, Ceres, 17 eolina, Proserpina, 17 eolinum, Odontostoma, 17 ephippium, Anomia, 143, 147-149 Epimenia, 361 Epitoniidae, 346, 454 Epitonium, 346 epomis, Agatrix, 311-313, 332, 333 epomis, Tribia, 333 erinacea, Ocenebra, 404, 405, 414, 415, 417 Eschatigenae, 242 esculentus, Tripneustes, 394 etterae, Eupleura caudata, 404 Eubranchidae, 441 Eubranchus, 114, 441, 443, 447 Eucidaris, 391, 392 Eudistoma, 268, 270 Eulimidae, 454 Euparypha, 195 Eupleura, 158, 404, 407, 408, 410, 411, 417 Eurystomella, 445 exiguus, Eubranchus, 114 exoleta, Turritella, 346 ezoensis, Tapes, 162 Facelinidae, 441 falcata, Margaritifera, 219, 221, 237 Falcidens, 361-363, 365, 372, 373, 376, 378, 380, 381 Fasciolaria, 313, 345, 414 Fasciolariidae, 307, 311, 325, 346, 454, 462, 466 Fasciolariinae, 325 favus, Cyclopecten, 35, 41, 48 femorale, Cymatium, 157 Ferrissia, 255, 266 festiva, Tritonia, 441, 444—446 Festuca, 263 Ficidae, 307, 311, 322, 385, 400, 454 475 Ficus, 3118312319 fidicula, Oenopota, 430 filiformis, Syringodium, 134 fimbriatus, Pachycerianthus, 444, 448 Fiona, 441, 443 Fionidae, 441 Fissurellidae, 354 flabellata, Lyonsia, 272 flammea, Cassis, 154, 386, 393, 394, 396-398 flava, Fusconaia, 219, 221, 227, 229, 250 flava, Littorina, 346 floridana, Lyonsia, 267-289 floridana, Thais, 404, 415 floridensis, Goniobasis, 63-68, 70, 71, 73, 74-81 florifer, Chicoreus, 158, 159 florifer, Murex, 404, 405 fluminea, Corbicula, 205-216 foliata, Pterorytis, 404 foliatum, Ceratostoma, 158 follyensis, Urosalpinx cinerea, 400, 404, 405, 412 Foraminifera, 376, 455, 459 fornicata, Crepidula, 135, 407, 408 Forreria, 158 Fragilariacea, 136 fragile, Periploma, 268 fragilis, Leptodea, 219, 250 francesae, Strombina, 323 francisensis, Micromytilus, 35, 41, 42 fratula, Limifossor, 363, 373, 375 frenchiensis, Lepton, 35, 41, 43, 44, 48, 51, 52, 55 fretalis, Entodesma, 268, 273, 277, 285 Friersonia, 241, 244 frondosus, Dendronotus, 441, 443, 445, 447, 448 fulgens, Cuthona, 447 fulgurator, Oliva, 328 fulvescens, Murex, 405, 415 funiculata, Subcancilla, 330 furcigera, Sphacelaria, 135 fusca, Coryphella, 447 Fusconaia, 217, 218, 221, 225-227, 229, 234, 239- 241, 250, 251 fuscus, Laevapax, 255, 257, 262, 263 fusiformis, Омета, 431 Fusininae, 325 Fusinus, 311-313, 325-327, 346 Fusitriton, 396 Fusiturricula, 310-313, 333, 338-340 galbana, Isochrysis, 100, 101, 106, 107, 269, 270 galea, Tonna, 346, 388, 389, 400 Galeodea, 385, 388, 389, 392, 396, 399, 400 Gastropoda, 10, 11, 35, 39, 48, 99-116, 153-160, 195-204, 206, 302, 349-357, 359, 403-422, 451—469 gatunensis, Рапатигех, 311, 319, 322 gatunensis, Paziella, 322 gatunensis, Phyllonotus, 322 gatunensis, Strioterebrum, 311, 335, 336 gatunensis, Terebra, 336 Genitoconia, 367, 380 Genkaimurex, 407 Geukensia, 224, 225 gibbera, Calyptogena, 186, 189 gibbosa, Clathrodrillia, 312, 313, 347 476 MALACOLOGIA Gibbula, 354, 355 gigantea, Megalonaias, 219, 221, 225-227, 229, 251 gigantea, Pleuroploca, 414 gıgas, Archivesica, 187 gigas, Callocardia, 187 gigas, Strombus, 346 glabra, Anomia, 147 glabrata, Ancilla, 312, 328, 346 glabrata, Biomphalaria, 202 glabratus, Australorbis, 263 *glaeserius, Archecharax, 21-*23, 24 glans, Prunum, 312, 347 glaucum, Strioterebrum, 336 Glebula, 219, 227, 229, 234, 242 globulosa, Odontostoma, 25 globulosa, Proserpina, 25, 28 Glossus, 182 Gomphina, 35, 41, 42, 44, 52 Gonidea, 217, 219, 228, 229, 234, 237-239, 241, 252, 253 Gonideinae, 234, 237, 238, 246, 249, 250, 252 Gonideini, 217, 234, 236-238, 246, 247, 250, 251 Goniobasis, 63-81, 83-98 Goniotrichum, 135 gonostoma, Turritella, 315 Gorgonacea, 445, 448 gouldiana, Pandora, 268 gouldii, Lyonsia, 268, 272, 275, 278 Gourmya, 138 grandis, Plagiobrissus, 390, 394, 397 granularis, Bursa, 388 granulatum, Phalium, 346, 391, 392, 394-397 guadalupensis, Pallacera, 346 gualtieriana, Natica, 407, 409 Gunnarea, 400 Gymnobela, 429 Gymnomenia, 362-365, 367, 368, 370, 373, 376, 378, 380 gyrina, Physa, 255, 256, 262, 263 haemastoma, Thais, 404, 408, 409, 411, 415, 417, 418 Halodule, 118, 132, 134 Hamea, 401 Haminoea, 135 hanleyi, Bedeva, 404, 405, 417 hannae, Proserpinella, 24 hartmani, Falcidens, 363, 376 Hastula, 347, 434 hedgpethi, Polycera, 441, 442 Helcion, 292 Helicina, 6, 14, 25, 26 *Helicinacea, 1, 2, *11-13, 16 Helicinidae, 1, 2, 5, 6, 11-16 Helicininae, 13, 14 Helisoma, 257, 262, 263 hembeli, Margaritifera, 219, 221, 226, 237 Hendersonia, 12, 14, 16 Hendersoniinae, 12-15 henekeni, Cypraea, 317, 319 henekeni, Mitra, 328 henekeni, Muracypraea, 317 henekeni, Siphocypraea, 311-313, 316-319, 341 henekeni, Subcancilla, 330 henekeni, Tiara, 328 henslowanum, Pisidium, 209 hepaticum, Cymatium, 157 hepaticus, Polinices, 346 Hermissenda, 99-116 heros, Lunatia, 404, 408, 410, 417 Hesperiturris, 457, 462, 464, 465 Heteranomia, 143, 145, 147, 149 heterodon, Alasmidonta, 243 heterodon, Prolasmidonta, 243 Heterodonta, 224 Heterogenae, 242, 244-246, 249, 251 Heterozostera, 36 Hexacorallia, 440 hiltoni, Phidiana, 441, 443 Hindsiclava, 335, 338, 339, 347, 427, 433—435 hodsoni, Persicula, 311, 313, 331-333 hodsoni, Rabicea, 331 hollisteri, Faciolaria, 313 Holothuroidea, 459 Homogenae, 249, 251, 252 Hopkinsia, 441, 442, 445, 446, 448 Hubertschenckia, 162, 164 hudsoni, Acanthodoris, 444 humerosa, Fusiturricula, 311, 339, 340 humerosa, Surcula, 340 humerosa, Turris, 340 hyalina, Lyonsia, 268, 270, 275, 276, 281 hydiana, Lampsilis, 219, 242 Hydrobia, 458 Hydrobiidae, 137 Hydrocenacea, 1, 11 Hydrocenidae, 1, 11 Hydrozoa, 440, 443, 446—448 Hyriidae, 245 hystricina, Onchidoris, 440, 442, 444 icterina, Elliptio, 219, 224, 229 illacidata, Mitra, 328 illacidata, Subcancilla, 311, 328, 329 illacidata, Tiara, 328 Imaclava, 428, 432, 433, 435 imbecillis, Anodonta, 219, 226, 243 implicata, Anodonta, 219, 226 inaequivalvis, Pandora, 268 inermis, Ophiodermella, 431 infucata, Quincuncina, 219, 226, 241, 250 infundibulum, Latirus, 346 inornata, Ocenebra, 404, 417 instabilis, Collisella, 300 integra, Physa, 263 interlineatus, Dendraster, 188 intermedia, Conomitra, 330 interruptolineata, Persicula, 313, 331-333, 347 iodinea, Coryphella, 441, 443, 445 iole, Fusiturricula, 311, 339, 340 lothia, 300 iris, Dendronotus, 441, 442, 444, 445, 447, 448 iris, Villosa, 219 ischna, Strioterebrum, 311, 312, 335, 336 ischna, Terebra, 336 islacolonis, Aphera, 311, 332, 333 islacolonis, Cancellaria, 333 INDEX, VOL. 20 477 islandica, Arctica, 163 Isochrysis, 100, 101, 106, 107, 114, 269, 270 jacksonensis, Nanomactra, 57 japonica, Ocenebra, 404, 417 jaquensis, Fusiturricula, 311-313, 333, 339, 340 jaquensis, Knefastia, 340 jaquensis, Pleurotoma, 340 jaspideus, Conus, 313 julietta, Oliva, 328 Katelysia, 35, 41, 43, 44, 51-55 Katherinella, 458 kawamurai, Akebiconcha, 162, 163, 186, 189 kawamurai, Calyptogena, 188, 190 Kelliella, 162, 183 Kelliellidae, 162 kellogii, Retusa, 457, 462 Kellyellidae, 162, 182 kennicottii, Suavodrillia, 431 kilmeri, Archivesica, 186 kilmeri, Calyptogena, 161, 177, 180, 183, 186 Knefastia, 313, 340, 426 krebsi, Cymatium, 346 kugleri, Strioterebrum, 335, 336 Kurtziella, 429, 432, 433, 435 kya, Doto, 441, 443 labiatum, Phalium, 391, 392 lacrimula, Siphocypraea, 317 lacteus, Polinices, 346 laeta, Paraborsonia, 341 Laevapex, 255, 257, 262, 263 laevigata, Cassidaria, 321 laevigata, Nitidella, 346 laevigata, Sconsia, 311-313, 319, 321 laevior, Conomitra, 330 Laila, 441, 442 Lamellariidae, 432 Lamellidens, 246 lamellosa, Nucella, 404 lamellosum, Epitonium, 346 Laminaria, 299 Lampsilidae, 249, 252 Lampsilinae, 234, 237, 238, 240, 249-252 Lampsilini, 217, 232, 234, 236-239, 241, 242, 244, 246, 247, 250-253 Lampsilis, 205-208, 214, 219, 227-230, 232-234, 237-239, 241, 242, 244, 250, 252 lanceolata, Elliptio, 219, 229, 241, 250 languilati, Unio, 243 lapillus, Nucella, 404-406, 408, 410-415 lasia, Calyptogena, 187 lasia, Phreagena, 162, 163, 187 Lasmigona, 219, 226, 234, 239, 243 lassula, Agladrillia, 311, 335, 338 Laternula, 268, 286 Latirus, 311, 313, 325, 326, 346, 456, 461—464, 466 lauta, Pisania, 346 leeana, Vesicomya, 187 lehneri, Conomitra, 311, 329-331 leopardus, Conus, 434 Lepeta, 292 Lepetidae, 291, 292 Lepidoptera, 25 Leptodea, 219, 229, 234, 245, 250 Lepton, 35, 41, 43, 44, 48, 51, 52, 55 Leucozonia, 346 lewisi, Lunatia, 404 lewisi, Polinices, 285, 412 Lexingtonia, 218 libya, Tiariturris, 415 Ligumia, 219, 229, 234, 242 limata, Lyria, 311, 313, 329, 331 Limifossor, 362-365, 367, 368, 370, 373, 375, 376, 378, 380, 381 Limifossoridae, 362 lineata, Littorina, 346 lineata, Monodonta, 354, 355 linguifera, Helicina, 25 linguifera, Proserpina, 25 linguleta, Enteromorpha, 135 Linidiella, 2, 6, 16-20 Lissarca, 35, 41, 42 littorea, Littorina, 285 Littorina, 124, 132, 285, 309, 346 Littorinidae, 309, 346 longa, Subcancilla, 330 longa, Vesicomya, 187 Longenae, 241, 244, 251 longissima, Calyptogena, 187 longissima, Cucullaea, 187 Lottia, 292 Lottiidae, 300, 302 Lucina, 183 lucunter, Echinometra, 390-392, 394-399 lunata, Mitrella, 135 Lunatia, 404, 408—410, 417 lutea, Acanthodoris, 444 lutheri, Pavlova, 100, 101, 106, 107 lutosum, Cerithium, 138, 140 Lymnaea, 201, 202, 255, 256, 262, 263 lyngbyaceus, Microcoleus, 135 Lyonsia, 267-289 Lyonsiella, 286 Lyonsiidae, 267-289 Lyria, 311, 313, 329, 331 Lyriinae, 331 Lytechinus, 390, 391 macgintyi, Distorsio, 346 machrochisma, Monia, 147 Macoma, 54, 392, 400, 458 macroptera, Pteropurpura, 159, 160 Mactra, 208 Mactridae, 454 maculata, Triopha, 441, 442, 446-448 madagascariensis, Cassis, 390, 394, 396, 397 *magnifica, Calyptogena, 161, “165, 167-169, 171, 172, 175, 177, 179-185 magnifica, Ectenagena, 165, 179 maiquetiana, Turritella, 315 Malleidae, 233 Mangeliinae, 428—430, 433-437 mappa, Conus, 334, 347 marcanoi, Despoenella, 26, 28 marcanoi, Proserpina, 26, 28 Marchia, 159 marense, Crucibulum, 311, 313-316 marense, Dispotaea, 315 478 MALACOLOGIA marensis, Fusinus, 311, 313, 326, 327 margaritacea, Neotrigonia, 35, 41, 43, 52, 54, 55, 57 Margaritaninae, 237, 252 margaritensis, Phyllonotus, 312, 346 Margaritifera, 219, 221, 225, 226, 230, 231, 233- 235, 237, 238, 241, 245, 247, 252 margaritifera, Margaritifera, 219, 225, 226, 237 Margaritiferidae, 225, 226, 234, 238, 249, 250 Margaritiferinae, 217, 231, 234, 235, 237, 238, 241, 243-247, 249-252 marginalba, Morula, 417 marginalis, Lamellidens, 246 marginatum, Prunum, 347 Marginellidae, 307, 311, 331, 346, 454 тапае, Tellina, 35, 41, 42, 44, 48, 51-55 marina, Zostera, 100 masoni, Fusconaia, 218, 219, 221, 225, 226 Mastogloia, 136 Mathildidae, 454 Mazatlania, 312, 313, 346 Megalonaiadinae, 234, 236, 250, 251 Megalonaias, 206, 219, 221, 225-227, 229, 232, 234, 236, 238-241, 251 megaradulatus, Scutopus, 362, 364, 365, 367- 369373, 375; 376; 378 melanobranchia, Phestilla, 100, 113 meleagris, Littorina, 346 Mellita, 390, 391, 394, 397 Melongena, 346 melongena, Melongena, 346 Melongenidae, 346 Melosira, 135, 136 Melosiraceae, 136 Meoma, 390, 394 mercatoria, Columbella, 346 Mercenaria, 224, 225, 267, 417 mercenaria, Mercenaria, 224, 225, 267, 417 Mesogastropoda, 10, 11, 63-81, 137, 385—402, 432 Mesogenae, 242, 249, 251 messorius, Murex, 346 Metachaetoderma, 363, 373 Michela, 457, 461-465 Microcoleus, 135 Micromytilus, 35, 41, 42 Microvoluta, 330 midiensis, Strioterebrum, 337 miliaris, Kelliella, 183 militaris, Hindsiclava, 427 minima, Batillaria, 138 Mitra, 328, 330, 341 Mitrella, 135 Mitridae, 307, 311, 328, 454 Mitrolumninae, 428, 433-435 modioliforma, Calyptogena, 161, 165, 177, 179, 187 modioliforma, Ectenagena, 179, 187 Modulidae, 117-142 Modulus, 117-142 modulus, Modulus, 117-142 Moira, 390 Mollusca, 35-62, 99, 154, 195-204, 302, 359, 380, 423, 458—460, 464—466 Monia, 143, 147 moniliformis, Melosira, 135, 136 Monochrysis, 114 Monodonta, 354, 355 monodonta, Cumberlandia, 219, 221, 225, 226, 237. Monoplacophora, 359 Monoplex, 400 Monotocardia, 10, 11 moorei, Latirus, 456, 461, 462 moraiensis, Calyptogena, 188 moraiensis, Unio, 188, 189 Morula, 417 Morum, 311, 313, 319, 320, 392 moskalevi, Problacmaea, 299 mosslandica, Cerberilla, 441, 445 muelleri, Zostera, 36 Muracypraea, 317 Murex, 153, 158-160, 312, 313, 346, 404, 405, 414, 415 Muricacea, 403-422 muricata, Onchidoris, 440-442, 444 muricatum, Vasum, 346 muricatus, Tectarius, 346 Muricidae, 153, 158, 307, 309, 311, 322, 346, 385, 399, 414, 454, 459 Muricinae, 322 Muricopsis, 404, 405 mus, Siphocypraea, 312, 314, 317-319, 346 muscarum, Cerithium, 127, 131, 135, 139, 140 musica, Voluta, 347 Mutelacea, 245 Mya, 206, 224, 225, 417 Myacea, 224 Mysella, 35, 41, 44, 51, 52, 54, 55 Mytilacea, 224 Mytilimeria, 267-289 Mytilus, 54, 206, 213, 214, 407, 413 Myurella, 434 Nacella, 292 nana, Cuthona, 113 nanaimoensis, Acanthodoris, 441, 442, 444 Nanomactra, 57 nassa, Leucozonia, 346 Nassariidae, 346, 454 Nassarius, 206, 346 nasuta, Ligumia, 219, 229, 242 Natica, 311, 312, 314, 316, 404, 405, 407, 409, 459 Naticacea, 403-422 Naticarius, 316 Naticidae, 307, 311, 316, 346, 385, 399, 454, 458, 459, 461, 462, 464-467 Navicula, 135, 136 Naviculacea, 136 neapolitana, Spurilla, 113 nebulosa, Littorina, 346 nelsoni, Ceres, 24, 5-7, 17 nemoralis, Cepaea, 201 Neogastropoda, 10, 11, 206, 432 Neomenia, 361 Neomeniomorpha, 361, 362, 364, 373, 376, 378, 380, 381 Neotrigonia, 35, 41, 43, 52, 54, 55, 57 nephele, Sconsia, 312, 321 INDEX, VOL. 20 479 Neritacea, 1, 11, 12 Neritidae, 11 Neritimorpha, 10-12 nesiotes, Lyonsia, 272 Neverita, 404 nicobaricum, Cymatium, 394, 396, 399, 400 niloticus, Trochus, 349-357 nipponica, Akebiconcha, 188 nipponica, Calyptogena, 186, 188-190 nitida, Lunatia, 409, 410 nitida, Natica, 405 nitida, Proserpina, 2-4, 8-10, 25 Nitidella, 346 nitidulum, Chaetoderma, 362-365, 367, 372, 373, 375, 376 Nitzschia, 135, 136 Nitzschiaceae, 136 nobilis, Architectonica, 346 Nodilittorina, 346 nodocarinatus, Hesperiturris, 457, 462 Noetiidae, 454 norvegica, Lyonsia, 272 Notoacmea, 292, 300 Notocallista, 35, 41, 44, 52, 54 Notocorbula, 458, 461, 462 novaezelandiae, Sigapatella, 55 Nucella, 404-406, 408, 410-415 nucleus, Planaxis, 346 Nucula, 54 Nuculanidae, 454 Nuculidae, 454 Nudibranchia, 99-116, 439449, 459 nuttalli, Муштепа, 268, 270, 283, 284, 288 Obelia, 446-448 Obliquaria, 242, 244 oblonga, Oliva, 312, 346 obscura, Doridella, 100 occidentalis, Distaplia, 284 Ocenebra, 404, 405, 414, 415, 417 ochracea, Collisella, 300 ochraceus, Unio, 244 Octocorallia, 440, 444-446 Octopus, 404, 459 Odontostoma, 17, 25, 26 Oenopota, 428, 430, 433-435 Okadaia, 404 Okeniidae, 441 Oliva, 311-313, 326, 328, 346 olivaceus, Eubranchus, 441, 443, 447 Olivella, 312, 346 oliviae, Spurilla, 441-445 Olividae, 307, 311, 327, 346, 454 Onchidorididae, 440, 441, 444 Onchidoris, 440-442, 444 Oniscidia, 319 oniscus, Morum, 392 Ophididae, 455 Ophiodermella, 431 Opisthobranchia, 10, 99, 111, 447, 459 optabilis, Conus, 312, 347 orbignyi, Archecharax, 21, 23 orbignyi, Cyane, 232 ordi, Dipodomys, 78 Oreaster, 391, 392 oregonensis, Argobuccinum, 396 oregonensis, Fusitriton, 396 Orthonymus, 219, 234, 240, 241 ostrearum, Muricopsis, 404, 405 Ostreidae, 454 ovata, Lampsilis ventricosa, 205-208, 219, 242 ovina, Festuca, 263 Owenia, 431 Pachycerianthus, 444, 448 pacifica, Ancula, 441, 442, 445, 446, 448 pacifica, Calyptogena, 161-163, 165, 177, 179, 180, 183, 185-190 pacifica, Corambe, 441, 442 Pagurus, 135 Pallacera, 346 pallida, Tenellia, 113 palustris, Lymnaea, 255, 256, 262, 263 panamensis, Calyptogena, 189 Panamurex, 307, 311, 319, 322, 340 Pandora, 268, 286 papillosa, Aeolidia, 99-116, 441—445 Paraborsonia, 307, 311, 339-341 paraguanensis, Turritella, 311-315 Paraterebra, 347 parilis, Bonellitia, 456, 462 parthenopeum, Cymatium, 346 Partula, 83 parva, Carunculina, 219, 239 Parviturbo, 313 patagonica, Entodesma, 268, 275 Patella, 292, 411 Patellacea, 291 Patellidae, 291, 292, 300, 303 Patelloida, 292 Patina, 292 Patro, 143-150 patula, Purpura, 346 pauperculus, Dermomurex, 346 paulettae, Knefastia, 340 Pavlova, 100, 101, 106, 107 Paziella, 322 Pecten, 206, 208 Pectinibranchia, 10 Pectinidae, 233 Pectinodonta, 292 pellucida, Gymnomenia, 373 pellucidus, Marchia, 159 pellucidus, Pterynotus, 160 Pennales, 136 Pennatulacea, 445, 448 perca, Biplex, 158 peregra, Lymnaea, 255, 262, 263 Periploma, 268, 286 perplexa, Olivella, 312, 346 perplicata, Amblema, 219, 227, 229 perrugata, Urosalpinx, 404, 405 Persicula, 311-313, 331-333, 347 perspectivum, Sinum, 404 Pervicavia, 434 petropolitana, Vokesula, 458 Phacoides, 182 Phaeodactylum, 99, 100, 102, 106, 109-112, 114, 115 Phaeophyta, 135 480 MALACOLOGIA Phalium, 153, 346, 391, 392, 394-397 Phascolion, 135 phaseolina, Thracia, 268 Phestilla, 100, 113 Phidiana, 441, 443, 445, 447 Pholadella, 286 Pholadomyidae, 286 Phos, 323 Phreagena, 162, 163, 187 Phyllonotus, 159, 160, 312, 322, 346 Physa, 255, 256, 262, 263 picta, Dirona, 441, 443, 445 picta, Polystira, 425 pileare, Cymatium, 346 pileolus, Toxopneustes, 391 pileolus, Tripneustes, 391 pilosa, Acanthodoris, 441, 442, 444 pilsbryi, Ficus, 311, 312, 319 Pinna, 286 pinnata, Fiona, 441, 443 pisana, Theba, 195-204 Pisania, 346 Pisidium, 209 pisum, Proserpina, 25, 28 Pitar, 286 Placuna, 143, 144, 147, 149 Placunanomia, 143 Plagiobrissus, 390, 394, 397 plana, Crepidula, 346 Planaxidae, 346 Planaxis, 346 planci, Acanthaster, 391, 392 planiliratus, Conus, 311, 313, 334, 335 *planior, Despoenella, *29, 30 *planior, Prosperina, 27-29-31 Planorbarius, 263 planulata, Proserpina, 25 planum, Crucibulum, 316 Plectomerus, 219, 227, 229, 234, 239-241 Pleurobema, 218, 219, 229, 234, 239, 240 Pleurobeminae, 234, 249, 250, 252 Pleurobemini, 217, 232, 234, 236-240, 242, 246, 247, 250-253 Pleuroceridae, 63-81, 83-98 Pleurofusia 338 Pleurophophis, 162 Pleurophopsis, 162-164 Pleurophoropsis, 162 Pleurophosis, 162 Pleuroploca, 414 Pleurotoma, 337, 338, 340 plicata, Amblema, 219, 226, 227, 229 plumbea, Kurtziella, 429 Pododesmus, 143-149 Policordia, 286 Polinices, 285, 346, 404, 406—408, 410, 412, 456, 458, 461-466 Polycera, 441, 442 Polyceridae, 441 Polychaeta, 459 Polygona, 325 Polyplacophora, 35, 39, 48, 359 Polystira, 311, 314, 315, 335, 337, 338, 347, 425 pomum, Murex, 160, 404, 405 pomum, Phyllonotus, 159 ponderosa, Calyptogena, 165, 179, 189 popei, Popenaias, 237 Popenaiadinae, 234, 236, 250 Popenaias, 234, 236 Porifera, 439 Porolithon, 349, 350, 352-354 Potamididae, 117, 136-138 praetenue, Cochlodesma, 268 Precuthona, 441, 443 probatocephalus, Archosargus, 135 Problacmaea, 292, 299 Prochaetoderma, 361-365, 367, 370-373, 375- 378, 380, 381 Prochaetodermatidae, 363 Prolasmidonta, 243 Pronucula, 35, 41, 43, 51-54 Proptera, 219, 221, 225-227, 229, 231, 234, 252 Propterinae, 249 Proserpina, 1-4, 6, 8, 9, 10, 12, 13, 16-20, 24-31 Proserpinella, 13, 17, 23, 24 Proserpinellinae, 13 Proserpinidae, 1-33 Proserpininae, 1 Proserpinus, 25 Prosobranchia, 2, 10, 12, 83-98, 114, 117-142, 198 Protobranchia, 183 proxima, Goniobasis, 83-98 Prunum, 312, 347 prunum, Prunum, 347 psammion, Eudistoma, 268, 284 pseudoargus, Archidoris, 114 Pseudodon, 238 Pseudodontinae, 234, 238 Pseudomelatominae, 424, 425, 433—435 pseudonana, Thalassiosira, 269, 270 Pseudovertagus, 137, 138 Pteriacea, 233 Pteriidae, 233 Pteropurpura, 159, 160 Pterorytis, 404 Pterynotus, 158, 160 Ptychobranchus, 219, 227, 229, 234, 239, 242 Ptychogenae, 242, 249, 251 ptychostoma, Helicina, 25, 26 pugetensis, Lyonsia, 208, 271, 272, 275 pugilis, Strombus, 346 pulchra, Proserpina, 25 pulchrum, Prunum, 312, 347 Pulmonata, 2, 10, 12, 195-204 pumilio, Strombina, 312, 323, 324, 326 puncticulatus, Conus, 313, 347 puntagordana, Tegula, 313 Purpura, 346, 404, 406 purpurata, Proptera, 219 pustulosa, Quadrula, 219, 229, 241, 251 Pycnogonida, 459 Pyganodon, 234 Pyramidellidae, 454 Pyramimitridae, 454 Pyura, 268 INDEX, VOL. 20 481 quadrangularis, Samarangia, 283 Scaphopoda, 39, 359, 454, 458, 459, 464 quadrispiralis, Strioterebrum, 311, 313, 335, 337 schepmani, Oliva, 311-313, 326, 328 quadrispiralis, Terebra, 337 Schizodonta, 224 Quadrula, 217, 219, 221, 225-227, 229, 234, 238- Scianidae, 455 241, 250, 251 Scleractinia, 454 quadrula, Quadrula, 219, 221, 225, 241, 251 Scobinella, 341 Quadrulidae, 249, 252 Sconsia, 311-313, 319, 321 Quadrulinae, 234, 249, 250 scripta, Oliva, 346 Quincuncina, 219, 226, 234, 239-241, 250 *scudderae, Despoenella, *26, 27, 30 quinquiesperforata, Mellita, 390, 391, 394, 397 *scudderae, Proserpina, *26-28, 30, 31 Rabicea, 331 Scurria, 292 radiata, Lampsilis, 219, 228, 229, 239, 242, 244 Scutibranchia, 10 raduliferum, Prochaetoderma, 370 Scutopus, 361-365, 367-370, 373, 375-378, 380, ramosus, Chicoreus, 159 381 raninus, Strombus, 346 Semelidae, 455 Rapana, 404406, 417 semicarinata, Goniobasis, 83-98 recta, Ligumia, 219, 242 semigranosum, Phalium, 391, 392 recticanalis, Latirus, 313 semilunata, “Natica” (Naticarius), 459 Rectidentinae, 234, 249 Septariidae, 11 recurvirostris, Murex, 313 Serranidae, 455 reflexa, Obliquaria, 244 setosum, Diadema, 391, 392 reticularis, Oliva, 313 severa, Natica, 404 reticulata, Cancellaria, 347 sibogae, Phestilla, 113 reticulata, Distorsio, 157 Sigapatella, 41, 43, 55, 57 Retusa, 457, 461—466 simplex, Anomia, 143, 147, 148, 407 Rhabdonema, 136 simplex, Goniobasis, 83-98 rhadina, Mitra, 330 simplex, Oenopota, 430 rhadina, Subcancilla, 311, 329, 330 Simrothiella, 381 Rhinoclavis, 127, 131, 137, 138 Sinum, 404, 459 rhodoceras, Acanthodoris, 444 Siphocypraea, 310-314, 316-319, 341, 346 Rhodopetala, 291-305 smithvillensis, Vokesula, 462 *Rhodopetalinae, 291-*302-305 Soleidae, 455 Rhodophyta, 135 Solemya, 164, 189 rhytiphora, Katelysia, 35, 41, 43, 44, 51-55 Solemyidae, 164 ricina, Drupa, 416, 418 Solen, 41, 42, 52-54, 57 Ringiculidae, 454 Solenogastres, 361 rivularis, Ferrissia, 255-266 solidissima, Spisula, 417 robustus, Scutopus, 362, 368, 369, 373, 377 soyoae, Akebiconcha, 189 rosacea, Hopkinsia, 441, 442, 445, 446, 448 soyoae, Calyptogena, 186, 188-190 rosaceus, Clypeaster, 390, 397 Sphacelaria, 135 rosea, Rhodopetala, 291-305 Sphaeriacea, 205, 230 rotundata, Glebula, 219, 227 Sphaeriidae, 230 rubela, Tectura, 295, 299, 301 Sphaerium, 208 rubicunda, Charonia, 396 spinosa, Cassis, 391, 392 rubricata, Lissarca, 35, 41, 42 spirata, Acanthina, 405, 408, 415 rubrocincta, Proserpina, 25, 26 Spisula, 417 rudis, Crassispira, 426 splendida, Lampsilis, 219 rufa, Cypraecassis, 391 Spongillidae, 459 rugosa, Cassidaria, 385 springvaleense, Crucibulum, 311, 314, 316 rugulata, Oenopota, 430 springvaleense, Dispotaea, 316 rustica, Thais, 346 spurca, Cypraea, 346 rustyus, Eubranchus, 441, 443, 447 Spurilla, 113, 441-445 Saccorhiza, 376 Spurillidae, 441 Sagdidae, 283 spurius, Conus, 347 sagenum, Buccitriton, 457, 462 squama, Monia, 147 salleana, Ceres, 2, 6, 17 squamula, Anomia, 149 salleana, Hastula, 347 squamula, Heteranomia, 143, 147 salleana, Proserpina, 17 Staffola, 17-20 Samarangia, 283 stagnalis, Lymnaea, 263 sapidus, Callinectes, 135 stearnsi, Vesicomya, 182, 183 Saxicavacea, 180 steinbergae, Doridella, 441, 442 saxicola, Entodesma, 267, 289 stenopa, Natica, 311, 312, 314, 316 scabra, Collisella, 300 stimpsoni, Conus, 334 482 MALACOLOGIA Stoastoma, 13 testudium, Thalassia, 118, 134 “Stoastomatidae,” 13 Tetragenae, 249, 251 Stolonifera, 445, 448 tetralasmus, Uniomerus, 219, 250 striata, Sconsia, 313, 321 tetraquetra, Tochuina, 441, 444—446 striata, Vesicomya, 182 texana, Notocorbula, 458, 462 Striatella, 136 Thaididae, 309, 346 Strictispirinae, 425, 433—435 Thais, 346, 404—406, 408, 409, 411, 415, 417, 418 Strioterebrum, 311-313, 335-337 Thalassia, 118, 134, 319 Strombidae, 117, 137, 346 Thalassiophylum, 299, 300 Strombina, 309, 311-313, 323, 324, 346 Thalassiosira, 269, 270 Strombus, 137, 346 Theba, 195-204 Strophitus, 245, 247 Theodoxus, 12 Stylommatophora, 195-204 thomasiana, Rapana, 404406, 417 suavis, Vesicomya, 187 Thracia, 268, 286 Suavodrillia, 431 Thyasira, 180, 189 Subcancilla, 311, 328-330 Tiara, 328 subdepressus, Clypeaster, 390 Tiariturris, 425 subramosus, Dendronotus, 441, 443 Tochuina, 441, 444—446 subrugosa, Collisella, 301 Tonna, 346, 388, 389, 400 subtentum, Ptychobranchus, 219, 229, 239 Tonnacea, 385—403 subtruncata, Mactra, 208 Tonnidae, 346, 385 Subulitacea, 432 Toxoglossa, 423—438 sulcata, Melosira, 136 Toxopneustes, 390, 391 sulcidentata, Eupleura, 404 trabeatoides, Michela, 457, 462 sulfureous, Linidiella, 19, 20 Trachycardium, 286 Surcula, 340 tramoserica, Cellana, 354 Surirellanceae, 136 trapexia, Anadara, 55 swifti, Cyane, 20 trialatus, Pterynotus, 158 swifti, Linidiella, 2, 6, 20 Tribia, 333 swifti, Proserpina, 19, 20 tribuloides, Eucidaris, 391, 392 sybaritica, Problacmaea, 299 tribulus, Murex, 159 symmetricus, Conus, 311, 334-336 tricornutum, Phaeodactylum, 99, 100, 102, 106, Synedra, 136 109-112, 114, 115 Syringodium, 134 Triculinae, 83 tabulata, Oenopota, 430 Tridacnidae, 233 Taenioglossa, 432—434 Trigoniidae, 55 Tagelus, 286 trilineata, Coryphella, 441, 443, 447 takanosimensis, Aeolidiella, 441—445 Trinchesia, 114 talpoideus, Limifossor, 362, 364, 365, 367, 368, trinitatensis, Thais, 346 3734373 Triopha, 441, 442, 446-448 tampaensis, Urosalpinx, 404, 405 Triophidae, 441 tampicoensis, Cyrtonaias, 244 Tripneustes, 386, 390-398 tankervillei, Ancilla, 312, 313, 327, 328, 346 triquetra, Eupleura, 411 Tapes, 162 triseriata, Lunatia, 404, 410 tasmanica, Heterozostera, 36 trispiralis, Strioterebrum, 311, 313, 335, 337 taurina, Paraterebra, 347 trispiralis, Terebra, 337 Tectarius, 346 Tritiaria, 323 tectum, Modulus, 117, 118, 137, 139 Tritogonia, 219, 227, 229, 234, 240, 241, 250 Tectura, 292, 295, 299, 301 Tritonia, 441, 444—446 Tecturidae, 292, 302 Tritoniidae, 441 Tegula, 313 trivolvis, Helisoma, 257 Tellina, 35, 41—44, 48, 51-55, 286 Trochidae, 349, 353 Tellinidae, 454 Trochus, 349-357 Tenellia, 113, 441, 443, 447 troscheli, Murex, 159 Terebra, 336, 337, 434 Truncaria, 307, 311, 324, 325 Terebridae, 307, 311, 336, 347, 423, 434, 436, 454 truncata, Laternula, 268, 286 teres, Lampsilis, 219, 239, 241, 242 Truncateilidae, 137 tertiolecta, Dunaliella, 100-102, 106-112, 114 trunculus, Murex, 415 tessellata, Aphera, 333 tuberculata, Cyclonaias, 219, 221, 225, 226, 241, tessellata, Cassis, 391, 392 257 tessellata, Littorina, 346 tuberculata, Nodilittorina, 346 tessellata, Persicula, 312, 347 tuberosa, Cassis, 154, 155, 385, 386, 390, 392- testiculus, Cypraecassis, 385, 391-399 394, 396-399 INDEX, VOL. 20 483 Tubularia, 446, 447 velero, Calotrophon, 312, 313, 322, 346 Tubulipora, 442 Veneracea, 224 tulipa, Fasciolaria, 346, 414 Venericardia, 35, 41, 42, 52, 54, 458 tumidus, Unio, 263 Veneridae, 162, 455, 458 tumulosa, Thais, 417 venezuelana, Ancilla, 311-313, 326-328 Tunicata, 439, 440 venezuelana, Eburna, 327 Turbinellidae, 346 venezuelana, Mitra, 330 Turbinidae, 454 venezuelana, Subcancilla, 311, 329, 330 Turbinimorpha, 10, 11, 13 venezuelensis, Parviturbo, 313 turricula, Oenopota, 430 ventricosa, Lampsilis, 205-208, 219 Turriculinae, 338, 425, 426, 433—435 ventricosa, Meoma, 390, 394 Turridae, 307, 311, 337, 423, 424, 432-437, 454, ventricosus, Tripneustes, 396, 390-393, 395-398 462, 464—466 ventrolineatus, Scutopus, 373 Turrinae, 337, 425, 433—435 Venus, 54 Turris, 337, 338, 340 Vermetidae, 137 Turritella, 126, 311-315, 346 verreauxi, Olivella, 346 Turritellidae, 307, 311, 313, 346, 454 verrucosa, Epimenia, 361 uchimuraensis, Adulomya, 164 verrucosa, Tritogonia, 219, 241, 250 undatum, Buccinum, 415 Verticordiidae, 268 undatus, Conus, 312, 347 Vesicomya, 162, 164, 182, 183, 187, 190 undulata, Alasmidonta, 219 Vesicomyacidae, 162, 164 undulosa, Gomphina, 35, 41, 42, 44, 52 Vesicomyidae, 161-194 unicolor, Brissus, 391 Viana, 15 unimaculata, Imaclava, 428 Vianinae, 1, 10, 13-15 Unio, 188, 189, 236, 243, 244, 253, 263 vibex, Nassarius, 346 unioides, Pleurophopsis, 162 Villosa, 219, 221, 225-227, 229, 234, 242, 250 Uniomerus, 219, 229, 234, 240, 250 viridis, Echinometra, 391 Unionacea, 205, 217-252 Vitrinellidae, 454 Unionidae, 217-253 Vokesula, 458, 461, 462 Unioninae, 234, 236-238, 249, 250, 252, 253 Voluta, 309, 310, 347 unipunctata, Striatella, 136 Volutidae, 307, 310, 311, 331, 346, 454 Urosalpinx, 285, 400, 404, 405, 407-415, 417 Volutomitridae, 307, 311, 330 vaginoides, Solen, 41, 42, 52-54, 57 vulgaris, Octopus, 404 valdiviae, Calyptogena, 189, 190 vulgata, Patella, 411 valdiviae, Vesicomya, 190 waccamawensis, Elliptio, 219, 229 vanbrunti, Echinometra, 391 wahlametensis, Anodonta, 219 vanhyningiana, Goniobasis, 64, 65, 67, 68, 70, 71, waltonense, Crucibulum, 315 73, 75-78 Wilkingia, 286 varicosa, Borsonia, 341 willistoni, Drosophila, 72-74, 76, 77 varicosa, Cordiera, 341 winckworthi, Vesicomya, 187 varicosa, Genkaimurex, 407 Wireniidae, 362, 380 varicosa, Mitra, 341 wrightii, Halodule, 118, 134 varicosa, Paraborsonia, 339, 341 yezoensis, Laminaria, 299 variegata, Charonia, 392 zebra, Cypraea, 346 variegata, Turritella, 313-315, 346 Zephyrinidae, 441 variegatus, Lytechinus, 390, 391 zeylanicum, Phalium, 391 variegatus, Toxopneustes, 390, 391 ziczac, Littorina, 309, 346 varium, Bittium, 135 Zonulispirinae, 429, 433-435 Vasum, 346 Zostera, 36, 100 Vayssiereidae, 403 zosterae, Polycera, 441, 442 ones ee № et 7 ey - | - VEL alar on u Br FED us du Mg Y mr Zu ‘er ee” Sr a v № ur а Ate lao) | му : ver A) MA Y 114 Y 12 Ua = wis poa bin Al Gale rav AA В u n SMS “4 13 y = ' o | TN О О un hl tag MÍN BA вы f | | у M | à ка MMS cd de et VP bat LINE ПЕРОВ u eee Aa ФА nh, M a А CAFE, ae 5 jar ue? YU rie » oe [a фай ne: a gv 4 > . QU EN re sé) Fe (WU À à 4 “i in YT! o IF SA av 067 A sd har ñ A А i “a р 4 » = ПОРА ea 11 Stee A =) FOUT Ayre р TR "| yin ram zh x В у E page. Each table must be submitted double- spaced throughout on a separate manuscript page. Do not use vertical lines. All References cited in the text must be list- ed (bibliographies including uncited items are unacceptable). Each reference should be cited accurately (the Editors will spot check for accuracy) and should be in the style used in recent issues of MALACOLOGIA—except that beginning with Vol. 16 journal titles will be cited complete and unabbreviated. For all manuscripts submitted henceforth, disregard the abbreviations in MALACOLOGIA, 1972, 11(2): 415—426. The journal uses the amper- sand (4) for “and”; “et al.” may be used in the text, but not in the References. In addition to the volume number, complete page numbers of articles and books must be cited. If plates or maps, etc., are not included in the pagina- tion they too must be cited. For books, the publisher and city are required. In systematic papers, synonymies should not give complete citations but should relate by author, date and page to the References. Voucher specimens. In systematic pa- pers, all new type-specimens must be de- posited in museums where they may be con- sulted by other scientists. Beginning with Vol. 16 and when appropriate, MALACOLOGIA will also require that voucher specimens from other kinds of research be deposited in mu- seums. Reprints. When they order 50 or more re- prints, authors will receive 25 additional re- prints gratis; additional copies may be or- dered at the time proof is returned to the Edi- torial Office. Later orders cannot be consid- ered. PAGE COSTS MALACOLOGIA requests authors with grant support to help pay publication costs. MALACOLOGIA requires subsidization for extra long papers. SUBSCRIPTION COSTS For Vol. 21, personal subscriptions are U.S. $17.00 and institutional subscriptions are U.S. $27.00. For information on Vol. 22, address inquiries to the Subscription Office. Vol. 20, No. 2 MALACOLOGIA CONTENTS M. LAZARIDOU-DIMITRIADOU et J. DAGUZAN Etude de l’effet du “groupement” des individus chez Theba pisana (Mol- | lusque Gastéropode Stylommatophore) .................................... ei L. R. KRAEMER The osphradial complex of two freshwater bivalves: histological evaluation Ch and functional Е NE RME EN o GE G. M. DAVIS and SS. L. H. FULLER A Genetic relationships among ecu Unionacea (Bivalvia) of North America ...... a W. D. RUSSELL-HUNTER, A. J. BURKY and R. D. HUNTER Interpopulation variation in calcareous and proteinaceous shell components insthe stream! limpet FermssiaknVvulanis) Sa... ae аа ce al В. $. PREZANT The arenophilic radial mantle glands of the Lyonsiidae (Bivalvia: Anomalo- desmata) with notes on lyonsiid evolution ...........:..... ee cle cs ele a а D. R. LINDBERG | Rhodopetalinae, a new subfamily of Acmaeidae from the boreal Pacific: Pete. anatomy and Systematics: „2.3... earn een. ee e ...2 № Е. J. PETUCH ER A relict Neogene caenogastropod fauna from northern South America ............... 3 FAN G. A. HESLINGA Larval development, settlement and metamorphosis of the tropical gastro- Wi DOT IRGENUS HOLCUS >. Varroa an ee laden ps RS 349 AMERICAN MALACOLOGICAL UNION SYMPOSIUM: FEEDING MECHANISMS OF PREDATORY MOLLUSCS Louisville, Kentucky. 24 July, 1980 A. J. KOHN Introduction. to: the SYMPOSIUM: 0... os Se peices la aaa TE ÓN 35 A. H. SCHELTEMA | Comparative morphology of the radulae and alimentary tracts in the Aplaco- | DGA: О Soe, rome Ee yn RS RE UC ce RARES RS 361 В. М. HUGHES and Н. P. 1. HUGHES | Morphological and behavioural aspects of feeding т the Cassidae (Топ- | nacea! Mesogasiropoda) ..:....2.- 2: en Sarnen а as ÓN 385 M. R. CARRIKER Shell penetration and feeding by naticacean and muricacean gastropods: СЕ О О O DES EE CEE 403 В. L. SHIMEK and A. J. KOHN Functional morphology and evolution of the toxoglossan radula ...:................. 423 J. NYBAKKEN and G. MCDONALD Feeding mechanisms of West American nudibranchs feeding on Bryozoa, € Cnidaria and Ascidiacea, with special respect to the radula ......................... 439 В. J. STANTON, Jr., Е. М. POWELL and Р. С. NELSON The role of carnivorous gastropods in the trophic analysis of a fossil com- (MUM: 0 О В О о о Le EEE 451 ро < a - BOOKBIN" 13 Co. МАУ1 5084 | ПА лу АВТ NR: Sa Ну И и чей, Jase SEEN wy ER A м are re di ZN ve и: А а PRE y à ts usr HA NL KAT мат ae Уи во а А У à En НН van QU PUN IN TS RN M de A A an ONE GA NE En rg. Bae eh ant ет HE Û DLA TT, бет O . ` des ААА у Fae eyes Ут: Puy s hund: NUN DRE HART OR Hayy iyi rus, + A Id abi x 8 pu Kock LOUER IO iets EUR CE: Cr Ÿ [NA he