г. ЕЕ EN SEX Nec othe Dow tr alae DTA ey var rere Sent Ia ie, 4 5 a THEN: beth de TR an ET DTA oP he AA, 1 Tomas at re LUE ae hard a NE se А Aue Nike le dern, в лам mae PRES Pr te et Taste Mo ee Eee ET sn en Term unten un Yeti teten im ton ae an ne daten EA: re Me th Aut nern Pan аль nr ah Ml Y ten MA RP on > e e Л e a a De Neon tee Me Fon! Re A een ET ee ee ee HARVARD UNIVERSITY e Library of the Museum of Comparative Zoology | a À u EN a. =) eee ice | ar | „№ | Е) +. in к к. FR be le. | № a mes sel : panies: « 2% E а ur = | , | a 20 an L Pa pen er и | u E № И} № _ ' al u FEAR ‘ih | VOL. 23 1982-1983 MALACOLOGIA International Journal of Malacology Revista Internacional de Malacologia Journal International de Malacologie Международный Журнал Малакологии Internationale Malakologische Zeitschrift Publication dates Vol. 21, No. 1-2—8 December 1981 Vol. 22, No. 1-2—24 June 1982 Vol. 23, No. 1—18 August 1982 MALACOLOGIA, VOL. 23 CONTENTS AVABOLETZKY" Developmental aspects of the mantle complex in coleoid cephalopods ...... 165 . E. BUROKER Sexuality with respect to shell length and group size in the Japanese OVSIERCrASSOSMOAIQIQAS ES iS ee ee ee ee 271 . G. BUTH & J. J. SULOWAY Biochemical genetics of the snail genus Physa: a comparison of popu- lAULONSHONWOsSDECIOS ee ee oc oie ia! bokeh ом: 351 . M. CHAMBERS On sibling species and genetic diversity in Florida Goniobasis ............. 83 . COPPOIS & C. GLOWACKI Bulimulid land snails from the Galapagos: 1. Factor analysis of Santa CruzelslanGeSpecieSp A ce ak ee nice ne eae ee ee 209 . C. EMBERTON, Jr. Environment and shell shape in the Tahitian land snail Partula (CLEMENT SSSR RE ee meg ae eet ert Ne 23 . T. HANLON The functional organization of chromatophores and iridescent cells in the body patterning of Loligo plei (Cephalopoda: Myopsida) ............... 89 MO НОСНВЕВЕ, Jr: The “kidneys” of cephalopods: a unique habitat for parasites .............. 121 WEEN Commentary on the American Malacological Union Cephalopod Sym- ОИ О Sect о eis sc opie Ste Mcrae se Casas Re Le 203 . M. HUMPHREY & R. L. WALKER The occurrence of Mercenaria mercenaria form notata in Georgia and South Carolina: calculation of phenotypic and genotypic frequencies ........ 75 . W. КАТ Reproduction т a peripheral population of Cyrenoida floridana (Bi- Yalvıa@yremaldidae)y о AR aa cy tes eee ree 47 . W. KAT Genetic and morphological divergence among nominal species of North American Anodonta (Bivalvia: Unionidae) ........................... 361 . М. LOUDA & К. В. MCKAYE Diurnal movements in populations of the prosobranch Lanistes nyas- sanus at Cape Maclear, Lake Malawi, Africa ............................. 13 . MARTOJA & M. TRUCHET Données analytiques sur les concrétions du tissu conjonctif de quel- quesigasteropodes dieaundouce Re ee une er ce ROULE TTC er 333 . MORTON The biology and functional morphology of the twisted ark Trisidos semitorta (Bivalvia: Arcacea) with a discussion on shell “torsion” in the COLEUS Seis ie o clo den O OS OS Oot Cie 375 (= fe MALACOLOGIA CONTENTS (cont.) . PACKARD Morphogenesis of chromatophore patterns in cephalopods: are morph- ological and physiological “units” the same? .... .....:02: 2... nu. 193 . R. PALMER Growth in marine gastropods: a non-destructive technique for inde- pendently measuring shell and body меюйе ...... a 63 sm: EJROPER Introduction to the American Malacological Union Cephalopod Sym- POSIUMi(1IIBO) 5: осла ae a LE A A 87 . H. SCHELTEMA On some Aplacophoran homologies and diets ............................ 427 . SEUGE & R. BLUZAT Effets des conditions d'éclairement sur la croissance de Lymnaea stagnalis (Gastéropode::Pulmoné)) »- - :.. оон jaa aa dl Spa 55 . SEUGE & R. BLUZAT Effets des conditions d'éclairement sur le potentiel reproducteur de Lymnaea stagnalis (Gastéropode Pulmone) .............................. 321 . L. SHIMEK Biology of the northeastern Pacific Turridae. |. Ophiodermella .............. 281 . W. SIGNOR Ш Burrowing and the functional significance of ratchet sculpture in tur- rttelliform gastropods" 2 TA are ste se eae Genes a ee eee 313 . T. SINGLEY Histochemistry and fine structure of the ectodermal epithelium of the Sepiolid-Squid Euprymna SCOlIOPCS ¿msc 2.0.2: CCE DATE 17% . THIRIOT-QUIEVREUX & В. $. SCHELTEMA Planktonic larvae of New England gastropods. V. Bittium alternatum. Triphora nigrocincta, Cerithiopsis emersoni, Lunatia heros and Grepidula;plana....=u.r. 22 as ae ee RER RE net FORCER 37 . G. THOMPSON On sibling species and genetic diversity in Florida Goniobasis ............. 81 . J. VERMEIJ Gastropod shell form, breakage, and repair in relation to predation by фе сгаБба@рра Sas See RS, oe A EE RTE 1 . A. VOSS & R. S. VOSS Phylogenetic relationships in the cephalopod family Cranchiidae (Oegopsida)) s.52:.kicn alc se 5 ca) 5 ео, о 397 . C. WILLAN New Zealand side-gilled sea slugs (Opisthobranchia: Notaspidea: Pleurobranchidae):.. RE ea pein ce tenn a eke fone eee 221 . E. YOUNG & J. M. ARNOLD The functional morphology of a ventral photophore from the meso- pelagic squid; ;Abralia Ingonura”. ле 135 MUS. IE: ZOOL, 1982 plik > - qa FR el AUG 241 Br. matic na и Journal of Malacology $5 tb | en * +A NS A Y U р LE в Internacional de Malacologia 34 1 Journal International de Malacologie \еждународный Журнал Малакологии ernationale 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 Assistants: JOHN B. BURCH MARY DUNN University of Michigan, Ann Arbor GRETCHEN R. EICHHOLTZ CHAMBERLIN ANNE GISMANN Maadi, A. R. 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 OLIVER E. PAGET Environmental Protection Agency Naturhistorisches Museum, Wien, Austria Washington, D.C. ROBERT ROBERTSON CHRISTOPHER J. BAYNE, President Oregon State University, Corvallis CLYDE F. E. ROPER Smithsonian Institution ELMER G. BERRY, Emeritus Washington, D.C. Germantown, Maryland W. D. RUSSELL-HUNTER, Vice-President KENNETH J. BOSS Syracuse University, New York Museum of Comparative Zodlogy Cambridge, Massachusetts NORMAN F. ЗОНЕ | United States Geological Survey JOHN B. BURCH Washington, D.C. MELBOURNE R. CARRIKER RUTH D. TURNER, Alternate University of Delaware, Lewes Museum of Comparative Zodlogy Cambridge, Massachusetts GEORGE M. DAVIS, Executive Secretary-Treasurer SHI-KUEI WU, President-Elect University of Colorado Museum, Boulder PETER JUNG Naturhistorisches Museum, Basel, Switzerland Institute meetings are held the first Friday in December each year at a convenient place. For information, address the President. Copyright © Institute of Malacology, 1982 1981 MUS EDITORIAL BOARD JA ALLEN Marine Biological Station, Millport, United Kingdom E. E. BINDER Museum d’Histoire Naturelle Geneve, Switzerland A. J. CAIN University of Liverpool United Kingdom P. CALOW University of Glasgow United Kingdom A. H. CLARKE, Jr. Mattapoisett, Mass., U.S.A. B. C. CLARKE University of Nottingham United Kingdom 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 Museum National d’Histoire Naturelle Paris, France М ЕВЕТТЕВ University of Reading United Kingdom E. GITTENBERGER Rijksmuseum van Natuurlijke Historie Leiden, Netherlands A. N. GOLIKOV Zoological Institute Leningrad, U.S.S.R. 5 GOUED Harvard University Cambridge, Mass., U.S.A. A. V. GROSSU Universitatea Bucuresti Romania T. HABE Tokai University Shimizu, Japan A. D. HARRISON University of Waterloo Ontario, Canada K. HATAI Tohoku University Sendai, Japan B. HUBENDICK Naturhistoriska Museet Goteborg, Sweden S. HUNT University of Lancaster United Kingdom 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. JALEVER Amsterdam, Netherlands A. LUCAS Faculté des Sciences Brest, France N. MACAROVICI Universitatea “Al. |. Cuza” lasi, Romania C. MEIER-BROOK Tropenmedizinisches Institut Tubingen, Germany (Federal Republic) H. K. MIENIS Hebrew University of Jerusalem Israel J. E MORTON The University Auckland, New Zealand R. NATARAJAN Marine Biological Station Porto Novo, India J. OKLAND University of Oslo Norway T. OKUTANI National Science Museum Tokyo, Japan W. L. PARAENSE Instituto Oswaldo Cruz, Rio de Janeiro Brazil J. J. PARODIZ Carnegie Museum Pittsburgh, U.S.A. М. Е. PONDER Australian Museum Sydney A. W. B. POWELL Auckland Institute & Museum New Zealand R. D. PURCHON Chelsea College of Science & Technology London, United Kingdom O. RAVERA Euratom Ispra, ltaly 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. F. STARMUHLNER Zoologisches Institut der Universitat Wien, Austria Y. | STAROBOGATOV Zoological Institute Leningrad, U.S.S.R. W. STREIFF Université de Caen France J. STUARDO Universidad de Chile, Valparaiso T. E. THOMPSON University of Bristol United Kingdom г. TOFFOLETFTO Societa Malacologica ltaliana Milano W. S. S. VAN BENTHEM JUTTING Domburg, Netherlands J. A. VAN EEDEN Potchefstroom University South Africa J.-J. VAN MOL Université Libre de Bruxelles Belgium N.H. VERDONK Rijksuniversiteit Utrecht, Netherlands B. R. WILSON National Museum of Victoria Melbourne, Australia C. M. YONGE Edinburgh, United Kingdom H. ZEISSLER Leipzig, Germany (Democratic Republic) A. ZILCH Natur-Museum und Forschungs-Institut Senckenberg Frankfurt-am-Main, Germany (Federal Republic) MALACOLOGIA, 1982, 23(1): 1-12 GASTROPOD SHELL FORM, BREAKAGE, AND REPAIR IN RELATION TO PREDATION BY THE CRAB CALAPPA! Geerat J. Vermeij Department of Zoology, University of Maryland, College Park, Md. 20742, U.S.A. ABSTRACT The crab Calappa hepatica (L.) attacks most of its gastropod prey by peeling, a process of chipping away the outer shell wall piece by piece in a spiral direction, beginning at the outer lip. Laboratory studies in Guam showed that 33% to 91% of attacks by Ca/appa were unsuccessful, depending on the prey species. A thick outer lip prevents Calappa from peeling many adult Strombus gibberulus. Thin-lipped Rhinoclavis spp. are protected from lethal peeling by a varix located 240° around the body whorl from the outer lip. Terebra may be protected by its very small aperture. The frequency of lethal breakage was studied in collections of “dead” shells from localities in the tropical Pacific. Postmortem breakage, as inferred from damage in drilled shells, was a minor contributing factor to observed frequencies of lethal breakage. The incidence of lethal breakage in the field was highly correlated with attack success of Ca/appa in the laboratory. Breakage-induced shell repair is common in tropical Pacific gastropods. The incidence of repair is highly correlated with the failure rate of Calappa in the laboratory. Spearman rank correlations between the frequencies of repair and lethal breakage were significant and positive within three of the four species and genera examined. Repair is a reliable indicator of the effectiveness of the shell as protection against lethal breakage, but it cannot be used as a reliable measure of the intensity of lethal breakage. Key words: predation; shell geometry; Calappa; tropical Pacific; sublethal injury. INTRODUCTION Shell breakage in molluscs is important both as a cause of death and as an agent of selection (Vermeij, 1978, 1979). The extent to which breakage is important in selection with respect to shell architecture is likely to vary within and between species both geographi- cally and temporally. Some species experi- ence breakage principally as a cause of death, whereas individuals of other species commonly survive breakage-inducing attacks when the shell has reached a sufficient (criti- cal) size. In general, a shell may be broken either by crushing or by peeling. By crushing | mean that a shell as a whole is compressed be- tween two apposing surfaces, such as the jaws of a vice or of a spiny diodontid porcu- pinefish, the pharyngeal bones of labrid fishes, and the massive crusher claws of many crabs (Liem, 1973; Vermeij, 1976, 1977, 1978; Zipser & Vermeij, 1978; Brown et al., 1979; Palmer, 1979; Yamaoka; 1978). Gonodactylid stomatopod Crustacea, which use the base of the second maxillipeds as a kind of hammer, also practice a form of shell crushing (Caldwell & Dingle, 1975). Experi- ments with various crushing predators and measurements of gastropod shell strength show that a compact spherical shape (low spire, small aperture), thick outer shell wall, strongly thickened outer lip, sturdy external sculpture, tight coiling, and certain features of crystal microstructure are effective deterrents against crushing (Papp et al., 1947; Kitching et al., 1966; Currey & Kohn, 1976; Vermeij, 1976, 1978; Hughes & Elner, 1979; Palmer, 1979; Vermeij & Currey, 1980). When prey approach or exceed the critical size for a par- ticular predator, damage may be confined to the lip, or it may be evident only as tooth- marks or as broken elements of sculpture on the shell's exterior (Vermeij, 1976; Zipser 8 Vermeij, 1978). Crushing predators are most diversified in shallow hard-bottom communi- ties, especially in the tropics (Vermeij, 1978; Palmer, 1979). Lip-peeling is the second type of gastropod shell breakage, in which the body whorl is 1Contribution no. 146 from the University of Guam Marine Laboratory. 2 VERMEIJ broken away piece by piece, beginning at the outer lip, and continuing in a spiral direction until the soft parts are exposed. Well-known lip-peeling predators include spiny lobsters (Palinuridae) (Randall, 1964; Vermeij, 1978) and sand-dwelling crabs of the genus Calappa (Shoup, 1968; Miller, 1975; Vermeij, 1978; Vermeij et al., 1980). Little is known about shell characteristics that prevent peeling predators from being successful shell-breakers. Traits that might be expected to protect gastropods against peel- ing include a thick lip, a narrow or very small aperture that is difficult to penetrate, and the ability to retract the edible soft parts far back into the shell (Vermeij, 1978). Retractility and aperture size are frequently associated with spire height. Species able to withdraw the soft parts far into the shell and away from the vulnerable outer lip are usually high-spired and have a small aperture (Hamilton, in prep.). Because calappids are widespread in soft-bottom habitats throughout the shallow- water marine tropics, an understanding of the shell characteristics that prevent successful peeling would contribute to the interpretation of geographical and temporal patterns in shell form of soft-bottom gastropods. The data that are required include observations on Calappa's attacks in the laboratory, estimates of the incidence of breakage-induced mor- tality in field populations of gastropods, and proof that gastropods in the field can survive attacks by Ca/appa and other shell-peeling predators. If Calappa were always successful in peeling a gastropod, there would be no se- lection between shells with and without traits that prevent lethal peeling. In this paper, | present the three kinds of data, with the fol- lowing additional questions in mind: (1) what is the relation between shell form and suscep- tibility to lethal lip-peeling? (2) How are the incidences of lethal breakage and of break- age-induced shell repair (unsuccessful pre- dation) related? (3) How much local variation exists in the incidence of repair independent of shell form? MATERIALS AND METHODS Crabs—Calappa hepatica (L.)—of various sizes were kept individually in aquaria with running ambient sea water and with enough coarse sand on the bottom for the crabs to bury completely. Each crab was offered a rep- resentative range of sizes of several prey species in order to establish the size of the largest individual of a given species that could be killed by a particular crab. Prey were re- moved as soon as they had been attacked or eaten (usually within a few hours). At any one time, a crab had a choice of several species of prey. All trials were carried out at the Univer- sity of Guam Marine Laboratory at Pago Bay, Guam, in the summer of 1975 and from Janu- ary to April, 1979. Estimates of the importance of lethal break- age as a cause of death were made from col- lections of “dead” shells (those no longer oc- cupied by a gastropod) from several shallow- water soft-bottom sites in the tropical Pacific. All available intact and broken shells, includ- ing apical fragments, were collected by hand. They were later measured for length to the nearest millimeter, and examined for signs of the cause of death. It was impossible to esti- mate the size of shells when only the apex was available. For the present analysis, only pieces and shells 5mm or longer were col- lected and examined. Breakage was considered to be lethal to the gastropod if the extent of injury was equal to or greater than the damage known to be fatal to the same species in laboratory trials with Calappa and other predators (see also Zipser & Vermeij, 1978). If lip breakage was less ex- tensive than would be necessary to kill an indi- vidual, it was assumed not to have caused that individual's death. Because many dead shells are sublethally broken, it is easy to overestimate the importance of breakage as a cause of death. Sometimes, shells are “lethally” broken while they are occupied by hermit crabs (Bertness, 1980) or when empty. Although it is difficult to assess the extent of this artifact, a lower limit of its contribution to the observed number of lethally broken shells can be ob- tained by examining drilled shells. Drilling by gastropods appears to be a cause of death for gastropods but not for hermit crabs. Any “lethal” breakage found in drilled shells must therefore have occurred after the death of the Original snail. A minimum estimate of post- mortem “lethal” breakage, f,, is thus the fre- quency of lethal breakage in drilled shells (number of broken drilled shells, пы, divided by number of drilled shells, nq). Taking the postmortem artifact into ac- count, | define the frequency of lethal break- age as follows: fo = (Mp/n)(1 — fp) (1) CALAPPA PREDATION ON GASTROPODS 3 where ny is the number of lethally- broken shells (drilled and undrilled) and n is the total number of shells in the sample. The frequency of lethal breakage as de- fined here is a reliable index of the contribu- tion of breakage as a cause of death of snails only if there is no bias in the preservation or collection of intact and broken shells. | sus- pect that observed frequencies of lethal breakage are conservative estimates of the importance of breakage to overall mortality. Waves and currents may selectively remove fragments and small whole shells from the population of vacated shells. Diodontids and other predators may crush shells so com- pletely that the apical fragments are unsuit- able as abodes even for competitively sub- ordinate hermit crabs. Some hermit crabs preferentially shun damaged shells (Abrams, 1980; M. D. Bertness, personal communica- tion). Probably the most credible estimates of the frequency of lethal breakage come from high-spired shells from sandy or muddy en- vironments where transport by waves and currents is limited. The frequencies in this paper are from such environments. The ma- terial was collected from the Pacific coast of Panama in the summers of 1976 and 1978, and at Indo-West-Pacific sites from January to July, 1979. “Dead” shells were always col- lected from the intertidal zone or slightly be- low, and never at the strand line above high water mark. The species composition of the dead shells was always similar to that of the living molluscs. Dead shells were analyzed Only when living representatives of the spe- cies in question were also found. Frequencies of shell repair were studied in detail at seven Western Pacific localities and at one Eastern Pacific site. All intact living and dead shells were inspected for signs of break- age-induced shell repair. Depending on shell shape, repairs were counted on 1, 2, or all whorls of the shell with the aid of magnifica- tion. Tiny repaired lip-chips were ignored. In a given sample of intact shells, the frequency of repair was defined as the number of scars per individual. This definition differs slightly from that of some earlier workers (Robba & Osti- nelli, 1975; Raffaelli, 1978; Elner & Raffaelli, 1980), who used the number of repaired shells divided by the total number of shells in a sample as their index of repair. The latter index can vary only between values of 0 and 1, whereas the index used in the present pa- per can take on any value greater than zero. Neither measure takes into account differ- ences between individuals in the number of scars. RESULTS Predation by Calappa— The outer lip of the shell is the first line of defense of a gastropod against peeling by Calappa. In previous work with the gastropods Nassarius luteostoma (Brod. & Sow.) and Strombina clavulus (Sow.) in Panama, | observed that a small (27.1 mm wide) Calappa convexa Saussure was capable of breaking lips up to 1.35 mm in thickness; however, lips thicker than 1.6 mm remained intact (Vermeij, 1978). In the pres- ent study, | offered 9 Strombus gibberulus L. with lips 1.5 mm to 2.3 mm thick to a 75.1 mm wide C. hepatica. None of the 3 S. gibberulus with a lip thicker than 2.0 mm was peeled. When a 41.3 mm wide С. hepatica was of- fered 7 Clypeomorus bifasciatum (Sow.) ranging in lip thickness from 1.0mm to 2.3 mm, and 4 Cerithium columna Sow. with lips 1.1 mm to 1.4 mm thick, only the five shells with lips less than 1.6 mm thick were peeled at least 180°. Calappa thus appears to be pre- vented from peeling by a thick outer lip. In a few instances, a thick lip did not pre- vent lethal breakage. Two shells each from among 14 Strombus mutabilis Swainson and 20 S. gibberulus that were lethally broken by Calappa were peeled from the anterior end or from a hole punched in the dorsum behind the outer lip. In these four shells, the thick lip was circumvented. A breach in the outer lip is no guarantee that a gastropod will be killed by Ca/appa. The high-spired cerithiids Rhinoclavis aspera (L.) and A. fasciata (Brug.) have thin adult lips (0.5 to 0.9 mm and 0.3 to 0.6 mm respective- ly), and more than 90% of attacks by Ca/appa resulted in peeling; however, only 25% of the 64 peels of R. fasciata, and 16% of peels of R. aspera (n = 25) were lethal. In every in- stance of unsuccessful predation, peeling ex- tended through circa 240° of angular distance from the outer lip to the first varix, but not beyond this externally and internally thick- ened structure. Successful peeling requires that the first varix be breached. Similar re- quirements hold for species of Cerithium, whose outer lip is usually thicker (1.0 to 2.1 mm) than in Rhinoclavis. Terebra affinis Gray (lip 0.3 to 0.6mm thick) has no thickened varices, but attacks on this species are usually unsuccessful. Of the 4 VERMEIJ 22 individuals attacked by seven crabs, only two (9.1%) were killed. Although more than 360° of shell must be peeled away in order to expose the soft parts, most unsuccessful peels extended only between 90 and 240°. It is unclear why peeling was discontinued. Possibly, the external tooth of the right claw is too large to fit into the small aperture of the Terebra as the outer shell wall is stripped back. In order to kill Strombus gibberulus, Calappa must peel through an angular dis- tance of at least 105° ifthe attack is initiated at the outer lip, as it usually was. Of 39 adult S. gibberulus that were attacked, 13 (33%) were killed. The others had small scratches on the outer lip, but the damage was not substantial enough to require repair. Only 8 of 31 juven- iles (26%) were broken lethally, and 5 (16%) suffered nonlethal shell injury. The remaining 18 individuals were not attacked, perhaps be- cause they were able to escape from Calappa. There was no indication that the sharply serrated operculum was an effective deterrent against crabs. Broad-apertured naticids (Natica, Polini- ces) were frequently peeled more than 180° by Calappa, but the damage required to kill these snails is slight. С. hepatica 6 to 8 cm in Carapace width can extract a portion of the body of Polinices tumidus (Swainson) (shell length 27 to 32 mm) without any shell damage or only a 10° peel. Extensive peeling probably results in a greater yield of flesh. As a rule, larger predators can kill larger prey. Table 1 shows, however, that this is not always the case with Calappa hepatica. Rhinoclavis fasciata has a critical size (shell length) of about 30 mm for crabs ranging in Carapace width from 33 to 78mm. Several crabs greater than 65 mm in width were un- able to kill R. fasciata as small as 24 mm in length. The 30.2 mm A. fasciata killed by a large (78.2 mm) Calappa would have sur- vived in spite of peeling had a hole not been pierced through the spire. Other R. fasciata offered to this crab ranged from 28.5 to 33.7 mm in length, and survived in spite of being peeled back to the first varix. Peeling alone, therefore, proved to be more effective for small crabs than for large ones, perhaps because the external tooth on the claw was too large for the apertures of many high- spired and coniform shells. Predation on Strombus gibberulus was more conventional; larger crabs were able to TABLE 1. Critical lengths of prey species for several individuals of Calappa hepatica. Prey shell lengths Crab Largest Smallest Prey species (mm) N Range given killed not killed Rhinoclavis aspera 42.6 7 20.7 to 38.1 31.6 20.7 44.0 Y 26.5 to 38.5 30.5 26.5 R. fasciata 33:3 14 18.0 to 33.6 29.8 26.8 41.3 15 23.8 to 29.5 28.0 24.5 62.8 10 23.0 to 35.4 30.0 30.5 67.0 6 24.2 to 32.0 — 24.2 74.9 4 24.3 to 33.1 — 24.3 75.1 9 23.8 to 35.6 — 23.8 78.2 7 28.5 to 33.7 30.2 28.5 Terebra affinis 338 й 16.3 to 32.9 — 16.3 41.3 6 23.8 to 33.6 23.8 24.2 67.0 9 23.8 to 39.0 — 23.8 78.2 Y 18.7 to 35.3 — 18.7 Strombus gibberulus 33:3 i 15.0 to 27.2 — 15.0 41.3 12 16.7 to 34.2 — 1677 54.9 13 27.4 to 40.0 39.0 27.4 67.0 7 27.8 to 37.6 27.8 30.0 74.9 10 18.6 to 46.7 31.0 ЭЙ 75.1 17 22.8 to 42.2 34.2 34.2 18:2 7 29.9 to 46.7 351 29.9 80.2 10 27.4 10 43.6 43.6 29.1 М = Number of ргеу offered. CALAPPA PREDATION ON GASTROPODS 5 kill larger prey, either Бу breaching the lip or by initiating a peel from the anterior end or from a dorsal hole. Substantial individual dif- ferences were evident both in the prey and in the crab predators, so that the correlation be- tween crab width and critical prey length is relatively weak (Table 1). Differences among predators of the same species in the ability to break shelis have been noted previously for other crabs (Zipser & Vermeij, 1978), and may be the rule among shell-breaking preda- tors. Although learning is likely to be important in a crab’s prey-handling behavior, no evidence of its effectiveness was found in this study. | saw no tendency for the critical size of any gastropod species to increase with the dura- tion of the crab’s captivity, nor did crabs inflict progressively less damage to prey that they could not kill. Lethal Breakage—Lethally broken shells are common in soft-bottom environments (Appendix A). Postmortem “lethal” breakage, as inferred from the incidence of breakage in drilled shells, was unimportant in most sam- ples. The frequency of lethal breakage ex- ceeded 0.10 in drilled shells only in one of the 6 samples in which there were at least 10 shells with a drill-hole. No postmortem break- age could be detected in 3 of the 6 samples. No estimates of postmortem breakage could be obtained for samples of Strombus spp. and for many species of Nassarius because of the low incidence of drilling. A relationship between the frequency of lethal breakage and shell architecture was clear at only one site (Venado Beach). Spe- cies with thick adult lips or narrowly elongated apertures had significantly lower frequencies of breakage than did species with thin lips and broad apertures. At the Indo-West-Pacific sites, the very high frequencies of breakage in thick-lipped species of Strombus destroy any relationship between shell form and break- age. Because most dead broken Strombus were recovered as apices, it was difficult to ascertain whether the snails were killed as juveniles or as thick-lipped adults. The pres- ent data on Strombus confirm and extend those reported earlier (Vermeij, 1979). Field estimates of the intensity of breakage- induced mortality were highly correlated with the success rate of Calappa on the same species in the laboratory. Success rates de- creased in the order Strombus gibberulus, Rhinoclavis fasciata, R. aspera, Terebra af- finis. The species were ranked similarly with respect to the incidence of lethal breakage at 4 localities in Guam (Appendix A). Shell Repair—The incidence of breakage- induced shell repair in a sample of shells is influenced by the following factors: (1) the likelinood that a shell will be attacked; (2) the probability that an attack will be successful; and (3) the probability that an unsuccessful attack will result in lip breakage requiring re- pair. The second and third factors depend on shell shape, strength, and size. The first may also be related to shell characteristics, espe- Cially if a shell-breaking predator learns to avoid prey that are of an unsuitable size or shape. The work with Calappa suggests that varices, high spires, and large size may be effective in preventing attacks by Calappa from being successful. A thickened lip not only repels an attacker, but obviates the need for repair if an unsuccessful attack leaves the lip intact. Larger shells should as a rule have higher frequencies of repair than smaller ones because they have been exposed to shell- breaking agents for a longer time (that is, scars accumulate), and because the proba- bility that an attack is successful decreases as the shell grows larger. An analysis of complete samples of intact shells (individuals of all sizes belonging to a given species) shows that repair is less com- mon among species with thickened adult lips than among those in which the lip remains rela- tively thin (Appendix A). This difference is sig- nificant at Wom Village (p < 0.003, Mann- Whitney U-Test) and Pujada Bay (p < 0.05). At Dodinga Bay, the 4 lowest frequencies of repair (out of a total of 7) are those of the 4 thick-lipped samples). The only thick-lipped species sampled in sufficient quantity at Tumon Bay (Strombus gibberulus) ranks third lowest among 9 species. At Venado Beach, there is no significant difference be- tween thick-lipped and thin-lipped species. Comparisons among thin-lipped species re- veal no patterns in the incidence of repair with respect to aperture shape, presence or ab- sence of an umbilicus, or height of spire. Of the 10 samples in which frequencies of repair can be estimated in more than one size class (Table 2), 8 show a rise in repair with increasing length, 1 (Polinices tumidus from Wom Village) shows a constant incidence of 0, and 1 (Rhinoclavis vertagus (L.) from Boear, Aru Islands) shows a mixed pattern. The tendency for repair to increase in fre- quency in larger shells was also found for 6 VERMEIJ TABLE 2. Shell repair in relation to shell length. Frequency of repair in size clsses 5-9 mm 10-19 mm 20-29 mm 30-39 mm Species Loc n f n f n f n f Rhinoclavis aspera Tb 23 .61 AC 12 .083 26 .23 R. fasciata Tb 9 Alk 38 .16 18 .56 Bi 15 0 72 .097 11 091 Cl 48 33 107 .36 67 55 R. vertagus B 8 0 17 .059 74 .19 Cerithium coralium Db 26 27 11 .82 Modulus catenulatus Vb 37 .27 Polinices tumidus Wv 13 0 16 0 P. uber Vb 12 .083 10 .20 Natica chemnitzii Vb 12 58 9 .67 Strombus labiatus Wv 28 .036 Pu 19 13 S. gibberulus Pu 11 0 Pyrene versicolor Db 12 31 Pu 12 50 Strombina clavulus Vb 11 2.00 Nassarius globosus Db 9 .44 Tg 31 .033 N. bicallosus Db 10 .40 25 .96 N. spp. Db 10 .60 N. luridus Tg 11 .091 Wv 22 27 N. distortus pu 10 1.50 N. subspinosus Pu 41 2 N. callospira Wv 10 .10 N. pullus Wv 1S .078 N. quadrasi Wv 13 .23 N. luteostoma Vb 14 .071 N. pagodus Vb 13 .46 N. versicolor Vb 12 .50 Olivella volutella Vb 15 .20 Cancellaria jayana Vb 11 155 Pilsbryspira aterrima Vb 17 47 Gemmula graeffei Db 13 78 Vexillum exasperatum Pu 22 .23 V. spp. Tb 12 33 Imbricaria spp. Pu 11 .27 Conus coronatus Pu И 35 С. arenatus Ри 13 ai C. muriculatus Wv 14 .14 C. ximenes Vb 10 .20 Terebra anilis Wv 10 1.10 T. affinis Tb 13 1.38 T. elata Vb 9 .44 T. spp. Vb 13 1.54 Otopleura mitralis Tb 10 .20 Pyramidella dolabrata Tb 10 .20 Key: CI Cocos Lagoon, Guam, February, 1979 n Number of shells ОБ Dodinga Bay, west coast Halmahera, Indonesia, July, f Frequency of repair 1979 | | Ри Pujada Bay, Mindanao, Philippines, July, 1979 Localities: Tb Tumon Bay, Guam, January to April, 1979 Ac Alupang Cove, Guam, April, 1979 Tg Tagbilaran, Baclayon, Bohol, Philippines, July, 1979 B Boear Island, Aru Islands, Irian Jaya, July, 1979 Vb Venado Beach, Panama, summers of 1976 and 1978 Bi Bangi Island, Ада, Guam, February, 1979 Wv Wom Village, Papua New Guinea, June, 1979 CALAPPA PREDATION ON GASTROPODS 7. TABLE 3. Shell repair, breakage, and attack suc- cess in relation to shell length in Rhinoclavis fasciata from Cocos Lagoon, Guam. Repair Breakage Attack (mm) n f n f success (OOo 14:9 12 .33 12 .58 1.75 15:0'to 199 36 .33 36 .44 1.33 2000 24.9 28 1.43 25 .28 0.58 25 Оо 29979 3359 .092 0.19 3010 t0'34:91 55.911 49 0 0 SOHUNOISI O2 75 12 .083 0.13 Attack success: Number of fatal attacks divided by number of scars on living and dead shells in given size class. n Number of individuals (living and dead) f Frequency of repair or breakage many Recent species of Terebridae (Vermeij et al., 1980) and for Conus sponsalis Hwass (Zipser & Vermeij, in press). Because larger prey are more immune to predation by peeling than are smaller individ- uals, increasing shell length should be asso- ciated with a decline in the incidence of lethal breakage as well as with an increase in the number of repairs per individual. The only field sample that was large enough to test this prediction was that of Rhinoclavis fasciata from Cocos Lagoon, Guam (Table 3). The data show that the frequency of fatal break- age falls steadily as shell length increases. Moreover, the ratio of lethal attacks to total number of attacks also decreases, as ex- pected. Only one shell longer than 30 mm sustained a fatal attack. This result is con- sistent with the laboratory observation on Calappa, which indicated that this crab is un- able to kill R. fasciata greater than 30 mm in length. The frequency of repair in the field is highly correlated with the failure rate of Calappa in the laboratory. Among the 4 species studied in detail at Guam, the incidence of scars in- creases in the order Strombus gibberulus, Rhinoclavis fasciata, R. aspera, Terebra af- finis (Appendix A). This ranking is identical to that for increasing failure rate. There appears to be considerable local intraspecific variation in the frequency of re- pair. This is well illustrated in Guam (Appen- dix A, Table 2), where repair varies by a factor of 6 in such species as Strombus gibberulus and Rhinoclavis fasciata. The possibility that these variations in repair reflect intraspecific differences in the amount of shell-breaking TABLE 4. Spearman rank correlation between fre- quency of repair and frequency of breakage. Group n Correlation Rhinoclavis spp. 7 +0.73 Strombus gibberulus 9 +0.75 Terebra spp. 6 +0.61 Nassarius spp. dit — 0122 n Number of samples mortality is analyzed in Table 4, in which rank correlations between the frequency of fatal breakage and the frequency of repair in com- plete samples are given for several species and genera. There is a significant positive cor- relation at the 0.05 level between repair and lethal breakage in Rhinoclavis, Terebra, and Strombus gibberulus, but in Nassarius the correlation is negative and not significant. DISCUSSION Ecologists have often assumed that the in- cidence of unsuccessful predation (shell re- pair in the present study) is correlated with, and is therefore a measure of, the proportion of an animal population that is killed by dam- age-inducing predators (Shapiro, 1974; Raf- faelli, 1978; Schall 8 Pianka, 1980). Schoener (1979) presented a theoretical analysis of tail- breaks in lizards. He showed that the propor- tion of individuals with injured tails is a meas- ure of predator inefficiency under certain con- ditions (continuous reproduction, size-inde- pendent probability of encounter between prey and agent of mortality, size-independent rate of mortality, injuring agent the only cause of death). When other causes of mortality are introduced, the proportion of injured individu- als in the population may sometimes be cor- related with the intensity of predation, de- pending on the temporal distribution of repro- duction. It is difficult to apply Schoener's theoretical work to breakage in molluscs, because break- age-induced mortality and the probability of encounter with shell-breaking agents are strongly dependent on gastropod prey size (see also Hughes & Elner, 1979; Elner 8 Raffaelli, 1980). Moreover, very little is known about the reproductive schedules of tropical gastropods. Comparisons of my laboratory work on Calappa with field estimates of repair and lethal breakage suggest a very high posi- tive interspecific correlation between the in- 8 VERMEIJ cidence of repair and Calappa's predatory inefficiency, and an equally strong negative interspecific correlation between the fre- quency of lethal breakage and crab failure. At the intraspecific and intrageneric level, how- ever, the frequencies of repair and lethal breakage are often positively correlated (Table 4). Rafaelli (1978) came to a similar conclusion. He found that the frequency of re- pair in Welsh Littorina rudis (Maton) was high in areas where agents of breakage (boulders and the crab Carcinus maenas (L.)) were abundant, and low in marshes and on cliffs where these agents were rare. These results suggest that the relationship between lethal and sublethal damage de- pends not only on the factors considered by Schoener (1979), but also on the relative strengths and abundances of predator and prey. As the strength of predators increases relative to that of the prey, the proportion of successful attacks rises and the frequency of reparable damage falls (Hughes & Elner, 1979; Elner & Raffaelli, 1980). With an in- crease in predator abundance relative to the prey, both the frequency of repair and the fre- quency of lethal damage will increase. The incidence of repair is thus correlated positive- ly with the incidence of lethal breakage only if variation in predation is expressed as varia- tion in the relative abundance of predator and prey. Variations in relative predator strength will have the opposite effect of producing a negative correlation between the incidences of lethal and sublethal damage. Even if a positive intraspecific or intrage- neric correlation between repair and lethal breakage exists, as it does in some gastro- pods, the frequency of repair is not necessar- ily a reliable measure of the intensity of break- age-induced mortality because the relation- ship is not linear. | expect that this situation is common with other types of predation as well, and that attempts by many ecologists to equate nonlethal predation with lethal preda- tion are optimistic or erroneous. Within assemblages of gastropods, the incidence of repair is a reliable guide to the effectiveness of the shell as a protective de- vice against locally prevailing agents of break- age. The high frequencies of repair in many tropical marine gastropods suggests that these species are surprisingly well adapted even against such relatively specialized pre- dators as Calappa (see also Currey & Kohn, 1976; Vermeij et al., 1980). The higher the frequency of repair, the greater is the likeli- hood that selection in favor of traits that pro- tect against lethal breakage is taking place. Although the frequency of repair is not a measure of the magnitude of this selection nor an indication whether selection is stabiliz- ing, directional, or disruptive, the presence of repair is a necessary condition for selection with respect to breakage resistance. A very low incidence of repair means either that shell-breaking agents are rare or that the agents, regardless of their abundance, are usually successful in fatally breaking the shell. A high frequency of breakage-induced scars means that many or most individuals in the gastropod population are exposed to the un- successful attacks of potentially lethal agents of breakage, and that characteristics of the shell (together with other defenses that come into play as the snail is subdued) are effective in protecting the gastropod against fatal breakage. Varices appear to be effective in limiting the extent of peeling. They are characteristic of the shells of many tropical gastropod families, including the Cerithiidae, Potamididae, Cas- sidae, Bursidae, Cymatiidae, and Muricidae (Linsley & Javidpour, 1980). Most other tropi- cal gastropod families are characterized by regularly spaced axial ribs, which may func- tion in the same way as do varices. These structures are common on hard and soft bot- toms alike, probably because sublethal dam- age by both crushing and peeling agents usually affects only the outer lip and the im- mediately adjacent parts of the body whorl. Thick-lipped gastropods, which are also very common in the tropics (many Cerithiidae, Strombidae, Cypraeacea, Tonnacea, Murici- dae, Columbellidae, Buccinidae, Nassariidae, Mitridae), have solved the problems posed by shell-breaking predators in a somewhat dif- ferent or complementary way. Their thin-lip- ped juvenile stages appear to be short com- pared to the thick-lipped phase during which damage to the lip is often limited to superficial scratches or chips (see Randall, 1964; Spight & Lyons, 1974; Yamaguchi, 1977). | have often wondered why crabs such as Calappa so often attack oversized or other- wise inappropriate prey. This behavior is not only typical of most shell-breaking predatory fishes and crabs, but indeed of most preda- tors from copepods to lions. Although a full review of this interesting topic is beyond the scope of this paper, it is worth noting that the inept behavior of most predators may be re- sponsible for the evolutionary acquisition by CALAPPA PREDATION ON GASTROPODS 9 the prey of effective antipredatory adapta- tions. If predators could increase their suc- cess rate and at the same time regulate their populations so as not to overexploit prey, predators could “concentrate” on other adap- tational dilemmas. Of course, it is these dilem- mas that ultimately regulate the predator's population, so that their solution would event- ually lead to the demise of the inept prey. The real world for the predator thus seems to bear little food of the right size or shape, and re- quires the prey to test less suitable items con- tinually. ACKNOWLEDGMENTS | thank Edith Zipser and Elizabeth Dudley for valuable assistance in the laboratory and field. Support and space were kindly provided by the staff of the University of Guam Marine Laboratory and by the scientists and crew on the Morro Expedition aboard the R.V. Alpha Helix. Two anonymous reviewers were ex- ceedingly helpful in improving this paper. The work was supported by grants from the Na- tional Science Foundation, Program of Bio- logical Oceanography. REFERENCES CITED ABRAMS, P. A., 1980, Resource partitioning and interspecific competition in a tropical hermit crab community. Oecologia, 46: 365-379. BERTNESS, M. D., 1980, Shell preference and uti- lization patterns in littoral hermit crabs of the Bay of Panama. Journal of Experimental Marine Bi- ology and Ecology, 48: 1-16. BROWN, S. C., CASSUTO, S. R. & LOOS, R. W., 1979, Biomechanics of chelipeds in some de- capod crustaceans. Journal of Zoology, 188: 153-169. CALDWELL, R. L. 4 DINGLE, H., 1975, Ecology and evolution of agonistic behavior in stomato- pods. Naturwissenschaften, 62: 214-222. CURREY, J. D. & KOHN, A. J., 1976, Fracture in the crossed-lamellar structure of Conus shells. Journal of Material Science, 11: 1615-1623. ELNER, R. W. 8 RAFFAELLI, D. G., 1980, Interac- tions between two marine snails, Littorina rudis Maton and Littorina nigrolineata Gray, a pred- ator, Carcinus maenas (L.), and a parasite, Microphallus similis Jägerskiold. Journal of Ex- perimental Marine Biology and Ecology, 43: 151-160. HAMILTON, S. B. in prep., Shell morphology and predator resistance in fresh-water gastropods. HUGHES, R. N. 8 ELNER, R. W., 1979, Tactics of a predator, Carcinus maenas, and morphologi- cal responses of the prey Nucella lapillus. Journal of Animal Ecology, 48: 65-78. KITCHING, J. A., MUNTZ, L. 8 EBLING, F. J., 1966. The ecology of Lough Ine. XV. The eco- logical significance of shell and body forms in Nucella. Journal of Animal Ecology, 35: 113- 126. LIEM, K. F., 1973, Evolutionary strategies and morphological innovations: cichlid pharyngeal jaws. Systematic Zoology, 22: 425-441. LINSLEY, R. M. & JAVIDPOUR, M. D., 1980, Epi- sodic growth in Gastropoda. Malacologia, 20: 153-160. MILLER, B. A., 1975, The biology of Terebra gouldi Deshayes, 1859, and a discussion of life history similarities among other terebrids of similar pro- boscis type. Pacific Science, 29: 227-241. PALMER, A. R., 1979, Fish predation and the evo- lution of gastropod shell sculpture; experimental and geographic evidence. Evolution, 33: 697- Walls: PAPP, A., ZAPFE, H., BACHMAYER, F. & TAUBER, A. F., 1947, Lebensspuren mariner Krebse. Konigliche Akademie der Wissenschaf- ten (Wien), Mathematische und Naturwissen- schaftliche Klasse, Sitzberichte, 155: 281-317. RAFFAELLI, D. G., 1978, The relationship between shell injuries, shell thickness and habitat char- acteristics of the intertidal snail Littorina rudis Maton. Journal of Molluscan Studies, 44: 166- 170. RANDALL, J. E., 1964, Contribution to the biology of the queen conch, Strombus gigas. Bulletin of Marine Science of the Gulf and Caribbean, 14: 246-295. ROBBA, E. & OSTINELLI, F., 1975, Studi paleo- ecologici sul Pliocene Ligure |. Testimonianze di predazione sui molluschi Pliocenici di Albenga. Rivista italiana Paleontologica, 81: 309-372. SCHALL, J. J. & PIANKA, E. R., 1980, Evolution of escape behavior diversity. American Naturalist, 115: 551-566. SCHOENER, T. W., 1979, Inferring the properties of predation and other injury-producing agents from injury frequencies. Ecology, 60: 1110- 11515: SHAPIRO, A. M., 1974, Beak-mark frequency as an index of seasonal predation on common but- terflies. American Naturalist, 108: 229-232. SHOUP, J. B., 1968, Shell opening by crabs of the genus Calappa. Science, 160: 887-888. SPIGHT, T. M. 8 LYONS, A., 1974, Development and functions of the shell sculpture of the marine snail Ceratostoma foliatum. Marine Biology, 24: 77-83. VERMElJ, С. J., 1976, Interoceanic differences in vulnerability of shelled prey to crab predation. Nature, 260: 135-136. VERMElJ, С. J., 1977, The Mesozoic marine revo- lution: evidence from snails, predators and grazers. Paleobiology, 3: 245-258. VERMElJ, С. J., 1978, Biogeography and adapta- 10 VERMEW tion: patterns of marine life. Harvard University Press, Cambridge: 332 p. VERMEIJ, С. J., 1979, Shell architecture and causes of death in Micronesian reef snails. Evo- lution, 33: 686-696. VERMElJ, С. J. & CURREY, J. D., 1980, Geo- graphical variation in the strength of thaidid snail shells. Biological Bulletin, 158: 383-389. VERMElJ, С. J., ZIPSER, E. 4 DUDLEY, E. C., 1980, Predation in time and space: peeling and drilling in terebrid gastropods. Paleobiology, 6: 352-364. YAMAGUCHI, M., 1977, Shell growth and mortality rates in the coral reef gastropod Cerithium nodulosum in Pago Bay, Guam, Mariana Islands. Marine Biology, 44: 249-263. YAMAOKA, K., 1978, Pharyngeal jaw structure in labrid fish. Publications of the Seto Marine Bi- ological Laboratory. 14: 409-426. ZIPSER, E. & VERMEIJ, С. J., 1978, Crushing be- havior of tropical and temperate crabs. Journal of Experimental Marine Biology and Ecology, 31: 155-172. ZIPSER, E. & VERMEIJ, G. J., 1980, Survival after non-lethal shell damage in the gastropod Conus sponsalis. Micronesica, 16: 229-234. APPENDIX A Frequencies of repair and breakage in some soft-bottom gastropods from the tropical Pacific. Locality and species Tumon Bay, Guam Rhinoclavis aspera R. fasciata DE Strombus gibberulus Otopleura nodicincta (A. Adams) O. mitralis (A. Adams) Pyramidella dolabrata (L.) E Vexillum spp. E Conus pulicarius Hwass Terebra affinis Cocos Lagoon, Guam Rhinoclavis fasciata Y Strombus gibberulus E Conus pulicarius Terebra affinis Bangi Island, Guam Rhinoclavis fasciata if Strombus gibberulus Alupang Cove, Guam Pyramidella dolabrata Rhinoclavis aspera Y Strombus gibberulus Pago River, Guam Rhinoclavis aspera Terebra affinis Wom Village, New Guinea Polinices tumidus Strombus canarium L. S. gibberulus S. labiatus (Roding) S. luhuanus L. 3444 Repair Mortality n f n na ‘Mea fo 31 .45 15 Ors AO eZ 67 .26 25 2 1 0044 24 25 53 0 О 268 13 .62 13 .23 14 23 15 .40 11 .18 25 5.1.32 222 .42 208 96 8 18) 17 .24 19 0 0 .84 16 .063 13 .92 98 .082 75 : 12 .083 11 0 0 .81 44 .16 24 4 0 .083 30 .033 120 0 0 e 44 .16 45 22 0 .16 53 .38 53 10 1 .070 27 #0 17 .059 12 0 0 .50 15 .067 34 .033 35 0 0 -55 15 .067 mm — MMM A445 m ++ ++ M A CALAPPA PREDATION ON GASTROPODS APPENDIX A (Continued) Locality and species Nassarius callospira (A. Adams) N. luridus Gould N. pullus (L.) N. quadrasi Hidalgo Vexillum spp. Conus muriculatus Sow. Terebra anilis (Roding) Bulla ampulla L. Pujada Bay, Mindanao Strombus gibberulus S. labiatus S. urceus L. Pyrene versicolor Sow. Nassarius distortus (A. Adams) N. subspinosus (Lamarck) Vexillum exasperatum (Gmelin) Imbricaria spp. Conus coronatus Hwass C. arenatus Hwass Tagbilaran, Bohol Strombus labiatus S. urceus Nassarius globosus (Quoy & Gaimard) N. luridus (Gould) Conus coronatus Dodinga Bay, Halmahera Cerithium coralium Kiener Nassarius bicallosus (E. A. Smith) N. globosus (Quoy & Gaimard) N. spp. Vexillum spp. Pyrene versicolor Sow. Gemmula graeffei (Weink.) Motupore, New Guinea Strombus gibberulus Majuro Lagoon, Majuro Atoll Strombus gibberulus Boear, Aru Islands, Irian Jaya Polinices tumidus Rhinoclavis vertagus Beaufort, North Carolina Terebra dislocata (Say) Venado Beach, Panama Polinices uber (Valenciennes) Natica chemnitzii Pfeiffer Modulus catenulatus (Philippi) Strombina clavulus Northia pristis (Deshayes) Nassarius dentifer Powys Repair n f 10 .10 24 29 13 .078 14 21 11 45 14 .14 10 NI 10 12 .083 23220 27 Sk 11 .091 12 .50 MESS 42 .26 22 23 11 27 12 135 18 .28 qa .091 30) 0 32 .059 11 .091 13540 37 41 35 75 9 .44 16 47 11 .81 13 3 15 ИЗ 4370 14 0 10 .10 148 .16 120 .083 27 13 26 .50 38 .26 138223 10 .40 20 55 61 43 10 114 130 27 Mortality Па Па 0 0 3 0 2 0 0 0 0 0 0 0 2 0 7 0 0 0 1 0 10 0 8 0 0 0 0 0 0 0 23 0 13 3 .095 .064 12 27 .58 .31 .40 .46 .30 STA .031 mm m VERMEIJ APPENDIX A (Continued) Locality and species N. luteostoma (Brod. & Sow.) N. pagodus (Reeve) N. versicolor (C. B. Adams) Agaronia testacea (Lamarck) Olivella volutella (Lamarck) Cancellaria jayana Keen Prunum sapotilla (Hinds) Conus ximenes Gray C. patricius Hinds Pilsbryspira aterrima (Sowerby) Terebra elata Hinds T. strigata Sowerby T. spp. Farfan Beach, Panama Terebra cracilenta Li Number of individuals Frequency Number of drilled individuals Number of drilled shells with lethal breakage Corrected frequency of lethal breakage thick lip elongated aperture Repair n f 21 .19 16 .67 42 .19 19 .16 17 .24 13 .62 1174 =18 13 23 17 74174 10 .40 10 1.00 14 METÍ 49 .29 12 55 io Salil MALACOLOGIA, 1982, 23(1): 13-21 DIURNAL MOVEMENTS IN POPULATIONS OF THE PROSOBRANCH LANISTES NYASSANUS AT CAPE MACLEAR, LAKE MALAWI, AFRICA Svata M. Louda and Kenneth R. McKaye Duke University Marine Laboratory, Pivers Island, Beaufort, NC 28516, U.S.A. ABSTRACT Diurnal variation in abundance of Lanistes nyassanus Dohrn (Prosobranchia: Ampullariidae) has been hypothesized to be due to directed daily movements up and down the shore at Cape Maclear, Lake Malawi. We tested this hypothesis and collected data on population density, dispersion, size distribution, and growth rates of L. nyassanus. Using SCUBA and following marked snails, we found the following. (1) No directed displacement up and down the slope occurs, but the number of snails buried in the sand is significantly higher in the morning than the afternoon. Most movement was crepuscular or nocturnal and averaged 2.8 m each day. Neither distance displaced nor direction of displacement suggest high vagility. (2) Growth, which aver- aged 1.0 cm/ring for adult snails, appears to be annual. The increment was correlated with increase in water temperature. (3) The life history strategy appears mixed, with a trade-off of predator defense for high growth rate in thin-shelled young and vice versa in adults. The esti- mated life span is between 5 and 10 years. (4) Fish predation may contribute to explaining: a) the characteristic morphological traits such as shell weight and relative dimensions; b) the skewed size distribution, which is composed of few thin-shelled juveniles compared to heavy-shelled adults; and c) the burying behavior of the snails. Key words: Lanistes; Ampullariidae; Lake Malawi; diurnal movements; vagility; growth rates; fish predation; life history strategy. INTRODUCTION Data on the movements of individuals are Critical to testing several major ecological and evolutionary hypotheses. Kozhov (1963), for example, suggests low vagility is a major fac- tor in speciation among benthic invertebrates, including gastropods, in Lake Baikal. How- ever, the mark and recapture data required to assess vagility and to test Kozhov's hypothe- sis are rare. Prosobranch gastropods in the deep lakes of the rift Valley in Africa are a particularly in- teresting group. The group has radiated in tropical as well as temperate deep lakes; however, tropical regions are richer than tem- perate areas in species of prosobranchs (G. E. Hutchinson, in press). For example, 26-30 of the 41 prosobranch species (63-73%) in Lake Tanganyika are endemic (Boss, 1978). Prosobranchs in Lake Malawi show a similar pattern. At least 15 of the 19 species reported (79%) are endemic (Mandahl-Barth, 1972). Little information, however, exists on the natu- ral history, distribution, movement or dynam- ics of these gastropods (Livingstone, 1981; D. H. Eccles, personal communication). (13) The available information suggests several testable ecological hypotheses with evolu- tionary implications. For example, daily varia- tions in relative abundance of the largest en- demic gastropod, Lanistes nyassanus Dohrn (Prosobranchia: Ampullariidae), are observed (Gray, 1980; С. T. and 1. Grace, personal communication; personal observation). To explain the observed daily variations in abundance and density, Gray (1980) hy- pothesizes that L. nyassanus exhibits a daily vertical movement up and down the littoral slope. Our purpose was: (1) to gather basic eco- logical information for Lanistes nyassanus at Cape Maclear; (2) to test Gray’s (1980) hy- pothesis of daily movement, and (3) to pro- vide an initial estimate of population vagility for this gastropod. LANISTES MONTFORT The Ampullariidae (=Pilidae) are medium, globose snails characterized by a taenioglos- sate radula (>10 mm) and a concentric oper- culum (World Health Organization, 1977). 14 LOUDA & MCKAYE The pallial cavity is divided into two compart- ments: the left one with a ctenidium and the other with a pulmonary sac (Mandahl-Barth, 1972; World Health Organization, 1977). Dur- ing siphonal respiration the left nuchal lobe is drawn out into an incomplete tube (McClary, 1964). Females are oviparous. Males bear a copulatory organ (verge) near the mantle edge. Ampullariids are distributed in freshwater throughout the warmer regions of the world and are represented by two genera (Pila, Lanistes) in southeast Africa (Mandahl-Barth, 1972; World Health Organization, 1977). The genus Lanistes Montfort is confined to Africa. Three species, Lanistes nyassanus, L. ellip- ticus and L. solidus, occur in shallow areas around Cape Maclear (Fig. 1), Lake Malawi (Gray, personal communication). Lanistes nyassanus Dohrn, endemic to Lake Malawi, can be distinguished from its most similar Cape Maclear . Blantyre FIG. 1. Map of Malawi with the Cape Maclear re- gion (Inset A), showing the location of our study site (a) at the Fisheries Research Station (FRS) and of the sampling site (b) for temperature profile. North is at the top of the map. congener, the deeper-occurring L. nasutus Mandahl-Barth, by being larger (up to 65 mm compared to 40 mm), heavier, and with a closed umbilicus. Young L. nyassanus (up to about 30 mm) have thin shells with a narrow umbilicus (World Health Organization, 1977). Crowley et al. (1964) suggest that there are two distinct habitat types for the lake-dwelling gastropods in Lake Malawi: permanent marshes with Lanistes ellipticus and L. ovum procerus and lake edge with Lanistes solidus and Lanistes nyassanus (see also Cantrell, 1979). These species all have heavy shells. Lanistes nyassanus is common in the sandy shallow lake edge habitat and the Potamoge- ton-Vallisneria beds of Cape Maclear. No field data apparently exist on feeding and nutrition of Lanistes nyassanus. How- ever, Yonge (1938) suggested that close rela- tives, Pila and nonendemic species of Lanis- tes from Lake Tanganyika, rasp algae from solid surfaces; both have a massive radula with short stout teeth. In Lake Tanganyika these species tend to live on higher plants in estuaries and inlets (Leloup, 1953). At Cape Maclear, Lake Malawi, L. nyassanus often faces into the current, with its foot and siphon partially extended and covered with copious mucus. Observations on another ampullariid genus, Pomacea, suggest this is a feeding posture. Pomacea can evidently feed on pro- tein absorbed on minute particles of organic matter (Johnson, 1952; Cheesman, 1956; McClary, 1964). The material probably in- cludes bacteria and algae (Hutchinson, in press). Ciliary feeding is not unknown among gastropods but is relatively rare among the Ampullaridae (G. E. Hutchinson, personal communication). METHODS Our observations were obtained on the sand beach adjacent to the Fisheries Re- search Station at Cape Maclear (34°50’E, 14°5'S), 12km W of Monkey Bay, Lake Malawi (Fig. 1), in January 1980, using SCUBA. These data are of three kinds: distri- bution and density, size and growth, and movement. Surface and 5m depth water temperatures, which peak in January-Febru- ary, are available for July 1977 to April 1980 from a nearby sampling station. The density and distribution of snails were determined in two ways. First, a 30 m transect line was anchored along the 1.5 m depth con- POPULATIONS OF LANISTES IN LAKE MALAWI 1S tour. The number of snails per m” on both sides of the line was recorded regularly at 24 hour intervals and occasionally at intermedi- ate intervals. Second, a 128 m? permanent grid of 4 rows of 8 quadrats (2 x 2 m each) was created. The rows were also along the slope contour. The first two rows (at 3.25 and at 3.75 m depth respectively) were on open sand. The third and fourth rows (at 4.25m and at 4.75 m respectively) were in a Pota- mogeton-Vallisneria bed. We recorded the number of snails in each quadrat on 20 Jan. 1980. We collected snails by swimming horizontal and vertical transects along the shore and col- lecting every third snail up to 50. Maximum shell height and width were measured (Fig. 2A). We also measured the aperture as an estimate of size since it frequently approxi- mates gastropod biomass more closely than total overall shell dimension, especially if A FIG. 2. Diagram of the measurements taken on shells of Lanistes nyassanus at Cape Maclear, Lake Malawi. A shows shell dimension where: Ly = total length, or overall height: Wy = total or body whorl width: Lz = aperture length or height: and, Wa = aperture width. В shows growth increment measurements on a transverse section where: pP = pre-penultimate ring, P = penultimate ring, and U = ultimate ring or growth increment. apical wear is common (Fotheringham, 1971; Louda, 1979). Growth increments were curvi- linear distance from aperature edge to first growth line, from first to second oldest growth line and, if present, from second to third oldest growth line (Fig. 2B). These were measured at their widest diameter. After the snails were measured, we marked and returned them within two hours to a cen- tral stake in a 100 m* movement grid at 2.5 m water depth on sand. Snails were marked by gluing a number on the top of the body whorl, epoxying a 10 cm piece of white yarn to the surface between the first and second whorls and putting red enamel paint along the edge of the aperture lip. The yarn allowed detection of buried snails. Recapture success was high (95%) so treatment did not affect survivorship. Specimens of Lanistes nyassanus have been deposited in the Peabody Museum, Yale Uni- versity. The snails’ linear displacement was meas- ured at 4 hr or at 24 hr from the position of last sighting. Each numbered snail's position was marked after every time interval by a num- bered flag. The distance between sightings was measured and the flag moved to the new snail position. Displacement from the new position was recorded at the next interval. Four consecutive days of observations were run 16-20 January 1980. Snail activity and position in or on the sand were recorded. RESULTS Lanistes nyassanus has been reported at depths from 12 to 28 m (World Health Organi- zation, 1977). This distribution may now be extended up into much shallower water since we were able to study substantial numbers of this species on sand and in adjacent Pota- mogeton-Vallisneria beds at depths of 1 to 5m at Cape Maclear (Fig. 1). For L. nyas- sanus in shallow water at the high point of the annual cycle of water temperature (McKaye, 1982), the interface of sand and weeds at 3.5m depth supported the highest densities (Table 1). In the permanent quadrats on sand at 3.5 m, density was 1.3/m” (N = 64 m’). The lowest densities of L. nyassanus were in quadrats in the deeper (4.5 m) Potamogeton- Vallisneria bed (0.4/m”, N = 64 m’). Density on the transect higher on the shore (1.5 m depth) on sand was intermediate (0.95/m”, N = 60 m”). Although densities in the morning (1100 hr) appeared to be lower than in the 16 LOUDA & MCKAYE TABLE 1. Density of Lanistes nyassanus (Number per m”) at Cape Maclear, Lake Malawi, in January, 1980. Time of Day N X SE. 95% C.l. Line Transect at 1.5 m depth, 30 x 2m 1100 60 0.9 0.13 0.64-1.16 1100 60 0.9 0.15 ~ 0:62-1:22 1800 60 1.0 1.10. .0.82-1.22 Permanent Grid at 3 to 5 m depth,! each row 8 x 2 x 2m quadrats Above Weeds 64 163 0.20 In Weed Bed 64 0.4 0.07 0.88-1.68 0.21-0.49 !Rows 1 and 2, above the weeds, were centered at 3.25 m and at 3.75 m depth respectively. Rows 3 and 4, in the Potamogeton-Vallisneria weed bed, were centered at 4.25 and at 4.75 m depth respectively. early evening (1800 hr) as originally sug- gested, the difference was not significant (Table 1). The heavy-shelled Lanistes nyassanus were smaller (Table 2: 4.2cm) than sug- gested usual (6.1-7.0 ст: World Health Organization, 1977). The largest individual we observed had a shell 5.1 cm high. The hy- pothesis that L. nyassanus is a continuously breeding population with the expected normal size distribution must be rejected. The size distribution was skewed toward large, adult snails at this time of year (Fig. 3). The small- est individual positively identified as L. nyas- sanus was 3.1 cm high. In addition, 3 thinner- shelled individuals, which measured 1.0 and 2.0 cm high and were most likely L. nyas- No. 2 3 a ет, SZ ECM) FIG. 3. Size distribution of marked Lanistes nyas- sanus Dohrn at Cape Maclear, Lake Malawi; fre- quency plotted at midpoints of 0.5 cm size classes, where Lr is the overall shell height. sanus juveniles, were found among the roots of the Potamogeton-Vallisneria bed at 5 m. Penultimate growth increment was not cor- related with shell size, either overall or aper- ture height (r = —0.03 and 0.02, respectively, N = 35 for each), eliminating the hypothesis that growth was a function of individual size. The 5 smallest adult snails (X = 3.5 ст height, S.E. = 0.11), in addition, had a penul- timate increment of 1.2cm (S.E. = 0.31) which was statistically equal to the penulti- mate increment of 1.3 ст (S.E. = 0.28) for the 5 largest snails (X = 4.8 cm, S.E. = 0.10). Thus adult growth was independent of overall adult shell size. There was a marginally sig- nificant correlation (r = 0.41, 0.10>p>0.05) between the penultimate growth ring and the pre-penultimate growth ring for the 18 heavy- shelled L. nyassanus with a third growth line (see Fig. 2), lending support to the hypothesis that individuals may vary in their growth rates. TABLE 2. Size and growth parameters for Lanistes nyassanus and the relationships among them (N = 37 heavy-shelled adult snails) at Cape Maclear, Lake Malawi, in January 1980. N x Maximum height 37 4.2 Maximum width 37 4.4 Aperture height ЭЙ 3.5 Aperture width 37 2.6 Growth increment Ultimate (Recent) 37 0.5 Penultimate 37 1.0 Pre-Penultimate 20 1.0 SIE: Correlation Coefficients 0.03 POPULATIONS OF LANISTES IN LAKE MALAWI 17 YN MD O E A x 0° > er: + © — De — a = (6) z = Ооо FIG. 4. Movement data for the third day о the consecutive 24-hour observations, 18-19 January 1980, taken at 1630 hr, showing original position (@), end position after 24 hours (@), and direction of displacement (>) for numbered Lanistes nyassanus on the movement grid at Cape Maclear, Lake Malawi. Buried individuals were circled. The dashed line to the right is the upper edge of the Vallisneria bed at 3m depth and the dashed line to the left represents the position of the 30 m transect line at 1.5 m depth. Movement, as estimated by linear dis- placement over a 4 hr period (Fig. 4), was greater in the late afternoon-early evening than in the morning or early afternoon (Fig. 5). In January, the average distance displaced from the previous sighting in 4 hr increased as light intensity increased and then decreased from 0830-1230, to 1230-1630 and to 1630- 2030 (with dusk by 1730 and dark by 1830 at 3m depth). In addition, the proportion of marked snails which were buried decreased from morning to evening (Fig. 5). Burial into the top layer of sand occurred at the time of day when snail density tended to decrease (Table 1) and less movement occurred (Fig. 5). Diurnal burial is an alternative hypothesis to explain the pattern of lower densities of snails in the morning than in late afternoon. Daily displacement over 4 days in January averaged 2.8 m/day (Fig. 6). Snail displace- ment over the first 24 hr period, which was (m) poling 0% Distance am pm eve FIG. 5. Distance displaced (6) and proportion of marked Lanistes nyassanus which were buried (6) after 4 hr, by time of day, in January 1980; am = 0830 to 1230 hr, pm = 1230 to 1630 hr, and eve = 1630 to 2030 hr, data plotted at midpoint of obser- vation period. 18 LOUDA & MCKAYE N= 35 32 29 32 11] (m) DISTANCE 1234 DAY FIG. 6. Mean distance displaced over 24 hr, on 4 consecutive days of observation on marked Lanis- tes nyassanus recorded at 1630 hr, 16-20 January 1980, at Cape Maclear, Lake Malawi: N = the number of snails sighted among the 37 marked in- dividuals available. hypothesized to represent a period of extra movement in response to either disturbance or concentration of snails at the central stake, did not differ from that on subsequent days (Fig. 6). The average distance that marked adult snails were displaced daily was remark- ably consistent (Fig. 6). Movement was random in direction. When the direction of a second displacement is compared to the direction of the first dis- placement no pattern emerges (Fig. 7). The one pattern discernable in any of the daily movement data was downward orientation in the first 24 hr period, i.e. after handling (Table 3). These data were only marginally signifi- cant (x = 3.46, df = 1, 0.10>p>0.05). Gray's (1980) hypothesis of general directed movement up and down the slope must be rejected. A related alternative hypothesis, that indi- vidual snails show patterns of movement, must also be rejected. A positive association of consecutive displacements would support NW W 2 DIRECTION NNE E SE DIRECTION 1 FIG. 7. Direction of consecutive 24 hr displace- ments per individual using all pairs of displace- ments for marked Lanistes nyassanus in January, 1980, at Cape Maclear, Lake Malawi. the hypothesis that individual snails were consistent either in moving or in sitting. A negative association would imply a long move is generally followed by a short one and vice versa. No correlation exists (Fig. 8). In addi- tion, there was no correlation (r = —0.13, N = 33) between individual adult snail shell size and mean distance displaced per day over the 4 days of observations (Fig. 9). These data support the hypothesis that large and medium snails move equally. Although the five small- est marked snails (X = 3.5 cm) had farther daily displacements (X = 3.1 m, S.E. = 0.89) than did the five largest (X = 2.3m, S.E. = 23), this difference was not significant (t = 0.87, 4 d.f.). DISCUSSION Information on the density of these snails is of interest for at least two reasons. First, the data provide an estimate of numbers and dis- TABLE 3. Direction of linear displacement by Lanistes nyassanus over 24 hr observation periods, at 1630 hr on 16-20 January 1980, at Cape Maclear, Lake Malawi. Downslope quadrants Upslope quadrants Time N (Hrs) (Snails) NW NE Total SE SW Total One-way x” A ae A See OS AAA RO 24 35 15 29 4 12 3.46 POPULATIONS OF LANISTES IN LAKE MALAWI 19 DISTANCE DISTANCE 1 (m) FIG. 8. Distance of consecutive 24 hr deplace- ments, using all pairs of displacements for marked Lanistes nyassanus in January, 1980, at Cape Maclear, Lake Malawi. tribution of this African prosobranch endemic to Lake Malawi. Second, shells of Lanistes nyassanus are an important spatial resource for an endemic, shell-dwelling cichlid fish and play an indirect role in the community dy- namics of the sand-dwelling fishes of Lake Malawi (McKaye, 1978). One factor, which is also of interest in un- derstanding the ecology of Lanistes nyas- Sanus, is the rate at which shells become available through patterns of L. nyassanus survivorship and mortality. If snail growth oc- curs during the period of warm water, as is suggested by an ultimate growth increment equal to half the penultimate increment at the middle of the warm water period, turnover time is on the order of 5-10 yr. The variation in this rough estimate is dependent on the rate of juvenile development to 3.0 cm size. Adult life appears to average from 3 to 6 growth periods minimum; 55% of our sample had four or more growth rings, probably representing 5-9 adult years, and the mean was 3.3 distinct rings per snail. Overall size and relative shell dimensions suggest Lanistes nyassanus is morphologi- cally adapted for disturbed habitats such as exposed open coasts of a large lake (G. E. Hutchinson, in press). First, it is large and heavy-shelled. Second, the short spire (Burch, 1968), and broad shell (Cain, 1977), with a short height compared to shell width er N = 128 Е Ss) ш 4 о . a 23: ER z г. |: E (a) 8 yA — 3.0 4.0 5.0 en: SIZE (cm) FIG. 9. Average distance displaced over 24 hr for marked Lanistes nyassanus by size category in January, 1980, at Cape Maclear, Lake Malawi, where ® = mean values for 0.2 cm size classes and ® = average (with 95% С.1.) for snails less than, equal to, or greater than the mean size of the adult population. (H/W = 0.95), and a low ratio of height to aperture (1.20), are typical of exposed area snails. Third, the relatively large (3.5 cm) and especially broad (2.7 cm) aperture (W/H = 0.74) is comparable to measurement of an- other gastropod on exposed shores, Gonio- basis livescens (Wiebe, 1926). Longterm, persistent fish predation pres- sure may be a contributing factor to the heavy-shelled morphology of Lanistes nyas- sanus and to the occurrence of shallow-water ‘thalassoid’ endemic gastropods in deep, an- cient lakes like Lake Tanganyika. A large and heavy marine-like shell, especially with a closed umbilicus, is an effective antipredator defense (Wright et a/., 1967; Vermeij, 1975: Vermeij & Covich, 1978; Calow, 1978). The heaviest freshwater gastropods occur in rela- tively shallow water in the ancient lakes; G. E. Hutchinson (in press) cites several examples and suggests that these are K-selected (MacArthur & Wilson, 1967). In addition, they may represent adaptation over a long period of coexistence to intense fish predation. The heavier shells of the shallower-water species of Lanistes, especially associated with the conspicuousness of L. nyassanus on the sand in shallow water, suggest that predation by molluscivorous, sand-dwelling cichlid fish may have been a significant selective pres- sure. This interpretation is further supported by the skewed size distribution of L. nyas- Sanus and by the occurrence of the thin- shelled juveniles in a more heterogeneous and protected microhabitat, the roots of the weed bed. 20 LOUDA & MCKAYE Freshwater snails, even iteroparous spe- cies, have a tendency toward relatively rapid growth and early reproduction when com- pared with marine species. This life history strategy is usually accompanied by a thin shell relative to marine gastropods (Vermeij, 1975; Calow, 1978). There may, thus, be a tradeoff for rapid development over predator defense. This hypothesis is difficult to assess with respect to Lanistes nyassanus but it is consistent with our data. Juvenile L. nyas- sanus appear to grow rapidly. A search, even in protected microhabitats, produced only a few relatively large (1.0-2.0 cm height) juven- iles at the middle of the annual cycle. Adult growth increments are less than those estimated for the young. The adult ratio of the mean penultimate growth ring to aver- age body whorl diameter is 0.23, about one- fifth. Penultimate and pre-penultimate incre- ments each represent approximately 10% of the shell circumference of the adult body whorl. The smallest adult (3.1 ст) had a 6.1 cm circumference up to the first growth line. Thus, the annual increment for the juven- ile phase, depending on the length of its ju- venile life, must have been: 6.1 cm/yr if the juvenile period was 1 year, 3.0 cm/yr if the juvenile period was 2 yr, or 2.0 cm/yr if the first growth phase was 3 yr. In any case, a 2.0-6.1 cm growth increment represents more rapid growth than the 1.0cm mean growth increment of adult snails. Lanistes nyassanus, thus, may present a mixed life history strategy: a juvenile period with excep- tionally rapid growth, during which suscepti- bility to predation is extremely high, and an adult period with slower growth during which predation mortality is decreased by the devel- opment of a heavy shell. Consequently we hypothesize that habitat heterogeneity is criti- cal for early survival and subsequent recruit- ment of L. nyassanus. The movement data were originally gath- ered to test Gray’s (1980) hypothesis that L. nyassanus moves into deeper water during the early hours and into shallower water in the afternoon and evening. No directed displace- ment was observed over 24 hr periods. Direc- tion of displacement over 4 hr at different times of day was also random. Gray’s (1980) hypothesis must be rejected for our study. Density, which averaged 0.95 snails/m” at 1.5m depth, did not vary significantly be- tween morning, afternoon and evening peri- ods. The original observations of fewer snails in the morning than in the afternoon, are best explained by the burial of adult snails into the top layer of sand. Burial was significantly higher during the morning than during the later afternoon or evening. Since burial is shallow, meter by meter counts by a SCUBA diver include buried snails. Yet, these individ- uals would not be recorded by an observor near the surface. Therefore the diurnal pat- tern does exist, but it is into and out of the sand rather than up and down the slope. We suggest that these snails bury to escape from visual predators. The consistency of the average distance displaced per day, the random orientation and the persistence of tagged individuals in one general 100-150 m” area provide the main components of an estimate of vagility hypoth- esized by Kozhov (1963) to be significant in gastropod speciation. There is no evidence of high vagility in our data. The distance dis- placed at this time of year averaged only 3 m and was nondirectional. Movement during the breeding season, which we expect to occur in August-September as the water begins to warm, needs to be examined. Dependence of juvenile survival on habitat heterogeneity should provide strong selection for directed movementtowardthe Potamogeton-Vallisneria beds for mating and egg-laying. Our observa- tions also suggest vagility between the adja- cent habitats, ¡.e. weeds and sand occur at a relatively high rate and in random sequence. The lack of directionality ofthe displacement observed is particularly interesting in light of the morphological capability of the family for air breathing; they have the option of branchial respiration via utilization of the pulmonary sac. Some, such as Pila species, are known to leave the water at night to forage (Prashad, 1925). We did find increased movement in the late afternoon and evening. We did not find an upslope, landward directional displacement as would be expected if the hypothesis that these snails move onto land at night to feed is cor- rect. ACKNOWLEDGEMENTS We thank all who contributed to this study. In particular W. N. Gray challenged us with his observations and interpretation. С. T. and 1. Grace encouraged us with their discussions and hospitality. G. E. Hutchinson helped us put our observations into perspective with his insight, experience and comments which he generously shared. P. N. Reinthal and D. S. Brown improved the manuscript. The Malawi POPULATIONS OF LANISTES IN LAKE MALAWI 21 Government gave permission to study the aquatic ecosystem at Cape Maclear and we appreciate the opportunity greatly. K. R. McKaye was supported by N.S.F. grant DEB- 7912338. LITERATURE CITED BOSS, K. J., 1978, On the evolution of gastropods in ancient lakes. /n: FRETTER, V. & PEAKE, J., eds., Pulmonates. Volume 2A: Systematics, Evolution and Ecology. Academic Press, Lon- don, p. 385-428. BURCH, J. B., 1968, Cytotaxonomy of some Japanese Semisulcospira (Streptoneura: Pleu- roceridae). Journal de Conchyliologie, 107: 1-51. CAIN, A. J., 1977, Variation in the spire index of some coiled gastropod shells, and its evolution- ary significance. Philosophical Transactions of the Royal Society of London, Ser. B, 277: 377- 428. CALOW, P., 1978, The evolution of life-cycle strat- egies in freshwater gastropods. Malacologia, 17: 351-364. CANTRELL, N. A., 1979, Invertebrate communities in the Lake Chilwa swamp in years of high level. In: KALK, M., MCLACHLIN, A. J. & HOWARD- WILLIAMS, C., eds., Lake Chilwa. Junk, The Hague, Netherlands, p. 161-173. CHEESMAN, D. F., 1956, The snail's foot as a Langmuir Trough. Nature, 178: 987-988. CROWLEY, T. E., РАМ, T. & WOODWARD, Е. R., 1964, A monographic review of the Mollusca of Lake Nyasa. Musée Royal de l'Afrique Centrale (Tervuren, Belgique). Annales, Serie 8, Sci- ences Zoologiques, 131: 58 p., 7 pl. FOTHERINGHAM, N., 1971, Life history patterns of the littoral gastropods Shaskyus festivus (Hinds) and Ocenebra poulsoni Carpenter (Prosobran- chia: Muricidae). Ecology, 52: 742-757. GRAY, W. М., 1980, Some unusual snails of Lake Malawi. Nyala, 5(2): 19-28. HUTCHINSON, G. E., in press, Gastropod mol- luscs of the littoral benthos. A Treatise on Lim- nology, Volume IV, Wiley, New York, 183 p. in manuscript. JOHNSON, B. M., 1952, Ciliary feeding in Poma- cea paludosa. Nautilus, 66: 1-5. KOZHOV, M., 1963, Lake Baikal and its Life. Vol. 11, Monographiae Biologicae, Junk, The Hague, Netherlands, 344 p. LELOUP, E., 1953, Gasteropodes. Exploration hy- drobiologiques du Lac Tanganyika (1946- 1947): Résultats Scientifiques (Institut Royal des Sciences Naturelles de Belgique), 3(4): 1-273, 13 pl. LIVINGSTONE, D. A., convenor, 1981, Paleolim- nology. In: SYMOENS, J. J., BURGIS, M. & GAUDET, J. J., eds., The Ecology and Utilization of African Inland Waters. United Nations Envi- ronmental Programme, Nairobi, p. 176-182. LOUDA, S. M., 1979, Distribution, movement and diet of Searlesia dira (Gastropoda) in the inter- tidal community of San Juan Island, Puget Sound, Washington. Marine Biology, 51: 119- 3 MACARTHUR, В. Н. 8 WILSON, E. O”., 1967, The theory of island biogeography. Monographs in Population Biology, No. 1. Princeton University, Princeton, New Jersey, 203 p. MANDAHL-BARTH, G., 1972, The freshwater mol- luscs of Lake Malawi. Revue de Zoologie et de Botanique Africaines, 86: 257-289. MCCLARY, A., 1961, Apparent geotactic behavior in Physa. Nautilus, 75: 75-83. MCCLARY, A., 1964, Surface inspiration and ciliary feeding in Pomacea paludosa (Prosobranchia: Mesogastropoda: Ampullariidae). Malacologia 2: 87-104. MCKAYE, K. R., 1978, Explosive speciation: the cichlids of Lake Malawi. Discovery, 13(1): 24-29. MCKAYE, K. R., 1982, Ecology and breeding be- havior of a cichlid fish, Cyrtocara eucinostomus on a large lek in Lake Malawi, Africa. Environ- mental Biology of Fishes: in press. PRASHAD, B., 1925, Respiration of gastropod mol- luscs. Proceedings of the 12th Indian Science Congress, Calcutta, Asiatic Society of Bengal: 126-143. VERMElJ, С. J., 1975, Evolution and distribution of left-handed and planispiral coiling in snails. Nature, 254: 419-420. VERMEIVJ, С. J. 4 COVICH, A. P., 1978, Coevolu- tion of freshwater gastropods and their preda- tors. American Naturalist, 112: 833-843. WIEBE, A. H., 1926, Variation in the freshwater snail, Goniobasis livescens. Ohio Journal of Science, 26: 49-68. WORLD HEALTH ORGANIZATION, Snail Informa- tion Center, 1977, A Field Guide to African Freshwater Snails. 4: Southeast African Spe- cies. Danish Bilharziasis Laboratory, Jaegers- bog Alle 1D, DK2920, Charlottenlund, Denmark, 20 p. WRIGHT, С. A., KLEIN, J. & ECCLES, D. H., 1967, Endemic species of Bulinus (Mollusca: Planor- bidae) in Lake Malawi (=Lake Nyasa). Journal of Zoology, 151: 199-209. YONGE, C. M., 1938, The prosobranchs of Lake Tanganyika. Nature, 142: 464-468. MALACOLOGIA, 1982, 23(1): 23-35 ENVIRONMENT AND SHELL SHAPE IN THE TAHITIAN LAND SNAIL PARTULA OTAHEITANA Kenneth С. Emberton, Jr. ! Department of Zoology and Microbiology, Ohio State University, Athens, Ohio 45701, U.S.A. ABSTRACT A multivariate analysis was performed on data from Crampton’s 1916 study of the arboreal snail Partula otaheitana in 50 valleys of Tahiti Nui, Society Islands. Populations in wetter, more shaded valleys tended to have shells more elongate in shape with a higher percentage of banding and with a reduced apertural barrier than populations in drier, more sunlit valleys. The correlations accounted for 20% of the variance in the three shell characters mentioned. Causa- tive factors of these variations remain to be investigated. Key words: Partula; snail; shell shape; environment; Tahiti; evolution. INTRODUCTION Cain & Sheppard’s (1950) landmark dis- coveries of some mechanisms of natural se- lection in shell coloration of the land snail Cepaea nemoralis explained some of the geographic variation in one seemingly non- selected polymorphism. Another well-studied land snail group, the partulids, however, re- mains enigmatic by its possession of appar- ently neutral characters. The family Partulidae is hypothesized (Kondo, 1973) to have invaded the Society Islands in three waves, the last of which (genus Partula) probably has reduced its predecessors (genus Samoana) to near ex- tinction. Partula now dominates in upland val- leys of the more westerly Society Islands. The genus is isolated by a hierarchy of barriers: onto individual islands by the Pacific Ocean; into specific valleys by dry or devegetated hogback ridges radiating from the center of each volcanic island; and into partially to total- ly discrete populations within valleys by dis- rupted patches of vegetation. Partula repre- sents a complex array of species, subspecies, morphs, and individual variants. H. E. Crampton, attracted by partulid diver- sity and variability, produced classic concho- logical studies on the partulids of Tahiti, Moorea, and the Mariana Islands (Crampton, 1916, 1925, 1932). Portions of his extensive data have been reanalyzed subsequently (Ludman, 1947; Bailey, 1956). His initial genetic investigations, in which he compared the shells of adult snails with the shells of embryos dissected from their uteri (Partula is Ovoviviparous), have been greatly augmented by recent hybridization experiments and field observations of Moorean species (Clarke & Murray, 1969, 1971; Johnson, Clarke & Murray, 1977; Murray & Clarke, 1966, 1968a, 1968b). They have shown that several shell characters, namely size, color, banding, and Sinistrality of two species of Partula from Moorea are highly heritable. Lipton & Murray (1979) compared courtship patterns of the two species and found evidence of behavioral reproductive isolation. It remains moot whether shell characters of Partula are subject to natural selection. Crampton, one of the first American scientists to teach Mendelism, was convinced that the extensive variation in Partula resulted solely from mutation and genetic drift and was un- affected by natural selection (Crampton, 1932). His reasoning was that: (1) habitat is uniformly humid tropical; (2) predation is ab- sent, except by rats introduced by man into coastal areas where Partula is scarce to ab- sent anyway; (3) the particular plant type which provides the decaying vegetation upon which Partula feeds is “a matter of indiffer- ence” (Crampton, 1916); and (4) different species of Partula occurring micro-sympatri- cally, even on the same leaf, in any given val- ley exhibit different shell morphologies and color patterns. Crampton's theory of random, unselected variation in Partula has been sup- ported by Huxley (1942) and contested by Present address: Committee on Evolutionary Biology, University of Chicago, Chicago, Illinois 60637, U.S.A. (23) 24 EMBERTON TAHITI NUI TAIARAPU FIG. 1. Tahiti, Society Islands. Lines represent ridge tops. The 50 labeled valleys were those used in this study (from Crampton, 1916). Cain & Sheppard (1950) and Ford (1964). The majority of attempts to correlate environ- mental variables with shell characteristics of Partula (Clarke & Murray, 1971; J. Murray, 1979, personal communication) have been unsuccessful. A single trend was discovered among the Partula of Moorea: there is a clear negative correlation between shell length and altitude. This paper presents evidence of a different nature, based on a multivariate analysis of Crampton’s (1916) data, of a linear relationship between general environmental trends (other than altitude) and shell char- acteristics of Partula otaheitana on Tahiti Nui (Great Tahiti). Crampton recognized six species of Partula on Tahiti. Of these, only P. otaheitana ap- proaches ubiquity in distribution. This species is endemic to Tahiti, abundant, and exhibits a wide range of shell morphologies and colora- tions. Crampton divided P. otaheitana into eight subspecies whose geographical distri- butions are mapped in Fig. 2. Sympatry of the nonhybridizing ‘subspecies’ affinis and rubescens in northern valleys makes Cramp- ton's designations immediately suspect. In- deed, Lundman (1947) proposed promoting rubescens to full species rank based on his reanalysis of Crampton’s data. However, be- cause Clarke & Murray’s (1969) hybridization experiments on Partula of Moorea resulted in lumping rather than splitting Crampton's spe- cies, and because no additional data have been published concerning the taxonomy of P. otaheitana, | shall adhere to Crampton's designations. The most obvious physical environmental variables that might exert selection pressure on Partula otaheitana are temperature and humidity. Partula is absent from drier regions ENVIRONMENT AND SHELL SHAPE IN PARTULA 25 3 9 A 7. > 7 \ 2 a 5; rubescens A e =. 9, A w 9%, 9, > atfinis \ af ) 3 ln E sinistralis sinistrorsa fe SE zen FIG. 2. Distributions of eight subspecies of Partula otaheitana on Tahiti, according to Crampton (1916). of the coasts and from deforested ridgetops, where direct sunlight produces greater tem- perature fluctuations. It is common to abun- dant in more shaded areas of consistently high humidity and lower temperature fluctua- tions, particularly in remaining patches of native vegetation (Crampton, 1916; A. Solem, personal communication). Partula is active generally under conditions of saturated hu- midity; much of the time the animal rests with the shell aperture sealed to the undersides of leaves (Crampton, 1916). Ideally one should have records of temperature and humidity for each area sampled. Since such data do not exist for Tahiti, it is possible only to estimate general trends on the island for those meteor- ological variables most heavily influencing temperature and humidity, namely insolation and rainfall. Average insolation in valleys of Tahiti Nui varies according to slope, aspect, shadowing by adjacent ridges, and cloudiness. Variation in all these factors is considerable due to the steeply pitched terrain. Rainfall differs mark- edly according to position on the island, valley side, elevation, and season. Moisture reten- tion varies drastically with vegetation cover. Much of the moisture carried by the nearly constant southeast trade winds precipitates on the windward side, resulting in about 635 cm per year on the windward coast and an estimated but inadequately measured 2590 cm per year on the highest slopes. The central peak produces a distinct rain shadow, so the leeward coast averages only about 200 cm per year (Crampton, 1916). Thus, av- erage rainfall would vary on Tahiti Nui accord- ing to both elevation and horizontal deviation of the mouth of the valley. Given these trends, the relative average insolation and rainfall can be estimated for any given valley on Tahiti Nui because of its circular shape, single central highland area, and regular series of radiating valleys (Fig. 1). Taiarapu (Tahiti Iti) was ex- cluded from analysis for the sake of simplicity. MATERIALS AND METHODS Crampton's field parties collected Partula otaheitana during four visits: February-March, 1906, June-July, 1907, June-August, 1908, and unspecified months in 1909. A total of 18,509 living adult snails were taken from 50 valleys. Crampton recorded the following set of shell data for each specimen: shell length, shell width, aperture length, aperture width. He calculated the ratios shell width/length, aperture width/length, and aperture length/ shell length for each specimen, and also re- corded whether the shell was sinistral or dex- tral, banded or unbanded. He classified each shell into one of five categories according to the degree of development of the parietal apertural barrier (1 = no barrier, 5 = very large barrier), and into one of several cate- gories according to shell color. Crampton's original measurements of indi- vidual shells could not be located. He pub- lished only the mean and standard deviation for each shell variable for a given subspecies in a given valley. With eight subspecies oc- curring, sometimes sympatrically, in 50 val- leys, this results in 62 subspecies-valleys for Tahiti Nui. | was able to calculate % sinistral and % banded for each subspecies-valley from Crampton's tables of those percentages broken down according to color form. Cramp- ton used a different color scale for each sub- species, so | standardized all mean color index values to the same scale he used for P. o. otaheitana: 1 = lightest, 4 = darkest. To illustrate the procedure by way of example, the sample of P. o. sinistrorsa in Tenaire Val- ley (Crampton, 1916, table 176) consisted of 907 adult shells of which 26.9% were classi- fied by Crampton in the color form apex, 25.7% in cestata, and 47.4% in phaea. Com- paring the colors of representative shells pic- tured in Crampton's (1916) plates, | judged that apex corresponded to a value of 1.0 on the P. o. otaheitana scale, cestata corre- sponded to 3.4, and phaea corresponded to 4.0. (These judgments must show subjective bias, but as long as the bias is consistent EMBERTON 26 L9'c Sees 19'bL $19 106 2085 8/'6 sg gl 05% 98 00 sıulye er ol с eyıaded LOL 6505 89 cB lL 6L OL 81'959 67 LL eL OZ ZBL 00 0001 susoseqni er CL с ецэдеа cv € 6/06 ct 62 59'9 £c8 c6 LS 15`6 BL OL 05% 053 00 sıulye 8t 9 v elyneyee 00°¢ G6 CS LO ZZ 499 99 8 cl 6S 99'6 LEOL 6t € G'6L 00 SIUIJE 9S 6/ с eINeled 861 90'159 Lt'82 191 6/6 8999 78`01 661 911 00 0'001 sugoseqns 9S 62 с eineled BEC G6 £G БЕДА 169 10`6 8t 6G £66 ggg, 0gE 20 00 sıulye 79 GG 14 YEA LOE G£'cS 8671 £8 9 9/'8 90°8S 0/6 0Z'9! 0€'€ er 8c ие 08 cl с BUSEUEW 19€ 00'ES 2192 191 186 00 09 ПРАВЕ 8981 05° 00 0001 SU9ISIANI 08 GL с EUSEUEW Ivo DVIS +0 9/ L£ 9 L£ 8 ÿL'2G 0£ 6 co Ot 00'€ 90! 00 sıulye 16 85 € 19181] 2071 OE LS 06 9/ 067 Ze Ol 0€'85 SO LL 8661 LL |. 00 0001 su99sagqnI 16 85 € 191811 сис cS cS Ov 82 cv9 218 89'695 136 ÿSGL ASS 6€ 00 sıulye 90+ Sc y lEUNIEE + 001 vo cS Gt 9/ lc 8 9.01 9979 98 ll vS OC 05° 00 0001 Su99S9qn1 90! GC y lEUNIEE y ЕТ ¡ARAS 99597 COL ct 6 ZAYAS 6c OL c8 Zt WEG LEG 00 sıulye Elk Le € edeJey cs kt 6t ES Ce vl +9 9 L68 Le 8S 0/6 ca Ol sa 00 vv sıulye cal сс | oousded 8rc 9905 eS pl OLE ce OL LE SG Oe Lt 8g£ OC ¡AA 00 0001 Sus9saqn AA! co L ooueded 80'¢ 19'59 £8 G/ 189 468 90 6S 066 eZ Ol LES Bar 00 sıulye 051 8 € oodiee + 8rvc c8 cs 1492 £6 9 10'6 10'6S 9001 LO'Z! 897 073 0° sıulye Op! 61 с nuoyy 870° 996r py Sl ДАЛ 8c OL +9 vG Gear: 99°02 961 00 000L suaassaqni Op! 6L с nuoyy 987 67'55 6€ 8/ iel с5'6 ce 6G ce OL Be Zl 8c € c 9c 80 ще orl G с ninenl Gral 8£'6b cl SZ COL 02/6 00'pS SS Ol Os’6L Sc? 00 0001 $и955э9п/ 97+ G с ninenl cg'c 99'235 pS Ll DIZ 126 ve'6S 9g£ OL Gt 11 pS 60 166 sııgewe OS! 01 € 20/141] v9c 6S CG 89 22 6, 7 9c 6 c9 6G 8t OL 9571 sec 00 0001 sııgewe 691 55 5 eie}Ieuod 06 C 69 IG 0677 LOT. 976 O€'8S 4901 8c'8l 8c | 6! 000L sııgewe sg 85 с ела 60 € G0'ZS £c 82 097 956 Lt 8S LZ'OL ce sl CLG 02 |'59 SI//qewe С er € EINWEH PAG Ly LG 1v 82 GLL £8 6 py ls 66 01 06+ 0/73 G'O L'GG eueJIayejo 821 8c | enejney es ее 2 E 7 5 = = O O O ® ® D = с Ф Oo = O = = _ un = mye al = = — x = a = oO @ a — Е Е = = = = 5. D D D < — 0 ' £ 3 3 = = er à 2 ® D 2 , = =. x = = a Oo O = О. = = 3 D o < а =] x = =: 2 = 2 = ® 8 à ce EOS Os a © 3 a =; = = x’ 0 o > Q — is Y) no © = 5 > 2 = E m eE ee m ЕЕ ЕЕ В И mar Br a Es mann un nn nn ‘Аэнел-заюэ4зап$ yoea 10} зэаеиел |jeys jo зиеэш pue зари! |EJUBWUOIIAUS :INN NiyeL jo SAaeA AYI4 ‘р JIGWL 27 ENVIRONMENT AND SHELL SHAPE IN PARTULA т к DL QR — RJ A _R __Q_R_ __ >= _ a o DEE $ O II IRA 951 €0'1S 16'SZ col 6+ 6 GLS gell ¿S'8! ce € 595 ce 0Z'06 ce 92 col 9v'6 BESS ve lt v9'8l 08'€ v'6€ 60 1 20'06 80 92 LOL 8001 0/`75 O0'EE 0106 ce € S'S€ 001 80 IS 86 S/ SOL 526 68 SG st Ol 60°81 68 € Lic 001 6t 0S 8991 60 Z ZA 9659 €c'Ol Zest LR т 16 BE 1206 LVALE LOL E16 APA 15'01 00'81 8s'€ c'GG 501 95716 06'SZ би д ¿86 v8'ss 89°01 0161 [083 ¿Sc 80"! 9315 81772 LL er Ob 21588 L6'OL S/'6! 975 Les 00"! Or'6r 0877 BIZ 09'6 0166 0701 6€ 61 65`5 0'€6 ДОЗ 00`15 0S'pZ 08'9 106 ZI-LS er ol SAL 00'p 00 001 OS'IS 05`58 0€Z 0/8 0S'09 05`01 SES 09° 00 ele RAS 7392: 2919 89'8 €2°8S 0/'6 6b'91 0S'c cock 8c'€ L9ES GE 92 cL'9 6/8 69 8S 08'6 99'91 0S'c lr вивиб! BSSEIO esseso 255219 255219 255819 255219 255819 255219 255219 SNEASIUIS SNEASIUIS SIEASIUIS SNENSIUIS SNENSIUIS SIEASIUIS SNENSIUIS SNENSIUIS SNEASIUIS SIPASIUIS SIPASIUIS BSIONSIUIS ESIONSIUIS 25/0515 BSIOHSIUIS 25/05/15 25/0455 BSIOHSIUIS ESIONSIUIS BSIONSIUIS sıuıye suaoseqns sıulye suaoseqni иде sıuıye suaoseqns oO NE — — YE INS E © cl OZ 88 LL 69 89 99 LZ S8 eZ 05 ооо чяяяяяяттяяяяттяяя я CU ED C9 ED ON ED EH HO QU inoedi] eundee] nnueung oodenieyw ау anyaded enoy 219J01O eJe my ednyeN ниепагел оеше! | elyaaja | EINJEN eUNIO e01111dO ndeyeiea | oodieA enjeua | pode d nyos | еплеце| eoleoyy IyeJee 4 ецчелецелел BUIUIEA PJPEN элеиэ1 LAB} L anewuidy anewuidy anewuidy ndoo ndoo 2u0ey епт епт п 28 EMBERTON across all samples, this method is justifiable.) Multiplying each score by the appropriate per- centage, then adding the wieghted scores, | calculated a mean color index of 3.03 for P. o. sinistrorsa in Tenaire Valley. For banded shells, color was scored for the unbanded portion of the shell only. | called the color index ‘darkness.’ Because of the many morphs, color could have been coded as more than one variable. However, Murray & Clarke (1976a, 1976b) have shown that, for two Moorean species of Partula, all genetic loci determining shell color are so tightly linked that they are best con- sidered a single “supergene.” On the working assumption that the same applies for P. otaheitana, color was restricted to a single variable. Thus, with an average value for every shell variable for every subspecies-valley, a 62 x 11 data matrix was filled (Table 1, columns 6-16). There are several problems with these data. In order to evaluate differences among valleys with regard to mean shell variables, the samples should be comparable in number of individuals, area sampled, location within each valley, and vegetation cover. Unfortu- nately, those conditions do not apply to Crampton's collections. Sample size ranged from 2 for P. o. rubescens in Faaripoo Valley to 988 for P. o. amabilis in Pirai Valley. Crampton gave no indication as to either the area collected or the time spent collecting in each valley. Presumably both varied greatly depending on terrain, vegetation cover, num- ber and identity of collectors, weather, local abundance, etc. Dates of collection ranged from February to August, but since Partula shows determinate growth (reaching a final adult size), variation in the date of collection seems irrelevant. Despite imperfections, the data were con- sidered worthy of analysis for several rea- sons. First, the general exploratory nature of this study permitted some leeway in experi- mental design. Second, although the errors in- troduced by noncomparable samples would obscure any clinal gradients in means of shell variables, the number of valleys was probably sufficiently large (50) that the mean of those errors would approximate zero, thereby leav- ing the clines unbiased. Third, the fact that samples were taken differently in each valley would similarly tend to iron out errors due to a dry spot in a wet valley or vice versa. Finally, it is unlikely, even in Crampton's time, that any sampling regime on Tahiti Nui could avoid all of the same deficiencies. Choosing a sample site or sites in each valley large enough to yield sufficient numbers of snails but small enough to avoid extraneous variation due to environmental gradients, interpopulational dif- ferences, and step clines would have been difficult then and impossible today due to habitat destruction. Lacking actual meteorological data, | de- vised indices for relative insolation and rela- tive rainfall. For relative insolation, the main axis of each valley was drawn by eye on Crampton's (1916) map of Tahiti Nui. The angular deviation, in degrees, of the axis from a north-south line constituted the insolation index for that valley (Fig. 3A). Thus the values ranged from 0 for north-south oriented valleys receiving the least sunlight (ignoring the effect of shading by the central peak) to 90 for east- west oriented valleys receiving the most sun- light. This variable was named ‘insolation.’ For relative rainfall, the deviation of the mouth of each valley from the southeast was calculated. The angle at which a line drawn from the estimated mouth of the valley to the estimated center of the island subtended a northwest-southeast line constituted the rain- fall index for that valley (Fig. 3B). Thus the values ranged from O for highest rainfall to 180 for lowest rainfall. The variable was named ‘low rainfall.’ Size of each valley might mediate the ef- fects of insolation and rainfall on temperature and humidity. Wider valleys might be warmer due to increased insolation. While it is logical to think that wider valleys would be drier, their greater extent of water-retaining flatlands re- tain runoff longer and hence they could be more humid (Crampton, 1916). Crampton categorized each valley according to size on a scale of 1 to 4 ranging from widest to nar- rowest valleys (Fig. 3C). | named this variable ‘valley smallness.’ Obviously, these three enviromental in- dices (Table 1, columns 2-4) are gross ap- proximations which ignore the wide range of insolation and rainfall present within any given valley. Nevertheless, their use can be justified by arguments similar to those used in justify- ing the use of mean shell measures for each subspecies-valley. First, in a search for gen- eral trends, some incertitude of data is per- missible. Second, although the indices theo- retically measure gradients occurring at sites with identical elevation, valley side, aspect, and vegetation, deviations of individual sites from the means of these criteria would tend to ENVIRONMENT AND SHELL SHAPE IN PARTULA 29 N MAIN AXIS OF FARAPA VALLEY INSOLATION INDEX FIG. 3. Calculation methods for three environment- al indices, as demonstrated for Farapa Valley. cancel each other out, leaving the gradients obscured but unbiased. Finally, climatic se- lection on snail populations is probably ef- fected by an unknown mix of longterm aver- age trends in insolation and rainfall (mean selection) and occasional cataclysms (crisis selection), neither of which is easily meas- ured. It may well be that the derived indices are the best estimates both of long-term trends and of cataclysms, short of setting up and monitoring a weather station in each val- ley at prohibitive expense. All statistical analyses employed BMDP-77 computer programs (Dixon & Brown, 1977). For the sake of simplification, the 11 shell variables were subjected to a factor analysis (Tables 2 and 3), a method which combines each group of highly intercorrelated variables MOUTH OF FARAPA Le VALLEY CENTER OF ~~ ^ \ ISLAND B RAINFALL ne in De PV XA ES \ Y de ih иг \ SE 2% De E e | N SA == E N 2 \ \ 4 ) x VALLEY SIZE INDEX to form a new single variable (factor). For ex- ample, the four measurements shell length and width, and aperture length and width, were highly intercorrelated, and therefore they were combined, each appropriately weighted, to form a single factor which | named ‘shell size.’ The total factor analysis resulted in condensing 11 shell variables to seven shell factors (Table 2). These were used as new and more independent variables for further analysis. The relationship between the three standardized environmental indices and the seven standardized shell factors was investigated by two different procedures. First, each shell factor in turn was regressed on the three environmental indices by step- wise linear regression. In order for this test to be valid, the data must conform with several 30 EMBERTON TABLE 2. Rotated factor loadings of seven named factors extracted from 11 shell Variables of Partula otaheitana. The seven factors explain 98.6% of the variance. Loadings <0.250 were replaced by zero. Oblique (D-Quartamin) rotation was used (see Table 3). Shell Shell Aperture % % Barrier Dark- size squatness roundness Sinistral Banded size ness Aper. width 1.011 0.0 0.0 0.0 0.0 0.0 0.0 Aper. length 0.960 0.0 0.0 0.0 0.0 0.0 0.0 Shell width 0.957 0.0 0.0 0.0 0.0 0.0 0.0 Shell length 0.839 0.0 0.0 0.0 0.0 0.0 0.0 AL/SL 0.0 0.925 0.0 0.0 0.0 0.0 0.0 Shell W/L 0.0 0.804 0315 0.0 0.0 0.0 0.0 Aper. W/L 0.0 0.0 0.984 0.0 0.0 0.0 0.0 % Sinistral 0.0 0.0 0.0 1.004 0.0 0.0 0.0 % Banded 0.0 0.0 0.0 0.0 1.000 0.0 0.0 Barrier size 0.0 0.0 0.0 0.0 0.0 0.998 0.0 Darkness 0.0 0.0 0.0 0.0 0.0 0.0 1.000 TABLE 3. Correlation matrix of shell factors of Partula otaheitana. Mean correlation = 0.239 (not significant), standard deviation = 0.157. eee eee — ЕЕ ЕЕ ЕЕ ЕЕ ЕЕ Shell Shell Aperture % % Barrier Dark- size squatness roundness Sinistral Banded size ness A A AE EE A AAA AA ee _ — Shell size 1.000 Shell squat. 0.361 1.000 Aper. round. 0:11 0.180 1.000 % Sinistral 0.441 — 0.381 — 0.127 1.000 % Banded —0.161 —0.374 —0.097 0.195 1.000 Barrier size —0.440 0.442 0.124 —0.210 — 0159 1.000 Darkness —0.438 0.065 —0.171 0.003 0.500 0.035 1.000 EE assumptions (Poole, 1974). The assumption of linear relationships between all shell- environment pairs was judged not violated based on visual examination of bivariate scat- tergrams. The assumption of orthogonality of predictor variables (environmental indices), a very commonly violated assumption (Green, 1979), was considered not violated on the basis of the low, non-significant correlations among the 3 indices, as shown below. cording to Green (1979), regression is a suf- ficiently robust technique that such minor vio- lation is tolerable, so long as the distribution of the variable is strongly unimodal, as was the case in this instance. Second, a canonical correlation analysis was performed between the set of shell factors and the set of environ- mental indices. The assumptions of canonical correlation are satisfied by meeting the as- sumptions of regression. Valley Insolation Low Rainfall Smallness RESULTS Insolation 1.000 : Low Rainfall #0.111ns 1.000 The seven shell factors were easily named Valley Small. 0.008ns -0.056ns 1.000 by examining their loading patterns (Table 2), The assumption of normality of each variable was tested by calculating kurtosis: a variable having a kurtosis value =3.00 was consid- ered normally distributed. The only variable violating this assumption was the factor ‘shell size, which had a kurtosis value of 4.47. Ac- which are relative measures of the contribu- tion of each variable to the additive construc- tion of each factor. The intercorrelations among the new shell variables (factors) were low (Table 3), indicating their relative inde- pendence. Only three of the seven shell factors were ENVIRONMENT AND SHELL SHAPE IN PARTULA Sil TABLE 4. Summary of stepwise regressions of seven shell factors on three environmental indices. Degrees of freedom were 1, 60 for F-to-enter. Significance levels are: *(p < 0.05), **(p < 0.01), ns (p > 0.05). NE EP EN EE SEE ES TRE % Variance Initial F-to-enter values explained by Low Valley Shell factor regression rainfall Insolation smallness Shell size 0.0 0.026"s 1.731NS 0.547nS Shell squatness 19.8 5.822” 9.091** 1.532NS Aper. roundness 0.0 0.080"S 0.506"S 1.04715 % Sinistral 0.0 0.174nS 0.075nS 0.155NS % Banded 20.9 5.042* 10.847** 1.689NS Barrier size 22.9 11.6972 5.933” 0.327NS Darkness 0.0 1.549nS 2.769"S 0.744nS A A AAA AAA A A AA O A significantly predicted by the environmental indices (Table 4). Neither shell size, aperture roundness, % sinistral, nor shell darkness was predicted. Shell squatness and apertural bar- rier size were both positively predicted by low rainfall and insolation. Insolation was most important in predicting shell squatness, whereas low rainfall was most important in predicting barrier size. Percent banded was negatively predicted by low rainfall and insola- tion, with insolation the most important. Each of the three significant regressions explained about 20% of the variance of its dependent variable. Valley smallness had no value as a predictor of any of the shell factors. A single significant canonical variable was extracted which expressed a correlation coef- ficient of 0.71 between the two sets (Table 5). Strong positive loadings of insolation and low rainfall correlated with strong positive load- TABLE 5. Canonical variable loadings for set of en- vironmental indices vs. set of shell factors. The sin- gle significant (p < 0.001) canonical variable gives a correlation coefficient of 0.710 between the two sets. First set Insolation 0.809 Low rainfall 0.674 Valley smallness —0.059 Second set Shell size —0.189 Shell squatness 0.620 Aper. roundness —0.071 % Sinistral —0.010 % Banded —0.648 Barrier size 0.646 Darkness —0.357 ings of shell squatness and barrier size and with strong negative loading of % banded. The results of stepwise regression and can- onical correlation were entirely complemen- tary. Together they suggest that populations of Partula otaheitana in wetter, more shaded valleys tend to have shells more elongate, with a higher percentage of banding, and with lesser development of a parietal barrier than populations from drier, more sunlit valleys. This trend is depicted in exaggerated form in Fig. 4. DISCUSSION AND CONCLUSIONS Several aspects of this study may obscure a true relationship between environment and shell characters. First, as previously dis- cussed, all variables used were averaged or indirectly estimated for each entire valley. Second, there were almost certainly a large number of other variables influencing shell characters (see Jones et al., 1977, for a dis- cussion of this problem with regard to Cepaea), including predation. For example, in contrast to Crampton’s conclusion of zero predation above coastal elevations, Kondo (1973) mentioned a specimen of Samoana burchi, a species no less heavily shelled than Partula otaheitana, collected at 1250 m on Tahiti Iti (where P. otaheitana also occurs), the apical half of which “had been eaten by some animal.” A. Solem (personal communi- cation) reports extensive predation on Partula by rats throughout Tahiti. Third, the two "sub- species” P. o. rubescens and P. o. affinis, which are sympatric in 11 northeastern val- leys, exhibit dissimilarities which may be in- terpreted as character displacement related to reproductive isolation mechanisms. P. O. 32 EMBERTON DRY, SUNLIT VALLEYS WET SHADED VALLEYS STRONG BARRIER SQUAT SHELL LITTLE BANDING WEAK BARRIER ELONGATE SHELL MUCH BANDING FIG. 4. Relational trends between enviromental indices and shell factors of Partula otaheitana on Tahiti Nui. Drawings exaggerate the trends by an approximate factor of 5. rubescens is exclusively sinistral and never banded; Р. o. affinis is predominantly dextral, sometimes banded (maximum 25%) and con- sistently smaller than P. o. rubescens (Crampton, 1916). Lipton & Murray (1979) found that opposite direction of coiling in P. suturalis is correlated with incompatibilities in mating behavior that “may result in partial re- productive isolation of dextral and sinistral animals.” Whatever the cause of divergence between P. o. rubescens and P. o. affinis, the result is to obscure environmental correla- tions. Whether these are species, semi- species, or systematically confused remains to be determined. It is noteworthy that, in spite of these ob- scuring aspects of the data, two of the three environmental indices explained (in a statisti- cal sense) 20% of the variation in three of the shell factors. Correlation, of course, does not imply causation. As Sokal (1978) put it, “Phenotypic distributions are believed to re- flect the distribution of selective forces through evolutionary history, but while un- doubtedly true in general, this is quite difficult to demonstrate in a specific case.... Al- though causal links between such distribu- tions are frequently assumed, . . . they are by their nature very difficult to prove.” With that air of caution, it is worthwhile to examine whether shell characters of Partula otaheitana which are correlated with an environmental trend could feasibly have been caused by that trend, based on related evidence. Apertural barriers in land snails have been assumed to function as protection against small invertebrate predators (Solem, 1976) or as aids in forming effectively placed mucous membranes to prevent desiccation (Ember- ton, unpublished). The apertural barrier of Partula otaheitana never reaches sufficient size to be very convincing as a block to pre- dators (Fig. 4), yet the possibility cannot be ruled out that it discourages attack by un- known invertebrate predators or parasites. Mucous membranes are formed only around ENVIRONMENT AND SHELL SHAPE IN PARTULA 33 the periphery of the aperture when the snails seal to the undersides of leaves, therefore the barrier is unlikely to be adaptive in this con- text. Both shell squatness and low incidence of banding in some desert snail species have been construed as adaptations to dry, sunny conditions (Yom-Tov, 1971; Schmidt-Nielsen et al., 1977). Partula is found only under hu- mid, shady conditions. Such conditions, how- ever, are relative. Partula may be sufficiently sensitive that slightly less humid regimes exert selective pressures similar in kind but not degree to those experienced by desert snails. Another hypothesis to explain some of the trends depicted in Fig. 4 was suggested by Alan Solem. Incremental shell growth to the edge of the aperture can occur only while the snail is extended, but thickening of the shell at any growth stage and size increment of the apertural barrier in adults can occur while the snail is retracted, sealed to a leaf, and quies- cent. Snails in drier valleys would spend a greater percentage of time in quiescence than snails in wetter valleys. Hence, assuming that new shell is formed by both active and quies- cent snails, those in drier valleys could have more squat, less elongate shells with a thicker apertural barrier as a growth byproduct. This hypothesis may invoke non-selected variation to explain phenotypic trends, and could be tested. For example, newly-born siblings could be reared under different humidity regimes. It is relevant that Pollard (1975) demonstrated a negative correlation between rainfall and shell thickness in Helix pomatia in a range of collections from southern England. Are the same phenotypic trends evident in other species of Partula? Partula hyalina, the next most widely distributed Tahitian Partula, has not been analyzed by the methods of this paper but appears to have a distribution of shell shapes similar to that herein demon- strated for Partula otaheitana. Based on a total of 463 measured adult shells, Crampton concluded that “(shells) in the north are somewhat shorter... far broader... (and) much stouter than the southern (forms)... [although] the relations mentioned are not in- variable by any means.” Partula hyalina, a species with an “unusually distinctive” pheno- type, unlike other Tahitian Partula, is wide- spread on other islands (Crampton, 1916). Partula clara, a species of much more re- Stricted range, on the other hand, showed the Opposite phenotypic distribution. Shells of the southern quadrant of Tahiti Iti “on the whole are the stoutest group” (Crampton, 1916). It may be relevant that Crampton deduced P. clara had been a rare species which had, within 50 years, profusely expanded its range and numbers. The other three species of Tahitian Partula were highly restricted in range. Conchological trends in Moorean Partula are also inconsistent. Lundman (1947), when mapping Crampton's data, discovered a cline in Partula taeniata on Moorea which approxi- mates the trends reported here for P. otaheitana on Tahiti Nui. Shells of P. taeniata were larger and longer in the northeast, broader in the extreme north, narrower in the south; aperture width/length graded from small in the southeast to large in the north- west; barrier size graded from absent in the south to weak in the north. Nearly all shells were dextral, and no trends in banding were apparent. On the other hand, P. suturalis, sympatric with P. taeniata, showed opposite, though weaker clines in shell size, length, and width, but it showed the same clinal trend as P. taeniata in aperture width/length. Dextral shells of P. suturalis occurred in the south- east, sinistral in the northwest and northeast. These conflicting trends lead to two possi- ble interpretations. First, the correlations be- tween environment and shell shape occur fortuitously as a byproduct of clines in shell shape produced by restricted gene flow, for example. Or, second, that the correlations oc- cur by way of direct causal relationship be- tween environment and shell shape which can be masked or overridden by rapid migra- tion, interspecific competition, etc. Choosing the correct one of these alternatives would be a difficult if not impossible task, but could have highly fruitful results. For example, it might be found that the degree to which envi- ronment and shell shape were correlated for a given species of Partula would indicate rela- tively how long that species had occupied its present range. That the relationships between shell varia- bles and environment in P. otaheitana on Tahiti Nui were invisible to Lundman s (1947) method of mapping characters speaks well for the power of the multivariate tests used in this paper. Green (1979) emphasized the value of multivariate analysis as an exploratory tool to detect basic trends in data. He recommended using more than one technique in order to safeguard against the pitfalls inherent in any one. The greater the number of dependent 34 EMBERTON variables in a series on which stepwise re- gression is performed, the greater the chance of concluding a significant regression exists when in fact it does not. Significance levels of the three significant regressions in Table 4 are probably low enough (p < 0.01, p < 0.001, p < 0.001) to assure that this error has not been made. The regression results are further reinforced by the clearcut results of canonical correlation analysis (Table 5). Be- cause canonical correlation maximizes the correlation between two sets of variables, it may introduce spurious relationships or ex- aggerate real ones (Pimentel, 1979). Step- wise regression guards against those dan- gers, especially because it indicates the strength of the relationship as percent ex- plained variance. To return to the original question of whether shell characters of Partula are subject to natu- ral selection, the question remains unan- swered by this analysis. Nor does it answer the question of whether the shell characters are adaptive (whether selected or not). At most, this analysis raises the possibility that adaptation and natural selection have played a role here. The distribution of partulid pheno- types is a fascinatingly complex one; the key to its understanding doubtless lies neither in the extreme non-selectionist view of Cramp- ton (1916) nor in a dogmatic adaptationist program (as criticised by Gould & Lewontin, 1979), but somewhere in between. Partula otaheitana can still be found in moderate to high densities at middle to upper elevations on Tahiti (A. Solem, personal com- munication). Ecological fieldwork on these contemporary populations would provide both a valuable update to Crampton’s work and а more empirical test of the trends detected in this analysis. Partula on Tahiti presents the same opportunity of an approachable natural experiment in evolution that first attracted Crampton over 70 years ago. The evolutionary relationships among spe- cies of Partula on Moorea are now reasonably well understood, thanks to over 15 years of exemplary work by Murray, Clarke, and co- workers (for a review, see Murray & Clarke, 1980). Moorea and Tahiti are successive vol- canoes formed over the same hot-spot and separated by movement of the Pacific Plate. Moorea is 1.65+0.13 million years old, whereas Tahiti Nui is only 0.65 + 0.22 million years old (Dymond, 1975). In many respects, then, Tahiti is a younger, less-eroded version of Moorea. Building a data base on Tahitian Partula comparable to that already existent for Moorean Partula would allow valuable and unique insights into the relationship between geomorphology and organic evolution. ACKNOWLEDGMENTS | thank those who offered encouragement and/or advice on this analysis: Arthur Cain, George Davis, Robert Dillon Jr., James Murray, Alan Solem, Gerald E. Svendsen, and Joseph Vagvolgyi. | am grateful to Ronald and Sharon Circe, Corpus Christi, Texas, for gracious hospitality while the first draft was being written and to Elliot Rosen, Ned Walker, Leigh Van Valen, Michael Wade, and especi- ally Alan Solem and two anonymous re- viewers for critically reading drafts of the manuscript. LITERATURE CITED BAILEY, D. W., 1956, Re-examination of the di- versity in Partula taeniata. Evolution, 10: 360- 366. CAIN, A. J. & SHEPPARD, P. M., 1950, Selection in the polymorphic land snail Cepaea nemoralis. Heredity, 4: 257-294. CLARKE, B. & MURRAY, J., 1969, Ecological genetics and speciation in land snails of the genus Partula. Biological Journal of the Linnean Society, 1: 31-42. CLARKE, B. & MURRAY, J., 1971, Polymorphism in a Polynesian land snail Partula suturalis vexillum. In: CREED, R. (ed.), Ecological Genetics and Evolution. Blackwell Scientific, Oxford, p. 51-64. CRAMPTON, H. E., 1916, Studies on the variation, distribution, and evolution of the genus Partula The species inhabiting Tahiti. Carnegie Institu- tion of Washington Publications, 228, 311 p., 24 pl. CRAMPTON, H. E., 1925, Studies on the variation, distribution, and evolution of the genus Partula. The species of the Mariana Islands, Guam and Saipan. Carnegie Institution of Washington Pub- lications, 228A, 116 p., 14 pl. CRAMPTON, H. E., 1932, Studies on the variation, distribution, and evolution of the genus Partula. The species inhabiting Moorea. Carnegie Institu- tion of Washington Publications, 410, 335 p., 24 pl. DIXON, W. J. & BROWN, M. B. (eds.), 1977, BMDP-77 Biomedical Computer Programs, P- Series. University of California, Berkeley, 880 p. DYMOND, J., 1975, K-Ar ages of Tahiti and Moorea, Society Islands, and implications for the hot-spot model. Geology, 3: 236-240. FORD, E. B., 1964, Ecological Genetics. Methuen, London, 335 p. ENVIRONMENT AND SHELL SHAPE IN PARTULA 35 GOULD, S. J. & LEWONTIN, R. C., 1979, The Spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist pro- gramme. Proceedings of the Royal Society of London, ser. B, 205: 581-598. GREEN, R. H., 1979, Sampling Design and Sta- tistical Methods for Environmental Biologists. Wiley-Interscience, New York, 257 p. HUXLEY, J. S., 1942, Evolution, the Modern Syn- thesis. Allen and Unwin, London, 645 p. JOHNSON, M. S., CLARKE, B. & MURRAY, J., 1977, Genetic variation and reproductive isola- tion in Partula. Evolution, 31: 116-126. JONES, J. S., LEITH, B. H. & RAWLINGS, P., 1977, Polymorphism in Cepaea: a problem with too many solutions? Annual Review of Ecology and Systematics, 8: 109-143. KONDO, Y., 1973, Samoana of the Society Islands (Pulmonata: Partulidae). Malacological Review, 6: 19-33. LIPTON, C. S. & MURRAY, J., 1979, Courtship of land snails of the genus Partula. Malacologia, 19: 129-146. LUNDMAN, B., 1947, Maps of the racial geography of some Partulae of the Society Islands based upon the material published by H. E. Crampton. Zoologiska Bidrag fran Uppsala, 25: 517-533. MURRAY, J. & CLARKE, B., 1966, The inheritance of polymorphic shell characters in Partula (Gas- tropoda). Genetics, 54: 1261-1277. MURRAY, J. & CLARKE, B., 1968a, Inheritance of shell size in Partula. Heredity, 23: 189-198. MURRAY, J. & CLARKE, B., 1968b, Partial repro- ductive isolation in the genus Partula (Gastro- poda) on Moorea. Evolution, 22: 684-698. MURRAY, J. & CLARKE, B., 1976a, Supergenes in polymorphic land snails 1. Heredity, 37: 253-269. MURRAY, J. & CLARKE, B., 1976b, Supergenes in polymorphic land snails Il. Partula suturalis. Heredity, 37: 271-282. MURRAY, J. & CLARKE, B., 1980, The genus Partula on Moorea: speciation in progress. Proceedings of the Royal Society of London, ser. B, 211: 83-117. PIMENTEL, R. A., 1979, Morphometrics. The Multivariate Analysis of Biological Data. Kendall/ Hunt, Dubuque, 276 p. POLLARD, E., 1975, Differences in shell thickness in adult H. pomatia L. from a number of locations in southern England. Oecologia, 21: 85-92. POOLE, R. W., 1974, An Introduction to Quanti- tative Ecology. McGraw-Hill, New York, 532 p. SCHMIDT-NIELSEN, K., TAYLOR, С. В. 4 SHKOLNIK, A., 1972, Desert snails: problems of survival. Symposia of the Zoological Society of London, 31: 1-13. SOKAL, В. R., 1978, Population differentiation: something new or more of the same? In: BRUSSARD, P. F. (ed.), Ecological Genetics: The Interface. Springer-Verlag, New York, p. 215-239. SOLEM, A., 1976, Endodontoid Land Snails from Pacific Islands. (Mollusca: Pulmonata: Sig- murethra) Part |. Family Endodontidae. Field Museum of Natural History, Chicago, Illinois, 508 Partula taeniata. p. YOM-TOV, Y., 1971, Body temperature and light reflectance in two desert snails. Proceedings of the Malacological Society of London, 39: 319- 326. MALACOLOGIA, 1982, 23(1): 3746 PLANKTONIC LARVAE OF NEW ENGLAND GASTROPODS. V. BITTIUM ALTERNATUM, TRIPHORA NIGROCINCTA, CERITHIOPSIS EMERSONI, LUNATIA HEROS AND CREPIDULA PLANA Catherine Thiriot-Quievreux! and Rudolf $. Scheltema2 ABSTRACT The veliger larvae of Bittium alternatum, Triphora nigrocincta, Cerithiopsis emersoni, Lunatia heros, and Crepidula plana from plankton at Woods Hole, Massachusetts, U.S.A., are described for the first time. Characteristics are given to distinguish veligers of Crepidula plana from those of C. fornicata. Photomicrographs taken with a scanning electron microscope show details of the larval shells, and the living veligers of two species were drawn extended. Key words: plankton; veliger larvae; New England; gastropods; Bittium; Triphora; Cerithiopsis; Lunatia; Crepidula. INTRODUCTION The larval stages of most prosobranch gas- tropods of the New England region (U.S.A.) are still unknown. Among the 42 species of prosobranch gastropods listed in the prelim- inary Woods Hole check-list (Russell-Hunter & Brown, 1964: 139-146), the mode of devel- opment for only 31 has been determined. Nineteen are known to have an indirect de- velopment with a planktonic larva, twelve de- velop directly without a planktonic stage, while the mode of development of the remain- ing eleven species is still unknown. Actually, the check-list of Russell-Hunter & Brown con- tains only the most common prosobranch species, less than half of those occurring in the Woods Hole region. Veligers of 13 among the 19 Woods Hole species known to have planktonic larvae have been described thus far in published ac- counts. These include Acmaea testudinalis (Muller, 1776) (Kessel, 1964); Anachis avara (Say, 1822) (A. H. Scheltema, 1969); Anachis translirata (Ravenel, 1861) (Scheltema & Scheltema, 1963, as A. avara; A. H. Schel- tema, 1969); Crepidula fornicata (Linne, 1758) (Werner, 1955); Lacuna vincta (Mon- tagu, 1803) (Lebour, 1937); Littorina littorea (Linne, 1758) (Thorson, 1946 and references therein); Nassarius (llyanassa?) obsoletus (Say, 1822) (R. S. Scheltema, 1962); Nas- sarius vibex (Say, 1822) (R. S. Scheltema, 1962); Nassarius trivittatus (Say, 1822) (Scheltema & Scheltema, 1965). In addition Station Zoologique, 06230 Villefranche-sur-Mer, France. there are illustrations of the larval shells of four other species, viz. Caecum pulchellum Stimpson, 1851, Sella adamsi (H. С. Lea, 1845), Cerithiopsis greeni (C. B. Adams, 1839), and Mitrella lunata (Say, 1826) (Thiriot-Quiévreux, 1980). The external morphology of the larvae of the six remaining species with planktonic development 15 known only from unpublished data or not at all. Thus, there remain even among the 42 common prosobranchs (including those for which the mode of development is still un- known) probably 15 to 20 undescribed plank- tonic veliger larvae. Although the prosobranch larvae of the Western North Atlantic are poorly known, the same is also true for other regions of the world (vide Robertson, 1974, for a summary of re- gional studies). Only in restricted localities such as Plymouth, England (Lebour, 1937) the western Baltic (Thorson, 1946), Banyuls- sur-Mer on the French Mediterranean coast (Thiriot-Quiévreux, 1969, 1972), the Bay of Naples (Richter & Thorson, 1975), the North Sea (Fretter & Pilkington, 1970), off the South Island (Otago) New Zealand (Pilkington, 1976) and in Kaneohe Bay, Hawaii (Taylor, 1975) are the gastropod veligers well enough known that illustrations, summaries and de- scriptive keys are available for identification of larvae. Yet an understanding of the popula- tion dynamics and ecology of species with plankonic stages requires first of all that the larvae be recognizable in the plankton. It is the goal of the present series of papers to 2Woods Hole Oceanographic Institution, Woods Hole, MA 02543, U.S.A. (37) 38 THIRIOT-QUIEVREUX AND SCHELTEMA describe veliger larvae of all common gastro- pods in the Woods Hole region so that they may be identified in plankton samples. METHODS Three methods can be used to identify and describe gastropod veligers. The first of these is to collect spawn from known females held in the laboratory and to grow their larvae through complete development (e.g. R. S. Scheltema, 1962; Christiansen, 1964; Struh- saker & Costlow, 1968). The second method is to collect larvae from the plankton and to rear them through metamorphosis to a juven- ile stage that can be identified (e.g. Lebour, 1937; Thorson, 1946; Thiriot-Quievreux, 1967a, 1967b, 1969). Finally a third method which has sometimes been successfully used is the comparison of unknown larval shells to the protoconch of identified juvenile or adult shells (e.g. R. S. Scheltema, 1971b, 1972; Thiriot-Quievreux, 1974, 1975; Thiriot-Quiev- reux & Rodriguez Babio, 1975). Previous con- tributions to this series have used the first method of rearing larvae from spawn. How- ever, there are certain gastropod species that will not readily deposit eggs in the laboratory either in adequate numbers or with sufficient regularity or whose egg capsules are un- known and hence cannot be collected in the field. Because of the difficulty in obtaining their spawn, most of the larvae newly de- scribed here were taken from the plankton and reared in the laboratory. Since the first paper in this series (R. S. Scheltema, 1962), scanning electron micro- scopy has revolutionized the study of gastro- pod larval shells (Fretter & Pilkington, 1971); Robertson, 1971; Thiriot-Quievreux, 1972, 1980; Richter & Thorson, 1975). In the present contribution we have included micro- graphs of the larval shell of each species dis- cussed using a Stereoscan 4 Scanning Elec- tron Microscope (Cambridge Scientific In- struments). The laboratory work was done in the Woods Hole region; samples were taken at the mouth of Hadley Harbor and the north- west entrance of Woods Hole channel. Plank- ton tows were taken with a net of 1/3-meter diameter and a mesh of 240 um. Samples were diluted to a convenient volume in a large finger bowl and subsamples were viewed un- der a binocular stereoscopic microscope (Wild M-5); all gastropod veligers were re- moved with a pipette. Larvae were placed in small 3-cm petri dishes and periodically fed Isochrysis or Phaeodactylum grown for this purpose in unialgal cultures. Shells to be used for scanning electron microscopy were kept up to 3 months in neutral 95% ethyl alcohol. Previous descriptions of larval species with- in the same genus often aid in the later identi- fication and description of closely related larvae. Each of the five species described here for the first time belong to genera for which other larval species are already known. This knowledge has aided in their identifica- tion. DESCRIPTIONS OF LARVAE Bittium alternatum (Say, 1822) (Figs. 1A-D, 4B) Shell: Dextral; 350 um long in fully-devel- oped veliger; 2-1/2 whorls; transparent, dark amber to brown; sutures of spire somewhat darker than rest of shell, columella dark brown; body whorl terminating in a rectangu- lar protuberance or “sinusigerous” lip (Figs. 1A, D; Fig. 4B) resembling that in Bittium re- ticulatum (Thorson, 1946; Thiriot-Quievreux, 1969, 1974). With the scanning electron mi- croscope the embryonic shell at the apex of the spire has a single smooth whorl (Fig. 1C); the second whorl has many small pustules particularly numerous near the sutures, and a single fine spiral granulated thread which decorates the lower part of the spire (Fig. 1A); on the body whorl of the larval shell are three — FIG. 1. Scanning electron micrographs of larval shells of Bittium alternatum (A-D) and Triphora nigrocincta (E-F). A) B. alternatum—larval shell showing spiral cords on body whorl. Scale = 100 um. B) B. alter- natum—shortly after metamorphosis. The edge of the sinusigerous lip is visible. Scale = 100 um. C) B. alternatum— apical view of larval shell. Scale = 40 um. D) B. alternatum—detail of body whorl showing sinusigerous lip. Scale = 40 um. E) T. nigrocincta— larval shell showing embryonic whorl and typical axial and spiral ribs. Scale = 100 um. F) T. nigrocincta—larval shell showing base of body whorl. Scale = 100 um. PLANKTONIC LARVAE OF NEW ENGLAND GASTROPODS V 39 40 THIRIOT-QUIEVREUX AND SCHELTEMA characteristic spiral cords (also seen with the optical microscope); the first is very marked and found on the upper half of the body whorl; the second and third are very fine and are on the lower part of the body whorl (Figs. 1A, D). After metamorphosis the postlarval shell growth completely surrounds the rectangular beak and two median spiral ridges are formed (Fig. 1B). Soft parts: Velum bilobed, lobes circular and colorless; foot with dark pigmentation; Operculum transparent; digestive gland greenish; two black eyes; two cephalic tenta- cles (Fig. 4B). Remarks: This species is said to be re- placed by Bittium varium (Pfeiffer, 1840) from Maryland and southward to Florida, the Gulf of Mexico and the West Indies to Brazil (Ab- bott, 1974, as Diastoma ). The larva of the southern species is described by Thiriot- Quievreux (1980) and differs in that it lacks the 3 spiral cords of B. alternatum. Triphora nigrocincta (C. B. Adams, 1839) (Figs: (EE) Shell: Sinistral, 600 um long when fully- developed; amber-brown; the specimens col- lected and examined had four whorls and ap- peared ready to metamorphose; embryonic portion of shell with 1 1/4 whorls, ornamented with closely spaced pustules which are denser at the periphery than in the center of the embryonic whorl; post-embryonic larval whorls with strong axial costae and one and then two spiral keels; the lower concave por- tion of the body whorl of the larval shell has a zone of axial threads followed by lines of tu- bercles which form concentric rings (Fig. 1F). The threads are most regular on the columella at the aperture. Soft parts: Velum colorless; pigmentation of foot black with white patch on base. Remarks: This species is apparently close- ly related to Triphora perversa (Linne, 1758) of Europe. Indeed, it sometimes is regarded as a subspecies of the latter (Johnson, 1934: 107; Abbott, 1974: 111). The larval shell also resembles superficially those described as Triphora perversa Thorson, 1946; Fretter & Pilkington, 1970; Thiriot-Quiévreux & Rod- riguez Babio, 1975; Richter & Thorson, 1975). However, the European Triphora perversa is apparently a complex of four distinct species; the species described here most closely re- sembles Triphora adversa (Bouchet & Guil- lemot, 1978: 350, figs. 15-16). The system- atics of the family Triphoridae in the Western Atlantic are clearly in need of further study. Larvae of the Triphoridae are known to be widely dispersed throughout the North and equatorial Atlantic and it is likely that at least some species will prove to be amphi-Atlantic in their geographic range (R. S. Scheltema, 197 1a). Cerithiopsis emersoni (C. B. Adams, 1838) (= subulatum “Montagu” of authors) (Figs. 2A-D) Shell: Dextral; length 700 ¡um in fully-devel- oped larvae; whorls ornamented with small tubercles arranged more or less in transverse rows (Figs. 2A, B); axial ribs at the end of the embryonic whorl. Later whorls lack transverse rows of tubercles but have instead regularly spaced opisthoclinal axial ribs; the periphery of the body whorl is characterized by a strong spiral thread (Figs. 2C, D). At the stage close to metamorphosis, the shell is dark red brown and has five whorls. Soft parts: Early planktonic larva has color- less bilobed velum, roundish foot without pigmentation, transparent operculum, and dark pigmentation on head. Late larva with velum slightly tetralobed and with fine edge of red-brown pigment. Small dots of dark pig- ment appear on head, the anterior part of body and foot. Remarks: Spawn was obtained from ani- mals held in the laboratory; in the natural en- vironment it is found on stones, sponges, and bryozoa. Egg masses convex, circular and gelatinous, 1.85 mm in diameter, smooth and very transparent at time of spawning; the mass rapidly becomes opaque from small de- bris which the adult places upon it. The eggs are resilient, yellow, inside a membrane and about 150 ит in diameter. An egg mass соп- tains about 50 eggs. The emergence of larvae from the egg mass occurs after 15 days at room temperature (ca. 20-23°C.). When the larva escapes it has a brown shell with 1-1/4 to 1-1/2 whorls (Fig. 2A). The species of Cerithiopsis here described differs from British species in details of orna- mentation (Fretter & Pilkington, 1970) but re- sembles the protoconch of a Pliocene speci- men of Cerithiopsis emersoni illustrated by Olsson & Harbison (1953, pl. 49, fig. 1). PLANKTONIC LARVAE OF NEW ENGLAND GASTROPODS V 41 FIG. 2. Scanning electron micrograph of larval shells of Cerithiopsis emersoni. A) Larval shell at the time of emergence from egg capsule, apical view. Scale = 40 um. B) Larval shell during early planktonic stage, apical view showing initial embryonic whorl. Scale = 40 um. C) Shell of intermediate-stage larva about midway in development. Scale = 100 um. D) Larval shell at time of metamorphosis. Scale = 100 um. Lunatia heros (Say, 1822) (Figs. 3A-C, 4A). Shell: Dextral; length at settlement 800 um; globose, holostomatous; attaining 1-1/2 whorls at metamorphosis (Fig. 3A) with dis- tinct umbilicus. There is an opaque spiral Operculum. Embryonic portion of shell has granular spiral lines which are more or less continuous and that appear as fine threads under light microscope; the remaining portion of the larvel shell is smooth. Soft parts: Velum is four-lobed (Fig. 4A). The lobes are slender and flexible when de- velopment is completed; they have a fine darkly-pigmented edge and elongated dark red pigment spots at their extremities. Dark pigmentation occurs above the mouth be- tween velar lobes. Remarks: This species can be readily dis- tinguished from other naticid larvae in the Woods Hole plankton by the spiral threads on the embryonic shell and the pigment spots on the end of each velar lobe. In all other com- monly occurring forms the embryonic shell does not have these spiral lines. The complete development has been pre- viously studied and described from larvae grown from egg collars deposited by snails in the laboratory (Dacy, 1965). The egg collars are described by Giglioli (1955). Development 42 THIRIOT-QUIEVREUX AND SCHELTEMA PLANKTONIC LARVAE OF NEW ENGLAND GASTROPODS V 43 in laboratory culture was completed in 31-54 days; the average shell length “from the edge of the aperture to the opposite side when the larva was lying on its side” was 792 um at the time of settlement (range 660-940 um). Lunatia heros larvae are remarkably similar to those of the European species Natica al- deri Forbes [= N. poliana Chiaje, N. pulchella Risso, N. nitida (Donovan), N. intermedia Philippi] illustrated by Thorson (1946: 218, fig. 130) and Fretter & Pilkington (1970: 17, figs. 20A-D). Crepidula fornicata (Linne, 1758) (Figs. ЗЕ, Е, 4F-H). Shell: About 760 um in length at settle- ment; colorless; transparent; the embryonic portion smooth. As growth progresses the apical end of embryonic shell becomes over- lain by the next succeeding whorl (Fig. 3F, Figs. 4F-H); shell more convex than Crepid- ula plana (see below); axial growth striae are distinct. Soft parts: Velum bilobed; dark pigment along edge and with a variable number of yel- low or yellow-green spots over its surface. Elongated yellow spots on mantle and bottom surface of foot; black pigmentation on head, foot and intestine; digestive gland translucent yellow. Remarks: The veliger of this species is de- scribed by Werner (1955) and by Fretter & Pilkington (1970, 1971). There are three spe- cies of Crepidula common in the New Eng- land region, viz. C. fornicata, C. plana and C. convexa. Only the former two have planktonic larvae; in Crepidula convexa development is completed within the mantle cavity. Since larvae of C. fornicata and C. plana are easily confused with one another both are consid- ered here even though the former has already been described elsewhere. Crepidula plana Say, 1822 (Figs. 3D, 4C-E). Shell: 650 um at settlement; quite trans- parent as in previous species; final whorl of completely developed larva is less swollen and consequently not so convex or globose as C. fornicata. Embryonic shell at apex of larval shell not overlain by the succeeding whorl as in C. fornicata (compare Figs. 4C-E with Figs. 4F-H). Growth striae of embryonic and also later larval shell numerous and dis- tinct. Teleoconch growth shows very fine spiral and axial striae forming a more regular pattern than in C. fornicata. Bandel (1975) described the shell of the newly emerged larva. Soft parts: Velum bilobed with darkly pig- mented edge; yellow spots on velum some- times absent; head and foot with yellow spots, intestine with diffuse dark pigment; kidney easily seen through shell owing to dark pig- mentation. Remarks: Larvae of the two species of Crepidula described above are most reliably distinguished from one another by their shells. The shell of C. fornicata is convex as a result of growth in which the first (i.e. embryonic) whorl is slightly overlain by the immediately succeeding whorl whereas the shell of C. plana is more nearly flat (i.e. not convex) ow- ing to the fact that the growth of the first post- embryonic whorl more or less is in the same plane as the embryonic whorl, i.e. the post- embryonic whorl does not conspicuously overgrow the embryonic shell. Less reliably the two are distinguished by pigmentation: yellow spots on mantle and a dark pigmented intestine in C. fornicata; no pigment on the mantle or intestine and dark kidney in C. plana. ACKNOWLEDGEMENTS We acknowledge with thanks the assist- ance of Jan A. Pechenik in the collection of field samples. Algal cultures for the work were maintained by Isabelle Williams to whom we are much obliged. Dr. Robert Robertson of the Academy of Natural Sciences of Phila- delphia was helpful in a number of ways and allowed us free use of the collections in his << FIG. 3. Scanning electron micrographs of Lunatia heros (A-C), Crepidula plana (D) and Crepidula fornicata (E-F). A) L. heros—intermediate-stage larval shell showing spiral threads on embryonic whorl, a character distinguishing this species from others of the genus in the Woods Hole region. Scale = 100 um. See also Fig. 4A. В) L. heros—larval shell at stage approximately the same as A, profile view showing characteristic operculum. Scale = 100 um. C) L. heros—detail of embryonic whorl showing beaded nature of spiral threads. Scale = 40 um. D) С. plana—larval shell at time of metamorphosis. Scale = 200 ит. See also Figs. 4C-E. E) C. fornicata—larval shell at time of metamorphosis. Scale = 200 ит. See also Figs. 4F-H. F) C. fornicata—larval shell at intermediate stage of development—profile view. Scale = 100 um. 44 THIRIOT-QUIEVREUX AND SCHELTEMA FIG. 4. A) Swimming veliger of Lunatia heros at an intermediate stage; the propodium is only partly devel- oped. The velar lobes have not reached their definitive proportions; each will grow longer and the pigmented spots will increase in size and intensity before settlement. B) Swimming veliger of Bittium alternatum. The velum lacks coloration; the mesopodium is darkly pigmented. The shell terminates in a broad rectangular protuberance, the sinusigerous lip. The larva shown is at a stage shortly before metamorphosis; develop- ment of the foot is almost complete. C) Crepidula plana—shell fifteen days after release from female and just after settlement; compare with shell of C. fornicata, F below. D) C. plana—outline of settled larva showing growth of second whorl; compare with C. fornicata, H below. E) C. plana—settled specimen showing coiling; compare with C. fornicata below. Post-larval growth results in “flat” adult form through coiling in a single plane. F) Crepidula fornicata—shell thirty-six days after release from female and recently settled, showing apex; compare to C above. G) C. fornicata—twenty-nine days after release of veliger from female, showing embryonic whorl partly overlain by succeeding whorl; see also F and H. The more or less helical growth results in a higher more nearly convex adult shell than in C. plana; compare E and G. H) C. fornicata— settled larva showing growth of spire; overgrowth of succeeding whorl shown by arrow; compare with D above. Scale = 1 mm, applies to Figs. C-H only. PLANKTONIC LARVAE OF NEW ENGLAND GASTROPODS V 45 care. Alison Stone Ament kindly provided us with larval shells of Crepidula fornicata and C. plana grown in culture in connection with her own research. These reared larvae confirmed differences between the two species found in plankton samples. A Stereoscan 4 Scanning Electron Microscope (Cambridge Scientific Instruments) was made available by the Cen- tre Oceanologique de Bretagne at Brest, France. This research was supported in part by a grant from the U.S. National Science Foundation OCE73-00439A02. This is contri- bution number 4098 of the Woods Hole Oceanographic Institution. LITERATURE CITED ABBOTT, R. T., 1974, American Seashells. Ed. 2, Van Nostrand Reinhold, New York, 663 p. BANDEL, K., 1975, Embryonalgehause karibischer Meso-und Neogastropoden (Mollusca). Abhand- lungen der mathematisch-naturwissenschaft- lichen Klasse, Akademie der Wissenschaften und der Literatur, Mainz, Jahrgang 1975(1): 1- 133: 2 Pill: BOUCHET Р. & GUILEEMOT; H., 1978, The Tri- phora perversa-complex in western Europe. Journal of Molluscan Studies, 44: 344-356. CHRISTIANSEN, M. E., 1964, Some observations on the larval stages of the gastropod, Nassarius pygmaeus (Lamarck). Pubblicazioni della Sta- zione Zoologica di Napoli, 34: 1-8. DACY, D., 1965, The larval development of Lunatia heros Say. Unpubl. M.S. thesis, Queens Univer- sity, Kingston, Ontario, 26 p., 23 figs. FRETTER, V. & PILKINGTON, M. C., 1970, Proso- branchia—Veliger larvae of Taenioglossa and Stenoglossa zooplankton. Fiches d'Identifica- tion du Zooplancton (Conseil Permanent Inter- national pour l'exploration de la Mer), sheets 129-132, 26 p. FRETTER, V., & PILKINGTON, M. C., 1971, The larval shell of some prosobranch gastropods. Journal of the Marine Biological Association of the United Kingdom, 51: 49-62. GIGLIOLI, M. E. C., 1955, The egg masses of the Naticidae (Gastropoda). Journal of the Fisheries Research Board of Canada, 12: 287-327. JOHNSON, C. W., 1934, List of marine Mollusca of the Atlantic coast from Labrador to Texas. Pro- ceedings of the Boston Society of Natural His- tory, 40: 1-204. KESSEL, M. M., 1964, Reproduction and larval de- velopment of Acmaea testudinalis (Muller). Bi- ological Bulletin, 127: 294-308. LEBOUR, M. V., 1937, The eggs and larvae of the British prosobranchs with special reference to those living in the plankton. Journal of the Ma- rine Biological Association of the United King- dom, 22: 105-166. OLSSON, A. A. & HARBISON, A., 1953, Pliocene Mollusca of southern Florida, with special refer- ence to those from St. Petersburg. The Academy of Natural Sciences of Philadelphia Mono- graphs 8, 457 p. PILKINGTON, M. C., 1976, Description of veliger larvae of monotocardian gastropods occurring in Otago plankton hauls. Journal of Molluscan Studies, 42: 337-360. RICHTER, G. & THORSON, G., 1975, Pelagische Prosobranchier-larven des Golfes von Neapel. Ophelia, 13: 109-185. ROBERTSON, R., 1971, Scanning electron micro- scopy of planktonic larval marine gastropod shells. Veliger, 14: 1-12. ROBERTSON, R., 1974, Marine prosobranch gas- tropods: Larval studies and systematics. Thalas- sia Jugoslavica, 10: 213-238. RUSSELL-HUNTER, W. & BROWN, S. C., 1964, Phylum Mollusca. XIV. In Smith, В. 1. (compiler), Keys to marine invertebrates of the Woods Hole Region—A Manual for the identification of the more common marine invertebrates. Marine Bio- logical Laboratory, Woods Hole, x + 208 p. SCHELTEMA, A. H., 1969, Pelagic larvae of New England gastropods. IV. Anachis translirata and Anachis avara (Columbellidae, Prosobranchia). Vie et Milieu, ser. A: Biol. Mar., 20(1-A): 94-104. SCHELTEMA, R. S., 1962, Pelagic larvae of New England intertidal gastropods. |. Nassarius obso- letus (Say) and Nassarius vibex (Say). Transac- tions of the American Microscopical Society, 81: 1-11. SCHELTEMA, R. S., 1971a, The dispersal of the larvae of shoal-water benthic invertebrate spe- cies over long distances by ocean currents. /n Crisp, D. (ed.), Fourth European Marine Biologi- cal Symposium p. 7-28, Cambridge University Press, Cambridge. SCHELTEMA, R. S., 1971b, Larval dispersal as a means of genetic exchange between geograph- ically separated populations of shallow water benthic marine gastropods. Biological Bulletin, 140: 284-322. SCHELTEMA, R. S., 1972, Eastward and westward dispersal across the tropical Atlantic Ocean by larvae belonging to the genus Bursa (Proso- branchia, Mesogastropoda, Bursidae). /nterna- tionale Revue der gesamten Hydrobiologie, 57: 863-873. SCHELTEMA, R. S. & SCHELTEMA, A. H., 1963, Pelagic larvae of New England intertidal gastro- pods. Il. Anachis avara. Hydrobiologia, 22: 85- 91. SCHELTEMA, R. S. & SCHELTEMA, A. H., 1965, Pelagic larvae of New England intertidal gastro- pods. Ill. Nassarius trivittatus. Hydrobiologia, 25: 321-329. STRUHSAKER, J. W. & COSTLOW, J. D., 1968, Larval development of Littorina picta (Proso- branchia, Mesogastropoda) reared in the labora- tory. Proceedings of the Malacological Society of London, 38: 153-160. 46 THIRIOT-QUIEVREUX AND SCHELTEMA TAYLOR, J. B., 1975, Planktonic prosobranch veligers of Kaneohe Bay. Unpubl. Ph.D. disser- tation, University of Hawaii, хш + 599 р. THIRIOT-QUIEVREUX, C., 1967a, Observations sur le developpement larvaire et postlarvaire de Simnia spelta Linne (Gasteropode, Cypraeidae). Vie et Milieu, 18: 143-151. THIRIOT-QUIEVREUX, C., 1967b, Descriptions de quelques veligères planctoniques de gastero- podes. Vie et Milieu, 18: 303-315. THIRIOT-QUIEVREUX, C., 1969, Caracteristiques morphologiques des veligères planctoniques de gasteropodes de la region de Banyuls-sur-Mer. Vie et Milieu, ser. B: Oceanographie, 20(2-B): 333-366. _ THIRIOT-QUIEVREUX, C., 1972, Microstructures de coquilles larvaires de prosobranchs au micro- scope electronique à balayage. Archives de Zoologie experimentale et generale, 113: 553- 564. THIRIOT-QUIEVREUX, C., 1974, Gasteropodes de la region de Roscoff. Etude particulière de la protoconque. Cahiers de Biologie Marine, 15: 531-549. _ THIRIOT-QUIEVREUX, C., 1975, Pyramidellidae et Retusidae de la region de Roscoff. Etude par- ticuliere des protoconques de quelques espè- ces. Cahiers de Biologie Marine, 16: 83-96. THIRIOT-QUIEVREUX, C., 1980, Identification of some planktonic prosobranch larvae off Beau- fort, North Carolina. Veliger, 23: 1-9. THIRIOT-QUIEVREUX, C. & RODRIGUEZ BABIO, C., 1975, Etude des protoconques de quelgues Prosobranches de la region de Roscoff. Cahiers de Biologie Marine, 16: 135-148. THORSON, G., 1946, Reproduction and larval de- velopment of Danish marine bottom inverte- brates with special reference to the planktonic larvae of the Sound (@resund). Meddelelser fra Kommissionen for Danmarks Fiskeri- og Hav- unders@gelser, ser. Plankton, 4: 1-523. WERNER, B., 1955, Uber die Anatomie, die Ent- wicklung und Biologie des Veligers und der Veli- concha von Crepidula fornicata. L. Helgolander wissenschaftliche Meeresuntersuchungen, 5: 169-217. MALACOLOGIA, 1982, 23(1): 47-54 REPRODUCTION IN A PERIPHERAL POPULATION OF CYRENOIDA FLORIDANA (BIVALVIA: CYRENOIDIDAE) Pieter W. Kat Department of Earth and Planetary Sciences, The Johns Hopkins University, Baltimore, Maryland 21218, U.S.A. ABSTRACT Reproductive physiology of a marginal population of the bivalve Cyrenoida floridana (Dall) was studied in an attempt to determine the extent of adaptation to temperature conditions at the geographical range limit. Results of the study indicate that the marginal population had extremely low spawning suc- cess, and that a second reproductive cycle was initiated in the fall but never completed. The second reproductive cycle is completed by populations in the ancestral parts of the range where winters are less severe. The disruptive effect of immigrant genes on adapted complexes are postulated to be responsible for the lack of reproductive adaptation exhibited by the population. Key words: reproduction; Bivalvia; Cyrenoida; peripheral populations; geographic range ex- pansions. INTRODUCTION Increase in a species’ range may be medi- ated by a number of factors. First, the species may be introduced into an open habitat by unpredictable events. In this category one can include small-scale, random colonizations of islands (or habitat islands) by species propa- gules (MacArthur & Wilson, 1967; Diamond, 1969; Simberloff & Wilson, 1969, 1970; Wil- liams, 1972; Simberloff, 1976) or large-scale, not necessarily random, invasions mediated by plate tectonics (Simpson, 1965; Mayr, 1970; Patterson & Pascual, 1972; Davis, 1979). Second, a species may expand its range over contiguous territory in response to ameliorating environmental conditions. This category would include range fluctuations brought about by periods of glaciation (Webb, 1969; Stehli & Wells, 1971; Clarke, 1973). Third, ranges can be increased by accidental or purposeful introductions by man (Elton, 1958). Hesse et al. (1951), however, point out that extension of the distributional range is also influenced by certain intrinsic attributes of the species, including genetic variability, physiological tolerance, and mode of repro- duction and dispersal. Such intrinsic limita- tions may restrict the ability of a species to expand its distributional range regardless of Opportunities presented. Grant & Antonovics (1978) mention that marginal populations provide suitable units for the determination of processes involved in range extensions. In the same vein, marginal populations allow for the identification of fac- tors limiting expansions of range. These fac- tors may be related, among others, to physio- logical tolerance, geographic barriers, re- source availability and biological interactions such as competition and predation (Crisp & Southward, 1958; Kinne, 1970; Jackson, 1972, 1974; MacArthur, 1972). Once the presence or absence of such factors has been established, it may be possible to pre- dict the extent of future range expansions. For instance, increase in range within ecological time can be predicted for a species whose peripheral populations show no evidence of physiological stress—such as drastic popula- tion fluctuations, reproductive difficulties and growth stunting (Stearns & Sage, 1980) or biotic interference such as competitive or predatory limitation (Crisp & Southward, 1958). Reproductive data gathered at a range margin can be used as a sensitive measure to ascertain the extent of adaptation of the pe- ripheral population to local conditions. Timing of initiation of gametogenesis and spawning together with percentage of individuals spawned in the peripheral population can be compared with similar data from central popu- lations to determine if the peripheral popula- tion is responding to novel environmental cues or whether the response remains essen- tially similar to that of central populations. This study focuses on certain maladaptive (47) 48 aspects of the reproductive cycle of a periph- eral population of the intertidal bivalve Cyre- noida floridana (Dall) located on Canary Creek Marsh near Lewes, Delaware, U.S.A. Following an initial range description by Dall (1896), in which the bivalve was limited to Florida and southern Georgia, several range expansions have been reported. These re- ports trace the progress of the species from its ancestral range through the Chesapeake and Delaware Bays (Johnson, 1934; Wass, 1972: Leathem et al., 1976). Also, J. P. E. Morrison (National Museum of Natural His- tory) collected the bivalve from a number of localities along the east coast from 1952 to 1954, including its northernmost limit in Cum- berland County, New Jersey (unpubl. data). This relatively rapid range expansion seems coincident with, and can probably be attri- buted to, the construction of the Intracoastal Waterway. This series of canals probably also provides avenues of dispersal for juveniles to peripheral areas subsequent to local extinc- tions caused by periods of severe climate. METHODS The study site, Canary Creek Marsh, is lo- cated northwest of the town of Lewes, Dela- ware, and has been described by Gallagher (1971), Sullivan (1971) and Elliot (1973). The largest percentage of the marsh surface is covered by halophytic plants, and Cyrenoida floridana is found among the roots of this vegetation, buried to a depth of about 1 cm. Highest densities of the bivalve occurred at a level of 1.4 m above Mean Low Water in soils covered with a thin layer of filamentous algae which improved moisture retention (Kat, 1978). Sampling of the marsh surface at locations with highest bivalve densities was conducted weekly from June to December, followed by a bi-weekly schedule from January to April. The Samples were washed over two sieves with 2.4 and 0.4mm meshes, and bivalves thus obtained were relaxed with magnesium chlo- ride and fixed in Bouin's solution. This fixative is mildly acidic and dissolved the shell. Spe- cimens measuring from 3.5 to 4.5mm were embedded whole in paraffin, sectioned at 6- 8 um and stained with Harris’ hematoxylin and Eosin Y (Humason, 1972). Attempts were made to find populations north of Lewes at locations described by Morrison (unpubl. data). No cyrenoidas were KAT found at any of these locations, even though Morrison's field notes indicated relatively high population densities. RESULTS Cyrenoida floridana is a simultaneous hermaphrodite: both types of gametes are produced concurrently during the entire re- productive life of an individual. Male and fe- male sex cells are found in the same gonadal follicle, although there is a suggestion that spermatogenic cells are more numerous in the dorsal regions of the gonads. Four major gonadal stages can be recog- nized in C. floridana: gametogenic, mature, spawned, and resorptive (Table 1). During early stages of gametogenesis the follicle walls are thick and compact, and the interior of the follicles present a disorganized picture, containing oogonia, spermatogonia, partition cells, pycnotic cells and amoebocytes. The follicle walls become progressively thinner as gametogenesis continues and the gonad expands in volume. Pycnotic cells and amoe- bocytes eventually disappear. Both oogenesis and spermatogenesis are gradual processes. The progressive stages of spermatogenesis can be recognized by the successive predominance of spermatocytes, spermatids, and finally spermatozoa. Similar- ly, a number of characteristic changes occur in the appearance of the oocytes as they ma- ture. At first, the nucleus is small, ranging TABLE 1. Reproductive state of Cyrenoida flori- dana on Canary Creek Marsh, April, 1976 to March, 1977 (n = 125). Individuals classified according to predominant gonadal condition. Month Resorptive of or Gameto- Year _ indifferent genic Mature Spawned J F + + M + A + M J + J af + + A + + S + + (9) > + N + D + REPRODUCTION IN A PERIPHERAL CYRENOIDA POPULATION 49 from 8 to 12 um in diameter. The nuclei of small oocytes stain darker than their cyto- plasm (Fig. 1). As the oocyte matures, the nucleus expands in diameter to measure about 25 um, and the chromatin in the nu- cleus disperses so that the cytoplasm stains darker than the nucleus (Fig. 2). The mature state of oogenesis occurs when large numbers of oocytes project into, or lie free within the lumina. Spermatogenic regions contain spermatids superimposed by clusters of spermatozoa. The follicle walls are gener- ally thin, and follicle cell nuclei widely dis- persed. The spawned stage is easily recognized by the presence of spawned ova in the intrala- mellar brood pouches. Total spawning was not observed in any individual, as many gametes were retained within the follicles af- ter spawning had occurred. Spawning to any degree was only observed in 9% (n = 65) of the bivalves collected during July, August and September. The method of fertilization was not directly observed in C. floridana, but it is probable that, upon detection of spermatozoa in the water column, oocytes are cleared from the follicles into the lamellar brood chambers where fertilization takes place. Fertilization may also be partially internal. This was evi- denced by the observation of embryos devel- oping within the gonadal follicles, which indi- cates incomplete spawning of a fertilized egg. Whether this internal fertilization results from self-fertilization is unknown; however, a large number of spermatozoa which had under- gone the acrosome reaction in response to mature oocytes (Dan & Wada, 1955; Pop- ham, 1974) were found within the follicles from early June to the middle of September. The resorptive stage is characterized by changes in the appearance of the gametes as well as the presence of amoebocytes within the follicles. Oocytes undergo autolysis during which the nucleus becomes almost transpar- ent, followed by convolution and rupture of the nuclear membrane. Those oocytes attached to the follicle wall also show signs of cytolysis, during which light areas appear in the yolk. As these changes take place, amoebocytes in- vade the follicles and phagocytize the remain- ing gametes. The reproductive cycle described above was repeated twice during the period of ob- FIG. 1. Immature oocytes with nuclei which stain darker than the yolk. Note immature oocytes and sper- matocytes attached to the follicle walls (scale bar 100 um). FIG. 2. Mature oocytes with nuclei which stain lighter than the yolk (scale bar = 100 um). servation. The first cycle began during the late winter, continued through spring and summer, and ended with resorption in early fall. The second cycle began with gametogenesis dur- ing the middle of fall, but maturity of the gonadal products was not reached, and re- sorption took place during the late fall. Spawning thus occurred only once, even though two cycles were initiated. The second gametogenic cycle is of con- siderable interest because of its lack of com- pletion. Most oocytes produced were irregular (Fig. 3), and cytolysis was evident in the larger oocytes after a short period of devel- opment. The follicle walls remained thick. DISCUSSION The low success of spawning observed in the peripheral population of Cyrenoida flori- dana (9%) could have been caused by several factors such as low density of indi- viduals (Mayr, 1970; Soule, 1973), unreliable temperature cues (Kinne, 1963; Korringa, 1957; Van der Schalie and Berry, 1973), an adverse effect of lowered temperatures on normal gametogenesis, and the possibility that self-fertilization caused reduced fertility (Paraense, 1959; Gee & Williams, 1965). More interesting is the observed interrup- tion of the second gametogenic cycle, which constitutes strong evidence of reproductive maladaptedness of the population. Reproduc- tive physiology of C. floridana has not been reported from other parts of its range, but examination of individuals collected near Brunswick, Georgia, in December and St. Marks, Florida, in July revealed developing juveniles in the demibranchs indicating com- pletion of the first and second reproductive cycles in central populations (Table 1). In addition, Subrahmanyam et al., (1976) men- tion that recruitment of juveniles takes place twice on Florida marshes. From these obser- vations it can be concluded that the bivalve spawns twice in ancestral portions of the range, while it is limited to a single spawning in Delaware: a rather common observation for wide-ranging species (e.g. Stauber, 1950; Loosanoff & Nomejko, 1952; Butler, 1955; Prosser, 1955; Korringa, 1957; Kinne, 1970; Jackson, 1974; Frank, 1975). The difference between this study and previous studies of REPRODUCTION IN A PERIPHERAL CYRENOIDA POPULATION 51 FIG. 3. Irregular oocytes produced during the interrupted second reproductive cycle. The oocytes both show signs of karyolysis during which the nuclear membrane becomes convoluted (right), then ruptures, followed by disappearance of the nucleolus (left) (Scale bar = 20 um). latitudinal variation in reproduction lies in the observation that a second reproductive cycle is still initiated by populations of C. floridana at the range periphery, but not completed. The activity of phagocytic amoebocytes de- termines if wastage of reproductive energy is involved in the second reproductive cycle, and hence if it should be selected against. Purchon (1968) and Yonge (1928) have indi- cated that amoebocytes may play a role in digestive processes by transferring nutrients from the digestive tract to epithelial cells. While histological sections of the resorptive stage revealed high numbers of amoebocytes outside the follicle walls in addition to those engaged in phagocytosis, there was no evi- dence of a concentration of these amoebo- cytes within the digestive system or among epithelial cells. While reproductive material was phagocytized rather than voided through the gonadal ducts as in the oyster, for in- stance (Galtsoff, 1964), which indicates that some redistribution of energy is likely, wast- age of this reproductive energy is neverthe- less indicated, as efficiency of transfer along a series of steps is low. During the severe winter of 1977, the popu- lation of C. floridana at Canary Creek suffered heavy mortality. A survey of the population taken before and after the winter showed that heaviest relative mortality occurred among the youngest age classes, and that population density decreased from an overall average across the marsh of 285 individuals/m” to 8.3 individuals/m”. Such high mortalities seem common among peripheral populations (Welch, 1968; Gallagher & Wells, 1969), and the absence of populations of the bivalve from areas sampled by Morrison in 1954 suggest that such fluctuations might be repetitive. Peripheral and marginal populations sub- ject to frequent extinctions can be re-estab- lished from two main sources (Slatkin, 1977). Colonists can either be drawn from the central Sa population at random (migrant-pool model) or originate from refugia near the periphery of the range (propagule-pool model). Marginal populations formed by the migrant-pool model can be expected to contain the same geno- types as found in central populations, and, as such, be ill-adapted to range periphery condi- tions during severe years assuming limited phenotypic compensation. Marginal popula- tions formed by the propagule-pool model can be expected to be better adapted, since their phenotypes have been subjected to physical selection pressures similar to those at the range margin (well-defined clines are associ- ated with this model). In the case of C. flori- dana, the migrant-pool model seems to best explain the lack of reproductive adaptation to local conditions. Survivors of catastrophic selection are probably extremely important in the formation of populations exhibiting physiological adap- tations to local conditions (Miller, 1956; Lewis, 1962; Harper, 1977). However, survivors may face serious problems in finding mates, espe- cially in sessile species. It has been pointed out that self-fertilization is a probable cause for reduced spawning success, and can con- sequently be expected to be selected against for this and other reasons in central popula- tions (Paraense, 1959; Moore & Lewis, 1965; Lewis, 1966; Antonovics, 1968). However, as Ghiselin (1974) points out, it is better to self- fertilize than not to fertilize at all. When popu- lation densities of adapted individuals (de- rived from the survivor group) build up to levels where cross-fertilization again be- comes possible, those individuals which pre- ferentially cross-fertilize gain a selective ad- vantage by virtue of their higher reproductive success. If, however, this population is flooded by individuals derived from the central population during climatically benign periods, cross-fertilization may reduce fitness by dis- rupting adapted gene complexes (Levin, 1970; Emlen, 1978). The extent of the distributional range of a species with little phenotypic (in this case re- productive) compensation is dependent on its capability to differentiate into locally adapted populations (Bradshaw, 1965). If such differ- entiation is prevented by gene flow and fluc- tuating selective regimes which may be sev- ere (directional) during some years and mild (non-directional) during others, such species will generally be restricted in distribution, and characterized by fluctuating range boundaries comprised of low-fitness populations. Clines KAT in unisolated species of this kind will tend to be indistinct, both due to the effects of gene flow and selection gradients which migrate over the geographic range of the organism. ACKNOWLEDGEMENTS | thank Drs. Steven Stanley, George Davis and Sally Woodin, Phil Signor, Scott Lidgard, Robert Hershler and two anonymous review- ers for comments on various drafts of this paper. Most data are drawn from a Masters thesis completed at the University of Dela- ware. Funding for subsequent research was provided in part by a Shell Foundation grant to the Department of Earth and Planetary Sciences. REFERENCES CITED ANTONOVICS, J., 1968, Evolution in closely ad- jacent plant populations. V. Evolution of self-fer- tility. Heredity, 23: 219-238. ANTONOVICS, J., 1971, The effects of a hetero- geneous environment on the genetics of natural populations. American Scientist, 59: 593-599. BRADSHAW, A. D., 1965, Evolutionary signifi- cance of phenotypic plasticity in plants. Ad- vances in Genetics, 13: 115-155. BUTLER, P. A., 1955, Reproductive cycles of na- tive and transplanted oysters. Proceedings of the National Shellfisheries Association, 46: 75. CLARKE, A. H., 1973, The freshwater molluscs of the Canadian interior basin. Malacologia, 13: 1- 509. CRISP, D. M. & SOUTHWARD, A. J., 1958, The distribution of intertidal organisms along the coasts of the English Channel. Journal of the Marine Biological Association of the United Kingdom, 37: 157-208. DALL, W. H., 1896, On the American species of Cyrenoida. Nautilus, 10: 52. DAN, J. C. & WADA, S. K., 1955, Studies on the acrosome. IV. The acrosome reaction in bivalve spermatozoa. Biological Bulletin, 109: 40-55. DAVIS, G. M., 1979, The origin and evolution of the gastropod family Pomatiopsidae, with emphasis on the Mekong River Triculinae. Academy of Na- tural Sciences of Philadelphia. Monograph, 20: 120 p. DIAMOND, J. M., 1969, Avifaunal equilibria and species turnover rates on the Channel Islands of California. Proceedings of the National Acad- emy of Sciences, 64: 57-63. DICKINSON, H. & ANTONOVICS, J., 1973, Theo- retical considerations of sympatric divergence. American Naturalist, 107: 256-275. ELLIOT, G., 1973, Great Marsh, Lewes, Delaware: The physiology, classification and geological REPRODUCTION IN A PERIPHERAL CYRENOIDA POPULATION 53 history of a coastal marsh. MS thesis, University of Delaware, 106 p. ELTON, C. S., 1958, The Ecology of Invasions by Animals and plants. Methuen, London. 146 p. EMLEN, J. T., 1978, Density anomalies and regu- latory mechanisms in land bird populations on the Florida peninsula. American Naturalist, 112: 265-286. FRANK, P. W., 1975, Latitudinal variation in the life-history features of the Black Turban Snail Tegula funebralis (Prosobranchia: Trochidae). Marine Biology, 31: 181-192. GALLAGHER, J. L., 1971, Algal productivity and some aspects of the ecological physiology of the edaphic communities of Canary Creek tidal marsh. PhD thesis, University of Delaware, 120 p. GALLAGHER, J. L. & WELLS, H. W., 1969, North- ern range extension and winter mortality of Rangia cuneata. Nautilus, 83: 22-25. GALTSOFF, P. S., 1964, The American Oyster, Crassostrea virginica Gmelin. United States Fish and Wildlife Service Fisheries Bulletin, 64: 1-480. GEE, J. M. & WILLIAMS, G. B., 1965, Self and cross-fertilization in Spirorbis borealis and S. pagenstecheri. Journal of the Marine Biological Association of the United Kingdom, 45: 275- 285. GHISELIN, M. T., 1974, The Economy of Nature and the Evolution of Sex. University of California, Berkeley, California, 346 p. GOULD, S. J. & ELDREDGE, N., 1977, Punctuated equilibria: the tempo and mode of evolution re- considered. Paleobiology, 3: 115-151. GRANT, M. C. & ANTONOVICS, J., 1978, Biology of ecologically marginal populations of Anthox- anthum odoratum. |. Phenetics and dynamics. Evolution, 32: 822-838. HARPER, J. L., 1977, Population Biology of Plants. Academic Press, London, 836 p. ESSE. А:, ALLEE; \\/.. С. & SCHMIDT, К. P:, 1951, Ecological Animal Geography. Wiley In- terscience, New York, 715 p. HUMASON, G. L., 1972, Animal Tissue Tech- niques. Freeman, San Francisco, 641 p. JACKSON, J. В. C., 1972, The ecology of molluscs of Thalassia communities, Jamaica, West Indies. |. Molluscan population variability along an en- vironmental gradient. Marine Biology, 14: 304— 337. JACKSON, J. B. C., 1974, Biogeographic conse- quences of eurytopy and stenotopy among ma- rine bivalves and their evolutionary significance. American Naturalist, 108: 541-560. JOHNSON, C. W., 1934, List of the marine Mol- lusca of the Atlantic coast from Labrador to Texas. Proceedings of the Boston Society of Natural History, 40: 1-204. KAT, P. W., 1978, The functional morphology and ecology of Cyenoida floridana Dall (Bivalvia: Cyrenoididae). MS thesis, University of Dela- ware, 182 p. KECK, R. T., MAURER, D. & LIND, H., 1975, A comparative study of the hard clam gonadal cy- cle. Biological Bulletin, 148: 243-258. KINNE, O., 1963, The effects of temperature and salinity on brackish water animals. |. Tempera- ture. Annual Review of Oceanography and Ma- rine Biology, 1: 301-340. KINNE, O., 1970, Temperature, animals, inverte- brates. /n KINNE, O. Ed. Marine Ecology Part |. Wiley Interscience, London, 326 p. KORRINGA, P., 1957, Water temperature and breeding throughout the geographical range of Ostrea edulis. Annals of Biology, 33: 1-17. LEATHEM, W., KINNER, P. & MAURER, D., 1976, Northern range extension of the Florida Marsh Clam Cyrenoida floridana (Superfamily Cyre- noidacea). Nautilus, 90: 93-94. LEVIN, D., 1970, Developmental instability and evolution in peripheral isolates. American Nat- uralist, 104: 343-353. LEWIS, H., 1962, Catastrophic selection as a factor in speciation. Evolution, 16: 257-271. LEWIS, H., 1966, Speciation in flowering plants. Science, 152: 167-172. LOOSANOFF, V. L. & NOMEJKO, C. A., 1952, Existence of physiologically different races of oysters, Crassostrea virginica. Biological Bulle- tin, 101: 151-156. MACARTHUR, R. H., 1972, Geographical Ecology. Patterns in the Distribution of Species. Harper & Row, New York, 269 p. MACARTHUR, R. H. & WILSON, E. O., 1967, The Theory of Island Biogeography. Princeton Uni- versity, Princeton, New Jersey, 203 p. MAYR, E., 1970, Populations, Species and Evolu- tion. Harvard University, Cambridge, Mass., 453 p. MCNEILLY, T. & ANTONOVICS, J., 1968, Evolu- tion in closely adjacent plant populations. IV. Barriers to gene flow. Heredity, 23: 205-218. MILLER, A. H., 1956, Ecologic factors that acceler- ate formation of races and species of terrestial vertebrates. Evolution, 10: 262-277. MOORE, D. M. & LEWIS, H., 1965, The evolution of self pollination in Clarkia xantiana. Evolution, 19: 104-114. PARAENSE, W. L., 1959, One-sided reproductive isolation between geographically remote popu- lations of a planorbid snail. American Naturalist, 93: 93-101. PATTERSON, B. & PASCUAL, R., 1972, The fossil mammal fauna of South America. /n: Evolution, Mammals and Southern Continents. KEAST, A., ERK, F. C., and GLASS, B., Eds. State Univer- sity of New York, New York. p. 247-309. POPHAM, J. D., 1974, The acrosome reaction in the sperm of the shipworm, Bankia australis Colman (Bivalvia: Mollusca). Cellular and Tissue Research, 151: 93-101. PROSSER, C. L., 1955, Physiological variation in animals. Biological Reviews, 30: 229-262. PURCHON, R. D., 1968, The Biology of the Mol- lusca. Pergamon, Oxford. 560 p. 54 SCHALIE, H. VAN DER & BERRY, E. G., 1973, Effects of temperature on growth and reproduc- tion of aquatic snails. Ecological Research Se- ries, Office of Research and Monitoring, U.S. Environmental Protection Agency, 165 p. SIMBERLOFF, D. S., 1976, Experimental zoogeo- graphy of islands: effects of island size. Ecology, 57: 629-648. SIMBERLOFF, D. S. & WILSON, E. O., 1969, Ex- perimental zoogeography of islands: the coloni- zation of empty islands. Ecology, 50: 278-296. SIMBERLOFF, D. S. & WILSON, E. O., 1970, Ex- perimental zoogeography of islands: a two-year record of colonization. Ecology, 51: 934-937. SIMPSON, G. G., 1965, The Geography of Evolu- tion, Chilton, Philadelphia. 249 p. SLATKIN, M., 1977, Gene flow and genetic drift ina species subject to frequent local extinctions. Theoretical Population Biology, 12: 253-262. SOULE, M., 1973, The epistasis cycle: a theory of marginal populations. Annual Review of Ecology and Systematics, 4: 165-187. STAUBER, L. A., 1950, The problem of physiolog- ical races with special reference to oysters and oyster drills. Ecology, 31: 109-118. STEARNS, S. C. & SAGE, R. D., 1980, Maladapta- tion in a marginal population of the Mosquito Fish Gambusia affinis. Evolution, 34: 65-75. STEHLI, F. G. & WELLS, J. W., 1971, Diversity and KAT age patterns in hermatypic corals. Systematic Zoology, 20: 115-126. SUBRAHMANYAM, C. B., KRUCZYNSKI, W. L. & DRAKE, S. H., 1976, Studies on the animal communities of two North Florida salt marshes. Part Il. Macroinvertebrate communities. Bulletin of Marine Science, 26: 172-195. SULLIVAN, M. J., 1971, Distribution and ecology of edaphic diatoms in the Canary Creek salt marsh. MS thesis, University of Delaware, 97 p. WASS, M. L., 1972, A checklist of the biota of Chesapeake Bay. Virginia Institute of Marine Science Special Scientific Report 65: 184 p. WEBB, S. D., 1969, Extinction-origination equilibria in Late Cenozoic land mammals of North Amer- ica. Evolution, 23: 688-702. WELCH, W. R., 1968, Changes in the abundance of the Green Crab, Carcinus maenas (L.) in rela- tion to recent temperature changes. United States Fish and Wildlife Service Fisheries Bulle- tin, 67: 337-345. WILLIAMS, E. E., 1972, The origin of faunas. Evo- lution of lizard congeners in a complex island fauna: a trial analysis. Evolutionary Biology, 6: 47-89. YONGE, C. M., 1928, Structure and function of the organs of feeding and digestion is Ostrea edulis. Journal of the Marine Biological Association of the United Kingdom, 14: 295-386. MALACOLOGIA, 1982, 23(1): 55-62 EFFETS DES CONDITIONS D'ÉCLAIREMENT SUR LA CROISSANCE DE LYMNAEA STAGNALIS (GASTÉROPODE PULMONÉ) J. Seuge et R. Bluzat Laboratoire de Zoologie—Batiment 442, Centre d'Orsay de l'Universite de Paris-Sud, 91405—Orsay Cedex—France RESUME Lymnaea stagnalis a été elevee au laboratoire depuis l'éclosion jusqu'à la mort sous diverses conditions d'éclairement: photophase longue (16/24 h), photophase courte (7/24 h), obscurite, intensité lumineuse variable (1300, 180, 50 et 10 lux), lumières colorées (ultraviolette, bleue, verte, jaune et rouge). En lumière blanche, l'élevage sous courte photophase se traduit par une augmentation de la longévité; la reduction de l'intensité lumineuse sous longue photophase produit le phénomène inverse. Seules les radiations rouges ont un effet positif sur la durée de la vie. A intensité lumineuse égale, la croissance de la coquille est nettement augmentée quand l'élevage se déroule sous photophase courte (13.4% en moyenne); ce phenomene est encore plus net quand les animaux evoluent à Гобзсище (23.7%). La réduction de l'intensité lumineuse ne se traduit par une diminution de la croissance que sous longue photophase. L'élevage des limnees sous des radiations ultraviolettes, vertes et jaunes modifie peu leur croissance. Par contre sous les radiations bleues la croissance est diminuée alors qu'elle est augmentée sous les radiations rouges. Dans tous les cas la croissance des coquilles peut être décrite par le modèle de von Bertal- anffy et l'étude des paramètres К et {< nous a permis de preciser les effets des différentes modalités d'éclairement. Mots clefs: Lymnaea; croissance; eclairement; photophase; intensité; obscurite; longueur d'onde. INTRODUCTION Nous avons constate chez Lymnaea stagnalis (Linn.) d'importants troubles de la croissance et de la fecondite lors d'intoxica- tions à long terme par divers toxiques (Bluzat et al., 1979; Bluzat & Seuge, 1981; Seuge & Bluzat, sous presse). Il nous a paru inter- essant de rechercher si des troubles du même ordre pouvaient être provoques par des facteurs externes. Dans un premier travail (Seuge & Bluzat, 1982), nous avons envisage l'influence de la mineralisation de l'eau. Dans le present arti- cle nous examinons les effets des conditions d'eclairement sur la croissance en etudiant successivement: la photophase, l'intensite lumineuse et la qualite de la lumière. Les con- sequences de ces differentes condition d'ele- vage sur la potentiel reproducteur seront en- visagees par ailleurs. Des travaux très anciens indiquent que les limnees ne perçoivent pas les radiations rouges (Graber, 1884 dans Franc, 1968) et (55) n'effectuent pas leur developpement em- bryonnaire sous les radiations vertes (Yung, 1878 dans Franc, 1968). Nous ne disposons pas d'autres informa- tions relatives à ce sujet en dehors des in- vestigations a court terme de van der Steen (1967) concernant la fecondite et d'un resume des premiers travaux de Bohlken et al. (1978). MATERIEL ET MÉTHODE L'animal utilise est le Mollusque Gastero- pode Pulmone Lymnaea stagnalis eleve depuis plusieurs generations au laboratoire dans l'eau de ville (pH: voisin de 7.5; durete totale: 230 mg CaCOs/litre; 20° + 0.5C; aeration permanente); l'alimentation est as- suree par des feuilles de laitue fournies en quantité suffisante et l'eau est renouvelee chaque semaine. Toutes nos experiences sont conduites avec des animaux groupes dans un bac de verre contenant deux litres d'eau. Trois cents jeunes limnees sont mises 56 SEUGE ET BLUZAT en experience le jour de leur eclosion; a cing semaines elles sont mesurées pour la premiere fois et seuls les 40 individus de plus grande taille sont conserves. Une deuxieme selection est operee de la méme facon un mois plus tard: chaque lot est alors constitue de 12 limnees qui sont maintenues en experi- ence jusqu'à ce que 50% d'entre elles soient mortes; l'etude de chaque lot est donc arrêtée quand le nombre de limnees vivantes est <5. Les elevages sous lumiere blanche (tubes Mazda ‘blanc industrie” TFR/40/BBL) sont realises soit en longue photophase, 16/24 heures d'eclairement (series 16) soit en courte photophase, 7/24 heures d’eclaire- ment (series 7). Differentes intensites lumi- neuses, mesurees a la surface des bacs, ont ete utilisees: 1300 lux (16 — 1 et 7 — 1), 180 lux (16 — 2 et 7 — 2), 10lux(16 — 3et7 =:3) et 50 lux (16 — 4). Une serie d'animaux a ete elevee а l'obscurite totale (0) a l'exception de courtes periodes necessaires à leur entretien. Les elevages sous lumières colorées n'ont ete conduits qu'en longue photophase (16/24h): les gammes de radiations sont obtenues par des tubes lumineux speciaux entoures d'un filtre de rhodoid “R.P.” de 0.25 mm d'epaisseur. Les tubes et les filtres utilises sont les suivants: Lumiere ultraviolette (U.V.): tubes TFA/4L/5 Mazda; pas de filtre; A = 380-440 nm; 360 lux. Lumiere bleue (B): tubes TF/40 Bleu Mazda; rhodoid n° 2005; À = 430-480 nm; 230 lux. Lumiere verte (V): tubes TF/40 Vert Mazda; rhodoid n° 2003; A = 500-550 nm; 360 lux. Lumiere jaune (J): tubes TF/40 Jaune Mazda; rhodoïd n° 222; A = 550-630 nm; 1300 lux. Lumiere rouge (R): TL/40/15 Philips; rhodoid n° 227; À = 630- 670 nm; 50 lux. Toutes ces intensites lumineuses ont ete mesurées à la surface des bacs; il est en effet impossible d'apprécier l'intensité exacte reçue par chaque limnée. Chaque mois la hauteur des coquilles est mesuree (Bluzat & Seuge, 1979); nous avons vérifie que la croissance pouvait être décrite selon le modèle de von Bertalanffy, Tt = tx (1—e-k(t-to)), en construisant la droite de Walford (1946), taille au temps t + 1 en fonc- tion de la taille au temps t; le parametre k est egal au logarithme neperien de la pente de cette droite et une determination graphique nous a permis d'évaluer to a —0.25 mois; T=, la taille maximale specifique, est calculee d'apres l'équation de von Bertalanffy. L'étude Statistique des resultats (analyse de variance et test de Student) nous permet d'apprécier le niveau de securite des differences observees. RESULTATS L'évolution de la taille des coquilles est rep- resentee graphiquement dans les Figs. 1 et 2; les donnees concernant les deux premiers mois sont detaillees dans le Tabl. 1. La Fig. 1 montre l'influence de la durée de la photophase (1a) ainsi que celle de l'inten- TABL. 1. Tailles moyennes (T) des coquilles et erreur standard à la moyenne (S. E.), а Гаде de 1 mois (première sélection) et à 2 mois avant (a) et après (b) la deuxième sélection. Age: 1 mois Conditions d'éclairage T (mm) SIE. 16 — 1 8.17 0.209 161552 10.52 0.296 16 =-3 6.37 0.275 16 — 4 10.17 0.184 7-1 7.59 0.133 7-2 7.76 0.102 TS 8.6 0.153 0 7.96 0.352 Ц. V. 12.15 0.219 B. 9.99 0.117 V. 10.74 0.173 J. 11.54 0.154 A 10.47 0.122 Age: 2 mois a b T (mm) 5. Е Т (тт) SIE 16.3 0.393 1925 0.468 19.52 0.357 22:5 0.3 13775 0.592 18.29 0.727 18.16 0.219 19.67 0.204 16.04 0.321 18.42 0.345 16.86 0.25 18.67 0.327 18.15 0.306 20.62 0.4 18.64 0.464 22.45 0.604 21.37 0.255 23.08 0.286 20.61 0.24 22.25 0.332 20.74 0.25 22.42 0.254 20.7 0.27 22.54 0.258 21.55 0.229 22.87 0.132 ÉCLAIREMENT ET CROISSANCE DE LYMNAEA 57 taille 50 temps en mois 10 12 14 16 taille nm nN o> col = > = nm = a = > taille we ee — E LA 5 > sa s ET x E 30 20 104 — Г =] © 16-1 16- À 416-3 | 016-4 J temps en mois A DAA AAA ET obs: T 22 4 6 8 10 12 14 FIG. 1. Croissance de la coquille de Lymnaea Stagnalis en fonction de la durée de la photophase et de l'intensite de l'éclairement: 1a: photophase (16 heures, 7 heures et O heure = obscurite); 1b: intensité lumineuse en photophase longue (16 — 1: 1800UX 16, — 2: 1680x6320) luxe 16) 4: 50 lux). 1c: intensite lumineuse en photophase courte (7 — 1: 1300 lux; 7 — 2: 180 lux; 7 — 3: 10 lux). L'interruption de l'axe des temps a 2 mois correspond a la deuxieme selection: le point de gauche indique la taille moyenne des limnées avant celle-ci; le point de droite precise la taille moyenne des 12 animaux les plus grands maintenus seuls en experience. SEUGE ET BLUZAT 58 L00'0>d (p—91) 100'0>9 (p—-91) 200>d>100 51650 ct Lp 50'0>9 >20 0 SAO) pl LL y Su ves’ | €'0p SS'0p L'o—d 86+ 0 LL 6 r ‘sut L6£ 0 L'8€ Ge 8€ L'o>d >S0'0 €0S'0 LE 8 A 10'0>d 1000 9990 S'S€ SS SE 100'0>d ZO) LL OL g SU 109'0 8'6€ 68'6€ 100'0>d ger 0 pl el An 100`0>9 (e —2) Lo'0>d >100'0 (£—2) 605'0 G6 St cc 9p 855`0 GS bl 0 100'0>d 100'0>d Е 91) 100'0>d (8-91) u(Ll-Z) €t80 19'ct 84 ct ‘S'U(L—Z) e6€'0 SS! Ss! в 10008 (2 91) L'o>d >G0'0 (2-91) u(L—Z) 6950 19 Lp 18 Lp "$ u(l-2) 9/£ 0 6/91 vl 2-1 100' des 8ct' 0 ct A 100'0>d 065`0 Zt pl L — 2 z0'0=0 S690 6 9€ 8£ 6€ 1000 -d 755`0 62`6 9 y — 91 z0'0=0 68/`0 €'9€ 95'85 100 0 -d 9/2 0 so! 9 В — 0} 10'0=9 1870 S'9€ LOLE 100'0>d NAO) o! 6 с — 91 — G9'0 68€ Ge 6€ — 8/7`0 GZ el LL | — 9L d ENS go Xu] co | d A (sioux) (sioui) эбеле|оэ.р эм эр OL = 99949 = 271 SUONNPUOY auueAow no aang sang "seseyjuajed эциа esidald шошз} э| Jane Jos | — 9| UIOWEL aj DONE JIOS ээллэ$ао 99U9J8JJIP E] эр иоцеэошиб!$ эр пеэлии :4 :(`3 'S) рлерие}$ пала LOS ya ээллэзао ayeWIXeW ajjle} :’qo хил ‘Ayuejeyag UOA ap vonenba | эр зэдэшелед :0] Ja y 'syeuBejs eeeuw/7 ap sajıınboo sap aoueSSI019 e] INS а BIA ap ээлпр ej INS диэшэлерэ,р зиошриоэ SAJUAJSYIP эр SJOHA ‘с TAEVL ÉCLAIREMENT ET CROISSANCE DE LYMNAEA 59 taille en mm 40 A 30 20 10 1 22 3 4 5 6 Y < Г] - M- = m--=7 a | - EE | > rk Sell EA El CN ee CU DE % === оо SS SOS EN ”— А Be tu + 8 9 10 11 12 13 14 FIG. 2. Croissance de la coquille де Lymnaea stagnalis en fonction de la qualité de la lumière (photophase longue); au cours des deux premiers mois la croissance de tous les groupes (sauf E = 16 — 1) est a peu près identique. Abscisse: idem Fig. 1. site lumineuse en photophase longue (1b) et en photophase courte (1c). La Fig. 2 presente l’évolution de la taille des limnées élevées sous diverses lumières colorées et permet la comparaison avec les témoins lumière blanche et les témoins ob- scurite. L'étude détaillée de la croissance de ces coquilles permet de conclure que le modele de von Bertalanffy reste applicable dans tous les cas. Le Tabl. 2 précise la durée moyenne de la vie des animaux de chaque groupe et rassemble les valeurs des paramètres К et Tx ainsi que celles des tailles maximales ob- servees. DISCUSSION Survie L'élevage de Lymnaea stagnalis se révèle possible sous toutes les conditions lumi- neuses eprouvees. Néanmoins, la duree de l'éclairement, son intensité et la qualite de la lumière modifient, dans un sens ou dans l'autre, l'espérance de vie des limnees (Tableau 2). A intensité lumineuse égale, une courte photophase augmente nettement la longevite et le résultat obtenu avec les animaux éleves dans une obscurite quasi-totale va dans le même sens. Par contre, il faut noter que les intensités lumineuses faibles (16 — 2) et très faibles (16 — 3 et 16 — 4) sont défavorables a la survie des animaux en photophase longue. Dans le cas des lumieres de couleur la duree de vie est soit normale: ultraviolet et rouge, soit plus courte que celle du temoin lumière blanche (16 — 1): bleu, vert et jaune. Le résultat obtenu avec la lumière rouge est assez paradoxal car, l'intensité lumineuse étant faible (50 lux), une durée de vie relative- ment courte, du même ordre que celle des groupes 16 — 3 et 16 — 4 (10 mois environ) 60 SEUGE ET BLUZAT etait previsible. La lumiere rouge excerce donc probablement une action specifique positive. Croissance En regle generale, nous avons constate une bonne homogeneite de l'evolution de la taille des coquilles des limnees dans chacun des lots mis en experience (Tableau 1 et Tableau 2; dans ce dernier seules figurent les S. E. relatives a la derniere taille moyenne calculee). Par ailleurs, il est important de souligner que la densite de population est restee stable (Tableau 2) pendant longtemps dans tous les cas et na pas pu influencer significativement la croissance (Forbes & Crampton, 1942). Dans le present article evolution de la vitesse de la croissance en fonction de Гаде des animaux n'a pas ete etudiee en detail; le calcul du parametre k par la methode de Walford (1946) en rend compte globalement. Nous examinerons successivement l'influ- ence de la longueur de la photophase, de lintensite lumineuse et de la qualite de la lumiere. Longueur de la photophase La Fig. 1a montre que pendant 10 mois la croissance des coquilles est tout a fait com- parable dans les groupes 16 — 1 et 7 — 1. Au dela il n'en va plus de même; les animaux eleves en courte photophase ont alors une croissance plus importante: les tailles maxi- males observees sont tres nettement differ- entes. La comparaison des series 16 — 2, 7 — 2 dune part, 16 — 3, 7 — 3 d'autre part con- duit exactement a la même conclusion. Le modele de von Bertalanffy permet de decrire dans tous les cas la croissance des coquilles; (etude des parametres К et Tx montre que leur valeur, pour une intensite lumineuse donnee, depend de la longueur de la photo- phase (Tabl. 2). Bohlken et al. (1978) ont signalé, chez Lymnaea stagnalis, que l'élevage sous courte photophase provoquait une croissance plus importante; nos resultats sont donc en plein accord avec celui de ces auteurs. Intensite lumineuse La Fig. 1b indique qu'en jour long une forte reduction de l'intensite lumineuse se traduit par une croissance plus faible; la difference est nettement sensible à Гаде de quatre mois. La Fig. 1c met au contraire en évidence que la reduction de l'intensité lumineuse n'a aucun effet sur la croissance des animaux eleves sous courte photophase: les tailles maximales observees ne diffèrent pas. L'etude des paramètres К et T= de l'équa- tion de von Bertalanffy confirme ces conclu- sions (Tabl. 2): (1) sous courte photophase les valeurs de k ne varient pratiquement pas quelle que soit l'intensite lumineuse: par contre, sous longue photophase, la valeur de к diminue avec celle de l'intensite |, ces deux variables etant liées par l'équation К = 0.0975 log | + 0.1763 où г = 0.997 et 0:01 = р = 0.001; (2) les valeurs de Tx sont très voisines des tailles maximales observées à l'exception de deux cas: 10 et 50 lux sous longue photo- phase (16 — 3 et 16 — 4); ceci souligne qu'une intensite lumineuse très faible sous longue photophase constitue une condition particulièrement aberrante qui perturbe pro- fondement la croissance, comme le montre par ailleurs l'augmentation de la variabilité des tailles (Tabl. 1). Cas de l'obscurite Un elevage à l'obscurité pratiquement permanente peut être envisage comme realise à la fois sous une photophase ex- tremement reduite et avec un eclairage d'intensite pratiquement nulle. Dans ce cas nous constatons que la croissance est par- ticulièrement importante (Fig. 1a); la taille moyenne maximale observee est de 45.95 mm (42.67 mm chez animaux 7 — 3) et la valeur du paramètre К particulièrement faible (0.338). Ce resultat parait tout à fait compatible avec celui obtenu sous courte photophase et tres faible eclairement et souligne qu'il existe encore une difference nette entre un elevage effectue sous une in- tensite aussi faible que 10 lux et celui realise a l'obscurite. Ces faits demontrent que les photorecepteurs de la limnée sont sensibles à des intensites tres faibles; cette conclusion doit être rapprochée de celles de van der Steen (1967) et de Lickey et al. (1977) puis- que ces auteurs ont precise que Lymnaea et Aplysia perçoivent respectivement des in- tensites lumineuses de 1 et 2 lux. Qualite de la lumiere Dans cette experience nous avons neglige volontairement le rendement energetique de chacune des gammes de radiations utilisees. Rappelons en outre que, si dans tous les cas la photophase est de 16/24 heures, l'intensite lumineuse n'est pas la même d'une lumière coloree a l'autre. ÉCLAIREMENT ET CROISSANCE DE LYMNAEA 61 L'examen des courbes de croissance (Fig. 2) et des tailles moyennes maximales ob- servees indique que seuls deux cas diffèrent significativement des témoins lumière— blanche (Tabl. 2): (1) la croissance est nette- ment plus faible (Tmx ob = 35.5 mm) sous lumière bleue; (2) elle est indiscutablement plus importante (Tmx ob = 42mm) sous lumière rouge. Il faut noter toutefois que la croissance semble toujours se dérouler normalement dans la mesure où les tailles maximales spécifiques calculées restent très proches des tailles maximales observées; par contre, les valeurs du paramètre k sont, malgre des variations d'assez faible ampli- tude, spécifiques de chaque gamme de radi- ations. Ces resultats nous permettent de répondre à l’une des questions que nous nous sommes posees: les limnées percoivent-elles les lumières ultraviolette, bleue, verte, jaune et rouge? Dans tous les cas la réponse est in- discutablement positive puisqu’aucun de ces resultats n'est identique à celui obtenu lors de l'élevage à l’obscurite. La faible croissance observée dans le cas de la lumière bleue pourrait être éventuelle- ment attribuée au niveau assez faible de l'intensite lumineuse (230 lux) mais la com- paraison avec les animaux 16 — 2 (180 lux) nous conduit à conclure qu'il n’en est rien et que les radiations bleues exercent un effet depresseur specifique sur la croissance. L'expérience qui s'est déroulée sous la lumière rouge constitue un cas particulier pour deux raisons: (1) la temperature s'est revelee être en moyenne de 21°C alors qu'elle a pu être maintenue a 20°C dans tous les autres cas, (2) l'intensite lumineuse etait particulièrement faible: 50 lux. Un élevage realisé sous 16/24 heures de lumière blanche à la température de 22°C (Tmx ob = 39.9 mm et k = 0.478) nous per- met de n'attribuer qu'un tres faible effet a l'élévation de la temperature dans Гехреп- ence sous lumière rouge et les résultats ob- tenus avec le temoin 16 — 4 (16/24h de lumière blanche—50 lux) soulignent l’exist- ence d'une action positive spécifique des radiations rouges sur la croissance. L'ensemble de ces experiences montre que la durée de l'éclairement, son intensité et sa qualité agissent sur le déroulement de la croissance, sans doute par l'intermédiaire de divers récepteurs photosensibles; à notre connaissance la littérature scientifique ne nous offre aucun point de comparaison directe. D'après les travaux de Geraerts (1973, 1975) et Joosse (1975) la croissance de la coquille de Lymnaea stagnalis est sous le contrôle d'une hormone (Growth Hormone) secretee au niveau des L.G.C. (Light Green Cells) des ganglions cerebroides. Nos re- sultats suggèrent que des perturbations de cette secretion hormonale pourraient être induites par la longueur de la photophase, l'intensite de la lumière et sa longueur d'onde. Cette hypothèse s'appuie sur les travaux de Wendelaar Bonga (1971) et Roubos (1975) qui mettent en evidence un rythme journalier dans l’activite des cellules neurosecretrices du cerveau de Lymnaea stagnalis. Dans des etudes anterieures nous avons deja enreg- istre d'importants troubles de la croissance provoques par des insecticides (Seuge et Bluzat, 1981) et des detersifs (Bluzat et Seuge, 1981); il nous parait donc très vrai- semblable que des perturbations du fonc- tionnement des cellules neurosecretrices puissent être entrainees par des facteurs abiotiques. CONCLUSIONS Le present travail a mis en évidence que la survie et la croissance de Lymnaea stagnalis sont perturbees par: (1) la duree de l'eclaire- ment journalier; (2) l'intensité lumineuse; (3) la longueur d'onde de la lumière. La croissance des coquilles est plus impor- tante (13.4% en moyenne) quand l'élevage est realise sous photophase courte (series 7 — 1, 2 et 3) que sous photophase longue (series 16 — 1, 2, 3 et 4); un veritable gigant- isme est obtenu quand l'élevage est effectue à l'obscurité: augmentation de la taille de 23.7% (temoin: moyenne des series 16 — 1, 2, 3 et 4). La croissance n'est pas affectee par la reduction de l'intensite lumineuse sous courte photophase; une conclusion opposee peut etre formulée dans le cas d'un elevage sous longue photophase ou la valeur du parametre К de l'equation de von Bertalanffy est propor- tionnelle a l'intensite lumineuse. Par rapport au temoin 16 — 1, en consider- ant la taille moyenne maximale observee, la croissance n'est pas significativement modi- fiée par l'élevage sous les lumières ultra- violette, verte et jaune. Par contre, elle est diminuée (8.7%) dans le cas de la lumière bleue et augmentee (8%) dans celui de la lumière rouge. Les valeurs du paramètre k mettent en évidence un effet specifique de ces radiations. 62 SEUGE ET BLUZAT TRAVAUX CITES BLUZAT, R., JONOT, O., LESPINASSE, G. & SEUGE, J., 1979, Chronic toxicity of acetone in the fresh water snail Lymnaea stagnalis. Toxi- cology, 14: 179-190. BLUZAT, R. & SEUGE, J., 1979, Effets de trois insecticides (Lindane, Fenthion et Carbaryl): toxicité aigué sur 4 espèces d'invertebres limniques; toxicite chronique chez le Mollusque Pulmone Lymnaea. Environmental Pollution, 18: 51-70. BLUZAT, R. & SEUGE, J., 1981, Effets à long terme de quatre detersifs chez le Pulmone d'eau douce Lymnaea stagnalis L.: intoxication des animaux des l'eclosion. Environmental Pollution, ser. À, 25: 105-122. BOHLKEN, S., ANASTACIO, S., LOENHOUT VAN, H. & POPELIER, C., 1978, The influence of day length on body growth and female reproduction activity in the pond snail Lymnaea stagnalis. General and Comparative Endocrinology, 45: 109. FORBES, G. S. & CRAMPTON, H. E., 1942, The effects of population density upon growth and size in Lymnaea palustris. Biological Bulletin, 83: 283-289. FRANC, A., 1968, Sous-classe des Pulmones /n: Traite de Zoologie, tome 5, fascicule 3, P. P. GRASSE (ed.), Masson, Paris, p. 325-607. GERAERTS, W. P. M., 1973, Effects on growth of endocrine centers in the cerebral ganglia of Lymnaea stagnalis. In: Abstracts of the Seventh Conference of European Comparative Endo- crinologists, p. 116. Budapest: AkademiaiKiado. GERAERTS, W. P. M., 1975, Studies on the endo- crine control of growth and reproduction in the hermaphrodite pulmonate snail Lymnaea stag- nalis. Thèse, Universite Libre d'Amsterdam, Pays-Bas, 126 p. JOOSSE, J., 1975, Endocrinology of Molluscs, Colloques Internationaux C.N.R.S., n° 251: Actualites sur les hormones d'Invertebres. Villeneuve d'Ascq, France, р. 107-123. LICKEY, M. E., WOZNIEK, J. A., BLOCK, G. D., HUDSON, D. J. & AUGTER, G. K., 1977, The consequences of eye removal for the circadian rhythm of behavioural activity in Ap/ysia. Journal of Comparative Physiology, 118: 121-143. ROUBOS, E. W., 1975, Regulation of neuro- secretory activity in the freshwater pulmonate Lymnaea Stagnalis (L.) with particular reference to the role of the eyes. A quantitative electron microscopical study. Cell and Tissue Research, 160, 291-314. SEUGE, J. & BLUZAT, R., 1982, Influence d'une intoxication par le lindane en fonction de la durete de l'eau chez Lymnaea stagnalis (Mol- lusca Pulmonata). Proceedings of the Seventh International Malacological Congress. Mala- cologia, 22: 15-18. SEUGE, J. & BLUZAT, R., sous presse, Chronic toxicity of three insecticides (Carbaryl, Fenthion and Lindane) in the freshwater snail Lymnaea stagnalis. Environmental Research. STEEN VAN DER, W. J., 1967, The influence of environmental factors on the oviposition of Lymnaea stagnalis (L.) under laboratory condi- tions. Archives neerlandaises de Zoologie 17: 403-468. WALFORD, L., 1946, A new graphic method of describing the growth of animals. Biological Bulletin, 90: 141-147. WENDELAAR BONGA, S. E., 1971, Formation, storage and release of neurosecretory material studied by quantitative electron microscopy in the fresh water snail Lymnaea stagnalis (L.). Zeitschrift für Zellforschung und mikroskopische Anatomie, 113: 490-517. ABSTRACT THE EFFECTS OF DIFFERENT LIGHT CONDITIONS ON THE GROWTH OF LYMNAEA STAGNALIS (GASTROPODA: PULMONATA) J. Seuge and R. Bluzat Pond snails, Lymnaea stagnalis, were reared in the laboratory, from hatching to death, under various light conditions: total darkness, two photophases (16/24 and 7/24 hours), four light intensities (1300, 180, 50 and 10 lux), five lights of different wavelengths (U.V., blue, green yellow and red). With white light, rearing under short photophase induced an increase in longevity; the de- crease of light intensity under long photophase led to an opposite result. The red light had a positive effect on life expectancy. At constant light intensity, shell growth was clearly increased when rearing was achieved under short photophase (13.4%); this result was even more evident when the snails were kept in darkness (23.7%). The reduction of light intensity induced a decrease of shell growth only under the long photo- phase. The rearing of pond snails under U.V., green and yellow radiations only slightly modified growth. Conversely, under blue light shell growth was reduced whereas it was increased under red light. In all cases, shell growth could be described by von Bertalanffy's pattern; parameters k and Tx allowed us to point out the effects of the various lights. MALACOLOGIA, 1982, 23(1): 63-73 GROWTH IN MARINE GASTROPODS: A NON-DESTRUCTIVE TECHNIQUE FOR INDEPENDENTLY MEASURING SHELL AND BODY WEIGHT A. Richard Palmer Department of Zoology, University of Alberta, Edmonton, Alberta T6G 2E9, Canada and Bamfield Marine Station, Bamfield, British Columbia VOR 1B0, Canada ABSTRACT A technique is described for obtaining non-destructive measurements of shell weight and body wet weight in marine gastropods. Shell weight is obtained by weighing whole animals in sea- water and using a regression between these values and destructively sampled dry weights of shell. This weight may then be subtracted from whole weight in air, to provide an estimate of body wet weight. Shell weights are more accurately estimated than body weights. However, the mean cumulative error of this technique for estimating body weights is 10.6% for Thais lamel- losa, 4.9% for T. canaliculata and 4.8% for T. emarginata. The possible application of this technique to other carbonate skeleton-producing invertebrates is briefly discussed. Key words: Gastropoda; non-destructive measurement; Mollusca; shells; growth; Thais; car- bonate skeletons; weight. INTRODUCTION Growth in molluscs may be assessed using several different quantities (Wilbur & Owen, 1964), the most common of which is one or more caliper measurements of shell size (Branch, 1974; Frank, 1965; Kenny, 1977; Randall, 1964; Spight, 1974; Yamaguchi, 1977). Other techniques, including laser dif- fraction (Stromgren, 1975) and total weight (Stickle & Duerr, 1970; Walne, 1958) have also been used to assess mollusc size. A drawback to these measures is that they measure attributes primarily of the shell and only indirectly those of the animal residing within it. Where shell shapes and thicknesses are very similar, such measures of size or weight are usually adequate if animals are growing actively. However, body size changes are not always paralleled by changes in shell size: decreases in body weight due to spawning or starvation will not be accompanied by concomitant decreases in shell size; shell growth may continue in some molluscs in the absence of feeding (Revelle & Fairbridge, 1957: 267; Rhoads & Young, 1970: 163; Zischke et al., 1970); and the spires of gastropod shells can often erode with increasing age (Spight et al., 1974). In these situations, shell size measurements will not provide an accurate estimate of body size. To circumvent these problems, | have de- veloped a non-destructive technique to sepa- rate body growth from shell growth in marine prosobranch gastropods. This technique re- lies upon two weight measurements: 1) a weight of the whole animal immersed in sea- water, which ultimately provides an estimate of shell weight (See Havinga, 1928, and Nishii, 1965, for immersed weight estimates of shell weight in oysters; and Bak, 1973, for application to corals; Lowndes, 1942, has ap- plied a similar technique to a number of in- vertebrates and vertebrates), and 2) a weight of the entire animal, shell plus body, in air. Subtracting the estimated weight of the shell from the total weight provides an estimate of the body weight, and the animal is still intact. Below, | describe the application of this tech- nique to three species of thaidid gastropods, Thais lamellosa (Gmelin, 1791), T. canali- culata (Duclos, 1832) and T. emarginata (Deshayes, 1839), all inhabitants of North American rocky intertidal shores from Alaska to California (Ricketts et al., 1968). In essence, this technique takes advantage of the specific gravity differences between shell and tissue. By weighing intact animals in two mediums of differing specific gravity (air and seawater), it is possible to separate the relative contribution of each component to the animal's total weight. It further takes advan- tage of two convenient attributes of gastropod molluscs: 1) the mantle is not attached to the shell, thus extrapallial fluid is not irrevocably trapped, and 2) it is possible to remove a sub- (63) 64 PALMER stantial amount of the pallial water without damaging the animal. Thus the whole weight may be reduced to shell plus body weight, with a minimal amount of residual extravis- ceral water. MATERIALS AND METHODS ‘Immersed weight,’ or the weight of whole animals in seawater, was obtained by placing them on а 3cm x 3cm VEXAR* plastic screen platform, supported by and suspended from a fine copper wire that could be hooked directly to the underside of a Mettler P153 balance. The balance was placed on a stand that straddled the container of seawater in which the animals were to be weighed. By taring the balance to compensate for the weight of the suspended platform, actual weights of the immersed animals could be read with no correction. Snails were intro- duced individually using a pair of forceps and weights recorded to the nearest 0.001 g. Because specific gravity differences were being used to separate shell weight from body weight, it was important to ensure that there was no air inside the mantle cavities of indi- viduals prior to weighing them under water, otherwise shell weights could have been un- derestimated. A procedure used to minimize this possibility was to completely immerse the animals for 24-48 hours prior to weighing, since most animals appeared able to clear their mantle cavities of air over this period. Animals were also ‘chased’ into their shells with the tips of the forceps immediately prior to placing them on the immersed platform; this acted to squeeze much, though probably not all, of any residual air out of the mantle cavity. Air was detected as bubbles released by the withdrawing animal in fewer than 5% of the animals; these individuals were noted. To obtain non-destructive estimates of shell weights from immersed weights, it was necessary to compute a regression of actual shell weight on immersed weight for all three species. For this purpose, | measured im- mersed weights for individuals from a size range of all three species. The shells were then broken open using a C-clamp to avoid uncontrolled shattering, and the fragments separated from the body and dried to constant weight at 80°C (Tables 2-4). The slopes of these regressions of destructively sampled shell dry weight on immersed weight (regres- sions 1-3, Table 1; Fig. 1) were then used to estimate actual shell weight from immersed TABLE 1. Least squares linear regressions for shell and body weight estimates. Weights are measured in grams. N = number of individuals. See Figs. 1 and 2 for plots of the data for regressions 1-3 and 8-10 respectively. Regression number Species Shell dry weight (Y) from immersed whole weight (X) N Regression equation В 1 Т. lamellosa 27 Y = 1.572Х + 0.0162 0.9998 2 Т. canaliculata 21 Y = 1.558Х — 0.0075 0.9995 3 T. emarginata 19 Y= 11.530X=0:0032 0.9997 Body immersed weight (Y) from body dry weight (X) 4 3 spp. pooled 16 Y = 0.202X — 0.0008 0.9946 Shell dry weight (Y) from corrected immersed whole weight (X) 5 T. lamellosa 27 Y = 1.598X + 0.0174 0.9999 6 T. canaliculata 21 Y = 1.600X — 0.0013 0.9996 7 T. emarginata 19 Y = 1.605X + 0.0017 0.9993 Ash free dry weight (Y) from estimated body weight (whole wt.-shell wt.) (X) 8 T. lamellosa 27 Y = 0.1043X + 0.0180 0.8050 9 T. canaliculata 21 Y = 0.1974X + 0.0141 0.9117 10 T. emarginata 19 Y = 0.2514X — 0.0029 0.9846 Log ash-free dry weight (Y) from log shell length (X) 11 T. lamellosa 27 Y = 2.940X — 5.426 0.9277 12 T. canaliculata 21 Y = 2.709X — 4.743 0.9471 13 T. emarginata 19 Y = 3.304X — 5.440 0.9759 NON-DESTRUCTIVE SHELL AND BODY WEIGHTS 65 weight of the whole animal for each species. It is clear from the high R° values (0.9995- 0.9998) that immersed weight can provide a very accurate, non-destructive estimate of actual shell weight. Some care must be exercised in the appli- cation of a single regression to different spe- cies, however, since not all the weight of a snail immersed in seawater is due to the shell. The different slopes in regressions 1-3 (Table 1; Fig. 1) reflect differences in the amount of tissue. To assess the contribution of tissue weight to immersed weight, | separated a small number of individuals of all three spe- cies from their shells and measured the im- mersed weight of the bodies alone. These were then dried to constant weight at 80°C. Regression 4 (Table 1) describes how much a given dry weight of tissue weighs when im- mersed in seawater before it has been dried (dry weights were used here because they are much more accurate than attempting to uniformly towel-dry animals for wet weights in Shell Weight (g) © air). If the tissue dry-weight values of Tables 24 are multiplied by the slope of this regres- sion, subtracted from the total immersed weights, and these corrected immersed weights (i.e. corrected for the contribution of body weight to immersed weight) regressed against destructively sampled shell dry weights, the differences between the three species disappear (regressions 5-7, Table 1). This is to be expected since the specific grav- ity of shell material should be essentially the same for all three species. However, since it is only changes in the amount of body weight relative to the weight of the shell that will af- fect the accuracy of the uncorrected estimate of shell weight, and also since the contribution of the entire body to immersed weight is at most 4% (4% for T. emarginata, less than 3% for T. lamellosa and T. canaliculata), such a correction will yield very slight differences in the estimated shell weights. In other words, since the entire body weight of an immersed animal amounts to less than 5% of the total 13.000 °-T. lamellosa °-7. canaliculata o-T. emarginata 10 2.0 3.0 Immersed Weight (g) 5.0 6.0 O 8.0 FIG. 1. The relationship between immersed weight of whole animals and destructively sampled dry weight of the shell for three species of Thais. The regression equations for these data are in Table 1, regressions 1-3. The high coefficients of determination (Table 1) indicate that shell weight is very accurately estimated from the weight of the whole animal immersed in seawater. PALMER 66 €90l 8810 0€'0 600'0 bte 2210 SUESW 692€ 2100 2600 99% +000 8510 0910 9510 5657 E100 5050 505`0 515'0 005`0 é (N°21 00'0L 010'0 000 000 0000 1720 1/50 120 eel 000 59250 2250 ZES'0 0€S'0 é 9°91 ЕТ 2000 41110 $750 1000 260 2620 1620 BIO 1000 0150 695`0 0/5`0 0/S'0 é A fO'cl 6500 tlcO 050 2000 59001 0900 0 9951 150`0 0261 GG6 | 986 1 696 1 w 6 Sc 7/5 4100 9570 000 0000 964+ 9611+ 9611 €20 1100 £6rE2 GGE 2 Bee z sge a w L 22 ESZL 1900 1870 020 2000 HIoı SIıoı 60+ 082 6500 8802 LSO 2 OIL горе w gle voO! 1500 8670 600 1000 Kr BbL'L Grrl Ее 0800 z2zez Sle? L0€'Z LSe'z } Lite 18€ 2100 6570 010 1000 860 860 2260 180 9100 9261 096 1 9161 166 1 } 6712 €€'S ¿200 2050 800 1000 zreı zreı гр SOL 9200 re 1St'e e8b 2 \6r 2 w 682 Gps 7500 290 020 £000 s2St 6251 9sı Lol 150`0 — 0$0`6 920 € ZSO'€ Ze0'e ui S'0€ ELLE 1970 бр 890 9100 WISE 9962 1892 892 910 ZES'S €SS'°S G6S'S 6bb'S } ze v6b 900 1560 950 S000 0851 LEI ZBEI oF 1700 СГБ 820€ 801€ 6bL'€ } 92€ 61 2 0800 ср 9/0 $100 уе ZIEL lest $79, 9900 S86'€ 266 € 8v6'€ rlor } 92€ €0‘ + 9500 5680 90 8000 v€Zt 081 BELL 1/0 8200 IS9E 149€ erge 079`5 w 6'£€ РОТЕ 221.0 90 90 20007 бт 050 2601 660 90 190 LLO'p 931 + lol + ui SE e6'p 0500 9101 570 6000 2366! 10861 9661 860 1700 L6lL5p v6l'P Oller 691 + } €'Se Ovel 9910 6521 270 6000 че Ole беге OSE 150 969% 759$ Gb9'p 88b'p } 19€ LES 0020 9081 010 000 тие 6012 бис в 0610 259% 8€/'p 8bs'p 609'p w 69€ сэ 9420 SIZE 050 2100 6862 EBEZ SES 66% 920 HBr'S 619'6 GS€'S 8/79 Ww 0'L€ VL 6210 291 550 0100 E6BZ 8882 8682 O61 610 0029 $129 eel 9 2979 Ww L'6€ POLL 9520 6102 050 0100 6968'S 98'S №6 16 5950 402 6tt' 2 peo l 69€ 2 w air Ge 9 8001 0785 20 $7100 9165 6065 6665 12/1 660 69LEL 62921 €Z9€l bSLEL Ww 025 LO'EL 9290 920% 110 2100 8689 2689 069 OSE #1590 Sel OEIL SIeSsı LOZ'pl Ww 629 1661 9G/0 98/5 150 6100 7609 565019 +809 SES 16/0 BSPEL bbSEl ZPO'EL УВЕ } 6 ES BLZ 6650 Helr 7/0 600 7299 1199 9999 681 7/20 269+L 908%: 08 PL 90S'pl } LSS Lt'6 5950 Str 6550 7200 01/9 6699 1229 S722 1750 BO6GbL 202 bl УФЕ ЗЕ } 0'9S 89'S 9970 189% 900 S000 MER e228 Les 9rı 1970 189ZL 19741 19921 2661 } 285 % ‘XEW | % ‘Xe! ues; с | % "Xen UBaN € с | X2S (ww) — _— — y¡Bua| 10119 “uno 10113 10113 IIeus (6) m Apoq pejeusz (6) siybiam рэзлэшил (6) siybiam ajoym ‘juasaid siued = ш ‘siuod où = y ‘sAep ом} ]se¡ эц} uo sjybiam ajoym о} 1o11d А|э]ерэшили paınseaw ajam sjybiem разлэщи ‘5Аер 9AISS999NS aaly] UO рашеао элэм sjybiam ajoym jeaday ‘езо/эше/ siey, 10, sisÁjeue 1o1e pue sjybiem разлэшии 'SI4UBIEM ajoym ‘г 3718V 1 67 NON-DESTRUCTIVE SHELL AND BODY WEIGHTS 289 5800 ¿IL 570 6000 6602 b6BOZ EOLZ 69+ 7/00 209% S0S'p WA Syst LSE A A a a A ес A Е eee % XEW | % XEN UEOW с | % XEW ZEN € | xas (wu) y¡Bua] 10118 “uno 10113 10113 eels} (6) ‘M Ápoq payewns3 (6) siubiam рэзлэшил (6) syyBiam ajoym esa. ‘siuad jjews = ,—W, ‘siued эбелэле = ш ‘siued ou = | '5Аер Om] 15е| eu} UO sjybiam ajoym о} лома Alayeıpawwı painseaw элэм sjybiam pasıaWu] ‘SEP anssasons ээлц} UO pauie]qo элэм sjybiam ajoym jeaday “eyeynoyeueD siey, 10) sisÁjeue 1018 pue sjybiam рэзлэшии ‘s}yBIam ajoym ‘€ FIGVL PALMER 68 LIL NY 600 0 ss'0 100'0 6c | 600 0 SUESW 89'6 E00°0 L€0°0 29"! 100'0 190'0 190'0 090'0 es"! 2000 0510 6ct 0 LEL'O 051`0 55'8 700`0 870`0 до 100'0 €Z0°0 €20'0 cL0°0 68 | 500`0 1910 L91'0 cal'0 6510 st st 0100 9900 G6 0 100'0 solo volo SOL'O S8'€ 600 0 0€c'0 Scc'O vee 0 L€c 0 evs S00 0 c60 0 08`0 100`0 9cl 0 910 Scl 0 geil 700`0 [8<'0 S8c'0 68c 0 88c 0 875 800'0 9r1°0 650 100'0 891'0 8910 691'0 cL |b 100 0 90ÿ' 0 LO+'O 80+ 0 Olv 0 00 + 900 0 O) vz! €00°0 ELO 540 IO cLO €00°0 8lv 0 917`0 ¡AO Oct'O cS + 600 0 661 0 c8 0 c00 0 Stc'0 9ÿc'0 pre 0 Oc! 200'0 8750 9150 9150 €8S 0 00'S 600'0 0810 00 0 000 0 ccc 0 AAA) coc 0 071 600 0 ST ASIC) 61S 0 8cS 0 8cS'0 c6 0 c00'0 8120 LEO 100'0 ¡PA 0430 ¡PRA 910 100'0 SE9'0 SE9'0 559`0 9€9'0 vl? 900'0 0830 00'0 000'0 ce 0 ce 0 ¿ce O 8/0 900'0 127240) 8920 127240) 6440 €9'c 800 0 705`0 Ze0 100'0 LEE'O 0€€'0 LEEO 980 200'0 0180 +08 0 1180 918'0 GL'e 600 0 RA 80) 00'0 000'0 595'0 595`0 €9€ 0 co | 600 0 v88 0 Li80 988 0 688 0 80 + 2100 /lr0 ce 0 100'0 9St'0 957`0 SGÿ 0 тт 910`0 voll 9601 AR voll gr Z10°0 ¿6€'0 se'0 c00 0 0ZS'0 LZS°0 695'0 эн $10`0 9821 $451 687 | 96c 1 plc LLO'O LOv'0 810 100'0 575`0 575`0 675`0 08 0 010`0 cbc"! sec! src"! Gvc'l 6€ | 900'0 cer 0 veo c00'0 c8S 0 €8S'0 1850 0€'0 700`0 9551 eee lh Seek 6€€'! LUZ 0€0'0 dat veo c00°0 v6t 0 S6r'0 £6t 0 9€'c 8c0 0 9211 6S!'! 281! cell AA 800 0 959'0 ¡AO 100'0 S9Z'0 59/`0 99250 8€ O 200 0 6c8 | cc8 | 6c8 | 9€8'| | Q oo pLe 8c0 0 6+/ 0 £L'O 100'0 008'0 66/ 0 008 0 gel 2200 696 | 1961 7961 H66 1 cle % ‘XEW L % xeW иеэи/\ с | % ‘XEW иеэи € г | X2S (Wu) AE ЕЕ убие| 10419 “uno 10113 10113 TES (6) yw Ароа payewiys3 (6) siybiam рэзлэшил (6) siybiam ajoym AA A A A e a PPP OO ES A ео ‘siuod jjews = —u, ‘siued эбелэле = ш ‘siuad ou = y ‘SÂep om} 1$е| ay], uo sjybiem ajoym о} 1oud Alayeıpawwı рэллзеэш ajam siybiam разлэшил ‘SPP anissagons aalyy uo paulejgo элэм sjybiam ajoym jeaday ‘eyeu/biewsa siey, 10, sısAjeue 10118 pue siybiam рэзлешии “siybiem 8JOUM ‘+ 379V1 NON-DESTRUCTIVE SHELL AND BODY WEIGHTS 69 immersed weight, the fractional differences in body weights among animals with the same shell weight will be much less and hence in- troduce little error. As the ratio of shell to body weight decreases, however, this error will in- crease, thus for animals with slight shells this technique may be less accurate. Whole weight in air (‘whole weight’ in Tables 2-4) was measured after several preparatory steps that removed most of the extravisceral water. In Nucella /apillus (Linne, 1758), this water may account for up to 39% of the total water of an attached animal (Boyle et al., 1979), and unless removed would be included as part of the body wet weight. Also, because animals will retract on their own to varying extents on different occasions, these preparatory steps permitted a much higher level of repeatability. Snails were first arrayed aperture up on 0.6 0.5 0.4 0.3 Body Ash-Free Dry Wt. (g) 0.2 0.1 1.0 2.0 Estimated Body Wt. (whole wt - shell wt, 9) paper toweling in the order in which they were to be weighed. Each animal was then ‘chased’ back into its shell by stimulating the foot with a small modeling brush. A soft ab- sorbent tissue (e.g. Kimwipe) was subse- quently pressed firmly up against the re- tracted foot with a pair of forceps to squeeze out nearly all the remaining water. The ani- mals were left on the toweling until all the shells in a group were visibly dry (approxi- mately 20-40 min) and then weighed on top of a Mettler P153 balance to the nearest 0.001 g (Tables 2-4). This whole weight cor- responded to shell plus tissue wet weight, plus any residual mantle water. The shell weight, as estimated from immersed weight (regressions 1-3, Table 1; Fig. 1), was then subtracted from the whole weight, thus pro- viding an estimate of body tissue wet weight. A more desirable correlation, though, was Q- 7. /amellosa 4- T. canaliculata o- 7. emarginata 3.0 4.0 FIG. 2. The relationship between estimated body weight (the weight of the whole animal in air minus the weight of the shell) and destructively sampled ash-free dry weight of the body for three species of Thais. The regression equations for these data are in Table 1, regressions 8-10. As indicated by the coefficients of determination (Table 1), this relationship is least accurate for the largest species (7. lamellosa, squares) and most accurate for the smallest species (7. emarginata, circles). T. canaliculata (triangles), an intermediate- sized species, exhibits an intermediate level of variability. Much of the variation about these regressions appears due to differences among individuals in the percent water and percent ash of the body. The regressions of body dry weight on estimated body weight (whole weight minus shell weight) and of body wet weight on estimated body weight exhibit much less scatter: coefficients of determination (R*) = 0.8591, 0.9217 and 0.9887 for body dry weight, and coefficients of determination (В?) = 0.9770, 0.9871 and 0.9989 for body wet weight respectively for the three Thais species, T. lamellosa, T. canaliculata and T. emarginata. 70 PALMER between this value of estimated body weight (i.e. whole weight minus shell weight), and ash free dry weight, since ash free dry weight is a better estimate of metabolizing or metab- olizable tissue. Regressions 8-10 (Table 1; Fig. 2) describe these relations for Thais lamellosa, T. canaliculata and T. emarginata respectively. Again, there is a close corre- spondence (R° values 0.8050—0.9846, Table 1), though not as precise as for shell weight. This two-step weighing procedure thus pro- vides independent, non-destructive estimates of body and shell weight, allowing either to be used to measure growth. The second step of this process, squeezing the water out of the mantle cavity, may trau- matize the animal to some extent. However, in a field monitoring experiment involving all three species (Palmer, 1980), subgroups of each species were either 1) weighed as above in addition to being tagged and meas- ured for shell length, or 2) only tagged and measured. There were no significant differ- ences for any of the species between the pro- portion of animals recovered from the two treatment groups over the course of the fol- lowing two weeks or after 2-1/2 months (Table 5), Suggesting that the trauma associ- ated with the weighing technique is slight for these species. Immersed weight on the other hand in- volves little more disturbance than dislodg- ment of the animals from the bottom. If they have been immersed for a sufficient length of time prior to weighing, there is no need to force any air out of the mantle cavity and they may be placed directly on the submerged weighing platform. The entire operation, ex- cept for the brief transfer, takes place under water. Another consideration regarding this tech- nique is its repeatability. A comparison of R* values from regressions 1-3 with those of re- gressions 8-10 (Table 1) indicates that the replicability of immersed weights is greater than that for whole weights (Tables 2-4). For Thais lamellosa (Table 2) the mean maximum percent error for whole weight is 2.5%. Im- mersed weights of Thais lamellosa vary by less than 0.022 g between successive weigh- ings and are in general much more accurate (mean percent error = 0.38%). For Thais canaliculata the mean maximum percent error iS 1.15% for whole weights and 0.39% for immersed weights. For Thais emarginata these errors are about the same, 1.29% and 0.55% for maximum whole weight error and immersed weight error respectively. Tables 24 also indicate what the potential cumula- tive error might be when estimating body weight as described above. DISCUSSION The principal advantage to length as a measure of gastropod size is the comparative ease with which it may be obtained. Caliper measurements of shell length even to an ac- curacy of 0.1 mm require only a few seconds and with practice are repeatable to 0.2 mm. They are also readily obtainable in the field with a minimum of disturbance to the animals. Weight measurements, on the other hand, to be of sufficient accuracy (and hence utility), almost invariably require that animals be brought back to the laboratory, thereby in- creasing the disruption of the animal's normal activity as well as requiring additional time to return them to the field. However, there are several limitations to shell length as a measure of animal size. First, aS gastropod shells age, the spires be- gin to erode. Spight (1974) consistently ob- served negative “growth” (change in shell length) over the winter in Thais lamellosa at Shady Cove which he recognized as being due to spire erosion (see also Spight et al., 1974). Second, as animals increase in size there is a progressively smaller change in length for a given change in body weight so given the limit on resolution of length change imposed by the repeatability of caliper meas- urements (0.1-0.2 mm), changes in weight will be more readily detected than changes in length. Third, in mature gastropods, where shell growth is almost negligible, body weight may still vary seasonally in association with spawning, reduced activity over the winter or increases or decreases in the food supply. These body weight changes would pass un- detected if only shell length is recorded. Fin- ally, if populations differ from each other in shell shape, or if there is shape variation among individuals within a single population, then a given length change in animals of the same initial length will be associated with dif- ferent changes in body weight. In the final analysis, the choice of weight or length to measure growth, if not set by logis- tical constraints, is determined by the kind of information desired. The correlation between log shell length and log body weight is gener- ally high for animals from a single population Tal NON-DESTRUCTIVE SHELL AND BODY WEIGHTS EE SSS 971 560'0 190 890 690 8/0 8/0 08'0 02:0 990 6850 850. 550 651 Ajuo цбиел MELO 22.0 080 980 08'0 08'0 8/0 0/0 000" 050 290 290 09 1ybiam + yBue7 ejeulbsewsa sıey| 81-0 9/00 Lt'O 05`0 [7'0 05`0 ¿yO Lv'0 8£ 0 Or'0 Or'0 050’ 920 232 Aluo yy6ue7 Z60'0 Ly'0 290 ro 9£ 0 vE'0 vE'0 ANA AAN 20 920 05 lybiam + чубие7 eyeynoyeued siey] en! 770`0 10 0 ЕО 50`0 100 ЕО 20`0 20:0. = 000 0607 Boom 500 62 Ájuo yi¡Bua7 €90'0 LLO 220 80'0 OL'O 8L'0 AO) ОО 600 VEO 8007 #200 09 lybiam + yIBU97 esojjawej sıeyl ES 2 as UESW 91/4 91/6 51/5 ct/S LL/S 6/S L/S 9/G G/G v/s pastajal jusu]eal] JSQUINN (Aep/ujuow) p81810981 ajeg P91910981 sjewue paseaja jo иоциодола A $158 релед шод anjen 1=} “UONBIASP PIEPUEJS = GS ‘8//92/+ UO paseaja элэм sjewiue pawey “(M,SS.221 ‘М,/г.8) ‘W'S'N ‘UOIBUIUSEM ‘зрие| уепг UBS ‘PUEIS| UBUIPeag je aseajai Buimo¡jo, sÁep зпомел ио sdnosb зиэидеед OM) JO цэеэ шоц P818A0981 SjeWIUe payeW JO SUODIOdOIY ‘G AIGVL 72 PALMER TABLE 6. Regressions of body weight change (Y) on shell length change (X). Shell lengths are in mm. Body weights are measured in grams. N = number of individuals. 7. lam. = Thais lamellosa. T. can. = T. canali- culata. T. em. = T. emarginata. Length change Regression Regression number Species Mean Range N equation R° 1 T. lam." —0.02 —O!5: to! 1:6 11 Y = 0.071X — 0.0425 0.1985 2 T. lam.2 —0.02 -10to 1.8 17 Y = 0.072X — 0.0434 0.3117 3 T. lam.3 6.43 3:310 10:3 27 Y = 0.045X — 0.0024 0.5635 4 T. lam.4 9.31 4.8 to 12.2 24 Y = 0.057X — 0.0865 0.7540 5 Tecan! 124 -0.3t0 4.1 28 Y = 0.018X + 0.0785 0.1178 6 T. em.! 1.24 — 07.1095.3 40 Y = 0.039X + 0.0061 0.7063 7 T. em.2 2.01 0.310797 9 Y = 0.025X — 0.0019 0.7575 1Deadman Island, field growth. 2Point George, Lopez Island, field growth. 3Collected from False Bay, San Juan Island; grown in cages. 4Collected from Turn Rock, San Juan Islands; grown in cages. (regressions 11-13, Table 1), indicating that length can provide a fairly reliable estimate of body weight. In addition, the correlation be- tween change in length and change in body weight can also be relatively high for animals in the field and in cages, as long as they are increasing in size (regressions 3, 4, 6 and 7, Table 6; but note regression 5). Consequent- ly, if growth rates are positive and of moderate magnitude relative to spire erosion, and if shell shapes are essentially the same, then length can provide an adequate measure of body size. If populations of different shape need to be compared, a single destructive correlation of body weight on length for each population can permit comparisons between them based on length measurements alone. If, on the other hand, there is a possibility that animals may lose weight and it is important to detect such a loss, or if there are significant differences in the rates of spire erosion among populations being compared, length measurement alone may lead to much lower resolution of growth rate differences. In principle this technique would be appli- cable to any organism composed of two major components of differing specific gravity. Weights of the organism first in a medium whose specific gravity is very close to that of one of the components, and then in a second medium which may or may not be of similar specific gravity to the second component, will, via the appropriate regressions, permit inde- pendent estimates of both body components. In practice, such separation may be less fea- sible for other carbonate skeleton-producing invertebrates (sclerosponges, corals, brachi- opods, bryozoans, bivalves and echino- derms), since for many the removal of extra- visceral water will be difficult to accomplish reliably or without damaging the animal. LITERATURE CITED BAK, R. P. M., 1973, Coral weight increment in situ. A new method to determine coral growth. Marine Biology, 20: 45-49. BOYLE, P., SILLAR, M. & BRYCESON, K., 1979, Water balance and the mantle cavity fluid of Nucella lapillus (L.) (Mollusca: Prosobranchia). Journal of Experimental Marine Biology and Ecology, 40: 41-51. BRANCH, G. M., 1974, The ecology of Patella Linnaeus from Cape Peninsula South Africa. Ш Growth Rates. Transactions of the Royal Society of South Africa, 41: 166-193. FRANK, P. W., 1965, The biodemography of an intertidal snail population. Ecology, 46: 831-844. HAVINGA, B., 1928, The daily rate of growth of oysters during summer. Journal du Conseil In- ternational pour l'Exploration de la Mer, 3: 231- 245. KENNY, R., 1977, Growth studies of the tropical intertidal limpet Acmaea antillarum. Marine Bi- ology, 39: 161-170. LOWNDES, A. G., 1942, The displacement method of weighing living aquatic organisms. Journal of the Marine Biological Association of the United Kingdom, 25: 555-574. NISHII, T., 1965, Examination of the underwater weight used for measuring the growth of the Japanese Pearl Oyster, Pinctada martensii (Dunker). Bulletin of the National Pearl Re- search Laboratory, 10: 1264-1282. NON-DESTRUCTIVE SHELL AND BODY WEIGHTS 73 PALMER, A. R., 1980, A comparative and experi- mental study of feeding and growth in thaidid gastropods. Ph.D. Thesis, University of Wash- ington, Seattle, 320 p. RANDALL, H. A., 1964, A study of the growth and other aspects of the biology of the West Indian topshell, Cittarium pica (Linnaeus). Bulletin of Marine Science, 14: 424-443. REVELLE, A. & FAIRBRIDGE, R., 1957, Carbon- ates and carbon dioxide. Geological Society of America Memoirs 67, Vol. 1: 239-296. RHOADS, D. C. & YOUNG, D. K., 1970, The influ- ence of deposit-feeding organisms on sediment stability and community trophic structure. Journal of Marine Research, 28: 150-178. RICKEMISNE NE CALVIN: JS НЕБСРЕТЫ, J: W., 1968, Between Pacific Tides. Ed. 4. Stan- ford University Press, 614 p. SPIGHT, T. M., 1974, Sizes of populations of a marine snail. Ecology, 55: 712-729. SPIGHT, T. M., BIRKELAND, C. & LYONS, A., 1974, Life histories of large and small murexes (Prosobranchia: Muricidae). Marine Biology, 24: 299-242. STICKLE, W. В. & DUERR, Е. G., 1970, The effects of starvation on the respiration and major nutri- ent stores of Thais lamellosa. Comparative Bio- chemistry and Physiology, 33: 689-695. STROMGREN, T., 1975, Linear measurements of growth of shells using laser diffraction. Limn- ology and Oceanography, 20: 845-848. WALNE, P. R., 1958, Growth of oysters (Ostrea edulis L.). Journal of the Marine Biological As- sociation of the United Kingdom, 37: 591-602. WILBUR, K. M. & OWEN, G., 1964, Growth. In: Physiology of Mollusca (WILBUR, K. M. & YONGE, C. M., eds.) Academic Press, N.Y., 1: 211-242. YAMAGUCHI, M., 1977, Shell growth and mortality rates in the coral reef gastropod Cerithium nodu- losum in Pago Bay, Guam, Mariana Islands. Marine Biology, 44: 249-263. ZISCHKE, J. A., WATABE, N. & WILBUR, K. M. 1970, Studies on shell formation: measurement of growth in the gastropod Ampullarius glaucus. Malacologia, 10: 423-429. MALACOLOGIA, 1982, 23(1): 75-79 THE OCCURRENCE OF MERCENARIA MERCENARIA FORM NOTATA IN GEORGIA AND SOUTH CAROLINA: CALCULATION OF PHENOTYPIC AND GENOTYPIC FREQUENCIES Celeste M. Humphrey! and Randal L. Walker? ABSTRACT Genotypic and phenotypic frequencies of the notata form of Mercenaria mercenaria were calculated from data provided by four studies: two natural populations from Georgia, one from South Carolina, and one hatchery brood. Phenotypic frequencies calculated for each study ranged from 0.76% to 2.25%. Gene frequencies calculated by Maximum Likelihood Estimation were 0.04% to 0.11%. There were no significant differences between samples of natural popula- tions. The natural populations and the hatchery brood are not comparable. Research applica- tions using these rare alleles as useful markers for breeding experiments with this valuable commercial species are briefly described. Key words: Mercenaria mercenaria form notata; gene frequency estimation; natural occur- rence. INTRODUCTION The hard clam, Mercenaria mercenaria (Linne, 1758), is a commercially important bi- valve indigenous to the Atlantic and Gulf coasts of North America. The shell usually is a uniform whitish color which frequently ap- pears tan to dirty gray as a result of sediment stains (Abbott, 1974). Mercenaria mercenaria form notata (Say, 1822) is a rare variant of M. mercenaria distinguished by brown bands or zigzag lines on the surface of the shell. It oc- curs throughout the range of M. mercenaria from the Gulf of St. Lawrence to Florida and the northern Gulf of Mexico (Abbott, 1974). M. mercenaria notata was designated a subspecies of M. mercenaria until Chanley (1961) crossed M. mercenaria mercenaria and M. mercenaria notata in the laboratory, and genetic analysis of the resulting offspring demonstrated that the notata marking is an individual variation controlled by a single gene or by several closely linked genes inher- ited as a Mendelian unit (Paul E. Chanley, personal communication). The commonly ob- served notata pattern is the heterozygous form and individuals homozygous for the notata allele have solid brown bands (Chanley, 1961). These two forms are illus- trated in Fig. 1. METHODS Literature on the occurrence of the notata variant is as scarce as the animal itself. Few papers devoted to the distribution and density of populations of M. mercenaria have noted the incidence of notata individuals in those populations. Phenotypic and genotypic fre- quencies of the notata variant of M. mercen- aria Were calculated using three studies which recorded the occurrence of the notata form in their samples: Eldridge et al. (1976) on the South Carolina coast; Walker et al. (1980), an intensive survey of Wassaw Sound, Georgia; and Humphrey (in prep.) on the Georgia coast. Further information was provided by the analysis of a brood spawned in the Vir- ginia Institute of Marine Sciences Eastern Shore Laboratory from 300 individuals col- lected in Wassaw Sound. The samples from South Carolina were col- lected using patent hydraulic tongs, and those from Georgia were collected either with rakes or by examining quadrats by hand; therefore the collected specimens were at least two years old due to the inability to detect smaller individuals by these methods. In addition to age-related attrition, habitat conditions in these areas tend to disfigure the shells. Indi- viduals are commonly located in anaerobic 1Department of Zoology, University of Georgia, Athens, Georgia 30602, U.S.A. Present address: Marine Extension Service, P.O. Box 13687, University of Georgia, Savannah, Georgia 31406, U.S.A. 2Skidaway Institute of Oceanography, P.O. Box 13687, Savannah, Georgia 31406, U.S.A. (75) 76 HUMPHREY AND WALKER FIG. 1. Heterozygous (A) and homozygous (B) phenotypes of Mercenaria mercenaria form notata from the southeastern United States; high contrast (Kodalith) photograph showing representative individuals of each phenotype. mud which tends to blacken the shells, and shell erosion is common. These factors make notata harder to detect which leads to the as- sumption that the percentages calculated for natural populations are minimum estimates. RESULTS AND DISCUSSION Eldridge et al. (1976) estimated the fre- quency of notata clams in South Carolina estuaries using the data of Gracy (1974). Out of 11 sites where notata was found, 19 of the 1539 individuals were notata (1.23%); sample percentages ranged from 0.71% to 2.17%. In the two studies of clam populations on the Georgia coast, the frequency of the notata variant was recorded. Walker et al. (1980) ina survey of clam populations from Wassaw Sound near Savannah, Georgia, examined fifty-seven stations which yielded a total of 2339 clams; of these, 13 samples contained a total of 18 individuals with notata markings (Table 1). Frequency of the notata variant cal- culated for this area using all stations was 0.77%; sample percentages ranged from 0.36% to 12.5%. Humphrey (in prep.) took six samples of clam populations from a wider geographic area. One sample was from Wassaw Sound, three were from Sapelo Island (approximately 48 km to the south), one was from Doboy Sound near the south end of Sapelo Island, and one was from St. Simon's Island. The total latitudinal range was approximately one degree (110 km). Of 1881 individuals, 21 (1.12%) were identified as notata, sample percentages ranged from 0.29% to 1.73% (Table 2). The hatchery spawned brood contained 234 (2.25%) notata individuals in a sample of 10,378; this is higher than the overall percentage in either state. Gene frequency estimation The best calculation for the estimation of gene frequencies is the Maximum Likelihood Estimation: pe V2B N p = gene frequency estimate; A = number of individuals homozygous for the gene esti- mated by p; B = number of heterozygotes; N GENETICS OF MERCENARIA MERCENARIA FORM NOTATA ur TABLE 1. Samples of Mercenaria mercenaria collected from the survey of Wassaw Sound, Georgia. 1977-79 (Walker et al., 1980). Sample* N No. notata % notata Gene frequency (q) 1 19 1 7.69 .0385 2 25 1 4.00 .0200 3 281 1 0.36 .0018 4 174 1 0.57 .0029 5 101 1 0.99 .0050 6 125 2 1.60 .0080 Y 82 1 1.22 .0061 8 25 2 8.00 .0400 9 8 1 12.50 .0625 10 71 2 2.82 .0141 11 51 1 1.96 .0098 12 144 3 2.08 .0104 13 12 1 8.33 .0417 Total (pooled) 2339 18 0.77 .0039 "Samples containing notata. TABLE 2. Samples of Mercenaria mercenaria collected from the Georgia coast (Humphrey, in prep.). Sample N No. notata % notata Gene frequency (q) St. Simon Is. 300 2 0.67 .0033 Doboy Sound 346 6 1273 .0087 Sapelo Is. | 310 5 1.61 .0081 Sapelo Is. Il 294 5 1.70 .0085 Sapelo 1$. Ш 348 1 0.29 .0014 Cabbage ls. (Wassaw Sound) 283 2 0.71 .0035 Total 1881 21 1.12 .0056 = total number of individuals in the sample; q = frequency estimate of the other allele; the variance of this estimate: pq 2N is the minimum variance for gene frequency estimation (Speiss, 1977). There were no homozygous notata detected in the Georgia studies, although individuals were closely ex- amined. The South Carolina study made no distinction between genotypes of notata indi- viduals; therefore, this method of gene fre- quency calculation, which requires the num- ber of heterozygotes, cannot be used for all three studies. If it is impossible to distinguish one of the homozygous genotypes, the maxi- mum likelihood estimation provides another method for calculation of gene frequencies. o Y/A N p — gene frequency whose homozygote is distinguishable; A = number of individuals of that homozygote; N = total number of individ- uals in the sample. The variance of this esti- D M edie, Gre AN ON + AN is larger than the minimum variance by a quantity proportional to q” (Speiss, 1977). For the data in these studies the estimate of q is so small that there is very little difference between estimates; for example the variance (a”) of p for the South Carolina study is: pq 15 Ba ON x .0000020018 and aN .0000020081 For these three studies the gene frequency estimates are the same by either method when calculated to the fourth decimal place. mate, South Carolina; Wassaw Sound; Georgia Coast; Hatchery brood; p = .9938, q = .0062, o” = .0000020 p = .9961, q = .0039, o? = .0000008 p = .9944, q = .0056, o” = .0000015 p = .9889, q = .0113, 0” = .0000005 78 HUMPHREY AND WALKER If the populations are in Hardy-Weinberg equilibrium then the expected numbers of each genotype can be calculated from the formula: Np” + 2Npq + Nq = М For the South Carolina Survey, the values (expected values given in parentheses) are: 1520(1519.97) + 19(18.97) + 0(.06) = 1539(1539.00) For Wassaw Sound: 2321(2320.79) + 18(18.17) + 0(.04) = 2339(2339.00) For the second Georgia Survey: 1860(1859.99) + 21(20.95) + 0(.06) = 1881(1881.00) In none of these studies is the expected occurrence of the homozygous notata form equal to one, so it is quite possible that there was an absence of homozygous notata rather than a failure to detect them. Chanley (per- sonal communication) has found homo- zygotes in nature but they are rare. In that case the simple gene counting formula: Asa. N can be used. This formula was used to calcu- late the gene frequencies in Tables 1 and 2. Heterogeneity between samples was calcu- lated by testing the phenotypic frequencies. South Carolina: 1520 19 = 1539 Georgia (Wassaw Sound): 2321 18 = 2339 Georgia coast: 1860 21 = 1881 Total: 5701 58 5759 With two degrees of freedom, the chi-square value was 2.35, 0.5 < p < 0.3. It is not mean- ingful to calculate expected numbers for the hatchery brood, nor can it be compared to the natural populations, since both the percent- age of notata and the actual number of indi- viduals that spawned to form the parental generation for this sample are not known. The notata variant is the only morphological character inherited as if controlled by a single gene that has been found in M. mercenaria. The uses of such a marker are numerous. For example, one application would be the mark- ing of offspring from controlled matings to determine their subsequent success. Since the allele is rare in nature, releasing notata larvae and checking adjacent populations in subsequent years could supply much needed information on larval migration, provided that migration is not affected by this allele. Meas- ures of genetic parameters such as gene fre- quency and heterozygosity are of particular interest to population geneticists working with electrophoretic alleles because they provide a known morphological marker for comparison with data derived from enzyme markers. Morphological markers of known genetic basis have many possible uses in practical shellfish research. As more species are suc- cessfully reared in hatcheries, use of these markers to identify individuals whose genetic composition results in desirable traits such as rapid growth, increases the efficiency of con- trolled breeding programs. Newkirk (1980) not only confirmed the genetic control of shell color in Mytilus edulis but also found that the gene for this color polymorphism was linked to growth rate. Research with the genetics of several com- mercial bivalves has been done (Kraeuter et al., in prep.; Innes & Haley, 1977), but very little work has been done with the notata vari- ant of M. mercenaria. Study of the natural oc- currence of notata in conjunction with the many studies done on populations of M. mercenaria would not require much effort, and the potential usefulness of this marker makes it desirable to acquire more knowledge about its characteristics in nature. ACKNOWLEDGEMENTS The authors thank the University of Georgia Marine Extension Service for its support and the use of its facilities, Dr. James W. Porter, Dr. D. M. Gillespie, Dr. Anne Michelle Wood, and Michael Castagna for critical evaluation of the manuscript. Special thanks go to Paul Е. Chanley for encouragement and assist- ance. This study was partially funded by Georgia Sea Grant NA79AA-D-00123. REFERENCES CITED ABBOTT, R. T., 1974, American Seashells. Ed. 2. Van Nostrand Reinhold, New York, 663 p. CHANLEY, P. E., 1961, Inheritance of shell mark- ings and growth in the hard clam, Venus mercenaria. Proceedings of the National Shell- fisheries Association, 50: 163-169. GENETICS OF MERCENARIA MERCENARIA FORM NOTATA 79 ELDRIDGE, P. J., WALTZ, W. & MILLS, H., 1976, Relative abundance of Mercenaria mercenaria notata in estuaries in South Carolina. Veliger, 18: 396-397. GRACY, R. L., 1974, Management and develop- ment of the shellfish industry in South Carolina. Annual Report Project 2179D in cooperation with National Marine Fisheries, 22 p. HUMPHREY, С. M. in preparation, Population genetics of Mercenaria mercenaria from Massachusetts to Florida. INNES, D. J. & HALEY, L. E., 1977, Inheritance of a shell color polymorphism in the mussel. Journal of Heredity, 68: 203-204. KRAEUTER, J. N., CASTAGNA, M., WALL, J. В. 8 KARNEY, R., in preparaton, Shell color and rib number inheritance in the bay scallop Argo- pecten irradians. NEWKIRK, G. F., 1980, Genetics of shell color in Mytilus edulis L. and the association of growth rate with shell color. Journal of Experimental Marine Biology and Ecology, 47: 89-94. SPEISS, E. B., 1977, Genes in Populations. Wiley, New York, 780 p. WALKER, R. L., FLEETWOOD, M. A. & TENORE, K. R., 1980, The distribution of the hard clam, Mercenaria mercenaria (Linne) and related predators in Wassaw Sound, Georgia. Georgia Marine Science Center Technical Report 80-8, 59 p. ws MALACOLOGIA, 1982, 23(1): 81-82 LETTERS TO THE EDITORS ON SIBLING SPECIES AND GENETIC DIVERSITY IN FLORIDA GONIOBASIS Two studies by Chambers (1978, Malaco- logia, 17: 157-162; 1980, ibid., 20: 63-81) are pioneer contributions to the understand- ing of genetic diversity and relationships in a genus that has always perplexed system- atists. Goniobasis is confusing because of the numerous described species (ca. 500), their morphological variation within and between populations, and the frequent instances of parallel evolution of shell characters among distantly related taxa. Chambers shows that in some instances genetic diversity exceeds shell differentiation and that taxonomic classi- fication based on isozyme analysis may be more nearly objective than classifications based on shell features. However, three items in these papers need reexamination. Chambers (1980: 65) identifies a species from the Ichetucknee River as G. athearni, which is correct. Earlier (1978) he identified this same population as a sibling of G. floridensis. In the more recent paper he shows that G. floridensis and G. athearni are distantly related genetically within Gonio- basis, but he continues to regard the Iche- tucknee population a sibling species of flori- densis. Because the Ichetucknee population of athearni is distantly related genetically to floridensis, they cannot be sibling species. Sibling species are very closely related taxa that have diverged sufficiently genetically to be reproductively incompatible, but are morphologically hardly distinguishable. If one relies solely on shell criteria, there are numer- Ous instances in which unrelated species of Goniobasis are hardly distinguishable from each other. Also, there are many instances in which species of Goniobasis are barely dis- tinguishable by shells from species of Pachychilus, Potodoma, Lithasiopsis, Juga, Semisulcospira, Doryssa (Pleuroceridae), or Tarebia, Melanoides, and Hemisinus (Thiaridae). Clearly similar shells in these species do not mean that they have not yet diverged in shell characters. More likely, simi- lar shell types have evolved due to similar strong environmental selection factors acting independently on unrelated species. G. athearni and G. floridensis are not sibling spe- cies in light of the genetic data that Chambers presents (1980, Tables |, Il). Yet this point is not satisfactorily addressed in the 1980 paper, and most likely it would not be noticed by authors gleaning the literature for ex- amples of sibling speciation. It is also appro- priate to note that G. athearni had not been found in the Ichetucknee River among numer- ous collections in the MCZ, UMMZ, and the FSM made there prior to 1970. Chambers’ study population was found only at a single station, a roadside park where U.S. Highway 27 crosses the Ichetucknee River. Currently the species has spread widely in the Iche- tucknee and has moved downstream into the confluent Santa Fe River. Apparently the Ichetucknee River population was introduced there after 1970. Biogeographic interpreta- tions based on its presence there are highly tenuous. Chambers (1980: 65) states that he found G. floridensis in Holmes Creek and the Choc- tawhatchee River, but did not find G. clenchi, which was reported from these places in ear- lier literature. However, his illustrations (p. 76; fig. 7 C-D) of specimens from the Chocta- whatchee River are typical clenchi. This is significant because he identified the Chocta- whatchee population as an isozymatically dis- tinct species of “G. floridensis.” The error is compounded on pp. 74-75, where he shows that “G. floridensis” and G. dickinsoni inter- grade in shell sculpture in the Chipola River but not in the Choctawhatchee River. The in- terpretations he gives are that “floridensis” and dickinsoni are far more diverse genetical- ly than they actually are, and biological speci- ation has occurred that previously was not recognized. Both interpretations are incorrect. Serious errors could be perpetuated in the scientific literature because of this misidentifi- cation. Finally, Chambers interprets taxonomic di- versity in Goniobasis by criteria established for the Drosophila willistoni group. In this group, semi-species, sibling species, and subspecies are definable on the basis of ap- parent genetic distances. Such is not the case in Goniobasis. Chambers’ data give inconclu- sive results. In some instances (clenchi- floridensis, clenchi-dickinsoni) interbreeding does not occur even though genetic distances are much less than between subspecies in (81) 82 THOMPSON Drosophila. Also a morphologically very dis- tinct geographic subspecies of floridensis (Wacasassa River) is not genetically distin- guishable from adjacent populations of typical-appearing floridensis. In other in- stances apparently great genetic distances occur between morphologically indistinguish- able populations of floridensis. Clearly the criteria of genetic distances used in the Drosophila willistoni group for defining spe- cific and infraspecific categories are not ap- plicable to taxonomic hierarchies within Goniobasis. The use of genetic distances by Chambers for showing phylogenetic relationships in Florida Goniobasis is a major advance in the development of an objective set of criteria for determining phylogenetic relationships. How- ever, the data base he presents is too ambig- uous at this point to permit clear-cut applica- tion to taxonomic classification and hierar- chies in Goniobasis. Fred G. Thompson Florida State Museum University of Florida Gainesville, Florida 32611, U.S.A. À. MALACOLOGIA, 1982, 23(1): 83-86 SIBLING SPECIES AND GENETIC DIVERSITY IN FLORIDA GONIOBASIS: A REPLY The snails of the genus Goniobasis are a quandary to the fastidious taxonomist and a source of fascination to the evolutionary bi- ologist. The complex patterns of geographic variation observed in the Florida Goniobasis led me to apply electrophoretic methods (also known as allozyme, isozyme or isoenzyme methods) in evolutionary and systematic studies of these snails (Chambers, 1978, 1980). Although one of the earliest applica- tions of electrophoretic methods to a taxo- nomic problem dealt with mollusks (Davis & Lindsay, 1967), these techniques are new to many malacologists. | am taking this oppor- tunity to respond to some specific questions that Thompson has raised about the applica- tion of these methods in studies of the Florida Goniobasis. These questions will be dis- cussed in the same order that Thompson has presented them. Sibling species Thompson's definition of sibling species re- quires that sibling species be very closely re- lated and show little genetic divergence. This definition is not consistent with the term sib- ling species as it was originally defined (Mayr, 1942) or with its subsequent use in the litera- ture. Thompson's definition is closest to that of Mayr (1969: 411) who defines sibling spe- cies аз“... closely related species which are reproductively isolated but morphologically identical or nearly so.” Mayr does not state in that work how “closely related” sibling spe- cies must be, but he does in other writings (Mayr, 1942, 1963, 1970, 1976). For example, in Mayr’s (1942) work that originally defined sibling species, he states that such species do not need to be phylogenetic siblings (p. 151). In a more recent review (1976: 513- 514) Mayr emphasizes that sibling species need not be genetically very similar and that their morphological similarity is not merely be- Cause these species are in statu nascendi and have not had time to diverge. Isoenzyme Studies of sibling species (Webster & Burns, 1973; Ayala, 1975; Nei, 1975: 185; Avise, 1976; Ryman et al., 1979) have confirmed Mayr's view that sibling species may display profound genetic differences. According to Avise (1976: 112), “Morphological similarity of (83) sibling species belies their large genetic dif- ferences.” Chambers’ (1978) use of the term sibling species for the genetically distinct spe- cies at the Ichetucknee River is therefore consistent with the original definition of the term and its subsequent usage in the litera- ture. Thompson should realize from reading Chambers [1978, 1980, and in press (prelimi- nary draft sent to Thompson in September, 1980)] that | recognize that shell features alone can be unreliable for classification and that convergence is common in gastropod shell features. For example, Chambers (1978: 161) specifically noted that the occurrence of sibling species in the Ichetucknee River “... perhaps is due to convergent evolution in shell sculptural characters.” One does not know for certain whether the Ichetucknee River species are convergent since no fossils are known that would indicate the physical appearance of the ancestors of either spe- cies. Thompson's list of genera that show simi- larity in shell sculpture are of evolutionary interest but not particularly relevent to a dis- cussion of sibling species since these taxa are mainly allopatric, with the unfortunate ex- ceptions of Tarebia granifera Lamarck and Melanoides tuberculata (Muller) which have managed to colonize some portions of North America (Dundee, 1974). The sibling species concept is clearly most relevant in areas where sympatric species may be confused with one another. | have observed T. granifera in the Ichetucknee River and M. tuberculata in Alexander Springs, Lake County, Florida. The shells of these species are not difficult to distinguish from those of the native Gonio- basis found at these localities. The large genetic distance between the sib- ling species in the Ichetucknee River and its significance were discussed in Chambers (1978). Since that work was cited in Cham- bers (1980) little would have been achieved by repeating that discussion in the latter paper, which was also published in Malaco- logia. Goniobasis floridensis (Reeve) and the population whose identification was sug- gested as Goniobasis athearni (Clench & Turner) by Chambers (1978, 1980) from the 84 CHAMBERS Ichetucknee River were referred to as sibling species in the 1980 paper because some in- dividuals from the Ichetucknee River are still difficult to safely place in the correct species without electrophoretic identification. A Recent Introduction at Ichetucknee River? Thompson's suggestion that the Ichetuck- nee River population referred to G. athearni by Chambers (1980) was recently introduced is a valid alternative hypothesis to that of Chambers (1978), which indicated that the seemingly large geographic distance between the Ichetucknee River population and related forms in the Apalachicola River drainage had parallels in other aquatic organisms, and that the headwaters of these two systems drain adjacent areas in Georgia (Jackson, 1975). The following additional observations on the introduction hypothesis are offered. Both G. floridensis and the G. athearni form were ob- served at or near the head of Ichetucknee Springs in 1974; this is 3 km distant from the site of the collection reported on in Chambers (1978). Thompson does not indicate how numerous are the pre-1970 collections of Goniobasis from the Ichetucknee River. One 1970 collection (Ohio State Museum no. 336) yielded both species at the site of the Cham- bers’ 1978 study. The only lot in the U.S. Na- tional Museum (USNM 515795) consists of three abraded shells collected in 1926 which cannot be identified with any certainty. Thompson suggests that G. athearni was introduced to the Ichetucknee River after 1970 and states, without presenting evidence, that this species has since spread downriver into the Santa Fe River. | collected this spe- cies at Blue Springs on the Santa Fe River (Gilchrist Co.) in January 1974. This site is about 20 km upstream from the confluence of the Ichetucknee River and the Santa Fe River. Movement of this species this far up- stream between 1971 and 1974 requires rates of locomotion exceeding measured rates for Goniobasis (Krieger & Burbanck, 1976; Mancini, 1978). If G. athearni was in- troduced to the Ichetucknee River as late as Thompson states, either these snails have dispersed on their own with extraordinary speed or there must have been a second in- troduction to account for the species’ pres- ence in Blue Spring early in 1974. Electrophoretic data (Chambers, 1980) in- dicate that the Ichetucknee River G. athearni has four alleles not detected in the Chipola River G. athearni sample. These alleles could be used as markers for locating the source population of the Ichetucknee River popula- tion if indeed it is a recent introduction. The introduction hypothesis could not be sup- ported if these alleles could not be found in reputed source populations. Biogeographic interpretations are usually tenuous; they are also an essential part of any study of geographic variation. The introduc- tion hypothesis for the Ichetucknee population is provocative, testable to some extent, and may well prove true; but its formulation does not absolve the researcher from the responsi- bility of providing prudent biogeographic in- terpretations such as that suggested by Chambers (1978). G. floridensis and G. clenchi The population from site 15 (Chambers, 1980) was referred to G. floridensis after com- parison with the following: 1. The population at Holmes Creek (Vernon, Washington Co.) which was referred to G. floridensis by Clench & Turner (1956). Thompson does not cite the earlier literature which, he maintains, reports G. clenchi from Holmes Creek. | am not aware of any pub- lished records of G. clenchi from that stream. 2. Four paratypes of G. clenchi (USNM 861587). 3. Description, with figure, of G. clenchi (Goodrich, 1924). 4. A lot of G. clenchi (USNM 668077) acquired from the Museum of Comparative Zoology, Harvard University. 5. Redescriptions of G. floridensis and G. clenchi by Clench 8 Turner (1956). These descriptions are very similar. G. floridensis shells are described as having a stronger peripheral cord, more indented suture, and less flattened whorls than shells of G. clenchi. The population sampled at site 15 was vari- able, but most individuals were closer to G. floridensis material and were well within the range of variation of the population at Holmes Creek at Vernon which was identified by Clench & Turner (1956) as G. floridensis. It is likely that G. clenchi and G. floridensis in the Choctawhatchee River Drainage are con- specific since populations referred to those different species overlap in shell characters. Chambers (in press) describes chromosomal divergence between G. floridensis from LEMERS TO! THE EDITOR 85 Holmes Creek and G. floridensis in the Chipola River drainage. Electrophoretic data Thompson makes several errors and misin- terpretations in his discussion of the electro- phoretic data bearing on the relationships be- tween G. clenchi and G. floridensis. Cham- bers (1980) did not refer to the site 15 popula- tion or any other population as “an isozy- matically distinct species,” but rather has warned against making such judgments based on electrophoretic evidence alone (p. 73) and discussed the conditions under which electrophoretic data are relevant to taxonomic work (pp. 75-76). Contrary to Thompson's assertion, Chambers’ (1980) data indicate lit- tle divergence between the site 15 sample and other Florida panhandle samples of G. floridensis. The interpretations that Thomp- son attributes to Chambers, “. . . ‘floridensis’ and G. dickinsoni are far more diverse ge- netically than they actually are, and biological speciation has occurred that previously was not recognized. ..,” do not appear in Cham- bers (1980) and are contrary to the views ex- pressed in that paper on G. dickinsoni and the Florida panhandle samples of G. floridensis. G. dickinsoni and G. floridensis were, in fact, shown to have diverged very little, which sup- ports an interpretation that these species are closer than recognized in previous literature (Clench & Turner 1956), and not more di- verse, aS maintained by Thompson. Including the site 15 sample with G. floridensis has a negligible effect on the total gene diversity within the complex because that sample, hav- ing only one unique allele (Acph-1'"), was very close to other samples. Semispecies, sibling species, and sub- species in the Drosophila willistoni group were not defined by genetic distances, as Thompson states, but by reproductive rela- tionships (Ayala et al., 1974). Those authors calculated genetic distances in order to meas- ure genetic divergence between samples at varying levels of taxonomic divergence. These data were chosen as a standard of comparison for describing genetic divergence between Goniobasis populations because the D. willistoni group is one of the best-studied groups where both taxonomic divergence, based on reproductive relationships, and genetic divergence between populations are known. Chambers (1980: 72-73, 75-76) dis- cussed the possible values and shortcomings of these comparisons. Chambers (1980) did not advocate the strict application of these comparisons for making taxonomic decisions but stated that “Electrophoretic methods alone cannot generate classification . . .” and warned that “... taxonomic decisions based solely on such comparisons [with D. willistoni] will often be erroneous” (p. 72). Chambers (1980: 76) clearly indicated the subjective nature of these comparisons when studying closely related species. Thompson objects that electrophoretic data (Chambers, 1980) do not distinguish a G. floridensis “subspecies” (which apparently has not been described) in the Waccasassa River. This form intergrades with a form with the “standard” G. floridensis sculpture pattern in the Wekiva River (Chambers, 1980, fig. 5). Both morphological and electrophoretic evi- dence indicate that the Waccasassa River “subspecies” is not so “very distinct” as Thompson believes. One must take care when using the term “phylogenetic” when referring to electropho- retic data. In particular, fig. 3 of Chambers (1980) is a dendrogram that summarizes the genetic distance values in table 2 and is not strictly a phylogeny, a term which is not used in that paper. Conditions that must be met for such a dendrogram to represent a phylogeny are set down in Nei (1975). The subjectivity of applying of electropho- retic data to taxonomic classification was Clearly indicated in the discussion of system- atic implications in Chambers (1980). Few sets of data, whether from electrophoretic, karyotypic, or morphological studies, permit “clear cut” determinations of taxonomic groupings of related allopatric populations. Electrophoretic, karyotypic, and morpho- logical data each give the researcher a view of a different aspect of an organism's biology and should not be viewed as antagonistic to each other. Each data set, however, should be evaluated according to criteria that take into account the theoretical constraints of the particular technique employed. Chambers (1978, 1980) identifies some of these criteria for the application of electrophoretic data to the classification of riverine snails. Given the enormous taxonomic problems presented by Goniobasis it would be unwise to discard any set of data that supplies a measure of an im- portant aspect of evolutionary divergence. 86 CHAMBERS REFERENCES CITED AVISE, J. C., 1976, Genetic differentiation during speciation, In: AYALA, F. J. (ed.), Molecular Evolution. Sinauer Associates, Sunderland, Massachusetts, p. 106-122. AYALA, F. J., 1975, Genetic differentiation during the speciation process. In: DOBZHANSKY, T., HECHT, M. K. & STEERE, W. C. (eds.), Evolu- tionary Biology, 9: 1-78, Plenum Press, New York. AYALA, F. J., TRACEY, M. L., HEDGECOCK, D. & RICHMOND, R. C., 1974, Genetic differentiation during the speciation process in Drosophila. Evolution, 28: 576-592. CHAMBERS, S. M., 1978, An electrophoretically detected sibling species of “Goniobasis flori- densis.” Malacologia, 17: 157-162. CHAMBERS, S. M., 1980, Genetic divergence be- tween populations of Goniobasis (Pleuro- ceridae) occupying different drainage systems. Malacologia, 20: 63-81. CHAMBERS, S. M., in press, Chromosomal evi- dence for the parallel evolution of shell sculpture pattern in Goniobasis. Evolution. CLENCH, W. J. & TURNER, R. 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. DAVIS, G. M. & LINDSAY, G. K., 1967, Disc elec- trophoretic analysis of molluscan individuals and populations. Malacologia, 5: 88-111. DUNDEE, D. S., 1974, Catalog of introduced mol- lusks of eastern North America (North of Mex- ico). Sterkiana, 55: 1-37. GOODRICH, D., 1924, Some old pleurocerids and a new one. Nautilus, 38: 43-48. JACKSON, D. R., 1975, A Pleistocene Graptemys (Reptilia: Testudines) from the Santa Fe River of Florida. Herpetologica, 31: 213-219. KRIEGER, K. A. & BURBANCK, W. B., 1976, Dis- tribution and dispersal mechanisms of Oxytrema (= Goniobasis) suturalis Haldeman (Gastro- poda: Pleuroceridae) in the Yellow River, Georgia, U.S.A. American Midland Naturalist, 95: 49-63. MANCINI, E. R., 1978, The biology of Goniobasis semicarinata (Say) [Gastropoda: Pleuroceridae] in the mosquito creek drainage system, South- ern Indiana. Ph.D. Dissertation, Univ. of Louis- ville. MAYR, E., 1942, Systematics and the Origin of Species. Columbia University Press, New York, 334 p. MAYR, E., 1963, Animal Species and Evolution. Belknap Press, Cambridge, Massachusetts, 797 p. MAYR, E., 1969, Principles of Systematic Zoology. McGraw-Hill, New York, 428 p. MAYR, E., 1970, Populations, Species and Evolu- tion. Belknap Press, Cambridge, Massachusetts, 453 p. MAYR, E., 1976, Evolution and the Diversity of Life. Belknap Press, Cambridge, Massachusetts, 721 p. NEI, M., 1975, Molecular population genetics and evolution. North-Holland, New York, 299 p. RYMAN, N., ALLENDORF, F. W. & STAHL, G., 1979, Reproductive isolation with little genetic divergence in sympatric populations of brown trout (Sa/mo trutta). Genetics, 92: 247-262. WEBSTER, T. P. & BURNS, J. M., 1973, Dewlap color variation and electrophoretically detected sibling species in a Haitian lizard, Anolis brevirostris. Evolution, 27: 368-377. Steven M. Chambers Office of Endangered Species, U.S. Fish and Wildlife Service Washington, D.C. 20240, U.S.A. and Department of Biology, George Mason University Fairfax, Virginia 22030, U.S.A. AMERICAN MALACOLOGICAL UNION SYMPOSIUM: FUNCTIONAL MORPHOLOGY OF CEPHALOPODS Organized by Clyde F. E. Roper, A.M.U. President William H. Hulet, Rapporteur Louisville, Kentucky 21 July, 1980 MALACOLOGIA, 1982, 23(1): 87 INTRODUCTION TO THE INTERNATIONAL SYMPOSIUM ON FUNCTIONAL MORPHOLOGY OF CEPHALOPODS Clyde F. E. Roper, Organizer National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560, U.S.A. When the broad diversity in morphology of cephalopods is considered, it seems surpris- ing that so little is known about the functional nature of morphological features in this highly evolved group of mollusks. The reasons for this, perhaps, are rooted in the past when the very problems that plague us today were even more unsurmountable: collections of speci- mens, especially of oceanic forms, were woe- fully inadequate; maintaining cephalopods alive for behavioral observations was almost impossible, with the occasional exception of Octopus and Sepia; microscopic and other laboratory instrumentation was primitive or non-existant. With the exception of a short burst of activity in the very late 19th Century and early 20th Century, led by the German master Professor Carl Chun, strikingly few morphological studies exist, and even fewer deal with functional aspects. In an attempt to begin to alleviate the gap in our knowledge, an international symposium entitled “Functional Morphology in Cephal- opods” was conducted during the 46th An- nual Meeting of the American Malacological Union held at Louisville, Kentucky (21 July, 1980). The symposium papers, presented by teuthologists with a broad range of research interests, reflect the current activities in func- tional morphology. Both the papers and the vigorous discussions that followed pointed out that many problems and gaps still exist and suggested new directions in research. Inter- estingly, a majority of the papers dealt with features of the skin—chromatophores, irido- phores, photophores, and epithelium. Other topics covered salivary glands, the mantle complex, and the “kidneys” as a unique habi- tat for parasites. These studies reflect the sig- nificant advances that have been made in the past two decades in our ability to capture and maintain cephalopods alive for observation and experimentation and to explore ever-finer structural details through transmission and scanning electron microscopy. This symposi- um provided a forum for the beginning of the modern age of functional morphology of cephalopods. (87) MALACOLOGIA, 1982, 23(1): 89-119 THE FUNCTIONAL ORGANIZATION OF CHROMATOPHORES AND IRIDESCENT CELLS IN THE BODY PATTERNING OF LOLIGO PLEI (CEPHALOPODA: MYOPSIDA) Roger T. Hanlon The Marine Biomedical Institute, University of Texas Medical Branch, 200 University Boulevard, Galveston, Texas 77550, U.S.A. ABSTRACT The tropical arrow squid Loligo plei is capable of a wide range of colorful body patterns used for camouflage and visual communication. These body patterns are produced by the appear- ance of different combinations of chromatic and postural components. This paper deals solely with the chromatic (color) components, which are broken down into groups of morphological units comprising one or two types of elements: chromatophores and iridescent cells located in the dermis of the skin. These elements are not distributed evenly across the body surface, but are organized into one of seven different morphological units. The chromatophore elements in L. plei are yellow, red or brown and may be arranged into any one of four morphological units. Each morphological unit is described in terms of its distribution across the skin surface as well as the sizes, colors and static morphological array (horizontal and vertical) of chromatophores within the unit. The iridescent cells—iridophores and reflector cells —are reflective elements positioned subjacent to chromatophores. They are arranged into three units that may appear either yellow- ish-white, pink or green; the horizontal distribution of each unit is described. The seven types of morphological units provide the capability of producing 16 differently shaped and colored com- ponents. Each chromatic component is described in terms of the functional morphology of the units that constitute it. In L. plei, the expression of chromatic components is a result of (1) the particular static morphological array of elements within differently constructed units, and (2) the selective nervous excitation of those morphological units. The resultant color of each component is postulated to aid in camouflage in shallow water, but in intraspecific signalling it is contrast and configuration of the components (and not color) that are important in transmitting information. Three aspects of sexual dimorphism related to chromatic behavior are demonstrated in L. plei, and chromatic expression within the genus Loligo is reviewed and compared. Key words: squid; Loligo; skin morphology; chromatophores; iridophores; color change; be- havior; camouflage. INTRODUCTION Camouflage and visual communication are highly advanced in cephalopods and they have developed a wide range of body pat- terns to convey this visual information. It is my purpose in this paper to illustrate in Loligo plei (Blainville, 1823) how the anatomically fixed elements of pattern and color change—the chromatophores and iridescent cells—are dif- ferently organized and distributed in the dermis of certain body areas to produce an assortment of spots, streaks, splotches, etc. used in body patterning. While a great deal of work has been performed on the structure and physiology of the chromatophore organ over the past 150 years, few studies have been aimed at describing the range of body patterns that the chromatophores and irides- cent cells produce. Body patterns of selected species of the genera Octopus (Cowdry, 1911; Hanlon & Hixon, 1980; Packard & Sanders, 1969, 1971; Packard & Hochberg, 1977; Warren et al., 1974), Eledone (Boyle & Dubas, 1981), Sepia (Holmes, 1940, 1955) and Sepioteuthis (Boycott, 1965; LaRoe, 1970; Moynihan, 1975) have been examined in varying detail. In these studies it has gen- erally been assumed for each species that (1) the anatomical array of chromatophores and iridescent cells is fixed, that is, it is recurrent and constant over the entire body surface, and (2) the appearance of body patterns is produced principally through the selective nervous excitation of groups of differently colored chromatophores. In Loligo plei the organization of chromatophores and irides- cent cells is not constant; it differs in specific (89) 90 HANLON areas of the body. Therefore, the actual ap- pearance of specific chromatic components results not only from selective nervous excita- tion of differently colored chromatophores, but from the size and distribution (horizontal and vertical) of chromatophores and iridescent cells on different areas of the body. To describe the organization of the chro- matic elements | have followed the hierarchi- cal classification of body patterning devel- oped by Packard & Sanders (1969, 1971) and Packard & Hochberg (1977), and further clari- fied by Packard (1982) in this symposium. Essentially they have determined that body patterns of cephalopods are produced by dif- ferent combinations of chromatic, textural and postural components. The chromatic com- ponents, in turn, may be broken down into morphological and physiological units com- prising two elements: chromatophores and iri- descent cells in the dermis of the skin. In this paper | will describe (1) the two elements of patterning in L. plei, (2) the functional mor- phology of seven types of morphological units, and (3) sixteen resultant chromatic components. The remaining aspects of body patterning in L. plei will be the subject of a future publication. These aspects include de- scriptions of postural components, behavioral movements and body patterns as well as the behavior associated with them. Loligo plei is commonly referred to as the tropical arrow squid and is distributed on the continental shelf and slope on the east coasts of North and South America from North Caro- lina, throughout the Gulf of Mexico and Carib- bean Sea, to Brazil (LaRoe, 1967; Voss et al. 1973; Cohen, 1976; Whitaker, 1978). It is a pelagic, schooling species that is abundant off the Texas coast in depths of 20 to 75 m, and it is commonly caught together with the long- finned squid Loligo pealei and Lolliguncula brevis (Rathjen et al., 1979). Loligo plei was originally described by Blainville (1823), and later made the type species of the genus Doryteuthis by Naef (1912). There is dis- agreement regarding its generic status (see Cohen, 1976), but | will rank Doryteuthis as a subgeneric designation and hereafter refer to this species as Loligo plei. MATERIALS AND METHODS Squids were collected under night lights with soft-mesh dipnets and transported in shipboard seawater tanks to laboratory aquaria. They were maintained in round 2 m diameter tanks or 10 m long raceway sys- tems (Hanlon et al., 1978). Over 500 indi- viduals were observed between 1975 and 1980. Squid survival in both systems aver- aged 20 days, with some animals surviving up to 84 days. Individuals ranged in size from 3 mm ML (mantle length) to 348 mm ML. The average sizes for squids in this study were 90 mm ML for females and 175 mm ML for males; maximal mantle lengths were 211 mm ML for females and 348 mm ML for males. Most observations and photographs were made through viewing ports built into the sides of the tanks, and the remainder from above the tanks. Close-up observations of the skin were made at 6, 12 and 25x under a Wild stereomicroscope, usually with the squids narcotized in 1% ethanol in seawater. Color videotapes were made through the microscope by adapting the video camera lens to the third viewing eyepiece and lighting the subject with fiber optics. All photographs were taken with flashed light using a Nikon F camera, 24, 55 and 105mm lenses, and a Nikon bellows unit for high magnification photographs of the skin. Frozen sections of pieces of skin were prepared to confirm the depth distribution of differently colored chro- matophores. Small skin patches were excised from live squids, frozen and cut in a Bright Model FS/FAS/M cryostat, then stained with a polychrome (Toluidine Blue and Basic Fuchsin), dehydrated with ethyl alcohol, cleared in xylene and mounted with Eukitt mounting media on glass slides. Frozen sec- tions were viewed and photographed with a Leitz Orthoplan compound microscope. Colors were standardized according to the Pantone Matching System (PMS, 1966) and in the text the PMS color number is given in parentheses. ELEMENTS OF BODY PATTERNING The elements of body patterning are the smallest individual morphological entities in the skin that are responsible for displaying color. In Loligo plei there are three colors of chromatophores—yellow, red, brown—and two types of iridescent cells—iridophores and reflector cells. The chromatophores produce pigmentary colors of relatively long wave- lengths by reflection of light after differential absorption in the pigment cell. The iridescent cells produce structural color of a wide spec- FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 91 trum of wavelengths by reflecting ambient light from the iridosomal platelets within the cell. It is the combined effects of all these ele- ments that result in color and pattern. For a more detailed account of how elements pro- duce color the reader is referred to Packard & Hochberg (1977: 193). Their descriptions are based upon Octopus, and some important dif- ferences in the skin of Octopus and Loligo should be pointed out: (1) chromatophores of Loligo are much larger and fewer per unit area, (2) in general Loligo has three color classes of chromatophores—yellow, red, brown—but no black (melanophores) or orange chromatophores as in Octopus, (3) in Loligo no white leucophores (Packard 8 Sanders, 1971) have been identified, (4) there are no grooves in the skin of Loligo that partition the chromatophores and iridophores into discrete, recognizable skin patches such as those found in Octopus, and (5) the skin of Loligo is smooth and can not produce the papillae that result in different skin textures in Octopus. In summary, the skin of Loligo is less differentiated and complex than that of Octopus, but the range of body patterns that Loligo can produce is nevertheless extensive. Chromatophores Many aspects of the morphology, ultra- structure and physiology of the chromato- phores of Loligo have recently been eluci- dated in considerable detail by Florey (1966, 1969), Cloney & Florey (1968), Florey & Kriebel (1969), Mirow (1972a) and Weber (1968, 1970, 1973). Chromatophores are controlled by nervous stimulation (Florey, 1966, 1969; Florey & Kriebel, 1969). The neu- ral connections of individual chromatophores are complex yet poorly understood, but it is known that individual chromatophores receive multiple innervation (Florey, 1966, 1969; Weber, 1968, 1970, 1973) and that certain groups of chromatophores may be linked by their musculature (Froesch-Gaetzi & Froesch, 1977). Boycott (1961), Young (1974, 1976, 1977) and Messenger (1979) have contrib- uted to an understanding of some of the neu- ral connections of chromatophores in the brain, but the details of the way the central nervous system controls the chromatophores are by no means clear. Fig. 1A shows the three basic color series of chromatophores found in Loligo plei: yellow (PMS 135), red (PMS 201) and brown (PMS 491). These should be regarded strictly as broad categories of color, since in living squids gradations of all three colors are seen. For instance, there are pink-colored chroma- tophores (PMS 204) of the same size range and depth as the red, but it is not clear wheth- er these represent a separate color series of chromatophores or nascent red chromato- phores with less pigment. An important con- sideration in color determination is that the quality of light reflected from chromatophores changes as the pigment granules are dis- persed or concentrated within the pigment container. For instance, when the chromato- phores are only partially expanded, yellow appears more orange (PMS 152), red ap- pears darker red or almost brown (PMS 492) and brown appears almost black (PMS 497). When frozen sections of chromatophores are examined microscopically there is consider- able gradation of color among all the chroma- tophores and distinction between red and brown is often difficult. The pigments of cephalopod chromatophores have not been identified biochemically and the number of color series of chromatophores is unclear (Packard & Hochberg, 1977). For Loligo, the general concensus among other workers (Cloney & Florey, 1968; Schelling & Fioroni, 1971; Mirow, 1972a; Weber, 1973; Williams, 1909) and myself is that there are only three color series—yellow, red and brown—with pink being considered in the red series. Iridophores and Reflector Cells In contrast to chromatophores, the mor- phology, ultrastructure and physiology of the iridescent cells of cephalopods are less well understood. Brocco & Cloney (1980) have re- cently investigated the morphology of Octo- pus reflector cells and compared them with sepioid and teuthoid iridophores. In accord- ance with their nomenclature the term reflec- tor cell will be used here to describe the uni- form iridescence of various eye parts of Loligo plei, since presumably their ultrastructure is similar to the reflector cells of the eye parts of Loligo forbesi (Denton & Land, 1971) in which the broad surfaces of the reflective platelets face the integument. Iridophores are non- pigmented cells that reflect, diffract and scat- ter light. They are characterized by their in- tracytoplasmic platelets of high refractive index and they are distinguished from reflec- tor cells by the orientation of their platelets, which are oriented on edge relative to the sur- face of the skin. Some aspects of iridophore 92 HANLON morphology in the squids Loligo opalescens and Loligo pealei have been clarified by Mirow (1972b) but because of complex mem- brane systems within the iridophore she was not able to construct a three-dimensional model of the cell. The physics of diffraction so thoroughly known in fish platelets (cf. Denton, 1970) has not been applied in studies of iri- dophores in cephalopods (but see Denton & Land, 1971) and therefore the mechanisms of reflection and the functional morphology of the iridophores in producing specific color ef- fects is yet unknown. These aspects are be- yond the scope of this work and | will limit my description to a macroscopic visual analysis of the distribution, color and transient appear- ance of iridophores and reflector cells in living L. plei. There appear to be three expressions of iridescence in Loligo plei. First is the bright, smooth reflective layer of reflector cells that is always visible on the eye parts and ink sac. Second is the shingled appearance of irido- phore splotches on the mantle and fins. Third is the shingled appearance of a continuous sheet of iridophores on the mantle collar, the dorsal head region and the first pair of arms. Mirow (1972b) described short and long irido- phores in Loligo but it remains to be deter- mined what their interactions are and which visual effects they produce. It is noteworthy that iridophores are present on the mantle of the hatchlings of Loligo (Arnold, 1967; Schell- ing & Fioroni, 1971; McConathy et al., 1980), but they are not conspicuous and they are more numerous on the ventral surface. Histo- logical sections of skin show that there are also iridophores in the dermis of the lateral (Fig. 9) and ventral surfaces of adult L. plei but they are not obvious in the living animals. CHROMATIC UNITS: THE STATIC MORPHOLOGICAL ARRAY The chromatic units contain the anatomical- ly fixed elements used in color change, i.e. the static morphological array. The chromatic units involving chromatophores are recogniz- able to a human observer both in their retracted and expanded states because of the relatively uniform and recurrent distribution of elements within them. The units of Loligo are not delineated by skin grooves as in Octopus, > FIG. 1. Aspects of body patterning in the tropical arrow squid Loligo plei. A) Standard Discoid Units, the most common and widespread arrangement of chromatophores. Brackets indicate the approximate dimensions of a single morphological unit. Note the considerable overlap among chromatophores of the three colors. The chromatophores are not yet at maximal expansion. From the dorsal mantle of a male, 220 mm ML. Line represents 1 mm. B) Tentacular stalk (center) and second arm spot (right) of a male squid, 165 mm ML. On the sialk note the Yellow-Brown Discoid Units and the areas devoid of chromatophores. Arrow points to one of the 7-8 spots characteristic of the chromatic component TENTACULAR STRIPE AND SPOTS. The second arm spot is produced by selective nervous excitation of a group of Modified Discoid Units. Note the overlap among all the brown chromatophores. Line represents 2 mm. C) Frozen sections of a Standard Discoid Unit. Yellow chromatophores lie above reds, and reds lie above browns. From a male squid, 108 mm ML. Line represents 200 um. D) Frozen section of a Lateral Flame Unit. Yellow chromatophores lie above browns, and browns lie above reds. Arrow indicates the layer of iridophores deeper in the dermis (see Fig. 9). Purple layer just below the epidermis is an excess of polychrome stain. From a male squid, 108 mm ML. Line represents 200 um. E, F, G) Different phases of expansion of the same group of Lateral Flame Units that make up the chromatic component LATERAL FLAME. See explanation in Discussion. From a male squid, 220 mm ML. Line represents 1 mm. H) Lateral view of a male Loligo plei, 201 mm ML, showing the following chromatic components to a nearby squid: TENTACULAR STRIPE AND SPOTS, ARM SPOTS, LATERAL FLAME, MID-VENTRAL RIDGE and DORSAL ARM IRIDOPHORES. I) Top view of a male Loligo plei, 175 тт ML, showing many of the same chromatic components in (H) towards a squid on its left (towards the bottom in the photograph). Note that LATERAL FLAME is expressed unilaterally only on the side towards the other squid. DORSAL MANTLE SPLOTCHES are prominent on the mantle, and the posterior portion of the conspicuous white testis is very lightly shaded by the chromatic component SHADED TESTIS. DORSAL ARM IRIDOPHORES is weakly expressed, and ARM SPOTS and STITCH- WORK FINS are shown bilaterally. J) Female Loligo plei, 121 mm ML, showing the chromatic components RING and DORSAL MANTLE SPLOTCHES. RING is produced by the selective nervous excitation of Standard Discoid Units on the mantle and fins. Note the silver-green appearance of the Reflector Cell units on the sclera of the eyeball. =. LA IN € 5 ee > Ze - к > < y En Be! Е : : es D'or 4 29. АР x ae Dek DY SOND on fOr T Perle sten tg ai р = oe Lc pa e 7 o RES ЕО == = с | FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 93 but they are nevertheless recognizable due mainly to the typed array of adjoining units. This “standard part” arrangement (Riedl, 1978) is theorized to be a result of the proc- esses of morphogenesis that insure the regu- lar spacing of dermal elements during onto- geny (Packard, 1982). The morphological units described herein must be distinguished from the physiological units described by Packard (1982); the physiological units repre- sent the concept of neurophysiological motor units that reflect nerve or muscle connection, not the anatomical placement of elements. This paper emphasizes the arrangement of the chromatic units, because the static mor- phological array of elements provides the framework for both the morphological and physiological unit structures, and these are the very bases of color and pattern change. In the initial stages of this study it appeared to me that observed differences in compo- nents were due mainly to selective nervous excitation and that the chromatophores were generally evenly distributed over the entire body surface. However, this could not com- pletely account for the wide variations of ap- pearance in different chromatic components, and upon closer observation it became ap- parent that in certain body areas the arrange- ment of chromatophores and iridophores was organized to produce specific visual effects. Accordingly | have grouped the two types of elements into seven chromatic units to facili- tate the description of sixteen chromatic com- ponents. This has inevitably involved some arbitrary decisions, but as a guide | have al- ways chosen the most obvious recurring ar- rangement of elements into which chromatic components can be most easily divided, and those that can be readily identified on a macroscopic level. Standard Discoid Unit This unit is characterized by a large brown chromatophore surrounded by a circlet of red chromatophores, with small yellow chromato- phores interspersed throughout (Fig. 1A). The term “discoid” describes the generalized di- mensions of the morphological unit, i.e. it is circular in the horizontal plane (1-4 mm in diameter) and flat in the vertical plane (50- 150 um). The static morphological array of retracted chromatophores in this unit is shown in Figs. 2A and 4. The Standard Discoid Unit is the most common unit on the body surface (Fig. 3). Adjoining units are most easily recog- nized by the large, central brown chromato- phore, but some red chromatophores in the circlet may be shared by the adjoining unit. On the dorsal mantle, most of the larger units have a large iridophore splotch centered be- neath the center brown chromatophore (Figs. 10, 11). In this type of unit brown chromato- phores are larger than reds, and reds are larger than yellows; this holds true both at full retraction and full expansion (Table 1). Their vertical distribution in the dermis, as shown by frozen sections, is: browns deepest, reds in- termediate and yellows shallowest (Fig. 1C). When retracted, each color class is generally identifiable by its minimal chromatophore diameter; e.g. all yellows are smallest and of the same approximate size. A great deal of Overlap occurs when all chromatophores in the array are at full expansion. Yellows and reds overlap with all colors, while browns overlap only with reds and yellows. Browns do not overlap with other browns because they are so widely and evenly spaced that it pre- cludes overlap. In adult squids the average Standard Discoid Unit comprises one brown chromatophore, approximately 7 to 13 reds, and 15 to 30 yellows; it may or may not have one underlying iridophore splotch. Modified Discoid Unit This unit is similar in most respects to the Standard Discoid Unit, except that some of the red chromatophores in the circlet are re- placed by brown chromatophores similar in size to the reds (Fig. 2B). The larger propor- tion of brown chromatophores per unit pro- duces an overall darker shade when full ex- pansion is achieved. The sizes of the chroma- tophores and their vertical distribution are similar to the Standard Discoid Unit. These units are found in two areas (Fig. 3): (1) over bright underlying organs, such as the highly reflective sclera on the back of the eye (Fig. 5) or the white testis of males, and (2) where very dark spots are produced, as on the second and third arms of males (Fig. 1B). Yellow-Brown Discoid Unit This unit is characterized by the absence of red chromatophores. Yellows are the same size as in all other types of units, while the browns are slightly larger than yellows, though not as large as browns in Standard and Modified Discoid Units. Overlap occurs commonly and yellows always lie above the 94 HANLON Chromatophore legend : Yeilow FIG. 2. Static morphological arrays of the four types of chromatophore units, each drawn to the same scale (line equals 1 mm) to accurately depict the distance between retracted chromatophores. All illustrations were traced from video tape recordings of live squids. More detailed descriptions are given in the text. A) Standard Discoid Units. Each brown chromatophore represents the center of one morphological unit. The central unit is the largest and presumably oldest one, and the brown chromatophore at its center is larger than surround- ing browns both in the retracted and expanded states. Drawn from the dorsal mantle of a male, 170 mm ML. B) Modified Discoid Units. Identical to (A) except that some brown chromatophores have replaced reds in the circlet around the larger center brown. C) Yellow-Brown Discoid Units, distinguished by the absence of red chromatophores. The brown chromatophores are smaller than browns in Standard and Modified Discoid Units. This illustration was traced from one group of adjacent units that made up one “dash” of the chromatic component STITCHWORK FINS around the periphery of the fins of a male, 185 mm ML. D) Lateral Flame Units, characterized by longitudinally oriented rows of yellow and brown chromatophores. The browns in these units are nearly as small as the yellows. Drawn from the lateral mantle of a male, 150 mm ML. NOTE: The diameters of the dots and circles are not all drawn accurately to the scale bar. In (A) and (B), the yellow chromatophore dots must be increased in diameter by 1.5 to give an accurate scale size of chromato- phores in the retracted state (see Table 1). In (C) and (D), the sizes are approximately accurate as shown. Distribution of this unit is confined to two areas (Fig. 3): (1) on the tentacular club and browns. The static morphological array of re- tracted chromatophores is shown in Fig. 2C. The brown chromatophores are evenly dis- persed relative to one another and form the centers of the units, which are 1-2 mm in di- ameter. Smaller yellows are evenly inter- spersed amid the remaining areas. There are generally 5-10 yellows and one brown per unit. stalk of juvenile and adult squids, and (2) around the periphery of the fins of males larger than approximately 105 mm ML. On the inside surface of the tentacular stalk, the units are grouped into seven or eight distinct “spots,” while the adjacent areas are without chromatophores (Fig. 1B). On the fins, the FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 95 Lateral Tentacle Ventral cm E Е FIG. 3. Distribution across the skin surface of the four types of chromatophore units in male Loligo plei. Iridescent cell unit types are not shown. In the top figure the dashed line represents the position of the white testis. (A) Standard Discoid Units. B) Modified Discoid Units. C) Yellow-Brown Discoid Units. D) Lateral Flame Units. E) Areas devoid of chromatophores. The drawing of the outside portion of the left tentacle is slightly enlarged. Females have Standard Discoid Units all over the body surface except for Modified Discoid Units over the top of the eyes and no chromatophores on the underside of the fins. Illustration by Charles Moen. units are grouped 2-3 units wide into a con- tinuous line, 2-5 mm inside the margin of the fin, that is present around the periphery of the fins. Lateral Flame Unit This unit, found only in mature males larger than approximately 70 mm ML, is markedly different in nearly every respect from the three types of discoid units. Adjoining units are or- ganized into long, longitudinally oriented rows of yellow, red and brown chromatophores separated by areas relatively free of chroma- tophores except for widely spaced large red chromatophores (Figs. 2D and 6). The rows are 1.5-3.0 mm wide. Yellow chromatophores are smallest, browns are the same size or 96 HANLON — - , — SEHE .. CZ LADA PL e . L 4 > > ‘ . . e > . = u bd u . > à . = ® e - + - u à ES e oa Ss o > = 7 . . . . = = . . ES ( . . ° a e . 7 = e . e ’ e . u 5 . = e o. . e . LA a o = . . a . . - o . ; o o à À 4 — FIG. 4. Photographs of Standard Discoid Units on the mid-dorsal mantle of live squids. Left: Retracted units from a male squid, 190 mm ML. Compare unit structure with the diagrammatic sketch in Fig. 2A. Line represents 1 mm. Right: Expanding units on a male squid, 147 mm ML. Brown chromatophores at the center of each morphological unit are identified by a small white dot at their centers. This photograph represents one transient appearance of very rapidly expanding chromatophores. There are many examples of units in different states of expression. Browns in similar states of expansion are members of the same physiological units (e.g. the four browns in the upper right hand corner connected by line). Note the many retracted chromatophores (smallest dark dots) that are members of physiological units different from those that are expanded. Line represents 1 mm. TABLE 1. Sample chromatophore diameters (in mm) of Loligo plei. Measurements taken from Standard Discoid Units on the dorsal mantle of a freshly dead male, 88 mm mantle length. Measurements were made with a micrometer eyepiece at 25x. Yellow Red Brown n = 20 n = 20 20 Retracted mean 0.07 0.14 0.24 range (0.040. 12) (0.08-0.24) (0.12-0.48) Expanded mean 0.35 0.68 1.07 range (0.12-0.60) (0.36-1.12) (0.80-1.52) Expansion factor mean 5.00x 4.86 x 4.46 max. (11.00x) (14.00x) (10.67x) slightly larger than yellows, and reds are very large. Their vertical distribution in the dermis differs from the three discoid units. It is: yellows shallowest, browns intermediate and reds deepest (Fig. 1D). When fully expanded, yel- lows always overlap browns and browns al- ways overlap reds. When retracted, yellows and browns are nearly identical in size and color and are generally indistinguishable; reds are larger and identifiable. At full expan- sion, the yellow and brown chromatophores form long distinct rows that are accentuated by the adjacent clear areas that have relative- ly few large red chromatophores. The distribu- FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 97 FIG. 5. Photographs of Modified Discoid Units on the dorsal head region over the eyes (arrows) of live squids. The rounded tip in the bottom of each photograph is the anterior portion of the mantle. Left: Retracted units on a male, 190 mm ML. The area within the dashed lines represents the approximate distribution of the Modified Discoid Units. Compare with the diagrammatic sketch in Fig. 2B. Line represents 2 mm. Right: Expanded Modified Discoid Units over the top of the left eye of a male squid, 147 mm ML. The expansion of these units results in the chromatic component SHADED EYE. The units are not yet fully expanded, but it can be seen that they are effective in blocking the reflection from the silvery Reflector Cells on the sclera of the back of the eyeball. Line represents 2 mm. tion of these units is outlined in Fig. 3 and may be noted in photographs in Figs. 1H, 23 and 27. The long, flame-like streaks of yellow and brown chromatophores bifurcate and later merge as they progress down the mantle. The exact placement of the rows differs slightly from animal to animal and may be as distinc- tive as a fingerprint, but on a macroscopic scale the visual effect is the same. The most startling visual expression results when the yellows, browns and reds in the rows are maximally expanded while the reds between the rows in the clear areas remain retracted. It is noteworthy that the rows are delineated by the distribution of the small yellow and brown chromatophores, and not in any de- gree by the large red chromatophores that are fairly evenly distributed throughout the entire area. This becomes obvious if one concen- trates on observing only red chromatophores in photographs or if red chromatophores are selectively traced from a photograph (Fig. 7). The sequence in Figs. 1E, 1F and 1G shows the progressive neural recruitment of yellows and browns that result in the flames. The de- lineation of rows is further illustrated in the transition zone between the Standard Discoid Units and the Lateral Flame Units (Fig. 8). The transition is accomplished by (1) the dis- appearance of the large, deeply placed brown chromatophores that are at the center of Standard Discoid Units, (2) the appearance of a continuous, longitudinal row of small con- spicuous brown chromatophores shallower in the dermis, just below the yellows, (3) the re- organization of the small, evenly distributed yellow chromatophores of the Standard Dis- coid Units into distinct rows and (4) the 98 HANLON . > E LL o . . ES = .. - ee ” o = pr e 0 . = ps e » . . . . ee 2 e e » © 4 La ‘ E à 2 . > . o o . . . pa . » ra . . . ° . . ee . > . 2 я . = ee к © > . 5 e » e-e 7 ‘ o . . «fe . . . . om . . o . . . e . . ® | A Ру . . e pa VARO © > e e a = e . ee .. 4 . . . = > > >: e bad ” > o FIG. 6. Photograph of retracted Lateral Flame Units that make up the chromatic component LATERAL FLAME. From the lateral mantle of a male squid, 190 mm ML. Compare with diagrammatic sketch in Fig. 2D and photographs of expanded units in Figs. 1E, 1F, 1G, 8 and 23. Line represents 1 mm. gradual reorganization of red chromato- phores from circlets in the Standard Discoid Unit into an evenly dispersed array of red chromatophores that are approximately the same size range as reds found in Standard Discoid Units. Reflector Cells This term refers collectively to the reflective layers on the ink sac and around the eye—the eyelids, iris and the sclera. These cells reflect light constantly and are visually conspicuous except when the overlying chromatophores are maximally expanded. The iris, eyelids and ink sac reflect silver, while the sclera on the top and back of the eyeball reflects silver- green (Fig. 1J). FIG. 7. Distribution of red chromatophores traced from a mirror image of Fig. 1F. Red chromato- phores are fairly evenly distributed among Lateral Flame Units and are not aligned into distinct rows as are yellow and brown chromatophores. Iridophore Splotches These splotches are distributed somewhat evenly over the dorsal mantle and fins and they act as a unit by showing the same de- gree of prominence. The splotches are dis- tinct, irregulary shaped, 1-3 mm in diameter, and they appear either pink, green or yellow- ish-white (Fig. 11). They are not always visi- ble, and they show varying degrees of promi- nence even when the overlying chromato- phores are retracted. The splotches are gen- erally centered beneath a large brown chro- matophore of the Standard Discoid Unit (Figs. 10411): Iridophore Sheets These iridophores appear as a continuous sheet of pink, green or yellowish white irides- cence that has a visual effect similar to that of the splotches. Like the splotches, they are not visible at all times. These units are distributed over the dorsal head region of both sexes. They are particularly evident on the first pair of arms in adult males, and they periodically render the entire arm brightly iridescent (Fig. 1H). Adult females have fewer and less con- spicuous iridophores on those arms by com- — FIG. 8. Transition zone between Standard Discoid Units of the dorsal mantle (extreme upper right hand corner) and Lateral Flame Units of a male squid, 220 mm ML. See text for explanation. Line represents 1 mm. FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 100 HANLON FIG. 9. Electron micrograph of iridophores from the dermis of the lateral aspect of the mantle of a male, 120 mm ML. Iridosomal platelets (black arrows) within iridophore cells are clearly evident. The ovoid inclusions in the cells lying between the iridophores and the muscle cells in the upper right hand corner are unknown structures that may be light-scattering organelles similar to leucophores (S. Brocco, per- sonal communication, 1980). White arrow points toward the epidermis. Black area on bottom left is the grid border. Line represents 2 um. Micrograph courtesy of M. R. Villoch. parison. On the mantle, iridophores appear to be densely clustered around the anteriormost margin of the mantle, and when expressed they appear as a collar of pink or green (Fig. 12). CHROMATIC COMPONENTS Components are, by definition, the parts that make up a pattern. Therefore, a body pat- tern can be described in terms of the relative position and intensities of the components regularly present (Packard & Sanders, 1971). Chromatic components come and go, leaving no trace in the static morphological array (Packard & Hochberg, 1977). Thus emerges FIG. 10. Dorsal iridophore splotch on the fin of a male squid, 190 mm ML. The shingled appearance is characteristic of the splotches. The overlying chromatophores are mostly retracted, and the white dot marks the center brown chromatophore of a Standard Discoid Unit under which the splotch is centered. The striated fin muscles appear as faint, oblique lines oriented top right to bottom left. Line represents 1 mm. FIG. 11. Video tape recording from the dorsal man- tle of a live male, 150 mm ML, showing the DORSAL MANTLE SPLOTCHES generally positioned beneath the brown chromatophores (marked with small white dots) of Standard Discoid Units. Line repre- sents 1 mm. FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 101 17 7+ За FIG. 12. Male squid, 201 mm ML, showing the chromatic component DORSAL ARM IRIDOPHORES. In this case the chromatophores of the first pair of arms are retracted while the remainder on the head and arms is expanded, resulting in the boldest ex- pression of the component. The Iridophore Sheets on the head can be seen, as well as DORSAL MAN- TLE COLLAR (arrow), DORSAL MANTLE SPLOTCHES and the outline of Lateral Flame Units on the man- tle. one of the general principles of patterning in cephalopods: that components vary in the ex- tent of their expression from barely percepti- ble to fully expressed (Packard & Hochberg, 1977). The changes in expression reflect neural control and, because of the “fine tun- ing” that nervous control provides, the compo- nents may change abruptly (in tenths of sec- onds according to Hill & Solandt, 1934) or they may grade one into another. For this rea- son it is difficult to make definitive classifica- tions of patterns, but by describing the dis- crete chromatic components that are com- monly repeated, one can obtain a basis for describing body patterns. The 16 chromatic components described here are based upon a wide size range of animals, and future work may reveal addi- tional components. Some components are specific to a certain size range of squids, some to a specific sex, and some are com- mon to all members of the species. When possible the components are described in terms of the types of units that constitute them. CLEAR This component (Fig. 13) is characterized by the complete retraction of all the chroma- tophores, which renders the squid trans- lucent. The chromatic units of Reflector Cells on the eye parts are very prominent, and Iri- dophore Splotches may be present. Many of the internal organs are visible through the translucent skin. When viewed from above, FIG. 13. Mating pair of Loligo plei. The female (top, 93 mm ML) shows the CLEAR chromatic component with DORSAL MANTLE COLLAR and DORSAL MANTLE SPLOTCHES. The white spot in mid-mantle is the ovary and the small dark spot below it (arrow) is the red accessory nidamental gland. The male (175 mm ML) shows ARM SPOTS, DORSAL ARM IRIDOPHORES, DORSAL MANTLE COLLAR and DORSAL MANTLE SPLOTCHES. The large white organ is the well-developed testis. 102 the testis of mature males and the ovary of mature females are clearly visible, as are the esophagus, stomach and mid-gut gland dur- ing feeding. The buccal mass and systemic heart are also visible. Laterally, the ovary, egg mass and red accessory nidamental gland (Figs. 13, 20) are clearly visible in sexually mature females. This component has been observed in males and females throughout the entire size range, from hatching (3 mm ML) to adult (348 mm ML). ALL DARK In this component all of the chromatophores are expanded and the squids appear uniform- ly dark over the entire body surface (Fig. 14). The overall hue is a deep red-brown. ALL DARK has been observed on squids of both sexes throughout the entire size range, in- cluding hatchlings. Three variations have been noted. First, small, young squids (less than 70 mm ML) produce a lighter red color than adults be- cause: (1) they have fewer chromatophores per unit area, (2) red and brown chromato- phores of young squids are very similar in size and (3) there appear to be more pink-colored chromatophores than in the adult. Second, a dark mottled variation was observed on three occasions in an adult male (140 mm ML) sit- ting motionless on the bottom. The mottling was produced when small, irregular splotches of chromatophores were retracted, while the units of Iridophore Splotches became very distinct. Third, a female (93 mm ML) once produced a pulsating, rippling wave of chro- matophores over the entire dorsal surface for 15 mins, ranging from light to dark red. — 2 er HANLON RING Considerable variation appears in this com- ponent, but it is always distinguished by three or four transverse rings around the mantle. The most common form consists of four man- tle rings (Figs. 1J, 15), the second anterior- most ring being normally the most prominent and well developed. This second ring is lo- cated anterior to the fin insertion, approxi- mately 3 of the way to the anterior margin. The first ring is next most prominent and lies midway between the anterior mantle margin and the second ring. These first two rings completely encircle the mantle and in most cases are best developed dorsally. The third ring bisects the anterior fin insertion and gen- erally extends transversely over the dorsal V2 to 23 of the mantle. The fourth ring is actually a transverse line of expanded Standard Discoid Units across the middle of the fin, but it gives the visual impression of a ring extending around the narrow posterior mantle. Regions between the rings and over the head and arms are normally clear, and Iridophore Splotches are usually prominent. This com- ponent has been observed on females of 34- 136 mm ML and males of 85-252 mm ML. The most common variation is the three- ring component, which is found in certain indi- vidual squids (mostly females) and is char- acterized by a single ring that is situated be- tween where the second and third rings would be in a squid with the four-ring component. Another variation occurs when the Modified Discoid Units between the eyes become dark while the remainder of the head and arms re- main clear; when viewed from above this pro- duced the visual effect of an extra ring. In still Y L À = и urn FIG. 14. A small female (67 mm ML) in ALL DARK. Note that the center brown chromatophores of some Standard Discoid Units on the mantle are only partially expanded, and the underlying iridophores of DORSAL MANTLE SPLOTCHES can be seen. FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 103 FIG. 15. Male (115 mm ML) swimming near the bottom in the RING chromatic component, with DORSAL MANTLE SPLOTCHES prominent. The head and arms are often clear when RING is shown. FIG. 16. Same female as in Fig. 13 (93 mm ML) in the three-ring variation of RING, with SHADED EYE and DORSAL MANTLE SPLOTCHES. The first and third mantle rings are poorly developed in this photograph. another variation, females in the three-ring component sometimes appear to have only one ring when the first and last are weakly developed due to the graded retraction of cer- tain elements in Standard Discoid Units on the periphery of the ring (Fig. 16). Any of the aforementioned variations may be seen with uniformly dark head and arms (Figs. 1J, 15). ACCENTUATED TESTIS The squid is uniformly dark, as in ALL DARK, except for a completely clear area of retracted chromatophores around the conspicuous, white testis of males (Fig. 17). In most cases the clear area extends well beyond the outline of the testis. This component has been ob- served in sexually mature males of 39-285 mm ML. Two variables regulate the apparent size and conspicuousness of the testis: (1) the size of the clear area and (2) the degree of expression of the surrounding Standard Dis- coid Units. Fig. 18 shows the clear area at its greatest dimension. In contrast, Fig. 17 shows a small clear area over the testis with maxi- mally expanded surrounding units. The com- bination that produces the most conspicuous component is produced when the clear area over the testis is largest and the chromato- phores of adjacent Standard Discoid Units are completely expanded. SHADED TESTIS In this component a male squid is clear ex- cept for the area over the testis, which is 104 HANLON FIG. 17. A male (133 mm ML) in ACCENTUATED TESTIS, ARM SPOTS and SHADED EYE. The testis of this male is not well developed and the clear area over the testis is not at its largest dimension. FIG. 18. A male (205mm ML) guarding recently laid eggs and showing ACCENTUATED TESTIS, SHADED EYE, ARM SPOTS (third arms), TENTACU- LAR STRIPE and weakly developed DORSAL ARM IRIDOPHORES. The clear area over the testis is at its greatest dimension. shaded by a latticework of brown chromato- phores (Figs. 1J, 19). The latticework is presumably produced by the selective nervous excitation of brown chromatophores in the Modified Discoid Units that cover the testis. Usually the center browns are only Ya expanded while the browns in the circlet are near maximal ex- pression; a few yellow chromatophores of each unit may also appear. These units pro- duce a stippled effect that obliterates the high- ly Conspicuous white testis by reducing its luminance. The interspersed yellows reduce the luminance of the bright testis so that it approximates that of adjacent areas on the dorsal mantle. SHADED TESTIS has been ob- served in sexually mature males of 20-285 mm ML. Only minor variations have been noted. In smaller squids (approximately 29-50 mm ML) the shaded area is more diffuse and covers a greater area than the testis. In squids this small the units are not as well defined and do not selectively delineate the testis. The shaded area may vary from the testis proper to a larger area such as that included in AC- CENTUATED TESTIS. It is noteworthy that on a few rare occasions females (50-90 mm ML) have shown a com- ponent, somewhat similar to SHADED TESTIS, that generally covers the posterior Ya of the mantle including the area of the ovary. | do not consider it a distinct component because (1) the ovary is not generally very conspicuous through the mantle, (2) the area covered by chromatophores is not well defined and looks more like DORSAL STRIPE (see below), and (3) it has only rarely been seen. In fact, fe- males appear to have no counterpart com- ponent homologous to ACCENTUATED TESTIS in males. SHADED EYE The highly reflective sclera of the back of the eye is shaded by Modified Discoid Units FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 105 FIG. 19. Same male as in Fig. 17 (133 mm ML) shown minutes later in SHADED TESTIS and DORSAL MANTLE SPLOTCHES. A Kit .- E = 4 iret FIG. 20. Female (89 mm ML) in DORSAL STRIPE and DORSAL MANTLE SPLOTCHES. Arrow indicates the bright red accessory nidamental gland. Note the brightness of the sclera of the eye when the overlying chromatophores are retracted. that are situated on the dorsal head region directly above them (Figs. 5, 14, 16, 17, 18, 27). When these units are fully expanded the browns overlap and provide complete shad- ing of the eyeballs. This component may be expressed in varying degrees that shade the eyes correspondingly. It has been seen on females and males as small as 40 mm ML. DORSAL STRIPE The squid is basically translucent with a dif- fuse longitudinal stripe of expanded chroma- tophores extending the entire length of the dorsal mantle (Fig. 20). The stripe is centrally aligned on the dorsal mantle and extends laterally Уз to 23 of the mantle width. The stripe is characterized by Standard and (in males) Modified Discoid Units in which the center browns are only partially expanded while the yellows, reds and browns are usually "2 to fully expanded. In most cases the Iridophore Splotches on the dorsal mantle are simultane- ously prominent. Occasionally the Modified Discoid Units between the eyes appear dark also. DORSAL STRIPE has been observed on females of 74-136 mm ML and males of 128- 226 mm ML. ARM SPOTS There are two expressions of ARM SPOTS. The first is a compact, well-defined dark 106 HANLON % sb te ze. FIG. 21. Dorsal base of the third right arm from a male (147 mm ML) showing the approximate area of Modified Discoid Units (within dashed line) that produces the chromatic component ARM SPOTS. The units are shown only partially expanded. The distribution of the Modified Discoid Units continues onto the ventral portion of the arm and covers the same approximate area as on the dorsal side. The skin is transparent, and when all these units are expanded simultaneously, the spot appears darker and more distinct than if only one surface had Modified Discoid Units. See Fig. 23 for the appearance of this distinct spot. Arrow indicates anterior direction. Line represents 1 mm. brown spot that occurs near the bases of the second and third arms (Figs. 1B, 21, 23, 27). These distinct spots are produced by a cluster of Modified Discoid Units in which all chroma- tophores are maximally expressed. The browns in this unit are morphologically ar- ranged in such a manner that they overlap one another when fully expressed to provide a very dark brown base for the spot. The yel- lows and reds contribute to the darkening as well. On the second pair of arms the Modified Discoid Units occur only on the dorsal surface of the arm, but on the third arms (Fig. 21), which are spread laterally as swimming keels, the units occur both on the dorsal and ventral surfaces of the outside edge of the arm, mak- ing the spot even more conspicuous. The second expression of ARM SPOTS is a diffuse area of expanded chromatophores that is centered at the base of each third arm, and there is considerable variation in its size and shape. It may appear as a diffuse red- dish-brown splotch covering up to Уз to V2 of each third arm (Fig. 13) or a large reddish- brown darkening of most or all of the second and third arms (Fig. 11). Any of the aforemen- tioned expressions may occur independently or in combination and they are all nearly al- ways exhibited bilaterally. The ARM SPOTS component has been seen on males of 41- 285 mm ML. Only twice were females observed showing similar markings. An egg-laying female (89 mm ML) once briefly showed spots at the bases of the second arms, and three females (100-114 mm ML) once showed uniformly dark arms for 30 secs, with the remainder of the body clear. STITCHWORK FINS This component appears as a series of dark dashes, or stitches, around the periphery of the fins (Figs. 11, 22, 23, 27). The stitchwork appearance results from the full expression of the Yellow-Brown Discoid Units that are grouped 2-3 units wide into a continuous line, 2-5 mm inside the margin of the fin. The fins are generally clear but in some cases the component appears when the adjacent fin areas are ALL DARK. Since the Standard Dis- coid Units on the outer Y of the fin periphery are somewhat dispersed relative to those on the mantle, STITCHWORK FINS is still con- FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS FIG. 22. Fin of a large male (220 mm ML) showing the distribution of Yellow-Brown Discoid Units that produces STITCHWORK FINS. Line represents 10 mm. FIG. 23. Lateral view of a male squid (174 mm ML) showing the chromatic components STITCHWORK FINS, LATERAL FLAME, MID-VENTRAL RIDGE, ARM SPOTS, TENTACULAR STRIPE AND SPOTS and DORSAL ARM IRIDOPHORES towards a squid on its right. The tentacles are not often extended in this manner. spicuous when ALL DARK is shown. On some squids the stitchwork arrangement is not so conspicuous and the component appears as a somewhat continuous line around the fin periphery; it is nearly always bilateral al- though in rare cases it appears unilaterally. Only males of 105-285 mm ML have been seen to exhibit it. MID-VENTRAL RIDGE Viewed laterally, this component appears as a distinct, dark red solid line extending the entire length of the midline of the ventral man- tle (Figs. 1H, 23). It is a protrusible flap of skin that is muscularly extended downward 1- 3 mm while the chromatophores are fully ex- 108 HANLON panded. The chromatophores along the mid- line are arranged into Standard Discoid Units (Fig. 24); beyond this line on each side are the Lateral Flame Units. When the skin is pro- truded downward the chromatophore units on either side are compressed “back to back, which results in a double layer of units that result in a solid dark appearance. It has been seen in males of 41-285 mm ML. Females of 50-90 mm ML have on rare oc- casions shown a component similar to MID- VENTRAL RIDGE. It appeared as a ventral stripe of expanded chromatophores (similar to DORSAL STRIPE) in the same position as MID- VENTRAL RIDGE, but without the ridge. LATERAL FLAME This component appears as longitudinal, flame-like streaks of expanded Lateral Flame . . . a ‹ < ат 2 ° . a . y er в er Er a . = + = В HT . . a . 5 LE . . amos % ag e г 4 > a ARE L PT . 5 2 : . i à . Di E 0 as 5 SAGE o о . PARLE . e 5 a di” e ei Te e LES A o . р . et z u к , n x ! ' À — \ . | Units (Figs. 1H, 8, 23, 27). It covers the entire lateral aspect of the mantle and is usually de- ployed unilaterally (Figs. 11, 23). It has been seen only in males from 70-285 mm ML. LATERAL BLUSH This component has two forms. Most com- monly it appears as a diffuse, dark, lateral stripe 4-5 mm below the fin insertion, extend- ing longitudinally from the arm tips to the man- tle tip, with the remainder of the squid clear (Fig. 25). The component arises from the ex- pansion of Standard Discoid Units, and the width of the stripe is usually 2-3 units wide. It is generally shown unilaterally although occa- sionally it is seen bilaterally. It has been ob- served on small males of 41-86 mm ML and females of 82-107 mm ML. As young male squids grow older, this component persists FIG. 24. Portions of the ventral midline of males showing the Standard Discoid Units that constitute the chromatic component MID-VENTRAL RIDGE. Left: Retracted units (within area between hash marks) from a live male, 190 mm ML. Right: Expanded units from a live male, 147 mm ML. Note the transition into Lateral Flame Units on the left and right of each photograph, and the waves of muscular contractions (arrow) produced by the same muscles that protrude this skin downward when the component is expressed. Complete expression of this component may be seen in Figs. 1H and 23. Lines represent 1 mm. FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 109 А FIG. 25. One form of LATERAL BLUSH in a small male (86 тт ML). The component ARM SPOTS is very weakly developed. FIG. 26. Another form of LATERAL BLUSH that is seen only in sexually mature females. Drawing by D. A. McConathy. and becomes the uppermost line of expanded chromatophores in LATERAL FLAME. In another form it appears as a diffuse splotch of expanded chromatophores on the anterior, lateral portion of the mantle (Fig. 26). This has been observed only on sexually mature females of 83-114 mm ML. TENTACULAR STRIPE AND SPOTS The stripe consists of a single, thin brown line of expanded chromatophores along the lateral margin of the entire length of each tentacle (Figs. 1H, 18, 23, 27). At the distal end of each tentacular club this line ends in a darkened tip. The chromatophores along the stripe and tip are mostly arranged into Yellow- Brown Discoid Units, except that isolated reds are found intermittently along the tentacular length. There are approximatlely seven or eight distinct brown spots placed along the middle third of the inside of each tentacular Stalk, posterior to the tentacular club (Figs. 1B, 2, 23, 27). These spots are composed exclusively of Yellow-Brown Discoid Units, the smaller isolated spots comprising approxi- mately 10-15 brown chromatophores and 6- 10 yellows. This component is most readily seen when the tentacles are extended; how- ever, it can also be seen amidst or through the other arms when the tentacles are held at arms length. It has been seen on males of 90-285 mm ML. In one rare instance it was seen on a female of 50 mm ML. DORSAL MANTLE SPLOTCHES This component is characterized by the fre- quent visual appearance of all units of Irido- phore Splotches (Figs. 11, 1J, 10, 11, 13, 14, 15, 16, 19, 20, 27). The splotches are evident in both sexes in sizes as small as 21 mm ML. DORSAL ARM IRIDOPHORES The component is produced by the bright, distinct lridophore Sheets that extend the length of the first pair of arms of males (Figs. ТЫ, 1 12, 13, 18; 23, 27). The chromato- phores о the Standard Discoid Units on these arms are completely or mostly retracted and the Iridophore Sheets reflect pink, bright white or whitish-yellow. It has been seen on males of 74-285 mm ML. Although this bilateral component involves no chromatophores, it is not present at all times and it appears that the squids have some control over its appearance (see Discussion). DORSAL MANTLE COLLAR The Iridophore Sheets that are clustered around the anteriormost margin of the mantle periodically appear as a pink or green collar (Figs. 12, 13, 27). It appears in young squids of both sexes from 21 mm ML. DISCUSSION Previous workers have rightly suggested that the body patterns of Loligo are of simpler construction than those of Octopus and Sepia 110 HANLON (Boycott, 1953, 1965; Holmes, 1955; Wells, 1962; Messenger, 1974). Nevertheless, Loligo plei exhibits a diverse range of chro- matic components that can be used in intra- specific signalling and camouflage. What is significant is that this diversity can be largely accounted for by the morphological arrange- ment of the elements. When the body patterns of Loligo plei are analyzed and described in terms of their con- stituent parts (i.e. components, units, ele- ments), it becomes manifest that, like Octo- pus, it is at the unit level that patterning is first developed, both in the morphological and physiological senses. In the following para- graphs | emphasize that the units are func- tionally organized to produce differently shaped and colored components that are unique not only to the species, but in many cases to each sex within the species. In L. plei, the expression of chromatic components is a result of (1) the particular static morpho- logical array of elements within differently constructed units, and (2) the selective ner- vous excitation of those morphological units. Functional Morphology of the Units of Patterning The 16 chromatic components seen thus far in Loligo plei, when combined with pos- tural components and behavioral movements, allow for a wide range of body patterns, each of which coincides with a particular aspect of behavior (Hanlon, 1981; in prep.). While this discussion is limited to the functional mor- phology of the constituent parts of body pat- terns, the following brief and very simplified explanations of the interrelationships between groups of various chromatic components and associated behavior are given to provide some perspective concerning the circum- stances in which they are expressed. Eight components are seen on calmly swimming squids and presumably aid in camouflage: CLEAR, RING, SHADED TESTIS, SHADED EYE, DORSAL STRIPE, DORSAL MANTLE SPLOTCH- ES and DORSAL MANTLE COLLAR. Two com- ponents are exhibited by squids that are closely approached by a predator or large ob- ject: ALL DARK and RING. Eight components are seen almost exclusively during intraspe- cific aggression among males: ACCENTU- ATED TESTIS, ARM SPOTS, STITCHWORK FINS, MID-VENTRAL RIDGE, LATERAL FLAME, LATERAL BLUSH, TENTACULAR STRIPE AND SPOTS and DORSAL ARM IRIDOPHORES. The Standard Discoid Unit is most widely dispersed over the body surface in both sexes and is responsible for producing the many variations of the components ALL DARK, RING, ACCENTUATED TESTIS, DORSAL STRIPE and LATERAL BLUSH. The Modified Discoid Units have a greater number of browns that provide an important function: a greater capability to darken an area of the skin, either to shade underlying reflection for camouflage (SHADED EYE, SHADED TESTIS), or to produce a dis- tinctly dark spot for visual communication (ARM SPOTS). Yellow-Brown Discoid Units are distributed in such a manner that they produce distinctiy dark brownish dashes (STITCHWORK FINS), Stripes or spots (TENTA- CULAR STRIPE AND SPOTS). For example, the spots of TENTACULAR STRIPES AND SPOTS are a result of the distribution of units into circular spots that are bordered by areas devoid of chromatophores or iridophores; they are not a result of selective nervous excita- tion. In summary, discoid chromatophore units are best suited for (1) uniformly darken- ing large skin areas, (2) forming loosely de- fined, large dark areas such as transverse rings and longitudinal stripes, (3) shading or highlighting organs, and (4) creating dark spots and splotches. In contrast, the Lateral Flame Unit is not circularly arranged but is functionally organ- ized into longitudinally oriented dark streaks highlighted by adjacent light areas. In addition to the different spatial distribution of elements within the unit, the depth distribution changes as well, with brown chromatophores now lying above reds (Fig. 1D). The small size of the yellow and brown chromatophores, and their concentrated distribution into well-defined rows, both enhance the distinct appearance of LATERAL FLAME. This arrangement of ele- ments is not suited for uniform darkening as are the circularly arranged units. This is illus- trated when males are in the ALL DARK chro- matic component, with all chromatophores maximally expanded, and the lateral flame streaks are still conspicuous (Figs. 12, 27). The organization of iridescent cells is nota- bly different from that of chromatophores. Some difficulty arises when attempting to organize them into morphological units, since they are deeper in the dermis and are not visible at all times; in contrast, chromato- phores are visible both in the retracted and expanded states. The only way to verify the distribution and morphology of all iridescent cells would be to make a histological survey of FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 111 the entire dermis of the squid. Consequently, they have been described from their visual appearance. The Reflector Cells are struc- tured to reflect all light constantly, whereas the iridophores probably are not. The Irido- phore Splotches on the mantle are not seen at all times, but mostly in combination with com- ponents that are used for camouflage, such aS CLEAR, SHADED TESTIS and DORSAL STRIPE. The slightly irregular spacing of the splotches and their reflection of pink, green and yellow light promote concealment when viewed from above against a variegated sub- strate (Cott, 1940). Many of the splotches are positioned beneath the central brown chroma- tophore of the largest Standard Discoid Units (Figs. 1J, 10, 11) and the regularity with which this occurs is reminiscent of the positioning of some black chromatophores over gaps be- tween underlying leucophores in the center of chromatic units in Octopus vulgaris (Froesch & Messenger, 1978). It is unclear whether the placement of these elements is related in morphogenesis, and whether there is coordi- nation between the chromatophores and iridophores of each morphological unit during patterning. The iridophore splotch has not been included in the description of the Stand- ard Discoid Unit for these reasons, and also because all units do not have an underlying irrdophore splotch. The Iridophore Sheets have a visual quality similar to that of the Iridophore Splotches, and histological investigation may determine that they are morphologically identical; the colors that they reflect are similar. These cells are organized to produce a continuous sheet of iridescence that is used for camouflage, ob- literative counter-shading and, among males, intraspecific signalling. The iridescence in the first pair of arms (DORSAL ARM IRIDO- PHORES) provides a particularly conspicuous signalling device that is far brighter than any chromatophore expression. This iridescence is evident even when the overlying chromato- phores are expanded, but to promote its con- spicuousness, the first pair of arms usually has the chromatophores retracted (Figs. 11, 12, 13, 23, 27), and the arms are often arched upward during intense agonistic displays (Fig. 1H). An important question arises at this point: is the appearance of iridophores controlled by the squid? Based upon my observation the answer is probably affirmative. Video tapes from a stereomicroscope clearly show irido- phores appearing then disappearing under identical lighting and viewing conditions. Series of photographs taken from a single port of the holding tank show the same squid expressing different iridophore units under unchanging light conditions. Additional evi- dence is that males show DORSAL ARM IRIDO- PHORES particularly during intraspecific ag- gressive bouts with other males, indicating that this component is controlled. Further- more, when neurotransmitter substances such as dopamine, acetylcholine and 5- hydroxytryptamine are applied to the skin of live, narcotized or freshly dead squids, both chromatophores and iridophores are often expressed. This phenomenon has also been observed in Loligo vulgaris in Naples (A. Packard, personal communication, 1979). This opens up the interesting possibility of whether the iridophores are under muscular control. Further investigation will be reward- ing, because at present the iridophores are thought to be static cells whose appearance is regulated only by the structure of the cell and the quality and angle of light falling upon them. Selective Nervous Excitation The neural organization of chromatophores is poorly understood because of the difficul- ties in tracing nervous interconnections. A fur- ther complication is evidence that chromato- phores may be also linked by their muscles, which means that the innervation of one chromatophore may have a muscular effect on neighboring chromatophores (Froesch- Gaetzi & Froesch, 1977). In Loligo plei, the yellow, red and brown chromatophores of a single morphological unit are innervated in different combinations. Proof of this may be seen when examining video tapes of close-up recordings of individual chromatophore units of live squids. Within a single Standard Dis- coid Unit, groups of reds are separately in- nervated from other groups of reds, as well as from the brown and yellow chromatophores. The same is true of the yellows. The single brown chromatophore of each unit fires syn- chronously with nearby browns of a similar size (Fig. 4). The appearance of a squid can change within a matter of seconds, as shown in Fig. 27. In these photographs the same squid pro- gresses from a relatively light, translucent body pattern to a dark body pattern by selec- tive nervous excitation. It is critical to note that it is not only the number of units expressed 112 CHERE. sb HANLON x eal er FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 18 that results in darkening, but the degree to which they are expressed. For example, in Fig. 27 (A-D), the dorsal mantle becomes progressively darker not because more Standard Discoid Units are expressed, but because the chromatophore elements within each unit are expanded to greater degrees. Within a single Standard Discoid Unit, a typi- cal representation of the progression from light to dark would be as follows: (1) all chro- matophores are retracted, (2) reds are ex- panded to "2 full size, (3) browns are then concurrently expanded to Ya full size, (4) yel- lows are expanded to Ya full size, (5) reds are expanded from Ya to full size, and (6) browns and yellows are also expanded to full size (i.e. all chromatophores are now maximally ex- panded). In other words, not only are more chromatophores of more colors being ex- panded, but each chromatophore is ex- panded to a progressively larger size until all are maximally expanded. Figs. 1E, 1F and 1G indicate how complex the neural connections in LATERAL FLAME may be. In Fig. 1E, the reds between rows are expanded, as are a few within rows. A few yellows are fully expanded, and some browns and yellows are partially expanded; in general the rows are weakly developed. In Fig. 1F, several more reds within the rows have ex- panded, and groups of yellows are expanded, especially in the bottom row, and the rows are clearly distinguishable. Note, however, that certain groups of yellows, browns and reds within the rows still remain unexpanded. From a functional standpoint, the complex morphological and physiological organization of Lateral Flame Units provides the capability of expressing LATERAL FLAME in varying de- grees of intensity. This is important because during intraspecific aggression, males use the unilateral expression of this component (Figs. 1H, 11) along with other components during agonistic displays with other males, and the degree of intensity of the LATERAL FLAME component is directly related to the degree of aggression during the display (Hanlon, in prep.). Another function of the complex organ- ization is to allow certain Lateral Flame Units to be used in the expression of other chro- matic components. For example, when large males are in RING, selected Lateral Flame Units are expressed in combination with se- lected Standard or Modified Discoid Units to produce a transverse band of expanded chromatophores around the mantle (Fig. AE): Significance of Color Cephalopods are almost certainly color blind (Messenger et al., 1973; Messenger, 1977, 1979) and therefore it is highly unlikely that color is of significance in intraspecific com- munication. Pattern and contrast are the in- formation transmitters among conspecifics, as postulated by others (Packard, 1972; Messen- ger et al., 1973; Messenger, 1974) and rein- forced by my observations of Loligo plei. The components used in intraspecific encounters (e.g. ACCENTUATED TESTIS, ARM SPOTS, etc.), especially LATERAL FLAME, are most conspicuous by their configuration, and are most effective as they become brighter. Color would be most effective in camou- flage patterns used to avoid predators with color vision, such as fishes. An important consideration in this regard is the water depth at which these colors are viewed. Only at the shallowest depths would the long-wavelength colors (especially red) retain their distinct hue because the longer wavelengths of light are more attenuated in sea water than the shorter wavelengths. Loligo plei is found occasionally on and around shallow coral reefs throughout the Caribbean and in these circumstances the warm-toned chromatophores, in conjunction with the iridophores that reflect all wave- lengths of light, are probably effective in achieving color as well as tone matching when the squids are sitting on or swimming near substrates that are colored brown, red, green or yellow. During mid-water swimming or schooling in the day, the iridophores may help achieve obliterative counter-shading (Cott, 1940). L. plei actively forages and feeds in the water column at night and the chromatic components are probably not effective in very low light levels. During the day, when the squids school near the bottom, the various A A Roo _Ée DO p_ o nd + FIG. 27. Example of selective nervous excitation changing the appearance of the same male squid (201 mm ML) within a few seconds. From A to E note the transient appearance of the conspicuous white testis and the chromatic components LATERAL FLAME, ARM SPOTS, TENTACULAR STRIPE AND SPOTS, STITCHWORK FINS, RING, DORSAL ARM IRIDOPHORES, DORSAL MANTLE SPLOTCHES and DORSAL MANTLE COLLAR. See Discus- sion for further explanation. 114 HANLON body patterns may be used for predator es- cape, camouflage and intraspecific communi- cation. The vertical distribution of differently col- ored chromatophores in Loligo is exactly op- posite to that of Octopus vulgaris (Froesch & Packard, 1979). In Loligo, it is yellow over red over brown (in Standard Discoid Units), whereas in Octopus it is brown over red over orange over yellow. The significance of this is unknown. From a functional standpoint, the chromatophores of Loligo act somewhat like a neutral density filter in a manner similar to Octopus, the principal differences being the slightly narrower color range and the vastly greater size of chromatophores in Loligo. With fewer and less dense chromatophores, Loligo is unable to produce chromatic com- ponents as complex and varied as Octopus vulgaris. However, the fine pattern reticulation and skin sculpture required of a benthic octo- pus are not needed by a schooling squid that spends most of its time in the water column. In contrast to Octopus, whose finely differenti- ated skin is constructed to produce subtle, re- fined patterns mainly for camouflage and pre- dator escape, Loligo uses most of its body patterns for intraspecific communication in which bold, coarse patterns are sufficient to visually communicate with a nearby squid. Hence, large chromatophores may be re- garded as an efficacious means of producing a chromatic component with the simplest skin differentiation. Sexual Dimorphism Three aspects of sexual dimorphism ге- lated to chromatic behavior may be demon- strated in Loligo plei: (1) the distribution of chromatic units in the dermis, (2) the range of chromatic components, and (3) the general behavior of each sex. Table 2 shows that most of the chromatic units and components of L. plei are shared by both sexes, but that males have acquired one unit and seven com- ponents that are unique to that sex alone. In contrast, females do not have any unique units or components. Among the shared units, males have a greater number and distribution of certain types. For example, both sexes have Modified Discoid Units over the eyes (SHADED EYE), but only males have them over the testis (SHADED TESTIS) and on the second and third arms (ARM SPOTS). In addi- tion to having Yellow-Brown Discoid Units on the tentacles, males have them on the pe- riphery of the fins (STITCHWORK FINS). Males have prominent Iridophore Sheets on the first pair of arms (DORSAL ARM IRIDOPHORES). The overall behavior of each sex is marked- ly different. Females are relatively passive TABLE 2. Breakdown of the chromatic units and components of Loligo plei that are found (1) only in males, (2) only in females, or (3) shared by both sexes. CHROMATIC UNITS Male only Lateral Flame Unit Female only Shared Standard Discoid Unit Modified Discoid Unit Yellow-Brown Discoid Unit Reflector Cells Iridophore Splotches Iridophore Sheets CHROMATIC COMPONENTS Male only ACCENTUATED TESTIS SHADED TESTIS ARM SPOTS STITCHWORK FINS MID-VENTRAL RIDGE LATERAL FLAME DORSAL ARM IRIDOPHORES Female only Shared CLEAR ALL DARK RING SHADED EYE DORSAL STRIPE LATERAL BLUSH TENTACULAR STRIPE AND SPOTS DORSAL MANTLE SPLOTCHES DORSAL MANTLE COLLAR E Al FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 1415 and docile laboratory animals, whereas males are strongly aggressive and spend a great deal of time fighting among themselves and courting females (Hanlon, in prep.). Of the male-only components, all but SHADED TESTIS are used during intraspecific aggres- sive encounters. Another dimorphic character is the mean and maximal size of adults; males are substantially larger. With these morpho- logical and behavioral differences it is quite easy to distinguish the sexes of living squids, even in certain cases where very young squids as small as 35 mm ML attain preco- cious sexual maturation in the laboratory. Comparisons Within the Genus Loligo Although this study represents the first de- tailed description of the organization of chro- matic elements in Loligo, several comparisons may be made with other members of the genus. Table 3 is a tabulation of incidental reports of various chromatic components that have been mentioned in the literature. The fact that all five of the species when observed alive showed CLEAR, ALL DARK and DORSAL MANTLE SPLOTCHES indicates that these are common components. MID-VENTRAL RIDGE, which is often clearly evident in preserved specimens, has been seen in four species be- sides Loligo plei, and evidence of LATERAL FLAME has been seen in two other species. The chromatic components ACCENTUATED MESUS SHADED TESTIS, SHADED EYE, TENTACULAR STRIPE AND SPOTS, DORSAL STRIPE, LATERAL BLUSH and DORSAL ARM IRIDOPHORES have not yet been reported in other species, but this undoubtedly is partly a result of the general difficulties encountered in keeping squids alive in captivity for long pe- riods or in making long-term observations in the field. It is likely that many of these com- ponents, as well as new ones, will be seen in other species when detailed observations are made. It is presently impossible to determine whether L. plei can show more than 16 com- ponents, or if it possesses a wider repertoire of chromatic components than other members of the genus, although it seems that L. plei has more chromatic expression than L. pealei or L. opalescens. More information is urgently needed for other species, for this will be in- valuable in understanding the role and ontogeny of behavior, reproductive strategies and phylogenetic relationships among cephalopods. ACKNOWLEDGEMENTS | am indebted to my many co-workers at The Marine Biomedical Institute who contributed greatly by helping capture, maintain and ob- serve live squids. Raymond F. Hixon, John W. Forsythe, Deirdre A. McConathy and Joseph P. Hendrix, Jr. all provided capable assistance that made the large scope of this work possi- ble. Special thanks are due to William H. Hulet and Raymond F. Hixon who continually pro- vided suggestions and reviewed several drafts of this manuscript. | am very grateful to Andrew Packard for the inspiration he provided through his work on body patterning and for helpful discussions concerning my work. | thank Professor J. Z. Young for first suggesting that | seriously con- sider the concept of units in the skin of Loligo. J. B. Messenger and Andrew Packard kindly reviewed the final draft and | thank them for their excellent critiques and improvements. Funding for this work was provided by Grants RR 01024-04 and 2P40 RR 01024-03 from the Division of Research Resources, Na- tional Institutes of Health, and from Marine Medical General Budget account 7-11500- 765111 of the Marine Biomedical Institute, University of Texas Medical Branch. A portion of this work was submitted in par- tial fulfillment of the requirements for the de- gree of Doctor of Philosophy from the Uni- versity of Miami, Rosenstiel School of Marine and Atmospheric Science (R.S.M.A.S.), and it constitutes a scientific contribution from R.S.M.A.S. and The Marine Biomedical Insti- tute, University of Texas Medical Branch. Lastly, | appreciate the efforts of Sharon K. Burton, who typed and assembled all drafts of the manuscript. LITERATURE CITED ARNOLD, J. M., 1962, Mating behavior and social structure in Loligo pealei. Biological Bulletin, 123: 53-57. ARNOLD, J. M., 1967, Organellogenesis of the cephalopod iridophore: cytomembranes in de- velopment. Journal of Ultrastructure Research, 20: 410421 BLAINVILLE, M. H. D. DE, 1823, Memoire sur les espèces du genre Calmar (Loligo, Lamarck). Journal de Physique, de Chimie et d'Histoire Naturelle, 96: 116-133. BOYCOTT, B. B., 1953, The chromatophore sys- tem of cephalopods. Proceedings of the Linnean Society of London, 164: 235-240. en] ‘aseq uly UO sadıys ‘sjods ul4 И peep peap OAI Г SA EZ61 ‘эем элн элн 9681 ‘ener 5иеб/пл 061/07 OAI 96161 ‘моли N peep 2961 '90ye7] эл! ¿SUIB -jew ul Buoje sjods weg W эл! 2961 ‘pjowy OAI] OAI] $7561 ‘UOSUSA9]S элн ‚эвиеш JOUajue ‘peau “sue Yıeg эл! элн OAI] OAI OAI 6061 ‘зшеним N эл! 1881 “USA ısjead 061/07 эл! элн 2161 'SydIUBpoy » UEUIUAON OAI] элн реэр реэр элн OAI] OAI] BA 0/61 80H] peep И peep OAI SA ON 1961 20He7 эл! эл! эл! 6961 ‘n0940g N peep 0561 ‘Seay 2 18d 061/07 O eee г sjuauodulo) JaylO YVTIOD SSHOLOIdS 3MNVWIW 390l8 SIOAS JdIHIS $4531 SILS3L ONIN УНУа HV319 129 À = 311NVN 3HOHdOOIHI TVH3LVI TIWHLNSA WHY WSHOG G30VHS 03LV Jouny/Ssalsads 1VSHOQ 1VSHOQ -OIW -NIN399V Е НЙ ЗЕ А И № ©, sjuaguoduo) эцешоц9Э AA PPP PPP rs ‘aluaAnl = Г ‘э|еше; = y ‘ele = yy 06/07 snueB ay, jo spinbs panıasaud 10 an ш usas sjuauoduoo эцешолцо jo soda ¡ejuapiou] “€ 37191 116 11174 FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS W SA ‘Peau pue sue weg 9AI| :реэц pue sue 12 эли :e[ppes ul ‘sjods uly эл! OUI] |едуэл “SADA OJIUMA эл! ‘aha mojaq 104$ ‘ajppes ul} pau ‘sjods uly IW peap IN peap peap IN реэр peap 3'N peap Эли 4'W Эли эл! Эли эл! эл! anı эл! эл! эл! Эли эл! эл! эл! эл! эл! 6261 “Meses 1eye8/q 061/07 6261 'Neses sınpa 06,07 7161 ‘Эщон 2561! ‘SSOA 0961 ‘5ээн 0761 'SOWIOH E261 JOEN 9681 ‘ецег Isaquoy 061/07 2161 ‘AelinH 95/61 ‘моли 25/61 MOIW 6961 ‘$р!9!- 7961 ‘UEMODOIN susosajedo 061/07 1161 ‘ЭМ 961 ‘ИЭМ 961: Зиэрлет 118 HANLON BOYCOTT, B. B., 1961, The functional organiza- tion of the brain of the cuttlefish Sepia officinalis. Proceedings of the Royal Society of London, Ser. B, 153: 503-534. BOYCOTT, B. B., 1965, A comparison of living Sepioteuthis sepioidea and Doryteuthis plei with other squids, and with Sepia officinalis. Journal of Zoology, 147: 344-351. BOYLE, P. R. & DUBAS, F., 1981, Components of body pattern displays in the octopus Eledone cirrhosa (Mollusca: Cephalopoda). Marine Be- haviour and Physiology, 8: 135-148. BROCCO, S. L. & CLONEY, R. A., 1980, Reflector cells in the skin of Octopus dofleini. Cell and Tissue Research, 205: 167-186. CLONEY, В. A. & FLOREY, E., 1968, Ultrastructure of cephalopod chromatophore organs. Zeitschrift fur Zellforschung, 89: 250-280. COHEN, A. C., 1976, The systematics and distribu- tion of Loligo (Cephalopoda, Myopsida) in the western North Atlantic, with descriptions of two new species. Malacologia, 15: 299-367. COTT, H. B., 1940, Adaptive Coloration in Animals. Oxford University Press, New York, 508 p. COWDRY, E. V., 1911, The colour changes of Octopus vulgaris. University of Toronto Studies, Biological Series, no. 10: 1-53. DENTON, E. J., 1970, On the organization of re- flecting surfaces in some marine animals. Philo- sophical Transactions of the Royal Society of London, Ser. B, 258: 285-313. DENTON, E. J. & LAND, M. F., 1971, Mechanism of reflection in silvery layers of fish and cephalo- pods. Proceedings of the Royal Society of Lon- don, Ser. A, 178: 43-61. FIELDS, W. G., 1965, The structure, development, food relations, reproduction, and life history of the squid Loligo opalescens Berry. California De- partment of Fish and Game, Fish Bulletin, 131: 1-108. FLOREY, E., 1966, Nervous control and spontane- ous activity of the chromatophores of a cephalo- pod, Loligo opalescens. Comparative Biochem- istry and Physiology, 18: 305-324. FLOREY, E., 1969, Ultrastructure and function of cephalopod chromatophores. American Zoolo- gist, 9: 429-442. FLOREY, E. & KRIEBEL, M. E., 1969, Electrical and mechanical responses of chromatophore muscle fibers of the squid, Loligo opalescens, to nerve stimulation and drugs. Zeitschrift für Vergleichende Physiologie, 65: 98-130. FROESCH, D. & MESSENGER, J. B., 1978, On leucophores and the chromatic unit of Octopus vulgaris. Journal of Zoology, 186: 163-173. FROESCH, D. & PACKARD, A., 1979, Octopus chromatophores accumulate nickel. Experientia, 35: 828-830. FROESCH-GAETZI, V. & FROESCH, D., 1977, Evidence that chromatophores of cephalopods are linked by their muscles. Experientia, 33: 1448-1450. HANLON, R. T., 1981, Courting and mating be- havior in the squid Loligo (Doryteuthis) plei. In: Advances in Invertebrate Reproduction, Elsevier, N.Y., p. 325. (Abstract). HANLON, R. T., in prep., Body patterning and as- sociated behavior in the tropical arrow squid Loligo plei (Myopsida, Cephalopoda). HANLON, R. T. & HIXON, R. F., 1980, Body pattern- ing and field observations of Octopus burryi Voss, 1950. Bulletin of Marine Science, 30: 749- 155: HANLON, R. T., HIXON, R. F. & HULET, W. H., 1978, Laboratory maintenance of wild-caught loliginid squids. In: BALCH, N., AMARATUNGA, Т. & O'DOR, В. K., eds., Proceedings of the Workshop on the squid Illex illecebrosus, Fish- eries and Marine Service Technical Report 833, p. 20.1-20.13. HILL, A. V. & SOLANDT, D. Y., 1934, Myograms from the chromatophores of Sepia. Journal of Physiology 83: 13-14. HOLME, N. A., 1974, The biology of Loligo forbesi Steenstrup (Mollusca: Cephalopoda) in the Plymouth area. Journal of the Marine Biological Association of the United Kingdom, 54: 481-503. HOLMES, W., 1940, The colour changes and colour patterns of Sepia officinalis L. Proceedings of the Zoological Society of London, Ser. A, 110: 17- 35: HOLMES, W., 1955, The colour changes of cepha- lopods. Endeavour, 14: 78-82. HURLEY, A. C., 1977, Mating behavior of the squid Loligo opalescens. Marine Behaviour and Physiology, 4: 195-203. JATTA, G., 1896, Fauna und flora des Golfes von Neapel. Monographie 23: | Cefalopodi (Sis- tematica). Friedlander, Berlin, 268 p. LAROE, E. T., 1967, A contribution to the biology of the Loliginidae (Cephalopoda: Myopsida) of the tropical western Atlantic. Master's Thesis, Uni- versity of Miami, Coral Gables, Florida, 220 p. LAROE, E. T., 1970, The rearing and maintenance of squid in confinement, with observations on their behavior in the laboratory. Ph.D. Disserta- tion, University of Miami, Coral Gables, Florida, 136 p. MCCONATHY, D. A., HANLON, R. T. & HIXON, R. F., 1980, Chromatophore arrangements of hatch- ling loliginid squids (Cephalopoda, Myopsida). Malacologia, 19: 279-288. MCGOWAN, J. A., 1954, Observations on the sex- ual behavior and spawning of the squid, Loligo opalescens, at La Jolla, California. California Fish and Game, 40: 47-54. MESSENGER, J. B., 1974, Reflecting elements in cephalopod skin and their importance for camou- flage. Journal of Zoology, 174: 387-395. MESSENGER, J. B., 1977, Evidence that Octopus is colour blind. Journal of Experimental Biology, 70: 49-55. MESSENGER, J. B., 1979, The eyes and skin of Octopus: compensating for sensory deficiencies. Endeavour, 3: 92-98. MESSENGER, J. B., WILSON, A. P. & HEDGE, A., FUNCTIONAL ORGANIZATION OF SQUID BODY PATTERNS 119 1973, Some evidence for colour-blindness in Octopus. Journal of Experimental Biology, 59: 77-94. MIROW, S., 1972a, Skin color in the squids Loligo реа! and Loligo opalescens. |. Chromatophores. Zeitschrift fur Zellforschung, 125: 143-175. MIROW, S., 1972b, Skin color in the squids Loligo реа and Loligo opalescens. Il. Iridophores. Zeitschrift fur Zellforschung, 125: 176-190. MOYNIHAN, M., 1975, Conservatism of displays and comparable stereotyped patterns among cephalopods. In: BAERENDS, G., BEER, C. & MANNING, A., eds., Function and Evolution in Behaviour, Essays in Honor of Professor Niko Tinbergen, F.R.S., Oxford University Press, New York, p. 276-291. MOYNIHAN, M. H. & RODANICHE, A. F., 1977, Communication, crypsis, and mimicry among cephalopods. In: SEBEOK, T. A., ed., How Ani- mals Communicate, Indiana University Press, Bloomington, p. 293-302. NAEF, A., 1912, Teuthologische Notizen. 4. Die Gattungen der Loliginidae. Zoologischer Anzeiger, 39: 741-745. NAEF, A., 1923, Fauna and flora of the Bay of Naples. Cephalopoda. Zoologische Station zu Neapel, Part |, /, Fas. Il: 293-917. NEILL, S. ST. J., 1971, Notes on squid and cuttle- fish; keeping, handling and colour-patterns. Pubblicazioni della Stazione Zoologica di Napoli, 39: 64-69. PACKARD, A., 1972, Cephalopods and fish: the limits of convergence. Biological Reviews, 47: 241-307. PACKARD, A., 1982, Morphogenesis of chromato- phore patterns in cephalopods: are morphologi- cal and physiological ‘units’ the same? Malaco- logia, 23: 193-201. PACKARD, A. & HOCHBERG, F. G., 1977, Skin patterning in Octopus and other genera. In: NIXON, M. & MESSENGER, J.B., eds., The Bi- ology of Cephalopods, Symposium of the Zoo- logical Society of London, 38: 191-231. PACKARD, A. & SANDERS, G., 1969, What the octopus shows to the world. Endeavour, 28: 92- 99. PACKARD, A. & SANDERS, G., 1971, Body pat- terns of Octopus vulgaris and maturation of the response to disturbance. Animal Behaviour, 19: 780-790. PANTONE MATCHING SYSTEM (PMS), 1966, Coordinated Matching System for Printing Inks, Universal Printing Ink Corporation, Pantone Press, New York, 144 p. RATHJEN, W. F., HANLON, R. T. 8 HIXON, R. F., 1979, Squid fishery resources and development in the northwest Atlantic and Gulf of Mexico. Proceedings of the Gulf and Caribbean Fisheries Institute, 31: 145-157. REES, W. J., 1950, Notes on Cephalopoda from the Caribbean. Proceedings of the Malacological Society of London, 28: 107-114. RIEDL, R., 1978, Order in Living Organisms, Wiley, New York, 313 p. SASAKI, M., 1929, A monograph of the dibranchiate cephalopods of the Japanese and adjacent waters. Journal of the Faculty of Agriculture, Hokkaido Imperial University, 20 (Suppl.): 1-357. SCHELLING, P. & FIORONI, P., 1971, Zur Entwick- lung der Chromatophoren bei Loligo vulgaris Lam. Verhandlungen der Naturforschenden Gesesellschaft Basel, 81: 55-68. STEVENSON, J. A., 1934, On the behavior of the long-finned squid (Loligo реа, Lesueur). Canadian Field-Naturalist, 48: 4-7. TARDENT, P., 1962, Keeping Loligo vulgaris L. in the Naples Aquarium. 7st International Congress on Aquarology, Bulletin de l'Institut Oceano- graphique, Monaco, Special no. 1A: 41-46. VERRILL, A. E., 1881, The cephalopods of the north-eastern coast of America, Part Il, The smaller cephalopods. Transactions of the Con- necticut Academy of Science, 5: 259-446, pl. 26-56. VOSS, G. L., 1952, A revision of the loliginid squids of the Western Atlantic. Master's Thesis, Univer- sity of Miami, Coral Gables, Florida, 83 p. VOSS, G., OPRESKO, L. & THOMAS, R., 1973, The potentially commercial species of octopus and squid of Florida, the Gulf of Mexico and Carib- bean area. University of Miami Sea Grant Pro- gram, Sea Grant Field Guide Series, no. 2: 1-33. WARREN, (|. R., SCHEIER, М. Е. & RILEY, D. A., 1974, Colour changes in Octopus rubescens dur- ing attacks on unconditioned and conditioned stimuli. Animal Behaviour, 22: 211-219. WEBER. W., 1968, Multiple Innervation der Chro- matophorenmuskelzellen von Loligo vulgaris. Zeitschrift für Zellforschung, 92: 367-376. WEBER, W., 1970, Zur Ultrastruktur der Chroma- tophorenmuskelzellen von Loligo vulgaris. Zeitschrift für Zellforschung, 108: 446-456. WEBER, W., 1973, Peculiarities of innervation in chromatophore muscle cells of Loligo vulgaris. Marine Biology, 19: 224-226. WELLS,M. J., 1962, Brain and Behaviour in Cepha- lopods. Heinemann, Toronto, 171 р. WHITAKER, J. D., 1978, A contribution to the bi- ology of Loligo pealei and Loligo plei (Cephal- opoda, Myopsida) off the southeastern coast of the United States. Master's Thesis, College of Charleston, Charleston, South Carolina, 186 p. WILLIAMS, L. W., 1909, The Anatomy of the Com- mon Squid Loligo pealii. American Museum of Natural History, Brill, Leiden, 92 p. YOUNG, y. Z., 1974, The central nervous system of Loligo. |. The optic lobe. Philosophical Transac- tions of the Royal Society of London, 267: 263- 302. YOUNG, J. Z., 1976, The nervous system of Loligo. II. Suboesphageal centres. Philosophical Trans- actions of the Royal Society of London, 274: 101- 167. YOUNG, J. Z., 1977, The nervous system of Loligo. Ш. Higher motor centres: the basal supra- oesophageal lobes. Philosophical Transactions of the Royal Society of London, 276: 351-398. MALACOLOGIA, 1982, 23(1): 121-134 F. G. Hochberg, Jr. Department of Invertebrate Zoology, Santa Barbara Museum of Natural History, 2559 Puesta del Sol Road, Santa Barbara, CA 93105, U.S.A. ABSTRACT The fluid-filled renal and pancreatic coela (“kidneys”) of cephalopods are an ideal environ- ment for the establishment and maintenance of parasites. Periodic, but incomplete, elimination of urine from the renal sacs provides an abundant source of nutrients rich in nitrogen and carbohydrate compounds. The epithelium of the convoluted renal appendages provides an excellent surface for attachment and the renal pores provide a simple exit to the exterior. As such the “kidneys” have been exploited by a number of phylogenetically distinct symbiotic organisms. These include: a virus, a fungus, three distinct groups of ciliates, dicyemid meso- zoans, a digenetic trematode, larval cestodes, and larval nematodes. Of these forms, the dicyemids and the apostome ciliate, Chromidina, are known to occur only in cephalopods and only in the “kidneys.” Benthic cephalopods are infected exclusively with dicyemids whereas pelagic cephalopods harbor only chromidinids. An overlap of hosts occasionally occurs when ciliates are acquired by planktonic stages of otherwise bottom-dwelling octopods and sepioids. The remarkable similarities in morphology and diphasic life cycle exhibited by the dicyemids and chromidinids indicate a high degree of convergence in response to selective pressures associ- THE “KIDNEYS” OF CEPHALOPODS: A UNIQUE HABITAT FOR PARASITES ated with life in the renal habitat. Key words: cephalopods; ‘kidneys’; Chromidina; convergence. INTRODUCTION The kidneys of any organism are an unusu- al place to find parasites, and in fact in the entire animal kingdom this organ has rarely been exploited. It is then with some interest that we view the rather remarkable situation in cephalopods where a majority of species and individuals of squids and octopuses harbor a diversity of parasites, several of which are found exclusively in the excretory organs. This paper reviews the morphology of the cephalopod excretory system and relates it to the parasites which occur there. The discus- sion primarily considers similarities between the chromidinid ciliates and the dicyemid mesozoans, but other renal parasites are briefly treated. MATERIAL AND METHODS Over 2,300 cephalopods representing 91 species in 56 genera and 22 families have been examined for parasites. In all cases the excretory organs were examined and any parasites encountered were removed. Para- sites were smeared and fixed on coverslips or prepared as whole mounts for later identifica- tion (see Hochberg, 1971 for details). Species descriptions have been published or are be- ing prepared. excretory organs; parasites; dicyemids; ciliates; A review of the literature revealed another 56 species of cephalopods in 29 genera for which parasite information has been addition- ally recorded. However, in only 6 of these genera has the presence or absence of “kid- ney parasites been specifically indicated. All genera considered are listed in Table 1. TABLE 1. Genera of cephalopod hosts and their “kidney” parasites. Parasites Hosts Chromidinid Dicyemid Nautiloidea Nautilus = = Coleoidea Sepioidea Spirula + - Euprymna = = Heteroteuthis + = Rondeletiola = + Rossia = + Sepietta = + Sepiola dE + Sepia + + Teuthoidea Alloteuthis = Loligo ++ Loliolopsis = = (121) 122 TABLE 1 (Continued) Hosts Lolliguncula Sepioteuthis Abralia Abraliopsis Enoploteuthis Pterygioteuthis Pyroteuthis Thelidioteuthis Octopoteuthis Moroteuthis Onykia Berryteuthis Gonatopsis Gonatus Bathyteuthis Histioteuthis Ctenopteryx Dosidicus Illex Ommastrephes Symplectoteuthis Todarodes Chiroteuthis Mastigoteuthis Bathothauma Cranchia Galiteuthis Helicocranchia Leachia Liocranchia Megalocranchia Phasmatopsis Sandalops Octopoda Grimpoteuthis Opisthoteuthis Japetella Bathypolypus Bentheledone Benthoctopus Eledone Graneledone Octopus Pareledone Pteroctopus Robsonella Scaeurgus Thaumeledone Ocythoe Argonauta Vampyromorpha Vampyroteuthis Chromidinid HOCHBERG Dicyemid Parasites + + + — + — + = + — + — р Е. + — + — = + - + ES — = + + = + + + — + + Sr - + — - ar + - + | CEPHALOPOD “KIDNEY” PARASITES: RESUME Historically, the first known reference to a cephalopod parasite is found in a brief note by Cavolini (1787). In the kidneys of an octopus (probably O. vulgaris) he discovered “little in- fusorial organisms, shaped-like eels, having a muzzle with a trembling beard, darting, divid- ing themselves into many portions.” This de- scription could easily apply to either the chromidinid ciliates or the dicyemid meso- zoans. Some 50 years following Cavolini’s re- port, Krohn (1839) first documented the dicyemids and this well Known group has been extensively detailed by numerous investigators. In 1881, Foettinger described the ciliate genus now referred to as Chromidina. This unusual parasite is known from only a few studies. Both these groups of ciliated, vermiform parasites are known only from cephalopods and characteristically occur only in the “kidneys.” | will present additional details on these two groups farther on in this paper. Several other parasites have been de- scribed which are either rare in cephalopods or rare in their excretory organs. Dobell (1909) reported on a small ovoid in- fusorian, Opalinopsis, from the “kidneys” of Sepia Officinalis. This ciliate is known from a number of cephalopods but is normally re- stricted to the digestive gland (see review in Hochberg, 1971). Its presence in the “kid- neys” is presumed to be an error which may have resulted from accidentally cutting the di- gestive gland while removing the renal organs. Raabe (1934) noted a fungus in the “kid- neys” of several specimens of Sepia offici- nalis and Octopus vulgaris in the Mediter- ranean. The filamentous thalli of this fungus invade the renal appendages and the blood spaces causing considerable damage to the host tissue. Tentatively identified as, “Aspergillus,” the fungus is a highly destruc- tive pathogen. This parasite must be quite rare, since it has never been reported or men- tioned again, in spite of the numerous cephalopods examined in the Mediterranean and elsewhere. A number of larval and adult helminths have been reported in cephalopods, but only one is known to regularly inhabit the “kid- neys.” Allison (1966), and later Shon 4 Powell (1968), described the digenetic trema- tode, Plagioporus maorum, and reported its PARASITES IN CEPHALOPOD “KIDNEYS” 123 presence in Octopus maorum and Robson- ella australis in New Zealand. The worms typically are found crawling on the renal ap- pendages within the renal sacs, or are located beneath membranes of nearby organs. This is the only sexually mature (not progenetic) digenean known from a marine invertebrate. Larval stages of digeneans (didymozoans), cestodes, and nematodes have all, on occa- sion, been observed in the “kidneys” of squids. This phenomenon is rare, and should not be confused with the natural condition. It occurs only when massive numbers of larval hel- minths, which normally inhabit the digestive tract, invade a cephalopod host, and pene- trate all the organs of the viscera (Hochberg, 1971). What appear to be virus particles have been observed in the nuclei of the epithelial cells of the renal appendages of several Octopus species from New Zealand, Florida, and California (Short & Hochberg, unpub- lished). This viral infection is found not only in the nuclei of the epithelial cells of the host but in the nuclei of the somatic cells of the dicyemid parasites as well (Short & Hoch- berg, 1969). More recently, representatives of two addi- tional families of ciliates have been recovered from the excretory organs of pelagic squids off Hawaii and California. Though distinct from Chromidina, they have not been identi- fied or described and will not be treated here (Hochberg, unpublished). DICYEMID MESOZOA The dicyemids are a puzzling group without definite affinities in the animal kingdom. They exhibit an impressive array of truly unique characters which hold a special curiosity for zoologists. Along with the orthonectids they have long been considered a class within the phylum Mesozoa (see Hyman, 1940; Grasse, 1961). In light of their dissimilar internal mor- phologies and the lack of homologies in stages of their life cycles, it is best to treat these two assemblages as separate phyla and to use the term “Mesozoa” to refer to their grade of organization only. The dicyemids are the most common and characteristic parasites of the excretory Organs of cephalopod molluscs. These mi- nute, vermiform organisms attach principally to the renal appendages while the remainder of their worm-like bodies float in the fluid-filled renal coelom. Occasionally they are found in decapods in the pericardium attached to the branchial heart appendages and also in the reno-pancreatic coelom attached to the di- gestive duct appendages. They live and re- produce in these organs, doing no apparent harm to the host. A total of 50 species of cephalopods repre- senting 18 genera are currently known to host dicyemids (Table 1). These parasites occur in sepioids, especially cuttlefishes and sepiolids. In the octopods, both cirrate and incirrate groups are parasitized. In the teuthoids, only Sepioteuthis, an epibenthic loliginid, is in- fected. Each host species usually harbors a specific dicyemid species or a complex of Species. Dicyemids exclusively parasitize benthic or epibenthic cephalopods. In temperate and polar waters, adult, benthic cephalopods are 100% infected, whereas in the tropics and off oceanic islands no cephalopods have been reported to be infected. In subtropical waters the incidence of infection varies but is always less than 100%. In other words, the distribu- tion of dicyemids in benthic hosts is by no means universal. Initial infection normally occurs in very young animals, either immediately follow- ing hatching, in cephalopods with demersal juveniles, or following settlement to the bot- tom, in those host species with planktonic larval stages. In all the animals | have exam- ined | have never encountered dicyemids in neritic or oceanic cephalopods. However, McConnaughey (1959) recorded a species of Dicyemennea as occurring in Loligo opales- cens and a single dicyemid was reported in a single specimen of /Шех illecebrosus by Aldrich (1964). These reports are probably in error considering that thousands of /llex and Loligo have been examined by many other investigators and all have been uninfected. Both Loligo and Шех are neritic genera which come inshore as adults only to mate, spawn and die. In essence, they are pelagic, and would predictably be free from dicyemids. To date, 67 species of dicyemids have been described. On the basis of the unde- scribed species in my collections and con- sidering the number of potential host species still to be examined it is possible to project a total of about 200 species in the phylum. Seven genera are currently recognized and placed in two families —DicYEMIDAE: Dicyema, Dicyemennea, Dicyemodeca, Pleodicyema, and Pseudicyema; CONOCYEMIDAE: Cono- cyema and Microcyema. 124 HOCHBERG The number and orientation of cells in each tier of the calotte, the presence or absence of abortive axial cells and the presence or ab- sence of syncytial stages determines the genus. The size of the adult stages, the num- ber of cells comprising the body, the shape of the calotte, the anterior extension of the axial cell, the presence or absence of verruciform cells and the structure of the infusoriform larva are characteristic for each species. When the dicyemids are examined closely, a simple structure is revealed. A single in- ternal, axial cell runs almost the entire length of the body in the nematogens and rhombo- gens. In total length these adult vermiform stages range from 500 to 10,000 um depend- ing on the species. Reproductive products are relegated to the interior of the axial cell of the parent. These cells function as nurse or fol- licular cells providing both protection and nourishment for the germ cells and develop- ing embryos. The axial cell is surrounded by a jacket of 20 to 40 large somatic or peripheral cells which are entirely ciliated externally. The head or anterior end is modified into a calotte, by means of which the parasite attaches to the host renal tissue. The actual shape of the calotte varies a great deal depending on the species. There is no trace of a differentiated digestive, circulatory, nervous, respiratory, glandular, or excretory system. No muscles, sensory receptors, or skeletal elements are present. In fact, nothing comparable to or- gans, tissues or glands are observed. The life cycle (Fig. 1) has been a contro- versial subject since Erdl (1843) and von Kolliker (1849) first observed the presence of two stages in the renal organs of the cephalo- pod host. Despite comprehensive study, the life cycle is still incompletely known (see re- views in Hyman, 1940; Nouvel, 1947; McConnaughey, 1951, 1968; Stunkard, 1954; Czihak, 1958; Grasse, 1961; Lapan & Morowitz, 1975). In its simplest expression it is composed of an alternation of essentially isomorphic, parent generations. The embryos of all known stages develop intracellularly until released through rupture of the parent's body wall. Cleavage is determinant. A definite cell number is attained early in development and subsequent growth is by cell enlargement. The mode of entry and the initiation of the infection is not known or has not been dem- onstrated experimentally. The earliest known stage observed in juvenile cephalopods is termed a stem nematogen. This stage differs from the typical adult vermiform stages princi- pally in having three axial cells instead of the usual one. However, all vermiform embryos produced by stem nematogens have only one axial cell. Immature hosts harbor populations of nematogens all of which contain elongate vermiform embryos in their axial cells. The embryos develop asexually from agametes (axoblasts) and resemble the parents when released. Constant proliferation of daughter nematogens eventually results in an enor- mous population of dicyemids which fills the renal organs of the cephalopod host. In older hosts the vermiform embryos are replaced by gamete producing infusorigens. The parent is now called a rhombogen. The resulting zygotes develop into ovoid embryos, which, when full grown are termed infusori- form larvae. The infusoriform larvae are anatomically the most complex of any stage in the life cycle. After breaking out of the par- ent's body the infusoriforms escape from the renal environment with passage of the urine. The fate of this dispersal stage and the phase(s) of the cycle which occur(s) outside the cephalopod host are still a mystery. Sev- eral authors have suggested that the infusori- form larvae or their released germinal cells must infect a secondary benthic host since they are not attracted to young cephalopods (see Nouvel, 1947; McConnaughey, 1951; Stunkard, 1954). On the other hand, Lapan & Morowitz (1975) recovered dicyemids in the renal organs of Sepia reared from eggs in isolated aquaria and exposed only to infusori- form larvae. This indicates that an intermedi- ate host may not be necessary. Twice during the course of an infection a change of phase takes place. The initial infec- tive phase is brief and when the stem nemato- gens are spent they disappear and are re- placed by nematogens. As the cycle pro- gresses all nematogens are eventually trans- formed into rhombogens during which stage gametic reproduction takes place. In octo- pods, the transition from nematogens to rhombogens is prolonged and a mixture of stages is often found (Hochberg, 1971) whereas in cuttlefish a rapid metamorphosis is completed at the time of sexual maturation of the host (Nouvel, 1933). Because the shift in phase is particularly evident in adult cephalopods, most authors have suggested that the hormonal flux associated with host maturation acts as the trigger. However, at the time of transition the renal organs are maximally crowded with parasites. Hochberg PARASITES IN CEPHALOPOD “KIDNEYS” 125 fertilization 9 Y meiosis Octopus outside cephalopod host FIG. 1. Life cycle of the dicyemid mesozoan, Dicyemennea, in octopus host. 1. larval stem nematogen; 2. stem nematogen; 3. vermiform embryo; 4. nematogen; 5. rhombogen; 6. infusorigen; 7. infusoriform larva released from parent. Density of parasites: A, low; B, high. (1971) postulated and Lapan & Morowitz (1975) later demonstrated that population pressure or crowding may actually be the key factor which initiates the shift from the nematogen to the rhombogen phase. CHROMIDINID CILIATES Next to the dicyemids, ciliates are the most frequently encountered parasites of cephalo- pods. However, only a few published studies deal with the many unusual forms that occur in the excretory organs, digestive glands and on the gills of squids and octopods. The genus Chromidina is restricted to a small group of vermiform ciliates which infect the renal organs of cephalopods. In the past, Chromidina has often been lumped with Opalinopsis, a genus of ovoid ciliates which infects the digestive glands of a number of cephalopods. The taxonomic position of these 126 HOCHBERG two genera has been the subject of consider- able debate. An affinity between these two highly specialized cephalopod parasites and the apostomes, which typically occur on crustaceans as epibionts, was first postulated by Chatton & Lwoff (1926) who two years later proposed this more definitely (Chatton & Lwoff, 1928). Their ideas with regards to this relationship, were expanded in 1930 and finally in 1931 they demonstrated morpho- logical stages in the life cycle of Chromidina that were very similar to stages in the life cycle of the apostomes, especially the foettin- geriids (Chatton & Lwoff, 1930, 1931). The extensive monograph by Chatton & Lwoff (1935) provides the only definitive study of Chromidina and the apostomes as a whole. As part of my doctoral research | reviewed the systematic literature and together with new material examined | reaffirmed place- ment of Chromidina in the order Apostomea but relegated the genus to its own family, the Chromidinidae, as separate from the Opalinopsidae (Hochberg, 1971). Though widely accepted as a well characterized and relatively homogeneous group, the apo- stomes are still an enigmatic assemblage without definite affinities in the accepted scheme of ciliate evolution (Corliss, 1979). A total of 23 species of cephalopods repre- senting 20 genera are currently known to har- bor chromidinid ciliates (Table 1). These cili- ates characteristically infect epi- and meso- pelagic squids and octopods. Infection of benthic or epibenthic hosts has been occa- sionally reported but in all cases the ciliates are found only in octopods which have plank- tonic larvae (i.e. Octopus salutii, O. vulgaris, Scaeurgus unicirrhus, and Eledone cirrhosa) or in sepioids whose young feed in surface waters (i.e. Sepia elegans, S. orbigniana and Sepiola rondeleti). The ciliates, which are contracted through association with crusta- ceans living in the water column, are brought to the bottom at the time of settlement. Once dicyemid infections are established, the cili- ates are slowly eliminated as the dicyemids multiply and fill the entire renal habitat (Hochberg, 1971; Nouvel, 1945). Only three species of Chromidina are de- scribed in the literature, though most earlier work has not been critically evaluated. On the basis of recent work at least ten different spe- cies are recognized and considering the potential hosts not yet examined, a total of perhaps twenty species may eventually be referred to the genus. Two basic body shapes are observed. Chromidina coronata has an inflated anterior end and a conspicuous crown of elongate cilia, whereas in C. elegans the anterior end is not swollen and the ciliary crown is lacking. In other ways the species are almost identical. Like the better known foettingeriids, Chromidina undergoes a complex polymor- phic life cycle which involves an ordered se- quence of distinct phases. Hochberg (1971) elucidated the two-host cycle as seen in Fig. 2. Young squids pick up the ciliates when they associate with or feed on swarms of pelagic crustaceans, such as euphausiids. At present the method of entry is not known. Within the cephalopod, the stages of the cycle show considerable modification and condensation as compared with the small, ovoid and less specialized foettingeriids (see especially Bradbury, 1966; and Chatton 8 Lwoff, 1935). In Chromidina, the vegetative and divisional phases are combined into long, thin tropho- tomonts. These vermiform individuals attach to the renal appendages or digestive duct ap- pendages by means of stiff thigmotactic cilia covering the anterior end. The posterior end which is actively involved in nutrient uptake and division, hangs free in the fluid-filled coelomic sac. Reproduction takes place by unequal, transverse fission or budding at the posterior end of the body. Chromidina, unlike the foettingeriids, does not encyst prior to division. Two distinct budding patterns are ob- served, monotomy and palintomy. In young hosts, the ciliates all produce large, single buds, termed apotomites, which re- semble the parents. When detached they are transformed directly into daughter tropho- tomonts. By means of this initial budding process the number of ciliates within the renal sac is greatly increased. Eventually, the renal habitat is saturated with ciliates. Chemical factors related to the density of the parasites or to host maturation probably trigger the second divisional phase as occurs in the dicyemids. In older hosts, the ciliates undergo palintomy, a multiple fission process which results in long chains of 8, 12 or 24 small buds. Eventually, tiny, ovoid dispersal stages, termed tomites, are produced which bear little resemblance to the parents. The tomites conjugate immediately after detach- ment from the parent tropho-tomonts and then exit through the renal pores with the passage of urine. Once in the sea the ciliates swim about searching for a new host. Upon contact with a PARASITES IN CEPHALOPOD “KIDNEYS” 127. Pterygioteuthis conjugation meiosis \ Y ==> FIG. 2. Life cycle of the apostome ciliate, Chromidina in squid (1-7) and on euphausiid hosts (8-9): 1. protropho-tomont; 2. 1” tropho-tomont; 3. production of apotomite via monotomy (single fission); 4. apotomite; 5. 2” tropho-tomont; 6. production of tomites via palintomy (multiple fission); 7. tomite detached from parent; 8. 1° phoront; 9. 2° phoront. Density of parasites: A, low; B, high. euphausiid or other appropriate crustacean host the tomites encyst on the mouthparts and setaceous appendages of the second host. Phoronts, or encysted resting stages of sev- eral sizes have been found indicating that the ciliates undergo a series of growth phases. Euphausiids are known to moult every few days. It is presumed that with each moult the ciliates excyst, feed on exuvial fluids in the cast off moult, grow and then reencyst on another crustacean. Eventually a size is attained which is infective to the cephalopod and the cycle begins again. The maximum length of vermiform stages in the cephalopod renal organs range from 400 to 2,000 um depending on the species. The infraciliature of the tropho-tomonts con- sists of a tight dextral helix, continuous with- out breaks from the anterior to the posterior pole. Typically 12-14 kineties are present. 128 HOCHBERG The macronucleus is an open network of chromatin which ramifies throughout the en- tire body. A tiny, spindle-shaped micro- nucleus is located in the posterior of the body in the region of the future fission plane. Mouth, rosette and contractile vacuole, typi- cally found in other apostomes, are absent in the vermiform stages of Chromidina. Occasionally hypertrophonts are found. De- scribed by Collin (1914, 1915), they may measure up to 5,000 um. These apparently are individuals that have penetrated the epithelium of the reno-pancreatic append- ages and entered the blood spaces within. Here they increase rapidly in size, probably because of the high osmotic pressures. The ciliates which are thus imprisoned and im- mobilized, appear degenerative due to attack by phagocytes from the blood of the host. With each successive division during palintomy the kinities are shortened and straightened. Fully developed tomites range in size from 15 to 30 um. They are pyriform in shape with a convex dorsal surface and a flat or slightly concave ventral surface. The oral field and contractile vacuole develop following detachment from the parent. The tomites of Chromidina closely resemble the tomites of many of the crustacean epibionts and it is at this stage that affinities between the apo- stome families, the foettingeriids and the chromidinids, are most evident. This affinity allows comparison of Chromidina with the foettingeriids which are considered to be the most primitive of the apostomes. This in turn provides insight into the adaptations and modifications which have occurred as a result of an endoparasitic existence. CEPHALOPOD EXCRETORY ORGANS: MORPHOLOGY AND FUNCTION Physical and chemical environments en- countered by parasites within hosts influence the selection of their form and life cycles. Similarities in form and reproductive strate- gies between the dicyemid mesozoans and the chromidinid ciliates, two phylogenetically unrelated parasites, suggest that the excre- tory organs are responsible for directing the course of their evolution. It is with this in mind that | will briefly review the main features of the excretory system of cephalopods which appear to be so ideally suited for maintaining parasites. The first detailed examination of the ex- cretory system of cephalopods was by Cuvier (1817) . The history of morphological investi- gations and many structural details are pro- vided by Turchini (1923) and more recently by Marthy (1968) and Schipp & Boletzky (1975). The secretory and reabsorptive functions of the excretory organs are reviewed by Harri- son & Martin (1965), Potts (1965), and Potts & Todd (1965) among others. Chemical compo- sition of the excretory fluids are reviewed by Lapan (1975). Additional details and a review of the literature is provided in Hochberg (1971). The term excretion refers to the sum of all activities involved in the formation and elimi- nation of wastes. Through this process an ani- mal regulates the fluid and chemical balance of its internal environment. As compared with other molluscs, cephalopods are very efficient in maintaining a rather constant internal medi- um. The elaboration of the excretory fluid or urine is actually the result of three separate physiological processes: a) ultrafiltration, b) secretion, and c) reabsorption. Wastes which are produced and eliminated by the cephalo- pod are only those compounds and molecules from which no further energy can be derived. These waste products, however, are not be- yond the use of other organisms, namely parasites. In cephalopods, excretion occurs specifi- cally in areas where the blood comes into in- timate contact with the transport-active epi- thelium. The main bulk of excretion is carried out by the renal complex (renal appendages and renal sac) in conjunction with the branch- ial heart complex (branchial heart append- ages and pericardium). In the decapods (squids, cuttlefishes and sepiolids) the proc- ess is aided by the digestive gland complex (digestive duct appendages and reno- pancreatic sac). Parasites infest all three sites. Since the excretory system of cephalo- pods is actually a complex of organs and since the renal component may not function in filtration it is best to avoid referring to these organs as a kidney. However, since the word is SO commonly applied to the excretory organs, | have continued to use it in quotation marks throughout this paper. When the ventral mantle of a cephalopod is cut open and the mantle walls laid back, a pair of conspicuous fluid-filled sacs is seen lying on the surface of the visceral mass. The renal organs are the principal excretory organs in cephalopods and the principal site for attach- ment of parasites. They are derived from PARASITES IN CEPHALOPOD “KIDNEYS” 129 mesodermal masses found in the coelomic complex of the developing embryos. As de- velopment proceeds one wall of the expand- ing renal coelom forms the thin wall of the urine-filled renal sac while the other wraps around the walls of the large veins which traverse the renal cavity. In the region of con- tact the walls of the vena cava become highly convoluted and eventually develop into the glandular renal appendages. In the append- ages a single-cell thick excretory epithelium separates the blood from the urine in the renal sac. On the lumen side, the renal appendages are covered by a microvillar brush border. Peristaltic activity of the circulatory system creates a flow of blood from the veins into the interior spaces of the renal appendages. Upon contraction, the muscular tissues of the appendages pump some blood back into the main venous vessels and compress the rest within the blind sacs of the highly folded ap- pendages. This action may provide pressure sufficient to effect some filtration although this is debatable. The glandular nature of the renal epithelium and its intimate contact with the blood suggests that the main function is secretory. Harrison, Martin, Potts and others have demonstrated active transport of a num- ber of inorganic and organic compounds through the epithelium of the renal append- ages. lonic resorption, important in the con- centration of urine, is presumed to occur in the renal sac. Water resorption probably does not occur in the area of the renal organ. Rhythmic contractions or pulsations of the renal appendages, as a whole, serve to con- stantly mix the fluids in the renal sac. Urine is periodically voided through tiny renal pores and then flushed from the mantle cavity dur- ing respiratory or locomotory contractions. Due to the dead spaces surrounding the renal appendages emptying is never complete. As a result fluid is always present. Accumulations of mucus, proteins and nitrogenous wastes may escape elimination for some time. The constant presence of a nutrient-rich fluid envi- ronment is considered essential for the main- tenance of parasites. The branchial hearts form a complex with the branchial heart appendages (or peri- cardial glands) in both octopods and deca- pods. The dark, muscular branchial hearts, located at the base of the gills, are accessory pumping structures which help to irrigate the gills. The lumen of these organs continues into the branchial heart appendages where it forms a system of blood sinuses. Each ap- pendage is surrounded by a pericardium which empties via the reno-pericardial canal into the renal sacs. Similar to the renal ap- pendages, the branchial heart appendages of sepioids and teuthoids are convoluted and are lined with a transport-active epithelium covered with a microvillar brush border. The high systolic pressure in the branchial hearts provides ample pressure for ultrafiltration to occur and in fact this region and not the renal appendages is the principal site for blood fil- tration. Cilia, present on the cells of the reno- pericardial canal, beat in the direction of the renal organs and are thought to additionally aid filtration by lowering the pressure inside the pericardial cavity. The resultant reno- pericardial filtrate or “primary urine” serves as the aqueous vehicle for excretory products formed by the renal appendages. Resorption of certain constituents such as glucose occurs in the pericardium and in the reno-pericardial Canal. A single species of dicyemid, Dicyemennea brevicephaloides, has been recovered from the branchial heart coela of Rossia pacifica, a small west American sepiolid (Hoffman, 1965). Other dicyemid and chromidinid para- sites probably occur in this site in sepioids and teuthoids but simply have been missed in routine autopsies. In octopods, the coelomic spaces surrounding the branchial heart ap- pendages are greatly reduced and the peri- toneum of the appendages is smooth (Witmer & Martin, 1973), hence parasites are unlikely to be present. In decapods, a third organ is present which serves an excretory function. These are the grape-like “pancreatic” or digestive duct ap- pendages (see Bidder, 1976) which cover the paired ducts connecting the “liver” or diges- tive gland and the caecum. The elaborations of these ducts are located in a single median sac which lies dorsal to and opens into the renal sac. Taken together the renal and the digestive duct appendages are contained within the reno-pancreatic coelom. The digestive duct appendages are com- posed of two layers of epithelial cells sepa- rated by a layer of blood vessels and sinuses. The low, outer epithelium is derived from the reno-pancreatic sac during development and is similar to the transport-active epithelium of the renal appendages, complete with micro- villar brush border. In function, the digestive duct appendages are thought to secrete inor- ganic and organic compounds and to concen- trate the urine by reabsorbing ions and water. 130 HOCHBERG In decapods, the two organs are contained within a common coelomic cavity, hence, when parasites invade this region they are found attached to both the renal and the di- gestive duct appendages. In octopods, this organ is embedded within the capsule of the digestive gland and, hence, not subject to in- fection by “kidney” parasites. In summary, the “kidneys” of cephalopods are ideally suited for the establishment and maintenance of parasites. As such, these organs meet the parasite’s requirements by providing: a) a substrate for attachment; b) a constant fluid bath; c) a source of nutrients; and d) a simple exit for release of dispersal stages. Vermiform stages of chromidinid ciliates and dicyemid mesozoans live and reproduce within the excretory organs of cephalopods. Specifically they are located on the renal ap- pendages, digestive duct appendages and occasionally on the branchial heart append- ages. In all cases the appendages to which the parasites are attached are convoluted and the epithelial substrate is transport-active and covered with a microvillar brush border. Dense thigmotactic cilia which interfinger with the microvillar surface of the appendages help to hold the parasites in place. Undula- tions of the body generated by ciliary beat create additional forces directed toward the surface of the excretory appendages. In many cases the anterior ends of both the dicyemids and chromidinids are imbedded between the convolution of the excretory appendages or they may even occupy individual depressions or “crypts” in the renal surface (Ridley, 1968). All the appendages mentioned above are surrounded by fluid-filled coelomic spaces into which the parasites hang. Periodic and incomplete emptying of these coelomic sacs ensure that the parasites are constantly sur- rounded by fluids. The parasites derive all their metabolic requirements from the dis- solved nutrients within the excretory fluids. A great deal of study has gone into unraveling the physiology of urine formation and the chemical composition of the urine, but little is known about the actual products utilized by the parasites. Since urine from uninfected cephalopods has never been analyzed we do not know what effects the parasites have on the excretory processes nor do we know how they modify the composition of the urine through discharge of their own metabolic wastes. Finally, in order to complete their life cycles, the parasites must find their way to the exte- rior of their host. When dispersal stages are produced within the renal organs of the cephalopod, they are easily voided through the renal pores along with the urine and then flushed out of the mantle cavity of the host. DISCUSSION Possession of similar characters, by phylo- genetically unrelated organisms, consequent upon a similar mode of living is termed con- vergence. The similarity in: a) site of infection; b) sedentary habits; c) general shape; d) body proportions; and e) diphasic life cycle, of both the dicyemids and the chromidinid ciliates is an example of such a phenomenon. Evidence suggests that the adaptive response of these two ciliated parasites to the excretory organs of cephalopods has resulted in the conver- gence of both their form and reproductive strategies. In essence, evolutionary proc- esses have been restricted by environmental factors which operated to select for develop- ment in specific directions. An elongate, vermiform shape; terminal (anterior) holdfast organ; and terminal (poste- rior) budding are not uncommon traits in the animal kingdom. These features are variously expressed in groups of hymenostome and astome ciliates, catenulate turbellarians, cestodes, and annelids among others. How- ever, as displayed by Chromidina, these fea- tures are unique within the apostome ciliates. Several additional features are displayed by the chromidinids which are not present in simpler, more primitive apostomes such as Gymnodinioides or Hyalophysa. These in- clude: their location in the excretory organs of cephalopods; their greatly increased size, polarized along the anterior-posterior axis; the absence of a rosette and functional cyto- stome; and unencysted divisional stages. Un- like the more typical foettingeriids, single buds or apotomites, produced by monotomy, are a constant feature of the life cycle. In Chromi- dina, protomites, produced by palintomy, are attached posteriorly to the parent in long, linear, multiple bud chains. Comparison of Chromidina with primitive or less complex apostomes, as mentioned above, provides insight into the modifications of morphology and life cycle which are cor- related with their endoparasitic specialization. Using these modifications as a guide, it is possible to postulate that similar changes PARASITES IN CEPHALOPOD 3005 “KIDNEYS” 131 FIG. 3. Convergence of chromidinid ciliates and dicyemid mesozoans. Parasites in young cephalopods (A & B); parasites in mature hosts (C & D). Chromidina from Pterygioteuthis (A & C); Dicyema from Octopus (B 8 D). 132 HOCHBERG must have occurred in the evolution of the dicyemid mesozoans, where no relatives are available for comparison. The diphasic life cycle, of both groups, is beautifully adapted to the requirements of their endoparasitic environment. Two distinct types of embryos (dicyemids) or buds (chromidinids) are produced within or at- tached to essentially isomorphic parents (Fig. 3). In young hosts, where there may be only a few initial parasite individuals, elongate daughter individuals (apotomites and vermi- form embryos), resembling the parents, re- main in the coelom of the renal organs when released. Endogenous proliferation of this sort functions to greatly increase the numbers of vegetative stages which eventually spread throughout the excretory organs. In older hosts, a change of phase triggers the produc- tion of small dispersal forms (tomites and in- fusoriforms) which are discharged from the host with elimination of urine. Widely broad- cast in the watery environment, these motile forms function to disseminate the parasites to new hosts. A convergence in form is also evident. In both groups of parasites the anterior end is often inflated and is always covered with short, stiff thigmotactic cilia. The parasites adhere, by this modified anterior end, to the brush-bordered, transport-active epithelium of the excretory appendages. The gradient, es- tablished when the parasites attach at one end while the other end hangs free in a fluid- filled coelomic cavity, results in the differentia- tion of distinct anterior and posterior regions. Growth, which occurs along this antero- posterior axis, imparts a vermiform shape to both organisms. This is most obvious in Chromidina when contrasted with other more typical apostomes. Almost without exception, the foettingeriids are small and compact ciliates. Only in Polyspira is an elongate, linear form observed, and then only during palintomy when chains of protomites are be- ing formed. Definite anterior and posterior re- gions are not distinguishable in this case (Chatton & Lwoff, 1935). The relative body proportions of adult vermiform stages of both dicyemids and chromidinids are almost identical. Nemato- gens, rhombogens, and tropho-tomonts gen- erally range in length from 1,000 to 5,000 um. Chromidina, in comparison with other apostomes, has undergone a many-fold increase in length. It can be assumed that the same holds true for the dicyemids, and that any resemblance to recent invertebrate forms is coincidental. The size of daughter individuals is equally similar. Vermiform embryos, produced within dicyemids, range from 100 to 300 um in length. Apotomites, produced by chromi- dinids, tend to be slightly larger. The similari- ty, in size and shape, of daughter stages is best witnessed by comparison of the tiny, ovoid dispersal forms. Both the tomites and the infusoriform larvae range in length from 20 to 40 um. Since these stages leave the host via the renal pores, it is assumed that their size is a function of the diameter of the open- ings of these tiny pores. In essence, they must be small enough to be expelled through the pores along with the urine. Concurrent infections rarely occur in nature. Cephalopods hosting the two parasites are normally spatially isolated. Chromidina typi- cally infects oceanic cephalopods which never contact the bottom, whereas dicyemids are known only from exclusively benthic or epibenthic hosts. The exploitation of the “kidneys” of cephalopods by these unusual vermiform parasites thus is facilitated and maintained by the habits of the hosts and the spatial separation of the infective stages. In the absence of competition, adaptation to the selective pressures within the excretory envi- ronment has favored convergence or the de- velopment of similar morphological and re- productive characteristics in these two unre- lated groups of parasites. ACKNOWLEDGEMENTS | thank Bayard McConnaughey, Elmer Noble, Robert Profant and Clyde Roper for reviewing the manuscript. | am especially grateful to Richard Young for providing space and facilities on several recent oceanographic cruises off Hawaii. While on these cruises | was able to wrap up the loose ends on many of the ideas expressed here. The drawings were translated from my rough sketches and rendered in final form by Jamie Calhoun. LITERATURE CITED ALDRICH, F. A., 1964, Observations on the New- foundland bait squid (Шех illecebrosus Lesueur, 1821), and the netting of squid in Newfoundland bays. Special Report to the Canadian Depart- ment of Fisheries, Industrial Development Branch, Ottawa, 22 ms. p. PARASITES IN CEPHALOPOD “KIDNEYS” 133 ALLISON, F. R., 1966, A new species of adult Allocreadidae from Octopus maorum Hutton. Record of the Canterbury Museum, 8: 81-85. BIDDER, A. M., 1976, New names for old: the cephalopod “mid-gut gland.” Journal of Zoology, 180: 441-443. BRADBURY, P. C., 1966, The life cycle and mor- phology of the apostomatous ciliate Hyalophysa chattoni n. g., n. sp. Journal of Protozoology, 13: 209-225. CAVOLINI, F., 1787, Memoria sulla Generazione dei Pesci et dei Granchi. Napoli. CHATTON, E. & LWOFF, A., 1926, Sur des para- sites branchiaux internes du Portunus depurator et sur leurs relations ontogenetiques probables avec les Infusoires (Opalinopsidae) des Cepha- lopodes. Comptes Rendus des Séances de la Societe de Biologie et de ses Filiales et As- sociees, 94: 282-285. CHATTON, E. & LWOFF, A., 1928, Sur la structure, l'évolution et les affinités des Opalinopsides (Cilies) des Cephalopodes. Comptes Rendus Hebdomadaires des Séances de l'Académie des Sciences, Paris, 186: 1382-1384. CHATTON, E. & LWOFF, A., 1930, Note prélimi- naire sur la systématique des Ciliés Apostomea (Foettingeriidae et Opalinopsidae). Bulletin de la Societe zoologique de France, 55: 296-327. CHATTON, E. & LWOFF, A., 1931, La conception des Ciliés Apostomes (Foettingeriides + Opalinopsides). Preuves de sa validite. Comptes Rendus Hebdomadaires des Séances de l'Académie des Sciences, Paris, 193: 1483- 1485. CHATTON, E. & LWOFF, A., 1935, Les Ciliés Apostomes. Morphologie, cytologie, éthologie, evolution, systematique. Premiere partie. Apercu historique et général. Etude monographique des genres et des espèces. Archives de Zoologie Experimentale et Generale, 77: 1-453. COLLIN, B., 1914, Sur les formes d'involution d'un Infusoire Cilié dans le rein d'un Céphalopode. Comptes Rendus Hebdomadaires des Séances de l'Académie des Sciences, Paris, 158: 891- 892. COLLIN, B., 1915, A propos de Chromidina elegans (Foettinger). Comptes Rendus Heb- domadaires des Seances de l'Académie des Sciences, Paris, 160: 406-408. CORLISS, J. O., 1979, The Ciliated Protozoa. Characterization, Classification and Guide to the Literature. Ed. 2, Pergamon Press, New York, 455 p. CUVIER, G., 1817, Mémoires pour servir à l'his- toire et a l'anatomie des Mollusques. Mémoire sur les Cephalopodes et sur leur anatomie. Paris. CZIHAK, G., 1958, Morphologie und Entwicklungs- geschichte der Wirbellosen. Mesozoa. Fort- schritte der Zoologie, 11: 1-15. DOBELL, C. C., 1909, Some observations on the infusoria parasitic in Cephalopoda. Quarterly Journal of Microscopical Science, 53: 183-199. ERDL, —., 1843, Uber die beweglichen Faden in den Venenanhangen der Cephalopoden. Archiv fur Naturgeschichte, 1: 162-167. FOETTINGER, A., 1881a, Un mot sur quelques In- fusoires nouveaux, parasites des Céphalo- podes. Bulletin de l'Académie royale de Belgique, 1: 887-895. FOETTINGER, A., 1881b, Recherches sur quel- ques Infusoires nouveaux, parasites des Cepha- lopodes. Archives de Biologie, 2: 345-378. GRASSE, P.-P., 1961. Classe des Dicyémides. In GRASSE, P.-P. (ed.), Traité de Zoologie, 4(1): 693-719, Plathelminthes, Mésozoaires, Ac- anthocephales, Nemertiens. Masson, Paris, 944 p. HARRISON, F. M. & MARTIN, A. W., 1965, Excre- tion in the cephalopod, Octopus dofleini. Journal of Experimental Biology, 42: 71-98. HOCHBERG, F. G., 1971, Some aspects of the biology of cephalopod kidney parasites. Ph.D. dissertation, University of California, Santa Barbara, 272 p. HOFFMAN, E. G., 1965, Mesozoa of the sepiolid, Rossia pacifica (Berry). Journal of Parasitology, 51: 313-320. HYMAN, L. H., 1940, Phylum Mesozoa. in HYMAN, L. H. The Invertebrates, 1: 233-247, Protozoa through Ctenophora. McGraw-Hill, New York, 126 р. KOLLIKER, A. VON, 1849, Ueber Dicyema para- doxum, den Schmarotzer der Venenanhange der Cephalopoden. Berichte von der König- lichen Zootomischen Anstalt zu Wurzburg, 2: 59-66. J KROHN, A., 1839, Uber das Vorkommen von Ento- zoen und Krystallablagerungen in den schwam- mingen Venenanhangen einiger Cephalopoden. Notizem aus dem Gebiete der Natur- und Heilkunde, 11: 213-216. LAPAN, E. A., 1975, Studies on the chemistry of the octopus renal system and an observation on the symbiotic relationship of the dicyemid Meso- zoa. Comparative Biochemistry and Physiology, 52A: 651-657. LAPAN, E. A. & MOROWITZ, H. J., 1975, The dicyemid Mesozoa as an integated system for morphogenetic studies. Journal of Experimental Zoology, 193: 147-160. MARTHY, H.-J., 1968, Die Organogenese des Coelomsystems von Octopus vulgaris Lam. Revue Suisse de Zoologie, 75: 723-763. MCCONNAUGHEY, B. H., 1951, The life cycle of the dicyemid Mesozoa. University of California Publications in Zoology, 55: 295-336. MCCONNAUGHEY, B. H., 1959, Dicyemennea nouveli, a new Mesozoa from central California. Journal of Parasitology, 45: 533-537. MCCONNAUGHEY, B. H., 1968, The Mesozoa. In FLORKIN, M. & SCHEER, B. T., Chemical Zoology, 2: 557-570, Porifera, Coelenterata, and Platyhelminthes. Academic Press, New York, 639 p. NOUVEL, H., 1933, Recherches sur la cytologie, la 134 HOCHBERG physiologie et la biologie des Dicyemides. Annales de l'Institut Oceanographique, Monaco, n.s., 13: 165-255. NOUVEL, H., 1945, Les Dicyémides de quelques, Céphalopodes des cótes francaises avec indica- tion de la presence de Chromidinides. Bulletin de l'Institut Océanographique, Monaco, 887: 1-8. NOUVEL, H., 1947, Les Dicyémides. 1'8 partie: Systématique, Générations vermiformes, In- fusorigene et sexualité. Archives de Biologie, 58: 59-219. POTTS, W. T. W., 1965, Ammonia excretion in Octopus dofleini. Comparative Biochemistry and Physiology, 14: 339-355. POTTS, W. T. W. 8 TODD, M., 1965, Kidney func- tion in the Octopus. Comparative Biochemistry and Physiology, 16: 479-489. RAABE, H., 1934, La degenerescence du rein des Céphalopodes provoquée par un Ascomycete. Bulletin de l'Institut Océanographique, Monaco, 610: 1-8. RIDLEY, R. K., 1968, Electron microscopic studies on dicyemid Mesozoa. |. Vermiform stages. Journal of Parasitology, 54: 975-998. SCHIPP, R. & BOLETZKY, S. VON, 1975, Mor- phology and function of the excretory organs in dibranchiate cephalopods. Fortschritte der Zool- ogie, 23: 89-111. SHORT, R. B. & HOCHBERG, F. G., 1969, Two new species of Dicyemennea (Mesozoa: Di- cyemidae) from Kaikoura, New Zealand. Journal of Parasitology, 55: 583-596. SHORT, R. B. & POWELL, E. C., 1968, Mature digenetic trematodes from New Zealand octo- puses. Journal of Parasitology, 54: 757-760. STUNKARD, H. W., 1954, The life-history and sys- tematic relations of the Mesozoa. Quarterly Re- views of Biology, 20: 230-244. TURCHINI, J., 1923, Contribution a l'étude de l'histologie comparée de la cellule renale. L'excretion urinaire chez les mollusques. Archives de Morphologie Generale et Experi- mentale, 18: 7-253. WITMER, A. & MARTIN, A. W., 1973, The fine structure of the branchial heart appendage of the cephalopod Octopus dofleini. Zeitschrift für Zellforschung und Mikroskopische Anatomie, 136: 545-568. MALACOLOGIA, 1982, 23(1): 135-163 THE FUNCTIONAL MORPHOLOGY OF A VENTRAL PHOTOPHORE FROM THE MESOPELAGIC SQUID, ABRALIA TRIGONURA R. E. Young! and J. M. Arnold? ABSTRACT The vertically migrating squid, Abralia trigonura, has at least two types of photophores in- volved in counterillumination. The most complex of these is described. We suggest that this photophore functions in the following manner. Innervated photocytes which contain crystalloids extract a component of the luminous reaction, presumably luciferin, from blood vessels via numerous finger-like processes. Energy for the reaction is supplied by banks of mitochondrial cells. Light is emitted by the crystalloids which are stacked to form a photogenic cone. The photogenic cone lies at the focus of a spherical proximal reflector. This reflector is an interference structure that selectively reflects light outward and contributes to color regulation by the alteration of its reflectance characteristics through changes in the diameter of its collagen rods. An interference filter, the axial stack, selectively transmits light and contributes to color regulation by altering the thickness of the fluid-filled spaces between its platelets. The torus and distal cap are ‘thick film” reflectors that slightly diffuse the highly directional emission beam. Another interference structure, the distal reflector, reflects outwardly light emitted be- tween the distal cap and the proximal reflector. Numerous chromatophores can conceal the photophore or aid in adjusting the radiance pattern of emitted light. The intensity of emitted light can be regulated over at least a 325-fold range, and the spectral emission maximum can vary between 480 and 536 nm. The angular distribution of emitted light can be regulated as well, but measurements have not been attempted. The photophore can rock on a fluid cradle which enables proper orientation regardless of the squid’s attitude in the water. The complexity and the alterability of the light emission properties of this photophore, in combination with one or two other types of photophores, indicate that this squid can match many of the varying patterns of intensity, color and radiance of downwelling light encountered in its oceanic habitat and, thereby, conceal itself by eliminating its silhouette from potential visual predators. Key words: photophore; squid; ultrastructure; bioluminescence; color. INTRODUCTION Abralia trigonura Berry, 1913 is a small mesopelagic squid endemic to the region of the Hawaiian Islands. This squid usually oc- cupies depths between 450 and 550 m during the day and the surface and 110 m at night (Young, 1978). At these depths the squid is often exposed to downwelling light that could silhouette it and, thereby, make it visible to potential predators. This squid, however, has the capability of eliminating its silhouette by counterilluminating with downward-directed bioluminescence at least under some labora- tory conditions (Young & Roper, 1977). Effec- tive counterillumination in the sea requires that the emitted light match the color, intensity and angular distribution of the downwelling light. During the day at depth only light inten- sity varies but at night, in near-surface waters, all three parameters can vary depending on the exact depth, moon phase and declination, and cloud cover. Photogenic organs capable of providing counterillumination under these variable conditions must be highly sophisti- cated. The photophores of related species of Abralia have been briefly examined by Hoyle (1894) and Joubin (1895). Mortara (1922) de- scribed in more detail the general features of the photophores of Abralia veranyi. However, without the aid of the electron microscope little about the functional morphology of the organs could be determined. MATERIALS AND METHODS Photophores fixed for electron microscopy came from live squid collected during a series 1Department of Oceanography, University of Hawaii at Manoa, 2525 Correa Road, Honolulu, Hawaii 96822, U.S.A. 2Pacific Biomedical Research Center, University of Hawaii, Kewalo Marine Laboratory, 41 Ahui Street, Honolulu, Hawaii 96813, U.S.A. (135) 136 YOUNG AND ARNOLD of cruises aboard the University of Hawaiis cac (distal) cap chromatophore research ship, R. V. KANA KEOKI off leeward cc (proximal) cup chromatophore Oahu, between 1975 and 1980. Fixation and co core embedding were carried out aboard ship. сос core cells Photophores were generally fixed for 1 hrin cr crystalloid 2.8% glutaraldehyde in filtered seawater buf- crc crystalloid cell fered with 0.01 M collidine at pH 7.5. Follow- dc _ distal cap ing a seawater rinse, the organs were post- dm dense cytoplasm fixed for 1 hr in 1% OSO: in seawater buffered dr distal reflector with 0.01 M collodine and adjusted to pH 7.5 ds dense substance with HCI. The photophores were then dehy- es extracellular space drated through an ethanol series and em- gr girdle bedded in Epon for thin sectioning. Sections Ic inner region of cap were stained with uranyl acetate followed by icc immature crystalloid cell lead citrate. Thin sections were viewed witha Im __ lamella Philips 201 electron microscope. m muscle Measurements of the luminescent emission mi mitochrondria spectrum were made according to procedures mt microtubules outlined in Young & Mencher (1980). Live п nucleus squid maintained in thin vinyl tubing were nc _ nucleus of core cell stimulated to luminesce by artificial illumina- ne nerve tion above the animal. A fiberoptic probe be- oc outer region of cap neath the squid transmitted the luminescence 0$ orbital space to a double monochrometer connected to a pc _ proximal cap photomultiplier tube (PMT). The output of the php photogenic plexus РМТ was fed into a photon-counting system pr proximal reflector connected to a desktop computer. Since ма- г root ter continuously flowed through the vinyl tube, fi ribbon the water temperature could be easily con- S (collagen) sheath trolled. sc sheath cell sur surface to torus ABBREVIATIONS у vesicles vc vesicle cell A Type A photophores ac apical chromatophore ax axial stack RESULTS axc axial stack cell axcl axial complex Abralia trigonura possesses five types of B Type B photophores photophores. Two types lie in a series on the бу blood vessel ventral surface of each eyeball and are not C Type C photophores involved in counterillumination. The most E photogenic cone abundant photophores (termed the Type A, A FIG. 1. Ventral head region of an adult Abralia trigonura showing the distribution of the three types (A, B, C) of integumentary photophores. The large white areas are ocular photophores. Scale bar is 4 mm. FIG. 2. Longitudinal section of the Type C photophore showing the general morphology. Compare with Fig. 5. Surface of the animal would lie above the top edge of the figure. The irregular outer convex surface of the posterior reflector is probably a fixation artifact. The large chromatophores surrounding the posterior re- flector absorb light that is not reflected by this color-selective reflector. Arrow—see Fig. 3. Scale bar is 50 um. FIG. 3. Cross-section of the Type C photophore at the level of the torus. Level of cut indicated by arrow in Fig. 2. Note the extensive orbital space and associated blood vessels. Scale bar is 50 um. FIG. 4. Longitudinal section nearly through the optical axis of the axial stack and photogenic cone showing the general relationships of the photogenic cone, the proximal and distal region of the axial stack and the zone of vesicles. Variations in spacing of the platelets of the axial stack is probably a fixation artifact. Scale bar is 5 um. ABRALIA PHOTOPHORES 137 Type B, and Type C) comprise the remaining three types. These types occur approximately in the ratio of 1(A):7(B):10(C) and they are intermingled and scattered over the ventral surface of the head, arms, funnel and mantle (Fig. 1). We describe here the structure of the Type C photophore which is not only the most abundant type but the most complex as well. UA Mi, Each Type C photophore is a small organ (140-180 um in diameter) that lies just be- neath the epidermis in a cup-like depression of the body musculature. For descriptive pur- poses the photophore can be subdivided into several components: 1) PROXIMAL CUP —a hemispherical region that forms much of the proximal half of the photophore; 2) DISTAL 138 YOUNG AND ARNOLD sur Lor N ÉS НЯ “ as), р 5 а $: HA = = АНИ > MATE 43) р = nx; B ee Bee OR APR NÉ Ta RE DOUTE SN ANAL e E FIG. 5. Schematic diagram of the Type C photophore. For clarity, nerves and mitochondrial cells are drawn only on opposite sides of the photogenic plexus. + FIG. 6. Longitudinal section of the photogenic cone showing the stacked crystalloids each in an individual cell. Note the close apposition of the crystalloids, and the dense material where the cytoplasm of the cell joins the crystalloid. Numerous microtubules can be seen in the crystalloid cells. Note the increase in size of the extracellular spaces toward the apex of the cone. Scale bar is 2 um. FIG. 7. Glancing section at the edges of the crystalloids showing the interconnecting extracellular spaces. Scale bar is 1 um. FIG. 8. High magnification of crystalloids showing the lamellar organization with similar alignment in adjacent crystalloids. Membranes bordering the extracellular spaces are associated with regions of dense cytoplasm (dm). Scale bar is 0.2 um. ABRALIA PHOTOPHORES E ne 140 YOUNG AND ARNOLD CAP—a bell-shaped structure that sits atop the proximal cup; 3) DISTAL REFLECTOR—a lamellar organ that surrounds much of the dome of the distal cap and whose broad sur- faces tilt slightly away from the optical axis of the photophore; 4) CHROMATOPHORES— several sets surround the photophore; 5) ORBITAL SPACE—large fluid-filled lacunae that surround the proximal cup. (See Figs. 2, 3, 5 for general morphology.) 1) PROXIMAL CUP—This portion of the photo- phore is composed of three primary regions: A. the photogenic plexus, B. the proximal re- flector, C. the axial complex (Fig. 2). A. Photogenic Plexus. This plexus is a highly vascularized region lying between the axial complex and the posterior reflector. The plexus is composed of several different cell types and tissues. Crystalloid cells. The fully differentiated crystalloid cells lie in the center of the photo- genic plexus. Each cell contains a single large crystalloid that lies freely within its cytoplasm (Figs. 4, 6, 9). Each crystalloid cell has many branches. The crystalloid is located at the center of the branching and the nucleus lies in one branch (Fig. 9). The crystalloid within each cell lies adjacent to those of neighboring cells. Together the crystalloids form a conical stack (conoid) which we term the photogenic cone (Figs. 4, 6). Individual crystalloids usual- ly have the shape of a thick half-disc (Figs. 6, 9). The two adjacent stacks of these half-discs form a conoid. For the most part, crystalloid cells branch only to one side of the photo- genic cone forming a “half-star” (Fig. 9). Nuclei of these cells are clustered in two small regions that lie on opposite sides of the cone (Fig. 9). The branches extend outward from the crystalloids, and subdivide into slender, finger-like processes that encircle blood ves- sels of the plexus, often forming a dense mat around the vessels (Fig. 10). The crystalloid cells are packed with dense arrays of microtubules that predominantly ex- tend between the crystalloids and the tips of the branches (Figs. 6, 9, 11). Near the crystal- loids radiating microtubules are intermingled with microtubules that tend to parallel the sur- face of the crystalloid (Figs. 6, 8). Adjacent to the crystalloids most microtubules blend into an electron dense, amorphous substance (Figs. 6, 7, 8). This substance is more exten- sive near the apical end of the cone. Occa- sionally structures resembling ciliary rootlets extend from this amorphous substance into the region of microtubules (Fig. 11). The crys- talloid cells also contain scattered tubular, branching, smooth endoplasmic reticulum which is most abundant near the crystalloids. Numerous vesicles, about the size of synaptic vesicles are observed in Glutaraldehyde- Osmium fixations (Figs. 6, 11). Mitochondria, however, are infrequently observed. The crystalloid is a highly organized crystal- like array of electron dense lamellae (Fig. 8). In section at right angles to the lamellae, the lamellae exhibit minor Z-shaped undulations, the result of alternating ridges or knobs. Ad- jacent crystalloids often exhibit parallel align- ment of their lamellae (Fig. 8). At their edges, the crystalloids blend abruptly into the cyto- plasm; they are not membrane-bound al- though they lie adjacent to the plasmalemma (Figs. 6, 8). The shape of the crystalloids near the apex of the cone is often more irregular than those at the base (Fig. 4). Near the crystalloids is a complex system of interconnecting extracellular spaces (Figs. 6, 7, 8). These spaces are large near the lateral edges of the crystalloids and are continuous with broad concavities or sometimes chan- nels that extend well into the adjacent cells. The spaces become progressively more ex- tensive toward the apex of the cone. Cell membranes bordering the channels are fre- quently distinguished by narrow, electron dense regions of cytoplasm (Fig. 8). The ma- terial within the extracellular spaces is not homogeneous. Electron light and dense areas intermingle often giving an indistinct striated appearance. The system of extracel- lular spaces is continuous with a broad, plate- like space that immediately underlies the axial stack (Fig. 7). The extracellular spaces of the axial stack are completely homogeneous electron light areas in contrast to spaces as- sociated with the crystalloids. Places can be found where the extracellular areas of these two regions join; however, they maintain their distinctiveness. Developing crystalloid cells which lack the nn FIG. 9. Cross-section through the crystalloids and the photogenic plexus. The crystalloid cells branch only to one side of the plexus. Their nuclei are grouped into two regions which lie on opposite sides of the plexus. Blood vessels and nerves are numerous in this section. Scale bar is 5 um. ABRALIA PHOTOPHORES 142 YOUNG AND ARNOLD FIG. 10. A crystalloid cell exhibiting multiple branching near the surface of a blood vessel. Note the numer- ous slender extensions of the crystalloid cell that cover the blood vessel. Scale bar is 2 um. FIG. 11. A crystalloid cell showing numerous adjacent nerves. Abundant microtubules can be seen within the crystalloid cell. Scale bar is 2 um. FIG. 12. Synapse between a nerve and a crystalloid cell. Scale bar is 0.1 um. ABRALIA PHOTOPHORES 143 dense arrays of microtubules and which fre- quently contain very small crystalloids are found in the fully differentiated photophore be- tween the photogenic cone and the posterior reflector (Fig. 5). Like the mature crystalloid cells these branched cells are in close contact with blood vessels. The branches that contact the blood vessels, however, are not as small Or aS numerous as are those of the mature cells. Mitochondria and Golgi bodies are com- mon in these cells and tubular, smooth endo- plasmic reticulum extends throughout the cytoplasm. The crystalloid cells of the devel- oping photogenic cone from an immature Type C photophore exhibited these same characteristics. Mitochondrial cells. The mitochondrial cells are highly branched structures that possess numerous large mitochondria (Fig. 15) The branches ramify throughout much of the photogenic plexus. Nuclei of these cells are located in the peripheral region of the plexus adjacent to the cells that form the proximal reflector. Because of the extreme tangle of the branches, individual cells cannot be traced very far from their nuclei. The branches are easily identified deep within the plexus by the lack of microtubules, the pres- ence of numerous large mitochondria, and often, by a more electron dense cytoplasm. Cytoplasm in the region of the nucleus con- tains abundant smooth endoplasmic reticu- lum and occasional Golgi bodies. Within the proximal region of the plexus the mito- chondrial cell processes are concentrated near the surface of the proximal reflector and FIG. 13. Nerves in the region of the cells that form the axial stack. Two nuclei of the vesicle cells are evident. Scale bar is 5 um. FIG. 14. Higher magnification of the nerves seen in Fig. 13 showing presynaptic vesicles. Scale bar is 1 um. 144 extend inward between blood vessels, photo- cytes and nerves. They do not penetrate, however, to the region surrounding the crys- talloids which is occupied only by crystalloid cells and occasional nerves. Blood vessels. Blood vessels are observed only in the region of the photogenic plexus and enroute to and from this tissue (Fig. 15). Relatively large vessels penetrate the photo- phore from the periphery at various points. They pass between the chromatophores that surround the proximal cup, over the reflector- secreting cells and into the photogenic plexus (Figs. 2, 3). There the vessels subdivide form- ing about 30-35 small (— 2.5 um) vessels that pass beneath the photogenic cone (Fig. 9). The morphology of these vessels is typical of exchange vessels (i.e. capillaries, see Brown- ing, 1979) (Fig. 10). A thick basal lamina is always present. The endothelial layer is in- complete and a single layer of deeply inter- digitating pericytes composes the vessel wall. Myofilaments, microtubules, smooth tubular ER, and mitochondria are common in the pericytes. No extensive extracellular matrix surrounds the vessels. Nerves. Numerous nerves packed with synaptic vesicles penetrate the photogenic plexus, presumably entering with the blood vessels above the margins of the proximal re- flector. A few small nerves pass into the re- gion of the axial stack (Figs. 13, 14). Most penetrate and entwine the branches of the crystalloid cells (Fig. 11). Rarely, however, do nerves penetrate to the immediate region of the crystalloids. Occasionally synapses are seen between the nerves and photocytes (Fig. 12). Other peripheral nerves are present that probably innervate the numerous chro- matophore and other muscles of the photo- phore. B. Proximal Reflector. The proximal reflector is cup-shaped with surfaces that appear to be portions of a sphere (Fig. 2). The reflector is an extracellular structure that consists of con- centric arrays of collagenous rods arranged into primary and secondary layers (Figs. 16, 17). There are about 12 primary layers, each YOUNG AND ARNOLD composed of 15-20 secondary layers. Each secondary layer consists of one layer of paral- lel rods. The parallel alignment of rods also exists between all secondary layers within a primary layer but usually they are not parallel to those of adjacent primary layers. The dif- ference in axial orientation of the rods be- tween primary layers varies randomly from near 0° to approximately 90°. A section at right angles to the rods reveals that each rod, ex- cepting those of the marginal layers, is sur- rounded by six equally spaced neighbors (Fig. 16). The rods lie in an extracellular medium and often appear to have a thin coating anda network of interconnecting filaments. The rods exhibit 53 nm periodicity in cross-band- ing which lies within the range of invertebrate collagen. The rods are circular in cross sec- tion and uniform in diameter (117 пт; SD = 6) in the central portion of the reflector, but at the periphery they are slightly larger in di- ameter (130 nm, SD = 7). The distance be- tween rods is generally affected by the fixa- tion and embedding process. The rods, in sections showing the greatest spacing, are separated by about 120 nm. Cells that secrete the collagen rods are lo- cated primarily adjacent to the periphery of the reflector (Fig. 17) although a few secretory cells can be found along the outer (convex) surface of the mirror. The secretory cells are characterized by an extensive granular endo- plasmic reticulum which frequently contains large cisternae filled with moderately electron dense material. C. Axial Complex. This complex occupies the center of the photophore between the photo- genic plexus and the distal cap (Fig. 2). The complex is formed of four distinct structures: the axial stack and the cells that form it, a zone of vesicles situated atop the stack, a core above the vesicles and a surrounding toroidal structure (Fig. 5). Axial stack. The axial stack consists of two adjacent multilayered structures, a proximal concave stack and a distal flat stack (Figs. 2, 4). Each stack contains approximately 50-60 electron dense layers (platelets) alternating —— FIG. 15. Distal region of the photogenic plexus. The mitochondrial cells dominate this region, although blood vessels are also common. The mitochondrial cells tangle between the blood vessels, crystalloid cells and nerves. The proximal reflector is on the left and the axial complex on the right. Scale bar is 5 um. FIG. 16. High magnification of the proximal reflector showing the banding pattern and spacing of the collagen rods. The consistent orientation of all rods within a primary layer is evident. Scale bar is 2 um. ABRALIA PHOTOPHORES 145 146 YOUNG AND ARNOLD A \ SAN er NS RE ANR ABRALIA PHOTOPHORES 147 with electron transparent layers (spaces). The exact shape of the proximal stack is difficult to determine with certainty. However, the con- centric layers apparently are portions of a sphere whose center lies at the center of the photogenic cone. Each electron dense layer consists of eight closely applied plasma mem- branes while each electron transparent layer is extracellular space (Figs. 18, 19). The axial stack is formed by approximately four stacked rings of surrounding cells (the axial cells) with a maximum of about 10-12 cells in each ring. Each platelet is formed by lamellar inter- digitations of the plasmalemma of at least several cells (Fig. 20). A single cell contrib- utes lamellae to many platelets. Each platelet is circular but is divided into segments. The membranes of each segment are separate from those of adjacent segments and, there- fore, do not extend the full width of the platelet (Fig. 19). Small knobs occur on the adjacent edges of the membranes from different cells and apparently contribute to some type of specialized junction although the details have not been resolved. The average thickneess of the electron dense layers is 79 nm (SD = 5, range = 72-91 nm). The thickness of the electron transparent layers is affected by fixa- tion and embedding and measurements are not meaningful. Vesicles. The vesicles, which occupy a zone between the axial stack and the core (Figs. 2, 4, 23) appear to be completely sepa- rate from any cells. They are large, closely FIG. 18. Edge of the distal region of the axial stack showing the typical membrane looping and the apparent continuity between extracellular spaces separating platelets (arrow). Each platelet can be seen to be formed of eight plasma membranes. Scale bar is 0.5 um. FIG. 19. Platelets in the axial stack of a developing photophore. The platelets are made of segments that abut against one another. The enlarged joints with easily visible membranes shrink in the mature photophore to approximately the same thickness as the platelets. Scale bar is 1 um. A FIG. 17. Section through the periphery of the proximal reflector showing the cells that secrete the collagen- ous rods. These cells are rich in rough endoplasmic reticulum (arrows). The collagen appears to be lain down parallel to the surface of the cells. Note the different orientation of collagen rods in different primary layers of the reflector. The muscle above the collagen-forming cells is one of the radial muscles that attaches to the torus. In the upper corner is one of the chromatophores that surrounds the posterior cup. Scale bar is 5 um. 148 YOUNG AND ARNOLD FIG. 20. Schematic drawing of how two cells might interdigitate to form segments of platelets, and how they produce the characteristic membrane looping at the edges of the platelets. packed, and sub-spherical in shape. The lumina of the vesicles are homogeneous and electron transparent. Lateral to the vesicles and situated between the axial cells and torus is a wedge-shaped ring of cells (Fig. 13). These cells extend to the vesicles but we have not found continuity between the two. However, since these cells may be responsi- ble for the formation of the vesicles, we tenta- tively name them the vesicle cells. A discon- tinuous sheath of collagen fibers covers the lateral surface of the torus and extends proxi- mally over the lateral boundaries of the vesi- cle cells (Figs. 13, 15, 23). Core. The core is a junctional area between the distal cap and the proximal cup (Fig. 23). Three cell types are present. A circlet of cells, the core cells, occupy much of the core and have their nuclei arranged in a ring at the base of the core where the latter surrounds the vesicles (Fig. 23). The lateral margins of the core cells intermesh in an irregular fashion with the cells of the torus. The second cell type, the junctional cells, have nuclei lying ina narrow, circular region (approximately one- fourth the diameter of the cap in the region of the girdle) that marks the separation between the core and the cap. Numerous cylindrical extensions from these cells (the roots) pene- trate the core cells (Fig. 23). Similar “roots” extend from cells in the distal cap between the junctional cells and into the core. The rela- tionship of the junctional cells to the cap cells is uncertain. We have been unable to confirm the presence of extensions of the former into the cap. In the distal portion of the core, the roots occupy over 60% of the cross-sectional area (Figs. 21, 24). Many of the roots pene- trate past the nuclei of the core cells to the lateral base of the core; others penetrate to and abut against the vesicles, and a few roots extend at oblique angles into the torus (Fig. 23). The roots carry a uniform array of rather evenly spaced microtubules although the density of microtubule packing sometimes dif- fers in different roots (Figs. 24, 25). Numerous fibrils extend between the microtubules. The roots have little else in internal structure and they appear quite electron transparent. The core cells also contain numerous micro- tubules and generally appear more electron dense than the roots (Figs. 24, 25). Torus. A torus containing membranous lamellae surrounds the core and the vesicles of the axial complex (Figs. 3, 21, 23). These lamellae are radially arranged around the optical axis of the photophore (Figs. 3, 21). The plane of each lamella intersects a hypo- thetical central axis. The lamellae are similar in Orientation to the radially arranged septa in a sea anemone. The lamellae are composed of electron dense layers of plasma mem- branes apparently formed by interdigitating membranes of adjacent cells as was found in the axial stack (Fig. 22). Large intracellular spaces, however, are lacking. The lamellae may be composed of just two membranes or as many as 40 to 50, and the spacing be- tween lamellae is highly variable. The lamel- lae are not completely flat or precisely orient- ed. The edges of the lamellae are irregular since the component membranes have differ- ent lengths. Nuclei of the cells that form the lamellae lie proximal and lateral to the lamel- lae. Muscle cells, presumably emanating from the large chromatophore at the base of the proximal cup, course radially over the surface of the cup to insert on the distal edge of the — FIG. 21. Cross-section through the torus and core showing irregular alignment of the lamellae. Scale bar is 10 ит. FIG. 22. High magnification of the medial end of a lamella from the torus showing its multimembran- ous structure and the characteristic membrane looping. Scale bar is 0.5 um. FIG. 23. Longitudinal section through the torus and core showing their relationship to the zone of vesicles and the distal cap. Nuclei of the junctional cells (n) are seen at the top of the core and one nucleus of a core cell (nc) is seen at the right of the vesicles. Scale bar is 10 um. ABRALIA PHOTOPHORES 149 YOUNG AND ARNOLD ABRALIA PHOTOPHORES 151 torus (Figs. 5, 17). Unlike most cephalopod muscle cells which have a solid core of mito- chondria (Fig. 30), these contain scattered mitochondria exhibiting no consistent posi- tional relationship to the myofilaments. 2) DISTAL CAP. The distal cap is a bell-shaped structure that can be divided into a central Oval region and a surrounding low bulbous girdle (Figs. 2, 26, 27). The size and shape of the cap varies somewhat between photo- phores, but the basic structure is constant. Nuclei are scattered throughout the cap and no lining membrane is present (Fig. 26). The most distinctive feature of the cap is the numerous long ribbons that appear to extend from its distal to its proximal margins (Fig. 26). The ribbons are formed of layers of mem- branes. The characteristic membrane looping at the edges of the ribbons indicates that they are formed, as in the case of the membrane platelets of the axial stack and the lamellae of the torus, by the interdigitating of plasmalem- ma from adjacent cells (Figs. 28, 29). Intracel- lular spaces between groups of membrane layers are virtually absent. In the central por- tion of the cap, the ribbons are narrow and thick, involving in some cases more than 100 membranes. In cross-section the ribbons in the central-most region of the cap have a somewhat random orientation and occupy about one-third of the area of the region (Figs. 23, 28). Peripheral to this region, the ribbons are broader and thinner and their orientation is more regular (Figs. 27, 29). In these ribbons there is a dominant trend toward an orienta- tion with the broad, flat surface parallel to the edge of the cap, and in cross-section they oc- cupy about one-fifth of the total area of the region. In the girdle the ribbons are fewer and thicker but very broad, and more regular still in paralleling the circumference of the cap (Fig. 27). They occupy 10-15% of the cross- sectional area of this region. The ribbons of the cap tilt toward the optical axis of the photophore with the angle of tilt increasing from the center to the periphery (Figs. 2, 26). The maximum angle of tilt is about 25-30° to << the optical axis. Although the tilt is not precise, the projected proximal ends of the ribbons in the central region would seem to converge at a point in the center of the axial stack. Nuclei are scattered throughout the cap. In the cen- tral region, the cytoplasm of the cap cells is packed with regularly arranged microtubules that run parallel to the long axis of the ribbons (Figs. 28, 29). In the girdle the microtubules are less regularly arranged and, while some have the same orientation as those of the central region, most lie at right angles to this direction and parallel to the circumference of the cap. Many muscles that originate on the body musculature surrounding the photophore course toward the distal cap and apparently spread over and attach to or near the surface of the cap (Fig. 31). However, we have not located specialized cell junctions that would define the exact areas of insertion. 3) DISTAL REFLECTOR. Almost completely surrounding the sides of the distal cap is an extensive lamellar structure similar in overall appearance to a series of typical cephalopod iridophores but differing in details of structure (Figs. 30, 31). Each iridophore bears numer- ous platelets separated by extra-cellular spaces. The plasmalemma of the iridophore forms both surfaces of a platelet which ex- tends outward from the nuclear region. The extracellular spaces lack structure and are electron transparent. The cytoplasm within a platelet is granular in appearance and is somewhat more electron dense near the cen- ter (Fig. 31). The thickness of the platelets varies by as much as 20-25% within a single platelet. The average thickness of the plate- lets is 136 nm (SD = 11). Each platelet is nearly formed into two layers by a single col- lapsed cisterna that extends almost the full length and width of the platelet (Fig. 31). Oc- casionally the cisterna is aligned with flat- tened vesicles that lie at a uniform distance from the plasmalemma in the region of the cell nucleus suggesting a stage in the formation of the cisterna. The platelets are imprecisely ar- FIG. 24. Cross-section through the core showing roots of the junctional cells and the axial-cap cells em- bedded within core cells. Numerous microtubules are present in all cell types. Scale bar is 1 ит. FIG. 25. High magification of core cross-section. The relationship of the cell membranes is evident. The numerous microtubules presumably have a structural function. Scale bar is 0.2 um. FIG. 26. Longitudinal section through the distal cap. The arrow indicates the optical axis of the photophore. The ribbons exhibit consider- able irregularity in alignment, and their tilt with respect to the optical axis increases laterally. Scale bar is 10 um. 192 YOUNG AND ARNOLD ABRALIA PHOTOPHORES 153 ranged into concentric rings around the dome of the cap. The flat surfaces of the platelets face the optical axis of the photophore and are generally tilted so that the proximal end of a platelet is somewhat closer to the optical axis than is the distal part (Fig. 5). The angle of tilt, however, is not consistent. Within a single longitudinal section through the re- flector, groups of platelets may vary in their angle of tilt from other groups by as much as 902 4) CHROMATOPHORES. Large chromato- phores cover the proximal cup (Figs. 2, 3, 17). Generally, one large chromatophore covers the apex of the proximal reflector and four overlapping chromatophores cover the sides. A series of chromatophores surrounds the sides of the distal cap. These chromato- phores, like those of the proximal cup, have a structure typical of cephalopod chromato- phores and are accompanied by large sheath cells (Fig. 30). A separate group of 4 or 5 small chromatophores, the apical chromato- phores, surround the distal tip of the cap (Fig. 5). The pigment granules of these chromato- phores are about Ya to 24 the size of the gran- ules in other chromatophores. Numerous muscles associated with the chromatophores encircle the photophore and some pass be- neath the girdle of the distal cap. 5) ORBITAL SPACE. Surrounding the proximal cup is a series of large apparently fluid-filled lacunae (Figs. 2, 3). These lacunae are all bound by thin membranes and lack associa- tion with muscle fibers. This space provides a thick cushion between the photophore and the body musculature. Observations on Fresh Photophores Observations have been made, with the aid of a dissecting microscope and lamp (artificial light), on the photophores of recently dead or dying A. trigonura. When a photophore is mechanically stimulated its muscles will fre- <<. quently contract. Repeated stimulation dem- onstrates that the photophore is capable of rocking in any direction. The maximum tilt of the optical axis was difficult to assess but ap- peared to be at least 45°. The photophore also could be rotated slightly on its optical axis. All chromatophores exhibited consider- able activity. At maximum expansion the cup chromatophores conceal the posterior cup except for a distal opening (i.e. the cap-core junction). This aperture is less than a third the diameter of the proximal portion of the photo- phore. The apical chromatophores were yel- low-brown in color which contrasted with the brown color of the other chromatophores. The former could converge (i.e. shift position) over the center of the distal cap and expand. If all chromatophores expanded at once the photo- phore would be completely concealed. Transmitted artificial light when viewed from the side of the rather firm but colorless distal cap, exhibits little distortion. Consider- able scattering of the light, however, is appar- ent from a top view (i.e. down the optical axis of the photophore). The axial stack is easily seen through the microscope when the distal Cap is dissected away. Artificial light incident on the distal end of the stack is generally re- flected as a small brilliant yellow disc. This color could result from the reflection of all but the blue component of the incident light. This reflected light swamps any transmitted light that may have bounced off the posterior re- flector. With the distal cap in place the reflec- tion off the stack is diffused as it passes through the cap. The degree of diffusion (scattering) varies between photophores and occasionally little if any diffusion of the re- flected light is seen. The proximal reflector is difficult to observe directly. Because of the spherical geometry of the reflector, most artificial light directed into the photophore is not reflected directly back. In addition the optical inhomogeneities of the distal cap (and presumably the torus) make clear viewing of the reflector difficult. How- ever, judging by the color, two sources of light FIG. 27. Cross-section of the distal cap showing the inner and outer areas of the central region and the surrounding girdle. The cross indicates the optical axis of the photophore. The differences in alignment, thickness and width of the multimembranous ribbons in the three region are apparent. Scale bar is 10 um. FIG. 28. Cross-section through inner area of the central region of the distal cap. Note thickness of multi- membranous ribbons, and the membrane looping at their edges. The numerous microtubules presumably have a structural function. Scale bar is 1 um. FIG. 29. Cross-section through outer area of the central region of the distal cap. The multimembranous ribbons are thinner and longer than those in Fig. 28; however, the pattern of membrane looping is similar. Scale bar is 1 um. YOUNG AND ARNOLD ABRALIA PHOTOPHORES 155 emerging from the photophore can be attrib- uted to reflection of artificial light off the pos- terior reflector: (1) a diffuse reflection gener- ally is observed from the photophore periph- eral to the reflection from the axial stack; (2) a series of brilliant tiny points of reflected arti- ficial light arises lateral to the diffuse reflec- tion. Light from these two regions gives the photophore its predominant color. With the distal cap dissected away bright points of light can be observed to be directly reflected off the posterior reflector. The color of this light is the same as that from the two previously men- tioned sources. This color is either blue or green, depending on the particular photo- phore. Occasionally when the cup chromato- phores are strongly retracted and the photo- phore is turned over, the outer (convex) sur- face of the posterior reflector is visible. As would be expected, artificial light reflected off this convex surface is the same color as that reflected off the concave surface. Generally, most Type C photophores under artificial light appeared predominately green in freshly captured animals although some- times a mixture of blue and green photo- phores was seen. In a few squid most Type C photophores were blue. On one occasion the color of the reflected artificial light was ob- served to change with time in a dead squid. After two days in 10°C seawater photophores that were initially green had become a dark blue, but, as the water warmed under micro- scope lights, the reflection became green again. The water was cooled to 0°C but the Type C photophores remained green. How- ever, as the water warmed slightly the photo- phores first turned blue and then returned to green as the water warmed further. Not all the Type С photophores exhibited these dramatic color changes, and a repeat of the tempera- ture changes had a negligible effect on the color. The predominant color of blue or green was due to light reflected off the proximal re- flector. Changes in the reflection off the axial stack shifted from pale yellow (blue photo- phores) to yellow-orange (green photo- phores) or, in photophores that did not make the full change, from the blue type to the green type, to lavender. In spite of the length -—— of time the animal had been dead, the skin had not turned opaque and chromatophore and photophore muscles were still responsive to mechanical stimulation. The distal reflectors were almost impos- sible to see under the microscope when view- ing directly into the photophore although oc- casionally some golden iridescence was ob- served. However, a small fibre-optic light guide (~60 um diam.) with one end placed atop the distal cap directed light into the organ and produced a bright blue-green oblique re- flection off the distal reflector. When the photophore was tilted and light struck the re- flector at near normal angles (roughly 60-70°), a red-orange reflection was ob- served. The endowment of distal iridophores varied between individual photophores. In addition, the arrangement of these irido- phores varied. In some photophores they closely surrounded the photophore and were covered laterally by large chromatophores. In other photophores, often adjacent to the first type, the iridophores were spread over a broad region half again the diameter of the proximal cup. Mechanical irritation failed to induce one type to transform into the other. Optical effects of the torus could not be detected. Observations on Bioluminescence 1. Intensity regulation. Young & Roper (1977) demonstrated that this squid could ef- fectively counterilluminate over at least 20- fold change in light level. We have measured the emission spectrum of the luminescence as it increased in intensity by 325-fold in re- sponse to changes in the overhead light. Young et al. (1980) found that a closely re- lated species, Abraliopsis sp. B, could adjust its bioluminescence during counterillumina- tion over a range of at least 15,000-fold. We would expect to find the same capability in A. trigonura if tested in a similar manner. 2. Color regulation. Young & Mencher (1980) demonstrated that the color of light produced during counterillumination under simulated day conditions (i.e. cold water en- vironment of 6-8°C) was a narrow unimodal FIG. 30. Cross-section of the distal region of the photophore showing the central region of the distal cap, a chromatophore with sheath cells, numerous large muscles which presumably orient the photophore, and the distal reflector. Scale bar is 10 um. FIG. 31. Platelets of the distal reflector. Collapsed cisternae form the distinctive double membrane in the midline of the platelet. The matrix of the platelet is heterogeneous. The platelets have a uniform thickness but imprecise alignment. Scale bar is 1 um. 156 YOUNG AND ARNOLD PERCENT OF MAXIMUM INTENSITY 500 GREEN 550 WAVELENGTH, nm YELLOW 600 FIG. 32. Luminescent emission spectra taken during counterillumination from the ventral surface of the head of A. trigonura. The overhead light which stimulated counterillumination was held at a constant intensity throughout the experiment. The system for detecting bioluminscence covers a large region of the ventral surface of the head and, therefore, it can receive light from all three types of photophores. The temperature of the water surrounding the squid was adjusted to a different temperature for each measurement. Curve A: water temp. 6-8°C; Curve B: water temp. 11-12°C; Curve C: water temp. 14-15”C; Curve D: water temp. 19-20°C; Curve E: water temp. 23°C. Curve A presumably represents the luminescent spectrum the squid would produce in its day habitat and Curve E the spectrum in its night habitat. Each curve represents six measurements. The two peaks of the night curve form in different ways. The shift of the peak at A to the right indicates that one set of photophores is changing the color of light it emits. The peak on the left gradually increases indicating a new set of photophores is on and increasing in intensity. band with peak emission at about 480 nm (full width at half maximum (FWHM) of 33 nm). Under simulated night condition (warm water environment of 23-25°C) a bimodal peak ap- peared with peaks at 440 nm (FWHM-55 nm) and 536nm (FWHM-46 nm). Under night thermal-conditions direct visual observations on a brilliantly lit squid revealed a mix of photophores of different brilliance suggesting that more than one type of photophore was involved. However, under day thermal-condi- tions at the same intensity of overhead light all photophores had a uniform brilliance suggest- ing a single photophore type was involved. Examination of the emission spectra during a gradual change between day and night tem- perature conditions confirmed that two sets of photophores are involved in producing the night colors. During this changeover, one of the night color modes (peak at 440 nm) ap- peared and gradually increased in height while the other night color mode (peak at 536 nm) developed from a gradual shift of the day mode to longer wavelengths (Fig. 32). Apparently, the mode at 440 nm is produced by a set of photophores turning on for the first time, while the mode at 536 nm is produced by the same set of photophores that produces the day peak at 480 nm. That is, the latter photophores can change the color of light they produce. The large number of active photophores observed during counterillumi- nation eliminates the relatively few Type A photophores as the source of the day color. Under the microscope unfixed Type B photo- phores reflected a blue-violet color indicating these were the source of the 440 nm peak. The Туре С photophores remain as the source of the 480 nm and 536 nm peaks. Additional color shifts have been measured under both day and night conditions at high light intensities (Young & Mencher, 1980). At the day temperatures the peak broadens slightly and at the night temperatures the 536 nm peak shifts back toward the blue re- gion as light intensity increases and combines with the 440 nm peak to form a complex curve (Fig. 33) ABRALIA PHOTOPHORES 157 100 LE 80 60 40 PERCENT OF MAXIMUM INTENSITY 20 400 VIOLET 450 BLUE N N De eee N== ES : --— 205 = 500 GREEN 550 YELLOW 600 WAVELENGTH, nm FIG. 33. Luminescent emission spectra taken during counterillumination from the ventral surface of the head of A. trigonura. Temperature was held constant at 23°C during the entire experiment. Beginning with curve A, which is curve E in Fig. 32, each subsequent curve was measured after increasing the overhead light. Each curve is the average of 1-6 measurements. Curve A: overhead light intensity approximately 2.5 x 10-3 uW/cm?; Curve В: overhead light intensity increased by 6 times; Curve С: overhead light intensity further increased 3.5 times; Curve D: overhead light intensity further increased 4.4 times. Note that the peak of A shifts to shorter wavelengths as the light intensity increases. 3.Angular distribution. We have no precise data on the angular distribution of the light except for visual observations that indicate it is highly directional (directed ventrally) when the squid is in cold water and somewhat less so in warm water in our experimental appa- ratus. DISCUSSION The site of light production is, almost cer- tainly, the crystalloids of the photogenic cone. Thus the crystalloid cells appear to be the photocytes. The cone's strategic optical loca- tion, its crystalloid nature and its close asso- ciation with blood vessels strongly support this assumption: 1) The cone sits at the focus of the spherical proximal reflector, and is cen- trally located with respect to the axial stack and the optical axis of the photophore. 2) The presence of crystalloids has been reported in the presumed photogenic cells of three other cephalopods (Bathothauma lyromma, Dilly 8 Herring, 1974; Watasenia scintillans, Okada, 1966; Enoploteuthis sp., Young, 1977). 3) In- tense vascularization is characteristic of all known photogenic tissue in cephalopods (e.g., Girsch et al., 1976; Arnold et al., 1974; Dilly & Herring, 1974; Okada, 1966). We do not know the chemical composition of the crystalloids. Similar crystalloids in the photophores of Watasenia scintillans are thought to be proteinaceous (Okada et al., 1934). We suspect the crystalloids of being composed predominantly of luciferase while luciferin, in some form, is constantly supplied to the crystalloids via the circulatory system. Young et al. (1979) found evidence for the presence of luciferin in the blood of the squid Symplectoteuthis oualaniensis. The intimate association between photocytes and blood vessels suggests that a component of lumi- nescent reaction, possibly luciferin, is sup- plied by the circulatory system. Browning (1979) suggests that in Octopus transport of molecules from the blood of less than 120A in diameter occurs through pericyte junctions. Luciferin extracted from the squid Watasenia is a small molecule with a molecular weight of 572 (Inoue et al., 1976). In Abralia the finger- like extensions of the crystalloid cells that sur- round the blood vessels could provide a large surface area for acquiring such molecules. 158 YOUNG AND ARNOLD We view the presence of dense arrays of microtubules within the dendritic photocytes as evidence for the intracellular transport of materials between the blood vessels and the crystalloids. While microtubules are generally considered to be cytoskeletal elements, they have also been implicated in the transport of cell structures. The presence of immature crystalloid cells near the proximal end of the photogenic cone indicates that crystalloids are continuously added to the stack as the photophore grows. The extracellular spaces and channels and/or the nearby tubular ER may function as a re- servoir for some component of the lumines- cent reaction. The scarcity of mitochondria in the crystal- loid cells seems inconsistent with their pre- sumed high metabolic activity. Presumably ATP is transported to the photocytes from the intertwining mitochondrial cells. Arnold et al. (1964) suggest a similar mechanism in the ocular photophores of the squid Pterygioteu- this microlampas. Two different types of iridophores are pres- ent in the Type C photophores: those of the axial stack and those of the distal reflector. Both are distinctly different in their structure from any previously described in cephalopods (i.e., Kawaguti & Ohgishi, 1962; Arnold, 1967; Mirow, 1972; Arnold et al., 1974; Young, 1977: Brocco & Cloney, 1980). A third inter- ference structure, the proximal reflector, is present; however, it cannot be properly termed an iridophore since the reflecting ap- paratus is entirely extracellular. To predict the reflectance and transmit- tance characteristics of iridophores, the re- fractive indices and thicknesses of the high (platelets) and low (spaces) refractive index layers must be known (Land, 1972). The re- fractive index of the spaces can be assumed to be that of tissue fluid, 1.33, as measured by Brocco & Cloney (1980) in Octopus. The re- fractive indices of the platelets in iridophores of Abralia are unknown. Brocco (see Brocco & Cloney, 1980), however, measured the re- fractive index of the proteinaceous platelets of Octopus iridophores as 1.42. Denton & Land (1971) found a value of 1.56 for platelets in squid and cuttlefish; however, Brocco & Cloney (1980) suggest the latter value may be high due to the method of measurement. The structure and composition of the platelets in Abralia differ from those measured. However, for lack of more accurate data, we assume a refractive index of 1.49 (i.e. midway between measured values) for the high refractive index structures in the iridophores. The axial stack could function either as a filter or a mirror. Direct observations of the reflection of artificial light off the distal end of the axial stack in fresh photophores suggest transmission of blue to green light (i.e. yellow or yellow-orange reflection). If the stack were to function in reflecting all bioluminescent light, it would reflect blue-green light or, per- haps, a much broader band (i.e. silvery reflec- tion). The axial stack therefore, selectively transmits light and functions as a filter. In Abralia the narrow transmission band meas- ured under day conditions indicates that the filter reflects light to either side of the band pass back toward the proximal reflector and in this manner controls the transmission band- width. The refractive index of the plates affects not only the wavelength of maximum reflection but also the reflection bandwidth (Land, 1972). For a refractive index of 1.49 and a large number of platelets, the bandwidth would be about 40 nm and at a lower refrac- tive index the width would be even narrower. (Smoothed bandwidths can be computed fol- lowing the procedure in Huxley, 1968: 241, except that for equation 28 the following is substituted: |R|? = 1 — V1 — 1/K°. This equa- tion is the counterpart of Huxley's equation 40 but applies to non-ideal stacks.) If the axial stack were formed of two uniform functional units, two reflection bandwidths of 40 nm each would be inadequate, judging from the measured emission spectra. Apparently, the slight variation in the thickness of the plates (and perhaps the spaces) functions to broad- en the reflection bandwidth. The exact re- flectance of the platelets is impossible to calculate since the widths of the spaces be- tween plates are highly variable, apparently due to fixation effects. However, for an ideal Va À stack the plate thickness predicts a re- flectance maximum at 471 nm [i.e. Amax = 2(па4а + Npdp), where nan) are the refractive indices of the light and dense layers in the stack respectively and d,d, are the thick- nesses of these layers (Land, 1972)], and a range of 429 to 542 nm based on the meas- ured range of thicknesses. Although we have reason to suspect that the stack cannot be ideal (see below), the calculations suggest that the stack is close to ideal. The iridophores of the distal reflector have thicker platelets with somewhat irregular spacings. If we consider this an ideal stack, ABRALIA PHOTOPHORES 159 maximum reflectance at normal angles of in- cidence to the broad surface of the platelet would be 811 nm. This value would shift to longer wavelengths for a non-ideal stack (al- though a second order peak at 2 Amax would occur in the visible region in the latter case [Land, 1972]). The iridophores, however, seem to function in reflecting oblique light. We roughly estimate that light from extreme spherical aberration off the posterior reflector or light exiting between the posterior reflector and the distal cap will arrive in the region of this reflector at angles of 30° to 70° to the optical axis of the photophore. If we consider the tilt of the platelets to be 7” to 17” to the optical axis (angles which would cause re- flected light to depart the region at half its ori- ginal angle to the optical axis) then the angles of incidence would be about 52” to 65” and reflectance maxima would vary between about 410 and 535 nm. For a non-ideal stack in which the spaces are wider than the plate- lets, these values would shift to longer wave- lengths. These values agree well with direct observations. The third interference system (the proximal reflector) is based on arrays of rods rather than layers of platelets. Unfortunately, a mathematical treatment of the properties of rod-based systems, apparently, has not been published (Land, 1972). However, Land (1972) suggests that the reflectance maxi- mum for normally incident light can be ap- proximated by “A = 2 nd, nd, 24 nd etc., where n is the mean refractive index of the structure and d is the separation of the centers of the rods in one plane to those in the next.” The latter measurement in our material, unfortu- nately, is unreliable due to fixation effects. Using the maximum distances measured, and assuming an ideal system, we calculate a pri- mary maximum at 680 nm and a secondary maximum at 340. At greater rod separation the secondary maximum would move into the visible region. A spherical interference reflector requires a complex construction as light of a given wave- length will strike the reflector at normal angles along the optical axis of the photophore but will strike at increasingly oblique angles lat- erally. Perhaps the slight lateral gradient of increasing rod diameter partially compen- sates for the different angles of incidence. Other than this lateral gradient, the rods are quite uniform in diameter. This uniformity indi- cates that the mirror acts as a narrow-band reflector. Direct observations with reflected artificial light confirm the narrow-band proper- ties. In contrast, the rod-based reflectors of the cat tapetum and of certain bird feathers that are broad-band reflectors contain rods of various sizes (Land, 1972). The silvery reflec- tor of certain cephalic photophores of the re- lated squid Abraliopsis contain rods that also vary greatly in size (personal observation). In order to understand the effects of other optical structures in the photophore we de- termined the approximate optical pathways within the photophore. Because of the large size of the photogenic cone relative to the posterior reflector, light cannot be considered to arise from a single point. Instead, we con- sidered an arbitrary array of 46 points evenly distributed throughout the cone to represent the light source. Also one cannot consider re- flection off the posterior reflector as arising at its front surface. Reflectance off an interfer- ence structure is exponentially related to the number of platelets within the reflector (Land, 1972). We divided the reflector into 15 equal layers which were considered to function as individual platelets in an ideal stack. The con- tribution of each layer was then determined according to the procedure in Land (1972: 82). In using this approach, we assumed that the entire thickness of the reflector was re- quired to produce 100% reflectance. If we ignore the optical effects of other accessory structures, representative optical pathways would appear as in Fig. 34, and the calculated angular distribution of the light would appear аъ in Fig. 35: The angular distribution of light from the photophore differs slightly from that of day- light in the habitat of the squid (Fig. 35). If the radiance pattern from a single photophore is to match that of daylight or nightlight (this is uncertain since it is the combined output of all photophores that is important), then acces- sory structures must compensate for the dif- ferences. The torus, the distal cap, the distal reflector and the chromatophores appear to be such structures. The presumed optically functional com- ponents of both the torus and distal cap are similar in structure to those of the axial stack. However, the thickness of their layers is high- ly variable and is generally too thick and ir- regular for interference effects. These struc- tures, which presumably act as “thick-film” rather than “thin-film” reflectors, should be ef- fective if the incident light is strongly oblique and if multiple reflections are allowed. The tilting ribbons of the distal cap will tend 160 YOUNG AND ARNOLD tn | EX Al SÁ EN = <= >>, a OES FIG. 34. Ray diagram showing representative light rays emerging from the center of the photogenic cone. Compare to Fig. 5. Heaviest straight lines represent ribbons of the distal cap. The effects of chromatophores, cap ribbons and distal iridophores on the ray paths are ignored. Note that most lateral- ly reflected rays pass through the torus. Note angles of incidence of rays on cap ribbons and dis- tal reflector, and the potential effect of chromato- phore movement. The cup chromatophores are drawn in a partially expanded position. to deflect, away from the optical axis, rays emerging from the posterior cup at angles less than roughly 15° to the optical axis (ac- cording to Fresnel's equations for reflectance) (see Fig. 35). The result is to spread some- what the directional beam arising from the photophore. Nearly all aberrant rays reflected from the spherical posterior reflector pass through the torus where they may encounter the vertical membrane lamellae. The surfaces of the lamellae will scatter light at low angles of inci- dence but would appear to affect azimuth more than the declination of the resulting rays. This could in turn affect the angle of in- cidence on the more precisely aligned ribbons at the periphery of the cap. Unfortunately, the imprecise geometry of both the torus and the cap make it impossible to determine exactly what the effect of the torus is. At any rate we consider that the torus scatters light to in- crease, in some manner the proportion of rays striking the cap ribbons at angles sufficiently high to allow reflection. In addition, a consid- erable portion of the light must be intercepted by the ends of the lamellar stacks of the torus FIG. 35. Polar diagrams. Each diagram represents a vector diagram in which the tips of all the vectors are joined to form a curve. The length of each vector represents light intensity and the angle of the vector represents the direction of the light. The dashed curve represents the angular distribution of daylight in the sea based on Denton et al., 1972. The large dot represents the photophore. The heavy solid curve represents the angular distribu- tion of light emitted from the photophore if the ef- fects of the torus, distal cap, distal iridophores are ignored and if chromatophores are retracted. The thin curve represents the effect on the previous curve if the cup chromatophores are expanded. Point A represents the effect of the edge of the posterior reflector on the emission beam and point B represents the effect of the edge of the expanded cup chromatophores. and cap. The fate of light guided within the stacks is uncertain due to their complicated terminations but the net effect must be to in- crease the scattering of the light. While we describe the functions of the torus and distal cap as those of a weak diffuser, the organiza- tion of these structures indicates that this is an oversimplification. ABRALIA PHOTOPHORES 161 The distal reflector, due to its slope to the optical axis, will intercept and redirect outward some of the directly emitted light passing be- tween the central portion of the cap and the proximal reflector when the posterior cup chromatophores are retracted (Fig. 35). Thus far this discussion has considered a more or less static photophore. We know, however, that both intensity and color are capable of being regulated and circumstantial evidence indicates that the angular distribu- tion of emitted light can also be varied. The enormous range over which light intensity can be regulated, combined with the requirements for adjusting intensity independent of the angular distribution of the light eliminates the chromatophores as the primary regulators of intensity. The presence of numerous nerves and occasional synapses associated with the photocytes argue for intensity control via di- rect nervous regulation. The Type C photophores appear to be re- sponsible for the measured shift in the “day” blue luminescence peak at 480 nm to the “night” green peak at 536 nm. This ability to change color is confirmed by visual observa- tions which show that colors reflected from artificial light in these photophores can either be blue or green in fresh animals and can exhibit reversible post mortem shifts between these colors. The axial-stack filter and the posterior reflector probably control the meas- ured color shift: the properties of both can be altered and all emitted light encounters one or both of these structures. Neither the axial stack nor the proximal reflector alone could account for the color shift. The effects of one cannot be physically separated from those of the other since much of the light that en- counters one or the other of these structures is intermingled in the distal cap. It is unlikely that changes at the site of the luminescent reaction contribute significantly to this shift. The inability of the squid to main- tain the green color at high light intensity sug- gests that the filters and reflectors in the “green mode” are selecting light from the tail of a fixed chemical emission spectrum. The simplest system for altering the trans- mission characteristics of the axial stack would involve a change in the spacing be- tween platelets. A change in the thickness of the spaces by about 20% will shift the reflec- tance maximum by about 55 nm. This could be accomplished by moving fluid from one portion of the stack to another. Where fluid is withdrawn the reflectance maximum would shift to shorter wavelengths. Where it is added the opposite would occur and a new bandpass would arise in-between. Continuity between some spaces Clearly exists (Fig. 13). The mechanism that would control fluid move- ment, however, is obscure. The presence of nerves adjacent to the axial cells (Fig. 8) indi- cates a potential means of regulation. This mechanism is not applicable to the posterior reflector which is an array of col- lagen rods suspended in a single fluid-filled extracellular space. In this case the change must occur in the thickness of the high, rather than the low, refractive index layers (the other option, an alteration in refractive index, would have only a minor effect due to the low range of refractive indices found in biological mate- rials). We suspect the mechanism involves changes in the hydration states of the col- lagen rods and/or simple osmotic swelling. Under severe conditions collagen can swell over threefold in diameter (Gustavson, 1956). In the photophore approximately a 20-30% increase in diameter presumably would be adequate. Temperature, salts and pH strongly affect one or both mechanisms. Resulting changes in refractive index and spacing would have an opposite but minor effect. Al- though reflectance changes were correlated with temperature changes in the dead animal, the color changes in life are not a passive response to temperature changes since in the living animal the color can shift from green to blue at constant high temperatures under the stress of high intensities of overhead light (Fig. 26). Apparently, the change is under be- havioral control. Perhaps nervous or hormon- al messages to nearby cells trigger alteration of pH, salts or some other factor in the fluid bath which in turn controls the hydration or swelling of the rods. The angular distribution of the lumines- cence apparently can also be regulated. The expansion state of chromatophores surround- ing the photophore must have strong influ- ence on the angular distribution of the emitted light (see Figs. 35, 36). In addition, between individual Type C photophores considerable variation is seen in the straightness of the rib- bons of the distal cap and some variation oc- curs in the shape of the cap. Large variations can be observed between adjacent Type C photophores in the spread of the distal reflec- tor. Numerous muscles lie in the region of these structures. Muscles attach near the dis- tal surface of the torus that presumably allow this structure to be altered. The nearby vesi- 162 YOUNG AND ARNOLD cles which easily distort may facilitate this movement. Thus the chromatophores, the distal lens, the distal reflector and the torus all have the potential of being altered. Unfortu- nately, the demonstration of such effects will be difficult. In addition, as the animal's attitude in the water changes, so must the radiance pattern of the luminescence with respect to the body axis. Presumably the lacunae of the orbital space and the numerous muscles that tilt the photophore enable its proper orienta- tion regardless of the squid's attitude. Although we are only beginning to under- stand the complex functioning of this minute photophore, we can speculate on its method of operation. The general features of which are summarized as follows. Light is produced by the photogenic cone which is under neu- ronal control. Luciferin, or other key com- ponents of the luminescent reaction, is sup- plied through the blood and collected by numerous extensions of the photocytes. The proximal reflector and the axial stack are in- terference structures that determine the color of the emitted light. Complex interaction be- tween chromatophores, proximal and distal reflectors and the torus and distal cap deter- mine the angular distribution of the emitted light. The light intensity, color and probably the angular distribution of the light can be regulated. Because of this flexibility, we ex- pect that these photophores, in combination with other types, allow the squid to conceal itself with bioluminescence under a variety of enviromental lighting conditions that are found in the squid’s open ocean habitat. ACKNOWLEDGMENTS We thank Patty O'Bryen and Thecla Ben- nett for preparing the tissues for electron mi- croscopy, and Peter Herring, Michael Land, Steven Brocco and John Shears for their helpful review of the manuscript. The senior author thanks S. Brocco for his suggestions on fixation techniques and his valuable com- ments on many aspects of this work. We thank Eric Hochberg for his shipboard assist- ance in obtaining the emission spectra of the luminescence and Frederic Tsuji for discus- sions on the biochemical implications of our findings. Huxley's equation 38 (see Huxley, 1968) was integrated by J. B. Nation. Finally, we thank the captain and crew of the R/V KANA KEOKI for their help in cooperation on the numerous cruises during which data were collected and photophores fixed and Sher- wood Maynard for organizing and supervising the cruises. This material is based on re- search supported by the National Science Foundation under grant Nos. OCE 7681030 and OCE 7825342 to the senior author. LITERATURE CITED ARNOLD, J., 1967, Organellogenesis of the cepha- lopod iridophore: cytomembranes in develop- ment. Journal of Ultrastructure Research, 20: 410-421. ARNOLD, J. М. & YOUNG, В. E., 1974, Ultrastruc- ture of a cephalopod photophore. |. Structure of the photogenic tissue. Biological Bulletin, 147: 507-521. ARNOLD, M. J., YOUNG, R. E. & KING, M. U., 1974, Ultrastructure of a cephalopod photo- phore. Il. Iridophores as reflectors and transmit- ters. Biological Bulletin, 147: 522-534. BROCCO, $. |. & CLONEY, В. A., 1980, Reflector cells in the skin of Octopus dofleini. Cell and Tissue Research, 205: 167-186. BROWNING, J., 1979, Octopus microvasculature: permeability to ferritin and carbon. Tissue and Cell, 11: 371-383. DENTON, E. J. & LAND, F. M., 1971, Mechanism of reflexion in silvery layers of fish and cephalo- pods. Proceedings of the Royal Society of Lon- don, 178: 43-61. DENTON, E. J., GILPIN-BROWN, J. B. & WRIGHT, P. G., 1972, The angular distribution of the light produced by some mesopelagic fish in relation to their camouflage. Proceedings of the Royal So- ciety of London, ser. B, 182: 145-158. DILLY, P. N. & HERRING, P. J., 1974, The ocular light organ of Bathothauma lyromma (Mollusca: Cephalopoda). Journal of Zoology, 172: 81-100. GIRSCH, W., HERRING, P. & MCCAPRA, F., 1976, Structure and preliminary biochemical characterization of the bioluminescent system of Ommastrephes pteropus (Steenstrup) (Mol- lusca: Cephalopoda). Journal of the Marine Bio- logical Association of the United Kingdom, 56: 707-722. GUSTAVSON, K. H., 1956, The Chemistry and Reactivity of Collagen. Academic Press, New York, 342 p. HOYLE, W. E., 1894, On the luminous organs of Cephalopoda. Report of the 63rd Meeting of the British Association for the Advancement of Sci- ence held at Nottingham (1893), p. 802-803. HUXLEY, A. F., 1968, A theoretical treatment of the reflexion of light by multilayer structure. Journal of Experimental Biology, 48: 227-245. INOUE, S., KAKOI, H & GOTO, T., 1976, Squid bioluminescence. III. Isolation and structure of ABRALIA PHOTOPHORES 163 Watasenia luciferin. Tetrahedron Letters, 34: 2971-2974. JOUBIN, L., 1895, Note sur les appareils photo- genes cutanees de deux Cephalopodes: Histi- opsis atlantica Hoyle at Abralia owenii (Verany) Hoyle. Memoires de la Societe Zoologique de France, 8: 212-228. KAWAGUTI, S. & OHGISHI, S., 1962, Electron mi- croscopic study on iridophores of a cuttlefish, Sepia esculenta. Biological Journal of Okayama University, 8: 115-129. LAND, M. R., 1972, The physics and biology of animal reflectors. Progress in Biophysics and Molecular Biology, 24: 77-106. MIROW, S., 1972, Skin color in the squids Loligo реа and Loligo opalescens. Il. Iridophores. Zeitschrift für Zellforschung und Mikroskopische Anatomie, 125: 176-190. MORTARA, S., 1922, Gli organi fotogeni di Abralia veranyi. R. Comitato Talassografico Italiano, 95: 1-18. OKADA, Y. K., 1966, Electron microscopy of Watasenia. In: Bioluminescence in Progress, p. 613-629, Princeton University Press, Princeton. OKADA, Y. K., TAKAGI, S. & SUGINO, H., 1934, Microchemical studies on the so-called photo- genic granules of Watasenia scintillans (Berry). Proceedings of the Imperial Academy of Japan, 10: 431-434. ROSS, K. F. A., 1967, Phase contrast and inter- ference microscopy for cell biologists. New York: St. Martin’s Press. YOUNG, R. E., 1977, Ventral bioluminescent coun- tershading in midwater cephalopods. Symposia of the Zoological Society of London, no. 38, 161-190. YOUNG, R. E., 1978, Vertical distribution and photosensitive vesicles of pelagic cephalopods from Hawaiian waters. Fishery Bulletin, 76: 583- 615. YOUNG, В. E., KAMPA, E. M., MAYNARD, $. D. MENCHER, F. M. & ROPER, C. F. E., 1980, Counterillumination and the upper depth limits of midwater animals. Deep Sea Research, 27A: 671-691. YOUNG, R. E. & MENCHER, F. M., 1980, Biolu- minescence in mesopelagic squid: Diel color change during counterillumination. Science, 208: 1286-1288. YOUNG, R. E. & ROPER, C. F. E., 1977, Intensity regulation of bioluminescence during counter- shading in living midwater animals. Fishery Bul- letin, 75: 239-252. YOUNG, R. E., ROPER, C. F. E., MANGOLD, K., LEISMAN, С. & HOCHBERG, Е. G., Jr., 1979, Luminescence from non-bioluminescent tissues in oceanic cephalopods. Marine Biology, 53: 69- U the MALACOLOGIA, 1982, 23(1): 165-175 DEVELOPMENTAL ASPECTS OF THE MANTLE COMPLEX IN COLEOID CEPHALOPODS S. v. Boletzky C.N.R.S., Laboratoire Arago, F-66650 Banyuls-sur-Mer, France ABSTRACT The embryonic development of the mantle complex in some decapod cephalopods is briefly described. Emphasis is placed on the relationship between the muscular fins and the shell sac. In the Decapoda (Sepioidea and Teuthoidea), the base of the fins tends to become detached from the actual shell complex by forming a pair of basal pouches that may completely separate from the shell sac. The embryonic development of the mantle-fin-shell complex in forms having an extremely reduced shell demonstrates the morphogenetic interdependence of the three components. Key words: Cephalopoda; mantle; fins; shell sac; development; morphology. INTRODUCTION When considering the form and function of the mantle in cephalopods the question arises as to how and when the particular form is achieved that corresponds to a particular function. The cephalopod mantle together with the funnel and collar acts as a pump; water drawn into the mantle cavity through the peripheral slits between the funnel pouch (or collar) and the mantle margin is expelled through the funnel tube by muscular contrac- tion of the mantle. This concentration and the resulting water jet can be more or less vigor- ous. The water is expelled gently during the regular water exchange in the respiratory cycle. More vigorous expulsion always has a locomotory effect; this is used for the typical jet propulsion, especially in attack and escape movements. The “dynamic lift” provided by rapid respiratory jets is characteristic of “hov- ering” in small animals. In the evolution of the “modern” coleoid cephalopods, the muscular mantle has largely replaced the protective shell, which has be- come an internal, residual “backbone.” Along with this replacement, the mantle has been equipped with a special device for stabiliza- tion and dynamic lift in hovering and slow lo- comotion, namely the muscular fins. In this paper, | briefly describe and discuss the early embryonic development of the com- plex comprising the muscular mantle, the fins and the shell sac. The cephalopods are characterized by an essentially direct development. The hatch- lings already have the complete mantle com- plex of the adult type. During the later stages of embryonic development, mantle contrac- tions become increasingly frequent and may serve respiration even before hatching, al- though to a limited extent. The general process of shell sac formation in cephalopod embryos has been described by Koelliker (1844), Naef (1928), and Spiess (1971). In all the decapods, a circular fold forms around the central part of the mantle rudiment at about stage VIII of Naef. This fold grows over the central area that is surrounded by the ring- or clasp-shaped rudiment of the muscular mantle; during this process the rudi- ments of the fins appear on each side of the now closing pore of the shell sac. Their bases are attached to the outer wall of this sac. In the octopods, shell sac formation begins earlier than in the decapods, at least in the Incirrata. The rudimentary shell sac is largely modified in its form, but not in its relation to the fin rudiments which appear even in the embryos of the finless Incirrata but do not continue to develop (Naef, 1928). A discus- sion of the octopodan relationships is pre- sented in an earlier article (Boletzky, 1978). The present description is limited to the co- ordinated processes of the mantle-fin-shell sac formation in decapods, especially the process by which the fins may detach from the actual shell sac. This phenomenon has briefly been mentioned by Naef (1923, 1928), but has not yet been described in detail. In a recent study of shell sac formation (Spiess, 1971), it is entirely ignored. (165) 166 BOLETZKY MATERIALS AND METHODS In addition to live observations of various embryonic stages, fixed embryos of different species were analysed. For scanning electron microscopy (SEM), embryos of Loligo vul- garis Lamarck, 1798 were fixed in buffered glutaraldehyde and post-fixed in OsO,, critical point dried and coated with gold-palladium. A series of ethanol-fixed embryos of Sepioteuthis lessoniana Lesson, 1830 from the Red Sea were imbedded in paraffin and sectioned. The sections were stained with Azan or Masson's trichrome. Embryos of Sepia officinalis Linnaeus, 1758 and Rossia macrosoma (Delle Chiaje, 1829) were fixed in Bouin’s and processed for histology as indicated above. Early organo- genetic stages of Rossia macrosoma were studied after fixation by detaching the embry- onic “cap” from the yolk mass and viewing it under the dissecting microscope in incident or transmitted light. OBSERVATIONS Loligo vulgaris When viewed under the SEM, the mantle rudiment at stage IX shows the peripheral tis- sue concentration giving rise to the mantle muscle and two slight elevations on either side of the shell sac pore, the fin rudiments (Fig. 1). They rise from the mantle surface during the ensuing stages when the shell sac becomes completely sealed (Fig. 2). These stages are characterized by a general con- traction of the embryo by which the organ rudiments are “assembled” into a more com- pact embryonic body. I FIG. 1. Ventrolateral view of Loligo vulgaris embryo at stage IX of Naef (SEM photograph). Arrows mark the fin rudiments. The two gill rudiments are indicated by arrow heads. The light patches (x) on the mantle border are ciliary cells. Scale bar = 50 um. MANTLE COMPLEX DEVELOPMENT IN CEPHALOPODS 167 FIG. 2. A. Right-side dorsolateral view of Loligo vulgaris embryo at stage XI (SEM), with one fin showing on top of the mantle rudiment (arrow head). B. Ventrolateral view of mantle and fins (partly damaged during preparation) in Loligo vulgaris embryo at stage XII. Scale bars in A and B = 100 um. 168 BOLETZKY FIG. 3. Ventrolateral view of mantle and oblique fins in Loligo vulgaris embryo at stage XIII (SEM). Note that fins remain unciliated on most of their surface. Scale bar = 50 um. FIG. 4. Juvenile Loligo vulgaris, 18 days old, in normal “hovering” position. Solid line indicates vertical axis in space, broken line corresponds to the plane of the fins (shown during a down beat). MANTLE COMPLEX DEVELOPMENT IN CEPHALOPODS 169 The fins lie in an oblique plane in relation to the longitudinal axis of the animal, closer to perpendicular than to parallel (Fig. 3). This is the position typical of the later embryonic and post-embryonic stages. When hovering, the hatchlings and early juveniles swim in an ob- lique position with the head down. The plane of the fins, oblique in relation to the body axis, is then roughly horizontal in space (Fig. 4). In rapid forward movements (e.g. attack), these terminal fins counteract the upward thrust of the reversed funnel tube and are thus essen- FIG. 5. Sepia officinalis embryos, reconstructed from histological cross sections, at stages XII, XIV and XVI. Arrow heads in the upper figure indicate the fin pouches forming under the fin rudiments. The stellate ganglia are marked black and the inner yolk sac is dotted (the posterior “appendix” of the yolk sac at stage XII is constricted by the closing mid-gut rudiment and will soon disappear). Arrow in lower figure indicates position of section shown in Fig. 6. 170 BOLETZKY у En > > 3_ : ® i ees " > „4 v N a A ea ie fon RS > OTRA = | ESS a SNS En - FIG. 11. Histological cross sections through the fin cartilage (c) and fin pouch (fp) in Rossia macrosoma, at late embryonic stage (XIX-XX, above) and in early juvenile, 20 days after hatching (below) at the same magnification. fin support (Cirrata) to the further reduction of both shell and fin (Incirrata). This “trend” does not of course explain what actually happened in the ancestor of the Incirrata. Other argu- ments will be necessary in formulating a coherent hypothesis for this crucial event. These arguments again will have to encom- pass all aspects of developmental coordina- tion, which is crucial in any constructional “deviation.” In the mantle complex of cephalopod em- bryos, such a deviation is not limited to gross morphology of muscular components in rela- tion to the shell; it includes also finer details in MANTLE COMPLEX DEVELOPMENT IN CEPHALOPODS 175 the structure of the integument. Indeed the absence of fins in incirrate hatchlings appears to be correlated with the presence of special integumental structures (the Koelliker organs) that act as auxiliary hatching equipment ad- justed to the particular egg structure of incir- rate octopods (Boletzky, 1978). A similar correlation between the morphol- ogy of the mantle complex and the differentia- tion of a particular hatching equipment exists in the decapods. In contrast to the incirrate octopods, all decapods have transitory cilia- tion that covers large parts of the embryonic integument. This ciliation, especially the bands of short cilia on the dorsal and ventral mantle surface of squid hatchlings, provides locomotion when the animal crosses the egg jellies (Boletzky, 1979). These specialized cili- ary bands are absent on the mantle integu- ment of sepiolid hatchlings (Boletzky, in press); they are “replaced” by an integumen- tal organ apparently involved in breaking the outer shell, which is rigid in Rossia eggs. The differentiation of the “terminal spine” on the mantle end is clearly correlated with the marked expansion of the rear part of the mantle and the corresponding reduction of the shell size described here. ACKNOWLEDGMENTS | am grateful to Mrs. D. Guillaumin and Mrs. M. Andre of the SEM laboratory of the Univer- sity of Paris (Laboratoire d’evolution des &tres organises, 105 Boulevard Raspail) for making the scanning electron micrographs. All histo- logical preparations were made by M. V. v. Boletzky to whom | give my warmest thanks. | also thank Dr. K. Bandel for making speci- mens of Sepioteuthis available. | gratefully acknowledge Dr. C. F. E. Roper and Dr. R. T. Hanlon for critically reading the manuscript. LITERATURE CITED BANDEL, K. & BOLETZKY, S. v., 1979, A com- parative study of structure, development and morphological relationships of chambered cephalopod shells. Veliger, 21: 313-354. BOLETZKY, S. V., 1978, Nos connaissances actu- elles sur le développement des Octopodes. Vie et Milieu, 28-29(1 AB): 85-120. BOLETZKY, $. v., 1979, Ciliary locomotion in squid hatching. Experientia, 35: 1051-1052. BOLETZKY, S. V., in press, Structure tegumentaire de l'embryon et mode d'éclosion chez les Cephalopodes. Bulletin de la Société Zoologique de France, 107. JELETZKY, J. A., 1966, Comparative morphology, phylogeny, and classification of fossil Coleoidea. Paleontological Contributions of the University of Kansas, Article Mollusca 7: 1-162. KOELLIKER, A., 1844, Entwicklungsgeschichte der Cephalopoden. Meyer & Zeller, Zürich, 180 p. NAEF, A., 1923, Die Cephalopoden. Fauna und Flora des Golfes von Neapel. 35. Monogr. (I, 1). NAEF, A., 1928, Die Cephalopoden. Fauna und Flora des Golfes von Neapel. 35. Monogr. (I, 2). PICKFORD, G. E., 1949, Vampyroteuthis infernalis Chun, an archaic dibranchiate cephalopod. II. External anatomy. Dana-Report, 32: 1-132. SPIESS, P. E., 1971, Die Organogenese des Schalendrüsenkomplexes bei einigen coleoiden Cephalopoden des Mittelmeeres. Revue Suisse de Zoologie, 79: 167-226. MALACOLOGIA, 1982, 23(1): 177-192 HISTOCHEMISTRY AND FINE STRUCTURE OF THE ECTODERMAL EPITHELIUM OF THE SEPIOLID SQUID EUPRYMNA SCOLOPES Carl T. Singley! University of Hawaii, Kewalo Marine Laboratory of the Pacific Biomedical Research Center, Honolulu, Hawaii 96813, U.S.A. ABSTRACT Histochemical analysis and electron microscopy reveal the presence of three distinct cell types comprising the ectoderm of Euprymna scolopes. The most prominent types were ovate- and goblet-shaped secretory cells. The secretory product of ovate cells was unreactive to periodate oxidation, alcian blue, azure A, colloidal iron and all iron diamine reactions, but was reactive with beibrich scarlet, bromphenyl blue, and the ninhydrine-Schiff test indicating the presence of basic proteins. In contrast, the product of ovate cells fixed in the process of active secretion was strongly metachromatic with azure A, alcianophilic at both pH 2.5 and pH 1.0, strongly basophilic, and was vigorouly reactive with all diamine sequences and colloidal iron. Furthermore, the ovate cell product remained periodate unreactive during secretion, but reactiv- ity with stains for basic proteins was lost. The ovate cell secretion product was judged by these criteria to be a strongly acidic protein-polysaccharide complex. The goblet cell secretion product was judged by these same histochemical tests to be a neutral polysaccharide having little if any protein. The third cell type was not obviously secretory and was found toward the outer surface of the epithelium. The mucous layer covering the epithelial surface stained as did the product of goblet cells. Electron microscopy revealed that the 30 um to 50 um thick epithelium was underlain by a 0.2 um thick basement membrane and below this a layer of orthogonally arranged striated collagen fibrils. The apical surfaces of all cells possessed microvilli covered with a thin layer of mucous material. Ovate cells had a single large vesicle filled with a fine granular material. In contrast, the secretory product of goblet cells consisted of large, electron dense membrane- bound granules. Goblet cells contained an extensive smooth ER-Golgi complex, whereas ovate cells possessed only small amounts of rough ER. Ovate cells had a meshwork of 5 nm to 8 nm microfilaments surrounding the large vesicle. Goblet cells contained small numbers of longi- tudinally oriented microtubules extending to the apical surface which were closely associated with secretory granules. The histochemically unreactive microvillous cells were revealed by TEM to contain very small membrane-bound vesicles containing a discontinuously electron dense material. The distribution of cell types was non-random. Ovate cells occurred in greatest numbers and on both dorsal and ventral surfaces. Goblet cells were found only in the dorsal epithelium and in greatest abundance on the head and back of the animal. Microvillous interstitial cells were found in both dorsal and ventral epithelia, but in greater numbers dorsally. The differential distribution and histochemical nature of the secretory products of the two glandular cells suggest that they play opposing roles in Euprymna’s camouflaging behavior. Key words: Cephalopoda; Sepiolidae; Euprymna scolopes; integument; histochemistry; be- havior. INTRODUCTION The sepiolid squid Euprymna scolopes Berry, 1919, like most members of the Sepiolidae (see Boletzky & Boletzky, 1970; Boletzky et al., 1971), is primarily nocturnal and burrows into the sandy ocean bottom dur- ing the day. In addition to burrowing in the sand, Euprymna is able to cause sand and other materials of the substratum to adhere to its dorsal surface. This behavior creates an effective camouflage while the animal is hunt- ing during the daylight or when it is disturbed from its hiding place. When the camouflage is no longer required the squid is able to release the adherent materials instantaneously. This sort of camouflaging behavior has not previ- ously been reported in a cephalopod. TCurrent address: Department of Zoology, The Ohio State University, 1735 Neil Avenue, Columbus, Ohio 43210, U.S.A. (177) 178 SINGLEY The character of Euprymna's camouflaging behavior suggests a mechanism involving the epidermis and its secretory activity. However, the nature and functions of cephalopod epi- thelial secretory products (indeed those of molluscs in general) have received little atten- tion regarding either their histochemical or specific biochemical characteristics (Hunt, 1970). The only available information con- cerning squid skin mucopolysaccharides (MPS) is provided by two brief studies. Anno et al. (1964) found that the MPS from the skin of Ommastrephes sloani pacificus is approxi- mately 70% unsulfated chondroitin. The skin of Loligo opalescens was found to contain MPS consisting of 70% highly sulfated chon- droitin sulfate and 25% chondroitin (Sriniva- san et al., 1969). There appears to be no in- formation on the biophysical properties of such secretions. This paper describes the results of an in- vestigation of the possible mechanism of sand adhesion and release by Euprymna scolopes. The nature of Euprymna’s epi- dermal secretions was studied using standard histochemical techniques. The histochemical characteristics of these secretions suggest that they may be involved in sand adhesion and release. The structure of the epidermis was Studied in further detail by transmission electron microscopy. Examination of the fine structure of the epidermis provides additional insight into the possible mechanisms involved in Euprymna’s camouflaging behavior. MATERIALS AND METHODS Adult Euprymna scolopes were collected in shallow intertidal areas at the island of Oahu, Hawaii and were maintained in the laboratory as previously described (Arnold et al., 1972). Histology and Histochemistry Animals were anesthetized by cooling to near 4°C, after which they were decapitated and rapidly immersed in one of the following: 2% calcium acetate-10% formalin (Ca-F) (Lillie, 1965), for 24 hr at 4°C; 0.5% cetyl- pyridinium chloride-10% formalin (CPC-F), for 24 hr at 20°C (Williams & Jackson, 1956); 0.5% cetylpyridinium chloride in Carnoy’s fix- ative (CPC-C), for 24 hr at 20°C to precipitate polyanionic polysaccharides; Lillie's (1949) acetic-alcohol-formalin (AAF), for 24 hr at 4°C; or 10% formalin in sea water, for 24 hr at 4°C. Samples of both dorsal and ventral epi- dermis were excised from the mantle while in the fixative. Samples of fins and arms were also removed and processed for histological comparison. To ensure the identical treatment of tissues subsequent to fixation, samples of epidermis fixed by each of the different methods were processed simultaneously. Tissues were de- hydrated through ethanol, cleared in xylene, and infiltrated with paraffin. Samples from each fixative were then embedded in the same block and sectioned at 5 um. This method of processing made it possible to evaluate the effectiveness of each fixative in preserving the secretory products of the skin and the effects of each fixative on the various staining procedures. Lillie's (1949) AAF provided the best cyto- logical detail and most vivid staining. All stain- ing reactions are described from material preserved by this method. No major devia- tions from these results were observed with the other fixatives. Neutral calcium acetate- formalin and formalin-sea water provided good preservation, but less intense staining. The CPC-containing fixatives produced ex- tensive cytological disruption, the loss of the epidermal surface and generally weak stain- ing. Various histochemical procedures were employed to elucidate the nature of the epi- thelial secretions. The procedures used were, except as noted, those outlined in Pearse (1968). The periodic acid-Schiff (PAS) technique alone (McManus & Mowry's [1960] modifica- tion, with prior acetylation, acetylation- deacetylation, or analine blockage of —OH groups) was used to demonstrate muco- substances having vicinal hydroxyls. Identifi- cation of mucosubstances presumably con- taining hexoses or deoxyhexose units, other than glucose, was accomplished using the periodic acid-para-diamine (pH 4.0) method (PAD) (Spicer, 1965). In addition, specific localization of periodate reactivity was con- firmed at the ultrastructural level using the periodic acid-thiocarbohydrazide-silver pro- teinate method (Thiery, 1967). To differentiate between neutral and acidic mucosubstances, the following metachro- matic and basic dyes were used: the mixed diamines (pH 4.0) technique (MD); 0.1% toluidine blue (TB) in 30% ethanol for 20 min (Kramer & Windrum, 1954); azure A (AA) in HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 179 various buffers at graded pH levels for 30 min; alcian blue G8X (AB) at graded pH levels for 2 hr and in combination with the PAS technique (Mowry, 1963); Mowry’s (1963) modification of Hale’s colloidal iron (Cl) for 2 hr; aldehyde fuchsin (AF) for 10 min (Halmi & Davies, 1953); the low iron (LID) and high iron dia- mine (HID) reactions for 18 hr (Spicer, 1965); and 1% alcoholic thionin (pH 1.0) as a general metachromatic stain for acidic mucins. Most of the preceding techniques were ef- fected concomitantly with supplementary pro- cedures and confirmatory tests which in- volved the chemical blockage, introduction or enzymatic removal of specific reactive groups in the mucosubstances. These procedures in- cluded both “mild” (4 hr at 37°C) and “active” (6 hr at 60°C) methylation, using methanol acidified to 0.1 N HCl, followed in control sec- tions by saponification for 20 min with 1.0% KOH in 70% ethanol (Spicer & Lillie, 1959). These methylation and methylation-saponifi- cation procedures were effected in conjunc- tion with AA, AB-PAS, AF-AB and HID-AB staining for selective blockage and restoration of various acid groups. Sulfate esters were introduced using a 1:1 mixture of sulfuric acid and acetic anhydride for 3 min or a 1:10 mixture of chlorosulfonic acid and pyridine at 70°C for 5 min (Kramer & Windrum, 1954; Pearse, 1968) prior to stain- ing with AA. The Bial reaction was performed as a test for the presence of sialic acid (Ravetto, 1964). In addition, digestion with Clostridium per- fringens neuraminidase (Sigma, Type VI) (Sialidase) was effected in conjunction with AA, AB, AF-AB and HID-AB staining for sialic acid containing mucosubstances. Sections were incubated at 37°C for 12 to 24 hr in 0.2 ml of 10 NFU/ml neuraminidase in citrate buffer at pH 5.0. Control sections were treated with buffer alone. Basic proteins were visualized with Beibrich scarlet (BS). Sections were stained for 1 hr at 20°C in 0.04% BS in phosphate buffer at pH 6.0 or Laskey’s glycine buffer at pH 8.0, pH 9.5 and pH 10.5 (Spicer & Lillie, 1951). In addition, comparative staining of proteins was effected using bromphenol blue (BPB) (for 2 hr at 20°C in 0.05% BPB, 1% HgCla and 2% acetic acid), acid solochrome cyanine (ASC) (for 10 min at 20°C in 1% ASC in 0.1 M citric acid, pH 2.1), and the ninhydrin-Schiff method (NS) (16-20 hr at 37°C in 0.5% ninhydrin and Schiffs reagent for 30 min). Electron Microscopy Adult squid were anesthetized in 0.5% urethane in sea water. Immobilized animals were immersed in 2.5% glutaraldehyde in Millonig's phosphate buffer, pH 7.4, plus 0.4 M sucrose for 15 min at room temperature, after which samples of dorsal and ventral epi- dermis were excised from the mantle while submerged in the fixative solution. Primary fixation was continued for an additional hour at 4°C. Samples of epidermis were then rinsed in buffer for 20 min, post-fixed in 1% OsO: in bicarbonate buffer (pH 7.4) for 1 hr at 20°C, dehydrated through a graded ethanol series and propylene oxide and embedded in EPON 812 (Luft, 1961). Sections, 60 nm to 70 nm thick were cut with a diamond knife using a Reichert OM-U2 or a Sorvall MT-2B ultramicrotome and mounted on uncoated mesh grids. Section were stained with uranyl acetate (Stempak & Ward, 1964) and/or lead citrate (Venable & Coggeshall, 1965). Obser- vations were made using a Philips EM 201 or EM 300 transmission electron microscope (TEM). KEY TO ABBREVIATIONS Stains Colors AA Azure A B blue AB Alcian Blue 8GX C grey ASC Acid Solochrom Cyanine G orange BPB Bromphenol Blue K black BS Beibrich Scarlet N brown Cl Colloidal Iron P purple D p-Diamine pk pink HID High Iron Diamine R red LID Low Iron Diamine Y yellow MD Mixed Diamines NS Ninhydrin-Schiff PAD Periodic acid-p-Diamine PAS Periodic acid-Schiff TB Toluidine Blue O RESULTS Histological Observations The epidermis of E. scolopes was observed to be a pseudostratified columnar epithelium composed of three morphologically distinct cell types and underlain by a thick basement membrane. The interstitial cell type was 180 SINGLEY polymorphic, appeared restricted to the distal surface of the epithelium and was distin- guished by a rounded nucleus. The ovate cell type was always the largest, ovate in shape, and contained a single, large secretory vesi- cle which flattened the nucleus against the basal surface. The goblet cell type contained darkly staining granules distributed from the region of the nucleus to the apical surface, producing the cells’ distinctive shape. Goblet cells typically had elliptical nuclei located near their basal ends and generally extended from the basement membrane to the surface of the Skin. Interstitial cells were the most abundant cell type in the dorsal epidermis, but were ob- served less frequently in the ventral epidermis where the ovate type was predominant. Ovate cells were observed in all regions of the epi- dermis with the exception of the distal sur- faces of the fins and arms. In contrast, goblet cells were found only in the dorsal epidermis and were especially abundant in the head and dorsal mantle. Goblet cells were observed less frequently in the dorsal skin of the arms and fins, and were completely absent in the dorsal-lateral surfaces of the fins and distal three-fourths of the dorsal arms where the epithelium was comprised of small, non- secretory cuboidal cells. Histochemical Observations Both the ovate and goblet cells secrete car- bohydrate-rich substances. In addition, a dis- tinct carbohydrate-rich layer covered the outer surface of the epithelium. The reactions of these cell types and the surface layer to specific histochemical staining procedures is summarized in Table 1-3 in which the figures TABLE 1. Staining reactions dependent on periodate oxidation of vicinal hydroxyls, and tests for basic protein. Cell type or tissue region Evacuating Histochemical method Surface layer Goblet cells Ovate cells ovate cells PAS 4PR 3-4PR 0 0 PAS-control V2pk 0-1pk 0 0 Acetylation (2 hr.)-PAS 0 1-2pk 0-1pk 0 Acetylation (9 hr.) -PAS 0 1pk 0 0 Acetylation (1 hr.)-Deacetylation-PAS 4PR 4PR 1-2pk 2pk Acetylation (9 hr.)-Deacetylation-PAS 4PR 4PR 1-2pk 2pk Analine-PAS 0 0 0 0 PAD (7 hr) 0 2YN 1C 4PK (24 hr) 1-V2C 2YN 1C 4PK (48 hr) 2-3C 3-4N 2C 4PK D (24 hr) without AP-oxidation 0 0 0 4PK MD 0-1N 2N 1N 4P Histochemical interpretation BPB-NR ASC (pH 2.1) (* NS Histochemical interpretation biphasic) Neutral mucosubstance with demonstrable vicinal- hydroxyls; suggests surface MPS are highly crosslinked. No demonstrable vicinal hydroxyls; acidic groups present. 0 2G 2-3R 0 0 1G 2-3R 0 0 2-3G 2-3R 2R 0 0-.5R V2-1R 0 2B 2B 4B 4B 1B 0-1B 3B 2G 1G 4B/G (*) 4G 4G/B (*) 0 0-1PB 2pk 1PB Some protein detectable Basic proteins in goblet cells, but present. absent or masked in surface layer. HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 181 TABLE 2. Alcianophilia at various pH levels and colloidal iron for acidic groups. A A EEE AAA CE EE RE EI IE SE EE a —_— A Cell type or tissue region Evacuating Histochemical method Surface layer Goblet cells Ovate cells ovate cells B (pH 2.5) 0 0-1B 0 3B B (pH 1.0) 0 0-1B 0 3B E (pH 2.6)-after Sialidase Digestion 0 0-V2B 0 3B AB (pH 2.6)-after Active Methylation 0 0 0 2-3pk AB-PAS 4PR 4PR 0 3B AB (pH 2.5)-PAS-after Active Methylation 4PR 3R 0 3B AB (pH 2.5)-PAS-after Active Methylation Saponification 4PR 3R 0 3B Cl 0-V2B у2В 0 2-3B CI-PAS 4PB 3-4PB 0 2-3B CI-PAS-after Active Methylation 0 0 0 0 Histochemical interpretation Periodate reactive with some Periodate unreactive; orthochromatic alcianophilia; highly sulfated slight Cl reactivity suggests mucosubstance. some free —COOH groups. TABLE 3. Azurophilia at various pH levels and tests for Sialic acid. Cell type or tissue region Evacuating Histochemical method Surface layer Goblet cells Ovate cells ovate cells TB 0 0-3B 2B 4PR TB after Active Methylation 0 0-2B 0-1B 2-3B AA (pH 0.5) 0 0 0 4PR (pH 1.0) 0 0 0 4PR (pH 3.2) 0 0 0 4PR AA (pH 1.0)-after Sulfation (alc) 0-1B 0-1B 0 3PR (pH 3.2)-after Sulfation (alc) 2B 1B 0 3B AA (pH 3.2)-after Sialidase Digestion 0 0 0 4P Bial 0 0 0 0 Histochemical interpretation Lack of metachromatic Metachromatic azurophilia azurophilia confirms lack of attributable to sulfate acidic sulfate groups; no groups; acid groups ap- sialic acid. pear masked in ovate cells. Observations of basophilia with various individual and combination dyes Thinonin 0 0 0 4R LID 0 0 0 4P LID-AB 0 0 0 4P PA-LID 0 0 0 4P HID 0 0 0 4P HID-AB 0 0 0 4P HID-AB after Salidase Digestion 0 0 0 4P PA-HID 0 0 0 4P AF 0 0 0 4P AF-AB (pH 2.5) 0 0-1B 0 4P AF-AB (pH 2.5)-after Sialidase Digestion 0 0-1B 0 4P Histochemical interpretation Neutral mucosubstance. Highly sulfated, but acidic groups appear masked in ovate cells. 182 SINGLEY represent visual estimates of the relative in- tensity of staining. Ovate cells which were fixed while in the process of secretion stain differently than non-secreting ovate cells so are treated as a separate cell type. Ovate Cells. The secretory vesicles of these cells were unstained by the PAS reaction as well as most other stains used in conjunction with prior periodate oxidation (Tables 1, 3). However, a light grey coloration was pro- duced by the PAD reaction and the acetyla- tion-deacetylation-PAS sequence produced staining not observed with PAS alone. No staining was observed with AA, AB, Cl or any other reaction to demonstrate acid groups. The secretory vesicle of ovate cells stained dark red with Beibrich scarlet at pH 6.0 to pH 9.5, but stained only light red at pH 10.5 (Table 1). Positive reactions for protein were also produced with BPB, ASC (pH 2.1) and the ninhydrin-Schiff reaction. Evacuating Ovate Cells. These, like ordinary ovate cells, were periodate unreactive. Stain- ing with BS was observed only at pH 9.5 sug- gesting the presence of basic protein (Table 1). The partially secreted cell product also stained strongly with BPB and NS. ASC (pH 2.1) staining in these cells was biphasic. That portion of the product remaining within the cells was orange whereas the portion extend- ing beyond the epithelial surface was blue. These cells exhibited strong orthochro- matic alcianophilia and were dark blue with Hale’s colloidal iron. In addition, they dis- played strong basophilia with LID, HID and AF. Staining was not reduced with prior neuraminidase treatment. The secretory pro- duct of these cells displayed strong y-meta- chromasia with TB and AA. Goblet Cells. The granular secretory product of goblet cells was strongly periodate-reactive as evidenced by the dark magenta coloration with PAS (Table 1). These cells reacted only moderately with BS and minimal reaction was observed with BPB. A strong reaction was ob- tained with ASC (pH 2.1) and, like the product of evacuating ovate cells, the staining was bi- phasic. The granular substance stained blue whereas the intergranular substance stained orange (Table 1). These secretory granules displayed weak alcianophilia, however, a strong reaction was observed after active methylation (Table 2). Reactivity with Cl was weak and no Staining was observed with TB or AA. In addition, there was no observable basophilia. Surface Layer. Staining reactions of the thin surface layer were essentially identical to those of the goblet cells with the exception that staining for basic proteins was weak or absent. Electron Microscopic Observations The dorsal epidermis ranges from 35 um to 50 um in thickness including the 0.2 um basement membrane (Fig. 1). The three major cell types described from light micro- scopic observation were readily distinguish- able with the TEM. The PAS-positive surface layer is comprised in part of a dense pile of microvilli (Figs. 2, 3). These microvilli are typi- cally 0.5 um to 0.6 um in length and 80 nm to 12 nm in diameter, are electron dense at their tips, and contain few oriented filaments (Fig. 3). Associated with the outer surface of the microvilli is a thin layer (100 to 300 nm) of fibrous mucous material. Between the micro- villi are Seen small (85 nm to 45 nm) electron dense granules (Fig. 3). Both the interstitial and goblet cells possess microvilli on their apical surfaces, whereas ovate cells lack microvilli. Interstitial cells con- stitute the largest proportion of the surface area of the skin by virtue of their greater num- bers and their broad apical surfaces (Fig. 2). The relative numbers and distribution of these cells is best appreciated by viewing sections near and tangential to the apical surface (Fig. 4). A complex system of nerve and glial cell processes is observed between the basement membrane and the secretory cells (Fig. 5) where they are most frequently associated with the bases of ovate type cells. Lying be- neath the basement membrane is a layer of orthogonally oriented collagen fibrils (Fig. 6), below which are the variously oriented layers of dermal muscle (Figs. 5, 6). Ovate Cells. Ovate cells of the dorsal epi- dermis are either roundly ovate or have the appearance of a drawstring purse with the nucleus flattened at the basal end (Fig. 1). Varying amounts of rough endoplasmic retic- ulum are observed within the cytoplasm sur- rounding the nucleus, but few Golgi bodies. The secretory material of these cells has a fine granular appearance (Fig. 7). Within the peripheral cytoplasm is a close reticulum of HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 183 FIG. 1. Electron micrograph of a cross section of dorsal epithelium. Sac-like ovate cells (oc) lie between interstitial cells (ic) which appear to deflect the lateral surfaces of the ovate cells. Interstitial cells extend microfilament-filled process toward the basement membrane (arrow). Bar = 10 ит. FIG. 2. Goblet cells (gc) often appear grouped, but are always interspersed among interstitial cells (ic). Note intracellular filaments in ic’s (pointer). Bar = 10 um. 184 SINGLEY FIG. 3. Microvillous surface at high magnification. The surface mucous layer is fibrous and contains no large granular component. Electron dense granules lie between the base of microvilli even in ic’s (arrow). Note the microfilamentous networks in the cell cortex (pointer). Bar = 200 nm. FIG. 4. This tangential section illus- trates the relative abundance of interstitial cells. All nuclei visible in this section are interstitial cell nuclei. This section is cut below the infolded apical ends of the goblet cells. Bar = 10 um. HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 185 FIG. 5. The basal region of the epithelium is often folded as illustrated here. The electron dense granules of glial cells (gl) are prominent beneath the ovate cells. Mantle muscles (m) underlie the epithelium. Bar = 5 um. FIG. 6. High magnification micrograph of the basement membrane (bm) illustrating the relationship of underlying collagen fibrils (c) and dermal muscle (M). Bar = 100 nm. 186 SINGLEY 5 nm to 8 nm filaments (Fig. 7). During se- cretory activity these filaments become rear- ranged and tightly packed. In addition, they appear to produce deep folds in the plasma membrane (Figs. 4, 8). Such folding begins near the apical end of the cells and prog- resses toward the basal end. The thickness of the peripheral cytoplasm in fully evacuated ovate cells is approximately twice that of pre- secretory ovate cells. Goblet Cells. Goblet cells are rounded at their bases, possess roughly ovate nuclei, taper in- ward toward their apical ends (Fig. 2) and have rounded cross sectional profiles except at their apices (Figs. 4, 9). Toward their apices the cells are infolded laterally, produc- ing a “pansy-like” profile in tangential section (Fig. 10). A branching profile is often seen in cells cut longitudinally (Fig. 11). The microvil- late surface of these cells is much smaller in area compared with the interstitial cells and the microvilli are frequently much shorter (Eigs. 9, 141): Numerous microtubules (Mt) are observed oriented along the long axes of goblet cells (Fig. 9). These microtubules are always most numerous in the portion of the cells which ap- pear to be actively secreting (Figs. 9, 10), and are frequently seen in close proximity to the secretory granules. The highly electron-dense secretory gran- ules are spherical and appear to vary con- siderably in size. The largest are typically 0.5 um to 0.7 um in diameter. The density of these granules is uniform and each is sur- rounded by a single unit membrane (Fig. 11). During active secretion the granules appear to break up into smaller units as they ap- proach the apical surface (Figs. 9, 10). The periodate reactivity of these granules and the surface mucous layer visualized by light microscopy is confirmed by the peri- odic acid-thiocarbohydrazide-silver proteinate (PATCSP) method of Thiery (1967) (Fig. 12). The density of silver grains is uniformly great- er over the large secretory granules than over the cytoplasm. The mucous layer at the sur- face also reacts strongly, with the grains found only over individual fibers. In contrast, the 35 nm to 45 nm granules lying between the microvilli are unreactive. Near the bases of the goblet cells one typi- cally observes a large rER-Golgi complex (Fig. 13). The Golgi is generally more massive than the rER and the Golgi saccules display a graded electron density. Interstitial Cells. The TEM reveals that the in- terstitial cells span the full thickness of the epidermis (Fig. 2). In most interstitial cells the nucleus and the majority of the cytoplasm is concentrated near the epithelial surface (Figs. 2, 4). The remainder generally extends in a narrow process which contacts the basement membrane. The cytoplasm of these cells is typically filled with thick (10 nm to 15 nm) and thin (5 nm to 8 nm) filaments (Figs. 2, 11) which are especially prominent in the basal processes (Fig. 13). Interstitial cells also possess small, mem- brane bound vesicles containing material with a non-uniform electron density (Figs. 8, 10). Although these granules may be for secretion, no conclusive evidence of their exocytosis has been observed. During the preparation of tissues for the TEM, particles of anomalous material were often retained on the microvillous surface of the epithelium (Fig. 14). Such particles appear to be firmly attached to the microvilliby means of the fibrous surface mucus. No particles were observed adhering to the ventral skin. DISCUSSION Histochemistry provides some indication of the general characteristics of Euprymna’s epithelial secretions. The secretory products of the two glandular cells appear to be pro- tein-polysaccharide complexes. The secretory material produced by goblet cells appears to be a neutral mucopolysac- charide. This material is highly periodate- reactive indicating the presence of vicinal hydroxyl groups (Table 1). The reaction of this material to PAD and MD staining confirm its periodate reactivity and do not suggest the presence of hexose or deoxyhexose units other than glucose. The specific localization of periodate reactivity in both goblet cell gran- ules and the surface layer is further confirmed by the PATCSP method (Fig. 12). The lack of staining observed in goblet cell granules with AA, AF, LID and HID (Tables 2, 3) suggest the absence of strongly acidic groups. Whereas slight reactivity observed with TB and Cl indicates the possibility that some carboxyl groups are present, is not con- firmed by AF or LID staining. These results suggest that the observed periodate reactivity is due to vicinal hydroxyls rather than closely associated hydroxyl and carboxyl groups. The goblet cell mucin displays moderate HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 187 FIG. 7. Micrograph illustrating orientation of filaments within the peripheral cytoplasm of two adjacent ovate cells (oc). Bar = 1 um. FIG. 8. Micrograph illustrating the folded aspect of the apical end of an ovate cell. Note the appearance of the peripheral filaments. Bar = 1 um. 188 SINGLEY FIG. 9. A goblet cell illustrating apical narrowing and the presence of an apparently actively secreting segment. Note the orientation of microtubules (mt) within this area and the dispersion of the large secretory granules. Bar = 1 ит. FIG. 10. Illustration of the “pansy-like” profile of a goblet cell cut tangential to the surface at the level indicated by the large arrow in Fig. 9. Note the abundance of microtubules in the actively secreting branch (arrow). The non-uniformly electron dense granules of interstitial cells (arrowheads) are most numerous near the surface of the epithelium. Bar = 1 um. HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 189 FIG. 11. Cross section at the apex of a goblet cell. The large membrane-bound granules are often seen in close proximity to microtubules (arrow). The apical ends of the cells typically form desmosomal junctions with neighboring cells. Bar = 1 um. FIG. 12. This section stained by the PATCSP method illustrates the localization of periodate reactivity. Silver grains are prominent over goblet cell granules and the surface mucous layer whereas the small granules between microvilli are unreactive (arrows). Bar = 1 um. 190 SINGLEY FIG. 13. Micrograph illustrating the rER-Golgi complex of goblet cells. In these cells the Golgi (G) was relatively more massive than the rough endoplasmic reticulum (rER). The adjacent interstitial cells (ic) contained large numbers of thin filaments which were often seen connected to highly folded membrane (arrow). Bar = 1 um. FIG. 14. Micrograph of the microvillous epithelial surface with an associated particle of unknown composition. The fibrous surface mucus appears to be attached both to the microvilli and the particle (inset, 30,000). Bar = 1 um. HISTOCHEMISTRY AND FINE STRUCTURE OF SQUID SKIN 191 reactivity to all stains for protein (Table 1). The positive staining with BS at all pH levels below pH 10.5 suggests the presence of neu- tral or acidic proteins. The biphasic nature of staining with ASC is consistent with the re- sults with BS. The blue coloration of the mucous granules suggests the presence of acidic protein, and is contrasted with the orange color produced by basic proteins in the cytoplasm. The secretory material of the ovate cells appears to be a highly sulfated protein-poly- saccharide complex. This material is perio- date-unreactive (Table 1) and is strongly re- active with tests for basic protein. Their nega- tive PAS reactivity and the lack of a positive reaction to tests for acidic groups by the ma- jority of ovate cells, contrasted with the very strong staining of this material in evacuating ovate cells suggests that acidic groups are present but masked in the unsecreted mate- rial. It is possible that a protein component is interfering with the staining of the polysac- charide moiety (Pearse, 1968). If this is the case, then in order to produce the observed color differences there must be an alteration in the protein-polysaccharide association when this material comes into contact with sea water. In this regard, the biphasic staining of granules fixed after partial exocytosis is of considerable interest. With ASC at pH 2.1 (Table 1) the portion of the secretion product in contact with the exterior stained blue, indi- cating a shift from acidophilia to basophilia. Similar shifts are seen with thionin, LID, HID and AF staining (Table 3). In addition, the y- metachromasia observed in evacuating ovate cells with AA and TB and comparison of the reactivity of ovate and evacuating ovate cells with AB and Cl (Table 2) suggests an un- masking of strongly acidic groups. The marked reactivity of evacuating ovate cells suggests further that these are sulfate groups. The lack of a positive Bial reaction with this, as well as other materials in the epithelium, is consistent with previous results indicating a lack of sialic acids in invertebrates (Ravetto, 1964). Inasmuch as the histochemical nature of the surface mucus parallels closely that of the goblet cell mucin, one might assume that this material consists largely of goblet cell secre- tions. The observed differences in their re- activity to stains for protein, however, argues that the secretory material is considerably modified. What these changes might consist of remains to be revealed by biochemical analysis. How such alterations might be in- volved in the functions of these materials dur- ing sand adhesion and release also remain to be studied. The observation that goblet cells are found only in the dorsal epithelium, which is the only area of the surface where sand adhesion oc- curs, suggests that the secretory product of these cells may impart the adhesive charac- teristic of the surface mucus. The presence of a massive, membrane- associated filamentous layer in the peripheral cytoplasm of ovate cells (Fig. 7) and the close proximity of gliointerstitial-nerve cell com- plexes (Nolte et al., 1976; Nicaise, 1973) to their bases (Fig. 5) suggests that the rate of secretion by ovate cells may be relatively rapid. By comparison, goblet cells contain no microfilamentous arrays and are seldom as- sociated with glial cells (Figs. 2, 9, 10). This, in addition to the presence of oriented micro- tubules close to the secretory granules and the fact that only one segment of these cells appears to be actively secreting, suggests that goblet cells may secrete at a slower rate than ovate cells. If such differences in rates of secretion in fact exist, it would explain the observation (unpublished) that subsequent to the release of adherent sand, there is a period of several minutes during which squids are refractory to sand adhesion. How then might sand release be accom- plished? Since ovate cells secrete a highly acidic mucoprotein and are distributed over most of the squid’s body, they would appear to be the most likely candidate for a means of accomplishing sand release (deadhesion). Highly sulfated mucins are generally “slimy” materials (Hunt, 1970), a character that fits the proposed role of these cells. Secretion of such a material by the ovate cells might cause the adhesive mucous layer and adherent sand to be lifted away from the microvillate surface. After release the adhesive character- istic of the surface would be restored. The folded aspect of the epidermal base- ment membrane viewed in cross section (Fig. 5) suggests the possibility that active defor- mation of the skin might also be involved in deadhesion. Such deformations could be mediated by the layers of dermal as well as mantle musculature and/or possibly by the interstitial cells. The presence of large num- bers of oriented thick and thin filaments within the basal extensions of interstitial cells (Fig. 2, 5, 13) suggests that these cells may possess contractile capabilities. Contraction of these 192 SINGLEY cells would not only result in a folding of the epithelial layer, but would also augment the secretory activity of the ovate cells. Unfortunately, histochemistry and the TEM provide a static picture of the structure of Euprymna's epithelial covering. A clear un- derstanding of the mechanisms involved in Euprymna's camouflaging behavior must await both the biochemical characterization of its epithelial secretions, including the putative secretory product of interstitial cells (Figs. 3, 10), as well as experimental evidence for the mechanisms of the control of secretory activ- ity. ACKNOWLEDGEMENTS This work was supported in part by PHS grant no. 9RO1 EY 00179 and NSF grant no. GB 22604 to Dr. J. M. Arnold. | also thank Dr. Arnold and Lois D. Williams-Arnold for their helpful comments and Dr. R. G. Kessel and Mike King for reviewing the manuscript. REFERENCES CITED ANNO, K., KAWAI, Y. & SEON, N., 1964, Isolation of chondroitin from squid skin. Biochimica et Biophysica Acta, 83: 348-349. ARNOLD, J. M., SINGLEY, C. T. & WILLIAMS- ARNOLD, L. D., 1972, Embryonic development and post-hatching survival of the sepiolid squid Euprymna scolopes under laboratory conditions. Veliger, 14: 361-364, 6 pl. BOLETZKY, S. VON & BOLETZKY, M. V. VON, 1970, Das Eingraben in Sand bei Sepiola und Sepietta (Mollusca, Cephalopoda). Revue Suisse de Zoologie, 77: 536-548. BOLETZKY, S. VON BOLETZKY, M. V. VON, FROSCH, D. & GATZI, V., 1971, Laboratory rearing of Sepiolinae (Mollusca: Cephalopoda). Marine Biology, 8: 82-87. HALMI, N. S. 4 DAVIES, J., 1953, Comparison of aldehyde fuchsin staining, metachromasia peri- odic acid-Schiff reactivity of various tissues. Journal of Histochemistry and Cytochemistry, 1: 447-459. HUNT, S., 1970, Protein-Polysaccharide Com- plexes in Invertebrates. Academic Press, New York, 329 p. KRAMER, H. & WINDRUM, G. M., 1954, Sulfation techniques in histochemistry with special refer- ence to metachromasia. Journal of Histochemis- try and Cytochemistry, 2: 196-208. LILLIE, R. D., 1949, On the destruction of cyto- plasmic basophilia (ribonucleic acid) and of the metachromatic basophilia of cartilage by the glycogen splitting enzyme malt diastase: a histo- chemical study. Anatomical Record, 103: 611- 633. LILLIE, R. D., 1965, Histopathologic Technique and Practical Histochemistry. McGraw-Hill, New York, 715 p. LUFT, J. H., 1961, Improvements in epoxy resin embedding methods. Journal of Biophysical and Biochemical Cytology, 9: 409. MCMANUS, J. F. A. & MOWRY, R. W., 1960, Stain- ing Methods. Histologic and Histochemical. Hoeber, New York, 423 p. MOWRY, R. W., 1963, The special value of meth- ods that color both acidic and vicinal hydroxyl groups in the histochemical study of mucins. With revised directions for the colloidal iron stain, the use of alcian blue G8X and their combina- tions with the periodic acid-Schiff reaction. An- nals of the New York Academy of Sciences, 106: 402-432. NICAISE, G., 1973, The gliointerstitial system of molluscs. /nternational Review of Cytology, 34: 253-332. NOLTE, A., REINECHE, M., KUHLMANN, D., SPECKMANN, E-J. & SCHULZE, H., 1976, Glial cells of the snail brain. Some remarks on mor- phology and function. In: SALANKI, J. Neuro- biology of Invertebrates. Gastropoda Brain. Tihany 1975. p. 123-138, Akademiai Kiado, Budapest. PEARSE, А. G. E., 1968, Histochemistry, Theoreti- cal and Applied. Vol. |, ed. 3, Williams & Wilkins, Baltimore, 759 p. RAVETTO, C., 1964, Histochemical identification of sialic (neuraminic) acids. Journal of Histochem- istry and Cytochemistry, 12: 306. SPICER, S. S., 1965, Diamine methods for differ- entiating muco-substances histochemically. Journal of Histochemistry and Cytochemistry, 13: 211-234. SPICER, S. S. & LILLIE, R. D., 1959, Saponifica- tion as a means of selectively reversing the methylation blockage of tissue basophilia. Journal of Histochemistry and Cytochemistry, 7: 123-125. SPICER, S. S. & LILLIE, R. D., 1961, Histochemical identification of basic proteins with Beibrich Scarlet at alkaline pH. Stain Technology, 36: 365-370. SRINIVASAN, S. R., RADHAKRISHNAMURTHY, B., DALFERES, E. R. JR. & BERENSEN, G. S., 1969, Glycosaminoglycans from squid skin. Comparative Biochemistry and Physiology, 28: 169-176. STEMPAK, J. C. & WARD, R. T., 1964, An im- proved staining method for electron microscopy. Journal of Cell Biology, 22: 697-701. THIERY, J-P., 1967, Mise en evidence des poly- saccharides sur coupes fines en microscopie electronique. Journal de Microscopie, 6: 987- 1018. VENABLE, J. H. & COGGESHALL, R., 1965, A simplified lead citrate stain for use in electron microscopy. Journal of Cell Biology, 25: 407- 408. WILLIAMS, G. & JACKSON, D. S., 1956, Two or- ganic fixatives for acid mucopolysaccharides. Stain Technology, 31: 189-191. MALACOLOGIA, 1982, 23(1): 193-201 MORPHOGENESIS OF CHROMATOPHORE PATTERNS IN CEPHALOPODS: ARE MORPHOLOGICAL AND PHYSIOLOGICAL ‘UNITS’ THE SAME? Andrew Packard Department of Physiology, University Medical School, Teviot Place, Edinburgh EH8 9AG, Scotland and Department of Biology, University of Victoria, Victoria, British Columbia, Canada ABSTRACT This paper discusses the differences between the two kinds of units in the skin of cephalopod molluscs that are recognized as entities of visible pattern generation—one anatomical and the other physiological—and points out that they are related in morphogenesis: ¡.e. to pattern generation in the developmental sense. The ‘chromatic unit,’ containing all the anatomically fixed elements (iridophores, leucophores, chromatophores) deployed in color change, should not be confused with the various transient physiological units (e.g. ‘motor unit’ neurophysiologically definable as a motoneurone and the particular elements it innervates). The anatomical unit develops through processes that parcel the skin up into compartments 1-4 mm across giving rise to the patches of Octopus and their equivalent (the large dark chromatophores and their satellites underlain by iridescent cells) in loliginid squids. Neurophysiological units too are poten- tially derivable as members of a developmental series; each is a chronological unit innervating a particular age-class of chromatophore elements identifiable by their size, color and relative positions in the morphological array but not confined to one particular anatomical unit. Examples are given of the different appearances produced by these units in the squids Lolliguncula brevis (Blainville, 1823) and Loligo plei Blainville, 1823 and in Octopus vulgaris (Lamarck, 1798) and Octopus rubescens (Berry, 1953). Nine rules are listed for the interpretation of chromatophore clusters based on Octopus studies but generalizable to other cephalopods. Key words: octopus; squid; skin; pattern; morphogenesis; electrophysiology. INTRODUCTION The question of what it is that allows octo- puses and squids to exhibit a range of pat- terns within the space of a few seconds is in one way easily answered. Distributed over the transparent body surface are chromatophores that expand or retract under nervous control: areas of expanded chromatophores turn dark or colored while the rest of the skin remains white or translucent. But in this relatively sim- ple statement lurks a dilemma for the biolo- gist, that has to do with the old dichotomy between form and function: anatomy and physiology. The arrangement of individual chromatophores is as much anatomically fixed in the skin as are the positions of light bulbs in an illustrated bill-board or the ar- rangement of muscles in the rest of the body, and it alters only with growth processes. But the patterns seen are transient phenomena that result from the various ways in which the chromatophore elements are switched on by electrical activities originating in the central nervous system. It would be foolish to try to account for the various displays that appear on an illuminated bill-board by giving a de- tailed description of the two-dimensional mat- rix of light bulbs making up the bill-board when such displays are a function of spatio- temporal connections encoded in a central programme. And yet a description of the mat- rix is necessary if one wants to account for the quality of the pictures displayed: for instance, the amount of detail or grain, and the colors available. In describing how patterns are gen- erated, there is no easy way of resolving the dilemma of whether to adopt a morphological or a physiological approach, for both are needed. This paper is intended a) to clear up a source of confusion that has already crept into the relatively young literature on body pattern- ing in cephalopods, resulting from use of the word ‘unit’ both in physiological and in mor- phological senses; b) to show that the prob- lem of choosing whether patterns are to be described in terms of the static morphological (193) 194 PACKARD array of chromatopores or in terms of the physiological responses (i.e. the events that play upon the array) can in part be resolved by recognizing that the two have common origins in morphogenesis. That is, by follow- ing the course of ‘pattern generation’ in the developmental sense. The word ‘unit’ (i) the term ‘motor unit’ has been used by Maynard (1967) for the cuttlefish Sepia, Florey (1969) for the squid Loligo and by Packard (1974) and Packard & Hochberg (1977) for Octopus, on the basis of neuro- physiological findings; (ii) the term ‘chromatic unit’ was coined by Packard & Sanders (1971) for the charac- teristic patch structure of the skin of octo- puses and used by Hanlon (1982) for squids. Motor units Both Florey and Maynard used the term motor unit—in the classical sense of a moto- neurone and the muscle fibers it innervates— to describe the responses obtained from elec- trical stimulation of nerve branches to the skin. They succeeded in isolating the all-or- nothing resonses of individual nerve fibres either by paring down the nerve branches or by adjusting the stimulus strength until it was clear that only single fibres were involved. The same rules hold for both the squid Loligo and the cuttlefish Sepia. One motoneurone innervates many chromatophores, and a sin- gle chromatophore can be innervated by more than one motoneurone. These two conditions are summarized in Florey's fig. 3. Multiple innervation of chromatophores (i.e. the fact that the muscles radiating from a chromatophore are not necessarily innervat- ed by branches of the same motoneurone) means that one and the same chromatophore can participate differently in different patterns. Maynard gave a dramatic example of this when he showed that some of the dark chro- matophores (melanophores) in the units he was studying are used either to enhance a white spot—through contraction of the muscle fibres on the side of the melanophore away from the spot, giving the spot a dark edge—or FIG. 1. Lower ‘lid’ of the left eye of a young Octopus rubescens (6 g body weight) showing a dark spot flashing on and off during partial alcohol anaesthesia. The constancy of this spot over several on/off cycles suggests it is part of a single motor unit (see text). Close inspection of the photographs reveals that the chromatophores involved in the response—seen in the resting, ‘off’ condition in 1, partly ‘on’ in 2 (all elements not yet invaded) and ‘on’ in 2—can be characterized by larger resting sizes than their immediate unresponsive neighbours. Most of the responding chromatophores occupy the slopes of a groove separating two patches. Photographs taken with Wild ‘Makro’ optics and automatic camera using flash illumination and fibre optics light source. Scale bar 0.5 mm. MORPHOGENESIS OF CHROMATOPHORE ‘UNITS’ 195 else to do the opposite, namely to help screen the white spot (through a second motoneu- rone that innervates the muscle fibres on the side over the spot). He cited this phenomenon as an illustration of what he called pattern- position separation. For our purposes of find- ing a suitable candidate to be classed as a unit of patterning, such “double agents,” shift- ing one way or another depending on which motoneurone is firing, are clearly unsuitable. Single chromatophores are not units of phys- iological pattern. The motor units of Florey and Maynard have the further characteristic that they play a recognizable part in generating whole body patterns. White spots on a dark ground are — (59) Ww EN resting grooves only patches & grooves patch centers @ -50 um e ~35 um } mean diameters (resting) 220) ae! FIG. 2. Drawings of the dorsal skin of Octopus vulgaris (adult) showing 3 kinds of physiological unit occupying a region of patch and groove units, seen in the pale (resting) condition in 1. A) As they might appear during natural patterning (scale bar 1 mm). B) Details of the responses in the area indicated seen at the level of individual chromatophores (scale bar 0.5 mm). Note that (a) in drawings 1-4 the same patches are present in A and the same chromatophores are present in B, (b) responses respect the patch and groove arrangement of the skin but are not confined to one patch and are not necessarily continuous with it, (c) responding chromatophores of the different physiologi- cal units, 2, 3 and 4 occupy complementary positions within a patch unit, (d) they belong to different resting size classes in descending order from the left. Responses were obtained by electrical stimulation of the skin. In these units, only dark chromatophores (melanophores) are involved. The hatched melanophores in B2 are intended as an indication that more than one motor unit takes part in the groove response. N.B. Chromatophores not involved in any of the three kinds of response are not figured. The drawing is a simplified version of the condition in life: e.g. there are fewer chromatopores than in life, and there are no shared chromatophores whereas in life 10-20% of the melanophores of neighboring size Classes are shared by such complementary units. (After Packard, 1978). The area studied is just proximal to the large white spot on the second arm of O. vulgaris. 196 PACKARD characteristic of some of the patterns of the cuttlefish; while they have other patterns in which the reverse is true: the white spots are overlain by dark screens. In other words, the motor units are also units of visible patterning. A similar picture emerges in Octopus both during spontaneous patterning (Fig. 1) and following direct stimulation of the skin, evok- ing axon reflexes. Electrical stimulation pro- duces darkening over definite areas that are recognizable as features of patterning dis- played by the octopus in its natural environ- ment (Fig. 2). A notable example is the eye- spot of the two-spotted octopus O. bimacu- loides (Packard & Hochberg, 1977). These features were originally referred to as chro- matophore fields (Packard, 1974). The ques- tion of how many nerve fibres might be in- volved was unanswered. Morphological units Various details of the make-up of the patch and groove structure called—for better or worse—a chromatic unit have now been given, particularly for parts of the dorsal skin of Octopus vulgaris (Packard & Sanders, 1971; Messenger, 1974; Packard & Hoch- berg, 1977; Froesch & Messenger, 1978). Variations in the arrangement of the patches between one species of Octopus and another are tabled in Packard & Hochberg (1977). The idea (Packard & Sanders, 1969; FIG. 3. Typical series of morphological units on dor- sal mantle surface of loliginid squid Loligo plei, each centered on a single large dark chromato- phore. Most chromatophore classes are physio- logically expanded. Area framed in middle of photo- graph is analyzed in Fig. 4. Scale bar 2 mm. The photograph is of the TV monitor screen taken during playback of a living squid exhibiting natural patterning recorded on videocasette by black and white TV camera through Wild dissecting microscope. Packard, 1972) that the patches, with their combination of reflecting elements and dark screening chromatophores (melanophores), enable the octopus to achieve part of its camouflage by tone-matching (through neu- tral density filtering of the light striking the skin) has now been experimentally estab- lished by Messenger (1974), Froesch & Messenger (1978), and Messenger (1979) who have drawn particular attention to the ar- rangement of the leucophores. (See Discus- sion.) The equivalent of the octopus patch and groove chromatic unit, which contains hun- dreds of chromatophores and many leuco- phores in an area little more than 1 mm?, is not immediately apparent in other cephalo- pods. Squid chromatophores are almost an order of magnitude larger than octopus ones, and local densities are of the order of ten per square millimeter compared with 300 to 400 in Octopus. However, in loliginid squids (Figs. 3-5) the dark chromatophores (brown and red) on the dorsal mantle and elsewhere are arranged as a series of continuous and inter- secting circles 1-2 тт across, each circle with a single large dark chromatophore at its FIG. 4. Rough classification of individual chromato- phores arrived at by inspection of the area indicated in Fig. 3 (Loligo plei). The classification is based on descending order of size and of pigment density. The numbers (2-5 etc.) give the putative succes- sion in which chromatophores were recruited into the morphological unit centered on 1 (and into neighboring units) and define their membership in the various age-dependent physiological units (see Fig. 5 and text: Rules for the interpretation of chro- matophore clusters). MORPHOGENESIS OF CHROMATOPHORE ‘UNITS’ 197 centre and with pale (yellow) ones inter- spersed (Fioroni, 1965; Hanlon, 1982). The circles have similar dimensions to the patches of the octopus, and, like the patches, they are, in many parts of the body underlain by con- spicuous accumulations of reflecting material (see Hanlon, 1982). Each circle is a morpho- logical unit with the same status as the patch and surrounding groove of Octopus. Each morphological unit contains all of the ele- ments that contribute to patterning in that part of the Skin. Morphogenetic Interpretation of Chromatophore Clusters The rules for the development of the mor- phological array (morphogenesis) of chroma- tophore clusters have been worked out for Octopus but appear to be general. They are based on inspection of photographs taken in the plane of the skin surface with macro- optics at different stages of ontogeny and dur- ing different phases of activity of the different classes of chromatophore: at rest, spontane- ously active and electrically stimulated. (The word recruitment is used in the sense of re- cruitment of new members into a population.) 1. General recruitment. Recruitment of chromatophores into the mantle, head and arm fields occurs from germinal cells spread throughout each expanding field. 2. Field effects. The rate of recruitment varies from one part of the field to another according to proximo-distal and dorso-ventral gradients established early in ontogeny (Fioroni, 1965) and to local conditions that alter with recruitment. FIG. 5. Details of the responses of dorsal mantle chromatophores of the squid Lolliguncula brevis recorded with black-and-white video-technique (see caption to Fig. 3) during natural patterning. The same chromato- phores are present in all six photographs; the central one appears torn and has been chosen as a marker. 1-3 serial expansion of different size classes. 1, all chromatophores at resting size (i.e. fully retracted); 2, responses confined to the larger resting-size classes (i.e. central chromatophore and chromatophores numbered 1-5 and 7-10 etc.); 3, all classes responding (N.B. in the expanded state yellow chromatophores are barely visible, but none remain in the resting state: cf. first photograph (1)). 4-6 partial responses of the array. 4 and 5 subdivision of the yellow chromatophores (i.e. smaller size classes) into complementary units whose boundary lies half way across the unit (compare photographs 4 and 5); 6 reduced responses of the large (dark) chromatophores (compare with 2) with near-maximal responses of the smaller size-classes (yellow). 198 PACKARD 3. Field subdivision. As each field ex- pands, sub-fields arise within it. 4. Position and size constancy. Once es- tablished, a chromatophore does not alter its position in the field nor are there appreciable changes in resting size. 5. Mutual relationship. Established chro- matophores influence the position, size and rate of differentiation of further chromato- phores arising in their neighborhood through processes of /ateral inhibition which result in chromatophores being spaced out relative to each other. 6. Size hierarchy and age. Chromato- phores can be ranked according to size. The size-hierarchy (based on resting diameters) in a cluster reflects the order in which the chro- matophores were born. The largest chroma- tophores are the earliest, the smallest are the latest. (N.B. The size hierarchy is maintained when all chromatophores in the cluster are uniformly expanded.) 7. Age and color. When first born, chroma- tophores are pale (usually yellow or orange in color) and become progressively darker with age as the density of the pigment increases. 8. Color and degree of expansion. The depth of a chromatophore's color is inversely proportional to its degree of physiological ex- pansion. 9. Separate responses. Chromatophores of a given size and color class are able to expand or retract independently from other size and color classes (Figs. 2, 5), i.e. each age Class has its own innervation and muscle connections. These largely descriptive rules for under- standing two-dimensional pictures in terms of the time dimension can be converted into Rules for the Conduct of Young Chromato- phores such as Rule 4 “Stay put once you have arrived,” Rules 5 and 6 “keep clear of neighbors already established and never grow larger than they.” Figure 4 is an example of how the rules can be applied for the interpretation of close-up photographs of the skin. Figure 6 is a visual Summary of the rules at work. Every structure is but a frozen stage of growth and the squid skin is not exceptional in this, but because comparatively few chroma- tophores are present and they are all size- and color-coded, the morphogenetic history of the skin can be followed with great clarity even in the adult. Here we find all four dimen- sions collapsed, as it were, into one plane. Figure 6 also answers the question posed in the title of this paper: Question: Are morphological and physio- logical units the same? Answer: See how they grow. DISCUSSION This is not the place to go into details of how the rules for pattern generation were ar- rived at, but, as they have not been published elsewhere, some discussion is appropriate. They are general rules that satisfy only part of the description. There are no values at- tached, and they say nothing about a number of interesting questions such as: 1) whether all yellow chromatopores eventually turn dark; 2) what is the influence of chromatophores upon the distribution of reflecting elements (leucophores and iridophores) which also contribute to a unit; 3) how do new units arise; 4) what accounts for the wide variety of units found in some species; 5) what are the details of innervation. 1) There are morphological units in which some of the yellow chromatophores are larger (in expanded diameter) than some of their darker neighbors. They may belong to an early age-class that is programmed to remain yellow/orange throughout ontogeny. 2) As Hanlon (1982) has shown, reflecting elements underly the central chromatophore of a morphological unit in loliginid squids much as the units are centered upon patches of re- flecting elements in octopuses (see above: ‘Morphological Units’). Presumably, both in octopuses and in squids, these reflecting ele- ments (leucophores and iridophores) are ex- posed to the same sources of positional infor- mation that determines the differentiation of chromatophores. The empty spaces unoccu- pied by leucophores lying immediately below dark chromatophores in the small arm patches of Octopus vulgaris (Froesch & Messenger, 1978) can be interpreted as meaning that the leucophores are born later than the chromatophore they appear to sur- round and outside its sphere of inhibition. 3) New morphological units arise as the two-dimensional field expands. It appears that local centres of pattern generation concentric with the morphological units have a radius of influence limited to one or two millimeters, and that new centres will arise in areas of the expanding field that lie beyond this distance. MORPHOGENESIS OF CHROMATOPHORE ‘UNITS’ 199 Morphological ‘Units’ UN Physiological ‘Units’ ul © ~ >» Il D I) 1 тт | < Эр а © > У FIG. 6. Diagrammatic representation of the pattern of recruitment of chromatophores to form a single ‘morphological unit’ in the dorsal skin of a young squid during a 5-fold increase in linear size of the region indicated. Resting sizes are slightly exaggerated. Five successive age-classes of chromatophore (I-V) differing in size and colour at any one moment (see text) are figured each theoretically with its own nerve supply (right). At first appearance, age-classes Il-IV are shown physiologically expanded, through contrac- tion of their muscle fibers. Members of age-class V (x) are still being born and their nerve supply is in the process of growing in. Note that the members of a particular age-class have smaller resting diameters than their predecessors, and that chromatopores retain their relative positions during ontogeny but turn darker with time. In the last frame, two new ‘morphological units,’ centered on chromatophores 3 and 4 (age-class 11), are seen in the process of formation. 200 Such a nascent unit is seen in the final state of the diagrammatic unit figured in Fig. 6. This explanation of the patterns seen in- vokes the principle of lateral inhibition (Rule 5) not only within a unit, i.e. at the level of the spacing between individual chromatophore (and reflecting) elements, but also between units. A complete mathematical model for the role of lateral inhibition in morphogenesis was developed by Meinhardt & Gierer (1974). 4) The different appearance of the units from one part of the skin surface to another presumably reflects different local epigenetic conditions such as changes in the amount and distribution of inhibitor substance present or alterations in the rate of diffusion of a mor- phogen. Sometimes one gets a direct insight into the processes at work by inspection. The white spot on the mantle of octopuses which contains units that are separately innervated from the area immediately ahead of the white spot, starts as a forward-facing crescent of reflecting material behind which the local density of chromatophores is lower than in front and to the sides of the crescent. It is as if the crescent marked the edge of a partial bar- rier in the skin to morphogens (or activator substances) diffusing in a posterior direction, the polarity of the field being set with refer- ence to the head (Rule 2). The area ahead of the white spot goes characteristically dark during patterning, the area over the white spot goes pale. The arm white spots have the same polarity (with respect to the head) as the mantle white spots (i.e. their polarity is rotated 180° with respect to the antero-posterior axis). The straight pale areas that form part of the lateral flame of Loligo plei (Hanlon, 1982) where none but a few large dark chromato- phores are present are strips of local inhibition where chromatophore genesis was complete- ly suppressed after the early chromatophores became established. The strips of inhibition appear to be laid across the otherwise con- centric arrangement of the units. 5) Because one cannot ‘see’ the nerves, the details of innervation of the living skin are hardest to know but potentially the most fasci- nating for a neurobiologist. The connections given in Fig. 6 are based on analysis of video- tape pictures of natural patterning of lolignid squids before and after alcohol anesthesia (see Fig. 5). They employ Rule 6 (relating chromatophore size and age) and Rule 9 (separate responses) and agree with the general findings for octopus chromatophores: namely, that each age-class of chromato- PACKARD phore has its own nerve supply and intercon- nected network of radial muscle fibers (Froesch-Gaetzi & Froesch, 1977). We know that brain cell numbers increase rapidly with growth (Packard & Albergoni, 1970) and it is assumed that the pattern of recruitment of chromatophores into the skin is paralleled by recruitment of motoneurones into the brain and outgrowth of their fibres into the skin. Thus the units shown are chronological units (or chronomere = new word), innervating particular age-classes of chromatophores and are not restricted spatially to one morphologi- Cal (i.e. spatial) unit, but will pick up chromato- phore muscle fibers (waiting to be innervated) wherever they find them, within the framework of competition from other ingrowing nerve fibres. As each chromatophore has many muscle fibres this explanation would allow for multiple innervation; but only by nerves of similar age-class (unless chromatophores acquire further muscle fibres as development proceeds). Most important for the interpretation of pat- tern, the idea of the motoneurones as chrono- logical units that successively intersect the morphological array as it develops accounts in rather simple terms for much of the seem- ing spatial complexity of the physiological units. ACKNOWLEDGEMENTS | thank the following for facilities and assist- ance: the Director and staff of the Naples Zoo- logical Station, Italy, the Director and staff of the Bamfield Marine Station, B.C., Canada and Dr. W. H. Hulet, Dr. Roger Hanlon, John Forsythe and the Marine Biomedical Institute, Galveston, Texas, with whose generous help the squids were studied and Tom Gore, photographer and Media and Technical Serv- ices of the University of Victoria, B.C., Cana- da. Early phases of the research were sup- ported by a Scientific Grant-in-aid from the Royal Society, London. The paper was pre- pared during tenure of a Lansdowne Fellow- ship at the University of Victoria—where | en- joyed the hospitality of Dr. G. O. Mackie's la- boratory—and presented at the July 1980 meeting of the American Malacological Union by Dr. Roger Hanlon who is closely associ- ated with part of the study. | thank him and Dr. J. B. Messenger for their comments on the text. MORPHOGENESIS OF CHROMATOPHORE ‘UNITS’ 201 REFERENCES CITED FIORONI, P., 1965, Die embryonale Musterent- wicklung bei einigen Mediterranen Tintenfischar- ten. Vie et Milieu, sér. A: Biologie Marine, 16: 655-756. FLOREY, E., 1969, Ultrastructure and function of cephalopod chromatophores. American Zoologist, 9: 429-442. FROESCH, D. & MESSENGER, J. B., 1978, On leucophores and the chromatic unit of Octopus vulgaris. Journal of Zoology, 186: 163-173. FROESCH-GAETZI, V. & FROESCH, D., 1977, Evidence that chromatophores of cephalopods are linked by their muscles. Experientia, 33: 1448-1450. HANLON, R. T., 1982, The functional organization of chromatophores and iridescent cells in the body patterning of Loligo plei (Cephalopoda: Myopsida). Malacologia, 23: 89-119. MAYNARD, D. M., 1967, Organization of central ganglia. In WIERSMA, С. A. G., ed., Invertebrate nervous systems, p. 231-255, Chicago: Uni- versity Press. MEINHARDT, H. & GIERER, A., 1974, Application of a theory of biological pattern formation based on lateral inhibition. Journal of Cell Science, 15: 321-346. MESSENGER, J. B., 1974, Reflecting elements in cephalopod skin and their importance for camou- flage. Journal of Zoology, 174: 387-395. MESSENGER, J. B., 1979, The eyes and skin of Octopus: compensating for sensory deficien- cies. Endeavour, new ser., 3: 92-98. PACKARD, A., 1972, Cephalopods and fish: the limits of convergence. Biological Reviews, 47: 241-307. PACKARD, A., 1974, Chromatophore fields in the skin of the octopus. Journal of Physiology, 238: 38—40Р. PACKARD, A., 1978, Cracking the code of the camouflage patterns of the common octopus, Octopus vulgaris. Poster presented at the // European Neuroscience Meeting, Florence. PACKARD, A. & ALBERGONI, V., 1970, Relative growth, nucleic acid content and cell numbers of the brain in Octopus vulgaris. Journal of Experi- mental Biology, 52: 539-552. PACKARD, A. & HOCHBERG, F. G., 1977, Skin patterning in Octopus and other genera. In NIXON, M. & MESSENGER, J. B., eds. The Bi- ology of cephalopods, Symposium of the Zoo- logical Society of London, 38: 191-231. PACKARD, A. & SANDERS, G. D., 1971, Body pat- terns of Octopus vulgaris and maturation of the response to disturbance. Animal Behaviour, 19: 780-790. MALACOLOGIA, 1982, 23(1): 203-208 COMMENTARY ON THE INTERNATIONAL SYMPOSIUM ON FUNCTIONAL MORPHOLOGY OF CEPHALOPODS William H. Hulet, Rapporteur The Marine Biomedical Institute, University of Texas Medical Branch, 200 University Boulevard, Galveston, Texas 77550, U.S.A. Why do we study animals? Several months ago, while busily operating on a squid brain, Professor J. Z. Young posed this question to several of us onlookers and quickly gave a reply: “We study animals to learn principles.” Principles that direct the course of living things. Functional morphology is at the begin- ning of the rainbow whose ultimate reward is an understanding of life processes. Indeed, most aspects of cephalopod biology are scarcely beyond the level of discovery and description. Exceptions, and there are sev- eral, the squid giant axon for one, have more often than not been related to some extremely unique aspect of anatomical structure. After an anatomical discovery has been described, the next step of investigation might be called a study of action mechanics. | once examined the ultrastructure of the Portuguese-man-of- war (Physalia) nematocyst and carefully de- scribed how the heavily armored thread is dis- charged from the capsule by turning inside Out (Hulet et al., 1974). | was able to induce free nematocysts to discharge by adding household Clorox. The mechanics were clear, but to this day the mechanism of nematocyst ®. ‘+ FIG. 1. А Loligo pealei hatchling minutes after emerging from the egg. Most of the dorsal and ventral ciliary bands are still intact. Between the fins the tip of the mantle is an actively protrusible organ. (203) 204 HULET FIG. 2. In Loligo pealei and other loliginid squids the anchor-shaped hatching gland sits between the two fins. The shaft of the anchor extends anteriorly over the dorsum of the mantle. COMMENTARY ON CEPHALOPOD SYMPOSIUM 205 discharge in the living Physalia remains elu- sive. Renal parasites, posterior salivary glands and ectodermal epithelium were the topics of three papers on descriptive morphology. The basic facts of structure presented in these papers elicited lively discussions and specu- lations on function. Many helpful recommen- dations for future study came from the sym- posium participants and it is worth noting that considerable attention was given to the future role of laboratory-reared cephalopods in un- ravelling many facts of cephalopod biology. The two papers on cephalopod chromato- phores clearly show that knowledge of chro- matophore function is only beginning to be acquired. Before this symposium, | am sure that few squid specialists realized that the yel- low, red and brown chromatophores of at least one species of squid, Loligo plei, were positioned by color at distinctly different depths in the dermis. Drs. Hanlon and Packard have reached beyond descriptive anatomy and have beautifully demonstrated that chromatophores, iridophores and leuco- phores are the instruments for expression for the highest level of communicative behavior in cephalopods. A clear and minutely detailed account of a complex ventral photophore in the deep-sea squid Abralia trigonura highlighted the paper by Dick Young and John Arnold. Experiments on living animals aboard ship complemented the anatomical studies and brought to life many interesting aspects of function in such a complex organ. Without doubt, this work ful- fills the symposium's ideal of a presentation on functional morphology. FIG. 3. Enzyme-laden cells (C) at the center of the hatching gland are ready to rupture and dissolve the chorion. The bands of cilia (CB) are in constant and rapid motion. Loligo pealei ready to hatch. 206 RUBET Dr. Boletzky has made significant progress in explaining the dissolution of the cephalo- pod mantle-fin-shell complex, and in his pre- sentation he described the anatomical sepa- ration of the developing fins from the onto- genetically receding shell complex (Fig. 1). Concomitantly, the internalization of the shell complex in developing coleoid cephalopods has freed the mantle for almost limitless muscular growth and made way for the emer- gence of the giant fiber system of nerves. The closed pore of the submerged shell complex gives way and is covered over by Hoyle's organ or hatching gland. The anchor-shaped appearance of the gland is characteristic of decapods (Fig. 2). The cells of the hatching gland (Fig. 3) are swollen with proteolytic enzyme that digest the chorion and possibly liquefy the mucinous jelly that surrounds the egg and is the last barrier to freedom for the squid hatchling. Once free of the egg, effec- tive passage through the jelly of the capsule is by means of numerous bands of motile cilia whose synchronous unidirectional beat pro- pels the hatchling outward (Figs. 4, 5) (Boletzky, 1979). Free-swimming squid hatchlings shed these cilia-bearing bands of cells within hours after hatching (Arnold & Williams-Arnold, 1980). Not all ciliated cells with the classical “9 + 2” arrangement of microtubules (Fig. 6), so abundant in the skin of loliginid hatchlings, are lost after the animal leaves the egg capsule. The hair cells of the statocyst are classical examples of functioning ciliated cells in adult cephalopods. Even in the skin, however, cili- ated cells abound in mature animals (Fig. 7). Some are thought to be chemoreceptors (Emery, 1975) and others found on the mantle and fins might function as mechanoreceptors not too unlike the lateral line system of fishes. The foregoing comments on structure and function of motile cilia are useful for a con- cluding remark to end this excellent sympo- sium on functional morphology of cephalo- pods. My final thought is a reminder for all of FIG. 4. The bands of moving cilia (CB) that nearly cover the mantle are shed within hours after hatching. Scanning electron micrograph. Loligo pealei. COMMENTARY ON CEPHALOPOD SYMPOSIUM us to exercise continuous vigilance for evi- dence of faulty repair in cells, tissues and organs. Cells can sustain a non-lethal injury, repair themselves and later express this faulty repair as a serious disturbance of function. So often the pursuit and study of an abnormality gives entry to knowledge of normal function. As illustrated in Fig. 6, the nine doublets of microtubules in all cilia are connected by arms containing the protein dynein which re- sembles the contractile protein myosin. Dynein acts as an adenosine triphosphatase (ATPase) and causes ciliary motion by form- ing temporary bridges between adjacent cili- ary tubules. Although ciliary motion is far more complex than the bridging function of dynein, the absence of dynein arms destroys 207 a coordinated ciliary beat and leaves the cilia immotile. One investigator interested in cilia was to discover the absence of dynein arms and unravel the mystery of why some children are born with the heart on the right side and suffer years of repiratory infections (Afzelius, 1976). Faulty cilia without dynein arms are to blame for situs inversus in the growing em- bryo and chronic bronchiectasis in later life. Kartagener described this debilitating illness nearly half a century ago (Kartagener, 1933). The discovery by Afzelius intensified work on the fine structure and chemistry of ciliary mo- tion. There are now many known causes of the immotile-cilia syndrome. All were waiting to be discovered. C FIG. 5. A section through a whole hatchling near the fins shows the extent of the ciliation. After hatching most of these cells (C) can be found free in the seawater medium spinning around like free-living ciliates. 208 HULET FIG. 6. A kinocilium from a Loligo pealei hatchling. The “9 + 2” doublets and two central microtubules are clearly visible. Dynein arms (DA), radial spokes (RS), nexin links (NL) are a few of the known struc- tural faults that can produce immotile-cilia syn- drome and its serious consequences. A transmis- sion electron micrograph. ACKNOWLEDGEMENTS | am grateful to Drs. Voss, Hanlon, Hixon and Roper who encouraged me to learn more about cephalopods than | ever thought pos- sible. | am again indebted to Margarita Villoch for the excellent scanning and transmission electron micrographs. LITERATURE CITED AFZELIUS, B., 1976, A human syndrome caused by immotile cilia. Science, 193: 317-319. FIG. 7. This and many other similarly ciliated cells were found on the mantle of adult specimens of Loligo plei. ARNOLD, J. M. & WILLIAMS-ARNOLD, L. D., 1980, Development of the ciliature pattern on the embryo of the squid Loligo pealei: a scanning electron microscope study. Biological Bulletin, 159: 102-116. BOLETZKY, S. v., 1979, Ciliary locomotion in squid hatching. Experientia, 35: 1051-1052. EMERY, D. G., 1975, The histology and fine struc- ture of the olfactory organ of the squid Lolli- guncula brevis Blainville. Tissue & Cell, 7: 357- 367. HULET, W. H., BELLEME, J. L., МУЗ, Е. & LANE, C. E., 1974, Ultrastructure of Physalia nematocysts. In: HUMM, H. J. & LANE, C. E., eds., Bioactive Compounds from the Sea, Dekker, New York, 1: 99-113. KARTAGENER, M., 1933, Zur Pathogenese der Bronchiektasien: Bronchiektasien bei Situs Viscerum Inversus. Beiträge zur Klinik der Tuberkulose und Spezifischen Tuberkolose- Forschung, 83: 489-501. WHY NOT SUBSCRIBE TO MALACOLOGIA? ORDER FORM Your name and address MES er р ae Send U.S. $17.00 for a personal subscription (one volume) or U.S. $27.00 for an institutional subscription. Make checks payable to “MALACOLOGIA.” Address: Malacologia, Academy of Natural Sciences Nineteenth and the Parkway, Philadelphia PA 19103, U.S.A. UNITAS MALACOLOGICA Eighth International Malacological Congress Organizing Committee: Laszlo PINTER (president); Endre KROLOPP (secretary); Karoly DATE: PLACE: TOPICS: LANGUAGES: EXCURSIONS: BABA; Tamas DOMOKOS; Levente FUKOH; Gyula KOVACS; Laszlo MERENYI; Ede PETRO; Andor RICHNOVSZKY; Andras VARGA First Circular August 29—September 3, 1983. Budapest, Hungary. The Congress is open to any topic of malacology (recent and fossil). Further- more, in case of sufficient number of people, we organize Colloquia in the fol- lowing subjects: 1. Quaternary malacology and fauna genesis 2. Non-marine mollusc biogeography 3. Sphaeriidae We organize meetings for the EIS and for museum curators, also the General Assembly of the UM. English, French and German. Preceding, during and after the Congress we shall organize collecting tours and excursions. Further details shall be issued in the Second Circular. Laszlo Pinter Hungarian Natural History Museum Budapest Baross u. 13. H-1088 Hungary 23(1) 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 1982 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 ¥2 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 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 (&) 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. 23, personal subscriptions are U.S. $17.00 and institutional subscriptions are U.S. $27.00. For information on Vol. 24, address inquiries to the Subscription Office. VOL. 23, NO. 1 MALACOLOGIA CONTENTS С. J. VERMEIL Gastropod shell form, breakage, and repair in relation to predation by the crab.EalappR 2... 2 2 sie bce intel es dd DER eR NO ем о S. M. LOUDA & K. R. MCKAYE Diurnal movements in populations of the prosobranch Lanistes nyassanus at Cape Maclear,. Lake Malawi; Africa... OT so scree AN K. C. EMBERTON, Jr. Environment and shell shape in the Tahitian land snail Partula otaheitana ...... C. THIRIOT-QUIEVREUX & R. S. SCHELTEMA Planktonic larvae of New England gastropods. V. Bittium alternatum, Triphora nigrocincta, Cerithiopsis emersoni, Lunatia heros and Crepidula plana ......... P. W. KAT Reproduction in a peripheral population of Cyrenoida floridana (Bivalvia: Gyrenaldıdae) IA A A Se eo CORRE J. SEUGE & R. BLUZAT Effets des conditions d'éclairement sur la croissance de Lymnaea stagnalis (Gasteropode:Pulmome) ico oe os ds aia о SN A. R. PALMER Growth in marine gastropods: a non-destructive technique for independently measuring: shell and body "weight . =: MA "een Bakes Se C. M. HUMPHREY & R. L. WALKER The occurrence of Mercenaria mercenaria form notata in Georgia and South Carolina: calculation of phenotypic and genotypic frequencies ................. LETTERS TO THE EDITORS F. G. Thompson. On sibling species and genetic diversity in Florida Goniobasis ......... SM Chambarssheply 2.2... 30 A A ee BO Seite SE PRES AMERICAN MALACOLOGICAL UNION SYMPOSIUM: FUNCTIONAL MORPHOLOGY OF CEPHALOPODS Louisville, Kentucky 21 July, 1980 EEE ROPER Introduction: A a Е AA ARA EEE ae R. T. HANLON The functional organization of chromatophores and iridescent cells in the body patterning of Loligo plei (Cephalopoda: Myopsida) ....................... F. G. HOCHBERG, Jr. The “kidneys” of cephalopods: a unique habitat for parasites .................. ; R. E. YOUNG & J. M. ARNOLD The functional morphology of a ventral photophore from the mesopelagic squid, Abralla' ingonurai: 2 rad ern. shades a fede die ei ata de ah NN S. V. BOLETZKY Developmental aspects of the mantle complex in coleoid cephalopods .......... C. T. SINGLEY Histochemistry and fine structure of the ectodermal epithelium of the sepiolid Squid EUprvirina: SCOIDPES: oct aaa wont ones dt passed wale HR NES A. PACKARD Morphogenesis of chromatophore patterns in cephalopods: are morphological and physiological ‘units’ the Same? of. =» «4 apo ain ande een sa AA W. H. HULET Commentary оп the Symposium‘... =. 000 RER bape rele at data ora eats PRE 23 ss... ss. OC ane Peet! . 198 WWD. LAA. LUE LIBRARY 1983 MAR - 8'233 te AA VA UNIVERSITY. (| ternationale Malakologische Zeitschrift MALACOLOGIA Editors-in-Chief: GEORGE M. DAVIS Editorial and Subscription Offices: Department of Malacology The Academy of Natural Sciences of Philadelphia Nineteenth Street and the Parkway Philadelphia, Pennsylvania 19103, U.S.A. Associate Editors: JOHN B. BURCH University of Michigan, Ann Arbor ANNE GISMANN Maadi, A. R. Egypt MALACOLOGIA is published by the INSTITUTE OF MALACOLOGY (2415 South Circle Drive, are: CHRISTOPHER J. BAYNE, President Oregon State University, Corvallis KENNETH J. BOSS Museum of Comparative Zodlogy Cambridge, Massachusetts JOHN B. BURCH MELBOURNE R. CARRIKER University of Delaware, Lewes GEORGE M. DAVIS, Secretary and Treasurer PETER JUNG, Participating Member Naturhistorisches Museum, Basel, Switzerland OLIVER E. PAGET, Participating Member Naturhistorisches Museum, Wien, Austria Copyright © Institute of Malacology, 1983 Ann Arbor, Michigan 48103, U.S.A.), the Sponsor Members of which (also serving as editors Editorial Assistants: MARY DUNN DAVID WATT x ROBERT ROBERTSON CLYDE: EXROPER Smithsonian Institution Washington, D.C. Syracuse ne New York NORMAN F. SOHL United States Geological Survey Washington, D.C. SHI-KUEI WU, President-Elect — J FRANCES ALLEN, Emerita Environmental Protection Agency Washington, D.C. ELMER G. BERRY, Emeritus Germantown, Maryland + 1983 EDITORIAL BOARD J. A. ALLEN Marine Biological Station, Millport, United Kingdom E. E. BINDER Muséum d'Histoire Naturelle Genève, Switzerland A. J. CAIN University of Liverpool United Kingdom P. CALOW University of Glasgow United Kingdom A. H. CLARKE, Jr. Mattapoisett, Mass., U.S.A. B. C. CLARKE University of Nottingham United Kingdom 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. 5 GOULD Harvard University Cambridge, Mass., U.S.A. A. V. GROSSU Universitatea Bucuresti Romania T. HABE Tokai University Shimizu, Japan A. D. HARRISON University of Waterloo... ~ Ontario, Canada LIBRAR K. HATAI en Tohoku University MAN Sendai, Japan B. HUBENDICK Naturhistoriska Museet Góteborg, Sweden S. HUNT University of Lancaster United Kingdom 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 Kobenhavn, Denmark A. J. KOHN University of Washington Seattle, U.S.A. Y. KONDO Bernice P. Bishop Museum Honolulu, Hawaii, U.S.A. J. LEVER Amsterdam, Netherlands A. LUCAS Faculté des Sciences Brest, France N. MACAROVICI Universitatea “Al. I. Cuza” lasi, Romania C. MEIER-BROOK Tropenmedizinisches Institut Tubingen, Germany (Federal Republic) H. K. MIENIS Hebrew University of Jerusalem Israel J. E. MORTON The University Auckland, New Zealand R. NATARAJAN Marine Biological Station Porto Novo, India J. OKLAND University of Oslo Norway T. OKUTANI National Science Museum Tokyo, Japan W. L. PARAENSE Instituto Oswaldo Cruz, Rio de Janeiro Brazil J. J. PARODIZ Carnegie Museum Pittsburgh, U.S.A. W. 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 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. F. 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. ТОЕРОЕЕТТО Societa Malacologica Italiana Milano R. D. TURNER Harvard University Cambridge, Mass., U.S.A. 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, 1983, 23(2): 209-219 BULIMULID LAND SNAILS FROM THE GALAPAGOS: 1. FACTOR ANALYSIS OF SANTA CRUZ ISLAND SPECIES! Guy Coppois? and Claude Glowacki3 ABSTRACT A biometrical comparison of Santa Cruz Island bulimulid land snails is made using factor analysis (7 variables, 140 specimens). The analysis points out two main factors, one represent- ing the overall dimensions of the shell, the other its degree of slenderness or plumpness. In a practical way, these two factors can be represented by the height of the shell and its terminal apical angle. The analysis allows the separation of the specimens into 28 taxa, mostly represent- ing different species, but it isolates some cases of infraspecific variations. Key words: Bulimulidae; Bulimulus; Naesiotus; biometry; factor analysis; speciation; Galapagos. INTRODUCTION Long term research on the speciation of the bulimulid land snails in the Galapagos archi- pelago is undertaken. It started by a three years stay in the islands where we were able to gather a large collection of these snails per- taining to most of the described species as well as undetermined taxa (species or varia- tions, some obviously not described). Collec- tions were made in most of the islands of the archipelago, but for this paper special atten- tion is directed to the malacofauna of Santa Cruz (Indefatigable) Island. Many different species can be found on this island. Some are extinct or on the way to extinction, but others are still present in numerous and healthy pop- ulations. They occupy various complex habi- tats often closely related to well defined plant zonation (Wiggins & Porter, 1971; Smith, 1966). Thanks to this species richness (about sixty species in one genus in the whole of the archipelago—Dall & Ochsner, 1928; Smith, 1966), the Galapagos Bulimulidae constitute an exceptionally suitable group for the study of speciation in invertebrates. Some species are easily identified by their obvious morphological characters; in other cases, less obvious morphological characters or wide intraspecific variations make it more difficult to separate specimens belonging to closely related species. Moreover, original descriptions are most of the time very brief and incomplete and almost never give any in- formation on intraspecific variations (Dall & Ochsner, 1928; Smith, 1972; Vagvolgyi, 1977). This makes a certain degree of uncer- tainty in species identification unavoidable, noticed even in large reference collections like that of the California Academy of Sci- ences at San Francisco. For many species one must compare the specimens to the types (often empty shells, when available) to obtain a correct identification. We had the opportu- nity to compare our material with the collec- tions of the California Academy of Sciences, the British Museum (Natural History) and the Naturhistoriska Riksmuseet of Stockholm. It has not always been possible to ascertain the correct name for all the taxa mentioned in this study. For clarity of the text, each taxon has been given a number. The identified taxa are listed with their corresponding numbers in the Appendix. The others, unnamed, will be sub- ject to further investigations. The frequent difficulty encountered in deter- mining with precision the bulimulid species of the Galapagos led us to analyze very carefully the material gathered by using factor analysis, a method giving in this case objective criteria based on the biometry of the shells and allow- ing an easier comparison of the various spe- cies in this group—limited in a first step to the species inhabiting Santa Cruz Island. In the first stage, only the biometrical variables of the shells were studied (Table 1). Some quali- Contribution no. 285 of the Charles Darwin Foundation for the Galapagos Islands. 2Zoologie Systématique, CP 160. 3Informatique Théorique, CP 212. 2-3Universite Libre de Bruxelles, 50, av. F.D. Roosevelt, 1050 Bruxelles, Belgium. (209) 210 COPPOIS AND GLOWACKI TABLE 1. Biometric measurements used in the factor analysis and their symbols in the following tables. Mean, standard deviation, minimum and maximum values observed for the repeated control measurements (n = 31) of one shell over a period of two years. oo) eee Soe Al Symbol Mean mm Standard deviation Min. mm Max. mm Height of shell H. shell 15.191 0.117 15.00 15.40 Maximum diameter of shell D. max. 7.976 0.082 7.83 8.12 Height of aperture H. apert. 5.608 0.104 5.42 5:77 Width of aperture W. apert. 4.925 0.072 4.82 5.09 Height of last whorl H. |. whorl 9.030 0.110 8.80 9.28 Terminal apical angle T. ap. angle 44.097 0.396 43.00 45.00 Embryonic apical angle E. ap. angle 72.484 0.926 71.00 74.00 CE De ot ou qe à D NS tative factors, like colour, type of apex and umbilicus, superficial characteristics of shells, presence or absence of teeth in the aperture, or anatomical peculiarities, will be examined later on. This study, focused on shell charac- ters alone, was necessary in order to obtain a correct identification of our specimens by comparing them with museum material and type-specimens which are almost only dry shells. MATERIALS AND METHODS SAMPLING AND MEASUREMENTS. The specimens used in this study were gathered between July 1973 and July 1976 by method- ical sampling on the whole surface of Santa Cruz Island. These samplings were made by picking up the snails found in quadrats (size: 1 m? and 0.25 m?) chosen at random in some areas, but also by local sampling in all the zones of the island or by quadrats regularly spaced along transects. One can safely as- sume that they represent all the bulimulid taxa (species or subspecies) that can be found on the island. The specimens belonging to each taxon were selected from lots corresponding to the various collecting stations. A detailed list of localities will be published later. For the purposes of factor analysis, only adult speci- mens were selected at random in restricted number (5 for every taxon) after having dis- carded juveniles and damaged specimens from the lots. Every shell was drawn using the Wild M5 stereoscopic microscope fitted with a camera lucida. Shells and the corresponding drawings are numbered individually, and the scale of every drawing was noted. This made possible, at any stage of the analysis, finding again individual specimens or drawings every time an observation had to be completed or verified. To facilitate drawings, every shell was placed in a small, flat-bottomed container filled with fine sand, and in a standard posi- tion, the three following points being put ina single horizontal plane (Fig. 1): 1) the apex (A), 2) the point marking the maximum ad- vance of the parietal area on the columellar surface of the previous whorl (B), 3) the point (C) corresponding to (B) on the outer edge of the aperture, (B) and (C) being situated on the same plane perpendicular to the apparent axis of the shell. It should be noted that in this position the surface determined by the aper- ture is seldom also included in a horizontal plane. In practice the standard position was obtained by successively focusing on points A, Band С at maximum magnification (50). At this magnification the depth of field is very narrow and the plane determined by these three points can be assimilated to the hori- zontal plane. The approximation was tested by repeated positioning of a shell and of a small object presenting a triangular plane sur- face on one of its sides (for the accuracy of the method see further on). The three points considered could be lo- cated easily whatever the type of shell; shapes may of course vary considerably from one species to another. Once a shell has been drawn, the terminal apical angle, the sides of which are tangents to the first and last whorls of the spiral, was traced on the draw- ing. Then we added an imaginary axis, pass- ing through the summit of the terminal apical angle and through the point (D) on the colu- mellar edge which was nearest the actual axis of the shell. This imaginary axis is never iden- tical with the actual axis. In optimal cases, the latter goes through the center of the colu- mella. All measurements were made along straight lines either parallel or perpendicular to the imaginary axis, from the axis to the outermost points of the shell and of the aper- ture. These measurements appear on Table BIOMETRY OF GALAPAGOS BULIMULIDS 211 T.ap. angle. FIG. 1. Drawing of a shell in a standard position for measurements: Bulimulus (Naesiotus) cavagnaroi Smith, 1972, xanthic form (taxon 2, paratype n° GC.652 of B. gilderoyi Van Mol, 1972; H. shell: 22.14 mm, D. max.: 13.78 mm). Legend: see text and Table 1. 1. It should be noted that the height of the last whorl of the shell was always measured along the imaginary axis. An “embryonic apical angle” was also constructed on every draw- ing; its sides were tangents to the first whorls of the shell. In most cases, this was found to be identical to the “maximal aperture apical angle.” As the sides of the embryonic apical angle could only be determined by closely spaced points on the drawing, its measure- ment was necessarily less accurate than that of other angles. The angles were measured with a protractor with an approximation of one degree. Length measurements were made on the drawings, using a sliding calliper (instru- mental precision: + 0.05 mm). To verify the accuracy of this method, measurements of the same shell (ref. GC 2, taxon 24) were taken 31 times at several months’ intervals over a period of two years. The mean, stan- 212 dard deviation as well as the minimum and maximum value observed for each biometric measurement for these controls are present- COPPOIS AND GLOWACKI TABLE 4. Percentage of total variance represented by each factor. ed in Table 1. The data thus collected were Factor Percent of Variation Cumulated percent reproduced on cards to be analysed by a CDC 6000 computer. L hs I FACTOR ANALYSIS. This type of analysis al- 3 7 99 1 lows study of intercorrelations between vari- 4 5 99 6 ables. Sensu stricto, this term applies only to 5 2 99.8 the treatment of aleatory variables which are 6 4 99.9 supposediy known in terms of probability. 7 1 100.0 This technique is essentially used to deter- mine, among a group of variables, those most intercorrelated, in order to eliminate redun- dancies in explanatory variables, either by eliminating some of the variables, or by creat- ing new variables which are linear combina- tions of the original variables. The sub-group of variables thus defined can be used for fur- ther analysis. In our case, the resulting sub- groups of variables or factors enabled us to compare the taxa found on Santa Cruz Island on biometrical data alone. One hundred and forty shells were used for the analysis (five for each taxon, see Appendix); their mean and the general variance, shown in Table 2, were used to reduce the raw biometric variables in order to estimate the correlation matrix of these variables (Table 3). TABLE 2. General mean, variance and standard deviation (140 shells). The next step was the factor analysis prop- er (Lebart & Fenelon, 1971, Dagnelie, 1975; Nie et al., 1975). The whole of the observation realized by the reduced variables can be visu- alised as a cluster of points centered in a multidimensional space. In this space, factor analysis will determine a main axis corre- sponding to the maximal variance of the clus- ter (axis F1). Other axes F2, Fs, Fs... corre- sponding to the next maximal variance can be determined perpendicularly to this axis F1. These oriented axes or vectors are known as “factors.” Table 4 shows the percentage of the total variance represented by each factor (shown in decreasing order). It will be noticed that factor 1 explains 77.5%, and factor 2, 20.9% of the total vari- ance. Together factors 1 and 2 explain 98.4% of this variance, the total of other factors ac- counting only for 1.6%. This means that in our multidimensional space, most of the observa- Standard Е x Variable Mean Variance deviation tions are clustered nearly in a single plane. Thus the geometry of Santa Cruz Island buli- H. shell 13.8862 15.17258 3.8952 mulid land snails is practically a bidimensional D. max. 7.8237 8.73202 2.9550 phenomenon, two axes being sufficient to H. apert. 5.4159 4.03206 2.0080 characterize it almost completely. An impor- W. apert. 4.8431 3.24144 1.8004 tant element of factor analysis lies in the inter- LA = un. “en ae Een pretation of the selected factors or, more pre- E ap angle 629143 28910200 17.0030 cisely, of the subspaces they determine. Gen- TABLE 3. Matrix of correlations between variables. erally the selected factors beyond the first two H. D. H. W. Н. 1. Е ар: Е. ар. shell max. apert. apert. whorl angle angle H. shell 1.00000 D. max .85296 1.00000 H. apert .92960 .94467 1.00000 W. apert .87934 .98505 .96935 1.00000 H. |. whorl 92792 97122 .98476 .97596 1.00000 T. ap. angle .06582 ‚56377 .35290 .50053 40312 1.00000 E. ap. angle .28562 .70446 53912 .66160 58314 .93391 1.00000 BIOMETRY OF GALAPAGOS BULIMULIDS TABLE 5. Correlations between initial factors and variables. Factor 1 Factor 2 H. shell .85081 —.51395 О. тах .99309 —.01170 H. apert .95995 — .23992 W. apert .99003 —.08080 Н. |. whorl .97700 —.19162 T. ap. angle .57075 81172 E. ap. angle 725 66386 TABLE 6. Correlations between factors and уап- ables after a Varimax rotation. Factor 1 Factor 2 H. shell .99289 —.04691 D. max .87889 46251 H. apert .95840 24604 W. apert .90910 40029 H. |. whorl .95040 29663 T. ap. angle .11546 98555 E. ap. angle ‚32393 93027 are not easy to interpret. Table 5 shows the existing correlations between factors 1 and 2 and the original variables. At this stage, it is clearly impossible to give a real meaning to the plane which has been defined. To make it possible we shall rotate factors 1 and 2 around the origin while keeping them in the same plane, in order to correlate them as well as possible with a sub-group of reduced variables. This was, in the present case, ef- fected by the well known VARIMAX method (Nie & al., 1975). Table 6 shows the correla- tions of the new factors with the observed variables. It will be clearly seen that factor 1 is closely correlated to: the height of the shell, the maximum diameter of the shell, the height of the aperture, the width of the aperture, the height of the last whorl; and factor 2 to: the terminal apical angle, the embryonic apical angle. Thus factor 1 represents mainly a measure- ment of the overall dimensions of the animal; factor 2 shows its degree of slenderness or plumpness. Since factor 1 is most closely related to the height of the shell, and factor 2 to the terminal apical angle, these two measurements can be used to describe accurately the geometry of the shell, on which they can easily be taken. The scores on factors 1 and 2 corresponding 213 TABLE 7. Coefficients used for calculation of indi- vidual factor scores. 1=1 Г-2 Ch shel 30506 —.23386 CD max 16488 08018 CH apert 23369 _ 05975 W.apert 18685 .03843 Ch.1.whorl 22075 -.02926 CT ap.angle 17127 53747 СЕ ap.angle 09774 46248 A EP DCE IE I EN NE EEE EE en to each observation were calculated from the coefficients given in Table 7 using the method given by Nie et al. (1975: 489). RESULTS AND DISCUSSION The distribution of the observations in the plane of factors 1 and 2 is shown on Fig. 2. On Fig. 3, the raw biometric data corresponding to each individual analyzed have been carried over into the plane of the variables “height of shell” and “terminal apical angle.” For clarity, in Figs. 2 and 3, the five observations pertain- ing to each taxon are enclosed in a cluster delimited by a line, a cross indicates the mean position calculated from the corresponding scores or data. Sketches of the corresponding shell shapes have been added on Fig. 2. One will readily notice the similarity of these two graphs, which show a total of 28 taxa identi- fied from Santa Cruz Island. Two extraneous taxa were added on Fig. 2 (indicated by ar- rows). They represent one species living on Isabela Island (B. (N.) pallidus Reibisch, 1892, taxon 51, from Alcedo volcano), and one found on Champion Island (B. (N.) plano- spira Ancey, 1887, taxon 49). The data per- taining to these taxa were not included in the global factor analysis to avoid any interfer- ence with the Santa Cruz data in calculations. Their position on Fig. 2, calculated from the scores of Table 7, are thus indicative. They are included to show that the variation in shapes of the bulimulids becomes even wider when taxa from other places in the archipel- ago are studied. Factor analysis shows that in most cases these taxa can be identified even if only the biometric data of the shells are used. Obvi- ously, an analysis including some qualitative characters of the snails will make possible an 214 Е> COPPOIS AND GLOWACKI FIG. 2. Factor analysis: distribution of the observations in the plane of factor 1 and 2. The clusters of points enclose the five observations corresponding to each taxon, the mean of each cluster is marked by a cross and sketches of the corresponding shells are represented for the 28 taxa from Santa Cruz Island. Two extraneous taxa (n° 49 and 51) were added (see text, taxa listed in the Appendix). even better separation between the different taxa (Coppois & Glowacki, in preparation). For example, taxa 9 and 12 are situated on largely overlapping areas of Fig. 2. But taxon 9 specimens, B. (N.) eos Odhner, 1950, have a light, smooth shell colored white or light brown, and no tooth in the aperture, or a very light parietal bump. Specimens of taxon 12, B. (N.) scalesiana Smith, 1972, have a thick, ro- bust shell with an irregular bumpy surface, es- pecially on the last whorl. The colour is white with a pink, or sometimes grayish apex; there is a tooth on the parietal wall in fully devel- oped adults. So, in spite of their superposition on Fig. 2, these two taxa can be separated beyond any doubt. After examination of the BIOMETRY OF GALAPAGOS BULIMULIDS 215 T. ap. angle. 60 40 30 20 10 10 15 20 mm. H. shell. FIG. 3. Raw data, height of shell by terminal apical angle (see text). For clarity the five observations corresponding to each taxon are enclosed in a cluster of points delimited by a line. The mean of each cluster is marked by a cross. The identified taxa are listed in the Appendix. type specimens of these two species and from our own conclusions there is good evi- dence that these two species are not synony- mous as stated by A. G. Smith and A.S.H. Breure (Smith, 1974). INTRASPECIFIC VARIATION. Although per- formed on a restricted number of snails (5 for each taxon) the factor analysis already pro- vides interesting information on intraspecific variations. It is obvious that the clusters of points representing each taxon could be more extended if more cases were entered in the analysis, but the general pattern of Fig. 2 would only be slightly modified. Some prelimi- nary tests have been made with samples of as many as 30 specimens, for taxa 1, 2 and 3; 216 the modifications are not significant. In taxon 6, B. (N.) akamatus Dall, 1917, smaller indi- viduals are plump, while larger ones are more slender, although all specimens measured were adults. Other clusters similarly stretched on an upper left-lower right line can readily be observed. Taxa 24 and 70 are clearly differentiated from their neighbours on Figs. 2 and 3 by their characteristic shell with a widely open umbili- cus and the apex marked by straight axial rib- lets, the spaces between these riblets show- ing fine spiral striae (like taxa 10, 16, 18, 19, 22, 23, 49, 51, 56, 68, 81). Their neighbouring taxa: 6, 9, 11, 12, 17, and also taxa 1, 2, 3, 4, 5, 7, 13, 15, 21, 60, 65, 69 have another kind of apex with fine undulating, sometimes con- verging axial riblets. Although one can easily separate the shells of taxon 24 from those of taxon 70, these last ones being more slender and having a small bump on the columellar surface, both taxa are in fact representatives of one species B. (N.) tanneri Dall, 1895, which shows a considerable intraspecific vari- ation when samples coming from different lo- calities are studied. These shells can vary in size and shape (plump or more slender), hav- ing or not a small bump on the columellar surface. This can be observed in the field even within short distances: the localities where the specimens of taxa 24 and 70 were collected are distant by less than 1km (Coppois & Glowacki, 1982). Colour variations occur in two species, B. (N.) eos Odhner, 1950 (taxon 9), for instance, has mixed populations of two colour forms: light brown or white shells. For B. (N.) cavag- naroi, there are three main colour forms that are easily separated: normal brown (taxon 1), “xanthic” yellow (taxon 2) and “white unband- ed” shells (taxon 3) (Smith, 1972). RELATIONSHIPS BETWEEN SHELL SHAPES AND HABITAT. Although ecology is not the main purpose of this paper, it seems interest- ing to make some preliminary remarks on some striking correlations that seem to exist between shell shape and habitat (the ecologi- cal approach of our work is still in prepara- tion). To give a better visualization of the vari- ation of shell shapes over the whole group of bulimulid taxa found on Santa Cruz Island alone, we shall refer to Fig. 4. On this figure, the drawings of the shells are centered on the crosses which locate the mean position of each taxon in the factor analysis (similar to Fig. 2). We shall also refer to the observations COPPOIS AND GLOWACKI of Smith (1966) and to the well known plant zonation (Wiggins 4 Porter, 1971; Van Der Werff, 1978). Among the taxa found in the dryer zones (Arid and lower Transition zones), those living in the open, on tree trunks and vegetation are small (taxa 7, 22, 56 and to some extent 81), with slender shells, a small aperture (the latter characteristic being consistent with the necessity to minimize evaporation) and no teeth. Those living under rocks are bigger (taxa 6, 10, 11, 23, 24, 60, 65, 70, 81), with sturdier shells and a relatively wider aperture. In humid zones (upper Transition zone, Scalesia forest, summits zone), the variation in size is definitely wider. The biggest forms (taxa 1, 2, 3 and 5) are all found in the wettest habitats and live in the humus layer on the ground, as taxon 17 from the summits zone which is slightly smaller. The smaller forms (taxa 13, 15, 19, 21 and 69) may spread to less humid zones (e.g. moist forest of the upper Transition zone); they also live hidden in the humus layer. All these taxa are char- acterized by a very globular shape and a me- dium wide aperture which is always reduced by the presence of teeth. Taxa 18 and 68, which live in the humus layer like the above ones, but in relatively drier zones, are small, with a less plump shell shape and no teeth. This clearly places them in an intermediate category, which tallies with the intermediate situation of their habitat between dry and hu- mid zones. A similar remark applies to taxa 4, 9, 12 and 16 which live exposed on the vege- tation like taxa 7, 22 and 56 described above, but in definitively moister habitats. CONCLUSION This preliminary study shows that factor analysis can be usefully applied to the sepa- ration of the bulimulid land snail taxa of the Galapagos Islands and to the identification of species. Although focused on the shells' bio- metry alone, this method is a suitable tool for comparing our specimens with museum ma- terial and type-specimens which are mainly dry shells. Further studies will include other morphological characters of the shells, ana- tomical and ecological observations and, we hope, will help to clarify the taxonomy of this confusing group of snails. BIOMETRY OF GALAPAGOS BULIMULIDS 217 FIG. 4. Variation in shape of the shells in the bulimulid taxa found on Santa Cruz Island. Factor analysis (see text). Sketches of the shells are centered on the mean position of each cluster of points. The taxa are listed in the Appendix. ACKNOWLEDGMENTS We thank the Charles Darwin Foundation for the Galapagos Islands, the Parque Naci- onal Galapagos' authorities, the Belgian Ministère de l'Education Nationale and Pro- fessor Jean Bouillon who made the field work possible. We are grateful to Dr. Barry Roth for his kind welcome, his help and facilities at the California Academy of Sciences, San Fran- cisco, and to Professors Guy Louchard and Jean-Jacques Van Mol for their advice and criticism during this work. We also thank Pro- fessor John F. Peake and Dr. Peter Mordan for access to the British Museum (Nat. Hist.) collections, and to Professor À. Andersson for 218 the loan of the Galapagos bulimulid collection of the Naturhistoriska Riksmuseet of Stock- holm. Special thanks go to Chantal De Ridder who patiently helped in the field work and pro- vided advice and encouragement. LITERATURE CITED ANCEY, M. C. F., 1887, Nouvelles contributions malacologiques. Bulletin de la Société Malaco- logique de France, 4: 293-299. BREURE, A. S. H. & COPPOIS, G., 1978, Notes on the genus Naesiotus Albers, 1850 (Mollusca, Gastropoda, Bulimulidae). Netherlands Journal of Zoology, 28: 161-192. COPPOIS, G. & GLOWACKI, C., 1982, Factor analysis of intraspecific biometrical variations of Bulimulus (Naesiotus) tanneri (Pulmonata, Buli- mulidae) in the Galapagos Islands. Proceedings of the Seventh International Malacological Con- gress. Malacologia, 22: 495-497. DAGNELIE, P., 1975, Analyse statistique a plusieurs variables. Presses Agronomiques de Gembloux, 362 p. DALL, W. H., 1893, Preliminary notice of new spe- cies of land-shells from the Galapagos Islands, collected by Dr. G. Baur. Nautilus, 7: 52-56. DALL, W. H., 1895, New species of shells from the Galapagos Islands. Nautilus, 8: 126-127. DALL, W. H., 1896, Insular landshell faunas, espe- Cially as illustrated by the data obtained by Dr. G. Baur in the Galapagos Islands. Proceedings of the Academy of Natural Sciences of Philadel- phia, 1896: 395-460, pl. 15-17. DALL, W. H., 1900, Additions to the insular land- shell faunas of the Pacific coast, especially of the Galapagos and Cocos Islands. Proceedings of the Academy of Natural Sciences of Philadel- phia, 1900: 88-106, pl. 8. DALL, W. H., 1917a, Preliminary descriptions of new species of Pulmonata of the Galapagos Islands. Proceedings of the California Academy of Sciences, (4)2: 375-382. DALL, W. H., 1917b, New Bulimulus from the Gala- pagos Islands and Peru. Proceedings of the Bio- logical Society of Washington, 30: 9-12. DALL, W. H. 8 OCHSNER, W. H., 1928, Landshells of the Galapagos Islands. Proceedings of the California Academy of Sciences, (4)17: 141- 185. LEBART, C. & FENELON, J.-P., 1971, Statistique et Informatique apliquées. Dunod, Paris, 426 p. NIE, N. H., HULL, C. H., JENKINS, J. G., STEIN- BRENNER, K., BENT, D. H., 1975, SPSS Statis- tical Package for the Social Sciences. Ed. 2. McGraw-Hill, 675 p. ODHNER, N., 1950, Studies on Galapagos buli- mulids. Journal de Conchyliologie, 90: 253-268, 1 pl. COPPOIS AND GLOWACKI REIBISCH, P., 1892, Die conchyliologische Fauna der Galapagos-Inseln. Sitzungsberichte und Ab- handlungen Naturwissenschaftlichen Gesell- schaft Isis in Dresden, 1892: 13-32, 2 pl. SMITH, A. G., 1966, Land snails of the Galapagos. In: The Galapagos. Proceedings of the Sym- posia of the Galapagos International Scientific Project. BOWMAN, В. 1. (ed.), University of Cali- fornia Press, Berkeley & Los Angeles, p. 240- 251. SMITH, A. G., 1972, Three new land snails from Isla Santa Cruz (Indefatigable Island), Gala- pagos. Proceedings of the California Academy of Sciences, (4)39: 7-24. SMITH, A. G., 1974, Galapagos bulimulids: a taxo- nomic correction. Nautilus, 88: 67. SOWERBY, С. B. |, 1833, Characters of new spe- cies of Mollusca and Conchifera, collected by Mr. Cuming. Proceedings of the Zoological Society of London, 1: 70-74. VAGVOLGYI, J., 1977, Six new species and sub- species of Naesiotus from the Galapagos Islands (Pulmonata: Bulimulidae). Proceedings of the Biological Society of Washington, 90: 764-777. VAN DER WERFF, H. H., 1978, The Vegetation of the Galapagos Islands. Doctoral dissertation. Rijksuniversiteit te Utrecht, Netherlands, 102 p., 12 pl. VAN MOL, J.-J., 1972, Au sujet d'une nouvelle et remarquable espèce de Bulimulidae des lles Galapagos (Mollusca, Gastropoda, Pulmonata). Bulletin de l'Institut Royal des Sciences Naturel- les de Belgique, 48(11): 1-7. WIGGINS, I. L. & PORTER, D. M., 1971, Flora of the Galapagos Islands. Stanford University Press, Stanford, California, 998 p. APPENDIX The generic name for the Galapagos Buli- mulidae is not accepted by everyone; two names were proposed, Bulimulus and Naesiotus. No really decisive arguments were presented by Breure & Coppois (1978) in their too short and tentative analysis of the genus Naesiotus. As this work leaves too many points unsolved, we decided on a more con- servative approach and use Bulimulus as a generic name and Naesiotus as a subgenus. What is needed is a new analysis, including anatomical observations of all the Galapagos species as well as their continental relatives. The taxa mentioned in this study are listed with the corresponding code number used on figures and in the text. Seven taxa were not identified (11, 15, 17, 21, 22, 56, 69) and will be subject to further investigations. BIOMETRY OF GALAPAGOS BULIMULIDS 219 Code number Bulimulus (Naesiotus): taxa from Santa Cruz Island cavagnaroi Smith, 1972 (normal brown colour) cavagnaroi Smith, 1972 (“xanthic” form) cavagnaroi Smith, 1972 (“white unbanded’’) blomberghi Odhner, 1950 ochsneri Dall, 1917a akamatus Dall, 1917a reibischi Dall, 1895 eos Odhner, 1950 adelphus Dall, 1917a scalesiana Smith, 1972 saeronius Dall, 1917b lycodus Dall, 1917a hirsutus Vagvolgyi, 1977 alethorhytidus Dall, 1917a duncanus Dall, 1893 (no certitude) NN = = a = oo CO © © O À D © © -J O O1 BR OX ND — O Sen = 24 tanneri Dall, 1895 60 olla Dall, 1893 65 cymatias Dall, 1917a 68 jacobi (Sowerby, 1883) 70 tanneri Dall, 1895 (modified form) 81 nesioticus Dall, 1896 Taxa from other places in the archipelago 49 planospira Ancey, 1887; Champion Island Si pallidus Reibisch, 1892; Alcedo volcano, Isabela A detailed list of localities will be published later. BULIMULIDAE (GASTEROPODES, PULMONES) DES GALAPAGOS: 1. ANALYSE FACTORIELLE DES ESPÈCES DE L'ÎLE DE SANTA CRUZ Guy Coppois et Claude Glowacki RESUME Une comparaison biométrique des Bulimulidae (Pulmonés terrestres) de l'île de Santa Cruz est réalisée au moyen de l'analyse factorielle (7 variables, 140 specimens). L'analyse met en évidence deux facteurs principaux, le premier représente une estimation des dimensions globales de la coquille, le second une estimation de sa corpulence. De manière pratique, ces deux facteurs peuvent être assimilés a hauteur totale de la coquille et à l'angle apical terminal. L'analyse permet la séparation des spécimens en 28 taxa, la plupart sont des espèces différentes mais quelques cas de variations infraspécifiques sont mis en évidence. MALACOLOGIA, 1983, 23(2): 221-270 NEW ZEALAND SIDE-GILLED SEA SLUGS (OPISTHOBRANCHIA: NOTASPIDEA: PLEUROBRANCHIDAE) R. C. Willan Zoology Department, University of Auckland, Auckland, New Zealand! ABSTRACT Five New Zealand species of the family Pleurobranchidae are recognized. Berthella ornata (Cheeseman) is endemic. Bathyberthella zelandiae is described as a new genus and species, and is also endemic. Berthella aurantiaca (Risso) is restricted to the Mediterranean Sea, but in New Zealand this name has been applied incorrectly and indiscriminately to two species differ- entiated here—Berthellina citrina (Ruppell & Leuckart) has a small shell, denticulate radular teeth, and possesses a prostate gland; Berthella mediatas Burn has a larger shell, smooth teeth and no prostate gland. Pleurobranchaea maculata (Quoy & Gaimard) is the sole species of its genus in New Zealand; Pleurobranchaea novaezealandiae Cheeseman and Pleurobranchaea novaezelandiae var. granulosa Bergh are synonyms of P. maculata. Key words: Gastropoda; Opisthobranchia; Notaspidea; Pleurobranchidae; New Zealand; taxonomy; revision. INTRODUCTION Considerable advances have been made during the past decade in studies of nomen- clature and relationships for New Zealand opisthobranchs. This paper reviews the order Notaspidea and in particular the family Pleu- robranchidae. It arises from my investigations into the identity and ecology of shallow-water pleurobranchs from New Zealand (Willan, 1975) plus studies conducted since that time on additional material. These examinations extended the coverage to include deep-water species and thus this work now encompasses the entire New Zealand notaspidean fauna. The Notaspidea have not been mono- graphed before in New Zealand. Cheeseman (1878, 1879) was the only worker to describe pleurobranchs from New Zealand as new species. Others recorded side-gilled slugs from this country under the names of estab- lished species (Bergh, 1900; Odhner, 1924). These subsequent names were incorrect but inevitably became incorporated into important checklists of the New Zealand molluscan fauna (Suter, 1913; Powell, 1937, 1946, 1957, 1961, 1976, 1979) and so became en- trenched. These incorrect taxa appeared in ecological works (Morton & Miller, 1968; Batham, 1969; Ottaway, 1977b). By noting that European names were being used incor- rectly for New Zealand species, Burn (1962: 134) highlighted the taxonomic chaos that was the legacy of these early works. Reviews and reappraisals of the genera Pleurobranchella Thiele (Willan, 1977), Pleurobranchopsis Verrill and Gymnotoplax Pilsbry (Willan, 1978) (none of which occurs in New Zealand) have already been pub- lished. This latter study was a prerequisite to this present paper in that it stabilized the taxonomy of Berthellina Gardiner. A review article on feeding within the Notaspidea that deals with two New Zealand species is cur- rently in preparation. MATERIALS AND METHODS Material Studied There are six species of the order Notaspi- dea in New Zealand; five belong to the Pleu- robranchidae and one, Umbraculum sinicum (Gmelin, 1791), belongs to the Umbraculidae. U. sinicum is uncommon in this country and is found in northern waters only. As it has been described adequately from Australia by Burn (1959) and Thompson (1970), it is not treated further in this paper apart from its inclusion in the key. All but one of the pleurobranchs occur in shallow water on the continental shelf. The taxonomic arrangement and sequence of 1Present address: Zoology Department, University of Queensland, St. Lucia, Queensland 4067, Australia. (221) 222 WILLAN TABLE 1. Classification of New Zealand pleuro- branchs. Family Pleurobranchidae Subfamily Pleurobranchinae Berthellina citrina (Rúppell 4 Leuckan, 1828) Berthella ornata (Cheeseman, 1878) Berthella mediatas Burn, 1962 Bathyberthella zelandiae n. gen. £ n. sp. Subfamily Pleurobranchaeinae Pleurobranchaea maculata (Quoy & Gaimard, 1832) 2 ЖИВЕЕ иание presentation in this publication are given in Table 1. Because of the past cursory descrip- tions of New Zealand pleurobranchs and lack of comparative details derived from anatomi- cal examinations of specimens from this country | have described each species fully. Such detailed descriptions are required for all pleurobranchs before there is critical reap- praisal of taxa as biologically meaningful enti- ties. Locality data of material examined for each species are given in the Appendix. Terminology Terminology used for describing the Notaspidea throughout this paper is as fol- lows (Figs. 1-4). The terms refer specifically to pleurobranchs. The body has a dorsal mantle and ventral foot; posteriorly the foot sometimes has a pedal gland ventrally and/or a caudal spur on the dorsal face. Body length (living speci- mens) is the distance between the middle of the anterior border of the oral veil and the posterior extremity of either the mantle or the foot, depending on which extends farther to the rear in the actively crawling animal. The prebranchial aperture, nephroproct and re- productive apertures are situated in front of the gill on the right side; the mouth opens an- teriorly beneath an expanded oral veil, which is derived from a forward-projecting flap of tis- sue that grows outwards from, and yet re- mains connected to, the oral tentacles during metamorphosis (Gohar & Abul-Ela, 1957). The rhinophores are scroll-like and located above the oral veil. The terminology of the gill requires stan- dardization (Fig. 4). The axis of the gill is the rachis, which is either smooth or tuberculate. From the rachis smaller side branches arise alternately; these pinnae (= primary lamellae or pinnules) consist of an axis and a series of minute, overlapping leaves down each side. These side leaves are here called pinnules (= secondary lamellae or plicae). A membrane attaches the ventral surface of the rachis, for the greater part of its length, to the side of the body. Morton (1972) felt there was not enough evidence to homologize the gill of the Notaspidea with that of other opisthobranchs. He established that opisthobranchs have a wide variety of gill structures (e.g. the plicati- dium of the Cephalaspidea and Anaspidea, the plume-like gill of the Notaspidea, and the circum-anal gills of the Doridacea) that re- place the ctenidium of prosobranch molluscs. Labelling of parts of the radular tooth fol- lows Bertsch (1977), since the teeth of pleuro- branchs are homologous to those of chromo- dorid nudibranchs. | follow Burn (1962) for terminology relating to the jaws and my label- ling of reproductive organs incorporates the terminology proposed by Ghiselin (1965) for an idealized, gonochoric opisthobranch sys- tem. Many differences that have been used as generic characters are, in my view, only of use at the specific level. | list below those pleurobranch characters that have taxonomic value, and have attempted to assess their usefulness. Full information on all these char- acters, for both living and preserved animals where possible, is required in any future de- scriptions of pleurobranchs. Mantle: Form, colour, extent and texture are distinctive and reasonably constant in liv- ing animals; these characters are most useful if there is some previous information on the degree of intraspecific variability. Rhinophores: Very similar in construction throughout the group, all are scroll-like with a lateral groove and basal dilation. Data on size and position are of generic (rather than spe- cific) value. Oral veil: A most useful character to ob- serve in living animals. Such features as rela- tive width, strength of lateral extensions, de- gree of sinuosity and presence or absence of papillae on the anterior margin are good spe- cific characters. All pleurobranchs have grooved edges to the lateral areas of the oral veil. Foot and pedal structures: The foot offers the same types of characters as does the mantle. On the anterior dorsal surface a broad mucous gland and transverse slit are present. Some species possess a dorsal caudal spur and/or ventral gland towards the rear. The appearance of these structures probably indi- NEW ZEALAND PLEUROBRANCHIDAE 223 FIGS. 1-4. Pleurobranch descriptive terminology. Berthellina citrina illustrated as a representative of the Pleurobranchidae. 1. Dorsal view of whole animal. 2. Lateral view of head; arrow indicates site of mouth. 3. Exterior of shell. 4. Lateral view of gill. Abbreviations. a.g. = mucous gland on anterior border of foot; a.m. = anterior margin of shell (i.e. edge facing anteriorly when shell is in place in living animal); ap. = apical region of teleoconch; d.gl. = digestive gland; f. = foot; g. = position of gill on right side beneath mantle; L. = left side of shell in place in living animal; m. = mantle; o.t. = ovotestis; o.v. = oral veil; pa. = pinna of gill; p.m. = posterior edge of shell (i.e. margin facing posteriorly when shell is in place in living animal); pt. = protoconch; pu. = pinnule of gill; R. = right side of shell in place in living animal; r. = rhinophore; ra. = rachis of gill; sh. = position of shell beneath mantle. 224 cates attainment of sexual maturity rather than similarity between species. The signifi- cance of these structures is discussed further at the end of this section. Gill: The gill has been employed as an indi- cator of genus, depending whether the rachis is smooth or tuberculate (but | shall later query the importance of this feature). Gener- ally the number of pinnae shows considerable intraspecific variation, particularly in relation to the size of the individual—juveniles have fewer pinnae and pinnules. Shell: Presence or absence is important at the generic level; generally shape varies con- siderably within any one species. The shell is never wholly, or even partially, uncovered by the mantle in any pleurobranch when alive (Willan, 1978). In the literature are records of specimens of normally-shelled pleurobranchs being found without shells—Pleurobranchus peroni and P. inhacae (Macnae, 1962) and Berthellina citrina (Thompson, 1970; Edmunds & Thompson, 1972). So absence of a shell, by itself, should not be thought of as significant when an unknown specimen is being checked against a published descrip- tion. Gut: Pleurobranchs show little variation in the basic structure of the alimentary canal. Guiart (1901) used Pleurobranchus mem- branaceus (Montagu) in his description of the gut in the Pleurobranchidae, and Thompson & Slinn (1959) figured the alimentary canal of this species. Both the gut and nervous system are characteristic at the generic level. Radula: Provides excellent diagnostic char- acteristics for delimiting genera (except Pleurobranchus and Berthella). Generally considerable intraspecific variation exists, particularly with respect to age of the individ- ual (as is the case in doridacean nudi- branchs—see Bertsch, 1976, 1977). Never- theless, radular details are most useful at the species level when the extent of this variability is understood. Jaws: As for the radula. Reproductive system: A highly important source of differential characters at higher levels. From species to species there is varia- tion in detail but it is easy to misinterpret de- velopmental changes in the genital organs as specific differences. Pedal gland: This enigmatic gland is found on many (maybe all) species of Pleuro- branchus, Berthella and Pleurobranchaea. It was first described by Vayssiere (1885), who attributed the first record of it to Delle-Chiaje WILLAN (1828). Vayssiere (1885: 111) was unable to suggest any function for the gland. Some steps have recently been made towards an understanding of the function of this gland. It is becoming clear that it has a sexual function. Thompson & Slinn (1959) showed that in Pleurobranchus membranaceus it is absent in juveniles but present in older specimens. Macnae (1962) noted the gland only in fully mature, sexually active specimens. My observations on Pleurobranchaea maculata suggest that the appearance of this gland indicates the onset of sexual maturity. P. maculata reaches full size towards mid- winter but there is no pedal gland. It appears when the water temperature rises in spring and remains until the animal dies the following summer. It thus seems that development of the gland takes place with attainment of sexu- al maturity. The shape of the gland varies considerably between species. | suggest that the gland produces a species-specific phero- mone that diffuses through the water and draws members of the same species together to mate. Pleurobranchs tend to be solitary and widely scattered within one area, as dictated by their food; individuals are seldom found to- gether except when copulating; it is possible, therefore, that there are chemical cues. HIGHER CATEGORIES Historical Overview Pilsbry (1896) attempted the first synthesis of the Notaspidea. His system was based mostly on external characteristics. He recog- nized the Umbraculidae, with two genera and two subgenera, and the Pleurobranchidae, with six genera and two subgenera. Vayssiere (1898,1901) published a large monograph dealing at length with the known pleuro- branchs; his classification scheme had ap- peared earlier (Vayssiere, 1896). The Pleuro- branchidae were divided into four genera, one (Pleurobranchus) being further split into four subgenera. A second large monograph was written by Bergh (1897-1905), this work being produced quite independently from that of Vayssiere. Bergh gave detailed anatomical accounts of specimens examined; however he did not treat all the known species as Vayssiere had done. Bergh recognized six genera which he did not subdivide further. | have followed Winckworth (1946) in citing the dates of the five parts of Bergh's Sempers NEW ZEALAND PLEUROBRANCHIDAE 225 Reisen, Malacologische Untersuchungen that deal with pleurobranchs. Odhner (1926) presented another scheme, again based on external characteristics— pedal gland, gill rachis and mantle margin; within the Pleurobranchidae he recognized two subfamilies with a total of five genera and two subgenera. Marcus (1971) has translated Odhner's key into English. With some small alterations Odhner's classification has been kept by later reviewers (Thiele, 1931; Burn, 1962; Franc, 1968; Thompson, 1976); | use it here because it allows incorporation of the greatest number of features, including obvi- ous external characteristics. Here | divide the Pleurobranchidae into two subfamilies with eight genera between them; recognition of subgenera is, in my opinion, premature at present. Mention must be made of Gardiner's (1936) brief paper, which because it reorga- nized generic nomenclature, was of great sig- nificance. The order Notaspidea Fischer All notaspideans have a single, elongate, plume-like gill on the right side of the body between the mantle and foot. All are carnivo- rous and have a broad radula with many teeth. All have rhinophores that are scroll-like and with a lateral, longitudinal slit. No species possesses parapodia. The order contains “tectibranchs” (with large, external, patelli- form shells (Umbraculidae)), and “nudi- branchs” (with shells that are internal or ab- sent (Pleurobranchidae)). The Notaspidea form a transitional link between the lower (e.g. the shell-bearing Cephalaspidea) and higher (e.g. the shell-less Doridacea) opisthobranch groups. Evidence suggests that the Notaspi- dea represent an intermediate grade of or- ganization. The shelled, umbraculiform mem- bers are so different anatomically from the pleurobranchs that both may well have reached the notaspidean state by independ- ent paths from tectibranch ancestors. Recently Thompson (1970, 1976) and Edmunds & Thompson (1972) have called this order Pleurobranchomorpha (originally a tribe of the order Opisthobranchia of Pelseneer (1906). | prefer to keep the name Notaspidea because it is brief and has been used consistently throughout the literature. Also, the other tribes which Pelseneer placed within the Opisthobranchia have been largely dismantled since. For example, Pelseneer in- cluded the Lophocercidae (now in Oxynoei- dae, Sacoglossa) and Limacinidae, Cymbulii- dae, Cavoliniidae (now in Thecosomata) in the tribe Bullomorpha. He included the Pneumodermatidae, Clionopsidae, Noto- branchaeidae, Thiptodontidae, Clionidae, Halopsychidae (now in Gymnosomata) in the tribe Aplysiomorpha. It would appear more appropriate to use something other than Pleu- robranchomorpha if a new name should be needed. Thiele (1931) combined the Notaspidea with the Nudibranchia sensu lato in a new order, Acoela, to reflect their close relation- ships, the nudibranchs having risen from the pleurobranchs (Odhner 1939). But since the Notaspidea are so different from other groups of opisthobranchs they should be given equal ranking with the Nudibranchia. Furthermore, the homogeneity of the order Nudibranchia itself is in doubt (Minichev, 1970; Minichev & Starobogataov, 1978). Hyman (1967) noted that Thiele’s taxonomy had failed to gain gen- eral acceptance, and most authors who have recently considered the arrangement of high- er groups within the Opisthobranchia have given the Notaspidea ordinal status equal to that of the Nudibranchia sensu lato (Hyman, 1967; Minichev, 1970; Morton, 1958, 1972; Nordsieck, 1972; Taylor & Sohl, 1962; Thompson, 1976). Franc (1968), however, lists the order separately as Pleurobran- chacea Deshayes, 1830. Members of the order have long been rec- ognized as falling into two distinct series, and | agree with Thompson (1976) that each war- rants no less than subordinal ranking. The first (Umbraculacea) includes the family Umbraculidae Gray (= Umbrellidae auct.) with three genera: Umbraculum Schumacher, 1817; Tylodina Rafinesque, 1819; Tylodinella Mazzarelli, 1897. The Umbraculidae is here diagnosed as fol- lows: Shell external, limpet-like, with protoconch minute and hyperstrophic, apex near centre, interior with a closed or incomplete muscle scar, periostracum often dense; body smaller or much larger than shell, mantle thin, margin serrated or tentaculate; foot with large flat sole, upper surface smooth or tuberculate; head with a pair of enrolled rhinophoral tenta- cles with sessile eyes at their bases, mouth with two pairs of small oral tentacles. Gill a long plume lying between mantle and foot on anterior and right side, adnate and bearing numerous bipinnate branches for greater part of its length, posterior end free and bipinnate; 226 anal papilla projecting behind attached por- tion of gill, penis anterior—external, lying in anterior sinus of foot, in median line in front of and below head (Umbraculum) or retractile, on right side in front of gill (Tylodina, Tylodi- nella). Radula very broad, bearing a great number of similar, crowded, needle-like teeth, with recurved simple cusps which lack sub- denticles; buccal armature consisting of lightly cornified polygonal plates which lack sub- denticles. Some authors (Pruvot-Fol, 1954; Burn, 1959) divide the Umbraculidae into two fami- lies—Umbraculidae Dall and Tylodinidae Gray, on the basis of the proportions of the shell and body, and the position of the head (either projecting or included in an anterior sinus of the foot), and differences in external genitalia, shell periostracum and muscle scars, radula and buccal armature. The Pleurobranchidae Menke The second series (Suborder Pleurobran- chacea) constitutes the family Pleurobran- chidae with seven recognized genera: Pleurobranchus Cuvier, 1805 (sensu Thomp- son, 1970 and Baba & Hamatani, 1971) Berthella Blainville, 1825 Pleurehdera Marcus & Marcus, 1970 Berthellina Gardiner, 1936 Pleurobranchaea Meckel in Leue, 1813 Euselenops Pilsbry, 1896 Pleurobranchella Thiele, 1925 The present work adds one more—Bathy- berthella n. gen. The Pleurobranchidae are defined as fol- lows: Gill on right side of body, extending back- wards in groove between mantle and foot, rachis smooth or tuberculate, side pinnae subdivided into pinnules, anterior part at- tached to body by a basement membrane, posterior end free; prebranchial aperture in front of gill. Shell internal beneath mantle, either small or absent; when present the shell is haliotiform or spatulate. Mantle smooth or tuberculate, either large and separated from foot (Pleurobranchus, Berthella, Pleureh- dera, Berthellina, Bathyberthella) or smaller than foot and merging with it anteriorly and/or posteriorly (Pleurobranchaea, Euselenops, Pleurobranchella). Head with trapezoidal oral veil projecting above mouth, its lateral edges longitudinally grooved; a pair of rolled rhino- phores above oral veil, also laterally grooved, rhinophores arising either together mid-ante- WILLAN riorly or separately where mantle merges an- teriorly into oral veil. Gut with extensible oral tube and muscular pharyngeal bulb inside which lie jaws and radula; two jaws placed laterally, composed of numerous similar im- bricated mandibular elements; radula broad, with or without rachidian; gut with unpaired, dorsal oral gland opening anteriorly into pharyngeal bulb and a pair of salivary glands; tubules from oral gland ramify to a greater or lesser extent throughout the body; stomach a large, unthickened, mid-ventral sac; anus opening on right side. Animals hermaphroditic with diaulic or triaulic reproductive systems; penis retractile, stout or long and slender, sometimes with enclosing flaps externally and with or without penial gland internally, smooth or papillose; vas deferens with or without prostatic portion; eggs laid in loosely ar- ranged, or coiled, spawn bands; larva hatch- ing as a pelagic veliger. Distributed world- wide in tropical and temperate (rarely in cold) areas. The Pleurobranchidae have two sub- families. Pleurobranchinae Menke, 1828 Shell internal; distinct mantle with free edges all round; rhinophores arising together mid-anteriorly; radula without a rachidian; mandibular elements cruciform with simple or denticulate blades; generally low activity (but a few species are able to swim). Pleurobranchaeinae Pilsbry, 1896 No shell; mantle reduced and continuous anteriorly with the oral veil, rhinophores sepa- rate, dorso-lateral; radula with or without rachidian; mandibular elements polygonal or scale-like and denticulate; generally high ac- tivity. Burn (1962) elevated both to familial rank essentially on the presence or absence of a shell; but as Edmunds & Thompson (1972) have shown for Berthellina citrina, this char- acter is not constant even within a species. My opinion is that these groups show such a unity of body design and gill plan and struc- ture of the gut and reproductive system as to warrant placing them within a single family. These similarities suggest their derivation from a common ancestor. Burn also split the Pleurobranchidae (sensu Burn, 1962) into two subfamilies—the Pleurobranchinae (large and with tuberculate gill rachis and pedal gland), and the Berthelli- nae (smaller with a non-tuberculate rachis NEW ZEALAND PLEUROBRANCHIDAE 227 and no pedal gland). These divisions are un- justified on the characters chosen since small species of Pleurobrachinae do exist, e.g. Pleurobranchus ovalis Pease (Thompson, 1970); some species of Berthella (subfamily Berthellinae) do have weak tubercles on the gill rachis, e.g. B. ornata (Cheeseman) (pres- ent observations); and several (if not all) spe- cies of Berthella have a pedal gland when they are sexually mature. Even though these groups do not warrant separation at the sub- familial level as advocated by Burn | agree with his distinctions and consider the “pleuro- branchine” and “berthelline” groups repre- sent natural lineages. The two monotypic genera Pleurehdera and Bathyberthella (the new genus described herein) together display all the characters required to span the gap between Pleurobranchus and Berthellina. | do not wish to imply, however, that either Pleurehdera or Bathyberthella is ancestral to either the “pleurobranchine” or the “berthel- line” groups. Relationships of genera within the Pleuro- branchinae remain confused. | recognize only four (Pleurobranchus, Berthella, Berthellina, Pleurehdera) and add a fifth (Bathyberthella). | support Thompson (1970) and Baba & Hamatani (1971) in considering Pleurobran- chus to encompass Oscanius Leach and Susania Gray. Two conflicting classifications for this subfamily have emerged from past studies; one based on radular characteristics unites Berthella and Pleurobranchus (e.g. Vayssiere, 1896, 1898; Odhner, 1926), the other relying on the nature of the mantle, gill and shell unites Berthella and Berthellina and excludes Pleurobranchus (Burn, 1962). The two systems are incompatible. As stated above, my studies so far have led me to sup- port the latter classification. Even though this stance is adopted, | acknowledge that Pleuro- branchus and Berthella are close to each other. At present the only satisfactory char- acter differentiating them is presence or ab- sence, respectively, of tubercles on the man- tle and gill rachis. When this character is set aside it seems impossible to draw a hard and fast line between them; indeed their radular and jaw structures seem identical. According to this criterion some species remain poised between Pleurobranchus and Berthella (the New Zealand Berthella ornata (Cheeseman) and North American Berthella americana (Dall) are examples). | suggest that difficulty of separating them has arisen more through lack of critical appraisal of species than lack of distinguishing characters. Species belonging to both genera require further examination (particularly for characters related to mantle, radula, jaws, reproductive and alimentary sys- tems) that will enable the boundary to be cut more sharply if these genera are to remain separate. Pleurobranchus and Berthella have long been upheld as separate. Both are enormous genera, easily the largest in the order, each has more than 50 named species although no one is sure how many biological species exist. Pleurehdera (with its type species, P. haraldi, still known from only a single speci- men from the Tuamotu Archipelago) and Bathyberthella are particularly important be- cause they possess characters linking them with both pleurobranchine and pleurobran- chaeine genera. One or both could be a key between the major genera or subfamilies. An appraisal of Bathyberthella with regard to other genera appears later in this work in con- nection with the description of the new spe- cies from New Zealand. Despite the natural affinity of Berthellina and Berthella there are sufficient characters presently recognized to adequately diagnose them as separate. All species of Berthellina have immediately recognizable radulae, their shells too are distinctive; they have a prostate gland and jaws with (almost always) smooth mandibular elements. Were it not for its man- dibular elements and pedal gland, Pleureh- dera haraldi would be classified as a Berthel- lina. In all these characters Berthellina spe- cies differ from the Pleurobranchus-Berthella group, which warrants their separation, but | would not remove Berthellina any further than recognizing it as a separate genus. There are probably not more than four valid species of Berthellina. The three genera in the Pleurobranchaei- nae (Pleurobranchaea, Euselenops, Pleuro- branchella) are well separated from each other by several characters. Pleurobran- chaea has numerous species, many poorly defined (Marcus & Marcus, 1966). Euselenops is monotypic—E. luniceps (Cuvier, 1817) be- ing widespread. The deepwater Pleurobran- chella has possibly four species (Willan, 19777): SYSTEMATICS Berthellina Gardiner, 1936 Berthella Vayssiere, 1896: 115 (non Berthella Blainville, 1825). 228 Berthellina Gardiner, 1936: 198. Type-spe- cies by original designation: Berthellina engeli Gardiner, 1936. Definition Relatively small pleurobranchs, body ellipti- cal and convex; mantle large, smooth, simple and free all round, without an anterior crenu- lation; pedal gland never present; gill rachis smooth; anus at posterior end of gill mem- brane; shell beneath mantle, small (1/4-1/5 length of body), triangular or ovate, carried anteriorly, occasionally absent; teeth or radu- la elongate, lamelliform, serrated on distal section of posterior edge; mandibular ele- ments cruciform on inner surface of jaw, smooth or indistinctly denticulate; vas defer- ens dilated into prostate gland; penis without accessory leaves. Remarks Species with these characters were attrib- uted to Berthella Blainville, 1825 by Pilsbry (1896), Vayssiére (1896, 1989) and Odhner (1926). Later authors used Berthellina for this genus (Gardiner, 1936; Odhner, 1939). The reasons for the change are given in the follow- ing summary. Blainville (1825) created Berthella to in- clude notaspidean opisthobranchs with lamel- late teeth and smooth jaw elements; he desig- nated Berthella porosa Blainville, 1825 as type-species. Vayssiere (1898: 271) pointed out that B. porosa Blainville was a junior ob- jective synonym of Berthella plumula (Mon- tagu, 1803) and therefore Montagu's Bulla plumula took precedence as type-species. But a re-examination of the characters of Berthella plumula by Gardiner (1936) showed that this species has smooth, hook-like radu- lar teeth and denticulate jaw elements. These characters had been applied to Bouvieria Vayssiere, 1896, so Bouvieria became a synonym of Berthella. Gardiner (1936) cre- ated Berthellina for those species with lamel- late radular teeth. It is probable that some of the six species of Berthellina recognized by Burn (1962) are not valid. Indeed Edmunds & Thompson (1972) and Thompson (1976) amalgamated B. engeli with B. citrina (Rúppell & Leuckart), thus giving that species' distribution as Indo- Pacific, Mediterranean and marginally North Atlantic (e.g., Britain). Marcus 8 Marcus (1967a) named a new subspecies (Berthel- WILLAN lina engeli ilisima), found from San Diego, California, to Guaymas, Mexico, but it was re- jected soon afterwards (Bertsch, 1970). If B. engeli does exist in North America and is synonymous with B. citrina, then the one spe- cies would have an almost world-wide distri- bution. Care is needed here, and | concur with Thompson (1977) that hasty lumping of Berthellina species at this stage would be un- wise. | think some reinstatements may be necessary in future (e.g. Berthellina engeli Gardiner). Marcus & Marcus (1957) trans- ferred Pleurobranchus (Oscanius) amarillius Mattox to Berthellina where it is now consid- ered a synonym of Berthellina quadridens (Morch) (Burn, 1962; Marcus & Hughes, 1974; Thompson, 1977). One species over- looked in Burn's (1962) list is Berthellina africana Pruvot-Fol (1953) from the Atlantic coast of Morocco. Abbott (1974) disrupted accepted nomen- clature by replacing Berthellina with Gymno- toplax Pilsbry. But the type-species of Gymnotoplax (Pleurobranchus americanus Verrill) is undoubtedly a species of Berthella. Abbott's alteration is therefore unacceptable (Willan, 1978). Berthellina citrina (Ruppell & Leuckart, 1828) (Figs. 1-6, 9, 10, 13-19) There are at least eight synonyms for this species. | do not give a synonymy here for three reasons. Burn (1962) has already pre- sented an adequate synonymy and discus- sion; also the species is so widespread that a full synonymy would be very lengthy. Finally, in New Zealand this species has been con- fused with another pleurobranch (Berthella mediatas Burn) and references do not dis- criminate between the two. Recent investigations have uncovered three more names to add to the synonymy of Berthellina citrina: Pleurobranchus cuvieri Bergh, 1898: 129— 131, pl. 11, figs. 19-27. Berthella borneensis Bergh, 1905b: 69-70, pl. 5, fig: 3; pl 11, figs: 45-47; Berthella minor Bergh, 1905b: 70-71; pl. 13, figs. 1-3. Marcus & Marcus (1970) redescribed Berthellina cuvieri (Bergh) from Madagascar, but their material easily falls within the range of variation of Indo-Pacific specimens of B. citrina. Narayanan (1970) studied material from the Gulf of Kutch, India, that he had pre- NEW ZEALAND PLEUROBRANCHIDAE 229 FIG. 5. Berthellina citrina. Lengths of both specimens approx. 40 mm. From 3-4 m, Leigh Harbour, North Auckland, 22 Nov. 1973. Photograph: G. W. Batt. viously assigned to Berthellina minor (Narayanan, 1969) and placed that name also in the synonymy of B. citrina. Live animal (Fig. 5) During this study over 80 living specimens of Berthellina citrina have been examined; this is an adequate number to permit a de- scription encompassing the natural variability of the species over its geographic range in New Zealand. Shape elliptical, mantle wrapping entire upper part of body except rhinophores and oral veil which project anteriorly; rhinophores exposed from just behind their point of fusion; foot generally does not appear behind poste- rior edge of mantle when slug is crawling. Colour varies from deep apricot-orange to pale lemon; white spots almost always on sur- face, generally more numerous on sides and back, less apparent in largest individuals (ap- prox. 50 mm long). Rhinophores and oral veil same colour as mantle surface. Mantle very delicate, smooth, soft, and translucent; ante- rior reddish-brown shell and posterior diges- tive gland, which appears as a black smudge, can be seen easily from dorsal aspect. When mantle margins are removed (e.g. through predation by Pleurobranchaea maculata), white spots become larger and more numer- ous after 5-6 days. Rhinophores elongate, diverging by 50- 60°, raised above oral veil, consisting of a spirally-wound sheet of tissue of 172 whorls, open at last turn to leave a narrow lateral slit. Distal margin of rhinophoral sheet can be moved to control size of slit. At posterior-later- al corner, slit expanded to form a pyriform aperture, water enters along slit and is ex- pelled through distal aperture. Eyes lie near, and just behind, basal aperture, covered by anterior edge of mantle. Oral veil extending forward from head as a trapezoidal sail, anterior margin broad, wavy—but not nearly as sinuous as in Berthella mediatas; thickened lateral borders of oral veil have a deep furrow that expands into a cavity internally at base of veil. Gill rachis smooth and cylindrical, base- ment membrane attaching gill to body for two- thirds of the gill’s length; 18-28 pinnae on up- per side of gill, first pinna always on upper side; mean number of pinnae for 15 speci- mens 23.7. Anus on dorsal side of gill, at hind end of basement membrane, opening on a 230 WILLAN slightly raised papilla; longitudinal folds in in- terior of rectum visible within anus. Prebran- chial aperture opens just in front of, and slight- ly above, front of rachis; a much smaller aper- ture, the renal pore, opens below and behind base of rachis. In specimens preserved in alcohol, mantle is creamish in colour. The delicate, translu- cent tissue becomes opaque, apparently thicker, and velvet-like, frequently obscuring shell. Shell (Figs. 6, 9, 10) Shell small (1/4-1/5 length of extended body), thin, flattened, triangular; lying above pericardium in front of digestive gland; visible through mantle in living specimens. Consist- ing of two distinct, and clearly demarcated regions—protoconch of 1% to 1172 whorls and teleoconch; protoconch translucent, lacking regular sculpture but possessing some irregu- lar calcification externally (Fig. 9); teleoconch spatulate, golden-brown or rich reddish- brown, older shells thicker and darker, clear apical region always present on teleoconch; first-formed teleoconch weakly convex and quadrangular, later growth unequal, produc- ing eventually a triangular, spatulate form. Shell entirely covered by glistening, transpar- ent periostracum that flakes off when shell dries; periostracum does not extend beyond shell margin. To the naked eye the surface sculpture is of irregular, concentric growth rings coalescing towards the margin in the adult shell to pro- duce broad, flat ridges. Fine surface sculpture is most obvious near teleoconch apex (Fig. 10) and appears as a series of radiating, punctate depressions separated by flattened areas. Punctae occur in radial files, with some concentric organization; occasionally three to five merge within the radial rows. However, concentric fusion (which corresponds with growth checks) can also occur, becoming more obvious in older portion of shell, obliter- ating much of the regular radial organization. All specimens of Berthellina citrina that | have examined (up to 5 cm) have had shells; this finding is contrary to that of Thompson (1970). He claimed that larger specimens (2cm and over in extended body length) lacked shells. In New Zealand, B. citrina does not show the linear relationship between shell length and body length that has been shown for this species in the Red Sea (Gohar & Abul- Ela, 1957). Five shells from specimens of Berthellina citrina taken at Leigh Harbour have been de- FIGS. 6-8. Shells of New Zealand pleurobranchs. 6. Berthellina citrina. Length 6.1 mm x width 3.3 mm. Specimen from Pananehe Is., Spirits Bay, 14 Jan. 1972; 7. Berthellina ornata. 13.8 x 9.5 mm. Specimen from Army Bay, Whangaparoa Pen., 10 Jan. 1974; 8. Berthella mediatas. 10.3 x 6.4 mm. Specimen from Army Bay, Whangaparoa Pen., 10 Jan. 1974. NEW ZEALAND PLEUROBRANCHIDAE 231 200 um $ Ma P Мася Vi т © ae | ES ee ENTRE | TERN. “Ane: + à + iy > 4 à Sort FIGS. 9-12. Scanning electron micrographs of shells. 9. Berthellina citrina, detail of protoconch. 10. Berthel- lina citrina, sculpture on apical region of teleoconch. 11. Berthella ornata, sculpture on exterior of teleo- conch. 12. Same, periostracum has peeled off and appears at top right. posited in the mollusc collection, Auckland In- stitute and Museum. Radula (Figs. 13, 14) Radula broad with two symmetrical halves, each forms one side of a V. The following radular description is applicable to an adult animal. Central tooth lacking. Lateral teeth all simi- lar, but showing gradual changes in length and denticulation across rows. Lateral teeth near midline 90 um high, elongate, weakly denticulate on upper third of posterior face, terminal cusp differentiated from rest of tooth by a relatively deep groove. Base of each tooth curving sharply backwards and slightly thickened at its attachment to radular mem- brane. Middle lateral teeth (Fig. 14) very large— 140 um high, comb-like, 15-20 strong denti- cles present on upper half of posterior face of blade; again terminal cusp longer than re- maining denticles, all denticles project in same plane. There is a group of one to three broader denticles immediately below cusp; re- maining denticles narrow, either needle-like or rounded, and frequently bifid at tips. Outermost lateral teeth (Fig. 13) shorter (approx. 100 um) than middle ones, and re- curved towards posterior face; 15-20 denti- cles on upper half of posterior face, smaller 232 WILLAN FIGS. 13-16. Radula and jaws of Berthellina citrina. 13. Detail of cusp and denticles on posterior face of a single tooth from middle lateral region of radula. 14. Outer lateral teeth and basal supporting structure within each row. 15. Mandibular elements on inner face of jaw. 16. Detail of same. than those on middle teeth, with weaker grooves separating them, seldom bifid; termi- nal cusp projects in plane of tooth, tip slightly flexed upwards, and separated by a relatively deeper groove from rest of denticles, which project obliquely from posterior face, although still in same plane. Examination of teeth with the SEM shows that the surfaces of the blade are smooth and there are no secondary structures on the den- ticles. There is a compression furrow on the shaft parallel to the anterior face. The furrow and the broad dorsal ridge beside it lock the blade in position laterally alongside adjacent teeth; there is thus a system for lateral sup- port along the rows during the feeding move- ments. Each tooth has an enlarged basal portion— a flange forming an oblique angle with the blade. On this flange, towards the base of the anterior face of the blade, is a raised ridge with a shallow depression at its base. A corre- sponding ridge on the next tooth in the same row fits into this depression, and into the de- pression on the flange of this tooth locks a ridge on the next tooth and so on. SEMs show NEW ZEALAND PLEUROBRANCHIDAE 233 the overlap of successive ridges down the row (Fig ТЗ): Any single formula for Berthellina citrina would be inadequate since the number of rows and the number of teeth per row in- crease as the animal grows. Specimens ex- amined had radular formulae in the range of 40 x 150.0.150 to 90 x 180.0.180. This is approximately the same as Burns (1962) range (60 x 120.0.120 to 95 x 200.0.200). Thompson (1970) published the first SEM photos of radular teeth of Berthellina citrina. Jaws (Figs. 15, 16) Two jaws, one on each side of the radula, are joined anteriorly at top and bottom by labi- al cuticle; anterior edges of cuticle recurved. Jaws rectangular, narrowing posteriorly to a point. Jaws composed of columns of interlocked small elements; at jaw surface elements ap- pear cruciform, with broad bases and elon- gate blades. Midway between base and blade are two rounded, lateral processes, one on either side; each process abuts against a sim- ilar process in the same row. “These proc- esses are not exactly opposite, those on one side being slightly in advance of those on the opposite side, thus determining the slight obliquity of the rows. Toward the dorsal and ventral margins of the mandible, the elements become somewhat more irregular in form and depart from the typical shape found in the central areas.” (MacFarland, 1966; part of description of Berthellina engeli Gardiner, but applicable to all species of Berthellina.) A mid-ventral spike on flattened base of each element locks into place behind lateral processes of two adjacent elements so sup- porting them from behind; overlapping tip of blade of element in the next row posteriorly acts to enforce connection from above. Blade of one element fits between base and lateral processes of two adjacent elements; raised and with dorsal flange so that dorsal surface presents only a series of similar flattened blades. SEM photos (Figs. 15, 16) give a view of this top surface. Mandibular elements are approx. 80 um long and up to 30 um wide at centre of jaws, base slightly narrower; element widens in a smooth curve to blunt tips of lateral proc- esses, element constricts just above lateral processes and broadens anteriorly to form blade. Blade elongate, straight-sided, narrow- ing gradually to a pointed tip. All jaws but one had no denticulation on the blade. The specimen in Figs. 15 and 16 had one or two definite denticles developed near the apex of many of the mandibular elements. Most had a single cusp, but a few showed a weaker cusp in a similar position on the other side of the blade. These weaker cusps lay in the plane of the mandibular elements and did not break the smooth outlines of the blade. They were not separated at their bases from the rest of the dorsal surface of the element and narrowed rapidly. Baba (1937) and Burn (1962) have also noticed these structures on the mandibular elements of Berthellina citrina. The presence of these rudimentary and ir- regular cusps on the mandibular elements of Berthellina citrina invalidates, in a sense, the diagnostic character of smooth-sided jaw ele- ments for the genus; but they are so different from those of Berthella and Pleurobranchus that the mandibular elements are still an im- portant taxonomic character. Reproductive system (Figs. 17-19) Reproductive apertures located on right side just in front of anterior end of gill, sur- rounded by a low, collar-like ridge. Most pre- served specimens examined had terminal portions of the reproductive ducts everted to some degree. Penial opening foremost, mid- dle opening marks vagina, and behind that is large aperture of oviduct. When genital papilla is everted, oviduct and vagina are separated by a broad flap of tissue projecting upwards from surrounding collar, but in retracted state they share a common atrium. Ovotestis yellowish, large, lying against right side of digestive gland, appearing glan- dular because of layers of immature eggs. Hermaphrodite duct divides into two as it enters ovotestis; each branch fine, white; after a short distance each divides repeatedly. Upon leaving ovotestis, hermaphrodite duct enlarges into an ampullar region which crosses beneath looped intestine and rises to dorsal surface of visceral mass where it ap- pears as a prominent tube, passing anteriorly across fluffy, yellow mucous gland. At anterior end of mucous gland, hermaphrodite duct constricts and forms a T-junction with a larger duct, anterior limb of the latter being proximal vas deferens; posterior limb enters nidamen- tal gland complex after a very short distance. Proximal vas deferens short, rapidly enlarg- ing to yellow prostate gland; distal vas 234 WILLAN У. FIGS. 17-19. Reproductive system of Berthellina citrina. 17. Composite view of structure of the reproductive organs. 18. Detail of structures associated with distal section of vas deferens and penis. 19. Detail of seminal receptacles in a mature animal. Abbreviations: amp. = ampulla; b.c. = bursa copulatraix; d.v.d. = distal section of vas deferens; ni. = nidamental glands; o.t. = ovotestis; ov. = oviduct; p. = penis; p.gl. = penial gland; pr. = prostate gland; p.s. = penial sheath; r.s. deferens much narrower, undergoing several loops and enters base of penis; a larger, much-coiled penial gland located there too (penial gland has also been called accessory prostate gland—Vayssiere, 1898, MacFar- land, 1966); penial gland elongate, thick- walled with longitudinal ridges, coiled about distal vas deferens, blind-ending. Beyond entrance of penial gland vas deferens is tight- = receptaculum seminis; v. = vagina. ly looped several times at base of penis. Penis short and curved, unarmed, projecting forward when everted, surrounded by stout muscular sheath. When everted, penis has weak concentric ridges on its surface and ends in a sharp tip. Fig. 18 gives a diagram of the anterior organs of the male section of the pallial gonoduct. Two prominent sacs open into vagina— NEW ZEALAND PLEUROBRANCHIDAE 235 bursa copulatrix which is spherical and thin- walled, and receptaculum seminis, which is smaller and thick-walled. Receptaculum seminis has a terminal, club-shaped dilation and a curved canal looping to enter vagina near its opening, where another swelling is present (see Fig. 19). No connection of va- gina with oviduct exists, save through the ex- ternal vaginal opening into its common exit close to the nidamental gland complex. Distribution The geographic range of Berthellina citrina in New Zealand is entirely northern (Aupouri- an). It extends down the east coast of the North Island, from Northland to the Bay of Plenty. There are no records so far of its oc- currence on the west coast (see Appendix). Elsewhere Berthellina citrina has been re- corded under various names from Australia (Quoy & Gaimard, 1832; O'Donoghue, 1924; Dakin, 1952; Burn, 1962; Thompson, 1970), New Caledonia (Risbec, 1928), Hawaii (Kay, 1979), Palau Islands (Marcus, 1965), Japan (Baba, 1937, 1949, 1969; Hirase, 1937; Usuki, 1969), Sri Lanka (Kelaart, 1883), South Africa (Macnae, 1962), Indonesia and Mauritius (Macnae, 1962; Marcus & Marcus, 1967a), Gulf of Kutch, India (Narayanan, 1970), Gulf of Elat (Marbach & Tsurnamal, 1973), and Red Sea (Gohar & Abul-Ela, 1957). Apart from B. oblongata (Audouin) and В. saidensis (O'Donoghue) from the Red Sea, B. citrina is the only species in this genus known in the Indo-Pacific (Burn, 1962). | can confirm Burn’s (1962) and Thomp- son's (1970) records of Berthellina citrina from New South Wales because | have col- lected specimens there myself. | have also collected B. citrina in southern Queensland and Vanuatu (New Hebrides). Dr. M. C. Miler found this same species in the Solomon Islands during the Royal Society В.$.1.Р. Ex- pedition of 1965. Discussion The identity of Berthellina citrina in New Zealand cannot be doubted. In B. citrina the combination of a small, spatulate shell, elon- gate radular teeth with denticulate posterior edges, mandibular elements with smooth (or weakly denticulate) blades, and prostatic dila- tion of the vas deferens are thoroughly dis- tinctive features allowing easy identification and separation from all other New Zealand pleurobranchs. Biologists, following Bergh (1900), have ap- plied the name “Bouvieria aurantiaca (Risso)” indiscriminately to any yellow or orange pleurobranch from New Zealand that lacked the dorsal, spotted pattern of Berthella ornata, so the two species answering this description (Berthellina citrina and Berthella mediatas) have not been hitherto distinguished. Table 2 lists the major characters which separate these two species. The only “B. aurantiaca” in the New Zealand literature that can be iden- tified as Berthellina citrina is that by Morton & Miller (1968). This is because these authors give an excellent colored illustration of a living animal (pl. 11, fig. 6). The inclusion of B. citrina by Gordon & Ballantine (1977, in Ap- pendix 3), was on my advice. “Bouvieria aurantiaca (Risso)” auct. is a valid species of Berthella from the Mediterranean Sea, and from the available literature it would appear close to Berthella mediatas Burn. Several recent publications have treated the biology of Berthellina citrina. A summary of development was presented by Gohar & Abul-Ela (1957). Usuki (1969) studied repro- duction, development and life history of Japa- nese specimens. Marbach & Tsurnamal (1973) made observations on feeding and acid secretion of specimens from the Gulf of Elat (Red Sea). | have not included any distribution records for Berthellina engeli Gardiner in the geo- graphic range list given above because | feel that it is premature to synonymize that spe- cies with B. citrina. In specimens of B. engeli that | have examined, the shells have been consistently more oval in shape, proportion- ately larger and positioned more posteriorly on the visceral mass. Berthella Blainville, 1825 Berthella Blainville, 1825: 370. Type-species by original designation: Berthella porosa Blainville, 1825 (= Bulla plumula Montagu, 1803). Cleanthus “Leach MS., 1819” Gray, 1847: 163. Published in the synonymy of Berth- ella Blainville, 1825. Bouvieria Vayssière, 1896: 66. Type-species by subsequent designation (Odhner, 1926: 22): Pleurobranchus aurantiacus Risso, 1818. Gymnotoplax Pilsbry, 1896: 20. Type-species by subsequent designation (Willan, 1978: 236 339): Pleurobranchus americanus Verrill, 1885. Berthellinops Burn, 1962: 135. Type-species by original designation: Berthellinops serenitas Burn, 1962. Abbott (1974) placed Cleanthus Gray as a synonym of Pleurobranchus Cuvier, 1804. Elsewhere (Willlan, 1978) | have already shown that the holotype of Pleurobranchus americanus Verrill, upon which Gymnotoplax Pilsbry is based, is a typical Berthella. Apart from a supposedly opposite (instead of al- ternate) arrangement of pinnae on the gill, Berthellinops Burn is a Berthella. Burn him- self later examined additional material and found this arrangement not to persist, and ac- cordingly would now treat Berthellinops as a synonym of Berthella (R. Burn, personal com- munication, July 1977). Definition Shelled pleurobranchs; body elliptical and convex; mantle large, simple and free, without an anterior indentation; pedal gland present; shell ovate, large (at least half length of body); gill rachis smooth or weakly tuberculate; radu- la with simple, curved or erect teeth; mandi- bular elements with smooth or denticulate blades; reproductive system lacking prostate gland. Remarks Members of the genus approach Berthel- lina in their relatively small size and (general- ly) smooth gill rachis; but in characters of the radula, jaws and reproductive system they ap- pear to show close affinities with Pleuro- branchus. Two Berthella species are found in New Zealand; both range throughout the country. There is the endemic Berthella ornata (Cheeseman), and secondly Berthella medi- atas Burn, a species shared with temperate southern Australia. Berthella ornata (Cheeseman, 1878) (Figs. 7, 11, 12, 20-31) 1879. Pleurobranchus ornatus Cheeseman: 175, pl. 15, figs. 1, 2—1879, Cheeseman: 378, pl. 16, figs. 1, 2.—1880, Hutton: 124.— 1896, Pilsbry: 206, pl. 47, figs. 22, 23.— 1898, Vayssiere: 337, pl. 14, figs. 18, 19.—1913, Suter: 550 (in subgenus WILLAN un Blainville).— 1915, Suter: pl. 77, 19. 6. 1924. Воимепа ornata Cheeseman; Odhner: 51, 86.— 1926. Odhner: 22.— 1937. Powell: 89, No. 1404.— 1953, Milligan: 134, 1139.— 1957, Powell: 114.—1961, Powell: 107. — 1964, Williams: 19, illust.— 1968, Morton & Miller: 167, 576, pl. 11, fig. 7— 1976, Powell: 112.— 1977, Gordon & Ballantine: 112.— 1979, Powell: 282, pl. 51, fig. 3. The description of Pleurobranchus ornatus Cheeseman (and Pleurobranchaea novae- zealandiae Cheeseman) first appeared in Proceedings of the Zoological Society of London for the year 1878. An identical de- scription was printed in the Transactions and Proceedings of the New Zealand Institute for the same year (but not published until 1879) (Pilsbry, 1896). Cheeseman’s descriptions were each accompanied by a figure, meticu- lously drawn by his sister Evelyn (A. W. B. Powell, personal communication, 1977). Powell (1939b) designated these drawings as iconotypes and he has recently republished them in colour (Powell, 1979). These icono- types are held in the Malacology Department, Auckland Institute and Museum. Live animal (Figs. 20, 21) Body oval to elongate; mantle smooth and soft, broad and circular. In actively crawling specimens (Fig. 22a) rhinophores and oral veil project beyond front edge of mantle, man- tle margin just covers point of fusion of rhino- phores; sole of foot often visible posteriorly. Mantle covers body as a broad cloak, lateral areas extend beyond foot to sweep the ground. In resting animals, tail and oral veil tuck beneath mantle, and rhinophores barely show (Fig. 22b). Anterior edge of mantle truncate, raised in midline to allow outward extension of rhino- phores, but not indented mid-anteriorly; pos- terior edge of mantle broadly rounded. Mantle thick in central area, thinner towards margins, not delicately translucent as in Berthellina citrina; margin entire, but often sinuous—par- ticularly on right side where upthrown poste- rior-lateral edge forms a temporary exit for ex- halant water current. Most animals have an apparently smooth mantle surface, but in some, large pores can be detected by eye on the surface. When an animal is removed from its substrate, the broad lateral areas of the mantle bend round the foot and hug the sides NEW ZEALAND PLEUROBRANCHIDAE 237 FIGS. 20, 21. Berthella ornata. 20. Length = 40 mm. From intertidal reef platform to left of Matheson Вау, near Leigh, North Auckland, 5 Oct. 1975. Photograph: R. C. Willan. 21. Length — 38 mm. From Pacific Bay, Tutukata, Northland, 8 March 1974. Photograph: G. W. Batt. 238 of the body to give a cylindrical and protective form. Colour characteristic, consisting of an even ground shading from pure white to pale (or rich) reddish-brown to orange-brown. Over- lying this are irregular blotches and spots, all a rich, dark, red-brown and very variable in shape and size. Some individuals show few (approx. 30) markings over mantle, more often entire mantle is covered. Markings tend to be largest and more concentrated on mid- dorsal region of mantle (area over shell and digestive gland), they decrease in size and number outward from the centre, and are small and scattered on anterior and lateral marginal areas. In some animals, particularly small ones (up to 10 mm in extended length), these spots extend out to mantle margins; adults have a more or less broad zone next to mantle margin that possesses nothing but background coloration; margin itself is opaque white. Digestive gland appears as a dark central smudge; shell not visible through mantle. When the animal is handled roughly and repeatedly, or when the mantle is prodded, the mantle produces a thick, milky-white re- pugnatory fluid; discharge can be localized or occur over the entire surface; secretion be- comes incorporated with clear body mucus and is left behind as slug crawls away. Foot large, oval, broad-soled, truncated in front and rounded behind; border minutely crenulate all round but crenulations are weak- est at anterior margin; sole sticky. The poste- rior quarter of sole in fully grown specimens has a large, but indistinct, circular pedal gland. All specimens examined had a creamy-white foot, with a thin, opaque white line at margin. No markings present on the foot sole, but lightly-pigmented blotches sometimes present on dorsal surface near tail. Oral veil broad, with thickened, moderately pronounced edges that are grooved laterally; anterior border sinuous, but not nearly as deeply embayed as in Berthella mediatas; very weakly crenulate along anterior margin. Oral veil creamy-white, always without mark- ings, though an indistinct white line along anterior border is present. Undersurface (in mid-anterior region) and lateral edges peri- odically touch the substratum in an explora- tory manner as animal crawls. Rhinophores Y-shaped, joined at bases and diverging by 30-45°; raised upwards at 90° from horizontal in an actively crawling ani- WILLAN mal, and at 45-60” when animal quiescent; each limb can be moved independently. Tips of rhinophores always white; in live speci- mens tips are active and mobile; very flexible as they move to test oncoming water. There is some degree of variability in coloration of low- er regions of rhinophores; these regions are often the same colour as mantle. Specimens with darkly-pigmented mantles usually have dark, reddish-brown or chocolate pigment in this region; extreme tips and fused basal areas are unpigmented. Small specimens (and adults with pale dorsal coloration) have rhinophores lightly pigmented with buff-brown or cream. Coloration of rhinophores changes to some extent with size of the individual. Most small specimens (less than 20 mm ex- tended length) had cream rhinophores (e.g. nine from Goat Island Bay, Leigh; May 1974). Larger specimens (greater than 30 mm ex- tended length frequently had rhinophores bearing darker pigmentation (e.g. 10 from Waiwera; January, July, September 1974). In living specimens rhinophores appear smooth externally, but in the preserved state strong contraction of these organs produces deep transverse grooves. When magnified, both oral veil and rhino- phores are seen to be covered with white, conical papillae. These papillae are evenly distributed over oral veil and are very dense towards tips of rhinophores. Sparse, white papillae are also present on mantle and tail. Berthella ornata has an exceptionally long gill, reaching backwards almost to level of tip of tail; gill attached to body wall for over half its length by basement membrane. Number of pinnae ranges from 18-25, first pinna always arises on upper side of rachis; mean number of pinnae counted on 12 adults—22; approxi- mately half the pinnae arise from posterior free end of gill. Pinnules are particularly large and give the gill a feather-like appearance. All specimens examined had creamy-white gills, devoid of markings. In living, and some well-fixed specimens, it is possible to see that the rachis is not smooth, but regularly covered with weak ridges or knobs where the pinnae join. These ridges continue on to the surface of the pin- nae which appear smooth. It would seem that these weak knobs are homologous with the better-developed tubercles of Pleurobran- chus Cuvier. Most preserved B. ornata ap- pear to have smooth rachises to their gills. Anus opens at side of body, on dorsal side of gill at hind end of basement membrane (i.e. NEW ZEALAND PLEUROBRANCHIDAE 239 about half-way along gill's total length); anus opens on a low, conical, backward-projecting papilla. Numerous glassy spicules (Fig. 23) are embedded in mantle, particularly at base of rhinophores where limbs diverge; spicules are 50-75 um in greatest total length; spicules consist of 2, 3, or 5 separate rays, each is straight or slightly curved, and terminates in a blunt tip, nearly all rays are equal in length. Although rays of spicules project in all direc- tions they do not diverge at right angles. Spic- ules very brittle, contraction of animal on death is enough to shatter most of then. Spic- ules effervesce in dilute acid suggesting they are, like the shell, calcareous. Shell (Fig. 7) Shell always present, generally rectangular and flattened, sides constricted towards api- Cal (posterior) part; slightly convex with ante- rior and posterior margins elevated from hori- zontal. Protoconch of 172 whorls, on left pos- terior-lateral corner; posterior flange of shell higher than level of protoconch. Lateral mar- gins of shell parallel, or showing a slight di- vergence anteriorly; anterior margin broadly rounded. Macroscopic sculpture is of irregular concentric folds, flattened, narrow and com- pressed towards margins. Microsculpture (Figs. 10, 11) consists of numerous, radiating ridges interspersed with broad depressions; ridges and depressions are parallel, directed in wavy lines that cross concentric folds; sur- face granular in texture. This sculpture is present only towards shell apex and becomes progressively weaker until, over the anterior third of teleoconch, it is obsolete. Outer surface of shell covered with a thin, iridescent periostracum (Fig. 12) which read- ily peels off; it is dull, whitish and chalky on shells of preserved specimens; periostracum does not exactly duplicate underlying wavy sculpture of shell, but appears punctate to- wards apex and smooth anteriorly. Shell thin, white towards apex, pale buff Over remainder; interior brownish-orange to- wards anterior edge, with a faint, red-brown streak towards left corner. Interior sculptured with raised, concentric ridges that narrow to- wards anterior margin. Four shells from specimens of B. ornata taken at Mahurangi Island, off Waiwera Beach, have been deposited in the mollusc collection, Auckland Institute and Museum. Radula (Figs. 24, 28, 29) Pale yellow-brown, broad and rectangular; maximum length 7.0 mm, width 4.0 mm when flattened. No rachidian in any row. Lateral teeth small and simple; inner and middle la- terals (Fig. 28) 40-60 um high, hook-like, Curved, with sharp-pointed apices; all have a relatively broad blade, both edges of which are smooth. Outer lateral teeth erect, straight- sided, with relatively narrow blades and rounded apices; outermost lateral teeth de- crease regularly from 45 um to 30 um (for ex- treme outer lateral tooth). There are many rows of teeth in the radula, giving the following range of radular formulae for adult specimens: 90 x 65.0.65 to 140 x 140.0.140. These for- mulae are high compared to those of other Berthella species. Jaws (Figs. 30, 31) Relatively small (approx. 2.0 mm long in adults), oval, with both ends rounded; jaws composed of numerous, cruciform, and rela- tively squat elements, each being 60-65 um long. Jaw elements broad-based, bases con- cave, widening to lateral processes; sides of blade smooth; each element terminates at a blunt, pointed apex (Figs. 30, 31); very rarely some elements have a weak denticle on side of blade about half way between a lateral process and apex. Apical thickenings on all mandibular elements are very apparent in both light microscopical views and SEM photos; thickenings rise from plane of top sur- face of element (i.e. that on inner surface of the jaw) and terminate in a sharp cusp; edges of blade near apex can also be raised. Such terminal thickenings have not been previously recorded for pleurobranch mandibular ele- ments. The jaw elements of Berthella ornata are broader and squatter than those of Berthellina citrina. Reproductive system (Figs. 25-27) Similar to that of Berthellina citrina, but with no prostatic enlargement of the vas deferens. The reproductive system of B. ornata differs in several other points from that of B. citrina, and these details are reported because of their specific importance. Although size of nidamental glands varies with age and maturity, two regions are dis- cernible. A dorsal, white and compact albu- 240 WILLAN amp. 25 FIGS. 22-27. Berthella ornata. 22. Whole animal from above, a = crawling, b = resting. Arrow indicates direction of exhalant current; 23. Spicules from mantle; 24. Radular teeth, a = inner lateral, b = middle lateral, c = outer lateral, d = outermost lateral; 25. Composite view of structure of reproductive organs; 26. Detail of albumen gland from ventral surface; 27. Detail of penis and penial sheath, the latter cut open. Abbreviations: al. = albumen gland; amp. = ampulla; b.c. = bursa copulatrix; o.t. = ovotestis; р. = penis; p.gl. = penial gland; p.s. = penial sheath; p.t. = apex of penis; r.s. = receptaculum seminis; v. = vagina; v.d. = vas deferens. NEW ZEALAND PLEUROBRANCHIDAE 241 FIGS. 28-31. Radula and jaws of Berthella ornata. 28. Inner and middle lateral teeth showing change in Structure across a row. 29. Detail of a single middle lateral tooth. 30. Mandibular elements on inner face of jaw. 31. Detail of the same. men gland (Fig. 26), consisting of a tightly- coiled tubular mass that is convoluted in ex- ternal apperance, and the mucous gland. This latter gland lies below the albumen gland; it is flattened, solid and cream in color. The vagina is grooved internally and passes as a long, straight canal to two semi- nal receptacles; vagina opens directly into a large sac-like bursa copulatrix whose dark- colored contents can be seen through the thin wall. A side branch arises from vaginal duct just before base of bursa copulatrix and loops to smaller receptaculum seminis. This con- centration is well back from vaginal aperture. There is a dilation in front of receptaculum seminis; receptaculum solid, roughly conical in shape, pressed against bursa copulatrix when in its natural position. Vas deferens much-convoluted, thin- walled, of a slightly larger diameter than hermaphrodite duct, pressed against mem- brane surrounding bursa; vas deferens lacks a prostatic glandular region. A penial gland arises from vas deferens at level of bursa; it is finger-like, flattened, thin-walled, not exten- sively coiled. Penial gland pressed against 242 membrane of bursa copulatrix, although not entwined with vas deferens. Beyond junction with penial gland, vas deferens runs beneath vagina, both lie within a common membra- nous sheath. Penis unarmed, of a smaller diameter than hermaphrodite duct; when re- tracted, penis remains coiled inside its sheath. Distribution Berthella ornata is found in clean, rocky situations on open or partially sheltered coasts throughout New Zealand. It ranges from the intertidal to shallow subtidal depths (approx. dm). Discussion A distinctive species of pleurobranch, Berthella ornata is the only endemic shallow- water New Zealand member of the order Notaspidea. It can be distinguished from its New Zealand congener, B. mediatas Burn, by colour, texture of mantle surface, gill length, and position of anus. B. ornata has smooth- bladed mandibular elements and relatively large, curved, middle lateral radular teeth; B. mediatas has strongly denticulate blades to the mandibular elements and relatively small, less curved, middle lateral teeth. Berthella ornata has several other charac- teristics important from a taxonomic view and which may prove significant in future studies of phylogeny within the genus: mantle colora- tion; gill; mandibular elements; radula; repro- ductive system. Other than Berthella ornata, few members of the genus have distinctive colour patterns on the mantle. Most species are uniformly dull yellow-orange or pale fawn, the mantle being more or less translucent with the dark diges- tive gland and light ovotestis showing through. A literature search revealed five other Berthella species with conspicuous darker markings on the mantle surface: Berthella ocellata (Delle-Chiaje, 1828) from the Mediterranean Sea. This species has a yellow-ochre to reddish-brown ground colour overlain with large, white or yellowish blotch- es; it has a much narrower shell than B. ornata, with a tall, projecting spire. Berthella stellata (Risso, 1826) from the Mediterranean Sea. It has numerous, small spots on a bright yellow background; there is WILLAN an unpigmented central area in the form of a cross. B. stellata has denticles on the blades of the mandibular elements. Berthella scutata (Martens, 1880) from Mauritius. This species is yellowish with large spots of dark purple-brown on the mantle and a few smaller spots on the oral veil, rhino- phores and sides of the foot; the mantle is evenly and finely granulose all over; like the previous species, B. scutata has denticulate mandibular elements. B. scutata may be a species of Pleurobranchus; Marcus (1977) considers it a nomen dubium. Berthella kaniae Sphon, 1972 from Panama and Mexico. Ground colour is deep yellow (Thompson, 1970). It has a whitish or pale fawn mantle, patterned over the entire surface with minute, pale brown, polygonal markings; this species has two short denticles on either side of the blade of each mandibular element. Berthella kaniae Sphon, 1972 from Pana- ma and Mexico. Ground colour is deep yellow to almost white; the mantle, gill, and veil, as well as the tips of the rhinophores and the area around the genital apertures are spotted reddish-brown. Like B. ornata, it has smooth blades to the mandibular elements, but in B. kaniae the teeth are of similar size and shape across each row of the radula, and the living animal has the capacity to autotomize parts of the mantle edge (Sphon, 1972)—a behaviour never observed in B. ornata. Two of the smaller Pleurobranchus species also have darker markings on the mantle— Pleurobranchus tessellatus Pease, 1868 from Polynesia, and P. ovalis Pease, 1868 from Tahiti and Australia. In Berthella ornata the gill plume is particu- larly long, extending for more than half the body length, the rachis is not smooth but weakly knobbed where a pinna branches off. These knobs appear homologous with the rachal tubercles of Pleurobranchus s.s. Fur- ther detailed studies must be made of the gills of Berthella and Pleurobranchus to see if a separation based on this single characteristic is as clear as has previously been assumed. Both Burn (1962) and Thompson (1970) di- agnose Berthella as having denticles on the blades of the mandibular elements. Yet there are none in B. ornata. A small group of other Berthella species has also been reported to have smooth-edged elements; these are: B. ocellata (Delle-Chiaje) (Vayssiere, 1898: 288); B. kaniae Sphon; B. tupala Marcus (Bertsch, 1975). NEW ZEALAND PLEUROBRANCHIDAE Berthella mediatas Burn, 1962 (Figs. 8, 32-44) 1900. Pleurobranchus aurantiacus Risso; Bergh: 210, pl. 20, figs. 34-38 (non Pleuro- branchus aurantiacus Risso, 1818). 1924. Bouvieria (Pleurobranchus) aurantiaca (Risso); Odhner: 51.—1939a, Powell: 217, No. 1232 (non Pleurobranchus aurantiacus Risso, 1818). 1955. Bouvieria aurantiaca (Risso); Powell: 118 (поп Pleurobranchus aurantiacus Risso, 1818). 1957. Pleurobranchus punctatus Quoy & Gaimard; Burn: 15 (non Pleurobranchus punctatus Quoy & Gaimard, 1832). 1962. Berthella mediatas Burn: 142.— 1966b, Burn: 271, No. 26.—1969, Burn: 80, No. 15 (listed as “Berthella mediata Burn (1962)”). Live animal (Fig. 32) Extended body length to 30mm. Mantle broad, with lateral margins extending beyond sides of foot; when animal is crawling, foot exceeds mantle in length; rhinophores and oral veil project well in advance of raised ante- rior margin. Mantle relatively thick, dull cream or pale brownish-orange, with a slight central darkening due to the underlying digestive 243 gland; occasionally a few small, diffuse, white spots are present as well; shell not usually visible. Mantle highly porous, pores large and very numerous over entire surface, visible in strong light without magnification. Glassy spicules similar to those of B. ornata are pres- ent amongst pores on mantle but require magnification to be seen. Foot hidden all round by mantle except posteriorly; dirty yellow- Orange in color and unspotted; anterior mucous groove makes front border darker; pedal gland present on sole, positioned to- wards one side, appears as a darker, thick- ened, triangular region, with base at back of foot. Oral veil large, narrow at insertion and broadening towards anterior border; tips of grooved lateral areas noticeably forward-pro- jecting. Anterior margin deeply sinuous, with a mid-anterior embayment. Oral veil pale, dirty yellow, lateral areas darker. Rhinophores long, projecting well beyond mantle’s anterior border; pale yellow or Orange. Gill relatively small, never extending behind tail when the animal is crawling, always cov- ered by mantle, attached for half its length to body wall, rachis narrow, smooth; 18-23 pin- nae arise from upper surface and alternate FIG. 32. Berthella mediatas. Length = 17 mm. From Cape Egmont, Taranaki, 16 July 1974. Photograph: G. W. Batt. 244 with a similar number below; pinnae bear smaller pinnules (in a specimen from Porto- bello | counted 12 pinnules on anterior edge of the fourth pinna). Anus opens forward of middle of gill membrane on upper side of gill, approximately between the second and third pinnae; anus not raised on a papilla. Shell (Fig. 8) Shell large, about half total body length; moderately convex, rectangular and not con- stricted towards apical (posterior) end; lateral margins nearly parallel; anterior and posterior margins rounded and a little produced beyond level of protoconch (Fig. 8). Protoconch small, white, just over one whorl, not projecting pos- teriorly but confluent with dorsal surface. One important distinguishing feature of shell is a whitish flange that projects beyond margin on columellar side (i.e. on left back corner when shell is seen dorsally). This ex- pansion is produced below the spire; it is only weakly developed in juvenile shells but very prominent in large shells. Sculpture on outside of shell is of numer- ous, concentric ridges; they often form broad, raised areas with flattened tops. Concentric sculpture is most strongly developed on early and middle regions of teleoconch. Ridges fre- quently intersect each other; each ridge there- fore is of a changing width; ridges run parallel with anterior and right margins, but arch strongly towards posterior margin and left hand edge. Sculpture most clearly visible on smooth area near apex, very similar to that of B. ornata; consisting of numerous, fine, wavy ridges and hollows that become obsolete to- wards the middle part of the shell. | have not yet examined shells with SEM. Shell light golden-brown, with faint, broad, radial streaks of orange-brown; covered externally with a glistening, iridescent periostracum. Shell con- cave internally, inner surface raised into nu- merous, irregular, rounded concentric folds separated by flattened areas. Three shells from specimens taken at Army Bay, Whangaparoa Peninsula, have been deposited in the mollusc collection, Auckland Institute and Museum. Radula (Fig. 42) Radula broad, rectangular, expanded to- wards youngest end; rather small in compari- son with radula of similar-sized Berthella or- nata (largest radula measures—3.0 x WILLAN 1.8mm). Rachidian lacking. Lateral teeth numerous, small and without denticulations; inner and middle laterals (Figs. 39-41) simi- lar, broad-based, erect, and ending in a smooth, hooked apex. Lateral teeth increase progressively from the innermost (8 um high) towards middle (20-30 um), then outer later- als increase to approx. 35 um high (Fig. 42). Outer laterals more erect than inner or middle laterals, blades long, straight, apices sharp, only slightly recurved. Extreme outer lateral teeth needle-like, getting progressively smaller. Numbers of teeth have been remarkably consistent in all radulae examined, with a range of 61 x 65.0.65 to 98 x 76.0.76 (10 radulae). The formula of 56 x 52.0.52 given by Burn (1962) would not appear significantly outside this range. Jaws Jaws approximately rectangular, anterior edges rounded. Mandibular elements (Figs. 43, 44) cruciform, sides of blades strongly denticulate. Base of each element narrow (approx. Ya of total length of 65-70 um), dis- tance between the lateral processes is 30- 35 ит. Strong denticles, which completely occupy sides of blade above lateral proc- esses are most characteristic feature of jaws; denticles coarse, numbers equal or unequal on either side (varying from 3 to 5); denticles point forward, separated by deep grooves; largest denticles immediately adjacent to sharp-pointed apex. Tips of lateral denticles and apices of elements themselves thick- ened. Specimens from the North Island and most South Island localities | have examined show little variation in the denticulation of the man- dibular elements, and Bergh (1900) illustrated identical elements in a specimen taken near the Chatham Islands. However, two of the jaws from South Island specimens (from Akaroa Harbour and Portobello) have mandi- bular elements without the characteristic den- ticulation of the blades and these elements are relatively broader (50-60 um long, 30- 35 um wide). Sides of blades are irregular but not denticulate, apices are blunt and unthick- ened; sides taper towards tips. Lateral proc- esses have a relatively larger area of contact with each other than in denticulate elements. Many more jaws require study before it is possible to say whether these variations in NEW ZEALAND PLEUROBRANCHIDAE 245 Р.. FIGS. 33-38. Berthella mediatas. 33. Radular teeth, а = outer lateral, b = middle lateral, с = inner lateral. 34. Group of mandibular elements on inner face of jaw. 35. Reproductive organs in situ from the undersur- face. 36. Detail of seminal receptacles in a mature animal. 37. Spawn coil. 38. Composite view of structure of reproductive organs. Abbreviations: al. = albumen gland; amp. = ampulla; b.c. = bursa copulatrix; d.gl. = digestive gland; ni. = nidamental glands; o.t. = ovotestis; p. = penis; p.gl. = penial gland; r.s. = receptacu- lum seminis; v.d. = vas deferens. mandibular elements are discontinuous as it would appear, or continuous. Reproductive system (Figs. 35, 36, 38) General arrangement similar to that of Berthella ornata. Receptaculum seminis of B. mediatas an irregular, club-shaped organ arising relatively farther back up vagina, its duct is longer and the receptaculum itself rela- tively larger. Penial gland elongate, thin- walled, with a recurved tip, relatively longer than that of B. ornata. Egg coil (Fig. 37) loosely spiralled, of about two whorls, flanged, white despite the yellow Or brown coloration of B. mediatas animals. Distribution Berthella mediatas has a continuous distri- bution throughout New Zealand. Most speci- mens have been collected from the under- sides of intertidal stones and so far there are very few sublittoral records. | suspect the spe- cies does not extend deeper than 20 m. There are several literature records of “Bouvieria aurantiaca (Risso)” from southern New Zealand localities: Port Pegasus, Stewart Island (Odhner, 1924); near the Chatham Islands (Bergh, 1900); Masked Island, Auckland Islands (Odhner, 1924); Auckland Islands (Powell, 1955). | have not been able to examine these specimens and NEW ZEALAND PLEUROBRANCHIDAE 247 the only description is Bergh's. However, these are probably all Berthella mediatas be- cause the only other yellow or orange pleuro- branch from shallow water, Berthellina citrina, does not extend farther south than East Cape (North Island). Bergh's (1900) description and figure of his specimen (preserved in sublimat- ed picric acid) are quite sufficient to identify it as B. mediatas, although his description is incorrect regarding the site of the anus. Bergh’s Chatham Islands locality was con- firmed by the specimen taken on Rangitira Island in 1977. Berthella mediatas also occurs in Australia: Tasmania; South Australia; Victoria; south Western Australia (Burn, 1962, 1966b, 1969, and personal communication). It is the com- monest pleurobranch along the Victorian coastline. Thompson (1970) did not record it from eastern Australia, and | found none there during visits in 1975, 1979 and 1980. The New Zealand specimens represent a new lo- cality record. Since B. mediatas is already known to have such a wide distribution around the temperate Australian coastline, | would still expect it to be found in southern New South Wales. Discussion Confirmation of the presence of Berthella mediatas in New Zealand has been achieved by examination of three Australian specimens kindly sent by Mr. R. Burn. Data are as fol- lows: Lorne, Victoria—1 under a stone in a channel on rock platform, R. Burn, 24 Nov. 1974; Warneet, Westernport Bay, Victoria—2 on undersides of stones in sandy area, R. C. Robertson, 15 Sept. 1968. The preserved specimens measured 10.2, 17 and 17mm long respectively. Not only do they agree with the New Zealand material, but also they have enabled clarification of several points raised by the original description (Burn, 1962). Pedal glands were present on both the larger speci- mens and the mandibular elements of all spe- cimens had three to five strong denticles on either side of the blade. Both these observa- tions contradict the original description and remove objections regarding the similarity of Australian and New Zealand material. In the Australian specimens the anus opens within ze the anterior third of the gill basement mem- brane, generally close to the insertion of the third pinna. This is even further forward than Originally described by Burn, he mentioned the anus as being “at the mid-length of the gill membrane.” The name “Bouvieria aurantiaca (Risso, 1818)” which has previously been applied in- discriminately in New Zealand to both Ber- thellina citrina (Ruppell & Leuckart) and Berthella mediatas Burn should be reserved for a species of Berthella from the Mediter- ranean Sea. The association of the name Berthella aurantiaca (Risso) with New Zea- land stems from Bergh’s (1900) usage. Since then the name has become entrenched and appears in works by Suter (1913), Odhner (1924), Powell (1937, 1939a, 1946, 1955, 1957, 1961, 1976, 1979) and Morton & Miller (1968). Bergh (1900) held the view that the occurrence of B. aurantiaca in New Zealand was indicative of one of the more widely- spread forms of opisthobranchs. Probably be- cause of Bergh’s reputation and lack of further detailed studies by others, this identification was never challenged. Bergh missed the sub- tle but significant characters that separate B. aurantiaca from B. mediatas, and then Suter (1913) compounded the errors by confusing B. mediatas and Berthellina citrina. Holding no doubt to the same view expressed above, Bergh (1898) reported Berthella aurantiaca from Norway, but this record was also subse- quently rejected (Odhner 1939). The two orange pleurobranchs that have been confused in New Zealand can easily be separated since they differ in numerous char- acters, the most significant of which are sum- marized in Table 2. Berthella mediatas shows considerable similarity to some of its congeners in other parts of the world. Fortunately, in New Zea- land its separation from B. ornata is straight- forward. B. mediatas appears to have its closest relations in Australia and Europe; separation from these species can be made by reference to Odhner's (1939) and Burn's (1962) thorough studies. The only species that | will distinguish it from is B. aurantiaca (Risso). According to descriptions given by Vayssiere (1898) and Odhner (1939), B. aurantiaca is apricot-orange when alive, al- FIGS. 39-44. Radula and jaws of Berthella mediatas. 39. Curved inner lateral teeth near midline of radula (midline at top left); note single row of malformed teeth. 40. View of rows of innermost lateral teeth on either side of central row viewed from above; note absence of central row. 41. Detail of curved inner lateral teeth; midline is at upper left; weak serrations on teeth in foreground are artifacts caused by scanning beam. 42. Detail of erect outer lateral teeth. 43. Low-power view of mandibular elements on inner surface of jaw. 44. Detail of same. 248 WILLAN TABLE 2. Distinguishing features between Berthellina citrina (Rüppell & Leuckart) and Berthella mediatas Burn. Character Mantle texture Shell Smooth, small pores length Anal opening Anterior margin oral veil Almost straight Pedal gland None Radula upper third Mandibular elements ally with weak denticles Prostate gland Present Gill Brownish-cream Spawn coil White Berthellina citrina Small, triangular, approx. 1/5 body Hind end of gill membrane Teeth erect, elongate, denticulate on Blades generally smooth, occasion- Berthella mediatas Many large pores Larger, auriculate, approx. 1/2 body length Anterior third of gill membrane Sinuous, deeply cleft mid-anteriorly Present in adults Teeth shorter, curved, smooth Both sides of blade usually strongly denticulate None Golden-yellow Golden most reddish on top. The mantle is translu- cent and reveals the very large shell beneath; the shell may be up to one-half of the length of the living animal. The anus is at the hind end of the gill membrane. Berthella aurantiaca ap- pears to be restricted to the Mediterranean Sea (Bergh, 1892; Vayssiere, 1898; Schmeckel, 1968; Barash & Danin, 1971). Bathyberthella Willan, n. gen. Definition Small pleurobranchs; abyssal; body ellipti- cal; mantle smaller than foot, free all round, without an anterior crenulation, smooth; pedal gland present; rhinophores arising together mid-anteriorly; anus at posterior end of gill membrane; shell internal, very large—cover- ing entire viscera, flexible, cuticular, without any calcification; radular teeth very numer- Ous, narrow, erect, smooth, similar across each row, rachidian lacking; mandibular ele- ments oval or elliptical at jaw surface, anterior margin irregularly denticulate; vas deferens dilated into a prostate gland; penis smooth, without accessory structures, penial gland present. Type-species: zelandiae Willan, n. sp. Bathyberthella The description of this new pleurobranch is included in this present paper so that it is complete for all the known New Zealand Notaspidea. Bathyberthella zelandiae is not likely to be taken often because of the depth at which it lives and its fragility. All Known specimens were collected by the New Zealand Oceanographic Institute’s re- search vessel “Tangaroa” whilst it was samp- ling the benthos of the Bounty Trough on the “Canyon Coral '79” cruise. All the specimens came from two stations close to each other in abyssal depths (>1600 m); they were taken with an epibenthic sled sampling device. A total of 43 specimens was sorted from the sta- tions. | was fortunate to be present when these pleurobranchs reached the surface in the trawl and were washed from the substratum of light grey ooze. All the animals were mori- bund when they were sorted from the sample. Those that appeared least damaged were placed in fresh sea water but they did not re- cover. Despite their moribund condition (all had partially everted buccal masses), the specimens were not dead and it was possible to give an account of the live animal from those that were recovered intact. Bathyberthella zelandiae Willan, gen. & sp. nov. (Figs. 45-56) Live animal Extended body length up to 40 mm. Speci- mens of all sizes from 5 to 40 mm were repre- NEW ZEALAND PLEUROBRANCHIDAE 249 sented in this collection. Body oval, globose, very flaccid. Mantle rather delicate, with a def- inite free border all round, edges thin; surface smooth, no pores apparent; spicules absent; mantle shorter than rhinophores in front, or tail behind. Foot spongy, pedal gland present on sole; pedal gland a large thickened area occupying posterior section of sole but not ex- tending to foot margin, oval, with posterior end wider, oriented at right angles to longitu- dinal axis of body. Eyes conspicuous, at base of rhinophores, usually large for an abyssal mollusc. Rhinophores rather short, covered with mi- nute papillae. Oral veil short, anterior margin smooth (i.e. lacking digitations); weakly em- bayed mid-anteriorly; upper surface papillate (papillae smaller and fewer than on rhino- phores). Gill relatively small in proportion to body length, never extending to level of hind end of mantle; free for about half its length; rachis narrow, smooth; 19-23 pinnae on (upper side of) rachis (mean for 8 specimens—21 pin- nae). Anus opens on upper side of gill at hind end of basement membrane; interior longitu- dinally ridged. Body creamish, salmon anteriorly; mantle translucent, cream, marked with small, vague, white flecks and speckles, yellow spots occa- sionally present; digestive gland appearing as a black smudge posteriorly; gill brownish; pro- boscis salmon. Shell (Fig. 45) Shell present beneath mantle in all speci- mens examined; very large (e.g. 25 x 16mm in a specimen of 30 mm preserved length); covering entire visceral mass (i.e. it reaches level of front of rhinophores anteriorly). Shell entirely cuticular (without any calcification), very thin and fragile, easily deformed in any direction, disturbance of liquid surrounding shell causes crumpling. Shell concave, broad- ly oval, a little narrower posteriorly; proto- conch not produced beyond posterior margin; shell lacks a posterior flange. Surface flat, concentric growth lines are the only sculpture present, chitin shows localized crumpling Caused by compression of overlying mantle during fixation, radial folds present towards margin (particularly anteriorly); a series of regular, undulating ridges present over a small area near apex but ridges are absent over rest of teleoconch. Shell not connected FIG. 45. Shell of Bathyberthella zelandiae. Length 23 mm x width 18mm. Specimen from 1676 m, northern side of Bounty Trough, 26 Oct. 1979. to body muscles or underlying integument. Shell shining, transparent or faintly yellow. Two shells from paratypes (NZOI Stn. $152) have been deposited in the wet section of the mollusc collection, National Museum of New Zealand. Radula (Figs. 46-49) Buccal mass extremely long, able to be pro- truded up to half body length. Radula square in appearance, rather short (up to 5 mm in length) and extremely broad through having a very great number of teeth per row. Innermost 40 rows slope very acutely towards midline, remaining rows nearly perpendicular to mid- line. Central tooth absent. Lateral teeth very numerous, fine, similar in structure across rows, tall and elongate, tapering gradually to a slightly recurved apex; no denticles on blade whatsoever; base triangular, with a thickened posterior area and thin, forward-produced ex- tension, base about five times as broad as middle section of blade. Inner lateral teeth average 120 um high (Figs. 46, 47). Middle laterals proportionately higher (155 um) having base not significantly enlarged. Outermost laterals shorter than middle laterals (95-115 um); teeth in outer- FIGS. 46-49. Radula of Bathyberthella zelandiae. 46. Lateral teeth near midline of radula (midline is just to left of centre). 47. Lateral teeth contiguous to those in Fig. 46. 48. Outermost lateral teeth from extreme edge of radula. 49. Detail of same. most 20 rows become very narrow (only 6 um wide) and acicular (Figs. 48, 49). Mean number of rows is 62.1 (standard er- ror = 1.4, radulae from six adults examined). The enormous number of closely-packed lat- erals prevents an accurate count being made of numbers of teeth per row with a light micro- scope. Counts of two radulae using a SEM have shown there to be between 210 and 240 lateral teeth. Therefore Bathyberthella zelandiae exhibits the following range of radu- lar formulae for adults: 58 x 210.0.210 to 67 x 240.0.240. Jaws (Figs. 50-53) Pair of jaws lines buccal mass, about 5 mm in length. Jaws lightly chitinized, blunt anteri- orly, tapering to a rounded point posteriorly. Jaws composed of numerous rows of mandi- bular elements; rows very irregular at surface of jaws, some elements run parallel or slightly NEW ZEALAND PLEUROBRANCHIDAE 251 FIGS. 50-53. Jaws of Bathyberthella zelandiae. 50. Group of mandibular elements of jaw from a region near centre. 51. Detail of same. 52. Group of mandibular elements towards edge of jaw; note asymmetry of denticles. 53. Detail of same. obliquely in groups of 10-12, other elements are considerably displaced from their neigh- bors to produce highly disorganized patterns; impression of inner surface of jaws is of rows of erratic elements (shagreened appearance). Elements themselves closely-packed (110 um apart at centre of jaw), elongate or polygonal, bearing a series of thickened, conical denticles along curved, anterior border, the median ones being strongest; elements at centre of jaw possess 4-10 strong denticles arranged symmetrically (Figs. 50, 51); elements towards edges of jaw possess 7-14 weaker denticles that are dis- posed asymmetrically on leading edge (Figs. 52, 53); denticles towards edges are taller and narrower than those near centre, those towards edges curve away from surface in a claw-like manner (Fig. 53). Denticle numbers are very variable, contiguous elements often have differing numbers; all denticles lack smaller subsidiary denticles. 292 WILLAN Reproductive system (Figs. 54-56) All specimens died with tip of penis pro- truded from summit of genital papilla; vaginal and oviduct apertures separate. Ampullar re- gion of hermaphrodite duct rather long, smooth-walled, white, flattened, pressed against nidamental glands; it constricts abruptly as it penetrates genital mass be- tween nidamental glands and bursa copula- trix; hermaphrodite duct gives rise to proximal vas deferens before terminating within nida- mental complex. Proximal vas deferens enlarges after a short distance to large prostate gland; pros- tate gland entwined with penial gland and both are compressed onto bursa copulatrix which they ensheath; prostate moderately thick and spongy; penial gland elongate, tubu- lar, exceedingly thin-walled, fragile, translu- cent, much-convoluted, dilated distally, wall puckered into pockets; penial gland entirely different in structure to prostate gland. Distal vas deferens long, narrow, thick-walled, it passes into large, conical penial sheath; sheath covered with papillae on outside. Vagina moderately long, walls ridged inter- nally, possessing a long, narrow muscle slip medially; passing to both bursa copulatrix and receptaculum seminis. Bursa copulatrix dis- FIGS. 54-56. Reproductive system of Bathyberthella zelandiae. 54. Detail of vagina and seminal recepta- cles. 55. Detail of male section of pallial gonoduct and associated organs. 56. Composite view of structure of reproductive organs, two points marked with asterisks are connected to each other. Abbreviations: amp. = ampulla; b.c. = bursa copulatrix; d.v.d. = distal section of vas deferens; g.p. genital papilla; ni. = nidamental glands; o.t. = ovotestis; o.v. = oviduct; p. = penis; p.gl. = penial gland; pr. = prostate gland; p.s. = penial sheath; p.v.d. = proximal section of vas deferens; r.s. = receptaculum seminis; v. = vagina. NEW ZEALAND PLEUROBRANCHIDAE 253 coidal or club-shaped (Fig. 54), walls exceed- ingly thin, sometimes with weak longitudinal folds. Receptaculum seminis narrow, tubular, much-coiled, approximately equal in diameter to vas deferens; a narrow twisted oviduct originates at base of receptaculum seminis and travels behind retractor muscle to lie at centre of nidamental glands. Distribution Bathyberthella zelandiae is only known from the northern slope of the Bounty Trough, southwest of New Zealand. Its bathymetric range is from 1640 to 1676 m. Type Data HOLOTYPE? 1676im; (45:52:39; 174° 04.9’E, northern side of Bounty Trough, S.W. of New Zealand (Stn. S152), R.C.W. on G.R.V. “Tangaroa,” 26 Oct. 1979 (NZOI, Reg. no. H-342). PARATYPES: seven specimens (all juve- niles), 1640 m, 45°46.0'S, 174°24.5’E, north- ern side of Bounty Trough, S.W. of New Zea- land (Stn. S150) R.C.W. on G.R.V. “Tanga- roa,” 26 Oct. 1979 (NZOI, Reg. no. P-571); 10 specimens collected with holotype at Stn. $152 (NZOI, Reg. no. P-572); 25 specimens collected with holotype at Stn. S152 (NM). Discussion Bathyberthella is a most significant genus within the Pleurobranchidae. This is because В. zelandiae possesses an unexpected amalgam of characters some of which are peculiar to it alone and others which relate it to genera of both the accepted pleurobranch subfamilies. The enormous, uncalcified inter- nal shell separates Bathyberthella at once from all other genera, but the morphology of the shell is probably related to the abyssal existence of the species. B. zelandiae has the external appearance of a member of the sub- family Pleurobranchinae, particularly a spe- cies of Berthella or Berthellina. The reproduc- tive system recalls Berthellina, the teeth are smooth as in Berthella yet elongate and nu- merous as in Berthellina. These pleurobran- chine features strengthen Burn’s (1962) con- tention that the small pleurobranchs (with smooth, non-emarginate mantles and smooth gill rachises) are closer to each other than either is to Pleurobranchus. The mandibular elements of B. zelandiae show a great like- ness to those of Pleurobranchaea (subfamily Pleurobranchaeinae). The bulk of characters favour placing Bathyberthella in the Pleurobranchinae. With- in that subfamily is another equally anomal- ous, yet important, monotypic genus— Pleurehdera. Pleurehdera haraldi is small; it has a smooth mantle and gill rachis; it pos- sesses a pedal gland, prostate gland and penial gland; its teeth are elongate, most pos- sess a single denticle near the apex but the outermost laterals are smooth for about one quarter of the row; the mandibular elements are variable (Marcus & Marcus, 1970). Some of the variation shown by the mandibular ele- ments of Pleurehdera haraldi resemble those of Bathyberthella zelandiae. Even allowing for the abyssal habitat of B. zelandiae, the differences between shells, radulae and re- productive systems of Pleurehdera and Bathyberthella indicate they are not con- generic. | interpret Pleurehdera and Bathyberthella as terminations of narrow lines produced dur- ing the radiation that followed the acquisition of the pleurobranch grade of organization by opisthobranchs. Both these genera stem from near the Berthella/Berthellina dichotomy. The genera Euselenops and Pleurobranchella il- lustrate analogous cases; they probably rep- resent terminations of narrow side lines that Originated on the pleurobranchaeine side of this radiation. | cannot interpret any of these monotypic (or very small) genera as ancestral to any large present day genus. This is be- cause all the small genera possess a mosaic of characters many of which are quite unlike those of their presumed derivatives. One point that emerges is that the oval type of mandibular elements with denticulate anterior borders (as in the Pleurobranchaeinae) pre- ceded the cruciform type of elements (as in the Pleurobranchinae). It is unlikely that Bathyberthella zelandiae could be confused with either of the two small, yellow or orange pleurobranchs from New Zealand that look superficially similar (Berth- ella citrina, Berthella mediatas) since both the latter occur in shallow water on the continen- tal shelf. Listing the many points of difference between these three species would involve repetition, in a comparative context, of diag- nostic characters for each genus and recon- struction of a table similar to Table 2 (p. 248) to incorporate B. zelandiae. Comparisons of the significant characters between these three species will be presented in the key and repe- 254 WILLAN tition of the more subtle points of distinction is unnecessary in view of the abyssal habitat of B. zelandiae and the scarcity with which it is likely to be encountered. Pleurobranchaea Meckel in Leue, 1813 Pleurobranchaea Meckel in Leue, 1813: 11. Type-species by subsequent monotypy (Blainville, 1825: 376): Pleurobranchidium meckelii Blainville, 1825. Pleurobranchidium Blainville, 1825: 372, 376. Type-species by monotypy: Pleurobran- chidium meckelii Blainville, 1825. Koonsia Verrill, 1882: 545. Type-species by monotypy: Koonsia obesa Verrill, 1882. Pleurobranchillus Bergh, 1892: 27. Type- species by subsequent designation (Willan, 1977: 153): Pleurobranchillus morosus Bergh, 1892. Meckel (in Leue, 1813) established Pleuro- branchaea without including nominal species. Blainville (1825) was the first author to de- scribe the species meckelii in the genus Pleurobranchaea, and it is this species ipso facto which becomes the type of Pleurobran- chaea by subsequent monotypy. According to the International Code of Zoological Nomen- clature, the spelling must revert to meckelii (from meckeli) since this is the correct original spelling (Art. 32(1), 1.C.Z.N. 1961). | have considered the synonymy of Koonsia and Pleurobranchillus elsewhere (Willan, 1977). Bergh (1897: 3, note 2; 1898: 64) himself sub- sequently recognized Pleurobranchillus as a synonym of Pleurobranchaea (see Vayssiere, 1901: 74; and Marcus & Marcus, 1957: 25). Definition Moderate-sized to large pleurobranchs with oval or oblong bodies that are blunt anteriorly; mantle reduced, smaller than foot, merging with oral veil anteriorly and tail posteriorly, covered with low tubercles; rhinophores far apart, inserted on either side of head at base of oral veil; oral veil with digitate processes along anterior margin; pedal gland present on posterior part of foot sole in sexually mature specimens; some species possess a caudal spur on dorsal side of tail; anus towards rear of gill basement membrane; shell absent; buccal mass relatively large; radula with cen- tral row, laterals curved, smooth-sided, most with an accessory denticle—either strongly or weakly developed; mandibular elements poly- gonal or rounded at jaw surface, denticulate along anterior edge; penis without accessory leaves; prostate gland present. Remarks As with all the Notaspidea, species of Pleurobranchaea often show considerable intraspecific variability and this has frequently resulted in the creation of spurious species. This cause for uncertainty has been com- pounded by inadequate description of new species based on poorly fixed material. Nev- ertheless there are subtle and consistent dif- ferences separating the valid species. Marcus & Marcus (1957) and Marcus (1957) listed the 18 described species of Pleurobranchaea; they later claimed that P. algoensis Thiele and P. japonica Thiele were unrecognizable (Marcus & Marcus, 1966). Four species have been added since the Marcus first lists: P. hamva Marcus (1957); P. gemini Macnae (1962), P. californica Mac- Farland (1966); P. gela Marcus & Marcus (1966). More material, however, has forced P. hamva to be incorporated into the synonymy of P. hedgpethi Abbott (Marcus & Marcus, 1967b). Much more work needs to be done on this genus; for example, to determine the status of such species as Pleurobranchaea capensis Vayssière, 1898, P. gemini Macnae, 1962, P. brocki Bergh, 1897 and P. agassizi Bergh, 1897. Some new species are known and await description. Because of their large size, high level of activity and variety of relatively complex be- haviours, species of Pleurobranchaea are now frequently used in physiological re- search, e.g. Davis and Mpitsos (1971), Davis et al. (1977), and this is further reason to re- examine the taxonomic status of the various entities. Pleurobranchaea maculata (Quoy & Gaimard, 1832) (Figs. 57-70) 1832. Pleurobranchidium maculatum Quoy & Gaimard: 301, pl. 22, figs. 11-14. 1878. Pleurobranchaea novaezealandiae Cheeseman: 276, pl. 15, fig. 3.—1879, Cheeseman: 378, pl. 16, fig. 3— 1880, Hut- ton: 124.— 1896, Pilsbry: 227, pl. 53, fig. 87.— 1897, Bergh: 150-152, No. 31 and 154-155 No. 33.—1913, Suter: 553.— 1915, Suter: pl. 36, fig. 2.—1933, Allan: 446.— 1949, Baba: 133, pl. 10, figs. 31, 32, NEW ZEALAND PLEUROBRANCHIDAE 299 34.— 1957, Burn: 12, 15.—1957, Marcus & Marcus: 26.—1965, Guang-Yu & Si: 266, 275; 1966, Marcus & Marcus: 177.—1969, Baba: 191—1976, Powell: 112.—1979, Powell: 282, pl. 51, fig. 2. 1896. Pleurobranchaea maculata (Quoy & Gaimard); Pilsbry: 227, pl. 53, figs. 88, 89.— 1897, Bergh: 153-154, No. 32.— 1901, Vayssière: 49-56, pl. 5, figs. 238— 247.—1913, Suter: 552.— 1915, Suter: pl. 23, fig. 17.—1924, Odhner: 52.—1937, Powell: 85, No. 1236.— 1954, Pruvot-Fol: 33.—1957, Marcus & Marcus: 26.— 1958, Burn: 6.— 1966a, Burn: 271.—1966b, Burn: 106.—1969, Burn: 80.—1970, Thompson: 192-195, fig. 10.—1976, Powell: 112.— 1977, Gordon & Ballantine: 40, 112.— 1979, Powell: 283. 1898. Pleurobranchaea novaezelandiae [sic] Vayssiere: pl. 15, fig. 28—1901, Vays- siere: 69-72.—1900, Bergh: 208, pl. 20, figs. 56, 57, pl. 21, fig. 69.—1920, Mesta- yer: 170-171.—1924, Odhner: 52.—1937, Powell: 85, No. 1234.—1961, Powell: 107.—1964, Williams: 20.—1968, Morton 4 Miller: 167, 576, pl. 11, fig. 8.—1969, Bath- am: 78.—1970, Thompson 196.—1972, Morton: 346.—1973, Miller 8 Batt: 19, fig. 78 (image reversed during printing).— 1977, Willan: 154.—1977a, Ottaway: 217- 218.—1977b, Ottaway: 125-130, fig. 1 (error pro P. novaezealandiae Cheeseman, 1878). 1900. Pleurobranchaea novaezelandiae var. granulosa Bergh: 209.—1913, Suter: 554 (as ssp. granulosa).—1937b, Powell: 85, No. 1235 (as ssp. granulosa).—1961, Powell: 107 (as ssp. granulosa). 1933. Pleurobranchaea dorsalis Allan: 445, pl. 56, figs. 4, 5.—1957, Marcus & Marcus: 26 1950. Pleurobranchaea maculata dorsalis Allan: 208, fig. 1. 1976. Pleurobranchaea granulosa Bergh; Powell: 112.—1979, Powell: 282. The original description of Pleurobran- chidium maculatum by Quoy & Gaimard (1832) was brief and dealt only with external features. The description was greatly ex- panded by Vayssière (1901) after a thorough re-examination of the type material. Vayssiere (1901) detected a mistake made by Quoy & Gaimard (1832) with respect to Pleurobran- chaea maculata. Vayssière had examined the (five) specimens collected by “l'Astrolabe” and he expressed astonishment that “Nou- velle-Zélande” was given as the locality on the labels on three of the five original speci- mens. Quoy & Gaimard (1832) stated that all had originated from Port Western to Jarvis Bay, Australia. Vayssière suspected that Quoy & Gaimard had erroneously labelled those three specimens; he did note, however, that New Zealand was included as part of the original locality citation for Pleurobranchaea maculata. My investigations have shown that there is only a single species of Pleurobranchaea in New Zealand, for which the name Pleuro- branchaea maculata (Quoy & Gaimard) has priority. No consistent differences can be found to warrant the continued separation of New Zealand specimens under the names of P. novaezealandiae Cheeseman or P. novaezelandiae granulosa Bergh (see be- low). Several other authors have already an- ticipated the synonymy of P. novaezea- landiae with P. maculata. In her discussion, Pruvot-Fol (1954: 33) concluded by reiterating Vayssière’s remark that the two might well be variants of a single species. Burn (1958) synonymized P. novaezealandiae and the Australian P. dorsalis Allan with P. maculata. Thompson (1970) suggested that all the Australian records of Pleurobranchaea are based on P. maculata (Quoy & Gaimard). Live animal (Fig. 57) Body large, elongate, oval, covered dorsal- ly by a relatively small mantle beneath which foot projects all round. Mantle confluent ante- riorly with oral veil, posteriorly it is continuous, over a relatively narrow area, with tail; sides of mantle free, left border held close to body, right slightly larger, held away from body to accommodate _ gill underneath. Mantle smooth, soft, upper surface (and that of oral veil) entirely covered with minute puckers and folds. Ground colour varies from pale grey to almost ash-black, broken by irregular, pale, grey-white areas to give a dappled-reticulate appearance. Raised areas sprinkled with nu- merous, minute, and almost microscopic, white dots; hollows between these raised areas darker. Some individuals uniformly dark, others have paler areas and a patchy appearance. Auckland east coast populations appear to show approximately equal propor- tions of dark and dappled morphs. Undersur- face of free mantle borders smooth, grey, 256 WILLAN FIG. 57. Pleurobranchaea maculata (Quoy & Gaimard). Length = 132 mm. Specimen from 7-11 m, off middle head of German Bay Hill, Akaroa Hbr., Banks Pen., May 1962. Photograph: M. C. Miller. regularly speckled with white dots. Mantle dis- charges a clear fluid of pH = 1. Oral veil trapezoidal, lower surface smooth- er than upper surface; it bears some raised pustulose areas forming weak reticulations; anterior border of veil broad, straight, wider than widest part of body when animal is crawl- ing. Anterior corners project laterally beyond front edge. Lateral areas marked dorsally by a white line. On undersurface, a groove extends length of lateral areas; here each area ap- pears as a brownish streak. All along anterior edge of oral veil are many, small, branched processes; therefore oral veil of Pleurobran- chaea species is more elaborate than that of Berthellina or Berthella species. The in- creased structural complexity is accompanied by increased activity; when animal is crawling actively, anterior border constantly ripples as it explores substratum in different areas. Rhinophores arise laterally where the man- tle passes into oral veil; mottled grey and white, tipped with white. Gill large, conspicuous, generally its hind end extends beyond overlying right mantle flap; basement membrane attached to gill for approximately half its length. Gill rachis bears 20-28 alternating pinnae (mean number for 9 specimens 23.0); pinnae relatively longer than in species of Berthellina or Berthella; pinnae bear large numbers of closely-packed pinnules (25-30 on anterior pinnae); rachis itself broad, flattened, with small, irregular knobs and lumps; however, rachises of pinnae round, smooth. In living specimens, gill is light ash-grey, with rachis and rachises of pinnae overlain with tiny black specks. This superficial speckled layer rubs off with ease in newly dead specimens, re- vealing a subepidermal layer of larger, opaque, white spots which does not rub off. Anus opens a short distance in front of hind end of the gill basement membrane. Pre- branchial aperture conspicuous, on a papilla before front of gill rachis, interior ridged. Foot broadly rounded anteriorly, more pointed behind; upper surface similar in colour and pattern to mantle; foot borders thin and with white puckers that are more regular than on mantle. Anterior dorsal surface with a clearly demarcated, broad, greyish mucous groove; this area being wider than area of the foot overhung by oral veil. Mucous groove has a deep, mid-anterior cleft perpendicular to its axis. Sole of foot very large, smooth, pale ash-grey, lightly speckled with white, margin NEW ZEALAND PLEUROBRANCHIDAE 257 white. Реда! gland present posteriorly in sex- ually mature individuals, lying beneath foot epithelium, length about 1/6 total length of the foot; pedal gland somewhat asymmetric and triangular with base (closest to tail) shorter than sides; base does not reach extremity of tail. The white buccal mass and dark digestive gland are visible through transparent tissue of foot sole. Caudal spur never present on dor- sal surface of tail. Pleurobranchaea maculata is able to float upside-down on the meniscus in a still con- tainer of water, then lateral and posterior foot margins show slight indentations. 58a 59 boat Animals are normally encountered at rest on undersides of stones, foot and body con- tracted and humped up; mantle rounded, gill extending beyond right border, pinnae twitch- ing occasionally quite independently of each other; oral veil contracted against anterior mantle edge; rhinophores held laterally in groove formed by oral veil and mantle. Radula (Figs. 58-63) Radula large (8.4 mm long for a specimen 37 mm in extended length), relatively broad, deeply grooved in midline, with many rows of | 58b FIG. 58, 59. Radula of Pleurobranchaea maculata. 58a. Representative teeth from a half row; 58b. Middle lateral tooth showing detail of basal attachment region; 59. Teeth from a half row (after Vayssière, 1901, pl. 5, figs. 238-240, 242). 258 WILLAN lateral teeth symmetrically on either side of central row. Radula of adult P. maculata has 40-50 rows. Rachidian inconspicuous, short (approx. 30 um high), narrow, with a rounded base and pointed apex; apparently constant in length in all rows. Careful preparation is necessary to keep rachidian on radula; de- spite its presence in all rows, it fell away in most cases during preparation of radula. There are 70-80 lateral teeth within each half-row; laterals undergo characteristic changes in size and structure across rows. Outermost lateral teeth (Fig. 63) small, peg- like, narrow, rounded apically with apices slightly recurved; extreme outermost teeth (number 80 counting from midline) smallest, approx. 48 um high. By 65th lateral, a small spike is present on inner side near base of tooth, cusp of tooth itself sharper; at this point teeth are up to 60 um high. Teeth continue to get larger, and lateral spike enlarges to a strong accessory denticle increasing in size, so that by 41st tooth (approx. mid-way across row) denticle is 64 um high, the tooth itself being 298 um high. In this middle region teeth have large, broad bases, sides parallel to point where denticle arises, beyond there sides curve inwards to sharp apex; lateral denticle straight-sided and erect (Figs. 61, 62). To the inside of this region, teeth again slowly decrease progressively although later- al denticle remains nearly same height. Close to rachidian there is a rapid decrease, teeth and denticles decreasing proportionately, second lateral approx. 60 um high and still bears a small accessory denticle. A radular formula typical of an adult is 43 x 74.1.74. The range of radular formulae for adult New Zealand specimens examined is 40-50 x 70-80.1.70-80. Radular formula in- creases as P. maculata grows; a juvenile (crawling length 5mm) from Cape Three Points, Akaroa Harbour, had a formula of 25 x 39.0.39; tooth height also increases propor- tionately with growth. System for attachment of teeth to basement membrane like that of Pleurobranchaea cali- fornica (MacFarland, 1966: 97), although the basal facets are knobbed rather than hook- like in P. maculata (Figs. 61, 63). System for support of teeth along rows also present; outer laterals (Fig. 63) have a socket towards base where right-angled elbow of base of tooth in front fits, same interlocking arrange- ment present between middle lateral teeth (Fig. 61). Tooth structure is a very constant and char- acteristic feature of Pleurobranchaea macu- lata, particularly the strong, accessory lateral denticle on middle lateral teeth. P. californica has very small, weak denticles (MacFarland, 1966). | have redrawn Vayssiere's (1901, pl. 5) illustration of isolated radular teeth from specimens of the type lot of P. maculata (Fig. 59); they show a strong resemblance to New Zealand material. Jaws (Figs. 64, 65) A pair of elongate, chitinous jaws lines in- side of buccal bulb; rectangular; hind portion expanded and rounded, anterior margin deeply sinuate, with a forward-projecting spike towards centre. Many tightly-fitting, polygonal elements, in the form of closely- packed columns make up jaws; surfaces of elements flattened, resembling interlocking paving stones; elements arranged in alternat- ing rows (about 60 per row), mostly hexago- nal though some are pentagonal or round. From the surface, base and sides of elements appear smooth, sides are parallel, slightly longer than the base; there are no lateral pro- jections. On broadly convex anterior face are 5-12 small, pointed denticles; denticles not positioned symmetrically about anterior edge. Central area on inner face is irregularly pustu- lose (Fig. 65). Average dimensions of ele- ments are: length 40 um; width 60 um. MacFarland (1966) studied the develop- ment of mandibular elements in Pleurobran- chaea californica. Reproductive system (Figs. 66-69) To expose the reproductive system in dis- section, the network of anastomosing tubules of dorsal accessory gland must first be re- moved; these fine tubules surround all organs of reproductive system and gut. Because the м FIGS. 60-65. Вадша and jaws о! Pleurobranchaea maculata. 60. Teeth on one half of the radula; outermost laterals are on the extreme left. 61. Detail of large, middle lateral tooth; note accessory denticle on each tooth. 62. Detail of same; longitudinal striations are probably dried mucus. 63. Detail of small, peg-like outermost lateral teeth. 64. Mandibular elements on inner face of jaw. 65. Detail of the same. NEW ZEALAND PLEUROBRANCHIDAE 259 260 WILLAN FIGS. 66-70. Reproductive system of Pleurobranchaea maculata. 66. Composite view of structure of reproductive organs. 67. Reproductive system of a paratype (after Vayssière, 1901, pl. 5, fig. 247). 68. Detail of fully everted genital organs. 69. Orientation of genital organs during reciprocal copulation, arrows indicate direction of sperm injection into bursa copulatrix of partner. 70. Spawn mass. Abbreviations: al. = albumen gland; amp. = ampulla; b.c. = bursa copulatrix; g.p. = genital papilla; m.p. = muscular pocket for vas deferens; mu. = mucous gland; o.t. = ovotestis; ov. = oviduct; p. = penis; p.l. = posterior lobe of everted genital papilla; pr. = prostate gland; p.r.m. = penial retractor muscle; v. = vagina; v.d. = vas deferens. structure of the reproductive organs differs in many ways from that of species of the Pleuro- branchinae, it is described here in full. Ovotestis dorsal, applied to, but not inter- lobed with, digestive gland; creamish-white, compact, appearing granular because of elongate acini stacked side by side. Inside ovotestis, fine white collecting ducts can be traced amongst acini; these smaller ducts join larger and larger ducts, finally to a still slen- der, sinuous, white hermaphrodite duct with ovotestis on its anterior ventral surface. Im- mediately upon leaving ovotestis, hermaph- rodite duct greatly increases in diameter and wall thickness to form ampulla. Ampulla lies next to foot musculature, much coiled and convoluted; it gradually enlarges as it passes towards ventral midline. Near midline, ampul- la turns forward (below nidamental complex), and abruptly narrows to about Ya of its previ- ous diameter, now becoming quite straight; it branches into two, and, after becoming a very short, but slightly wider proximal vas defer- ens, enters prostate gland. Oviduct is the other of these branches; at its point of origin from hermaphrodite duct (be- side the prostate gland), oviduct is large, its walls deeply constricted, of a greater diameter than hermaphrodite duct, thickest approx. Vs of the distance along its length, after that di- ameter is halved. Oviduct next travels as a straight, whitish tube beneath bursa copulatrix to join much larger vagina before the latter passes into bursa copulatrix. Bursa copulatrix varies in position and shape; it lies, in situ, in midline, either dorsal- ly, above the buccal mass and visceral gan- glia, or ventrally, below and next to them. NEW ZEALAND PLEUROBRANCHIDAE 261 Bursa generally discoidal or club-like, flat beneath and convex above. Vagina com- pletes a half loop on dorsal side of bursa be- fore passing into it; end of the vagina connect- ed to bursa is neither swollen nor constricted. Recepaculum seminis absent. Vagina slowly increases in diameter as it passes, in a semicircle, beneath all the repro- ductive organs, to exterior. Vagina is tube of greatest width in reproductive system; it is whitish, its walls are tough and unconstricted. Nidamental gland complex situated besides vagina—mucous gland orange, relatively thin- walled; albumen gland larger, more solid, cone-shaped, walls much-convoluted. Vas deferens is a continuation of hermaph- rodite duct from the point at which oviduct de- parts; proximal vas deferens is visible for only a very short distance before entering prostate gland. Prostate gland irregular in outline and size, spongy, consisting of many tightly- packed vesicles. Distal vas deferens leaves prostate gland, its walls thickened and shin- ing. Just beside the penis, distal vas deferens Curves backwards into an elongate pocket of clear and tough tissue; pocket passes back- wards above prostate gland, female glands and digestive gland, then travels to foot where retractor muscles attach it to floor of body cav- ity; this pocket is the most conspicuous part of entire genital system. Distal vas deferens lies within pocket, it often loops over itself, it re- tains its same diameter throughout entire course after leaving prostate gland. Towards the end of its course back inside pocket vas deferens expands gradually into penis. Penis is long, narrow, smooth-walled, quite smooth, capable of eversion to a great distance. The organization of the reproductive sys- tem in New Zealand Pleurobranchaea mate- rial presents one of the strongest reasons for believing it to be conspecific with P. maculata (Quoy & Gaimard). The system detailed above (Fig. 66) is almost identical with that described by Vayssière (1901) for specimens from the type lot (Fig. 67). The reproductive system in this genus is highly diagnostic for each species. In Pleuro- branchaea meckelii the pocket for the vas deferens is shaped like a large, inverted cone and the retractor muscle is inserted on the body wall; in this species too, the oviduct has a short proximal tube followed by two ovoid swellings; the shape of the bursa copulatrix is also different. The enigmatic caudal spur is present on the post-dorsal surface of the foot. Behaviour at copulation and oviposition Not only is the organization of genital or- gans important in species recognition, but also reproductive behaviour is highly species- specific. Pleurobranchaea species tend to be solitary, so mating encounters occur seldom and have been reported very infrequently in the literature. Davis & Mpitsos (1971) give de- tails of copulatory behaviour in P. californica. The following account describes the mating behaviour of P. maculata in detail. When two mature specimens of Pleuro- branchaea maculata encounter each other, one of two behaviour patterns results. Upon contact one animal rapidly everts its oral tube and makes feeding thrusts towards the other; often removing a piece of flesh. Most attacks of this kind appear to be directed at the tail region. This type of behaviour most often re- sults from initial contact of the oral veil of one specimen, with some other part of the other specimen. The attacked animal may let go of the substratum and swim with head-to-tail flexions to avoid further attacks. The second response often occurs when two specimens meet head first. After the oral veils contact each other, forward motion slows, both partners raise their right mantle edges, and partially evert their genital organs. This is a very characteristic sexual posture. It is often followed by mutual circling during which the genital organs are fully everted, and the long, whip-like penis is rapidly everted and thrust in the general region of the reproductive apertures and gill. In the laboratory, following an encounter of this type, one partner frequently fails to recip- rocate and crawls away. But if both animals are receptive at that time, circling and penial thrusting continue. The penis is retracted and fully everted repeatedly but the posterior swelling that exposes the vagina remains everted and distended. Behind the vagina this swelling becomes lobed and narrows to a tail- like extremity (Fig. 68). When the penis of one partner is correctly thrust towards the vaginal entrance, the penis enters and passes a considerable distance into the vagina (the tip probably reaches the base of the bursa copulatrix). Copulation is most often reciprocal by this stage, and Fig. 69 shows the position of everted genitalia and penial insertion in such a reciprocal copu- lation. Copulation lasts less than two minutes. Oviposition lasts one or two hours. The egg 262 mass (Fig. 70) is usually in the form of a loose and irregular coil of approximately 172 whorls, but this pattern is not circumscribed as in the Pleurobranchinae; some egg masses are ir- regulary looped, others open. Larger individu- als lay larger coils. The egg mass is circular in cross section and is composed of a clear, ge- latinous outer area inside which the long string of white eggs is tightly coiled; coiling is generally oblique to the outer wall. The egg mass is attached to the substratum by a nar- row, opaque-white strip which frequently per- sists after the mass has disintegrated. Mestayer (1920) and Graham (1941) have recorded observations and given photo- graphs of aquarium spawning in Pleurobran- chaea maculata. Larvae hatch after approximately 10 days; they are planktonic veligers with a small cup- like shell and operculum. Veliger shells are 180-200 um long and approx. 120 um wide. In aquaria, settlement was observed after six days but could be delayed for at least one week. The larvae can feed in the plankton; larvae, experimentally fed with a suspension of Dunaliella primolecta, ingested these cells. Further aspects of the planktonic life and set- tlement of veligers were not studied. Distribution Pleurobranchaea maculata occurs through- out New Zealand. It lives both intertidally and subtidally (to at least 300 m) and appears to be equally abundant through that entire bathymetric range. Specimens are most fre- quently encountered nesting in depressions on the undersurfaces of stones; when uncov- ered they immediately start crawling to safety. Pleurobranchaea maculata is more toler- ant than other pleurobranchs of waters carry- ing suspended silt, and thus appears in har- bours and estuaries where the others are ab- sent. This is probably also correlated with food; P. maculata is an opportunistic carniv- ore and can take advantage of a wide range of prey species, whereas other pleurobranchs feed on species of sponge or ascidians which are themselves confined to clear water situa- tions. Pleurobranchaea maculata also occurs in southern Queensland (Burn, 1966a), New South Wales (Allan, 1950; Thompson, 1970), Victoria, Australia (Burn, 1957, 1966, 1969), China (Tchang-Si, 1934; Guang-Yu & Si, 1965), Sri Lanka (White, 1948) and Japan (Baba, 1937, 1949, 1969; Baba & Hamatani, 1952). WILLAN Discussion Some discrepancies exist in the description of specimens of Pleurobranchaea from New Zealand. For instance, in the diagnosis of P. novaezealandiae, Cheeseman (1878) stated: “branchial plume often over an inch in length, and free for half that distance; pectinations about 17, finely ciliated.” This number of pin- nae is well below the average (23) for the species based on material | have examined, but the distal pinnae are so small Cheeseman probably missed them in his count. Quoy 8 Gaimard's (1832) original figure of Pleurobranchaea maculata, later reproduced by Pilsbry (1896), Vayssiere (1901) and Suter (1915), was inexact. It was drawn from a moribund, or dead, specimen as evidenced by the everted penis. Vayssiere (1901) gave a detailed account of the reproductive system, jaws, and radula of the type material, all of which correspond well with the same organs for New Zealand specimens. Cheeseman's (1878, 1879) original illustration of P. novae- zealandiae, reproduced subsequently by Pilsbry (1896), Vayssiere (1898) and Powell (1979), is slightly inaccurate too; the mantle appears not to be confluent with the oral veil, but this is merely due to the angle from which the animal has been drawn. Bergh (1900) originally differentiated Pleurobranchaea novaezelandiae var. granulosa on the basis that the back and tail were covered with small, round and oval granulations (0.5-1.0 mm diameter). This dif- ference in surface texture only reflects the de- gree of variation of this character in P. macu- lata and it is accentuated by the fixation pro- cedure adopted. Baba (1937) realized this and included P. novaezelandiae var. granu- losa in P. novaezealandiae. Other differences do exist between var. granulosa and novae- zealandiae according to Bergh (1900) in the shape of the jaws and ampullae of the salivary glands; but these differences do not merit taxonomic separation. They can be explained by intraspecific variation and by fixation tech- niques respectively. Bergh gave no illustration for his single specimen from French Pass. | have examined specimens in the Suter Col- lection (National Museum of New Zealand) from Te Onepoto, labelled as P. novaezea- landiae granulosa, and can find no consistent differences from typical P. maculata. Pleurobranchaea maculata is similar to the Mediterranean species P. meckelii, but there are several points of difference: sexually ma- ture P. meckelii has a caudal spur, there is NEW ZEALAND PLEUROBRANCHIDAE 263 none in P. maculata; in P. maculata the re- tractor muscle is attached posteriorly to the floor of the body cavity whereas in P. meckelii it is attached to the body wall dorso-laterally; there is a significantly greater number of denticles (11-15) along the anterior edge of the mandibular elements in P. meckelii; the shape of the rachidian tooth is also different in the two species. Pleurobranchaea maculata differs from P. californica in details of the radula and reproductive system. An Ectoparasitic Nematode from Pleurobranchaea maculata Whilst examining the external features of specimens of Pleurobranchaea maculata col- lected from beneath subtidal stones in Leigh Harbour (23 Nov. 1973), | discovered that the slugs each carried several ectoparasitic nematodes. One had 18 nematodes, most were on the undersurface of the foot with a few on the mantle, none was on the rhino- phores, oral veil or gill. Later | found more than 30 embedded in the foot sole of a P. maculata from Goat Island Bay. The nematode has a long and relatively broad body which is bent to form a loop. There are pronounced adhesive organs on the tail. Specimens had their tails buried a few millimeters in the P. maculata tissue, and when removed most carried away some mollusc flesh. Dr. W. C. Clark examined the nematodes and kindly informed me they probably be- longed to the order Monhysterida but this could not be confirmed because all were im- mature. The presence of these nematodes on Pleurobranchaea maculata raises questions regarding their life history patterns and possi- ble specificity. KEY TO NEW ZEALAND NOTASPIDEA 1. Shell external, circular and flattened, limpet-like; body large and warty Shell internal or absent; body elongate, slug-like Umbraculum sinicum (Gmelin, 1791) . Shell present beneath mantle; rhinophores arise together; mantle large, smooth; simple DOC toro VE: e Rue (ay NOTES Sli о ee ов 3 Shell absent; rhinophores widely separated, at sides of head; mantle small, greyish, its surface puckered and wrinkled; anterior border of oral veil with numerous branched Papiliaem er ado Pleurobranchaea maculata (Quoy & Gaimard, 1832) Fig. 57 . Shell relatively large (up to Y2 body length), auriculate; radular teeth simple, without AENA ION IH I ee RE. RE AR een 4 Shell relatively small (1/5 to 1/4 body length), triangular; radular teeth with a series of denticulations on posterior margin; mandibular elements (generally) smooth; prostate gland present; mantle lemon-yellow to apricot with scattered white specks; no pedal gland'ontsole of foot... Berthellina citrina (Rúppell & Leuckart, 1828) Figs. 1, 5 . Mantle uniform in colour, dull orange or creamish-yellow with a few white specks; mandibularelements:denticulale nro. on se cece stare ee en. 5 Mantle spotted with large, chocolate-brown blotches; anus at rear of gill membrane; mandibular elements smooth .......... Berthella ornata (Cheeseman, 1878) Figs. 20-22 . Mantle with numerous, large pores; shell calcareous; anus near front end of gill mem- brane; mandibular elements cruciform with denticulate blades; prostate gland absent, found intertidally and shallowly subtidally (to ca. 20 m) .. Berthella mediatas (Burn, 1962) Mantle smooth, not conspicuously porous; shell very large, cuticular and without calcifi- cation; anus at rear of gill membrane; mandibular elements oval with denticles on ante- rior border; prostate gland present; occurs only in very deep water (>1000 m) A em are Le tien Bathyberthella zelandiae Willan, gen. et sp. nov. 264 WILLAN ACKNOWLEDGMENTS The bulk of this investigation was conduct- ed at the University of Auckland in 1974. | am grateful to Dr. M. C. Miller for supervision, for literature and use of comparative material from his opisthobranch collections. He has also critically read the manuscript. | thank also the following people for access to collections and loans of specimens: Mr. R. Burn; Mr. W. O. Cernohorsky; Dr. F. M. Climo; Dr. A. W. B. Powell. | also acknowledge with thanks the efforts of the many friends who assisted greatly with this study by collecting pleuro- branchs for me to examine. The SEM photos were taken by Mrs. H. Silyn-Roberts and Mr. B. A. Marshall, the photographs printed by Mr. R. Anderson. | am much indebted to Mrs. R.-M. С. Thompson (NZOI) and Mrs. W. Esam (NM) for typing the manuscript to carefully. REFERENCES CITED ABBOTT, R. T., 1974, American Seashells. Ed. 2. Van Nostrand Reinhold, 663 p. ALLAN, J. K., 1933, Opisthobranchs from Australia. Records of the Australian Museum, 18: 443- 450. ALLAN, J. K., 1950, Australian Shells. Georgian House, Melbourne, 470 p. BABA, K., 1937, Opisthobranchia of Japan (1). Journal of the Department of Agriculture, Kyushu Imperial University, 5(4): 236 p. BABA, K., 1949, Opisthobranchia of Sagami Bay Collected by His Majesty The Emperor of Japan. lwanami Shoten, Tokyo, 194 p. BABA, K., 1969, List of the Pleurobranchidae and Pleurobranchaeidae from Japan. Collecting and Breeding, 31: 190-191. In Japanese. BABA, К. & HAMATANI, |., 1952, List of the species of Opisthobranchia from Kii, Middle-Japan. The Nanki-Seibutsu Supplement 1: 1-11. The Nanki Biological Society, Wakayama, Japan. BABA, К. & HAMATANI, |., 1971, Description of Pleurobranchus semperi (Vayssière, 1896) from Osaka Bay, Middle Japan. Veliger, 13: 325-329. BARASH, A. & DANIN, Z., 1971, Opisthobranchia (Mollusca) from the Mediterranean waters of Israel. /srael Journal of Zoology, 20: 151-200. BATHAM, E. J., 1969, Benthic ecology of Glory Cove, Stewart Island. Transactions of the Royal Society of New Zealand (Biological Sciences), 11: 73-81. BERGH, L. S. R., 1892, Opisthobranches provenant des campagnes du yacht l'Hirondelle. Résultats des Campagnes Scientifiques accomplies sur son yacht par Albert I, Prince Souverain de Monaco, 4: 1-35. BERGH, L. 5. R., 1897, Die Pleurobranchiden.—- Pleurobranchea. In C. G. Semper, Reisen im Archipel der Philippinen von Dr. C. Semper, Malacologische Untersuchungen, 7(1)[1-3]: 1- 158, 12 pl. Wiesbaden. BERGH, L. S. R., 1898, Die Pleurobranchiden.— Pleurobranchus. Sempers Reisen, 7(3): 117- 158. BERGH, L. S. R., 1900, Ergebnisse einer Reise nach dem Pacific (Schauinsland 1896-1897). Zoologischen Jahrbücher Abtheilung für Sys- tematik, Geographie und Biologie der Thiere, 13(3): 207-246. BERGH, L. S. R., 1902, Die Pleurobranchiden.— Acteon ... Oscaniella. Sempers Reisen, 7(7): 313-382. BERGH, L. S. R., 1905a, Die Pleurobranchiden.— Berthella... Chelyonotus. Sempers Reisen, 9(2): 57-119. BERGH, L. S. R., 1905b, Die Opisthobranchiata der Siboga-Expedition. Siboga Expedition Re- ports, 50: 1-248. BERTSCH, H., 1970, Opisthobranchs from Isla San Francisco, Gulf of California, with the description of a new species. Contributions in Science, Santa Barbara Museum of Natural History, 1(2): 1-16. BERTSCH, H., 1975, Distributional and anatomical observations of Berthella tupala (Opisthobran- chia: Notaspidea). Nautilus, 89(4): 124-126. BERTSCH, H., 1976, Intraspecific and ontogenetic radular variation in opisthobranch systematics. Systematic Zoology, 25: 117-122. BERTSCH, H., 1977, The Chromodoridinae nudi- branchs from the Pacific Coast of America—Part 1. Investigative methods and supraspecific tax- onomy. Veliger, 20: 107-118. BLAINVILLE, H. M. D. DE, 1825, Dictionnaire des Sciences Naturelles, 41. BURN, R., 1957, On some Opisthobranchia from Victoria. Journal of the Malacological Society of Australia, 1(1): 11-29. BURN, R., 1958, Further Victorian opisthobranchs. Journal of the Malacological Society of Aus- tralia, 1(2): 20-36. BURN, R., 1959, Comments on the Australian umbraculacean Mollusca. Journal of the Mala- cological Society of Australia, 1(3): 28-30. BURN, R., 1962, On the new pleurobranch sub- family Berthellinae (Mollusca: Gastropoda). A revision and new classification of the species of New South Wales and Victoria. Memoirs of the National Museum, Melbourne, 25: 129-148. BURN, R., 1966a, Some opisthobranchs from southern Queensland. Journal of the Malaco- logical Society of Australia, 1(9): 96-109. BURN, R., 1966b, Opisthobranchia. Port Philip Survey 1957-1963. Memoirs of the National Museum, Melbourne, 27: 265-364. BURN, R., 1969, A memorial report on the Tom Crawford collection of Victorian Opisthobran- chia. Journal of the Malacological Society of Australia, 1(12): 64-106. CHEESEMAN, T. F., 1878, Descriptions of three new species of opisthobranchiate Mollusca from New Zealand. Proceedings of the Zoological Society of London, 1878: 275-277, pl. 15. NEW ZEALAND PLEUROBRANCHIDAE CHEESEMAN, T. F., 1879, Descriptions of three new species of opisthobranchiate Mollusca from New Zealand. Transactions and Proceedings of the New Zealand Institute, 11: 378-380, pl. 16. (Article and figure reprinted from Proceedings of the Zoological Society of London.) DAKIN, W. J., 1952, Australian Seashores. Angus & Robertson, Sydney, 266 p. DAVIS, W. J. & MPITSOS, G. J., 1971, Behavioral choice and habitation in the marine mollusk Pleurobranchaea californica MacFarland (Gas- tropoda, Opisthobranchia). Zeitschrift für vergleichende Physiologie, 75: 207-232. DAVIS, W. J., MPITSOS, G. J., PINNEO, J. M. & RAM, J. L., 1977, Modification of the behavioral hierarchy of Pleurobranchaea. 1. Satiation and feeding motivation. Journal of Comparative Physiology, A, 117(1): 99-125. DELLE-CHIAJE, S., 1828, Memorie sulla storia e notomia degli animali senza vertebre del regno di Napoli, 3: 232 p. (Not seen.) EDMUNDS, M. & THOMPSON, T. E., 1972, Opisthobranchiate Mollusca from Tanzania. IV. Pleurobranchomorpha, Dendronotoidea and Arminoidea. Proceedings of the Malacological Society of London, 40: 219-234. р FRANC, A., 1968, Opisthobranchia. п GRASSE, P. P., Traité de Zoologie 5, Mollusques gastéro- podes et scaphopodes. Paris, 1083 p. GARDINER, A. P., 1936, Engel's paper on “The English species of the Pleurobranchidae.” Journal of Conchology, 20: 195-198. GHISELIN, M. T., 1965, Reproductive function and the phylogeny of opisthobranch gastropods. Malacologia, 3: 327-378. GOHAR, H. А. F. & ABUL-ELA, |. A., 1957, The development of Berthellina citrina (Mollusca, Opisthobranchiata). Publications of the Marine Biological Station, Al Ghardaga, Red Sea, 9: 69-84. GORDON, D. P. & BALLANTINE, W. J., 1977 [‘1976’], Cape Rodney to Okakari Point Marine Reserve. Review of knowledge and bibliography to December 1976. Tane, 22: 146 p. GRAHAM, D. H., 1941, Breeding habits of twenty- two species of marine Mollusca. Transactions and Proceedings of the Royal Society of New Zealand, 71: 152-159. GRAY, J. E., 1847, A list of the genera of Recent Mollusca, their synonyma and types. Proceed- ings of the Zoological Society of London, 15: 129-219. GUANG-YU, L. & SI, T., 1965, Etude sur les mol- lusques Pleurobranchidae de la côte de Chine. Oceanologia et Limnologia Sinica, 7: 265-277. GUIART J., 1901, Contribution à l'étude des gasteropodes opisthobranches et en particulier des céphalaspides. Theses Presentées a la Faculté des Sciences de l'Université de Paris. Le Bigot Frères, Lille, 219 p. HIRASE, S., 1937, On a new Berthella from Japan. Proceedings of the Malacological Society of London, 22: 177-181. HUTTON, Е. W., 1880, Manual of the New Zealand 265 Mollusca. A Systematic and Descriptive Cata- logue of the Marine and Land Shells, and of the Soft Mollusca and Polyzoa of New Zealand and the Adjacent Islands. Colonial Museum and Geological Survey Department, James Hughes, Wellington, 224 p. HYMAN, L. H., 1967, The Invertebrates, Volume 6, Mollusca 1. McGraw-Hill, 792 p. KAY, E. A., 1979, Hawaiian Marine Shells. Reef and Shore Fauna of Hawaii. Bernice P. Bishop Museum Special Publications, 64(4): 653 p. KELAART, E. F., 1883, New and little known spe- cies of Ceylon nudibranchiate molluscs, and zoophytes. Journal of the Royal Asiatic Society of Great Britain and Ireland, Ceylon Branch, Colombo, 3(9): 76-125. MACFARLAND, F., 1966, Studies of opisthobran- chiate mollusks of the Pacific Coast of North America. Memoirs of the California Academy of Science, 6: 546 p. MACNAE, W., 1962, Notaspidean opisthobran- chiate molluscs from Southern Africa. Annals of the Natal Museum, 15(15): 167-181. MARBACH, A. & TSURNAMAL, M., 1973, On the biology of Berthellina citrina (Gastropoda: Opisthobranchia) and its defensive acid secre- tion. Marine Biology, 21: 331-339. MARCUS, Ernst, 1957, On Opisthobranchia from Brazil (2). Journal of the Linnean Society of Lon- don, Zoology, 43: 390-486. MARCUS, Ernst, 1965, Some Opisthobranchia from Micronesia. Malacologia, 3: 263-268. MARCUS, Eveline, 1971, Key of genera and spe- cies of the Pleurobranchidae—from Odhner, 1926. Molluscan Digest, 1(2): A5-A6. (English translation of Odhner's key). MARCUS, Eveline, 1977, On a pleurobranchid from the Mediterranean. Veliger, 19: 418—421. MARCUS, Eveline & HUGHES, H. P. 1., 1974, Opisthobranch mollusks from Barbados. Bulletin of Marine Science of the Gulf and Caribbean, 24: 498-532. MARCUS, E. & MARCUS, E., 1957, Sea-hares and side-gilled slugs from Brazil. Boletim do Instituto Oceanografico, Universidade de Sao Paulo, 6(1-2): 3-33. MARCUS, E. & MARCUS, E., 1966, Opistho- branchs from Tropical West Africa. Studies in Tropical Oceanography, University of Miami Institute of Marine Sciences, 4(1): 152-208. MARCUS, E. & MARCUS, E., 1967a, American Opisthobranch mollusks. Studies in Tropical Oceanography, University of Miami Institute of Marine Sciences, 6: 256 p. MARCUS, E. & MARCUS, E., 1967b, Some opis- thobranchs from Sapelo Island, Georgia, U.S.A. Malacologia, 6(1-2): 199-222. MARCUS, E. & MARCUS, E., 1970, Some gastro- pods from Madagascar and West Mexico. Mala- cologia 10(1): 181-223. MECKEL, J. F., 1913, /n LEUE, Pleurobranchaea novo molluscorum genre. Dissrn. Inaug. Acad. Halle: 11. MESTAYER, M. K., 1920, A note on Pleurobran- 266 chaea novae-zelandiae Cheeseman. New Zea- land Journal of Science and Technology, 3: 170-171. MILLER, M. C. & BATT, G. W., 1973, Reef and Beach Life in New Zealand. Collins, 141 p. MILLIGAN, E. N., 1953, Marine molluscs of Little Barrier, and marine molluscs of Hen Island. Tane, 6: 134-139. MINICHEV, Y. S., 1970, On the origin and system of nudibranchiate molluscs (Gastropoda Opisthobranchia). Monitore Zoologica Italiano, N.S., 4: 169-182. MINICHEV, У. 5. & STAROBOGATOV, Y. I., 1978, On the systematic arrangement of euthyneuran snails. Malacological Review, 11: 67-68. MONTAGU, G., 1803, Testacea Britannica, or Natural History of British Shells, Marine, Land, and Freshwater. White, London, 291 p. MORTON, J. E., 1958, Molluscs. Ed. 1. Hutchin- son, 244 p. MORTON, J. E., 1972. The form and functioning of the pallial organs in Akera bullata with a discus- sion on the nature of the gill in Notaspidea and other tectibranchs. Veliger, 14: 337-349. MORTON, J. E. & MILLER, M. C., 1968, The New Zealand Sea Shore. Collins, 638 p. NARAYANAN, K. E., 1969, On the opisthobranchi- ate fauna of the Gulf of Kutch. Proceedings of the Symposium on Mollusca, India, 1: 188-213. NARAYANAN, K. E., 1970, On a species of the genus Berthellina (Opisthobranchia: Notaspi- dea) of the Gulf of Kutch. Journal of the Marine Biological Association of India, 12(1-2): 210- 213. NORDSIECK, F., 1972, Die Europäischen Meeresschnecken Opisthobranchia mit Pyra- midellidae, Rissoacea. Fischer, Stuttgart, 327 p. ODHNER, N. H., 1924, Papers from Dr. Th. Mor- tensen's Pacific Expedition 1914-1916. XIX. New Zealand Mollusca. Videnskabelige dansk naturhistorisk forening i Kjpbenhavn, 77: 90 р. ODHNER, N. H., 1926, Die Opisthobranchien. Further Zoological Results of the Swedish Ant- arctic Expedition 1901-1903, 2(1): 100 p. ODHNER, N. H., 1939, Opisthobranchiate Mol- lusca from the western and northern coasts of Norway. Kongelige Norske videnskabernes Selskabs Skrifter, 1: 93 p. O'DONOGHUE, C. H., 1924, Report on Opistho- branchia from the Abrolhos Islands, Western Australia, with description of a new parasitic copepod. Journal of the Linnean Society (Zool- ogy), 35: 521-579. OTTAWAY, J. R., 1977a, Predators of sea anemones. Tuatara, 22: 213-221. OTTAWAY, J. R., 1977b, Pleurobranchaea novae- zelandiae preying on Actinia tenebrosa. New Zealand Journal of Marine and Freshwater Re- search, 11: 125-130. PELSENEER, P., 1906, In LANCASTER, R., ed. Treatise on Zoology, Part 5, Mollusca. Adams & Charles Black, London, 355 p. WILLAN PILSBRY, H. A., 1896, Manual of Conchology, Philadelphia, 16: 175-226. POWELL, A. W. B., 1937, The Shellfish of New Zealand. Unity Press, 100 p. POWELL, A. W. B., 1939a, The Mollusca of Stewart Island. Records of the Auckland Insti- tute and Museum, 2: 211-238. POWELL, A. W. B., 1939b, Biological primary types in the Auckland Museum. Records of the Auck- land Institute and Museum, 2: 239-259. POWELL, A. W. B., 1946, The Shellfish of New Zealand. Ed. 2, Whitcombe & Tombs, 106 p. POWELL, A. W. B., 1955, Mollusca of the Southern Islands of New Zealand. Cape Expedition Series Bulletin, 15: 151 p. POWELL, A. W. B., 1957, Shells of New Zealand. Ed. 3, Whitcombe & Tombs, 202 p. POWELL, A. W. B., 1961, Shells of New Zealand. Ed. 4, Whitcombe & Tombs, 203 p. POWELL, A. W. B., 1976, Shells of New Zealand. Ed. 5 revised, Whitcoulls, 154 p. POWELL, A. W. В., 1979, New Zealand Mollusca: Marine, Land and Freshwater Shells. Collins, 500 p. PRUVOT-FOL, A., 1953, Etude de quelques opisthobranches de la Cote Atlantique du Maroc et du Senegal. Travaux de l'Institut Scientifique Cherifien, Zoologie, 5: 1-105. PRUVOT-FOL, A., 1954, Mollusques Opisthobran- ches. Faune de France, 58. Lechavalier, Paris, 460 p. QUOY, J. R. C. & GAIMARD, J. P., 1832-1835, Zoologie: Mollusca. Voyages de decouvertes de l'Astrolabe, sous le commandement de M. J. Dumont d’Urville, Vol. 2 and Atlas. Paris. RISBEC, J., 1928, Contribution a l'étude des nudi- branches Néo-Calédoniens. Faune des Colonies Francaises, 2: 328 p. SCHMEKEL, L.-R., 1968, Ascoglossa, Notaspidea und Nudibranchia im litoral des Golfes von Neapel. Revue Suisse de Zoologie, 75: 103- 155: SPHON, G. G., 1972, Berthella kaniae, a new opisthobranch from the Eastern Pacific. Nautilus, 86: 53-55. SUTER, H., 1913, Manual of the New Zealand Mol- lusca: with an Atlas of Quarto Plates (1915). Government Printer, Wellington, 1120 p. TAYLOR, D. W. & SOHL, N. F., 1962, An outline of gastropod classification. Malacologia, 1: 7-32. TCHANG-SI, 1934, Contribution a l'étude des opisthobranches de la côte de Tsingtao. Contri- butions from the Institute of Zoology, National Academy of Peiping, 2: 149-165. THIELE, J., 1931, Handbuch der systematischen Weichtierkunde, 1: 778 p. Jena. THOMPSON, T. E., 1970, Eastern Australian Pleurobranchomorpha (Gastropoda, Opistho- branchia). Journal of Zoology, 160: 173-198. THOMPSON, T. E., 1976, Biology of Opistho- branch Mollusca, Vol. 1. Ray Society, 151: 207 p. NEW ZEALAND PLEUROBRANCHIDAE THOMPSON, T. E., 1977, Jamaican opisthobranch molluscs 1. Journal of Molluscan Studies, 43: 93-140. THOMPSON, T. E. & SLINN, D. J., 1959, On the biology of the opisthobranch Pleurobranchus membranaceus. Journal of the Marine Biologi- cal Association of the United Kingdom, 38: 307- 324. USUKI, 1., 1969, The reproduction, development and life history of Berthellina citrina (Rúppell 8 Leuckart), (Gastropoda, Opisthobranchia). Sci- ence Reports of Niigata University, Ser. D (Bi- ology) 6: 107-127. VAYSSIERE, A., 1885, Recherches zoologiques et anatomiques sur les opisthobranches du Golfe de Marseille. Premiere Partie, Tectibranches. Annales du Musée d'Histoire Naturelle de Marseille, Zoologie, 2(3): 181 p. VAYSSIERE, A., 1896, Description des coquilles de quelques espèces nouvelles ou peu connues de Pleurobranchidés. Journal de Conchyliol- ogie, 44: 113-137. VAYSSIERE, A., 1898, Monographie de la famille des Pleurobranchidés |. Annales des Sciences Naturelles: Zoologie, 8: 209-402. VAYSSIERE, A., 1901, Monographie de la famille des Pleurobranchidés (Deuxième et dernière partie). Annales des Sciences Naturelles, Zool- ogie, 8(12): 1-85. VERRILL, A. A., 1882, Catalogue of marine Mol- lusca added to the fauna of the New England Region, during the past ten years. Transactions of the Connecticut Academy of Arts and Sci- ences, 5: 447-587. WHITE, K., 1948, On a collection of marine mol- luscs from Ceylon. Proceedings of the Malaco- logical Society of London, 27: 199-204. WILLAN, R. C., 1975, /dentity and Feeding of New Zealand Notaspidean Opisthobranchs (Mol- lusca: Gastropoda). Unpublished M.Sc. thesis, University of Auckland, 164 p. WILLAN, R. C., 1977, A review of Pleurobranchella Thiele, 1925 (Opisthobranchia: Pleurobran- chaeinae). Journal of Conchology, 29: 151-155. WILLAN, R. C., 1978, An evaluation of the notaspidean genera Pleurobranchopsis Verrill and Gymnotoplax Pilsbry (Opisthobranchia: Pleurobranchinae). Journal of Conchology, 29: 337-344. WILLIAMS, E. G., 1964, Molluscs from the Bay of Plenty. Whitcombe & Tombs, 41 p. WINCKWORTH, R., 1946, On Bergh's Malacolo- gische Untersuchungen. Proceedings of the Malacological Society of London, 28: 20-22. APPENDIX This Appendix lists the New Zealand pleuro- branch material that | have examined during this study. Habitat data are included where possible. Molluscan collections housed in the National Museum of New Zealand, Wellington 267 (NM), New Zealand Oceanographic Institute, Wellington (NZOI), Auckland Institute and Museum (AIM), and Geology Department, University of Auckland (AUG), have been in- spected. Dr. M. C. Miller of the Zoology De- partment, University of Auckland, has gener- ously given me access to his opisthobranch collections. Comparative Australian material was kindly provided by Mr. R. Burn of Victoria. | collected most of the live specimens on which the morphological descriptions here are based, both intertidally, and subtidally by scuba diving, and these are signified by the letters “R.C.W.” in this Appendix. Berthellina citrina NORTH ISLAND: Spirits Bay, Northland—1 washed live on to beach near Pananehe Island, R.C.W., 14 Jan. 1972; 18m, Black Point, N.E. side of Karikari Pen.—2 copulat- ing, and a freshly-laid egg coil, R.C.W., 11 Feb. 1978; 10 m, off N. side of Jolliffe Point, Matai Bay, N. end of Doubtless Bay—1 be- neath a stone, sand substrate, R.C.W., 30 March 1978; 11 m, off S.W. side of Motuta- pere Is., Cavalli Is—1 beneath a stone, R.C.W., 29 Dec. 1978; 8.5 m, large bay at N. side of Hamaruru Is., Cavalli Is.—1 beneath a stone, R.C.W., 31 Dec. 1978; 15 m, Nursery Cove, Aorangi Is., Poor Knights Is. —abun- dant on undersides of stones and also several copulating pairs, beneath kelp forest; 20 m, entrance to Riko Riko Cave, Aorangi Is., Poor Knights Is.—4 on undersides of stones, fine bryozoan substrate, R.C.W., 10 Dec. 1973; 15 m rock slope to left of Riko Riko Cave, Aorangi Is., Poor Knights Is.—abundant on undersides of stones, R.C.W., 31 Dec. 1978; 60 m, off Deep Water Cove, Bay of Islands— 1 shell, dredged in grey, muddy ooze, R.C.W., 7 April 1973; 6-8 m, off Castle Rock, Tauri- kura, Whangarei Hbr.—1 dredged amongst deposit of Glycymeris, Pecten, Oxyperas, Tawera, some fine broken shell and small stones, also ascidians, bryozoans and hy- droids, M. С. Miller, 18 Мау 1961; McGregor's Bay, Whangarei Heads—2 on 14 May 1961 and 3 on 5 May 1961, both M. C. Miller; High Is., Whangarei Hbr.—on undersides of rocks at low water level, 2 on 14 Jan. 1959, 3 on 15 May 1961 and 3 on 15 Dec. 1961; Ocean Beach, Whangarei Heads—4 (including one copulating pair), beneath intertidal stones, M. C. Miller, 14 Dec. 1961; 15-16 m, Northern Bay, Little Barrier Is—2 on undersides of stones, R.C.W., 25 Jan. 1977; 15-18m 268 “Sponge Garden” and “North Reef,” off N.W. tip of Goat Is., Leigh—2 on 7 Aug. 1973, 22 on 11 May 1974, 5 on 14 May 1974, 6 on 21 Sept. 1975, 1 on 27 May 1976, several on 28 Sept. 1976 and several on 29 Sept. 1976, all R.C.W., 20 m, ‘The Canyon,” at S.E. corner of Goat Is., Leigh—2 and several spawn coils, R.C.W., 24 Dec. 1975; “Echinoderm Reef Flat,” Goat Island Bay, Leigh—5 on 19 Dec. 1960, 3 on 28 Aug. 1961, 2 on 26 Oct. 1961, and several on 21 May 1967, M.C. Miller; “Echinoderm Reef Flat,” Goat Island Bay, Leigh—1 on 14 May 1974, 9 on 24 May 1974, all R.C.W., all specimens found on under- sides of stones and rocks in shallow pools and channels near low water mark; 12- 15 m, near “Knot Rock,” c. 100 m offshore from W. end of Goat Island Bay, Leigh—2 on 25 June 1977, 1 on 5 July 1977, 3 on 29 Aug. 1977, and 1 on 20 Sept. 1977, all R.C.W.; 3-4 m, Leigh Hbr.—9 on 22 Nov. 1973 and 4 on 14 March 1974, all found on the under- faces of stones on substrate of gravel and coarse sand, R.C.W.; 4.5 m, Matheson Bay, Leigh—1 and spawn coil, R.C.W., 19 Jan. 1976; Ti Point Channel, Leigh—4 on under- sides of stones, on coarse sand-gravel sub- strate, R.C.W., 16 May 1974; 43 m, Aldermen Is.—1, 21 Nov. 1971 (NM); 59-74 m, off E. side of Mayor Is.—1 shell, amongst pebbles, shell grit and algae, 22 Jan. 1979 (NM); Omaio Bay, E. Bay of Plenty—1, R. K. Dell, 12 March 1962 (NM); little bay before Cape Runaway, East Cape—1, on underside of an intertidal stone, K. R. Grange, 29 March 1975. Berthella ornata NORTH ISLAND: Henderson Pt., S. end of Rarawa Beach, Northland—1 on underside of a low-tidal, weed-covered rock, R.C.W., 18 Jan. 1970; Mahinepua Peninsula, near Whangaroa—1, R.C.W., 7 Jan. 1966; “The Gap,” Mahinepua Peninsula, near Whang- aroa—2 on undersides of stones in the sub- littoral fringe, R.C.W., 7 Jan. 1970; Church Bay, Tutukaka—4 on undersides of low-tidal stones, J. D. Willan, 17 Oct. 1970; Pacific Bay, Tutukaka—1 on underside of a stone, E.L.W.S. level, C. Grange, 8 March 1974; High Is., Taurikura, Whangarei Hbr.—3 on undersurfaces of rocks at low-water mark, 7 Jan. 1958, 1 on 14 Jan. 1959, 2 on 16 Jan. 1959, several pairs copulating on 13 Dec. 1961 and 2 on 15 May 1961, all M. C. Miller; 6 m, E. side of High Is., Taurikura, Whangarei Hbr.—1 on undersurface of a stone, R.C.W., WILLAN 21 June 1975; McGregor's Bay, Whangarei Heads—1 beneath an intertidal stone, M. C. Miller, 14 May 1961; “Echinoderm Reef Flat” Goat Island Bay, Leigh—3 on 28 Aug. 1961 and 1 on 26 Oct. 1961, both M. C. Miller, and 1 on underside of a stone on substrate of clean, coarse sand, middle of low-tidal plat- form in a rock pool, 22 May 1974 and 9 on 24 May 1974, both R.C.W.; 6m off Goat Is. Leigh—1 on underside of a stone on sub- strate of coarse gravel and sand, D. Rowe, 30 June 1974; 12-15m near “Knot Rock,” c. 100 m offshore from W. end of Goat Island Bay, Leigh—1 on underside of a stone, R.C.W., 29 Aug. 1977; Omaha, Leigh—1 on underside of a low-tidal stone, E. N. Gardner, 16 Sept. 1974; Kawau Is., Hauraki Gulf—1, C. Wormald, 13 Aug. 1967; Beehive ls., off Kawau Is., Hauraki Gulf—1, R.C.W., 26 Sept. 1965; Mahurangi Is., off Waiwera Beach—1 on 30 Aug. 1969, 8 on 12 Jan. 1974, 1 on 22 July 1974 and 1 on 15 Sept. 1974, all on undersides of stones in pools, at low-water level, R.C.W.; 11m, eastern side of Tiritiri Matangi Is., Hauraki Gulf, 1 on underside of a stone, R.C.W., 29 Feb. 1976; Army Bay, Whangaparoa Pen.—2 under a rock ledge. E.L.W.S. level, in a rock pool, under Cysto- phora retroflexa, R.C.W., 10 Jan. 1974; Takapuna, Auckland—1, H. Suter (Suter Colln., NM M17843); Beacon Rocks, Mount Maunganui, Bay of Plenty—2 shells, E. N. Milligan (AGU): Breaker Bay, Wellington—1 under a stone in brown algal association, W. R. B. Oliver, 15 Sept. 1923 (W. R. B. Oliver Colln., NM). SOUTH ISLAND: Headland Pt. Portobello Pen.—1 on 21 Aug. 1962, crawling across muddy floor of a cave, 0.3-0.5 m below low- water level, and 1 on 22 Aug. 1962, on under- surface of a rock at low-water level, both M. C. Miller. Berthella mediatas NORTH ISLAND: Ocean Beach, Whang- arei Heads—1, M. C. Miller, 14 Dec. 1961; High Is., Taurikura, Whangarei Hbr.—2 seen but only 1 collected, on undersurface of a rock, low tide level, M. C. Miller, 16 Jan. 1969; “Echinoderm Reef Flat,” Goat Island Bay, Leigh—1 on the underside of a stone, E.L.W.S. level, R.C.W., 1 June 1977, and 1 in holdfast of Laurencia sp. in a rock pool on inner part of reef, R.C.W., 23 June 1979; Army Bay, N. side of Whangaparoa Pen.—4 on undersides of low tidal stones in rock NEW ZEALAND PLEUROBRANCHIDAE 269 pools, R.C.W. and J. D. Willan, 10 Jan. 1974; Takapuna, Auckland—1, H. Suter (Suter Colln., NM M17843; specimen in lot with 2 Berthella ornata); Narrow Neck Reef, Auck- land—1 spawning, H. Suter, 24 July 1906 (NM); Cape Egmont, Taranaki—1 on under- side of low-tidal stone, K. R. Grange, 16 July 1974; Pencarrow Head, Wellington—2, W. F. Ponder, 20 Oct. 1956 (NM); Point Howard, Wellington—1, W. F. Ponder, 11 Nov. 1958 (NM); Lyall Bay, Wellington—4, R. K. Dell, 8 Jan. 1950 (NM). SOUTH ISLAND: Cape Three Points, Akaroa—4 on 17 May 1962, 1 on 18 May 1962 and 1 on 22 May 1962, all on undersur- faces of rocks at low tide level, M. C. Miller; Aquarium Pt., Portobello, Otago Hrb.—4 on 17 Jan. 1961, 1 on 18 Jan. 1961 and 1 on 17 Aug. 1962, all on undersurfaces of rock at low water level, М. С. Miller; Quarantine Is., Portobello, Otago Hbr.—2 amongst moveable stones in a sheltered inlet, M. C. Miller, 20 Aug. 1962. CHATHAM IS: East Bay, South East (Rangatira) 1$.—1 under a stone, in low tidal rock pool, E. C. Young, 2 Jan. 1975. Pleurobranchaea maculata NORTH ISLAND: Russell, Bay of Islands— 1, L. J. Mather, May 1965 (NM); 6-8 m, S. of Castle Rock, Tutukaka—1 on substratum of coarse broken shell and sand, M. C. Miller,19 Jan. 1959; N. end of Oakura Beach—1 on underside of a stone at low tide, M. C. Miller, 17 Jan. 1960; beach at end of reclamation, near channel, N. side of Tutukaka Hbr.—1 crawling on mud amongst rocks; low water level, K. R. Grange, 14 May 1972; 8-10 m, “Waterfall Reef,” coast near Goat Is., Leigh— several small juveniles on undersides of stones on 26 March 1977 and many on 22 Dec. 1978, both R.C.W.; 18-20m, ‘Deep Point” and “The Canyon” at southeastern tip of Goat Is., Leigh—2 and several egg masses on 21 Sept. 1975, 1 juv. on 24 Dec. 1975, many and fresh spawn on 4 June 1976 and many on 12 Dec. 1978, all R.C.W.; 9-17 m, “North Reef,” off N.W. tip of Goat Is., Leigh— 1 on underside of a stone on 11 May 1974 and 1 juv. on underside of a stone on 7 May 1977, both R.C.W.; 18 m, “Sponge Garden,” off N.W. tip of Goat Is., Leigh—2, R.C.W., 28 Sept. 1976; 6-7 m, off S.W. tip of Goat Is., Leigh—1, D. K. Rowe, 30 June 1974; “Echinoderm Reef Flat,” Goat Island Bay, Leigh—2 on undersurfaces of stones at low- tide level, 28 Aug. 1961 and several seen on undersurfaces of rocks, in pools just above low-tide level, 26 Oct. 1961, both M. C. Miller; 3m, Leigh Hbr.—5, R.C.W., 22 Nov. 1973; 9m, Matheson Bay, Leigh—many, R.C.W. and M.S. Leighton, 21 Dec. 1978; 2-3 m, Ti Point Channel, entrance to Whangateau Hbr., Leigh—1 on 14 May 1974 and 1 on 22 May 1978, both R.C.W.; Mahurangi Is., Waiwera— many on undersurfaces of low-tidal stones in pools, 30 Nov. 1973 and many in similar habi- tats on 12 Jan. 1974, both R.C.W.; Army Bay, N. side of Whangaparoa Pen.—2 on under- surfaces of stones in low-tidal pools on 27 Dec. 1973 and many in similar habitats on 10 Jan. 1974, both R.C.W.; Matakatia Bay, S. side of Whangaparoa Pen.—1 on undersur- face of a stone in a low-tidal pool, R.C.W., 30 Nov. 1973; Surfdale, Waiheke Is.—8, R.C.W., 25 Nov. 1973; 5-6 m, between Waiheke Is. and Pakatoa Is—1, P. R. Bergquist, 14 March 1974; 8-9m, Rakino Channel—1 on substratum of mixed shell and sand, M. C. Miller, 16 Nov. 1960; Rangitoto Channel, off North Head—1 and several spawn masses, M. C. Miller, 16 Sept. 1967; Takapuna Beach—1, H. Suter (Suter Colln., NM); West Tamaki Reef, Waitemata Hbr.—1 on under- side of a stone at low-tide level, M. C. Miller, 8 Dec. 1962; Tamaki Str.—1 in bottom fish trawl, W. Tong, Sept. 1974; bay E. of Little Blowhole, South Head, Manukau Hbr.—1 found at E.L.W.S. level, D. Brambley, 2 Sept. 1966; 4.5m, Bowentown entrance, Waihi— W. Salmons, Oct. 1977; 14m, approx. 1.2km offshore from Tataraimaka Historic Site Headland, S. of New Plymouth, Taran- aki—1 juv. on undersurface of a stone, J. Nicholson, 15 Feb. 1978 and 1 juv., amongst fouling hydroids and algae on buoy rope, close to surface, R.C.W., 5 May 1978 and 6 juvs., 10.5-12 m on undersurfaces of stones, R.C.W., 4 July 1979; Island Bay, Wellington— 1, N. M. Adams, 5 Oct. 1970 (NM); Lyall Bay, Wellingion—1, C. Hale, Nov. 1952 (NM); 120 m, Cook Str. trawling grounds, 1 taken by M. V. “Thomas Currell,” 27 Feb. 1964 (NM). SOUTH ISLAND: Tasman Bay, Nelson—1, M. Young, Nov. 1934 (NM); 12m, Golden Bay, Nelson, N. Z. Marine Dept. (NM); Port Underwood, Marlborough Sounds—3, R. Ponder, May 1962 (NM); Port Levy, Banks Pen.—1, W. R. B. Oliver, 1 April 1907 (Oliver Colln., NM); Te Onepoto, Lyttelton—1 (Suter Colln., NM); 8-12 m, off Middle Headland of German Bay Hill, Akaroa Hbr.—2 on sub- 270 WILLAN stratum of mud and coarse, broken shell, M. C. Miller, 11 May 1962; Cape Three Points, Akaroa Hbr.—several under stones at low water, and crawling amongst algae below low water, M. C. Miller, 22 May 1962; Oputereinga Pt., Akaroa Hbr.—1 in a large pool amongst mussels, M. C. Miller, 20 May 1962; Yacht Club shore, Akaroa Hbr.—sever- al on undersides of stones, M. C. Miller, 21 May 1962; 280 m off Cape Campbell—1, F. Abernethy, 14 Nov. 1962 (NM). STEWART ISLAND: Golden Bay, Paterson Inlet—1, R. K. Dell, 29 Oct. 1948 (NM); 7 m, Big Glory Cove, Paterson Inlet, M. V. “Munida” Stn. 67-37. E. J. Batham, 15 Feb. 1967 (NM); 20 m off southern corner of Native Is., Paterson Inlet—1 and several spawn masses, R.C.W., 8 Feb. 1977. MALACOLOGIA, 1983, 23(2): 271-279 SEXUALITY WITH RESPECT TO SHELL LENGTH AND GROUP SIZE IN THE JAPANESE OYSTER CRASSOSTREA GIGAS Norman E. Buroker Bureau of Biological Research, Rutgers, The State University, Piscataway, New Jersey 08854, U.S.A. ABSTRACT The sex ratios for two populations (Big Beef Creek and Dabob Bay, Hood Canal, Washington) of the Japanese oyster, Crassostrea gigas, were determined with regard to shell length and group size (i.e., the number of oysters attached together as a group). In this study the proportion of males was found to decrease with increasing shell length as is the case for protandrous hermaphrodites and was also found to increase with increasing group size. Also, an investigation was made of individual sexuality with respect to group size. In addition to singleton oysters (i.e., one oyster with no shell attachment to conspecifics), a wide variety of group sizes was found in both populations ranging from two to sixteen oysters per cluster. This study revealed that for doubleton oysters (i.e., two oysters per group) there was a highly signifi- cant incidence of male-female individuals as opposed to male-male and/or female-female indi- viduals. This social relationship found among doubleton oysters is not apparent for oysters contained in group sizes greater than two individuals per cluster. The male-female relationship in doubletons is discussed with respect to its evolutionary significance. Key words: female-male doubleton; group size; oyster; sex ratio; social behavior. INTRODUCTION The determination of sex in oysters has been and still is a controversial area of re- search (Coe, 1932a, 1932b, 1934, 1940; Haley, 1977). However, recent advances have been made for oysters (Haley, 1977) as well as for other marine invertebrates (Hoag- land, 1978). In sex-labile marine invertebrates such as the oyster, it is thought that the dura- tion of a sex phase (male or female) results from simple Mendelian segregation of multi- ple sex genes, whose action is additive (Bacci, 1965). Haley (1977) has provided evi- dence that the American oyster, Crassostrea virginica (Gmelin, 1791), displays at least three polymorphic sex loci. Each locus segre- gates two (male and female) alleles which have an additive effect in the individual. This type of model predicts that when male and female alleles of an individual are in equal number, the individual has the ability of being either sex and perhaps readily prone to sex reversal. It follows that when there are inci- dences of an unbalanced number of male or female alleles, the sexual phase of an individ- ual would be dictated by the sex allele in greatest number (Bacci, 1965). However, there exists a large volume of literature which indicates that sex determination in sex-labile marine invertebrates is also influenced by the environment (Bacci, 1965; Hoagland, 1978; Reinboth, 1975) and the social behavior of individuals (Hoagland, 1978). Sex determination in Crassostrea species has been influenced by such environmental conditions as the amputation of gill lamellae (Amemiya, 1936), crowding (Burkenroad, 1931, 1937); Coe, 1932a; Needler, 1932), nutrition (Coe, 1932a, 1934; Amemiya, 1925), geographical location (Katkansky & Sparks, 1966), and steroids (Mori et al., 1969). Indica- tions of social interactions in sex determina- tion of oysters can only be inferred from the available data (Burkenroad, 1931); however, studies involving social behavior between in- dividuals of oyster populations have not been conducted. | have in this paper presented evi- dence of social influence on the sex determi- nation of the Japanese oyster, Crassostrea gigas (Thunberg, 1795), as well as examined individual sexuality with respect to shell length and location to conspecifics. In the Pacific Northwest of the United States of America, C. gigas spends the non- planktonic period of its life as a sessile inter- tidal adult. Sexual reproduction occurs when gametes of the two sexes are broadcast into the surrounding water where fertilization oc- curs upon contact. Consequently, for suc- (271) 272 cessful fertilization to occur it is necessary for these sedentary adults to be in close proxim- ity as well as exhibit sexual heterogeneity be- tween contiguous individuals. Planktonic larval development lasts from two to three weeks which provides ample opportunity for dispersion (Quayle, 1969). When the time of settlement approaches, the gregarious (Cole & Knight-Jones, 1949; Knight-Jones, 1952; Hidu, 1969) oyster larvae select suitable at- tachment sites (usually near conspecifics) where they will settle, metamorphose, and grow into adults. Due to the gregarious nature of the plank- tonic larvae during settlement, this species consists of dense populations. The physical structure of these populations can be parti- tioned into different group sizes. The group size will be defined as the number of oysters which are attached together by shell fusion and/or attached to the same substratum. A particular group or cluster of oysters may de- velop by: (1) new recruits settling on an adult; and/or, (2) two or more recruits settling in juxtaposition and fusing their shells together as they grow older. An oyster that is not physically attached to another individual can be denoted a singleton. Two oysters whose shells are fused together can be considered a doubleton (or given a group size of two); three oysters can be considered a tripleton, and so on. MATERIALS AND METHODS Oyster samples were collected during the reproductive season (July through August) of 1975 and 1976 from Big Beef Creek and Dabob Bay, Hood Canal, Washington. The BUROKER sex and length of each animal within the vari- ous groups were determined. The shell length of an individual was measured from the umbo to its ventral valve edge. The sex of an indi- vidual was determined by microscopic exami- nation to be either male (m), female (f), or undetermined (ud), depending on whether sperm, eggs or no gametes, respectively, were found in the gonad. The population sex ratio (i.e., males divided by the total number of males and females) has been recorded for various shell lengths. Also, the observed and expected number of groups with particular sex ratios were determined. The expected number of groups was computed from the ex- pected group frequency, which is obtained from the expression [р + (I — p)]", where pis the frequency of males in the sample, (| — р) is the frequency of females, and п is the group size. RESULTS Sex Ratio, Shell Length and Group Size The data from this study have been tabu- lated with regard to group size and the cor- responding shell length of the oysters within the populations (Tables 1 and 2). The propor- tion of males in the two populations (for con- secutive years) declined with increasing shell length class. The data support the known pat- tern of sexuality in Crassostrea species: pro- tandry (Amemiya, 1925, 1929; Burkenroad, 1931; Coe, 1934; Needler, 1932). In addition to providing support for pro- tandry in Crassostrea gigas, the data in Tables 1 and 2 record the proportion of males per group size. As an example, for group TABLE 1. The relationship between oyster sexuality, shell length, and group size for Crassostrea gigas from the Big Beef Creek population, Hood Canal, Washington. The sex ratio (s/r) is recorded with respect to shell length and group size. The number of males (m), females (f), and undetermined sex types (ud) are recorded for each shell length and group size. n is the total number of individuals. 1975 Group size 1976 Group size 1.60-3.99 m Shell length (cm) Sex 1 2 3 4 59 0 0 0 1 2 f 0 оо 0 0 ud 0 0 1 0 9 4.00-4.99 m 0 0 0 0 1 f 0 оо о 1 ud 0 0° 200 4 2 5.00-5.99 m 1 03823. 5) 3 f 0 1 1 0 0 ud 0 оо о 0 s/r 1 2 3 4 516 п s/r 1.000 3 6 1 6 20 36 1.000 0 0 0 0 0 0 12 7 14022460 0.500 10 4 3 8 7% 32 0.970 0 0 1 0 0 1 0 2 1 2 if 12 0.800 4 7 4 4 4 23 0.920 0 1 0 0 1 2 0 0 1 1 0 2 273 CRASSOSTREA GIGAS SEXUALITY TABLE 1 (Continued). 1976 Group size 1975 Group size s/r 5-9 Sex Shell length (cm) 34 0.850 6.00-6.99 33 0.750 0 0 m 7007-99 45 0.703 m 8.00-8.99 44 0.657 1 m 9.00-9.99 34 0.420 2 9 0.375 2 m 10.00-10.99 23 0.411 1 7 0.304 1 m 11:00—11.99 19 0.388 2 9 4 0.308 т 12.00-12.99 оо oo 19 0.475 2 m 13.00-13.99 5 0.185 0.200 m 14.00-14.99 10 0.400 1 3 0 0.000 m 15.00-15.99 4 0.308 0 4 0 0.000 m 16.00-16.99 oo oo 3 0.300 1 1 0 0.000 m 17.00-17.99 oo oo 0.200 0 0 0.000 m 18.00-18.99 oo oo 0.500 1 0 0.000 m 19.00-19.99 Oo 0.071 4 0 0.000 20:00---- TOTALS: 53 367 30 264 136 136 767 49 19 aie se 19 18 75 .67 125 102 38 69 13 127 24 63 82 13 158 43 16 11 14 а 13 14 _0 28 43 El un Ol Males Females Undet. Total 96 57/2 276 184 38 259 12 WE Cf 9 .22 39 .48 .36 .58 .64 .59 .49 [m/(m+f)] 274 BUROKER TABLE 2. The relationship between oyster sexuality, shell length, and group size for Crassostrea gigas from the Dabob Bay population, Hood Canal, Washington. The sex ratio (s/r) is recorded with respect to shell length and group size. The number of males (m), females (f), and undetermined sex types (ud) are recorded for each shell length and group size. ee ee KE 1975 Group size 1976 Group size Shell length — — (cm) Sex 1 2 3 4 510 п СУД 1 2 3 4 516 on s/r 1.40-3.99 m 0 0 0 2 0 2 1.000 0 0 1 0 0 1 1.000 f 0 0 0 0 0 0 0 0 0 0 0 0 ud 2 1 4 2 8 1% 1 6 6 5 44 62 4.00-4.99 m 2 6 1 0 1 10 0.526 1 2 1 1 1 6 0.857 f 6 2 0 0 1 9 0 0 0 1 0 1 ud 3 1 3 2 4 13 0 1 0 1 12 14 5.00-5.99 m 3 4 1 1 7 16 0.667 1 1 1 1 5 9 0.643 f 4 2 1 0 1 8 1 0 2 0 2 5 ud 6 2 3 3 61720 0 0 1 1 4 6 6.00-6.99 m 2 7 if 0 2 18 0.563 5 1 3 3 6 18 0557 f 7 2 3 0 PLU 4 3 0 0 2 9 ud 1 0 S 0 2 6 0 1 2 0 5 8 7.00-7.99 m 1 6 4 4 0 15 0.484 3 16 9 1 i 36. 0/57 f 6 76 2 0 1 16 1e) 7 2 0 5 27 ud 1 1 0 0 0 2 0 2 0 0 4 6 8.00-8.99 m 1 7 3 2 О’ 13 0520 10 23 20 1 17 71 0.664 f 7 3 2 0 (A 10 AS 5 0 6 36 ud 0 2 1 0 0 3 1 4 1 0 1 7 9.00-9.99 m 3 3 9 2 1 14 0:400 10’ 28 10 5 15 58 0.576 f 10 4 <) 2 2-21 14 13 1 5 50 ud 0 1 0 0 1 2 0 2 4 0 0 6 10.00-10.99 m 1 5 if 0 1 14 0.438 9 19 7 3 10 48 0.527 f 8 6 3 1 0 18 8 18 8 4 54s ud 1 0 0 0 0 1 1 0 0 ie 2 4 11.00-11.99 m 9 4 2 2 0 17 0.386 22 8 4 12 38 0.447 f 15 6 4 1 1 27 6 21 7 3 10 47 ud 2 1 0 0 0 3 1 1 1 1 2 6 12.00-12.99 m 3 3 6 0 072112720:375 0 if 4 4 7 2210647 f 9 5 3 3 0) 20 0 8 0 0 AZ ud 3 0 1 0 0 4 1 0 1 1 0 3 13.00-13.99 т 4 2 0 0 2 8 0.276 3 1 4 0 8 16 0.500 f 3 8 if 0 37221 0 7 2 2 5 a6 ud 1 0 0 0 0 1 2 1 0 1 2 4 14.00-14.99 m 1 2 2 0 0 5 0.417 1 2 2 0 6: = 11 101579 f 1 6 0 0 0 té 2 2 2 1 1 8 ud 1 0 0 0 0 1 0 0 1 0 0 1 15.00-15.99 m 2 0 0 0 0 2 0.400 3 5 0 0 3 11 0.688 f 0 2 0 1 0 3 0 1 1 0 3 5 ud 1 0 0 0 0 1 0 0 0 0 1 1 16.00-16.99 m 1 0 0 0 0 1 0.500 0 0 0 0 0 0 0.000 f 0 1 0 0 0 1 1 1 1 1 2 6 ud 0 0 0 0 0 0 0 0 0 0 0 0 17.00-17.99 m 0 0 0 0 0 0 0.000 0 1 1 0 2 4 0.571 f 0 0 0 0 1 1 1 1 0 0 1 3 ud 1 0 0 0 0 1 0 1 0 0 0 1 18.00-18.99 m 0 0 0 0 0 0 0.000 0 0 0 0 1 1 0.333 f 0 0 0 0 0 0 2 0 0 0 0 2 ud 0 0 0 0 0 0 0 0 0 0 0 0 CRASSOSTREA GIGAS SEXUALITY 275 TABLE 2 (Continued). DNs... aa, 1975 Group size 1976 Group size Shell length (cm) Sex 1 2 3 4 510 п s/r 1 2 3 4 5-16 n s/t II 2 A A A A A E ЕН 19.00-19.99 m 0 0 0 0 0 0 0.000 0 1 0 0 1 2 0.400 f 0 0 0 0 0 0 0 1 0 0 2 3 ud 0 0 0 0 0 0 0 0 0 0 0 0 20.00---- m 0 0 0 0 0 0 0.000 2 0 0 0 2 4 0.250 f 0 0 0 0 0 0 2 1 1 1 NE ud 0 0 0 0 0 0 0 0 0 0 0 0 TOTALS: Males 337 2497738 213721147 147 50 119’ УТ 23710377356 Females 76 54 28 LS 67 100 44 14 60 285 Undet. A A PET E 75" 25, PD A ВЕ OO Total 132 112 81 28 37 400 122 238 132 48 240 780 [m/(m+f)] 30 47. 57 62 254 45 43 (54 52 627.63 56 TT, sizes of one individual the population sex ratio is below 0.50, indicating a greater incidence of female oysters. But, with increasing group size, the population sex ratio rises above 0.50, indicating an increase in male individu- als. Keeping this in mind as well as the de- crease in the proportion of males with increas- ing shell length, it appears that large singleton oysters are more often female than male and that clusters of oysters usually consist of large females with many small male individuals. Tables 1 and 2 also give sex ratios in each of the two populations for consecutive years. The proportion of males was less in 1975 for both populations than it was in 1976 (for the same populations). Three possibilities can ac- count for these annual differences in sex ratio. They are: (1) new recruitment into the popula- tion in 1975, which would reflect an incease in the number of males for 1976 according to the protandry theory; or, (2) delayed sex change from male to female in 1976 perhaps due to unfavorable environmental conditions (i.e., a depressed food resource); or, (3) early sex change from male to female in 1975, due to favorable environmental conditions. The data from Tables 1 and 2 favor the first alternative because of the large number of oysters from the smaller shell length found in 1976 (cf. Big Beef Creek) as compared to 1975. Sexuality and Group Size The sexuality of oysters within the various group sizes (i.e., singleton, doubleton, triplet and quadruplet groups) were studied (Table 3). Samples of these group sizes were again collected from the Big Beef Creek and Dabob Bay populations in 1975 and 1976. In this study the observed number of males and fe- males per group was compared with the ex- pected group frequency assuming a random distribution of both sexes. Of the four group sizes examined only the doubleton group size (i.e., two oysters per group) exhibited a signifi- cant deviation between observed data and expected values (Table 4). The doubleton group size consistently showed a greater abundance of the male-female type than ex- pected for a random distribution. Group Size Distribution The distribution of the various group sizes was examined by random quadrat sampling of the big Beef Creek and Dabob Bay popula- tion in 1976 (Table 4). The number of oysters per group ranged from one to sixteen. The observed number of groups along with the ex- pected number of groups (assuming a geo- metric distribution) has been tabulated as well as the observed and cumulative frequency of the groups for each population. These data indicate that the observed group size frequencies between the two pop- ulations coincide well with each other. Also, the observed group size frequencies clearly indicate that the number of groups decreases with increasing group size. Singletons oc- curred in the greatest number. Assuming a geometric distribution for the number of vari- Ous groups expected within a population, it 276 BUROKER TABLE 3. Four different group sizes of the Japanese oyster Crassostrea gigas were studied with respect to the observed and expected (i.e., based on group frequency) number of males (m) and females (f) per group. The chi-squared values for goodness-of-fit of observed to expected values, degrees of freedom, and proba- bility of occurrence are listed. Where p is the proportion [males/(males + females)] of males in the popula- tion and (| — р) is the proportion of females. п is the number of groups sampled. zx ga dl SINGLETON GROUP m f Group frequency p (Ip) Year sampled 1975 1976 Population BBC DB BBC DB Chi-squared 2.778 16.963 0.016 2.470 af 1 1 1 1 Probability >.050 <.001 >.800 >.100 n 9 109 252 iz DOUBLETON GROUP mm mf ff Group frequency p? 2p(1—p) (I-py2 Year sampled 1975 1976 Population BBC DB BBC DB Chi-squared 14.535 29.666 20.486 37.068 df 1 1 1 1 Probability <.001 <.001 <.001 <.001 n 44 57 79 117 TRIPLET GROUP mmm mmf mff fff Group frequency ps 3p2(I—p) 3p(I—p)2 (I—p)3 Year sampled 1975 1976 Population BBC DB BBC DB Chi-squared 3.632 4.614 4.588 1.474 df 2 2 1 2 Probability >.100 >.050 >.020 >.300 n 10 19 14 32 QUADRUPLET GROUP mmmm mmmf mmff mfff ffff Group frequency p4 4p3(I—p) 6p2(I—p)@ 4р(!-р)3 (I-p)4 Year sampled 1976 Population BBC DB Chi-squared 1.020 0.113 df 1 1 Probability >.300 >.700 n 11 5 would appear that a deficiency of doubletons exists for both populations. This deficiency appears to be offset by an excess in group sizes of greater than two oysters per cluster. The deviation of the observed data from the expected geometric distribution is statistically significant for both populations. DISCUSSION As in other sessile marine invertebrates which have a free-swimming larval stage in their life histories (Scheltema, 1971), Crassostrea oyster larvae terminate their planktonic period by testing and finally choos- ing an attachment site on which to metamor- phose and commence a secondary life. The location or proximity of the attachment site to conspecifics would appear to be very impor- tant for the propagation of oyster species. Evidence for gregarious behavior in plank- tonic oyster larvae during settlement is docu- mented for Ostrea species (Cole & Knight- Jones, 1949; Knight-Jones, 1952) as well as Crassostrea species (Hidu, 1969). This type of behavior between oyster larvae, and be- tween larvae and adults would most certainly infer a communication system (e.g., via chemotactic secretion) (Galtsoff, 1930; Mackie & Grant, 1974; Crisp, 1974). The evo- lution of a communication system as exhibited CRASSOSTREA GIGAS SEXUALITY 277 TABLE 4. The distribution of oysters per group size in 1976 is tabulated for two oyster populations in Hood Canal. The observed number of oyster/group was obtained by random sampling of each population: four one-quarter square meter quadrats. The expected numbers were obtained by assuming a geometric distri- bution with means of 0.588 and 0.561 for the Big Beef Creek and Dabob Bay population, respectively. The chi-square values for goodness-to-fit of observed to expected values, degrees of freedom, and probability of occurrence are given. Also, the cumulative frequencies of observed groups are recorded. Big Beef Creek Dabob Bay Number of groups Frequency Number of groups Frequency Oyster/Group Obs. Exp. Obs. Cumul. Obs. Exp. Obs. Cumul. 1 114 114.0 0.588 0.588 88 88.0 0.561 0.561 2 30 47.0 0.155 0.742 25 38.7 0.159 0.720 3 18 19.4 0.093 0.835 16 17.0 0.102 0.822 4 17 8.0 0.088 0.923 6 7.4 0.038 0.860 5-16 15 5.6 0.077 1.000 22 5.9 0.140 1.000 TOTAL 194 194.0 157 157.0 Chi-square 32.117 49.000 df 4 4 Probability <.001 <.001 AA A A PY EE A A AAA AAA GG A by such gregarious behavior and the synchro- nized release of gametes by males and fe- males (Galtsoff, 1938a, 1938b; 1940) facili- tates successful reproduction in oyster popu- lations. The study of group sizes with regard to the individual sexualiy of oysters within the groups provides additional evidence (i.e., the high incidence of heterosexual doubletons) in support of a communication system. The sexuality of contiguous individuals prior to reproduction could be influenced if these individuals were of dissimilar size or age. For example, the male-female doubleton phe- nomenon (Table 3) described in this study could result from a sex or age dominance re- lationship as long as the appropriate sex alleles are present. If the opposite sex alleles are not present in the two individuals of the doubleton, then incidences should occur where the same sex is represented in these doubletons regardless of dissimilar shell size or age. Consequently, it is possible that fe- males could sequester (via chemical secre- tions) the partner of the doubleton into be- coming a male for a given reproductive cycle. Previous work with steroids on C. gigas indi- cate that estrogen greatly facilitates ovary and testes respiration while testosterone does not activate respiration in either gonad type (Mori, 1968). Therefore, estrogen (or some deriva- tive), when secreted by the female, may be responsible for stimulating gonad develop- ment in other individuals of a group. Other evidence for the existence of a com- munication system within doubletons sur- faced when the observed group-size distribu- tion was compared with that expected from a geometrical distribution (Table 4). A defici- ency of doubletons and an excess of group sizes greater than two oysters per group in the group-size distribution suggests that the two-unit (doubleton) group-size is a short- lived phenomenon. That is, the male-female relationship established through communica- tion may be responsible for attracting plank- tonic oyster larvae to settle on the doubleton and thereby generating larger group-sizes. When populations of C. gigas are parti- tioned according to their permanent attach- ment to other conspecifics, an interesting similarity appears between this species and the marine mesogastropod Crepidula forni- cata. That is, they both form multi-individual stacks; however, with respect to getting the sexes together there is an important differ- ence between the two species. C. fornicata is mobile as an adult young male while C. gigas is sedentary. Other similarities are also evi- dent. Both species are gregarious, display a planktonic larval stage of development, are sex-labile, are protandrous hermaphrodites, and have environmentally- and socially-influ- enced sex determination and hence sex ratios. These characteristics appear to be typical of mollusc species which disperse as larvae, are largely sedentary as adults and have patchy substrates over time and space (Hoagland, 1978). When fertility increases more rapidly with age or size in one sex than in the other, it 278 would be to the individual's advantage (i.e., an increase in fitness) to assume that sex last (Ghiselin, 1969; Leigh et al., 1976). This may be the case for some oyster genera (e.g., Crassostrea, Saccostrea) where the evi- dence supports protandry, but, some other relationship between sex and age exists in Ostrea where species display rhythmic sexual cycles. Usually when age- (or size-) specific sex ratios are investigated in protandrous hermaphrodites, there is often a fairly sharp change in the sex ratio at a particular age (Warner, 1975). However, this sharp change does not appear in species where environ- mental and/or social influences in sex deter- mination is evident (e.g., Crepidula fornicata; Hoagland, 1978; Crassostrea gigas, Tables 1 and 2). In fact evidence has been presented where sex reversal in C. gigas and C. virginica has occurred in the reverse direction from female to male (Amemiya, 1929; Burkenroad, 1937; Needler, 1942), while Ostrea species exhibit a rhythmic sex change cycle (Orton, 1927; Hopkins, 1937). However, it should be pointed out that these incidences occur after the initial sex reversal from male to female has already been completed. Evident- ly, after age-specific sex determination has been invoked, there is a time in the oyster's life when communicating individuals can influ- ence sex determination (i.e., at least in doub- letons). In conclusion, it appears that the Japanese oyster is like other Crassostrea species with respect to individual sexuality. That is, C. gigas exhibits a protandrous hermaphroditic pattern of sexuality by change in sex types (i.e., from male to female) with increasing shell length (or age) and increasing group size. Finally, it would appear that sex determi- nation in this oyster is influenced to some ex- tent by a social relationship between some contiguous individuals. ACKNOWLEDGEMENTS | am grateful to Messrs. Dick and Earl Steele who kindly provided me access to their private oyster beds during this study. Also, | thank the Washington State Sea Grant Pro- gram, the College of Fisheries, University of Washington and the Charles and Johanna Busch endowment, Bureau of Biological Re- search, Rutgers University, for their support of this study. Melbourne Carriker, Herbert Hidu, K. Elaine Hoagland and Robert Prezant kindly reviewed an early draft of the manuscript. BUROKER LITERATURE CITED AMEMIYA, I., 1925, Hermaphroditism in the Portu- guese oyster. Nature, 116: 608. AMEMIYA, |., 1929, On the sex-change in the Japanese common oyster, Ostrea gigas (Thun- berg). Proceedings of the Imperial Academy Tokyo, 5: 284-286. AMEMIYA, |., 1936, Effect of gill excision upon the sexual differentiation of the oyster, Ostrea gigas (Thunberg). Japanese Journal of Zoology, 6: 84-85. BACCI, G., 1965, Sex Determination. Pergamon, Oxford, vii + 306 p. BURKENROAD, M. D., 1931, Sex in the Louisiana oyster, Ostrea virginica. Science, 74: 71-72. BURKENROAD, M. D., 1937, The sex-ratio in al- ternational hermaphrodites, with especial refer- ence to the determination of rate of reversal of sexual phase in oviparous oysters. Journal of Marine Research, 1: 75-84. COE, W. R., 1932a, Inheritance of sex in oysters. Proceedings of the Sixth International Congress of Genetics, 2: 26-28. COE, W. В., 1932b, Sexual phases in the American oyster (Ostrea virginica). Biological Bulletin, 63: 419-441. COE, W. R., 1934. Alternation of sexuality in oys- ters. American Naturalist, 68: 236-251. COE, W. R., 1940, Divergent pathways in sexual development. Science, 91: 175-182. COLE, H. A. & KNIGHT-JONES, E. W., 1949, The settling behaviour of larvae of the European flat oyster Ostrea edulis L., and its influence on methods of cultivation and spat collection. Minis- try of Agriculture and Fisheries, Fisheries Inves- tigations, ser. 2, 17(3): 39 p. London: His Maj- esty’s Stationery Office. CRISP, D. J., 1974, Factors influencing the settle- ment of marine invertebrate larvae. In: Chemo- reception in Marine Organisms, GRANT, P. T. & MACKIE, A. M., eds., Academic Press, London and New York, p. 117-265. GALTSOFF, P. S., 1930, The role of chemical stim- ulation in the spawning reactions of Ostrea virginica and Ostrea gigas. Proceedings of the National Academy of Sciences, 16: 555-559. GALTSOFF, P. S., 1938a, Physiology of reproduc- tion of Ostrea virginica. |. Spawning reactions of the female and male. Biological Bulletin, 74: 461-486. GALTSOFF, P. S., 1938b, Physiology of reproduc- tion of Ostrea virginica. 1. Stimulation of spawn- ing in the female oyster. Biological Bulletin, 75: 286-307. GALTSOFF, P. S., 1940, Physiology of reproduc- tion of Ostrea virginica. Ш. Stimulation of spawn- ing in the male oyster. Biological Bulletin, 78: 117-135. GHISELIN, M. T., 1969, The evolution of hermaph- roditism among animals. Quarterly Review of Biology, 44: 189-208. HALEY, L. E., 1977, Sex determination in the CRASSOSTREA GIGAS SEXUALITY 279 American oyster. Journal of Heredity, 68: 114- 116. HIDU, H., 1969, Gregarious setting in the American oyster, Crassostrea virginica Gmelin. Chesa- peake Science, 10: 85-92. HOAGLAND, K. E., 1978, Protandry and the evolu- tion of environmentally-mediated sex change: a study of the Mollusca. Malacologia, 17: 365- 391. HOPKINS, A. E., 1937, Experimental observations on spawning, larval development and setting in the Olympia oyster, Ostrea lurida. Bulletin of the Bureau of Fisheries (U.S.A), 48: 439-503. KATKANSKY, S. C. & SPARKS, A. K., 1966, Sea- sonal sexual pattern in the Pacific oyster, Crassostrea gigas, in Washington state. Fish- eries Research Papers, 2: 80-89. Washington Department of Fisheries. KNIGHT-JONES, E. W., 1952, Reproduction of oysters in the Rivers Crouch and Roach, Essex, during 1947, 1948 and 1949. Ministry of Agricul- ture and Fisheries, Fisheries Investigations, ser. 2, 18(2): 47 p. London: His Majesty’s Stationery Office. LEIGH, E. C., Jr., CHARNOV, E. L. & WARNER, R. R., 1976, Sex ratio, sex change, and natural selection. Proceedings of the National Academy of Sciences, 73: 3656-3660. MACKIE, A. M. & GRANT, P. T., 1974, Interspecies and intraspecies chemoreception by marine in- vertebrates. In: Chemoreception in Marine Or- ganisms, GRANT, P. T. & MACKIE, A. M., eds., Academic Press, London and New York, p. 105— 141. MORI, K., 1968, Effect of steroid on oyster. II. In- crease in rates of fertilization and development of Estradiol-17B. Bulletin of the Japanese So- ciety of Scientific Fisheries, 34: 997-999. MORI, K., MURAMATSU, T. & NAKAMURA, Y., 1969, Effect of steroid on oyster. III. Sex reversal from male to female in Crassostrea gigas by Estradiol-17B. Bulletin of the Japanese Society of Scientific Fisheries, 35: 1072-1076. NEEDLER, A. B., 1932, Sex reversal in Ostrea virginica. Contribution of the Canadian Biology and Fishery. 7: 285-294. NEEDLER, A. B., 1942, Sex reversal in individual oysters. Journal of the Fisheries Research Board of Canada, 5: 361-364. ORTON, J. H., 1927, Observations and experi- ments on sex-change in the European oyster (O. edulis), Part |. The change from female to male. Journal of the Marine Biological Association of the United Kingdom, 14: 967-1055. QUAYLE, D. B., 1969, Pacific oyster culture in Brit- ish Columbia. Bulletin of the Fisheries Research Board of Canada, 169: 192 p. REINBOTH, R., 1975, /ntersexuality in the animal kingdom. REINBOTH, F., ed., Springer, Berlin, Heidelberg and New York, 448 p. SCHELTEMA, R. S., 1971, Larval dispersal as a means of genetic exchange between geographi- cally separated populations of shallow-water benthic marine gastropods. Biological Bulletin, 140: 284-322. WARNER, В. R. 1975, The adaptive significance of sequential hermaphroditism in animals. Amer- ican Naturalist, 109: 68-82. MALACOLOGIA, 1983, 23(2): 281-312 BIOLOGY OF THE NORTHEASTERN PACIFIC TURRIDAE. 1. OPHIODERMELLA Ronald L. Shimek University of Washington, Friday Harbor Laboratories, Friday Harbor, WA 98250, U.S.A. ABSTRACT Behavior, environment, and reproduction of Ophiodermella inermis and O. cancellata were studied. Ophiodermella inermis was found to be a specialist vermivore upon the polychaete Owenia fusiformis. It lives in intertidal to shallow subtidal habitats characterized by the nearby presence of dense O. fusiformis beds. Predation by O. inermis may locally affect the Owenia populations, and the polychaete assemblage in general, as it can eat from 8-16% of the standing crop of this dominant polychaete. Ophiodermella inermis shows little capability for distance chemoreception, and finds prey by moving up-current until it encounters a worm. Predation on O. inermis by Cancer gracilis and C. productus, which may limit recruitment, is significant. The snail has a size refuge from C. gracilis but not C. productus. In the intertidal area where both crabs are common, recruitment is very rare and adult mortality is high, resulting in a declining population. Growth is rapid when the animals are small and effectively ceases when they reach 31 mm in length. Rapid juvenile growth may be an adaptation to heavy juvenile mortality. Ophiodermella inermis reproduces from October to July, with a peak of egg capsule deposition in February. Each female deposits from one to nine egg capsules per night and each capsule contains several hundred eggs. Encapsular development time is about 50 days. There are no nurse eggs, and all eggs develop to leave the capsule as planktotrophic veligers, which develop for at least five more weeks. Settlement was not observed. Ophiodermella cancellata is only found subtidally in silty or shell fragment habitats. It is also a specialist predator upon an oweniid polychaete, Myriochele oculata. In the subtidal habitat examined, recruitment is temporally patchy. Adult mortality is continuous, relatively few snails exceeding 12.0 mm long. Alternative hypotheses explaining this mortality are proposed. This species appears to mate in summer and settle in winter, but capsule deposition was not ob- served. Both species occur in habitats characterized by stable sediment particle distributions. Polychaete assemblages may change drastically and turrid predation may be a causative agent for these changes in the subtidal habitats. Environmental stress effects are unimportant subtidally, but drastically affect the intertidal population of O. inermis, causing behavioral changes and some mortality. Key words: Ophiodermella; Turridae; ecology; diet; habitat; subtidal; vermivores. INTRODUCTION Toxoglossan gastropods are often impor- tant components of tropical ecosystems, and the biology of Conus, in particular, is reason- ably well known (Kohn, 1959, 1967, 1968; Kohn & Nybakken, 1975; Leviten, 1976, 1978; Nybakken, 1978). In many tropical ecosys- tems the most obvious toxoglossans are in the families Conidae and Terebridae, and members of a third family, the Turridae, are less evident. In colder water, boreal, and deep-sea ecosystems, however, the Turridae are the only toxoglossan family represented and many turrids are important components of the fauna (Hartman, 1955; Jones, 1950; Parker, 1964; Wade, 1972; Rex, 1976). While the morphology of a few turrids has been ex- amined (Franc, 1952; Robinson, 1960; Smith, 1967a, 1967b; Sheridan et al., 1973; Shimek, 1975), the immense size of the group has pre- cluded an adequate understanding of the re- lationships within it (Shimek & Kohn, 1981). Ecological information is lacking, with few published observations on feeding, habitats, or reproduction (Pearce, 1966). In some habitats of the Puget Sound region of Wasington, turrids are relatively abundant. The most common turrids in these areas are generally members of the genus Ophio- dermella. | examined turrid populations to de- termine habitat, dietary requirements, and aspects of predatory and reproductive be- havior. Several major questions were ad- dressed: 1) What is the relationship of diet to the potential dietary resources present? 2) (281) 282 SHIMEK Are these animals dietary specialists or gen- eralists? 3) What is the effect of these preda- tors upon their prey populations”? 4) Are there any peculiar traits that limit their choices of habitats or prey? 5) Is there any obvious rela- tionship between reproductive biology and distribution? In addition to attempting to an- swer these specific questions, | made obser- vations of natural history attributes to try to make some generalizations on turrid biology. MATERIALS AND METHODS Two species of Ophiodermella are found in the region: the smaller, O. cancellata (Car- penter, 1864) ranges from southern Alaska to California (Grant & Gale, 1931); the larger can be referred to as O. inermis incisa (Car- penter, 1864), type-locality Neah Bay, Wash- ington, but for this study | will refer to it as O. inermis. There is some doubt whether O. incisa is a separate species (McLean, per- sonal communication). Ophiodermella inermis (Hinds, 1843) ranges from southern California to the Pacific Northwest (Grant & Gale, 1931; Cernohorsky, 1975). (Fig. 1). Study sites Ophiodermella inermis was studied most intensively at an intertidal location on the east shore of Port Washington Narrows (47° 32’ 39” N, 122° 39’ 15” ММ) and in a shallow subtidal location near Windy Point, Dyes Inlet (47° 37' 25" N, 122° 40’ 30" W). Additional O. inermis animals were collected at Alki Point in Seattle (47° 39’ 48” N, 122° 26’ 06” W) (Fig. 2). The Port Washington Narrows (PWN) study site is oriented north-south, is about 180m long and averages 20 m wide, although it is almost 40 m wide at the widest point. This site was sampled in six areas: high, +0.5 m; mid- dle, 0.0 m; and low, —0.5 m to —0.8 m; in both sand and cobble areas. All tidal heights are in respect to datum, 0.0m, defined as mean lowest low water. The sampling areas were chosen to cover all habitats where turrids were seen in preliminary observations. The beach was bounded at both ends by rocky, reef-like areas and subtidally by an area of high currents and moving “sand dunes” inter- spersed with areas of large shell fragments. Repeated preliminary observations in the rocky and subtidal areas indicated that no tur- rids are found outside the beach area, and therefore the turrid population was treated as a closed population. The beach itself could be divided into sandy or cobble areas. The cob- ble area, in the center of the beach and bounded on each side by sand, is 60-80 m long and consists of stones 0.5-2.0 m in di- ameter embedded in sand or gravel sub- strate. The Windy Point (WP) area is completely subtidal, from —1.5 m to —9.0 m below MLLW and is bounded on the upper edge by a dense bed of sand dollars, Dendraster excentricus (Eschscholtz, 1831), and on the lower edge by a bed of sea pens, Ptilosarcus gurneyi (Gray, 1860). No turrids were found among the sand dollars and only an occasional indi- vidual was seen in the sea pen bed. The site was not bounded laterally, and appeared to extend over 2 km with no noticeable change. Four habitats were sampled: a gently sloping 20-30 m wide upper area (subarea upper bench) from —1.5 m to —3.0 m; a 10 m wide slope from —3.0т to —6.0m (subareas upper and lower slope) and a gently sloping 50 m wide lower area from —6.0 m to —9.0 m (subarea lower bench). The subareas or habi- tats were determined by depth and steepness of slope. The upper and lower benches are virtually level. The steeply sloping area was divided by depth into upper and lower halves based upon differences in algal cover. The upper bench and slope areas were covered in summer with a thick, 1.0 to 1.5 m deep at high water, mat of ulvoid algae, mostly Entero- morpha. The algal cover on the lower slope and bench consisted of scattered Neoagard- hiella baileyi (Kutz.) Wynne & Taylor, 1973, Anhfeltia concinna J. Agardh, 1847, and Desmarestia ligulata (Lightfoot, 1777). Ophiodermella cancellata was studied most intensively subtidally off the University of Washington Friday Harbor Laboratories dock, San Juan Island, Washington (48° 32' 38" N, 123° 00' 50” W) (Fig. 2). This study site (FHL) was the most topographically diverse of the major study sites containing at least five visually distinct habitats: areas of rock, wood chips, shell fragments, and upper and lower areas of silty mud. This site was located in Friday Harbor Bay at depths of -10 m to -25 m below MLLW, and consisted of two permanent 100 m tran- sect lines and the nearby substrate, up to 100 m from the lines. No discontinuities limit- ed the site except at the upper edge where the boundary was the lower edge of an eel- grass, Zostera marina L., 1753, bed. Prelimi- OPHIODERMELLA BIOLOGY 283 FIG. 1. A. Ophiodermella inermis. B. Radular teeth of O. inermis. C. O. cancellata. D. Radular tooth of O. cancellata. Scale bars in A and C are 5 mm; in B and D, 100 um. Arrows indicate healed fractures. 284 SHIMEK O 20 km FIG. 2. Major study sites in Washington state. FHL = Friday Harbor Laboratories, San Juan Id.; PWN = Port Washington Narrows; WP = Windy Point. nary observation established that no turrids were found in the eelgrass bed. Preliminary observation also indicated no turrids, indeed virtually no macroscopic biota of any sort in the wood chip areas, conse- quently no sampling was done in those areas, otherwise the substrate appeared acceptable to turrids at least 100 m from the transects. No infaunal sampling was done in the rocky areas, characterized by large boulders and areas of bedrock, often several meters across, even though some turrids were seen in those areas. The tops of boulders and bed- rock areas, the cracks in them, and the inter- stices between them were covered with silt, generally in very shallow layers. These pock- ets of sediment were too small to sample ef- fectively. The remaining habitats were all sampled quantitatively for infaunal polychaetes and turrid distribution. The upper (—10m to —12 m) and the lower (—17 т to —20 m) silt areas, distinguished on the basis of depth related biotic factors such as diatom and algal cover, were otherwise similar topographically. These areas were not often confluent, being separated by shell fragment and rocky areas, but when they were they intergraded. The shell fragment area was similar to the silt areas, but with many more visible shell frag- ments. The shell fragment zones, usually near the rocky areas, could be found as far as 30 m from the rocks. No shell fragment areas were found above —13 m, and they seemed to be found only in areas of gentle slope. Habitat analyses At the three major study sites the physical and biological components of the major habi- tats were examined in detail. Two bimonthly replicate 0.018 m” cores were taken about 1 m apart to a depth of 10 cm in each habitat and fixed in 2-5% sea water formalin with Rose Bengal. Sediment analysis was per- formed by removing an aliquot of each sam- ple, sieving it to remove all particles greater than 1.8 mm, which were then separated into size classes determined by standard @ nota- tion (Holme & McIntyre, 1971). The remainder of each aliquot was subsampled and two or three replicates were analyzed with a settling tube (Emery, 1938). A weighed subsample of the aliquot was sieved to remove the silt-clay fraction which was dried and weighed. Sam- ples were analyzed graphically (Inman, 1952) and median particle size and sorting coef- ficient were computed. The remainder of the sample was washed through a 0.5 mm sieve and the animals were collected and sorted by taxon. Gastropods and polychaetes were counted and identified to species when possible. Other taxa were OPHIODERMELLA BIOLOGY 285 identified to class and counted but not de- tailed further. The two replicate samples were pooled for further biological analysis making the total area per sample 0.036 т". Sediment parameters were tabulated and compared between the areas using the Wil- coxon Rank-Sum method (Hollander & Wolfe, 1973). Seasonal variability of the sediment parameters was insignificant, so no statistical comparison on the seasonal basis was done. In situ measurements of PWN snail and habitat temperatures were taken with a YSI portable electric thermometer with a 3 mm di- ameter probe. Polychaete assemblage abundances for each habitat were determined by quantitative infaunal sampling. The assemblages were compared between and within habitats on a seasonal basis using the “D” index (Whitta- ker, 1952; Whittaker & Fairbanks, 1958; Schoener, 1968; Pielou, 1974) to measure similarity and Н” (Kohn & Nybakken, 1975) to measure heterogeneity. The June and De- cember polychaetes from each habitat were divided into prey and non-prey fractions based upon turrid fecal determination, dried, weighed, incinerated at 550°C, and re- weighed. Turrid distribution, collection and processing At the three major sites periodic transect studies from November, 1973, until Decem- ber, 1975, were used to determine turrid dis- tributions, seasonal or other changes in those distributions, and to provide a reference for all quantitative infaunal samples. At the PWN sites, the transects were temporary because of human disturbance. Intersampling period deviation was minimized by placing the tran- sects with reference to mapped landmarks. At the FHL site two 100 m transects were per- manently emplaced. Habitat types and turrid positions relative to the transects were noted. Turrids were collected, washed in sea water, placed individually into marked jars filled with filtered sea water, and maintained at 10°C for two to seven days. The animal was then removed from the jar, its length, width, and aperture length measured with calipers to the nearest 0.1 mm, and blotted dry. The shell was Cleaned and an identifying code was ap- plied with standard drawing ink which was covered with a layer of clear fingernail polish. After this layer had dried the animal was re- turned to fresh sea water at 10°C and ob- served to assure no noticeable effects of the measuring and marking procedure. An ink line was placed at the outer edge of the outer lip of the aperture to assess growth. This line was not covered with fingernail polish, as the solvent in the polish was lethal in some cases if applied near the aperture. The animal was then transferred to a “holding” aquarium where it was maintained in an artificial habitat similar to the normal one. All apparently healthy animals were returned to their habitat, although seldom to their point of capture, within two weeks. Measuring and marking mortality was less than one percent. Recaptured marked snails allowed estima- tion of growth rates. Changes in total length were used to calculate growth rate. Shell shrinkage was caused by fracturing the outer lip and siphonal canal by crabs, and/or apical erosion. Body whorl height, as measured by aperture length, and total length were highly correlated; but as the aperture length is short, similar changes in length were relatively larger and more variable. Animals were as- signed size classes on the basis of pregrowth size, and the mean growth rate in um day-1 was calculated. For O. inermis, the WP popu- lation allowed a second determination of growth rate, as the size-frequency histogram of the population contained a number of dis- tinct peaks, presumbly corresponding to year- ly recruitment. This histogram was assumed to be the result of several overlapping normal distributions, and was separated into com- ponent distributions following the method given in Bliss (1967). A similar procedure was used to calculate growth rate for O. cancellata at FHL. The recapture rate of the WP popula- tion was so low that no independent verifica- tion of these assumptions could be made. Mark-recapture data also allowed estima- tion of population size at the PWN area which was geographically bounded, and where the population was considered to be closed. Pop- ulation size was estimated by Jolly's Stochas- tic Multiple-Recapture Method (Poole, 1974). These data were plotted and analyzed graph- ically. Similar analyses were not attempted at either the WP or FHL areas, as they were not bounded, and too few recaptures per unit time were made to make the population estimates meaningful. At PWN seasonal snail distribution was determined by plotting snail position on maps of the areas. A 60 x 15 m scale rectangle was plotted on the same maps with the vertical midline being the boundary between the sand and cobble habitats, and the horizontal mid- 286 SHIMEK line being the —0.3m contour line. These lines divided the center of the sampling area into four equal area rectangles, each 30 x 7.5 m, one each in the sand and cobble area from —0.8 m to —0.3 m, and —0.3 to +0.3 m. Seasons were defined as summer (May, June, July) and winter (November, Decem- ber, January). The number of turrids plotted in each area was summed for each season, di- vided by the number of surveys per season, and tested using the log-likelihood ratio to determine any seasonal differences between substrates or heights. Following measuring and marking, any par- ticulate material remaining in the collecting jar was placed on a slide, dried, mounted in poly- vinyl lactophenol (A. Kohn, personal commu- nication), and examined. Identification of all fecal material was attempted. Feces con- sisted of mucus, radular teeth of the same animal, diatom frustules, and polychaete re- mains. Preliminary gut analysis by dissection Owenia fusiformis Peu | 5 um O AM Ventral uncini FIG. 3. Owenia fusiformis, ventral view, modified from Dales (1957); and ventral uncini from Myriochele oculata, M: and O. fusiformis, O. B = bands of uncini. C = collar width. S = shoulder length of an uncinus. OPHIODERMELLA BIOLOGY 287 indicated that polychaetes swallowed whole were the only prey. Animals collected for these gut analyses were preserved immedi- ately after capture in 70% ethanol and boiled for 10 minutes upon return to the laboratory. It is unlikely that any rapidly digested animals were undetected by these analyses. Thus, only polychaete remains consisting of setae, jaws, and occasional cuticular strips were ac- cepted as indicators of feeding. These re- mains were identified by comparison with slides prepared of known polychaetes, and comparison with description and drawings of Berkeley & Berkeley (1952), Woodwick (1963), Blake (1966, 1971), Hartman (1969), Blake & Woodwick (1971), Foster (1971), Banse (1972), Banse & Hobson (1974), and Hobson & Banse (1981). Some captured O. inermis were starved for one week, placed in a dish with their most common prey, Owenia fusiformis delle Chiaje, 1844, and the time from ingestion to defecation was noted. Owenia fusiformis diameter at the top of the collar (Fig. 3) was measured for five large and five small worms. In addition the shoulder lengths of 10 uncini from each worm were measured. The uncini shoulder length corre- lated well with the collar diameter; г” = 0.96; (worm collar diameter (um) = —1963 +396 (uncinus shoulder length (um))). To test whether different sized worms were being eaten by different O. inermis populations, fecal samples from each population were randomly chosen, examined, and uncini shoulder lengths were measured. Sample means were compared with a “t”-test. Laboratory experiments Ophiodermella inermis was tested in a “Y” choice chamber (Fig. 4) to determine its re- sponse to currents and ability to sense prey. In the response to prey experiments, 15-20 Owenia fusiformis were placed in a small cloth bag in a randomly chosen arm of the apparatus, with an empty bag in the other. Ten snails were placed at the base of the stem of the “Y,” a current of 2.5 cm sec—1 was applied to both arms, and the experiment was run for at least eight hr. Similar experi- ments were run without bags or worms to test the response to current, and an additional set of unbaited experiments was run with a cur- rent of 5.0 cm sec-1 in one arm with no cur- rent in the other. The animals were scored as having made a choice when they were more FIG. 4. “Y” choice chamber for chemo- and rheo- taxis experiments. Arrows indicate the direction of water flow. N = No-Choice area. C = Choice areas. All arms were 30.5 cm long. Cross-hatched area is the starting position for the snails. than two centimeters into an arm at the com- pletion of the run. The tube was cleaned be- tween each set of snails by scrubbing to re- move all traces of mucous trails. Results were compared with log-likelihood tests and cumu- lative binomial probabilities. Animals from PWN were offered to two po- tential predators, Cancer gracilis Dana, 1852, and C. productus Randall, 1839, to determine if the crabs would eat the snails. Individual O. inermis were placed in an aquarium with an individual crab. After 24 hr the aquarium was examined to see if the crab had eaten the snail, or chipped the shell attempting to eat it. The crabs had been starved for one week prior to the experiments, and most of the crabs attempted to eat the snails, conse- quently no post-experimental verification of crab hunger was attempted. Attempted unsuccessful crab predation can be assessed by counting the number of healed fractures (Vermeij et al., 1980). Ran- domly chosen Ophiodermella inermis from both PWN and WP areas were assessed for the number of healed fractures by the method of Vermeij et al. (1980). The total number of healed fractures (Fig. 1) was determined, and subdivided into those in the top 10 mm of the shell, measured from the apex to the suture line 10 mm from the apex, and those in the remainder of the shell. Only shells of the WP habitat size range, 15.0 mm to 30.6 mm, were 288 SHIMEK > r SO cm O pS — SAND 90: cm | НЕ FIG. 5. Choice chamber for substrate choice experiments. А. Starting агепа. compared. The difference in the means of these samples were compared using a ‘t’- test. A substrate choice chamber was construct- ed from an aquarium with a plexiglas partition extending from one side to 8cm from the other side on the midline, thus dividing the aquarium into two equal area chambers. The partition had holes drilled in it to insure access from one chamber to the other. A starting arena cleared of sediment was maintained at the partition (Fig. 5). One chamber was filled to a depth of 2 cm with sand, sieved to insure particle size distribution from 0.250mm to 0.500 mm. The other half had shell fragments sieved to insure particle size distribution in ex- cess of 2.00 mm. Both sediment types had all noticeable biota removed. Turrids of one spe- cies were placed in the starting arena, and one week later they were collected and their positions noted. These data were analyzed using cumulative binomial probabilities, ex- cluding all animals in the starting arena, on the walls, or partition. Egg capsules were collected in the field or from containers that turrids were stored in. Capsule dimensions (Fig. 6), the egg number per capsule, and the egg diameter were measured. Developmental stages were deter- mined for field collected capsules. The cap- sules were examined periodically and upon hatching, the veligers were maintained in fil- tered sea water containing either /sochrysis sp. or Dunaliella sp. Water in the cultures was FIG. 6. Egg capsule of Ophiodermella inermis showing measurements taken. H = height. L = length. W = width. Dotted lines indicate embryos. changed periodically. The apparently mature veligers were given a variety of potential sub- strates to metamorphose upon. OPHIODERMELLA BIOLOGY 289 RESULTS Habitat descriptions Port Washington Narrows: The PWN study area is on the east shore of Port Washington Narrows about 2km N of Bremerton, Washington. Algal cover on the beach varied seasonally. In the summer, ulvoids in a single layer covered most of the beach up to +0.3 m. In the winter, algal cover was less than 5% in most areas. In the lowest intertidal areas, patches of Gigartina papillata (C. Agardh, 1821, and Opuntiella californica (Farlow, 1877) are found, and in the cobble areas are occasional patches of Hedophyllum sessile (C. Agardh, 1824). This study site is a narrow tidal channel characterized by periods of high currents, up to 7.5 km hr-1, and thus the sediment is well aerated and no black, sulphide-smelling anoxic areas were en- countered, except occasionally at a depth of nine to ten centimeters in the summer in the higher areas. Substrate temperatures are highest in summer (Fig. 7). During the winter, substrate temperatures approximated water temperatures and averaged about 9°C. No periods of freezing were encountered during the study. Sediment physical parameters, median particle size and sorting coefficient were sub- stantially different among the six PWN habi- tats. Compared with one another on a habitat by habitat basis (Tables 1, 2), the results indi- cate that the beach is patchy in both particle size distribution and sorting. The high sand areas contain some rocks, and all areas have holes dug by clammers, up to several hun- dred per tidal period, which are filled with un- consolidated sediments. Generally, the higher 2 5 4 вовне хРОЗЕВ FIG. 7. Temperature of the exposed substrate at Port Washington Narrows. Values were taken оп зиппу days only in the months indicated and are the means of from two to five values. The range of values is shown for July, 1974 only. Similar ranges were seen for all months. 290 SHIMEK TABLE 1. Physical parameters—all habitats. Port Washington Narrows (PWN) Habitat High cobble (HC) Middle cobble (MC) Low cobble (LC) High sand (HS) Middle sand (MS) Low sand (LS) Windy Point (WP) Upper bench (UB) Upper slope (US) Lower slope (LS) Lower bench (LB) Friday Harbor Laboratories (FHL) Upper mud (UM) Shell fragments (SF) Lower mud (LM) Median particle size (mm) 2.11 1.02 0.42 1.10 0.46 0.43 0.42 0.29 0.36 0.40 1729 1.10 0.71 Sorting coefficient 3.79 2.91 1.65 2:92 2.20 2.49 1:51 1.54 1.45 1.31 2.69 210 2.79 TABLE 2. Comparison of differences in physical parameters. Tested with Wilcoxon Rank-Sum. O = not signifcant; + = significant at a = 0.05; ++ = significant at a = 0.01. Habitat abbreviations as in Table 1. MC PWN WP РНЕ 5Б НС ++ ++ Mes Le ++ 0 ++ 0 ++ 0-40 20 0: 9h:0 0 0 HS ++ ++ Habitat № 15 . “UB US ++ ++ ++ ++ + + | ++ + + + | O0 0 + 0 | gece o | ++ ++ и ++ 0 ++ 0 ++ + ++ 0 | ++ ++ | ++ ++ | ++ ++ Median particle size (Md й) LS ++ ++ 0 ++ + + LB ++ ++ 0 ++ 4 ++ | | | } ! h | t UM SF LM ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ oo D---0W0 =30=0==>=000 PD CO EH 5_ÁERR—— EEE OPHIODERMELLA BIOLOGY 291 sediments are coarser and less sorted than the lower ones. Superimposed on this pattern is the tendency for the cobble area sediments to be coarser and less sorted than those at the corresponding tidal height in the sand areas, although these differences are not al- ways significant. The beach topography did not change sea- sonally. The area is well protected from wave action; most waves are caused by boat traffic in the adjacent channel. There is some fresh water runoff and erosion, primarily in the win- ter when channels up to five centimeters deep are dug several times each winter. Generally, these channels were filled in by tidal action within two days. Seasonal polychaete trends are summa- rized on a habitat by habitat basis (Tables 3, 4). The sand and cobble areas have different polychaete assemblages, although overlap is high. In summer, the high cobble area has a very high density of Pygospio elegans Claparède, 1863 and a low diversity of organ- isms in general. The remainder of the summer PWN samples are all similar, D = 0.637, re- flecting the proportions of Owenia fusiformis, 800 LS MS L m 450 E A . 900 N 450 U M B 900 E R 450 / 5 1125 5 900 U д 450 В Е 900 M E 450 = E R 35 10 35 10 3 5 COLLAR DIAMETER Platynereis bicanaliculata (Baird, 1863), Malacoceros glutaeus (Ehlers, 1897), Axiothella rubrocincta (Johnson, 1901), and Capitellids. In winter, the assemblages be- come more distinct. The high cobble assem- blage is still characterized by high densities (>25,000 m-2) of Pygospio elegans. The high sand assemblage loses many of the rarer species and the O. fusiformis density drops from over 500 т- 2 to less than 20 т-2. The middle and lower areas remain similar, D = 0.620, with O. fusiformis, Axiothella rubrocincta, and capitellids as the dominant organisms. The mean size of O. fusiformis, the principal turrid prey, peaks in late sum- mer. A large component of small worms is added to the population in late summer-early autumn, coincident with the mortality of large individuals (Fig. 8), thus O. fusiformis appears to be an annual species. The summer prey biomass is concentrated in large individuals, while the winter populations contain many small worms. Windy Point: The Windy Point sediment physical para- meters (Tables 1, 2) indicate that the WP sub- 35 10 35 10 [100 pm] FIG. 8. Size frequency distributions of Owenia fusiformis in Puget Sound habitats. Port Washington Narrows habitats: LS = lower sand, MS = middle sand, LC = lower cobble, MC = middle cobble. Windy Point habitats: UB = upper bench, US = upper slope. NC = none collected. £Gt'0 2490 £0t'0 GT9°0 LEO 999`0 €8€°0 66c'0 cee 0) Itc'0 | 882°0 8Gc'0 a ; a = x 0S2°0 9/6`0 = u Ag, | a M И 3$ THA 209°0 €65°0 767`0 Tre 0 182°0 ÿLe"0 L62°0 WA 0z2"0 £vc'0 téc 0 L8v°0 §S9°0 663`0 657`0 91 9S2°0 ¿620 8£c'0 IT/°0 €8t'0 815°0 877`0 $1 dM 812°0 6€2°0 6220 [59:0 529.0 L0S°0 069°0 sn JEWLUNS чото с9Т"0 SIT 0 146°0 v85"0 ¿190 609°0 an ÿ0£"0 [Z2°0 S8T'0 005*0 $1 v82"0 v62"0 £8T'0 6€€°0 SW 882°0 062°0 £92°0 [67`0 SH L02°0 v02°0 v9T"0 Gvt'0 3 4 NMd 00€*0 sv2"0 390 Orr’ 0 IW 6S0°0 £80°0 €v0°0 ell 0 JH SW HA dM NMd ‘| alge, ul se иоцелелаае уемаен ‘(8961 “sausoyos) а, = хери! deueng 'sdeueno эбеашез$е ejeeuoÂlod uosees x jeyqey x JEIIGEH € 318V1 292 293 OPHIODERMELLA BIOLOGY ee оЭд——д—д—д—————————————ЮЮ—————а—д—ад——————————ооддд v 61 6€ 69 Ed 90 VL rol 96 Gs ol 9`5 ¿81 ‘dds Aaid-uoN Ol GE Ss! e! 20 20 70 Lal 5'0 Ут L'O> Z0 ‘dds Asid puni (гб) зчбем Aug 991} USE UPOW ¡A el 7 es occ LL Sve 58'© Ov'e 857 977 €t'O 00"! 1H lerze 0052 4165'8 8992 228 ггг‘. 558% 9082 6956 с/с 1986 299€ z—W Jequinu UESW "sajeeyoAjod jje—ejep abejquiassy ‘8 01 SEINE | 5981 ‘эрэледе! suebaja оа$об/Аа 1 (1681 ‘51э!493) зпавзи/б S018909E/EN G v | | | | N x | | Ol (0821 ‘sniouqe4) ebuo, эиог}3 pret 'efeiyo эпэр sIiuuojsn/ EIU8MO (1561 '“suoqinag) sisuayabnd sojdojo9s - _ _ (9061 ‘2100N) SA9/Q eIpueuy € = - (5981 ‘pureg) eyeynoyeueriq sjasaUuAje|d = -ox co © © + wearer Nn оо | © | ow — © © O — © v 6 (LOGL ‘uosuyor) BJOU/IDOIQNI е/эщоху - - ¿261 ‘Аээжеа eJo/d epulsAo г LOGL ‘UOSUUOF 5/еэлод snpodiwasy € (6061 ‘2100N) S/NU8] snjsewojoN 9 (2v61 ‘чешиен) ejasiqwe snjsewoipeyy 8 (08/1 ‘sniouqe4) ejeydeo ejjaydeo saloads M S M S M 5 M S M S M 5 uosees s1 SIN SH on OW OH тераен ‘59109э4$ us} do] ‘миел [еэнэшптм “y DONNO | ¡| ONONO¡|O— | TOO = | | | | | o a | | | | +0 N SO oo "au | | | | D ‘| эае1 ul se suonenalqae yeygey ‘sisAjeue abejquiasse ajeeyoAiod NMd ‘+ 3718VL 294 SHIMEK TABLE 5. WP polychaete assemblage analysis; habitat abbreviations as in Table 1. De A A A AA AAA NI ee ee ee ee el A. Numerical rank, top species. Habitat UB US LS LB Season S W S W S W S W Species Capitella capitata (Fabricius, 1780) Mediomastus ambiseta (Hartman, 1947) Notomastus tenuis (Moore, 1909) Tharyx multifilis Moore, 1909 Glycinde picta Berkeley, 1927 Lumbrineris spp. Blainville, 1828 Axiothella rubrocincta (Johnson, 1901) Nephthys caecoides Hartman, 1938 N. ferruginea Hartman, 1940 Platynereis bicanaliculata (Baird, 1863) Scoloplos pugettensis (Pettibone, 1957) Owenia fusiformis delle Chiaje, 1844 Eteone longa (Fabricius, 1780) Malacoceros glutaeus (Ehlers, 1897) Polydora socialis (Schmarda, 1861) Others B. Assemblage data—all polychaetes. Mean number т-2 3,611 2,750 3,519 2,630 H” 2.22 2.52) 2:39" 22:84 2.83 2.59 2.58 73102 Mean ash free Dry weight (g m2) Turrid prey spp.* 4.0 0.8 3.0 0.9 3:3 5.4 1.0 0.3 М оп-ргеу spp. 3.6 16.8 15 75 06 38 34 zen III ee — ро | œ © y (Mona 6905 | bb Г On! ON Maar 5 | lwooon | — lIsuouo!lomno-—- ı=-wmoaoolw|» | w ors | ora | | JR a | — | nm im O ND OO BR © | | — о | | Го INC = INA ао о! о © bh. on | — о er = o © Y *Including Polydora socialis. areas are similar to one another and to the PWN lower and middle cobble, and lower sand subareas. The polychaete fauna of the upper slope and bench areas is similar to the PWN lower habitats, particularly in winter (Tables 3, 5). There is little overlap between the lower WP habitats and any of the PWN habitats particu- larly as regards the important prey species, Owenia fusiformis. Only in the upper bench area was O. fusiformis in any appreciable densities. The secondary prey species, Poly- dora socialis (Schmarda, 1861), was abun- dant in all WP habitats especially in the sum- mer, although it was less common in the win- ter when O. fusiformis, if present, was more abundant. Particularly in the summer, P. socialis contributes substantially to the prey biomass, which can be somewhat misleading, as it is a distinctly minor food component. All of the prey biomass in the lower areas was due to Polydora socialis. As at PWN, the poly- chaete assemblages of the upper and lower areas became more distinct in winter. Friday Harbor Laboratories: The sediment analyses of the FHL habitats indicate that the sampled areas are quite homogeneous (Tables 1, 2). The lower mud area is the only one of the unconsolidated sediment areas to be somewhat different, with a distinctly smaller median particle size. How- ever, the difference is not statistically signifi- cant. Even with many more visible shell frag- ments, the median particle size of the shell fragment area is similar to that of the upper mud area. The sorting coefficients of all the areas are similar. The other physical factors affecting the FHL areas do not vary between the habitats. Tem- perature varied yearly from 6° to 11°C. Salinity was not measured in situ, but measurements at the nearby laboratories indicated that the normal salinity of 27-29°/.. did not vary wide- ly. Some seasonal variation in salinity was present due to flooding of some large local rivers. However, such events were not of long duration, and were generally surface phe- nomena not reaching the — 10 m depths nec- essary to impact these areas. Current velocity 1m above the bottom seldom exceeds 20 ст sec- 1, peaking at about 70 cm sec”! only during periods of maximum tidal ex- change. The FHL polychaete fauna is diverse, and OPHIODERMELLA BIOLOGY 295 TABLE 6. FLH polychaete assemblage analysis; habitat abbreviations as in Table 1. QT A. Numerical rank, top ten species. Habitat Season S Drilonereis falcata Moore, 1911 = Capitella capitata (Fabricius, 1780) — Mediomastus ambiseta (Hartman, 1947) Chaetozone setosa Malmgren, 1867 Cirratus cirratus Müller, 1776 Tharyx multifilis Moore, 1909 Axiothella rubrocincta (Johnson, 1901) Nephthys ferruginea Hartman, 1940 Pholoe minuta (Fabricius, 1780) Laonice cirrata (Sars, 1861) Prionospio cirrifera Wirén, 1883 P. steenstrupi Malmgren, 1867 Polycirrus caliendrum Claparède, 1868 Terebellides stroemi Sars, 1863 Others B. Assemblage data. All polychaetes. Mean number т-2 778 H" 2.95 Mean ash free Dry weight (g m2) Turrid prey spp. Non-prey spp. 35) lon! очьнбёзооа | — о UM SF = = 0) = 0) = All less than 0.1 — = = 6 4 9 = > = es 6 5 = = 8 = 2 4 4 6 = 10 9 = 5 3 4 3 3 1 2 lp 1 1 = = = 7 = 2 = = == = 3 4 8 6 5 7 = 7 10 = = 1 3 2 Y 9 5 6 8 2 = 10 = 9 8 = 7,8 9 = 910 10 1,056 1,264 723 1,556 1,204 2.56 2:66 2.79 SANS 22 8.8 3.1 2.5 0.8 46.0 Oo Bs a a a relatively sparse (Tables 3, 6). The most abundant species, such as Axiothella rubro- cincta and Tharyx multifilis (Moore, 1909), are present in many of the samples. Less com- mon but often very important species such as Myriochele oculata Zachs, 1922, the major turrid prey, are often poorly represented. The similarity between the infaunal samples from all of the habitats is high in summer, but ex- cept for the shell fragment habitat, there are many changes by winter. Coupled with this discrimination some species are associated with only one habitat, such as Drilonereis falcata Moore, 1911 in the lower mud, and Prionospio cirrifera Wirén, 1883, in the shell fragments. Biology of Ophiodermella Ophiodermella inermis is the only turrid found at PWN, although it was associated with two other turrids, Oenopota levidensis (Carpenter, 1864) and Kurtziella plumbea (Hinds, 1843) at WP. An estimate of O. inermis population size is available only from PWN, where the population was effectively contained by physical factors. The population apparently varied annually from a low in mid- winter to a high in late spring or early summer (Fig. 9). Few juveniles were collected during the sample period, only 5 of the 1125 animals collected for fecal sample examination were less than 15 mm long. Juveniles were proba- bly not missed during the sampling; turrids as small as 3.0 mm long were regularly collected in other sampling areas. At WP where smaller O. inermis animals were found, they looked and appeared to behave much like the adults, supporting the contention that few juveniles were present at PWN. The PWN population consists of a narrow shell length distribution with few individuals differing from the mean length by more than 5.0 mm (Fig. 10). Mean shell length did not vary significantly from season to season, and remained the same for at least 15 months. Recaptured O. inermis at PWN allowed estimation of growth rate. Ani- mals were assigned size classes on the basis of pregrowth sizes, and the mean growth rate, in um day-1, was calculated. Two growth trends are evident; smaller animals grow much faster, and individuals reach a determi- nate adult size, approximately 31 mm in length, which few individuals exceed. The de- crease in growth rate with increasing length is described by a power curve: growth rate (um day-1) = 5.14 x 1011 x (pregrowth length (mm))-7.751; r2 = 0.79. Exclusion of values of shrinkage and zero growth changes the curve only slightly: growth rate (um day-1) = 6.33 296 SHIMEK DJIFMAMJIJSIASONDJIF MAM J 4000 3000 2000 1000 F/ O 973 19 74 ISS FIG. 9. Population size estimates of the Port Washington Narrows population of Ophiodermella inermis. Bars are one standard error on either side of the estimated population size. The curve was fitted by a series of power curve equations. 40 | 1975 20 N u m N 60 1974 | Е r 40 (O2 20 730% 40 Length (mm) FIG. 10. Size frequency distribution of Ophio- dermella inermis collected at PWN. Data are for the three month period ending 31 January, of the indi- cated year. Lower N = 360, Upper N = 195. Arrows indicate the mean size. x 109 x (pregrowth length (mm))-6.260; r2 = 0.86. These curves are based upon a relative- ly narrow size range, the smallest recaptured animals at PWN had a pregrowth length of about 23 mm, and are probably not valid for smaller sizes. Nonetheless, they indicate the growth rate of small animals may be rapid. Examination of the WP population size fre- quency distribution shows several peaks (Fig. 11), and the histogram can be separated into a series of normal distributions with distinctly different means. If the PWN population is as- sumed to be a mature population with little recruitment, and the characteristics of that size frequency distribution are used as the basis for separating the larger “mature” com- ponent of the WP population, a series of arbi- trary distributions result (Fig. 11). If the peaks of these distributions can be assumed to rep- resent successive yearly recruitments, and if the animals at WP can be assumed to grow at the same rates as the nearby PWN popula- tion, both sets of data can be pooled to com- pute growth rates for a larger suite of animals. A logarithmic equation describing the change in growth rate related to size was calculated by fitting the above data to a formula of the general form: y = a + b In x, where y = growth rate (um day~'), x = pregrowth rate OPHIODERMELLA BIOLOGY 20 [О 207250 Length (mm) 40 FIG. 11. Size frequency distribution of Ophio- dermella inermis at Windy Point. All = Total num- ber collected, N = 254, 1975. A-D Hypothesized recruitment cohorts from previous years. A = 1974, М = 7;B = 1973, N = 95; С = 1972, N = 38; D = 1971 and earlier, N = 114. Arrows indicate means. (mm), and a and b are constants. The result- ing equation: y = 80.463 — 22.371 (In x), describes the change in growth rate well, r2 = 0.964, over the larger range of size classes, from 21 to 30 mm. If this curve is an accurate representation of the change in growth rate with increasing size, the time necessary to grow from 21.5 mm to 25.4 mm is 1.1 years, and from 25.5 mm to 27.7 mm is 0.9 years. These sizes are the mean sizes of cohorts in Fig. 11. More importantly, this more general equation can provide a better estimation of growth rates for smaller sizes than could be done with data from the PWN population alone. If this rela- tionship is valid down to the smallest size, and if a shell length of 1.5 mm is assumed as the size of the settling veliger, it will take 0.56 years to reach 10.0 mm in length, and 2.3 years to reach 21.5 mm. Consequently, growth is probably very rapid in young O. 297 inermis, and slows as they mature. No defini- tive measure of size at sexual maturity is available, but the smallest female O. inermis to spawn was 22.5 mm in total length. Many of the “mature” snails collected at PWN 500-700 days after marking showed essentially no change in length, growing less than 1.0 mm, and they were only slightly more likely to increase than to decrease in length. Decreases were caused by apical erosion, probably from mechanical abrasion, and aperture destruction by unsuccessful pre- datory attempts by crabs. Mechanical erosion of the shell was also caused by the use of the shell as burrow substrate by spionid poly- chaetes of the genus Polydora. In some cases, the shell was substantially eroded by the worms. In two cases, | collected snails whose shells were perforated by the worm burrows exposing the digestive gland. Al- though these snails survived the measuring and marking procedure apparently without harm, they were not recovered subsequently. The growth pattern exhibited by O. cancel- lata is not similar to that in O. inermis. Mark- recapture data were too sparse to indicate confidently growth rates or changes in growth rates. Many of the data indicated shrinkage, caused primarily by apical erosion. If apical erosion was occurring at a similar rate in the two species, the effect of increasing error would be greater on the smaller O. cancel- lata. There is a trend for growth rate in length to slow as the animals get larger, but it is not significant because of the large variances in- volved. Measurements of lip growth were un- successful with O. inermis, presumably as mechanical abrasion of the beach sand re- moved the marks. Lip growth measurements were obtained for a few O. cancellata, how- ever, and they show no correlation with the growth rate in shell length, and a negative correlation with increasing length (Fig. 12). Examination of the size frequency data from the FHL O. cancellata population indi- cates that the population size frequency dis- tribution changed dramatically over the course of the study (Fig. 13). The mean size of the population in summer, 1974, is signifi- cantly larger than the mean size in the sum- mer of 1975. The intervening winter popula- tion is distinctly bimodal, indicating both mor- tality of the larger individuals, and the recruit- ment into the population of the smaller ones. The rate of change in the mean size of the smaller size cohort over the year is about half as much as is necessary to grow from the mean of the small sized cohort to that of the 298 SHIMEK Length(mm) FIG. 12. Outer lip growth rate versus pregrowth length for field recaptured Ophiodermella cancel- lata. Calculated curve: у = 60.92 — 21.46 In x, г2 077 larger size cohort in only one year. If the two cohorts of winter 1974-1975 are taken to be successive recruitments they must be at least two years apart. If the larger sized cohort is assumed to be mature animals, and the small sized cohort is assumed to be recruited at least two years later, the summer, 1975, dis- tribution can be divided into its components (Fig. 13). This assumes a relatively constant adult mortality which seems to be borne out by the gradual decrease in the size of the modal group of larger individuals. My collec- tion of O. cancellata was size dependent; | seldom collected any animals smaller than 3.0 mm, as they were difficult to distinguish from the surrounding sediment particles. It is evident by the large group of smaller animals crossing that collection threshold in the sum- mer of 1975 that recruitment was good for that cohort, which probably represents individuals that settled the previous winter. Predation Large crabs, Cancer productus and C. gracilis are present at the PWN and WP study sites, and both will attempt to eat O. inermis (Table 7). Successful predation by C. gracilis upon adult O. inermis is unlikely, but C. productus can eat them. Both crabs are abundant at PWN and although no quantita- tive estimates of abundance could be ob- tained as they tended to remain below the water line, C. gracilis appeared more com- mon than C. productus. Ranging over the en- 32 71975 ra Al AS EZ 00 5 KO) ГЭ Length(mm) FIG. 13. Size frequency distributions of Ophioder- mella cancellata at FHL from animals collected in summer, 1974, N = 68; winter, 1974-75, N = 88; and summer, 1975, N = 177. Hypothesized recruit- ment cohorts are indicated; the 1970 and earlier cohorts are white, 1972 is black, and the 1973 cohort is cross-hatched. Means are indicated by arrows. OPHIODERMELLA BIOLOGY TABLE 7. Results of crab predation experiments. Crab Cancer Cancer Ophiodermella inermis gracilis productus Eaten 0 5 Not eaten Attacked 1 2 Not attacked 7: 5 G = 9.51, P< 0.005. Thus, C. productus is significantly more successful at attacking and eating O. inermis than is C. gracilis. tire beach at high tide, they have access to any O. inermis present, perhaps excluding buried individuals. There is no evidence that O. inermis is any more attractive to either crab than other gastropods present at PWN. None- theless, substantial predation pressure could be inferred due to several factors. First, O. inermis is relatively large, only Nucella lamel- losa (Gmelin, 1791) and Polinices lewisii (Gould, 1847) are larger gastropods on this beach. Second, it is abundant, having densi- ties up to 0.9 animals m2. Third, the fre- quency of repaired shell fractures indicates that attempted predation is common. Fourth, these frequencies of shell repair are amongst the highest recorded (Vermeij et al., 1980). Attempted crab predation as measured by TABLE 8. Repaired crab-fracture frequency. 299 healed shell fractures (Fig. 1) is significantly more common at PWN than at WP (Table 8). Cancer productus is less abundant at WP, restricted to areas where debris in the habitat creates suitable refuges. Near these areas crab predation may be important. Cancer gracilis is seasonally abundant (Table 9) but is probably not a predator on adult O. inermis, although it may be an important mortality factor for juveniles. An O. inermis shell that has a total length of 15 mm or a calculated age of about one year, has an apical length of 10 mm if the apical length is measured from the apex of the shell to the shoulder suture at the top of the body whorl. The number of healed fractures in the top 10 mm of apical length therefore corre- sponds to the number of healed fractures in animals up to 15 mm total length. The number of healed fractures is not significantly different for the two areas and is relatively low (Table 8), indicating few unsuccessful attacks on juveniles. The number of healed fractures on the remainder of the shell is significantly greater at PWN. It is likely that this rate of increased attempted predation on PWN adults is reflected in increased successful predation on small PWN individuals, which probably accounts for their rarity in the area. Predatory pressure on the O. cancellata population was difficult to assess. Large crabs are rare in the lower mud habitat where O. cancellata is most common, only two adult C. Mean number of repaired fractures Number Shell length Area Top 10 mm of spire Remaining spire examined range (mm) PWN 15515) && 4] 5172 6.97 + 2.54 31 15.0-30.6 WP 1:09 == 1.33 4.24 + 2.46 34 15.2-30.6 t-test on difference of means: t = 1.642; ns. t = 4.364: Р. =. 0.001. TABLE 9. Distribution of Cancer gracilis at Windy Point, 1975. Mean number of C. gracilis observed per 25 m2 Month J F M A M UB 0 1 = 1 2 US 2 1 = 0.5 14 LS 1 1 = 1 3 LB 0 0 = 0.5 2 Number of surveys 2 2 0 2 3 J J A S O N D 1 5 3 0 0 0 0 0.5 4.5 ss 0 2 0 1 1 5.5 0 3 1 0 0 1 2.5 0 1 0 0 0 2 3 2 1 1 1 1 300 SHIMEK magister Dana, 1852 were seen in over 150 hr of subtidal observation, and no adult C. productus were ever seen in the habitat, al- though both were relatively common less than 200 m away. Juvenile C. productus are seen in the area occasionally, and presumably they could be attacking the snails. Six of the 60 O. cancellata recaptured after marking showed evidence of recent shell fracturing, and older fractures were not uncommon (Fig. 1). Large hermit crabs, Pagurus ochotensis Brandt, 1851, and P. armatus (Dana, 1851), are com- mon in the habitat, and might attack the snails. Thirteen of the 48 O. cancellata shells collected with hermit crabs in them had been drilled by naticids, and both Natica clausa Broderip & Sowerby, 1829, and Polinices pallidus Broderip & Sowerby, 1829, are found in the area, but most shells showed no evi- dence of the cause of death. Fish predation may be a factor; the ratfish, Hydrolagus colliei (Lay & Bennett, 1839) is sometimes found in the area and is known to prey on gastropods (Miller et al., 1978), but none could be col- lected for gut analysis. A large asteroid, Luidia foliolata Grube, 1866, occurs sporad- ically in the lower mud and ingests small gas- tropods and infauna. Five were collected and their gut contents examined: no turrid remains could be identified, although other small gas- tropods characteristic of the habitat were rep- resented. It is also possible that my manipula- tion was resulting in mortality of O. cancellata. Of the 48 O. cancellata shells that were col- lected with hermit crabs in them, only two were previously marked, however, indicating that this factor is probably not major. Environmental stress effects have little in- fluence at either of the subtidal sites; how- ever, some physical factors can have a pro- nounced effect at PWN. Temperature effects are especially important. During the first sum- mer low tide with uninterrupted sunlight in 1974 (20-VI-74), 42 of 111 snails collected were dead or moribund. Following this initial sunny low tide, few animals were collected exposed on the beach during the remaining summer low tides. Subsequent observations determined the temperatures on the beach (Figs. 7, 14). Shell temperatures closely ap- proximated substrate temperatures, and al- though tests were not done to determine tem- perature tolerances, animals collected with shell temperatures in excess of 30°C or from substrates of similar temperatures were gen- erally dead or dying. No periods of freezing were observed during the three winters (1973-74, 1974-75, 1975-76) that PWN was visited. | was present on the beach during all winter tides when the low tide exposed snails, consequently, cold stress is probably less im- portant than heat stress. Many O. inermis at PWN were observed and subsequently collected in areas of fresh water runoff. These animals appeared to suf- fer no ill effects upon their return to a normal saline environment. Due to beach topogra- phy, the maximum time for fresh water im- mersion was about three hr. No apparent ef- fects were observed upon snails left in fresh water in the laboratory at ambient tempera- tures (5°C, winter; 20°C, summer) for up to three hr. Habitats Ophiodermella inermis ranges completely over all habitats at PWN, and although there are some significant seasonal differences in distribution (Table 10), it is unlikely that these changes are due to tracking of the prey popu- lation. Although there is a slight positive corre- lation between prey and turrid distribution in the winter, there is none in the summer. The Causative agent for the shift in distribution re- mains unclear, but it may be an artifact of the summer behavioral change. Ophiodermella inermis was the most abun- dant turrid at WP, although Oenopota levidensis and Kurtziella plumbea are also present. Of the 254 O. inermis whose posi- tions were determined, 71 (28%) were seen on the upper bench, 65 (26%) on the upper slope, 57 (22%) on the lower slope, and 61 (24%) on the lower bench. Each habitat ac- counted for 25% of the surveyed area. Thus, there is no defined use of one habitat. The turrid assemblage at FHL is large and diverse. In addition to O. cancellata, Kurtziella plumbea, Oenopota elegans (Moller, 1842), O. excurvata (Carpenter, 1864), O. fidicula (Gould, 1849), O. levidensis, O. turricula (Montagu, 1803), O. pyramidalis (Strom, 1788), and Clathromangelia _ interfossa (Carpenter, 1864) were also found. Transect studies indicate that O. cancellata does not use all of the habitats equally. Within 1 m of the FHL transects are 144 m2 of the lower mud or silt habitat, 138 m2 of the shell frag- ment habitat, 20 m2 of the rock habitat, and 48 m2 of the upper mud habitat. During the period of the study, these habitats represent- ed 36%, 34%, 5%, and 25%, respectively, of the area surveyed. Of the O. cancellata found OPHIODERMELLA BIOLOGY 301 within 1 m of the transects: 277 (68%) were found in the lower mud, 112 (28%) were found in the shell fragments, 11 (3%) were found in the rocks, and 7 (2%) were found in the upper mud. The distribution of O. cancel- lata is significantly different from the distribu- tion of the habitats (G = 253.6; P < 0.005), and shows a distinct bias for the lower mud habitat. A similar result is apparent in the mi- nor sites where O. cancellata is found (Table 11), all of which are similar to the lower mud. During the substrate preference tests O. inermis showed a slight, but insignificant, ten- dency to avoid areas of shell. The results of | FOURS 2 в 4 Exposed FIG. 14. Exposed shell, substrate, and air temperatures, of Ophiodermella inermis at PWN. Full mid-day sun, no wind, 18 July, 1974. Standard deviations shown for shell only, similar variations were seen for all values. All points are means of five values. e = substrate. A = shell. © = air. TABLE 10. Ophiodermella inermis distribution differences at PWN. Three level nested analysis of variance comparing habitats (sand, cobble); heights within habitats (middle, low); and seasons within heights (sum- mer, winter). Degrees of Source of variation freedom Among habitats 1 Among heights within habitat 2 Among seasons within heights 4 Within seasons, Error 92 Total 99 Sum of Mean squares squares E 23.040 23.040 1.426 n.s. 32.320 16.160 0.168* n.s. 218.740 54.685 4.976 PESOL0015 1012.860 11.009 1286.960 *Calculated with a recomputed MS (MS' = 96.026) due to mixed model ANOVA (Sokal & Rohlf, 1969). "F0.001(4,120) = 4-95. 302 SHIMEK TABLE 11. Minor study sites with Ophiodermella cancellata. PE Area Position Depth (m) Bottom type Abundance Je eee E a zz eee Iceberg Point, Lopez ld. 48° 24’ 54" N, 122° 53' 24” W 73- 91 Silt, shell fragments + Lopez Sound 48° 28’ 48" N, 122° 50’ 06” W 49 Silt + Smallpox Bay, San Juan Id. 48° 32’ 26” N, 123° 09’ 47" W 18 Silt + Parks Bay, Shaw ld. 48° 33’ 45" N, 122° 59’ 20” W 9 Silt +++ Upright Channel 48° 34’ 18" N, 122° 52' 30” W 36- 55 Silt, shell fragments Ar Potato Patch 48° 35’ 00" N, 122° 50’ 42"W 42-55 Silt, shell fragments + North Shore, Shaw Id. 48° 35’ 29" N, 122° 57' 28" W 18- 40 Silt ++ McConnell and Reef Is. 48° 36’ 00" N, 123° 01 04” \/ 15-22 Silt, shell fragments + Jones and Yellow Is. 48° 36’ 00” N, 122° 02’ 00” W 22-115 Silt, shell fragments + North Pass 48° 36’ 36" N, 123° 00' 42"W 18- 29 Silt + Deer Harbor, Orcas Id. 48° 36’ 54" N, 123°00'09"W 15-22 Silt + Cowlitz Bay, Waldron Id. 48° 41' 30” N, 123° 03’ 18” W 9- 55 Silt + + = present; ++ = common; +++ = abundant TABLE 12. Results of substrate choice experi- ments. —— ne Substrate Turrid Sand Shells p* Ophiodermella inermis 13 9 0.523 O. cancellata 8 11 0.648 III IE *Two tailed cumulative binomial probability of a deviation as large or larger if P(sand) = P(shells). the tests with O. cancellata show a slight pref- erence for shell fragments over the sand (Table 12). These results are probably more useful for O. inermis than O. cancellata, as the sand versus shell choice is more char- acteristic of the substrate choices at the PWN and WP sites, than at FHL. However the par- ticle size distribution in the upper mud ap- proximates the particle distribution in the sand area of the choice chamber, and the shell por- tion of the choice chamber is representative of some portions of the lower bench habitat at WP and the shell fragment habitat at FHL. Diet During the study, 1125 O. inermis were col- lected from PWN and their feces, if any, were microscopically examined (Table 13). Of these, 369 had feces containing polychaete remains; 363 of these were identifiable, and 353 of these were from Owenia fusiformis. This polychaete has only two types of setae, long capillary notosetae, and minute biden- tate hooked uncini (Fig. 3). There may be several hundred thousand of these character- istic uncini per worm (Macintosh, 1915). As the notosetae of O. fusiformis are not abso- lutely distinctive, a fecal sample was recorded as identified only if both uncini and notosetae were present. Owenia was fed to eight Ophiodermella inermis and the mean time from ingestion to defecation at 8-10°C was determined as 1.3 + 0.5 days. Using this in- formation, the fecal sample data, and the esti- mate of population size, an estimate of the mean yearly predation upon O. fusiformis by Ophiodermella inermis can be calculated (Table 14). As at PWN, O. inermis at WP is primarily eating Owenia fusiformis (Table 13). The overlap between the diets from the two sites is 0.994. Examination of worm uncini from the feces of snails collected at the two sites, and calculation of worm collar diameter indicates that the snails from both sites are eating worms in the same size range (Table 13). Be- cause of the small sample size for certain months at WP, assessment of the impact of the snails on the polychaete population is not possible. Nonetheless, O. inermis appears to be feeding at a relatively constant rate similar to the rate at PWN. Twelve O. inermis animals were collected from a sandy beach on the south side of Alki Point, Seattle, Washington, on 10 July, 1975. Two defecated and had eaten O. fusiformis. Two individual O. inermis were collected in the San Juan Islands, both subtidally from sandy substrates superficially similar to the WP area. Neither produced any identifiable fecal remains. Ophiodermella cancellata is also a vermi- vore, but it feeds less often than O. inermis (Table 15). Although the diversity of diet is 303 OPHIODERMELLA BIOLOGY ‘AUO sway! Aaıd paynuap! шо paje¡naje), в Бъъъъннтн ‘SOZIS илом UI ээцалаир зиеэшиб!$ ON 9b VEC + LS ESB 2 dM 18'80p + 96'576 д NMd — (шт) ezis 1209 илом иеэш рэзепо!еэ paulwexe sajdwes ¡esaj jo 1aquiny Baly 'USJea SIUNOJISN, BIUSMO JO JejaWweıp ле|оэ иеэш pajenojeo ‘а 766`0 = а :seidwes 1298} шо} Чешэло Aejaip Je]iqeH G'9€ 0'0 8°8S g'8c 00€ DIE ZOE 1/95 > 6559] бицеэ % v | i | L рэуциерип с | | (1981 ‘ерлешц9$) 5/е/20$ e10pAJOy 58 Ol с G 6 pl GE Le | tr8L ‘aleiyD энер sywwoyisny виэмо ГЕО Selsads 4914 vbc C ZE 0 de 0 Oc GE €S GE о 69 9 PeuILEx8 Spin] JO JOQUNN dM [55 c6E 661 0'6p 05 11: 682 982 16S bzZE — €6¢ 605 бицеэ % 9 | | | | 2 рацциарий 8 с | с Z | (1981 ‘ерлешцо5) $//е!20$ “y | | 5881 ‘Igoyer ejeqojespenb “y | L ÿ£6L ‘EAOHUSUUY в/оэши 210pAJO4 595 Lv 29 LC | 0 81 9c 05 Le ly GG prsl ‘elelyo ejjep siunoyısn, eluamo gro Saloads Adıd Soll Oc! LEE 6r v 6 92 86 88 89 0 #9 cb! pauluexa Spuun} JO JOqUINAY NMd a A A E E E E E жи [2101 a N O 5 У Г Г N У W = р "Чиош Jad sajdwes |еэа; paynuap] “y жвБзбжь ь ь + + у-у ьъууъннзтнньн ———— ‘SIWBUI е/эшлароцао—иоцешлозии Mejeig ‘EL 318V1 304 SHIMEK TABLE 14. Estimate of the effect of Ophiodermella inermis predation. A. Owenia fusiformis population Beach Mean number Total estimated mean area (m2) Owenia (m2) population of Owenia, 1974-1975 Sand 2.3 x 103 8.1 x 102 1.9 x 106 Cobble 3.3 x 103 3.4 x 102 0.4 x 106 Total 3.6 x 103 2.3 x 106 B. Ophiodermella inermis population Mean yearly fraction eating Owenia: 0.314 Mean digestion time per Owenia: 1.3 + 0.5 days, at 8-10°C. Date Estimated Ophiodermella population size May, 1974 4193 May, 1975 2023 C. Predation effects Estimated number of Owenia eaten if digeston time is: 1 day 2 days Per day, May, 1974: 1318 659 Per day, May, 1975 635 318 From May, 1974 to May, 1975: 3.6 x 109 1.8 x 105 Fraction of total Owenia population: 0.16 0.08 Pan en TG GT Tr Tr ee ne ПиНчниЕ Ех ети ЕЕ заикание somewhat greater than in O. inermis, O. cancellata is also a dietary specialist, and also primarily eats an oweniid polychaete, Myriochele oculata. The uncini of M. oculata are also quite distinctive which facilitated fecal analysis (Fig. 3). Far fewer O. cancellata are feeding per unit time compared to O. inermis, and relatively more prey items are unidentifi- able, perhaps reflecting the increased dietary diversity. Chemo- and rheoreception Ophiodermella inermis from PWN were tested in a choice chamber to determine the extent of distance chemoreception and/or current effects (Table 16). The results are un- ambiguous; current is needed for any re- sponse; in the absence of current the animals do not move. The response is greater, how- ever, if both bait and current are present. They move up current, do not make any choices that lead them to the baited arm, and they respond randomly in the choice cham- ber. Ophiodermella inermis reproduction Collection and confinement for fecal sam- ple examination act as stimuli to egg capsule deposition. The peak of capsule deposition was during February, although deposition oc- curred from October through July (Table 17). After fecal sample examination and marking, the animals were released into an aquarium for up to two weeks prior to their return to their own areas, and many deposited egg cap- sules. The capsules were identical to those collected in the field. Since the behavior in the holding tank corresponded to the observed natural behavior, the deposition of egg cap- sules probably corresponded with natural de- position. Encapsular development averaged about seven weeks. Trochophore and early veliger stages are passed in the capsule, growing and becoming very mobile within the capsule. There are no nurse eggs, and the number of veligers hatched appeared to be the same as the original number of eggs in the capsule. After hatching, the veligers are very mobile 305 OPHIODERMELLA BIOLOGY ‘Ajuo sway! Aaıd paynuapi шо peyeingjeo, CL АБ ЛИ 69 ве дс 109 Sif ren ES ois бицеэ % 6 с | | | с | | ajaeyoAjod payuepiuf С | | payyuapiun ‘pılÄS | | (L98L ‘ерлешц9$) sye120s e10p/Ajo4 LE с с с 9 | e 5 S 9 С € 226, ‘SU2EZ EJE/N20 ajeyooÁyy L£'O salads Adıd 289 x pl ve EP м 05 zer 56 № 26 os рэишиеха spin] jo Jaquiny — „H №01 а М O S У Г Г W У W 3 ll SSS 'yyuow Jod sajdwes 1898 рецпиер! ‘взе/езиео ejjeuwepoiydg—uoneuojul Lejeig ‘SL AIGVL 306 SHIMEK TABLE 16. Choice experiments. A. Chemotaxis. Conditions No choice Chosen arm baited Unbaited Total run i. No bait—equal water flow in both arms* 78 10 12 100 ii. Bait—equal water flow in both arms 30 38 За 100 iii. Bait—no water flow in either arm 79 12 9 100 Test Choice No choice pre Ai. 22 77 7.95 x 10-9 АЙ. 70 30 3.93 x 10-5 АЙ. 21 79 2.62 x 10-9 B. Rheotaxis. Conditions No choice Chosen arm flow No flow Total run i. No bait—water flow in one arm 75 22 3 100 ii. No bait—no water flow in either arm’ 95 4 1 100 Test Choice No choice p** Bi. 25 72 282% 107 Ви. 5 95 6.26 x 10-23 С. Probabilities of choices made. Potential choices Test A B Ai. Control 10 12 Aii. 38 32 Ali. 12 9 Bi. 22 3 Bii. 1 4 2-tailed cumulative binomial probability lf Pray = Pia, = 0.5 0.832 Non-significant 0.664 Non-significant 0.550 Non-significant 157% 105 Highly significant 0.776 Non-significant * = Choices are: right vs. left arms. ° = 1-tailed cumulative binomial probability if P(choice) = P(no choice) and feed actively for up to five weeks. They live for another two weeks in a relatively inac- tive condition (Table 18). This inactive period probably corresponds to the time when the animals would normally settle. | was not able to induce metamorphosis. A variety of settle- ment substrates were offered: sand, silt, empty O. fusiformis tubes, Phyllochaetop- terus tubes, and bare glass, but no settlement occurred. Spawning and development of O. cancel- lata were not observed although copulation was seen several times in the field from July through September. The smallest O. cancel- lata collected was 1.8 mm long and was col- lected on 20-11-75. Most (53 of 85, 64.4%) small (3.0 to 5.5 mm) O. cancellata were col- lected from April through June, suggesting a settlement time of late autumn to mid-winter. DISCUSSION Both species of Ophiodermella are preda- tory specialists upon oweniid polychaetes. The degree of specialization in these species is uncommon, yet there does not appear to be a close correspondence between turrid distri- bution with any polychaete species on the fine scale in any of the habitats examined. For ex- ample, at WP O. inermis was common in all habitats, while Owenia was concentrated in OPHIODERMELLA BIOLOGY 307 TABLE 17. Ophiodermella inermis reproductive information. A. Field observations Area: Date: Number seen: I. Copulation PWN 11-X1-73 “several” PWN 25-X1-73 3 PWN 9-X11-74 1 PWN 19-X1-74 1 WP 24-V-73 1 |. Capsule deposition PWN 26-V-74 3 PWN 13-V-76 1 B. Laboratory observations |. Capsule deposition Month Number deposited Number of O. inermis examined October 2 49 November 0 354 December 0 122 January 1 118 February 20 193 March 4 40 April 4 103 May 6 141 June 1 133 July 1 96 TOTAL 39 1253 TABLE 18. Ophiodermella inermis egg capsule characteristics. Number of capsules examined Mean dimensions (mm) length 4.68 + 0.99 width 3.95 + 0.77 height 1.44 + 0.29 Mean egg number diameter (um) 208 + 61 222 + 15 Mean number of days encapsular period: 50.6 + 13.05 Maximum number of days of post-hatching survival: 50 the upper bench area, and was only sporad- ically found elsewhere. Ophiodermella inermis is very mobile; during one 30 min pe- riod, | observed one marked individual to move 4 m along a transect. Consequently, the snails could easily vary their habitat and for- age in the upper areas from the lower ones. Although the chemo- and rheotaxis experi- ments indicate scant powers of distance chemoreception (Table 16), the tendency of the animals to move up current if prey are present presumably aids them in locating prey. Ophiodermella inermis clearly does not select habitats due to sediment particle size except in a gross manner; it is seldom found on large (diameter > 50 cm) boulders, and was never seen in bedrock areas. It does re- spond to degree of exposure: after encounter- ing the high temperatures typical of the first summer low tides, the behavior patterns of the animals change drastically. Prior to this 308 SHIMEK exposure, most of the population is uncov- ered at low tide. After the first few exposed tidal periods, the upper limit of the population apparently drops from +0.5 m to +0.2 m, and the majority of the population remain buried through the low tide period, emerging shortly before the incoming tide passes over them. During the late winter and early spring, the animals do not bury and are left stranded around the beach for the duration of the low tide. Collection of the animals is facilitated in late winter and spring when the animals are exposed and visible. In many cases, they ap- peared to be actively foraging while exposed. This seasonality of behavior results in the ap- parent cycling on the population (Fig. 9). This behavior pattern is triggered by exposure on sunny warm days, and presumably is main- tained by repeated periods of high insolation. It is obviously a protective behavior pattern, and although some animals perish initially, relatively few are exposed to the worst of the summer heat. If habitat and prey utilization patterns can be attributed to active preferences, the rele- vant cues can be hypothesized on the basis of comparing resources available and utilized. If this is so, O. cancellata, by virtue of its almost absolute specialization on Myriochele oculata, one of the rarest polychaetes sam- pled (2 out of 11,667 worms sampled), must be responding primarily to the distribution of its principal prey, and secondarily to any sub- strate parameters. This hypothesis is sup- ported by the lack of any substrate preference in experimental conditions. Unfortunately the distribution of M. oculata is unknown here, hence no correlation with its predator's distri- butions can be made. While it cannot be con- clusively shown that prey with no identifiable remains were not being eaten by O. cancel- lata, the extreme dietary specialization other- wise seen in both O. inermis and O. cancel- lata argues against this. On the basis of the O. cancellata distribution being strongly skewed to the lower mud, it can be assumed that M. oculata is most common there. Although dis- tance chemoreception was not tested in O. cancellata, the rarity of the prey suggests that this species may have better capabilities in this regard than O. inermis. These two turrids may react differently to prey population densities. The extreme spe- cialization on abundant prey such as seen at PWN and WP is consistent with optimal forag- ing theory (Emlen, 1966, 1968; MacArthur & Pianka, 1966; Schoener, 1971; Hairston, 1973) which predicts that where food is not limiting, specialization should occur. The PWN site contains several years’ supply of food, and presumably no food limitation exists at WP either, because of the abundance of prey in the upper bench and slope areas. Predatory effects are difficult to substantiate without experimental manipulation. Nonethe- less, some of the annual drop in the WP upper slope O. fusiformis populations is likely the result of the turrid. This is certainly unlike the situation at FHL, where O. cancellata preys upon the rare Myriochele oculata. The turrid is quite mobile, and | have starved some for up to two months without any apparent ill effects. Both of these properties would be expected in predators whose prey is rare, although many carnivo- rous gastropods survive prolonged starvation regardless of prey density. Furthermore, the low fraction of turrids feeding could be indica- tive of the low prey density and long search time. That the prey density is low there is no doubt; the cause of this low density is the per- tinent question. Is O. cancellata such an efficient predator that it can effectively keep its prey at a very low density? If so, is the turrid food limited? Or is O. cancellata simply adept at finding low density prey? The turrid population recruits well, but apparently not every year (Fig. 13). Since Owenia fusiformis is an annual (Fig. 8), it is reasonable to suggest that M. oculata is as well. If this is the case, perhaps the rela- tionship between O. cancellata and its prey is a cyclic one where the predator regularly ex- terminates the prey and then itself goes ex- tinct locally. This might result in the pattern seen in the O. cancellata size frequency dis- tribution, and account for the missing year class. Perhaps veligers respond to the pres- ence of the prey as a settlement cue, and during a low period in the population of M. oculata they do not settle. If the predators can starve for long periods, especially as pre- reproductive juveniles, when the energy re- quirements for gamete production would be non-existent, they might be able to survive from annual pulse to annual pulse of their relatively transitory prey. As the prey popula- tion would respond to predator population cycles, but out of phase, this cyclic interaction would be quite stable, and could result in periods of very low population of both the predator and the prey. The smaller individuals could devote more time and energy to seek- ing prey and might be substantially more suc- OPHIODERMELLA BIOLOGY 309 cessful than the reproductive individuals at finding sufficient food to maintain growth. If the preproductive effort involved substantial energy utilization for mate seeking, gamete production, and egg capsule production, the reproductive success would be expected to decline, perhaps as precipitously as we see here, especially in a food limited situation. The infaunal polychaete diversity is rela- tively high at all three major sites, and the mean polychaete size is small. The structure of the soft-sediment community may be large- ly dependent upon the activities of these or- ganisms (Sanders, 1958, 1960; Woodin, 1974, 1976; Gray, 1974; Rhoads & Young, 1970). Turrid predation may substantially af- fect this assemblage of worms, especially if the turrids can render their prey rare, or if they can drastically affect the population of a dominant organism. Owenia fusiformis, a dominant component of the infauna at PWN and the upper WP areas, can certainly be af- fected by Ophiodermella inermis, at least on the small scale. These hitherto ignored preda- tors of an essentially unknown trophic level (Trevallion et al.,1970) may be having sub- stantial effects on the community as a whole. The effects of predation upon both of these turrids is important and difficult to assess. Brachyuran predation is significant at PWN and at least on small individuals at WP. Ma- ture O. inermis, greater than 27 mm long, probably have escaped predation by Cancer gracilis, although C. productus can certainly prey upon them. The relative rarity of C. productus at WP probably indicates that the large O. inermis there have a size refuge from predation that is lacking at PWN. Indeed, the relative abundance of small O. inermis at WP may be due largely to the lack of efficient pre- dation from the larger crab. The highly com- pressed unimodal size frequency distribution of O. inermis at PWN appears largely due to predation by crabs upon new recruits and other small individuals. The population struc- ture reflects a large successful settlement and recruitment swamping the predators, coupled with a rapid early growth rate. Indeed, the ap- parently rapid early growth rate of O. inermis would be strongly selected for by predation on smaller animals. That the larger animals have at least a partial refuge from predation is evi- dent from the healed and repaired fractures of the kind caused by attempted crab predation. This model of crab predator assumes the smaller C. gracilis to be the most abundant predator, but still leaves the population at the mercy of C. productus. The burying behavior of intertidal O. inermis in the summer, and after prey ingestion, presumably also aids in escaping predation. Because of the ability of C. productus to eat turrids of any size, as well as the periodic early summer exposure mor- tality, the population at PWN depends upon massive periodic turrid recruitment or mas- sive crab mortality to maintain itself. If no re- cruitment occurs for 5 or more years, the pop- ulation is likely to go extinct. Subsequent to the majority of this study, the PWN was sur- veyed twice in the autumn of 1980, and a total of two Ophiodermella inermis were found for ten man-hours of search. In a comparable period of 1973, over 400 animals were col- lected, suggesting that the population is now effectively extinct. The WP population, on the other hand, should be able to survive, due to the largely size selective nature of the predation, the lack of environmental stress effects, and the rarity of the larger predator. Surveys of this area in December, 1980, and January, 1981, showed an apparently normal population. Predation effects upon O. cancellata at FHL are more difficult to determine. The snail population is obviously subject to some adult mortality factor; very few animals are 12.0 mm long or longer, and the larger ani- mals consistently disappear from the popula- tion. Naticid predation is certainly a major fac- tor. Because of the small size of the species, and the evidence of shell breakage in about 10% of the recaptured animals, crab preda- tion is also a factor, although its magnitude is unclear. The fact that most of the empty shells recovered had no evidence of the cause of death made it impossible to unambiguously determine whether malnutrition induced by reproductive stress, asteroid predation, dis- ease, or a determinate life span produced the observed distribution of dead snails. The long larval stage of O. inermis, up to four months, half spent in an egg capsule, and half as a planktonic veliger, may be related to the extreme dietary specialization. The long larval stage may allow sufficient development of the specialized toxoglossan radula to per- mit the capture of Owenia fusiformis by the recently metamorphosed juveniles. If this is the case, settlement of the juveniles should coincide closely with the settlement of juvenile prey polychaetes, presuming that small snails must eat small worms. For O. inermis this ap- pers to be the case, most of the turrids ap- pearing to be ready to settle no earlier than 310 SHIMEK late spring and no later than mid-summer. This is the period when Owenia fusiformis shows a substantial increase in density, pre- sumably from settling juveniles. Recruitment of O. inermis may be very patchy. The uni- modality of the PWN site frequency data (Fig. 10) suggests one, or at most two closely spaced recruitments several years prior to 1973, and none since. These turrids provide examples of dietary specialization unparalleled in temperate inter- tidal gastropods. Furthermore, the specializa- tion upon tubicolous polychaetes of only one family, the Oweniidae, and the basic similari- ties of the radular teeth of these two turrids (Fig. 1), and other members of the subfamily Borsoniinae, invite investigation of the dietary requirements of other members of the sub- family, and functional analysis of this toxo- glossan radular tooth. Because of the special- ization of the turrids upon only one prey spe- cies per predator, very fine partitioning of the dietary resource base is possible, and coupled with the diversity of potential prey polychaetes in temperate and boreal uncon- solidated sediment ecosystems, this may be responsible for the phenomenal adaptive radiation seen in many genera of boreal tur- rids. ACKNOWLEDGMENTS Portions of this paper are taken from a dis- sertation submitted in partial fulfillment of the requirements for a Ph.D. in the Department of Zoology at the University of Washington. | thank the members of my graduate commit- tee, Dr. K. Chew, Dr. R. Paine, and especially Dr. Alan Kohn, Dr. Eugene Kozloff, and Dr. Paul lllg for their many helpful suggestions. Much of the field work for this was done with SCUBA, and while | cannot thank all my div- ing partners individually, Paul Raymore, Carl Nyblade, Larry Moulton, Steve Bloom, Ken Sebens, Kathy DeRiemer, and Ed De Martini were especially helpful. Over 75 other people were diving partners during this study and without them all it could not have been done. Hal Scheidt of Bremerton, Washington, im- measurably helped me throughout the study, and assisted in locating the Dyes Inlet study sites. Mr. V. B. Carter of Bremerton, Washing- ton, graciously allowed the use of the PWN area. Elsie Marshall of Seattle, Washington, es- pecially, and other members of the Pacific Northwest Shell Club, provided much infor- mation on the distribution and variety of tur- rids in this region. Much of this information consisted of valuable specimen shells with collection data that were freely and enthusi- astically given to me to use. These data, while not utilized directly in this study, helped guide me to study sites and dredging areas and have helped form my ideas of turrid distribu- tion. | am indebted to Dr. Robert Fernald and Dr. Dennis Willows, the past and present direc- tors, and to Dr. Richard Strathmann and Dr. Eugene Kozloff, acting directors of the Uni- versity of Washington Friday Harbor Labo- ratories for allowing the use of laboratory facilities. | thank Dr. J. Nybakken, Dr. J. H. McLean, V. O. Maes, and an anonymous reviewer, for making many helpful suggestions upon re- viewing earlier drafts of this paper. A special debt is owed to Dr. Alan Kohn, who provided the use of his laboratory, equip- ment, and expertise in toxoglossan biology and who, through a course at F.H.L. provided my initial opportunity to examine and wonder about turrids. Many of my companions at the University of Washington discussed various aspects of tur- rid biology and made many helpful sugges- tions; especially notable among this group were the late Paul Leviten, Carl Nyblade, Larry Moulton, Paul Raymore, and Steve Bloom. This work was partially supported by an N.S.F. Grant 75-03303 to Dr. Alan Kohn, by an N.S.F. doctoral dissertation grant GA-41814, by grants from the Friday Harbor Laboratories, by two scholarships from the Pacific Northwest Shell Club and by support from the Zoology Department of the Univer- sity of Washington. | dedicate this paper to Roxie Lynn Shimek, for without her constant encouragement and support | could not have completed it. LITERATURE CITED BANSE, K., 1972, On some species of Phyllodoci- dae, Syllidae, Nephtyidae, Gonididae, Apisto- branchidae, and Spionidae (Polychaeta) from the Northeast Pacific Ocean. Pacific Science, 26: 191-222. BANSE, K. & HOBSON, K. D., 1974, Benthic er- riantiate polychaetes of British Columbia and Washington. Bulletin of the Fisheries Research Board of Canada, 185: 1-111. BERKELEY, E. & BERKELEY, C., 1952, Annelida OPHIODERMELLA BIOLOGY 311 Polychaeta, Sedentaria. Canadian Pacific Fauna, No. 9b (2): 1-139. BLAKE, J. A., 1966, On Boccardia hamata (Web- ster), new combination (Polychaeta, Spionidae). Bulletin of the Southern California Academy of Science, 65: 176-184. BLAKE, J. A., 1971, Revision of the genus Poly- dora from the east coast of North America (Poly- chaeta: Spionidae). Smithsonian Contributions in Zoology, 75: 1-32. BLAKE, J. A. & WOODWICK, K. H., 1971, A review of the genus Boccardia Carrazzi (Polychaeta: Spionidae) with description of two new species. Bulletin of the Southern California Academy of Science, 70: 31-42. BLISS, С. 1., 1967, Statistics in Biology. McGraw- Hill, New York, 1: 558 p. CERNOHORSKY, W. O., 1975, Ophiodermella inermis (Hinds, 1843) (Gastropoda: Turridae). Veliger, 17: 318. DALES, R. P., 1957, The feeding mechanism and structure of the gut of Owenia fusiformis delle Chiaje. Journal of the Marine Biological Asso- ciation of the United Kingdom, 36: 81-89. EMERY, K. O., 1938, Rapid method for analysis of sands. Journal of Sedimentary Petrology, 8: 105-111. EMLEN, J. M., 1966, The role of time and energy in food preference. American Naturalist, 100: 611- 617. EMLEN, J. M., 1968, Optimal choice in animals. American Naturalist, 102: 385-389. FOSTER, N. M., 1971, Spionidae (Polychaeta) of the Gulf of Mexico and the Caribbean Sea. Studies on the Fauna of Curacao and Other Caribbean Islands, 36: 1-183. FRANC, A., 1952. Notes écologiques et anatomi- ques sur Philbertia purpurea (Montagu) (Moll. Ctenobr.). Bulletin du Museum National d’His- toire Naturelle (France), (2)24: 302-305. GRANT, U. S., IV & GALE, H. 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: 1- 1036. GRAY, J. S., 1974, Animal-sediment relationships. Oceanography and Marine Biology Annual Re- views, 12: 223-261. HAIRSTON, N. G., 1973, Ecology, systematics, and speciation. Breviora; Museum of Compara- tive Zoology (Cambridge), 414: 1-21. HARTMAN, O., 1955, Quantitative study of the benthos of the San Pedro Basin, southern Cali- fornia. Allan Hancock Pacific Expedition, 19: 1- 186. HARTMAN, O., 1969, Atlas of sedentariate poly- chaetous annelids from California. Allan Han- cock Foundaton, 812 p. HOBSON, K. D. 8 BANSE, K., 1981. Sedentariate and archiannelid polychaetes of British Columbia and Washington. Canadian Bulletin of Fisheries and Aquatic Sciences, 209: 144 p. HOLLANDER, M. 8 WOLFE, D. A., 1973, Non- Parametric Statistical Methods. Wiley, New York, 503 p. HOLME, N. A. & MCINTYRE, A. D., 1971, Methods for the study of marine benthos. Blackwell, Ox- ford and Edinburgh, 334 p. INMAN, D., 1952, Measures for describing the size distribution of sediments. Journal of Sedimen- tary Petrology, 22: 125-145. JONES, N. S., 1950, Marine bottom communities. Biological Reviews, 25: 283-313. KOHN, A. J., 1959, The ecology of Conus in Hawaii. Ecological Monographs, 29: 47-90. KOHN, A. J., 1967, Environmental complexity and species diversity in the gastropod genus Conus on Indo-West Pacific reef platforms. American Naturalist, 101: 251-259. KOHN, A. J., 1968, Microhabitats, abundance and food of Conus on atoll reefs in the Maldive and Chagos Islands. Ecology, 49: 1046-1062. 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. LEVITEN, P. J., 1976, The foraging strategy of vermivorous conid gastropods. Ecological Monographs, 46: 157-178. LEVITEN, P. J., 1978, Resource partitioning by predatory gastropods of the genus Conus on subtidal Indo-Pacific coral reefs: the significance of prey size. Ecology, 59: 614-631. MACARTHUR, R. H. 8 PIANKA, E. R., 1966, On optimal use of a patchy environment. American Naturalist, 100: 603-609. MACINTOSH, W. C., 1915, The British marine annelids, 3(1) Polychaeta, Opheliidae to Ammocharidae. Dulau, London, p. 353-368. MILLER, B. S., SIMENSTAD, C. A., MOULTON, Е. L., FRESH, К. L., FUNK, К. C., KARP, W. A. 8 BORTON, S. F., 1978, Final Report of The Puget Sound Baseline Project—Nearshore Fish Sur- vey. Fisheries Research Institute, Univeristy of Washington, Seattle, Wa., 220 p. NYBAKKEN, J. W., 1978, Population characteris- tics and food resource utilization of Conus in the Galapagos Islands. Pacific Science, 32: 271- 280. PARKER, R. H., 1964, Zoogeography and ecology of some macro-invertebrates, particularly mol- lusks in the Gulf of California and the continental slope off Mexico. Videnskabelige Meddelelser fra Dansk Naturhistorisk Forening i Kóbenhavn, 126: 1-178. PEARCE, J. B., 1966, On Lora trevelliana (Turton) (Gastropoda: Turridae). Ophelia, 3: 81-91. PIELOU, E. C., 1974, Population and community ecology: principles and methods. Gordon & Breach, New York, 424 p. POOLE, R. W., 1974, An Introduction to Quantita- tive Ecology. McGraw-Hill, New York, 532 p. REX, M. A., 1976, Biological accommodation in the deep sea benthos: comparative evidence on the importance of predation and productivity. Deep Sea Research, 23: 975-987. 312 SHIMEK RHOADS, D. C. & YOUNG, D. K., 1970, The influ- ence of deposit-feeding benthos on bottom sedi- ment stability and community trophic structure. Journal of Marine Research, 28: 150-178. ROBINSON, E., 1960, Observations on the toxo- glossan gastropod Mangelia brachystoma (Philippi). Proceedings of the Zoological Society of London, 135: 319-338. SANDERS, H. L., 1958, Benthic studies in Buz- zards Bay. |. Animal-sediment relationships. Limnology and Oceanography, 3: 245-258. SANDERS, H. L., 1960, Benthic studies in Buz- zards Bay. Ill. The structure of the soft-sediment community. Limnology and Oceanography, 5: 138-153. SCHOENER, T. W., 1968, The Anolis lizards of Bimini: resource partitioning in a complex fauna. Ecology, 49: 704-726. SCHOENER, T. W., 1971, Theory of feeding strate- gies. Annual Review of Ecology and Systemat- ics, 2: 369-404. SHERIDAN, R., VAN MOL, J.-J. & 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, R. L., 1975, The morphology of the buc- cal apparatus of Oenopota levidensis (Gastro- poda, Turridae). Zeitschrift für Morphologie der Tiere, 80: 59-96. SHIMEK, R. L. & KOHN, A. J., 1981, Functional morphology and evolution of the toxoglossan radula. Malacología, 20: 423-438. SMITH, E. H., 1967a, The proboscis and oesopha- gus of some British turrids. Transactions of the Royal Society of Edinburgh, 67: 1-22. SMITH, E. H., 1967b, The neogastropod stomach, with notes on the digestive diverticula and intes- tine. Transactions of the Royal Society of Edin- burgh, 67: 23-42. SOKAL, R. R. & ROHLF, F. J., 1969, Biometry. Freeman, San Francisco, Ca., 776 p. TREVALLION, A. R., EDWARDS, R. C. 4 STEELE, J. H., 1970, Dynamics of a benthic bi- valve. In STEELE, J. H., ed., Marine Food Chains. Univ. Calif, Press, Berkeley and Los Angeles, p. 285-295. VERMEI, С. J., ZIPSER, E. & DUDLEY, E. C., 1980, Predation in time and space: peeling and drilling in terebrid gastropods. Paleobiology, 6: 352-364. WADE, B. A., 1972, A description of a highly di- verse soft bottom community in Kingston Harbor, Jamaica. Marine Biology, 13: 57-69. WHITTAKER, R. H., 1952, A study of summer foli- age insect communities in the Great Smoky Mountains. Ecological Monographs, 22: 144. WHITTAKER, R. H. & FAIRBANKS, C. W., 1958, A study of plankton copepod communities in the Columbia Basin, southeastern Washington. Ecology, 39: 46-65. WOODIN, S. A., 1974, Polychaete abundance pat- terns in a marine soft-sediment environment: the importance of biological interactions. Ecological Monographs, 44: 171-187. WOODIN, S. A., 1976, Adult-larval interactions in dense infaunal assemblages: patterns of abun- dance. Journal of Marine Research, 34: 25-41. WOODWICK, K. H., 1963, Taxonomic revision of two polydorid species (Annelida, Polychaeta, Spionidae). Proceedings of the Biological So- ciety of Washington, 76: 209-216. MALACOLOGIA, 1983, 23(2): 313-320 BURROWING AND THE FUNCTIONAL SIGNIFICANCE OF RATCHET SCULPTURE IN TURRITELLIFORM GASTROPODS Philip W. Signor III Department of Geology, University of California, Davis, California 95616, U.S.A. ABSTRACT Burrowing by some Indo-Pacific, high-spired gastropods is aided by low amplitude asymmetri- cal shell sculpture. This sculpture functions as a ratchet, resisting back-slippage when the snail is inserting its foot into the sediment while allowing free forward movement of the shell. In experiments, when ratchet sculpture is removed or covered the snails take longer and more burrowing cycles to burrow into the sediment, compared to unmodified animals or experimental controls. Species that have evolved ratchet sculpture are found in the Cerithiidae (the genus Rhinoclavis Swainson, 1840), Mitridae and Terebridae. As in burrowing bivalves with sculpture serving a similar function, ratchet sculpture remains approximately constant in size during the ontogeny of the snail. Ratchet sculpture is highly distinctive because it combines negative allometry and strong asymmetry, suggesting that ratchet sculpture may also be used in inferring the life habits of fossil gastropods. Key words: Gastropoda; functional morphology; sculpture; burrowing; Terebra; Mitra; Rhinoclavis. INTRODUCTION Infaunal bivalves have evolved a variety of shell sculptures that aid the bivalves in bur- rowing. These sculptures include growth-con- formable concentric ridges (e.g. Anomalo- cardia brasiliana (Gmelin, 1791); Stanley, 1981), radial ribs (e.g. Codakia Scopoli, 1777; Seilacher, 1973), combinations of concentric and radial ribs (e.g. Donax dentifer Hanley, 1843; Seilacher, 1973) as well as discordant (neither spiral nor axial) asymmetrical ridges (e.g. Divaricella von Martens, 1880; Strigilla Turton, 1882; Stanley, 1969, 1970; Seilacher, 1972, 1973). Also, at least one genus of bi- valves has evolved asymmetrical modifica- tions of the periostracum which may aid in burrowing (Sinonovacula Prashad, 1924; Seilacher, 1972). Thus far similar sculptures have not been described in the Gastropoda, nor has the function of the sculpture actually been tested in the majority of cases. Many turritelliform gastropods are active burrowers. Like most burrowing bivalves and gastropods (Trueman & Ansell, 1969, and references therein), turritelliform species bur- row with a series of discontinuous move- ments. The shell and part of the animal act as a penetration anchor as the snail's foot probes the sediment. The foot then forms a terminal anchor and the shell and remainder of the animal is drawn forward. This burrowing cycle is repeated as the snail continues to burrow. A spike-like shell is not an optimal anchor, but the anchoring effect of the shell can be improved with sculpture. This sculp- ture would serve a ratchet-like function, in- creasing the effectiveness of the penetration anchor by reducing back-slippage while mini- mizing resistance to forward movement through the sediment. Several species of Indo-Pacific gastropods have evolved asym- metrical sculpture that appears to serve this ratchet function. The objective of this project was to test the hypothesis that shell sculpture aids in burrowing by turritelliform gastropods by increasing the anchoring effect of the shell. PROCEDURES | selected two relatively common species of infaunal Indo-Pacific gastropods for the ex- perimental portion of this study. These spe- cies, Terebra dimidiata (Linnaeus, 1758) (Neogastropoda: Conacea) and Rhinoclavis aspera (Linnaeus, 1758) (Mesogastropoda: Cerithiacea), occur sympatrically in the Indo- Pacific region but subsist on different foods, burrow in different ways, and have different sculptures. Terebra dimidiata is an infaunal predator (313) 314 SIGNOR which feeds on enteropneusts (Miller, 1966, 1975). This species burrows solely with the foot, in a manner similar to that described for Terebra gouldi Deshayes, 1859 (Miller, 1975). The shell of this species bears a single cuesta or asymmetrical ridge on each whorl; a second cuesta is formed at the sutural ramp (Fig. 1A, B). The cuestas on each whorl are of approximately equal size and remain about the same size on successive whorls. Eight specimens of 7. dimidiata from Piti Bay, Guam, were used in the first experiment. The snails were maintained in running sea- water aquaria with coarse carbonate sand substrates. Seawater temperature was fairly constant throughout the experiments (27- 29°C). Each snail was timed for three trials burrowing into the substrate; the number of burrowing cycles required for the animals to bury themselves was also recorded. (For the purposes of this experiment, burrowing time was defined as the length of time from initial penetration of the sediment until the apex of the shell was drawn into the substrate.) This procedure was repeated for the subsequent trials. The sculpture of each snail was then FIG. 1. A. Terebra dimidiata from Piti Bay, Guam. Scale bar is 1 cm. B. Scanning electron micrograph (SEM) of sculpture on T. dimidiata. Cuesta in center is the sutural ramp. Note cuestas are approximately equal in height. Cuestas are shingled toward the apex (to lower left), shallow slope is toward aperture. Scale bar is 1 mm. С. Rhinoclavis aspera from Pago Bay, Guam. Scale bar is 1 cm. D. SEM of А. aspera sculpture. Tubercles are inclined toward apex. Scale bar is 1 mm. GASTROPOD RATCHET SCULPTURE 315 filled with a quick-drying paste and allowed to dry. Excess paste was removed and the paste was covered with a lacquer and allowed to dry. The animals were wetted periodically dur- ing this process, which lasted up to seven minutes, and were returned to the aquaria for one day prior to the burrowing trials. For a control on the experimental treatment, the lacquer and paste were removed, the animals rested one day, and then timed burrowing. The experimental treatment added only a few percent to the weight of the animal. Any effect from this increase in weight was offset by an amount equal to the weight of the water displaced by the paste and lacquer. A second control, on lacquer alone, was not possible because the lacquer tended to collect in the cuestas and round out the sculpture. This bead of lacquer dried only very slowly and required keeping the snail out of the water for an extended period. The one control tests for possible long term effects of the total treat- ment on the snails. The behavior of the ani- mals showed no obvious changes when the snails were coated with paste and lacquer. Rhinoclavis aspera is an infaunal cerithiid (Fig. 1C) (Houbrick, 1978). The diet of this species is unknown but Houbrick (1978) has reported algae and sand grains in the stom- achs of some specimens. This species’ sculp- ture consists of asymmetrical tubercles mounted on transverse (collabral) ridges (Fig. 1D). R. aspera uses both its large spatulate head and foot to penetrate the sediment. In burrowing, the head is placed upon the ante- rior dorsal surface of the foot and both are inserted into the sediment. The head is lifted up, the foot thrusts down, and the animal pulls the shell and body forward. This cycle is repeated as burrowing continues. Rhinoclavis aspera could not be treated with the same experimental technique used for 7. dimidiata because of the high relief of the sculpture. To cover the tubercles would require large amounts of paste which would concomitantly increase the area of the ante- rior cross-section of the shell. This would in- crease the burrowing time regardless of the effect of covering the sculpture. Fifty-one specimens of R. aspera were col- lected from Pago Bay, Guam, and divided into three groups of 17 specimens each. The snails were placed in running seawater aquaria with coarse carbonate sand sub- strate. The initial group was not modified. The tubercles were carefully filed from the second (experimental) group. About five minutes were required to gently file off the tubercles without removing the transverse ridges. In the third group of snails, the control group, the snails were filed for five minutes but only the sculpture on the dorsal side of the apex was removed. As the apex is the last part of the shell drawn into the substrate, scupture in this area should be the least important in burrow- ing. Each snail was rested for one day before burrowing trials commenced. The snails were timed for three trials each burrowing into the sand; the number of burrowing cycles was also recorded for each burrowing. Animals were rested briefly between burrowings to prevent tiring. RESULTS Analysis of the data from the experiment with Terebra dimidiata is complicated by a correlation between burrowing times and shell length (г = .68, N = 19, P < .001) (Fig. 2). The number of burrowing cycles is also corre- lated with shell length (r = .62, N = 19, P < .002). Therefore, the data were analyzed using ANOVA techniques treating shell length as a covariate of burrowing time. Results show that Terebra dimidiata re- quired more time and burrowing cycles to bur- row when the sculpture was covered, com- pared to the initial or control treatments (Fig. 3). The F-values obtained in the pair-wise comparisons of burrowing times of the un- modified and control groups with the experi- mental group were 26.74*** and 29.19””*, re- spectively (N = 8). There were no significant differences between the times and number of burrowing cycles taken to burrow by the un- modified and control groups (the F-value for the comparison of burrowing times between the unmodified and control groups was .817). Similar results were obtained in the analysis of the number of burrowing cycles taken to bur- row by the different groups. Rhinoclavis aspera also takes significantly more time and burrowing cycles to burrow when the sculpture is removed, compared to either the control or unmodified groups (Fig. 4). As with the previous experiment, the analysis treats burrowing time as a covariate of shell length. The F-values obtained in the pair-wise comparison of the unmodified and control groups with the experimental group were 9.20** and 5.92*, respectively (N = 17). Comparison of the unmodified and control group yielded an F-value of .008. Similar re- 316 SIGNOR 400 TEREBRA DIMIDIATA 300 à © D о 2 | > E 200 À 2 : т — lOO 7 40 50 60 70 8 90 100 SHELL LENGTH (тт) FIG. 2. Burrowing times for Terebra dimidiata. Shell length is highly correlated with burrowing time. Each point is the average of three burrowing trials for a single animal. 4001 TEREBRA DIMIDIATA > — 300 | + NS) D | 2 200 | Lu I} ? = во + 100 o UNMODIFIED à CONTROL à EXPERIMENTAL po 40 50 60 70 80 90 SHELL LENGTH (mm) FIG. 3. Results of burrowing trials with artificially smoothened Terebra dimidiata. Note the uniform increase in burrowing times when the ratchet sculpture is covered (experimental group). Each point is the average of three burrowings by one individual. Vertical lines connect data from three treatments of one specimen. GASTROPOD RATCHET SCULPTURE 317 RHINOCLAVIS ASPERA 400 300 200 TIME (sec) 100 © UNMODIFIED a CONTROL 4EXPERIMENTAL I 29 30 SS 40 45 SHELL LENGTH (mm) FIG. 4. Results of burrowing trials with artificially smoothened Rhinoclavis aspera. Each point is an average of three burrowings by one individual. sults were obtained in the analysis of the number of burrowing cycles the snails re- quired to burrow when the sculpture was re- moved. DISCUSSION The consistent results obtained with the two different experimental methodologies strongly supports the hypothesis that ratchet sculpture aids burrowing by increasing the effective- ness of the snail's penetration anchor. During the course of the experiments, | frequently ob- served the penetration anchor of smoothened Snails to fail, causing the shell to slip back- wards. Back-slippage was immediately fol- lowed by a brief cessation of burrowing, which partially accounts for the greatly increased burrowing times of artificially smoothened ani- mals. An alternative to the methodologies em- ployed here would be to compare the length of single burrowing cycles between the un- modified, experimental and control groups. However, | did not utilize this methodology because the lengths of the burrowing cycles increase as the snails penetrate deeper into the sediment. Any analysis using length of burrowing cycles would have required an ad- ditional correction for the depth to which the snail had penetrated the sediment. This seemed an unnecessary complication to the experimental design. Several other factors potentially could bias the results reported here. These factors are handling of the animals (Brown, 1971), time- specific activity patterns (Webb et al., 1959; Miller, 1966) and the use of artificial or re- duced light in the experiments (Bell & Frey, 1969). In order to avoid these sources of bias, care was taken to ensure that the possible 318 effects of each of these factors were spread over all treatment groups. Linsley (1978) noted that size does not seem to be a major factor determining loco- motory rates in most gastropods. Apparently, this generalization does not extend to burrow- ing, at least by shell-draggers. Large Terebra dimidiata penetrate the sediment faster, in absolute terms, than small individulas (Fig. 2). This pattern in high-spired gastropods is simi- lar to that in bivalves, where there is a strong correlation between shell length and burrow- ing time (Stanley, 1970). The sculpture present on the species studied here is negatively allometric during the ontogeny of the organisms. The height of the sculpture remains approximately constant in size during ontogeny while the distance be- tween tubercles or cuestas increases in con- stant proportion to other features of the shell. In T. dimidiata, the sculpture increases in height at a rate of about 1.1, while the whorl expansion rate of the species is about 1.2 (see Raup, 1966, for a discussion of whorl expansion rate). In R. aspera from Guam, the height of the tubercles increases until they reach about .5 mm and remain approximately constant in height thereafter. This pattern of constant sculpture size during ontogeny matches the pattern observed among many bivalves and crustaceans which have ratchet sculpture (see Seilacher, 1972, 1973). One remaining puzzle is why so many spe- cies of burrowing snails (and bivalves!) lack ratchet sculpture. | have found ratchet sculp- ture on fewer than ten percent of the more than two hundred infaunal marine turritelliform species | have examined. The majority of the burrowing bivalves examined by Stanley (1970) lack any ratchet sculpture. Are most species failing to optimize or is ratchet sculp- ture advantageous only under limited circum- stances? Stanley (1969) suggested that ratchet sculpture should not be expected in bivalves dwelling in muddy sediment because the mud would be too fluid to allow the sculp- ture to engage the substrate. Further re- search is necessary to resolve this problem. Several burrowing gastropods have asym- metrical shell sculpture that does not remain constant in size during ontogeny (e.g. Bullia sp., Terebra crenulata (Linnaeus, 1758); Fig. 5). This sculpture may also aid in burrowing, but | have not been able to test this hypothe- sis. Savazzi (1981) has shown experimentally that the efficiency of ratchet sculpture is de- creased when the sculpture becomes large SIGNOR FIG. 5. Species with sculpture resembling ratchet sculpture. A. Bullia sp. B. Terebra crenulata. Scale bars are 1 cm. relative to the sediment grain size. Savazzi's results suggest that asymmetrical, isometric, sculpture may serve some function besides assisting in burrowing. Most species of burrowing turritelliform gastropods lack pronounced shell sculpture, presumably because the increased resis- tance to penetration of the sediment resulting from sculpture would inhibit burrowing. This problem is avoided by species with ratchet sculpture because ratchet sculpture is asym- metrical (smoothed in the direction of burrow- ing), in contrast to the symmetrical sculpture usually found on epifaunal species (see Savazzi, 1981, for a discussion of the relative merits of symmetrical and asymmetrical sculpture). | would expect epifaunal species which occasionally burrow (e.g. Cerithium nodulosum Bruguiere, 1789) would burrow faster when artificially smoothened because of the reduction in friction with the sediment. This is exactly the opposite of the experi- mental results presented here, as burrowing times increase when ratchet sculpture is re- moved. Experimental modification of species with other types of sculpture remains to be attempted but it does not seem likely that symmetrical sculpture will aid in burrowing. The combination of strong asymmetry and negative allometry during ontogeny make ratchet sculpture a highly distinctive morphol- GASTROPOD RATCHET SCULPTURE 319 ogy. Thus far, | have observed ratchet sculp- ture only in actively burrowing species among the Mitridae, Terebridae and Cerithiidae (the genus Ahinoclavis). This suggests that ratch- et sculpture will be useful in the interpretation of the life habits of fossil gastropods. Unfor- tunately, the use of ratchet sculpture in paleo- ecology will be severely limited because ratchet sculpture is comparatively uncom- mon. Some non-burrowing crabs (e.g. Grapsus grapsus (Linnaeus, 1758); Schmal- fuss, 1978a) and bivalves (Bankia setacea (Tryon, 1863); Roder, 1977) have evolved sculpture similar to ratchet sculpture, which the organisms use to wedge themselves against hard substrates. However, this sculp- ture is not negatively allometric during ontog- eny which, along with other adaptations to hard substrates, permits these organisms to be distinguished from burrowers. As mentioned in the introduction, ratchet sculpture has also evolved in the Bivalvia. Stanley (1969, 1970) and Seilacher (1972, 1973) have documented occurrences of ratchet sculpture in a variety of modern and extinct infaunal species. Ratchet sculpture has also evolved in the Brachiopoda (Seilacher, 1972), Crustacea (Seilacher, 1972, 1973; Schmalfuss, 1978a; Savazzi, 1981) and Trilobita (Stitt, 1976; Schmalfuss, 1978b), and has, in most cases, been inferred to aid in anchoring the organism in an uncon- solidated substrate. A similar structure has also been noted in the carpoid echinoderms (or calcichordates) (Jeffries, 1975; Derstler, 1975; Jefferies & Lewis, 1978). Particularly interesting is the observation that Ca/appa (Weber, 1795), a burrowing crab that preys on both Rhinoclavis aspera and Terebra dimidiata (Vermeij, 1978; Vermeij et al., 1980) has also evolved ratcheted sculpture on its Carapace (Schmalfuss, 1978a). The evolution of ratchet sculpture among these different taxa is one of the unusual examples of con- vergent evolution among marine inverte- brates. CONCLUSIONS Artificially smoothened specimens of Terebra dimidiata and Rhinoclavis aspera take more time and burrowing cycles to bur- row compared to controls or to unmodified animals. The sculpture on these species ap- pears to perform a ratchet function, prevent- ing back-slippage by locking the shell to the substrate while the snail is inserting its foot into the sediment but allowing free forward movement of the shell through the sediment. Ratchet sculpture is distinct from other sculp- tures because it combines strong asymmetry with constant size during the ontogeny of the animal. Ratchet sculpture may be useful for interpreting life habits of fossil gastropods, because it is morphologically distinct and limited to actively burrowing species. Other authors have documented conver- gent features which evolved in the Brachio- poda, Crustacea, Trilobita and possibly the Echinodermata. Thus, ratchet sculpture is a common solution to the functional needs of a diverse range of burrowing organisms. ACKNOWLEDGMENTS The research was funded in part by an H. T. Stearns Fellowship from the Geological So- ciety of America and a Grant-in-Aid of Re- search from Sigma Xi, the Scientific Research Society. | thank M. Colgan, P. Kat, A. Seil- acher, C. Signor, B. Smith, S. M. Stanley and members of the Guam Shell Club for aid, dis- cussions and criticisms. G. J. Vermeij and an anonymous reviewer suggested several im- provements to this paper. This is contribution 149 of the University of Guam Marine Labora- tory. REFERENCES CITED BELL, B. M. & FREY, R. W., 1969, Observations on ecology and the feeding and burrowing mecha- nisms of Mellita quinquiesperforata (Leske). Journal of Paleontology, 43: 553-560. BROWN, A. C., 1971, The ecology of the sandy beaches of the Cape Peninsula, South Africa, Part 2: The mode of life of Bullia (Gastropoda: Prosobranchiata). Transactions of the Royal Society of South Africa, 39: 281-319. DERSTLER, K., 1975, Carpoid echinoderms from Pennsylvania. Geological Society of America Abstracts with Programs, 7: 48. HOUBRICK, R. S., 1978, The family Cerithiidae in the Indo-Pacific, Part 1: The genera Rhinoclavis, Pseudovertagus and Clavocerithium. Mono- graphs of Marine Mollusca, 1: 1-130. JEFFERIES, R. P. S., 1975, Fossil evidence con- cerning the origin of the chordates. Symposia of the Zoological Society of London, 36: 253-318. JEFFERIES, R. P. S. & LEWIS, D. N., 1978, The English Silurian fossil Placocystites forbesianus and the ancestry of the vertebrates. Philosophi- cal Transactions of the Royal Society of London, ser. B, 282: 205-323. 320 SIGNOR LINSLEY, R. M., 1978, Locomotion rates and shell form in the Gastropoda. Malacologia, 17: 193- 206. MILLER, B. A., 1966, Studies on the biology and ecology of some Indo-Pacific Terebra from Eniwetok Atoll. Unpublished M.S. thesis, Univer- sity of Hawaii, Honolulu, 85 p. MILLER, B. A., 1975, The biology of Terebra gouldi Deshayes, 1859, and a discussion of life history similarities among other terebrids of similar pro- boscis type. Pacific Science, 29: 227-241. RAUP, D. M., 1966, Geometric analysis of shell coiling; general problems. Journal of Paleontol- ogy, 40: 1178-1190. RODER, H., 1977, Zur Beziehung zwischen Kon- struktion und Substrat bei mechanisch bohren- den Bohrmuscheln (Pholadidae, Teredinidae). Senckenbergiana Maritima, 9: 105-123. SAVAZZI, E., 1981, Functional morphology of the cuticular terraces in Rania (Lophorania) (brachyuran decapods; Eocene of N.E. Italy). Neues Jahrbuch fur Geologie und Palaontologie, Abhandlungen, 162: 231-243. SCHMALFUSS, H., 1978a, Structure, patterns and function of cuticular terraces in Recent and fossil arthropods. Zoomorphologie, 90: 19-40. SCHMALFUSS, H., 1978b, Constructional mor- phology of cuticular terraces in trilobites, with conclusions on synecological evolution. Neues Jahrbuch für Geologie und Paleontologie, Abhandlungen, 157: 164-168. SEILACHER, A., 1972, Divaricate patterns in pelecypod shells. Lethaia, 5: 325-343. SEILACHER, A., 1973, Fabricational noise in adap- tive morphology. Systematic Zoology, 22: 451- 465. STANLEY, S. M., 1969, Bivalve mollusk burrowing aided by discordant shell ornamentation. Sci- ence, 166: 634-635. STANLEY, S. M., 1970, Relation of shell form to life habits in the Bivalvia (Mollusca). Geological So- ciety of America, Memoir 125, 296 p. STANLEY, S. M., 1981, Infaunal survival: alterna- tive functions of shell ornamentation in the Bi- valvia (Mollusca). Paleobiology, 7: 384-393. STITT, J. H., 1976, Functional morphology and life habits of the Late Cambrian trilobite Stenopilus pronus Raymond. Journal of Paleontology, 50: 561-576. TRUEMAN, A. R. & ANSELL, A. D., 1969, The mechanisms of burrowing into soft substrate by marine animals. Oceanography and Marine Bi- ology, Annual Review, 7: 315-366. WEBB, H. M., BROWN, Е. A., Jr. & BRETT, W. J., 1959, Fluctuations in rate of locomotion by llyanassa. Biological Bulletin, 117: 431. VERMEIL, С. J., 1978, Biogeography and Adapta- tion. Harvard University Press, Cambridge, Massachusetts, 416 p. VERMEIJ, G. J., ZIPSER, E. & DUDLEYNE NC: 1980, Predation in time and space: Peeling and drilling in terebrid gastropods. Paleobiology 6: 352-364. MALACOLOGIA, 1983, 23(2): 321-331 EFFETS DES CONDITIONS D’ECLAIREMENT SUR LE POTENTIEL _ REPRODUCTEUR DE LYMNAEA STAGNALIS (GASTEROPODE PULMONE) J. Seugé et R. Bluzat Laboratoire de Zoologie— Bâtiment 442, Centre d'Orsay de l'Université de Paris-Sud, 91405—Orsay Cédex—France RESUME La fécondité de Lymnaea stagnalis a été étudiée sous courte et longue photophase, sous différentes intensités lumineuses, sous cing lumières colorées et à l'obscurité. La fécondité cumulée des limnées élevées sous courte photophase (LD 7:17) est toujours plus élevée (9.6% à 89.5%) que celle des animaux élevés sous longue photophase (LD 16:8). Cette différence peut être expliquée par un accroissement de la longévité qui conduit à un plus grand nombre de pontes par animal. La fécondité n'est proportionnelle à l'intensité lumineuse que sous longue photophase. Par rapport à leur témoin respectif, la fécondité cumulée n'est réduite que sous la lumière jaune (—13.4%); elle est augmentée dans les 4 autres cas: +35.7% (bleu), +47.4% (vert), +61.8% (U.V.) et +77% (rouge). A l'exception du cas de la lumière jaune, la durée de la période de ponte et le nombre de pontes par animal sont augmentes. Nos résultats montrent que les cing lumières colorées sont perçues par les limnées. Une réduction de la fécondité, due essentiellement a un effet sur le nombre de pontes pondues par animal, est observée lors de l'élevage a l'obscurité. Le ‘Potentiel Limnée” (P.L. = volume de la coquille x fécondité cumulée) est plus élevé quand les animaux sont élevés sous jours courts. Le P.L. n'est proportionnel à l'intensité lumineuse que chez les limnées élevées sous jours longs et il atteint une valeur paradoxale chez celles qui sont élevées à l'obscurité. Le P.L. est réduit chez les animaux élevés sous lumière jaune; il est au contraire augmenté dans les quatre autres cas de lumière colorée. Ceci met en évidence un effet spécifique de ces lumières. Mots-clés: Lymnaea stagnalis; fécondité; photophases; intensités lumineuses; lumières colorées; obscurite. INTRODUCTION Dans des travaux antérieurs, nous avons montre, chez Lymnaea stagnalis (Linn.) l’ex- istence d'importants troubles de la croissance et de la fécondité d'une part lors d'intoxica- tions à long terme par divers toxiques (Bluzat et al., 1979; Bluzat & Seuge, 1981; Seuge & Bluzat, 1979 et sous presse), d'autre part lors de la variation d'un facteur du milieu tel que la mineralisation de l'eau (Seuge, 1980; Seugé & Bluzat, 1982a). Dans une étude précédente (Seugé & Bluzat, 1982b), nous avons ex- amine l'influence de diverses conditions d'eclairement sur la croissance de ce Mol- lusque; nous nous proposons ici d'exposer les effets sur sa fécondite de la photophase, de l'intensité lumineuse en lumière blanche, de l'obscurité et de cinq lumières colorées. A notre connaissance, les données bibli- ographiques relatives à ces sujets sont peu nombreuses chez les Gastéropodes: Deschiens & Byan (1956), Deschiens (1957), Gaud (1958), de Witt & Sloan (1960), Yung (1911; Franc, 1968: 498), Graber (1884; Franc, 1968: 498, 499), Liche (1934a, 1934b) et Carmichael (1933; Franc, 1968: 499), analysées par Franc (1968), Van der Steen (1967), Bohlken et al. (1978) et Dogterom (1980): les informations qui peuvent en être recueillies sont souvent contradictoires. Les differences de méthodologie, la durée d’ob- servation et l'espèce étudiée en particulier, peuvent expliquer, en partie, les divergences constatées mais les auteurs modernes (Van der Steen, 1967; Bohlken et al., 1978; Dog- terom, 1980 & Seuge, 1980) s’accordent pour souligner, chez Lymnaea stagnalis, l'impor- tance des conditions d’eclairement sur la re- production. MATERIEL ET METHODE L’animal utilise est le Mollusque Gastero- pode Pulmone Lymnaea stagnalis eleve (321) 322 SEUGE ET BLUZAT depuis plusieurs generations au laboratoire dans l'eau de ville (pH: voisin de 7,5; dureté totale: 230 mg CaCO;/litre; 20° + 0,5C; aeration permanente); l'alimentation est as- suree par des feuilles de laitue fournies en quantite suffisante et l'eau est renouvelée chaque semaine. Toutes les expériences ex- posees ci-dessous sont réalisées simultané- ment avec des animaux groupes dans un bac de verre contenant 2 litres d'eau. Trois cents jeunes limnées sont mises en experience le jour de leur éclosion; à cinq semaines elles sont mesurées pour la premiere fois et seuls les 40 individus de plus grande taille sont conserves. Une deuxième sélection est operee de la même façon un mois plus tard: chaque lot est alors constitue de 12 limnées qui sont maintenues en expérience jusqu'à ce que 50% d'entre elles soient mortes; l'étude de chaque lot est donc arrêtée quand le nombre de limnées vivantes est <5. Dans tous les cas les animaux sont exposés dès l'éclosion aux conditions lumineuses etudiees. Des élevages sont réalisés sous lumière blanche (tubes Mazda “blanc industrie” TFR/ 40/BBL) soit en longue photophase LD 16:8 series 16, soit en courte photophase LD 7:17 series 7 et dans les deux cas sous différentes intensites lumineuses: — 1200 lux: series 16-1 et 7-1, — 180 lux: series 16-2 et 7-2, — 10 lux: series 16-3 et 7-3, — 50 lux: serie 16-4. En plus un lot a été élevé à l'obscurité quasi- permanente (serie 0). Les élevages sous lumières colorées ont tous ete realises sous une photophase LD 16:8: les gammes de radiations sont ob- tenues par des tubes lumineux spéciaux entourés d'un filtre de rhodoid “R.P.” de 0,25 mm d'épaisseur. Les tubes et les filtres utilisés sont les suivants: —lumière ultraviolette (U.V.): tubes TFA/4L/5 Mazda; pas de filtre; À = 380- 440 nm; 360 lux (témoin: 16-2). —lumière bleue (B): tubes TF/40 Bleu Mazda; rhodoid n° 2005; À = 430- 480 nm; 230 lux (temoin: 16-2). —lumière verte (V): tubes TF/40 Vert Mazda; rhodoid n° 2003; À = 500- 550 nm; 360 lux (témoin: 16-2). —lumière jaune (J): tubes TF/40 Jaune Mazda; rhodoid n° 222; À = 550-630 nm; 1300 lux (temoin: 16-1). —lumiere rouge (R): tubes TL/40/15 Philips; rhodoid n° 227; À =630-670 nm; 50 lux (temoin: 16-4). Dans tous les cas les intensités lumineuses ont ete mesurées au niveau de la surface de l'eau; il est impossible d'apprécier avec ex- actitude la quantité de lumière reçue par chaque animal en fonction de sa position dans l'aquarium (dissimulation sous la salade, niveau de Гепюпсетеп{ entre autres). Les pontes sont prelevees et denombrees chaque semaine; le nombre des oeufs dans les pontes est compté sous stéréomicro- scope. Les resultats sont donc enregistres chaque semaine, a partir des premieres pontes et jusqu'à la mort de 50% des animaux dans chacun des 105; les moyennes hebdomadaires (nombre d’oeufs/nombre d'animaux; nombre de pontes/nombre d'animaux; nombre d'oeufs/nombre de pontes) permettent de calculer pour chaque lot en effectuant la somme de 4 moyennes hebdomadaires la valeur mensuelle du: —nombre moyen d'oeufs par animal (f), —nombre moyen de pontes par animal. L'évolution de la fécondité cumulée (F) et du nombre cumulé de pontes (P) est ensuite étudiée en fonction de l’âge puis en fonction de la taille; par ailleurs, le nombre moyen d'oeufs par ponte (Nm) est calculé pour toute la période de ponte. Le rendement biologique de l'espèce est enfin apprécié par le calcul du produit: volume de la coquille (У en mm3) x fécondité cumulée (F), auquel nous donnons le nom de “Potentiel Limnée” (P.L.) (Seugé, 1980). RÉSULTATS La Fig. 1 montre l'influence de la durée de la photophase, de l'intensité lumineuse et de — FIG. 1. Analyse de la fécondité en fonction du temps (nb w/an: nombre d'oeufs par animal): influence de la longueur de la photophase sous trois intensités lumineuses. 1a = intensité lumineuse 1200 lux; 1b = intensité lumineuse 180 lux; 1c = intensité lumineuse 10 lux et obscurite; Partie inférieure des graphiques; fecondité mensuelle (f); Partie médiane des graphiques: fécondité cumulée (F); Partie supérieure des graphiques: pourcentage de variation de F (LD 7:17 = séries 7)—par rapport à F (LD 16:8 = séries 16): 1a et 1b.—par rapport à F (0): 1c. 1d, e et f: variations mensuelles du nombre moyen de pontes émises par limnée sous longues (séries 16) et courtes (séries 7) photophases et sous différentes intensités lumineuses: 1200 lux (1d), 180 lux (1e), 10 lux et obscurite (1f). ÉCLAIREMENT ET POTENTIEL REPRODUCTEUR DE LYMNAEA 323 o . . Yo de variation ode variation Nu e temps en mois % de variati ÓN AAA nb pontes/an. temps en mois 324 SEUGÉ ET BLUZAT l'obscurité sur la fécondité: 1a partie inférieure du graphique 1a (1200 lux) indique que la fécondité mensuelle fluctue en fonction du temps sous les deux photophases; la partie médiane, fécondité cumulée en fonction de l'âge, souligne que les animaux 7-1 pondent au total plus que les autres. De même les Fig. 1b (180 lux) et 1c (10 lux) amènent à des conclusions identiques dans fecondite a 1500 fécondite 1000 500 le cas des faibles intensites; le graphique 1c presente en plus l'évolution de la fécondité des animaux éleves à l'obscurité et demontre qu'ils pondent au total un peu plus que les animaux 16-3 mais beaucoup moins que les animaux 7-3. Une seule des deux composantes de la fécondité, le nombre moyen de pontes emises par animal, est réellement modifiée: mois FIG. 2. Analyse de la fécondité en fonction du temps: influence de la couleur de la lumière. 2a et 2b: évolution de la fécondité mensuelle (f); 2c et 2d: pourcentages de variation de la fécondité cumulée en fonction du témoin respectif (le moment de la mort de celui-ci est précisé par une fleche). ÉCLAIREMENT ET POTENTIEL REPRODUCTEUR DE LYMNAEA les courbes 1d, e et f en présentent les varia- tions mensuelles dans chaque cas. La Fig. 2 montre l'influence des diverses gammes de longueurs d'onde sur la fécon- dite. Les graphiques 2a et 2b précisent l'évo- lution de la fécondité mensuelle dans les dif- férents groupes. Dans trois cas (U.V., Jaune et Rouge, Fig. 2a) cette évolution se traduit, comme pour la lumière blanche, par la suc- cession d'une phase ascendante, d'un maxi- mum de fécondité et d'une phase descen- dante; ces trois phases n'existent pas dis- tinctement quand les limnées sont élevées sous des lumières bleue et verte ainsi qu'à l'obscurité (Fig. 2b). Les graphiques 2c et 2d indiquent, en pourcentages, les fluctuations de la fecondite cumulée en fonction du temps par rapport à celle du témoin respectif; à l'exception du cas de la lumière jaune ces variations sont impor- tantes, le plus souvent dans le sens positif. 325 Les données de la fécondité cumulée en fonction de l’âge permettent, pour simplifier, de tracer des droites qui sont rassemblées dans la Fig. 3a: la fécondité des animaux élevés à l'obscurité se distingue nettement de celle des autres groupes. Les graphiques 3b, с et d montrent l’evolu- tion de la fécondité cumulée en fonction de la taille après transformation log-log: cette représentation souligne que pour une taille donnée la fécondité varie avec les conditions d'éclairement. Le Tableau 1 permet de confronter le bilan de la fécondité des limnées élevées sous les différentes conditions d'éclairement; il rap- pelle en outre la taille moyenne maximale observée des coquilles (Seugé & Bluzat, 1982b) et indique enfin la valeur du “Potentiel Limnée” dans chaque cas. Le Tableau 2 rassemble les résultats rela- tifs à la fertilité des oeufs, à la durée moyenne TABLEAU 1. Fécondité des limnées élevées dans différentes conditions d'éclairement (12 limnées par groupe). Age Groupes début Durée expérimentaux ponte ponte mx OD et témoins (mois) (mois) F z P Nm (mm) P.L. x 107 16-1 2,25 11 7882 9,2 78,5 100,4 38,9 4,3 16-2 2 8 6696 8,8 67,4 99,3 36,5 3,4 16-3 2,5 8 4655 7,9 49 95 36,3 2,5 16-4 2,25 7/ 4723 Vath 54,5 86,8 36,9 2,4 EE PP A EP ES SS EE OE SE Oe Е D ZE EE 2: 3Ы 7-1 3 13 8637 12,9 87,6 98,6 42 6,6 7-2 2,75 12 8915 9,5 78,2 114 41,7 722 7-3 2,5 13 8821 8,3 81,9 107,7 42,7 745 ETC TP DORE EP OP EME CEST EE RES EOS DS Pei PE SEG D DPF ERREURS JT ST RNA TE DR то 0 2,5 13 5107 55 49,2 103,8 45,9 5:3 U.V. (Т 16-2) 2 12 10832 10,7. 108,6 99,7 39,8 7,2 Bleu (T 16-2) 2 9 9089 14,1 94,6 96,1 35:5 3,8 Vert (Т 16-2) 2 9 9871 14,1 85,6 115,3* 38,1 4,8 Jaune (Т 16-1) 2 9 6827 10 81,2 84,1* 40,3 Sh i/ Rouge (T 16-4) 2 12 8362 10,2 92,9 90 42 6,3 PL SE ET NE PEER SE PEE SE OST DE EAE SES SSE SS eS ees Se ee Groupes 16-1, 16-2, 16-3 et 16-4: longue photophase (LD 16:8) pour des intensités 1200, 180, 10 et 50 lux. Groupes 7-1, 7-2 et 7-3: courte photophase (LD 7:17) pour des intensités de 1200, 180 et 10 lux. Group 0: élevage à l'obscurité. Groupes U.V., Bleu, Vert, Jaune et Rouge: lumières colorées (LD 16:8); pour chaque cas le témoin lumière-blanche est indiqué entre parenthèses. Valeurs moyennes par limnée de: —F: fécondité cumulée; — 7: pente de la droite log (F) = z log (T) + log (a); —P: nombre cumulé de pontes émises par animal; —Nm: nombre moyen d'oeufs par ponte; P.L.: “Potentiel Limnée.” —*: la comparaison des moyennes mensuelles montre que les séries J et 16-1 d'un côté, V et 16-2 d'un autre diffèrent significativement (test t avec P< 0.01). 326 SEUGE ET BLUZAT АЕ (log) A | а RN ae 4 | 8 9 10 1 12 une 5 35 40 я taille (log) 40 FIG. 3. Analyse de la fécondité cumulée (F). 3a: en fonction de l'âge des animaux, cas des animaux élevés sous les lumières colorées et à l'obscurité; 3b, с et а: en fonction de la taille (T) après transformaton log. log. des données; influence de la photophase en forte (3b) et faible (3c) intensité lumineuse; influence des lumières colorées (3d). de l'embryogenèse et a la fréquence des CONCLUSIONS ET DISCUSSION anomalies observées (embryons doubles ou multiples); précisons que les oeufs ont été Le potentiel reproducteur des limnées incubés dans les conditions exactes où les soumises à différentes conditions lumineuses parents ont été élevés. sera analyse en envisageant la durée de la ÉCLAIREMENT ET POTENTIEL REPRODUCTEUR DE LYMNAEA 327 TABLEAU 2. Fertilité des oeufs, temps moyen d'éclosion (T.M.E.) et pourcentages d'anomalies (embryons doubles et multiples) en fonction des différentes conditions lumineuses (incubation des oeufs dans les conditions d'origine). Pour les groupes: idem Tableau 1. Groupes Fertilité Anomalies expérimentaux EE PA et témoins Nombre d'oeufs % eclosion TME (jours) Nombre d'oeufs % 16-1 1466 83,6 22 1606 0,68 16-2 1865 92,1 21 5825 0,21 16-3 1125 86,2 22 4328 0,15 16-4 1204 86,7 22 7982 0,11 7-1 1605 80,9 21 3829 0,18 7-2 1911 91,6 21 7216 0,18 7-3 1155 86,3 22 8087 0,05 E A A A DS A A ET SLE an un mem Obscurité 989 66,5 22 2277 0,66 AAA A AA A AAA A A A u AAA nn U.V. 2286 97,1 25 5615 0,19 Bleu 3305 85 20 5944 0,15 Vert 2525 91,2 19 7424 0,45 Jaune 3490 97,7 22 7932 8,76 Rouge 3153 92,2 19 3518 0,17 periode de ponte puis la fecondite; enfin, nous tenterons de realiser une synthese des effets de la lumière sur la croissance et la fécondité. Le fécondité peut être étudiée selon deux methodes; d'abord en envisageant l'évolution des données mensuelles (f) puis celle des données cumulées. L'étude de la fécondité cumulée (F) peut être effectuée en fonction de l’âge des animaux ou en fonction de leur taille. Dans le premier cas plusieurs méth- odes d'appréciation peuvent être envisagées, soit après un temps donné (5 à 6 mois par exemple), soit à la mort des témoins, soit à la mort des animaux expérimentaux. Selon nous, la comparaison du potentiel repro- ducteur des animaux expérimentaux et témoins ne doit être effectuée qu'après leur mort pour prendre en compte, entre autre, le problème de la longévité. Dans le second cas la fécondité cumulée (F) peut être exprimée en fonction de la taille (T) sous la forme F = aTz selon la formule proposée par Levina (1973) où z est la pente de la droite log (F) = z log (T) + log (a). Période de ponte L'âge des limnées lors des premières pontes (Tableau 1) varie seulement en fonc- tion de la longueur de la photophase: un retard de trois semaines est observé dans le cas de la photopase courte (LD 7:17). Cette conclusion est en accord avec celle de Bohlken et al. (1978) chez Lymnaea stagnalis mais est différente de celle de de Witt & Sloan (1960) chez Physa pomilia. Dans les condi- tions extrêmes d'intensité lumineuse (16-3, 7- 3) la longeur de la photophase ne semble plus intervenir. La durée de la periode de ponte, qui varie entre 7 et 11 mois sous jours longs, est nette- ment plus longue aussi bien sous jours courts (12 ou 13 mois) qu'à l'obscurité (13 mois). Relativement au témoin lumiere-blanche respectif toutes les possibilités sont ob- servées dans le cas des lumières colorées: raccourcissement de la période de ponte (lumière jaune: 2 mois), allongement faible (lumiére bleue et verte: 1 mois) ou important (lumiére U.V.: 4 mois et rouge: 5 mois). Remarquons que, dans tous les cas, les animaux pondent jusqu’a leur mort; les differ- ences notées ci-dessus sont donc dues à l'in- fluence de la photophase sur leur longévité comme nous l'avons montré par ailleurs (Seugé, 1980; Seugé & Bluzat, 1982b). Influence de la photophase L'évolution de la fécondité mensuelle (f) se traduit par une courbe en cloche (Fig. 1a, b et с); un maximum d'environ 1100 oeufs par 328 mois est atteint (16-1) et en régle generale la fécondité des animaux eleves en jours longs (LD 16:8) est superieure à celle des animaux eleves en jours courts. Au contraire l'examen de la fécondité cumulée (F) fait ressortir que, dans tous les cas, la production d'oeufs est, au total, plus importante en courte photophase (LD 7:17) qu'en longue photophase (Tableau 1): (F) est augmentée de 9.6% (16-1, 7-1), de 33.1% (16-2, 7-2), de 89.5% (16-3, 7-3). Pour l'essentiel, cette augmentation est due à celle du nombre de pontes (P) par animal liée à la plus grande longévité des animaux. Par ailleurs, la Fig. 3 (b et c) permet de constater que, pour une taille donnée, les animaux élevés sous courte photophase pondent moins que les autres; cet effet ne tend à s’effacer que dans la série 7-1 où la valeur anormalement elevee da la pente z (12.9) traduit nettement que l'équilibre mis en evidence entre la croissance et la fécondité (Geraerts, 1976a, b) n'existe plus dans ce cas (Seuge, 1980). Notre méthode d'étude, se basant essen- tiellement sur la fécondité cumulée, peut sans doute expliquer nos divergences avec les auteurs ayant aborde ce problème (Van der Steen, 1967; Bohlken et al., 1978; Dogterom, 1980 et Joosse, 1980). Influence de l'intensité lumineuse Alors qu'aucun effet n'est enregistré en jour court (LD 7:17), la fécondité cumulée parait directement liée à l'intensité lumineuse en jour long (LD 16:8): —15% (16-1, 16-2), —40% (16-1, 16-4) et —41% (16-1, 16-3). La comparaison (Tableau 1) des séries 16-1, 16- 2, 16-3 et 16-4 indique que la diminution de la fécondité est due à la réduction du nombre de pontes (P) émises par animal liée à une longevite plus faible. La confrontation des résultats des séries 16-3, 7-3 et O (Fig. 3c) montre qu'une in- tensite lumineuse aussi faible que 10 lux est perçue tres différemment de l'obscurité. Cette conclusion rejoint celle de Van der Steen (1967) qui démontre, pour sa part, que les limnées sont sensibles a une intensité lumi- neuse de 1 lux. Influence de l'obscurité La fécondité mensuelle est caractérisée par des valeurs demeurant inférieures à 500 oeufs par animal pendant toute la période de SEUGE ET BLUZAT ponte. La fecondite cumulée des animaux 0 est toujours faible: environ —42% en moyenne relativement aux animaux éleves sous courte photophase. Cette diminution est due spécifiquement à un nombre plus réduit de pontes (P) puisque, dans ce cas, une dif- ference de longévite ne peut être invoquée. De plus les oeufs pondus et incubes ont une fertilité significativement abaissee (Tableau 2); L'influence de l'obscurite sur la fécondité des Gasteropodes a deja ete envisagée: elle semble tres variable: elle est sans effet chez Australorbis glabratus (Deschiens & Byan, 1956) et chez Bulinus truncatus (Deschiens, 1957, et Gaud, 1958) mais a une action depressive bien marquee chez Bulinus con- tortus (Deschiens & Byan, 1956) et chez Lymnaea stagnalis (Van der Steen, 1967). Un arrêt complet de la ponte n’a, malgré tout, jamais ete observe; ce fait peut etre expliqué par le travail de Roubos (1975) qui démontre que chez des limnees aveuglées l'hormone d'ovulation continue d'être secrétée mais que son rythme d'élaboration est supprime. Influence des lumières colorées Dans l'ignorance des mécanismes récep- teurs en jeu, nous n'avons pas tenu compte, lors de cette première investigation, du rende- ment energetique de chacune des gammes de radiations utilisées (Seliger & McElroy, 1965). Nous nous sommes limités à établir des comparaisons avec des témoins élevés sous lumière blanche (LD 16:8), à des in- tensités lumineuses de même ordre que celles réalisées pour les lumières colorées. L'étude de la fécondité mensuelle (Fig. 2a,b) montre que le pic de fécondité maxi- male correspond rigoureusement à celui du témoin dans le cas de la lumière jaune; ce pic est plus tardif pour les animaux élevés sous lumières U.V. et rouge. Dans le cas des lumières bleue et verte la courbe est atypique. Le niveau des moyennes mensuelles maxi- males (jusqu'à plus de 1400 oeufs par limnée dans les cas U.V. et vert) souligne que nos conditions d'élevage ne génent en rien l'ex- pression de la fécondité des animaux. L'étude de la fécondité cumulée en fonction de Гаде (Figs. 2c,d, За) permet de conclure qu'à l'exception du groupe élevé sous lumière jaune la production d'oeufs est plus forte sous les lumières colorées que celle du témoin respectif: +36% (bleu), +47% (vert), +62% (U.V.) et +77% (rouge); la faible élévation de ÉCLAIREMENT ET POTENTIEL REPRODUCTEUR DE LYMNAEA température observée dans un seul cas (lumière rouge: eau = 21°C) ne peut ex- pliquer ces résultats. Dans deux cas (lumière jaune et lumière verte) le nombre moyen d'oeufs par ponts (Nm), soit diminué de 16% (16-1, J) soit augmente de 16% (16- 2, V), rend compte des variations de la fécondité. Dans les trois autres cas, la fecon- dité plus élevée est liée, au contraire, a un nombre plus important de pontes émises par animal (P) qui, lui-même, peut être expliqué tantôt par une longévité plus importante (U.V. et rouge) tantôt par un effet spécifique (bleu). La Fig. 3d, fécondité en fonction de la taille, fait ressortir d'une part, l'effet spécifique de la lumière bleue (fécondité plus importante pour une taille donnée), d'autre part, que toutes les lumières colorées sont perçues par les limnées (les diverses valeurs de z étant large- ment supérieures à celle qui caractérise l'ob- scurité). Dans la mesure où nous avons établi précédemment, que sous longue photo- phase, une diminution de l'intensité lumi- neuse se traduit par une baisse de la fécon- dité les résultats obtenus dans les élevages sous lumières colorées soulignent l'existence d'un effet spécifique de ces gammes de radia- tions. Précisons par ailleurs, que, si la fertilité des oeufs n'est pas atteinte, la durée de leur embryogenèse est modifiée (Tableau 2) dans trois cas; nos résultats relatifs à la fertilité sont en contradiction avec ceux de Yung (1878) qui signale même l'impossibilité d'obtenir un développement sous lumière verte. Dans l'ensemble de ces expériences seul l'élevage sous lumière jaune a provoqué un fréquence anormalement élevée (8.76%) d'anomalies de l'embryogenese. A notre connaissance deux auteurs seule- ment ont étudié l'influence de la couleur de la lumière sur la fécondité des Gastéropodes: Gorf (1963) chez Viviparus viviparus et Van der Steen (1967) chez Lymnaea stagnalis. D'après le premier auteur les lumières orange et rouge stimulent la ponte alors que, pour le second, l'élevage sous les lumières verte et rouge nentraine aucune modification de la fécondité mais, selon nous, le temps d'obser- vation extrêmement court (3 jours) explique facilement sa conclusion. Nous disposons par ailleurs de quelques informations anciennes sur le comportement de Lymnaea sous des lumières colorées: pour Graber (1884: Franc, 1968: 498, 499) cet animal percoit le bleu mais pas le rouge; selon Liche (1934a,b) la limnée distingue le 329 bleu et le rouge; enfin d'après Carmichael (1933: Franc, 1958: 499) les Pulmonés ne perçoivent pas la lumière U.V., sauf peut être Limax flavus. “Potentiel limnée” Dans le but d'intégrer les effets des condi- tions d'éclairement sur la croissance (Seugé & Bluzat, 1982b) et sur la fécondité (presents résultats), nous avons tente de déterminer le rendement biologique des limnées à la fin de leur vie, ce qui permet d'établir des comparai- sons entre les différents groupes considérés. Dans ce but, nous calculons le produit du vol- ume moyen atteint par la coquille au terme de l'expérience (V) par la fécondité cumulée (F): РА IE: Les valeurs du P.L. (Tableau 1) montrent que, dans nos experiences, les courtes photophases (LD 7:17) sont de ce point de vue de meilleures conditions de vie que les longues photophases (LD 16:8): l'augmenta- tion enregistrée est de 53,5% (1200 lux), 111,8% (180 lux) et 300% (10 lux). Une rela- tion entre la valeur du P.L. et celle de l'inten- site lumineuse n'a ри être établie que dans le cas des élevages sous longue photophase; il augmente alors de 36% entre 10 et 180 lux et de 26,5% entre 180 et 1200 lux. Les limnées élevées à l'obscurité, malgré leur faible fécondité, ont un rendement biologique (5.3 x 107) intermédiaire entre celui des animaux maintenus sous longue photophase (2.5 x 107) et celui des animaux élevés sous courte photophase (7.5 x 107). Ces résultats nous conduisent à conclure que l'élevage à l'ob- scurité ne peut être rapproché d'un élevage sous faible intensité lumineuse ni en photo- phase longue, ni en photophase courte. Dans le cas des lumières colorées, nous constatons, par rapport au témoin respectif, une augmentation des valeurs du P.L.: de 12% (bleu), 41% (vert), 112% (U.V.) et 158% (rouge); un cas fait exception, celui de la lumière jaune où la valeur du P.L. est plus faible que celle de son témoin (— 14%). Ces variations sont imputables soit à un effet dépresseur sur la fécondité (jaune) soit à une augmentation de la croissance et de la fécondité (U.V., vert et rouge), soit enfin à la résultante de deux effets antagonistes (bleu). Différents articles montrent qu'en règle générale les invertébrés sont sensibles aux conditions lumineuses (Charles, 1966; Gardiner, 1972; Lickey et al., 1976; Stoll et al., 1976). Chez les Insectes, celles-ci ont une 330 SEUGE ET BLUZAT influence déterminante sur le cycle biologique (Saunders, 1976) mais les travaux les plus récents n'ont pas encore pu elucider la totalité des mécanismes en jeu. Notre étude souligne l'importance de ces facteurs sur la longévité, la croissance (Seugé & Bluzat, 1982b) et la fécondité (présents résultats) du Mollusque Lymnaea stagnalis élevé au laboratoire ce qui peut con- tribuer à faciliter la compréhension du com- portement de cette espèce dans son milieu naturel. REMERCIEMENTS Nous remercions le Professeur J. Gener- mont pour ses conseils, Madame O. Jonot pour la réalisation de l'illustration et Madame E. Jézéquel pour la dactylographie de cet article. TRAVAUX CITÉS BLUZAT, R., JONOT, O., LESPINASSE, G. & SEUGE, J., 1979, Chronic toxicity of acetone in the fresh water snail Lymnaea stagnalis. Toxi- cology, 14: 179-180. BLUZAT, R. & SEUGE, J., 1981, Effets à long terme de quatre détersifs chez le Pulmoné d'eau douce Lymnaea stagnalis L.: intoxication des animaux des l'éclosion. Environmental Pollution, 25: 105-122. BOHLKEN, S., ANASTACIO, S., LOENHOUT, H. VAN & POPELIER, C., 1978, The influence of day length on body growth and female reproduc- tive activity in the pond snail Lymnaea stagnalis. General and Comparative Endocrinology, 34: 109. CHARLES, G. H., 1966, Sense organs. In WIL- BUR, К. М. & YONGE, С. M., eds., Physiology of Mollusca. Academic Press, N.Y. and London, 2: 455-521. DESCHIENS, R., 1957, Sur la perpétuation des élevages de mollusques vecteurs de Bilharzio- ses à l'obscurité. Bulletin de la Société de Path- ologie Exotique, 50: 229-233. DESCHIENS, R. 4 BYAN, H., 1956, Comportement d'élevages de mollusques vecteurs de Bilharzio- ses à l'obscurité. Bulletin de la Société de Path- ologie Exotique, 49: 658-661. DOGTEROM, G. E., 1980, Spontaneous and neu- rohormone induced ovulation in Lymnaea stagnalis kept under various experimental con- ditions. Haliotis, 10: 49. FRANC, A., 1968, Sous-classe des Pulmones. In GRASSE, P. P., ed., Traité de Zoologie, 5(3): 325-607, Masson, Paris. GARDINER, M. S., 1972, The biology of Inverte- brates. McGraw Hill, New York, 954 p. GAUD, J., 1958, Rythmes biologiques des mol- lusques vecteurs des Bilharzioses. Facteurs saisonniers et climatiques influencant le cycle de reproduction de Bulinus truncatus et de Planor- barius metijensis en Afrique du Nord. Bulletin of the World Health Organization, 18, 751-769. GERAERTS, W. P. M., 1976a, Control of growth by the neurosecretory hormone of the Light Green Cells in the fresh water snail Lymnaea stagnalis. General and Comparative Endocrinology, 29: 61-71. GERAERTS, W. P. M., 1976b, The role of the Lat- eral Lobes with control of growth and reproduc- tion in the hermaphrodite fresh water snail Lymnaea stagnalis. General and Comparative Endocrinology, 29: 97-108. GORF, A., 1963, Die Einfluss des sichtbaren Lichtes auf die Neurosekretion der Sumpfdeckel- schnecke Vivipara vivipara L. Zoologische Jahrbucher, Abteilung fur allgemeine Zoologie und Physiologie der Tiere, 70: 266-277. JOOSSE, J., 1980, Endocrinology of reproduction in Basommatophoran Pulmonate snails. Haliotis, 103572: LEVINA, O. V., 1973, Fécondité des Moilusques d'eau douce Lymnaea stagnalis et Radix ovata. Zoologicheskii Zhurnal, 52: 676-684. En Russe. LICHE, H., 1934a, Les réactions photiques de la Lymnaea stagnalis L. Comptes rendus Acadé- mie de Cracovie 7: 5. LICHE, H., 1934b, Uber die photischen Reaktionen bei der Schlammschnecke Lymnaea stagnalis L. Bulletin International Académie Polonaise des Sciences, Classe des Sciences mathématiques et naturelles, ser. B Il, 5-7: 233-249. LICKEY, M. E., BLOCK, G. D., HUDSON, D. J. 8 SMITH, J. T., 1976, Circadian oscillators and photoreceptors in the Gastropod, Aplysia. Photo- chemistry and Photobiology, 23: 253-273. ROUBOS, E. W., 1975, Regulation of neurosecre- tory activity in freshwater pulmonate Lymnaea stagnalis L. with particular reference to the role of the eyes. A quantitative electromicroscopical study. Cell and Tissue Research, 160: 291-314. SAUNDERS, D. S., 1976, Insect clocks. Perga- mon, Oxford, 279 p. SELIGER, H. H. & MCELROY, W. D., 1965, Light: physical and biological action. Academic Press, New York, London, 417 p. SEUGE, J., 1980, Ecophysiologie de Lymnaea stagnalis: étude des effets des insecticides de synthèse. Thèse, Université de Paris-Sud, France, 227 p. SEUGE, J. & BLUZAT, R., 1979, Toxicité chronique du Carbaryl et de Lindane chez le Mollusque d'eau douce Lymnaea stagnalis L. Water Re- search, 13: 285-293. SEUGE & BLUZAT, R., 1982a, Influence d'une in- toxication par le Lindane en fonction de la durete de l'eau chez Lymnaea stagnalis. Proceedings ÉCLAIREMENT ET POTENTIEL REPRODUCTEUR DE LYMNAEA 331 of the Seventh International Malacological Con- gress; Malacologia, 22: 15-18. SEUGE, J. & BLUZAT, R., 19825, Effets des condi- tions d'éclairement sur la croissance de Lymnaea stagnalis (Gastéropode Pulmoné). Malacologia, 23: 55-62. SEUGE, J. & BLUZAT, R., sous presse, Chronic toxicity of three insecticides (Carbaryl, Fenthion and Lindane) in the fresh water snail Lymnaea stagnalis. Hydrobiologia. 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. STOLL, G. J., SLOEP, P., DUIVENBODEY, V. VAN & WOUDE H. A. VAN DER, 1976, Light sensitivity in the pulmonate gastropod Lymnaea stagnalis: peripherally located shadow-receptors. Pro- ceedings; Koninklijke Nederlandsche Akademie van Wetenschappen, 79: 510-516. WITT, R. M. DE & SLOAN, W. C., 1960, The initia- tion of ovoposition in fresh water snails. Ana- tomical Record, 137: 349. EFFECTS OF LIGHT CONDITIONS ON THE REPRODUCTIVE POTENTIAL OF LYMNAEA STAGNALIS (GASTROPODA: PULMONATA) J. Seugé and R. Bluzat ABSTRACT The fecundity of the snail Lymnaea stagnalis reared under short and long photophases, various intensities, five coloured lights and darkness was studied. The overall fecundity of snails reared under short photophase (LD 7:17) was always greater (9.6% to 89.5%) than that of snails reared under long photophase (LD 16:8). This difference could be explained by an increased longevity which led to an increased number of egg masses per snail. The fecundity was not related to the light intensity during short days whereas it was proportional to this factor during long days. The overall fecundity was only reduced (— 13.4%) under the yellow light and was increased under the others: +35.7% (blue), +47.4% (green), +61.8% (U.V.) and +77% (red) when each experimental group was compared to its control. The yellow light excepted, the number of egg-masses laid per snail was always enhanced and the laying period was increased signifi- cantly under U.V. and red lights. Our results show that the five coloured lights were discerned by the snails. Rearing in continu- ous darkness induced a decrease of the fecundity due essentially to a specific effect on the number of egg-masses laid per snail. The “Lymnaea Yield”: P.L. (shell volume x overall fecundity) was higher when the snails were reared during short days. The P.L. was proportional to the light intensities studied only in the snails reared during long days and it reached a paradoxical value in the snails reared in dark- ness. The P.L. was reduced under the yellow light and in the four other cases of coloured light it was increased; this points out a specific effect of these coloured lights. MALACOLOGIA, 1983, 23(2): 333-349 DONNÉES ANALYTIQUES SUR LES CONCRÉTIONS DU TISSU CONJONCTIF DE QUELQUES GASTEROPODES D'EAU DOUCE Micheline Martoja! et Michel Truchet? RÉSUMÉ Les concretions du tissu conjonctif de cinq espèces de Gastéropodes Prosobranches et Pulmonés d'eau douce ont été analysées par spectrographie des rayons X, émission ionique secondaire et spectrométrie Raman, sur coupes histologiques de ба 10 um d'épaisseur. Le carbonate de calcium a été identifié chez Valvata cristata et Planorbarius corneus, le phosphate de calcium chez Viviparus viviparus et Bithynia tentaculata; de petits cristaux a carbonate de calcium et des sphérocristaux à phosphate coexistent chez Lymnaea peregra. La composition peut être simple (calcite pure chez Va/vata) ou complexe (quatre sels et six éléments décelés chez Lymnaea). Elle est sans rapport avec la position systématique des espèces. Mots-clés: Gasteropodes; conjonctif; concretions; calcium; microanalyse. INTRODUCTION Depuis leur découverte par Barfurth (1881), les spherocristaux du tissu conjonctif des Gastéropodes, qui peuvent être groupés autour des artères ou dispersés dans le manteau, la masse viscérale et le pied, sont considérés comme étant formés de carbonate de calcium. A cet égard, il est classique de les opposer à ceux de la glande digestive qui sont constitués de phosphate de calcium (voir Manigault, 1960). La presence de carbonate de calcium a été confirmée depuis lors (Kapur & Gibson, 1968; Timmermans, 1969; Rich- ardot & Wautier, 1972; Sen Gupta, 1977), mais les méthodes d’analyse plus précises montrent que leur composition élémentaire peut être complexe (Tompa & Watabe, 1976; Sminia et al., 1977; M. Martoja et al., 1980). D'autre part, le tissu conjonctif fait parfois fonction de rein d’accumulation et ses con- cretions sont alors de nature purique. L'exemple de Pomatias pourrait n'être pas unique; il en irait de même des cellules inter- Stitielles de la glande digestive d’Helix (Abolins-Krogis, 1960) et du tissu conjonctif péri-rénal de Viviparus (Andrews, 1979). Il nous a donc paru interessant de repren- dre l'analyse de ces sphérocristaux par des procédés physiques récemment adaptés à l'étude de coupes de tissus animaux, Comme exemple, nous avons choisi des Gastéro- podes d'eau douce. C'est, en effet, chez ceux-ci que les concrétions du tissu conjonctif sont les plus spectaculaires, ainsi que Га signalé Cuénot (1899). MATÉRIEL ET MÉTHODES La durée des analyses spectrales à la microsonde Raman et la disponibilité encore très limitée de l'appareil ne nous ont pas permis d'examiner un grand nombre d'ani- maux. La recherche de variations affectant les bioaccumulations n'a donc pas ete abordée. Nous avons préféré multiplier les espèces et notre étude a porté sur trois Prosobranches Mésogasteropodes et deux Pulmonés Basommatophores récoltés et fixés dans les conditions suivantes: Viviparus viviparus (L.) (Viviparoidea), 1 individu fixé au formol, récolté en septembre dans le canal du Nivernais; Valvata cristata Mull. (Valvatoi- dea), 2 individus fixés l’un au formol, l’autre par le mélange de Carnoy, récoltés fin octobre dans le Leman; Bithynia tentaculata (L.) (Rissooidea), 2 individus fixes par le melange de Carnoy récoltés l'un en juin dans la rivière Saône, l’autre fin octobre dans le Léman, Lymnaea (Radix) peregra (Mull.) (Lymnaeoidea), 2 individus fixés par le mélange de Carnoy, récoltés dans le Léman, Рип en juillet, l’autre en octobre; Planorbarius (Coretus) corneus (L.) (Planorboidea), 2 individus fixés l’un par le mélange de Carnoy, Institut Oceanographique, 195 rue Saint Jacques, 75005 Paris, France. Laboratoire d'Histophysiologie fondamentale et appliquée, 12 rue Cuvier, 75005 Paris, France. 1-2Equipe de Recherche Associée n° 570 du C.N.R.S. (333) 334 l'autre par le glutaraldéhyde, récoltés en septembre dans une mare de l'Ile de France. (Nomenclature et orthographe conformes à celles qui figurent en Macan, 1969.) Les animaux ont été fixés sitôt après leur capture, puis inclus à la paraffine et débités en coupes de 6 à 10 um d'épaisseur. Les sphérocristaux ont été repérés sur des coupes colorées par le rouge solide-picro- indigocarmin, voisines de celles qui ont été analysées par procédés physiques. Dans cer- tains cas, nous avons fait appel à des méthodes histochimiques: la coloration par la laque d'alizarine (Dahl et McGee-Russell) et la méthode de substitution par l'argent (Von Kossa) ont été employées pour la mise en évidence des accumulations calciques. Les méthodes au ferricyanure ferrique (Schmorl) et à l'argentométhénamine (Gomori) ont été utilisées pour la recherche des déchets puri- ques (voir Pearse, 1972, pour l'exposé de ces méthodes). Des examens ultrastructuraux ont été pratiqués sur des poumons de Planorbe fixés au glutaraldéhyde (2% tampon phosphate pH 7,4) et post-osmiés (1% tampon phosphate pH 7,4); les coupes ont été contrastées à l'uranium-plomb. Trois procédés d'analyse physique ont été appliques: 1. Analyse par spectrographie des rayons X Les coupes ont été analysées par spectro- graphie dispersive en longueur d'onde, avec une microsonde CAMECA MS 46, dans les conditions suivantes: tension d'accélération, 15 kV; intensité de la sonde, 50 nA; cristaux K.A.P. (Potassium acid phtalate) et P.E.T. (pentacrythritol). 2. Analyse élémentaire par emission ionique secondaire Utilisant la pulvérisation cathodique et la spectrométrie de masse, la méthode permet de caractériser et de localiser les éléments et leurs isotopes à l'échelle de la microscopie photonique (Castaing & Slodzian, 1962; Truchet, 1975). Dans un appareil CAMECA SMI 300, équipé d'un filtre électrostatique, les échantillons ont été bombardés à oxygène O>*, sous 5,5 kV et 7,5 мА avec une forte defocalisation (30/10). Le diaphragme de la lentille était de 200 um, le diaphragme de champ de 3700 um, la bande passante de 10V et le convertisseur réglé pour un gros- MARTOJA ET TRUCHET sissement de 115 environ. La résolution spatiale était de 1 um environ et la résolution en masse de 300. L'analyse spectrale a porté sur l'émission positive de 1 à 240. Les images ont ete faites sur les alcalins, Na et K, et sur les alcalino-terreux, Mg, Ca, Sr et Ba. 3. Analyse moléculaire par trometrie Raman microspec- Constitué d'un laser, d'un microscope optique et d'un monochromateur (Delhaye & Dhamelincourt, 1975), Гарргей ultilise l'effet Raman. Il permet de determiner la nature moléculaire de tout composé suffisamment concentré dans des volumes de quelques ums. Nous avons employe l'appareil JOBIN- YVON MOLE 77, équipé d'un laser a argon ionise. Les raies 457,9 пт, 488nm et 514 nm ont servi de radiations excitatrices. Dans tous les cas, l'analyse a été faite avec un objectif 100 de grande ouverture numer- ique afin de standardiser les conditions de collecte du flux et de recueillir le maximum de lumière. Ce mode de collection dit “rétro- Raman a grand angle solide” ne permet pas d'effectuer des mesures de polarisation. Le diaphragme de champ a servi a délimiter le volume analysé et a réduire la fluorescence parasite. La largeur des fentes du mono- chromateur, les réglages du compteur de photons et de l'enregistreur, les vitesses de défilement, adaptés a chaque cas, sont pré- cises dans les légendes des spectres. Les figures 1, 2 et 3 représentent les spectres Raman de trois échantillons de référence. Kkk kk Figures 1 à 3, 5 à 8 et 10 et 11 (spectres Raman): L'axe des abscisses représente le nombre d'onde (fréquence), compté en ст- 1 depuis la raie excitatrice. L'axe des ordonnées représente l'intensité de la raie et du fond de lumière parasite, exprimée en nombre de photons par seconde (cps). RÉSULTATS 1° Viviparus viviparus Le tissu conjonctif qui entoure le complexe réno-péricardique renferme de gros sphéro- cristaux isolés (Fig. 4A). Selon Cuénot (1899), ils seraient formés de carbonate de CONCRÉTIONS DU TISSU CONJONCTIF 335 FIG. 1. Phosphate de calcium (spectre Raman): produit pur. Excitation: 514.5 nm; 600 mW (filtrée). Fentes: 750 um. Comptage: 104 cps, 4s. Enregistrement: 100 mV; 20 ст-1/тп et 100 cm-1/cm (sans dia- phragme). calcium mais, d’après analyse histochimique, Andrews (1979) estime qu'ils sont constitués de calcium et d’acide urique ou d'urates. Les deux éléments dominants décelables par spectrographie des rayons X en sont le calcium (jusqu’a 4000 chocs/sec.) et le phosphore (jusqu’à 2500 chocs/sec.); il s'y ajoute des traces de magnésium. Analyses aux longueurs d'onde 514,5 et 457,9 nm, les sphérocristaux n'ont montré entre 850 et 1800 cm-1, qu'un pic centré sur 965-970 cm-1, correspondant à la vibration de valence symétrique P-O du phosphate de calcium (Fig. 5). Par sa largeur, la raie indique un état cristallin médiocre: le phosphate serait donc à l’état d'une poudre de fins cristaux. Rien dans la région 1000-1100 ne permet de soupconner la présence de carbonate. De même, la recherche de composés puriques dans la région 575-700 est restée négative. Ces derniers, s'ils sont présents ne peuvent donc être que peu concentrés. La fluores- cence est assez forte; elle atteint son mini- mum, 2500 coups/secondes, après une heure d'irradiation environ. Centrée vers le rouge, elle est très étale et demeure impor- tante à 457,9 nm. Il n’est donc pas exclu que les sphérocristaux referment d'autres com- posés moins concentrés. Les divergences entre nos résultats et ceux d’Andrews (1979) nous ont conduits à recher- cher les déchets puriques par voie histo- chimique chez d'autres individus fixés par le formol ou le mélange de Carnoy. Cette détec- tion n’a jamais donné de réaction franche- ment positive. Tous les animaux ayant été autopsiés a la fin de l'été, une décharge périodique pourrait être à envisager. Au contraire, dans tous les cas, le calcium a pu être révélé par ce même type de méthodes. 2° Valvata cristata Le céphalopodium et la masse viscérale contiennent une énorme quantité de petits sphérocristaux souvent groupés. Leur réparti- tion, beaucoup plus vaste que dans les autres espèces peut même s'étendre aux masses musculaires. Leur composition ne semble pas avoir été étudiée jusqu'à present. L'analyse en a été faite a 514,5 et 488 nm (Fig. 6). Après 15 minutes d'irradiation, le fond de fluorescence n'est plus que de 350 coups/sec. et tombe en moins d'une heure au taux le plus bas jamais observé sur coupe histologique: 80-100 coups/sec., pour 750 mW au laser. Dans ces conditions, chaque 336 MARTOJA ET TRUCHET | 1080 | (12000 eps FIG. 2. Carbonate de calcium (spectre Raman): produit pur cristallisé en Calcite. Excitation: 514.5 nm; 600 mW (filtrée). Fentes: 750 ит. Comptage: 104 cps, 1s. Enregistrement: 100 mV; 50 ст-1/тп et 100 cm- 1/cm (sans diaphragme). concrétion analysée fournit un spectre identi- que a celui d'un échantillon témoin de calcite, sans aucune autre raie. L'enregistrement ef- fectué entre 2800 et 3400 cm-1, dans la re- gion des liaisons C-H non spécifiques, démontre l'absence de matière organique. Ces concrétions sont donc formées de calcite pure. L'accumulation de calcite est sujette à des variations individuelles. En effet, dans le second exemplaire récolté le même jour au même point, les sphérocristaux se présen- CONCRÉTIONS DU TISSU CONJONCTIF 337 } } | | | 6000 | | | | 4000 710 If 200 N | 1090 (14500 cps} Se ne = AAA SSE ART aa ee === A A A ee | | A BEE EEE FIG. 3. Carbonate de calcium (spectre Raman): produit pur cristallise en Aragonite. Conditions d’enregistre- ment identiques à celles de la calcite. taient sous forme de “fantômes” d’ailleurs bien visibles; le calcium n’a pu y être mis en evidence par histochimie. Aucune raie n'a ete obtenue sur le spectre Raman, ce qui indique que les concretions sont vides ou qu'elles ne renferment qu'un mélange de substances trop peu concentrées pour fournir un signal. 3° Bithynia tentaculata Nous avons observé dans le pied, des amas de tout petits cristaux sous-jacents au tégument et de sphérocristaux plus gros, à l'intérieur de la masse musculo-conjonctive (Fig. 4b). Ces derniers avaient été vus par Garnault (1887) qui les considérait comme homologues de ceux de Pomatias elegans, mais constitués de carbonate de calcium. Dans les deux cas, la spectrographie des rayons X a décelé du calcium et du phos- phore. Dans les sphérocristaux les teneurs ponctuelles atteignent 3000 chocs/sec. pour le calcium et 1500 chocs/sec. pour le phos- phore. 338 MARTOJA ET TRUCHET FIG. 4. Aspect morphologique des sphérocristaux. A. Viviparus viviparus (formol; réaction au ferricyanure ferrique; 230). En cartouche, détail d'un sphérocristal (x 550). В. Bithynia tentaculata (Carnoy; coloration par la laque d'alizarine; x550). С. Planorbarius corneus (glutaraldéhyde-osmium; uranium-plomb; x 2600). En haut à gauche, apparaît un sphérocristal en partie vidé contrastant avec celui qui occupe le reste du cliché. 339 CONCRETIONS DU TISSU CONJONCTIF ‘| WO 099 E Je 629 e sesuajul seres sine] Jed uewey эщешодоэ4$ ua sesiajoeses ‘sanbund sjeyoap ap ээиэзае.| jueuouw | WO 00/ e 6/5 чобэн :ayonoyes ug (sua, эшбелцае!р) wo/, — шо Oz je UW/, WOO} ‘AW 005 ‘uewensIBeiuz ‘$ y ‘sdo yO! :ebeydwog “ur 009 :sajuay (831114) Мш OZ ‘WU 6'/spy :иоцеуохз “¡eual-ued jjouoluos nssy np xnejsuoolayds :(uewey эдээ4$) smedinA пива ‘6 ‘914 340 800+ +— 400 200 MARTOJA ET TRUCHET FIG. 6. Valvata cristata (spectre Raman): amas de sphérocristaux. Excitation: 488 nm; 600 mW (filtrée). Fentes: 700 um. Comptage: 103 cps, 1 $. Enregistrement: 100 mV; 50 cm 1/mn et 100 cm 1/cm (dia- phragme ferme). L'analyse Raman s’est montrée difficile en raison de la fragilite du materiel et d’une fluo- rescence importante. Les radiations 457,9 nm et 488 nm, absorbées, n'ont pu être utilisées. L'irradiation a 514.5 nm a dû être modérée: a 900 cm-1, le fond de fluorescence reste de 1500 coups/sec., pour une puissance au laser de 150 mW seulement. Dans ces condi- tions, le rapport pic/bruit de fond est mauvais. La région explorée va de 900 cm-1 à 1100 pour les amas de petits cristaux et de 500 à 1300 pour les sphérocristaux. Les petits cristaux ne donnent qu'un faible signal vers 960-970 ст-1 et les sphérocristaux, un spectre un peu mieux résolu où apparaissent trois signaux а 965, 1070 et 1140 ст-1 (Fig. 7, A et B). Le pic situé à 970 cm-1 qui se retrouve CONCRÉTIONS DU TISSU CONJONCTIF 341 4000 1060? ie 965-970 iit 4 FIG. 7. Bithynia tentaculata (spectre Raman): A. Petits cristaux sous-tégumentaires. B. Sphérocristaux de la zone musculo-conjonctive du pied. Excitation: 514.5 nm; 150 mW. Fentes: 600 um. Comptage: A: 104 cps, 1 s.; В: 103 cps, 4 $. Enregistrement: A: 500 mV, 20 cm- 1/mn et 20 cm- 1/cm. В: 500 mV, 20 cm {/mnet 50 cm-1/cm. dans les deux types d'accumulation, corre- spond à la raie principale du phosphate de calcium; comme chez Viviparus, son étale- ment montre que les cristaux sont de petite taille. Les raies 1140 et 1070 sont difficiles à interpréter. La réponse à 1140, déjà ren- contrée sur d’autres échantillons riches en phosphate de calcium, pourrait traduire la présence de la trame de cristallisation ou de sites de nucléation et correspondre à une vibration de valence, C-O par exemple. La raie 1070 pourrait être due à un nitrate ou à un carbonate de calcium, bien qu'elle ne corresponde exactement ni à l’un (1054) ni à l’autre (1088). Toutefois, c'est la presence de carbonate qui est la plus probable puisque, sur des spectres d'os, le pic de la calcite qui accompagne le phosphate est décalé vers 1080 (R. Martoja et al., 1981). Il est possible, en effet, comme le suggère le pic 1140, qu'un autre composé soit lié a ces sels et qu'une interaction moléculaire provoque un léger déplacement des modes caractéristiques, même pour les vibrations de valence. De tels déplacements ont déjà été observés pour les phosphates (Daudon et al., 1980). L'absence de tout signal dans la région des 600 cm-1 permet d'affirmer que les déchets puriques ne représentent pas les composants essentiels des sphérocristaux. Les sphérocristaux peuvent être vidés de leurs sels minéraux. Ainsi, chez l'animal autopsié en juin, la teneur ponctuelle en calci- um déterminée par spectrographie des rayons X n'excédait pas 30 chocs/sec. alors que les sphérocristaux restaient visibles a l'examen histologique sous forme de “fantômes.” Les conclusions de Garnault (1887), ob- tenues au moyen de méthodes microchimi- MARTOJA ET TRUCHET 342 (eue, эшбелцае!р) wo/, wo 06 Ja иши; _W9 OZ ‘AW 001 зиэшедебелиз ‘s y ‘sdo yO! :ebe}dwog ‘шт 009 :sajuay (991114) MW 00Z ‘WU SSL py :uoNeyax3 ‘paid np aaouofuoo-o¡nosnu auoz ej эр xneysuoo1euds :(uewey эдээ4$) елэбэлэа eaeUWAT 'g ‘014 ms aed CONCRÉTIONS DU TISSU CONJONCTIF 343 FIG. 9. Lymnaea peregra (images d'émission ionique): sphérocristaux de la zone musculo-conjonctive du pied. A. Calcium (Ca). B. Strontium (Sr). C. Baryum (Ba). Noter la forte émission de calcium par les 3 spherocristaux, et les légères différences de localisation entre Sr et Ba au sein des deux sphérocristaux situés en haut de la figure. Intensités d'émission ionique (en Ampères: A) et temps de pose (en secondes: $): 40Ca+: 1,5.10- 14 A; 20 $. 88Sr+: 10-16 A; 15 mn (900 s). 138Ва+: 4.10- 17 A; 15 тт (900 $). ques très rudimentaires, ne sont donc pas confirmées. Le phosphate de calcium est le constituant dominant des sphérocristaux. 4° Lymnaea peregra Nous avons analyse les concretions de la zone musculo-conjonctive du pied ou nous avons retrouvé les sphérocristaux et les mini- sphérules en poussière que Prenant (1924) avait signalées chez Lymnaea stagnalis. Dans cette dernière espèce, les sphérocrist- aux de la masse viscérale ont fait l’objet d'une étude récente (Sminia et al., 1977). Par spectrographie des rayons X, nous n'avons détecté dans les mini-sphérules que du calcium (1000 chocs/sec.) et du phos- phore (40 à 120 chocs/sec.) alors que les sphérocristaux ont montré la composition suivante: Ca, 4000 à 8000 chocs/sec.; P, 400 à 1000 chocs/sec.; S, 20 chocs/sec.; Mg, 10 chocs/sec.; Ba, 5 chocs/sec. En raison d'une fluorescence importante, les sphérocristaux n'ont pu être analysés qu'après une longue irradiation et sous une illumination de forte intensité (700 nW). De 200 à 1800 cm- 1, douze pics apparaissent sur le spectre (Fig. 8). On peut attribuer sans difficulté les signaux 965 au phosphate de calcium, 1460 a Гоха!ае de calcium, 1045 au nitrate de calcium et 1000 au sulfate de calci- um. Les pics 1320 et 1420, déja observés chez d'autres Invertébres (Ballan-Dufrancais et al., 1979), comme les pics 440, 850 et 1725, pourraient étre dús a la trame organi- que, abondante si Гоп juge par l'intensité de la région 2900 ст-1 qui est celle des hydro- carbures non spécifiques. Rien ne signale l'existence de déchets puriques dans la zone des 600-650 cm-1. L’analyseur ionique décèle les quatre alcalino-terreux, Mg, Ca, Sr et Ba, mais les plus fortes intensités d'émission restent celles du calcium (Fig. 9, A, B, C). Le strontium et le baryum se rencontrent d’ailleurs partout ou le calcium est abondant, en particulier dans les zones pigmentées du tegument. A noter que le strontium, trop peu concentré, échappe a la microsonde alors que l’analyseur ionique le met facilement en évidence. Quoi qu'il en soit, les deux méthodes confirment la complexite des sphérocristaux qu'indiquent les nom- breuses raies Raman. Malgré une forte fluorescence, l'analyse 344 MARTOJA ET TRUCHET 8000 6000 FIG. 10. Lymnaea peregra (spectre Raman): petits cristaux de la zone musculo-conjonctive du pied. Excita- tion: 488 nm; 260 mW (filtrée). Fentes: 750 um. Comptage: 105 cps, 1,5s. Enregistrement: 100 mW; 50 ст-1/тп et 100 ст-1/ст (diaphragme fermé). En cartouche: Region 350 a 700 cm! d'un autre point montrant un pic a 460 susceptible d'être attribué à la silice. d’amas de petits cristaux a été réalisée à 488 et 457,9 nm. Dans les deux cas, les pics suivants sont apparus; 275, 1080, 1375 et 1595 cm- 1. Sur plusieurs spectres, un pic a pu être caractérisé à 710 cm 1 et, en un seul point, un pic bien défini s'est manifesté a 460 cm- 1. Les réponses a 275, 710 et 1080 cm-1 caractérisent le carbonate de calcium sous forme de calcite. Par leurs positions en fréquence, leurs formes et leurs intensités relatives, les pics a 1375 et 1595 cm-1 cor- respondent au graphite sous une forme assez élaborée (Fig. 10). Enfin, le pic a 460 cm“! pourrait traduire la présence de silice. Aucun autre sel de calcium ou d'un autre alcalino- terreux, aucun déchet purique ne peut étre identifié. En particulier, aucun phosphate n'apparaît sur les spectres bien que la micro- sonde détecte une certaine quantité de phos- phore. La lumière diffusée par les phosphates étant beaucoup plus faible que celle des car- bonates, leur présence n'est pas exclue, mais il ne peut s'agir que de traces. La structure des petits cristaux est donc plus simple que celle des spherocristaux: ils sont formés essentiellement de calcite et de graphite, ce dernier pouvant étre responsable de la couleur jaune-brun des amas cristallins. Comme dans les especes précédentes, les sphérocristaux sont susceptibles de se vider de leur contenu minéral. Dans l'individu autopsié en juillet, l'analyse n’a fourni qu'un tres faible signal vers 970 cm-1 indiquant la présence de phosphate de calcium et quelques signaux vers 2900 cm-1 corre- spondant à la trame organique. Ces sphéro- CONCRETIONS DU TISSU CONJONCTIF 345 1400 1200 1000 800 600 400 200 500 1000 1500 FIG. 11. Planorbarius corneus (spectre Raman): sphérocristaux de la paroi du poumon. Excitation: 514.5 nm; 350 mW (filtrée). Fentes: 600 um. Comptage: 104 cps, 0,6 s. (diaphragme fermé). En cartouche: Autre concrétion, région de 150 à 4200 cm” 1, montrant le maximum de fluorescence (4000 cps) à 610 nm, ou 3100 cm- 1, dans l'orange. cristaux doivent garder une composition très complexe, mais aucune des substances présentes n'est assez concentrée pour fournir un signal détectable en Raman simple. A l'examen histologique, ils présentent un aspect éclaté différent des fantômes ren- contrés dans les autres espèces. Nous n'avons pas observé de variations au niveau des amas de petits cristaux. L'originalité de Lymnaea peregra réside en 346 une ségrégation des sites de précipitation des phosphates et des carbonates. Dans les deux cas, le présence de déchets puriques est à écarter. Notre résultat global concorde avec celui de Sminia et al. (1977) pour L. stagnalis. Utilisant des méthodes chimiques, ces auteurs ont pu caractériser les ions calcium, magnésium, carbonate et phosphate que nous avons nous aussi identifiés. Alors qu'ils les situent dans un seul type de bioaccumula- tion complexe, l'analyse in situ nous a permis de les rapporter à deux catégories distinctes de concrétions, ce qui confirme les observa- tions de Prenant (1924). L'existence de graphite dans les petits cristaux échappe actuellement à toute inter- prétation. Le seul cas analogue relaté à ce jour est celui d'un Poisson (Delhaye et al. 1979). 5° Planorbarius corneus L'abondance exceptionnelle de “carbonate calcique” dans la paroi du poumon du Planor- be a été mentionnée par Cuenot (1899). Chez un autre Planorbidae, Helisoma duryi eudiscus, le carbonate de calcium a été iden- tifié par méthode histochimique dans les con- crétions du pied (Kapur & Gibson, 1968). L'analyse des sphérocristaux de la paroi du poumon par spectrographie des rayons X a décelé du calcium (2800 a 3000 chocs/sec.), du potassium (10 chocs/sec.) et du soufre (5 chocs/sec.). La spectrométrie Raman a été effectuée à 514,5 пт; la fluorescence, centrée sur le rouge-orange, est peu importante. A l'excep- tion d'un spherocristal qui n’a donne aucune réponse, la raie 1088, caractéristique du car- bonate de calcium apparaît constamment et la raie 710 assez souvent (Fig. 11). La vibra- tion de réseau n'a été observée qu'une fois à 282. Aucune autre raie ne se démarque du bruit de fond. En particulier, la région 2800- 3600 cm-1 qui corespond à la matière or- ganique est vide. Cependant, un éclat de calcite pure de même taille que les sphéro- cristaux a donné un signal 20 fois plus intense а 1088 dans des conditions d'enregistrement identique. Conformément aux données bibliograph- iques, les sphérocristaux du Planorbe con- tiennent du carbonate de calcium. Dans l’un, la calcite a pu être identifiée. Dans les autres, compte tenu des intensites des pics 1088 et 710, la raie de la calcite aurait dú être ob- servée: son absence indique qu'il pourrait MARTOJA ET TRUCHET s'agir d'aragonite dont la raie vers 210 est moins intense. Contrairement a Valvata, le carbonate de calcium nest pas à l’état pur et la comparaison avec l'échantillon de refer- ence suggère une concentration de 5%. En l'absence de raie, il n’a pas été possible de determiner la nature des autres constituants mais, d'après l'allure generale du spectre, ils doivent être nombreux et peu concentrés. La microsonde décèle d’ailleurs les éléments К et S. Enfin, la variabilité observée dans les autres espèces existe ici à l’intérieur même de l'individu (Fig. 4C) puisque se côtoient des sphérocristaux vides, d'autres contenant de la calcite et d'autres probablement de l'arago- nite. DISCUSSION— CONCLUSION Nos résultats ne nous permettent pas d'in- tervenir dans le débat concernant le где physiologique des concretions; nous ren- voyons donc aux mémoires de Curtis & Cowden (1979) et de Richardot (1979) pour l'exposé du probleme et la bibliographie. Nous confirmons seulement l'existence d'une variabilité individuelle et d'une diversité des formes minéralogiques d'un même sel. La variabilité individuelle pourrait être due à des conditions alimentaires propres à chaque animal, l'inanition ayant pour effet de vider les sphérules de leur contenu calcique (Vovelle & Grasset, 1979; De With & Sminia, 1980). La diversité des formes minéralogiques ob- servées chez le Planorbe corrobore les ré- sultats de Richardot & Wautier (1972) qui ont montre que le carbonate de calcium amorphe, Гагадопйе et la calcite se rencontraient chez Ferrissia wautieri. Nous n’avons pas observe de differences entre les Pulmonés et les Prosobranches. Or, si le rein des premiers conserve la structure fondamentale de l'organe renal des Gastero- podes (Bouillon & Delhaye, 1970), les trois Prosobranches, Bithynia (Lilly, 1953), Valvata (Cleland, 1954) et Viviparus (Andrews, 1979) se singularisent par un rein modifié qui ne semble pas pouvoir assumer les fonctions d'un rein normal, Il serait donc logique que le tissu conjonctif de ces Prosobranches fasse office de rein d'accumulation et qu'il s'y dépose des composés puriques, comme Га pensé Andrews (1979). Si nous n'y avons décelé que des sels de calcium, nous n’excluons pas l'idée qu'il s'y adjoigne des déchets puriques, a d'autres moments du cycle annuel. On sait que chez Pomatias, les CONCRÉTIONS DU TISSU CONJONCTIF 347 TABLEAU 1. Résumé des principaux resultats. Nombre de Principal sel types de Region Aspect des de calcium concrétions examinée concrétions identifié Viviparus 1 conjonctif spherocristaux phosphate péri-rénal Valvata 1 ensemble du petits carbonate conjonctif sphérocristaux Prosobranches petits cristaux phosphate Bithynia 2 conjonctif du pied sphérocristaux phosphate petits cristaux carbonate Lymnaea 2 conjonctif du pied sphérocristaux phosphate Pulmonés Planorbarius 1 conjonctif sphérocristaux carbonate péri-pulmonaire composés puriques coexistent avec de la calcite et du phosphate de calcium amorphe (Martoja, 1974) et que, chez les Gastéro- podes, il y a une “multiplicité des facteurs internes et externes qui influent sur l'excrétion azotée” (Daguzan, 1980). La composition élémentaire des bioac- cumulation calciques n'est liée ni à l'habitat puisque tous nos animaux sont dulcicoles, ni au rythme saisonnier puisque tous ont été autopsiés la même époque, ni à la position systématique puisque le carbonate et le phosphate de calcium ont été trouvés indif- féremment dans les deux sous-classes. Le cas extrême est représenté par la Limnée, capable d’accumuler à la fois le carbonate et le phosphate dans deux types distincts de concretions. A l'heure actuelle, nous ne sommes pas en mesure d'interpréter ces resultats. La présence de phosphate de calcium dans la glande digestive a été reconnue très tôt, par les zoologistes du siècle dernier. Récem- ment, ce composé a été identifié dans les concrétions du rein de Gastéropodes et de Bivalves (Martoja, 1975; Doyle et al., 1978; Hignette, 1979). Nos résultats présents per- mettent d'ajouter a cette liste, le tissu con- jonctif de certains Gastéropodes dulcicoles. En conclusion, le principe de la spécificité minéralogique valable pour la coquille, ne s'applique pas aux sphérocristaux du tissu conjonctif. D'autre part, il n’y a pas lieu d’op- poser de façon systématique les dépôts cal- caires de la glande digestive ou du rein à ceux du tissu conjonctif, Celui-ci contient certes souvent du carbonate de calcium mais des sphérocristaux de phosphate de calcium auquel sont associés éventuellement d’autres sels, peuvent aussi s'y former. RÉFÉRENCES BIBLIOGRAPHIQUES ABOLINS-KROGIS, A., 1960, The histochemistry of the hepatopancreas of Helix pomatia (L.) in relation to the regeneration of the shell, Arkiv for Zoologi, sér. 2, 3: 159-201. ANDREWS, E. B., 1979, Fine structure in relation to function in the excretory system of two species of Viviparus, Journal of Molluscan Studies, 45: 186- 206. BALLAN-DUFRANCAIS, C., TRUCHET, M. & DHAMELINCOURT, P., 1979, Interest of Raman laser microprobe (MOLE) for the identification of purinic concretion in histological sections. Biologie Cellulaire, 36: 51-58. BARFURTH, D., 1881, Der Kalk in der Leber der Heliciden und seine biologische Bedeutung. Zoologischer Anzeiger, 4: 20-23. BOUILLON, J. & DELHAYE, W., 1970, Histophysio- logie comparée des cellules rénales de quelques Gastéropodes Pulmonés terrestres et dulçaqui- coles. Annales des Sciences Naturelles, Zoolo- gie, sér. 12, 12: 1-26. CASTAING, R. & SLODZIAN, G., 1962, Micro- analyse par émission ionique secondaire. Journal de Microscopie, 1: 395-410. 348 MARTOJA ET TRUCHET CLELAND, D. M., 1954, A study of the habits of Valvata piscinalis and the structure and function of the alimentary canal and reproductive system. Proceedings of the Malacological Society of London, 30: 167-202. CUENOT, L., 1899, L'excrétion chez les Mollus- ques. Archives de Biologie, 16: 49-95. CURTIS, S. K. & COWDEN, R. R., 1979, Histo- chemical and ultrastructural features of the aorta of the slug (Limax maximus). Journal of Mor- phology, 161: 1-22. DAGUZAN, J., 1980, Principales caractéristiques de l'appareil excréteur et de Гехсгейоп azotee chez les Mollusques Gasteropodes. L'Année Biologique, 19: 367-393. DAUDON, M., JAESCHKE-BOYER, H., PROTAT, M. F. & REVEILLAUD, R. J., 1980, La micro- sonde Mole et l'analyse des calculs urinaires. Perspectives et réalités. L'actualité chimique, Compte Rendu des premières journées d'étude sur les applications de la microsonde MOLE, 25-29. DELHAYE, M. & DHAMELINCOURT, P., 1975, Raman microprobe and microscope with laser excitation. Journal of Raman Spectroscopy, 3: 3343. DELHAYE, M., DHAMELINCOURT, P. & WALL- ART, F., 1979, Analysis of particulates by Raman microprobe. Toxicological and Environ- mental Chemistry Reviews, 3: 73-87. DE WITH, N. D. & SMINIA, T., 1980, The effects of the nutritional state and the external calcium concentration on the ionic composition of the haemolymph and on the calcium cells in the pulmonate freshwater snail Lymnaea stagnalis. Proceedings of the Koninklijke Nederlandse Akademie van Wetenschappen, ser. C, 83: 217- 227. DOYLE, L. J., BLAKE, N. J., WOO, C. C. & YEVICH, P., 1978, Recent biogenic phosphorite: concretions in Mollusk kidneys, Science, 199: 1431-1433. GERNAULT, P., 1887, Recherches anatomiques et histologiques sur le Cyclostoma elegans. Actes de la Société Linnéénne de Bordeaux, 41: 10- 158. HIGNETTE, M., 1979, Composition des concre- tions minérales contenues dans les reins de deux mollusques lamellibranches: Pinna nobilis L. et Tridacna maxima (Rodding). Comptes Rendus Hebdomadaires de l'Académie des Sciences, ser. D, 289: 1069-1072. KAPUR, S. P. & GIBSON, M. A., 1968, A histo- chemical study of calcium storage in the foot of the freshwater gastropod, Helisoma duryi eudiscus (Pilsbry). Canadian Journal of Zool- ogy, 46: 987-990. LILLY, M. M., 1953, The mode of life and the struc- ture and functioning of the reproductive ducts of Bithynia tentaculata (L.). Proceedings of the Malacological Society of London, 30: 87-110. MACAN, T. T., 1969, A key to the British fresh- and brackish-water gastropods with notes on their ecology. Freshwater Biological Association, Sci- entific Publication n° 13, ed. 3, 46 p. MANIGAULT, P., 1960, Coquille des Mollusques: structure et formation. In Traité de Zoologie, GRASSE, P. P., ed., Masson, Paris, 5(2): 1823- 1844. MARTOJA, M., 1974, Remarques sur l'excrétion d'un Gasteropode Prosobranche adapté au milieu terrestre Pomatias (= Cyclostoma) elegans. Comptes Rendus Hebdomadaires de l'Académie des Sciences, sér. D, 287: 923-926. MARTOJA, M., 1975, Le rein de Pomatias (= Cyclostoma) elegans (Gasteropode Proso- branche): données structurales et analytiques. Annales des Sciences Naturelles, Zoologie, sér. 12, 17: 525-558. MARTOJA, M., VU TAN TUE & ELKAIM, B., 1980, Bioaccumulation du cuivre chez Littorina littorea (L.) (Gastéropode Prosobranche): signification physiologique et écologique. Journal of Experi- mental Marine Biology and Ecology, 43: 251- 270. MARTOJA, R., MELIERES, F., RAYNAUD, C. & TRUCHET, M., 1981, Données physico- chimiques et cristallographiques sur les calculs rénaux de la nephropathie au chlorure mercur- ique d'un Mammifere, le Lapin. Comptes Rendus Hebdomadaires de l'Académie des Sciences, ser. Ill, 292: 857-862. PEARSE, A. G. E., 1972, Histochemistry theoreti- cal and applied. Livingstone, Edinburgh, 2: 1518 p. PRENANT, M., 1924, Contributions à l'étude du calcaire, |. Quelques formations calcaires du conjonctif chez les Gasteropodes. Bulletin biologique de la France et de la Belgique, 58: 331-380. RICHARDOT, M., 1979, Calcium cells and groove cells in calcium metabolism in the freshwater limpet Ferrissia wautieri (Basommatophora Ancylidae), Malacological Review, 12: 67-78. RICHARDOT, M. & WAUTIER, J., 1972, Les cel- lules a calcium du conjonctif de Ferrissia wautieri (Moll. Ancylidae), Description, minéralogie et variations saisonnières. Zeitschrift fur Zellfor- schung und Mikroskopische Anatomie, 134: 227-243. SEN GUPTA, A., 1977, Calcium storage and distri- bution in the digestive gland of Bensonia monticola (Gastropoda: Pulmonata): a histo- physiological study. Biological Bulletin, 153: 369-376. SMINIA, T., DE WITH, N. D., BOS, J. L., VAN NIEUWMEGEN, М. E., WITTER, M. P. 4 WONDERGEM, J., 1977, Structure and function of the calcium cells of the freshwater Pulmonate snail Lymnaea stagnalis. Netherlands Journal of Zoology, 27: 195-208. TIMMERMANS, L. P. M., 1969, Studies on shell formation in Molluscs. Netherlands Journal of Zoology, 19: 417-528. CONCRÉTIONS DU TISSU CONJONCTIF 349 TOMPA, A. S. & WATABE, N., 1976, Calcified scopie et Biologie Cellulaire, 24: 1-22. arteries in a gastropod. Calcified Tissue Re- VOVELLE, J. & GRASSET, M., 1979, Approche search, 22: 159-172. histophysiologique et cytologique du rôle des TRUCHET, M., 1975, Application de la micro- cellules a sphérules calciques du repli opercu- analyse par emission ionique secondaire aux laire chez Pomatias elegans (Müller) Gastéro- coupes histologiques: localisation des principaux pode Prosobranche. Malacologia, 18: 557-560. isotopes de divers éléments. Journal de micro- ANALYTICAL DATA ON CONNECTIVE TISSUE CONCRETIONS OF SEVERAL FRESH-WATER GASTROPODS Micheline Martoja and Michel Truchet SUMMARY Concretions within the connective tissue of five species of fresh-water prosobranch and pulmonate gastropods have been investigated by X-ray spectrometry, secondary ion emis- sion analysis and Raman spectrometry on paraffin wax sections cut between 6 and 10 um. Calcium carbonate has been characterized in Valvata cristata and Planorobarius corneus, calcium phosphate in Viviparus viviparus and Bithynia tentaculata; small carbonate crystals and phosphate spherocrystals exist together in Lymnaea peregra. The composition may be simple (pure calcite in Valvata) or complex (4 salts and 6 elements found in Lymnaea). К has no relation to the systematic position of the species. MALACOLOGIA, 1983, 23(2): 351-359 BIOCHEMICAL GENETICS OF THE SNAIL GENUS PHYSA: A COMPARISON OF POPULATIONS OF TWO SPECIES Donald С. Buth! and John J. Suloway? ABSTRACT Vertical starch gel electrophoresis was employed to compare two population samples each of Physa gyrina and Physa anatina collected near Urbana, Illinois. Eleven enzyme systems were examined and gene products of 14 presumptive loci were resolved. Six of the loci (aspartate aminotransferase, Aat-A; two esterases, Est-1 and Est-2; glucosephosphate isomerase, Gpi-A; glycerol-3-phosphate dehydrogenase, G-3-pdh-A; and phosphoglucomutase, Pgm-A) were polymorphic. The Aat-A, G-3-pdh-A, and Pgm-A gene products exhibited species-specific electrophoretic patterns. Average heterozygosity values ranged between 5.1 and 6.7%. Intra- specific genetic distances were D = 0.012 or less. The genetic distance between P. anatina and P. gyrina was D = 0.45. Key words: Physidae; Physa; allozymes; electrophoresis; polymorphism; heterozygosity. INTRODUCTION Since the first demonstration by Hubby & Lewontin (1966) and Harris (1966) that ge- netic polymorphism at structural loci can be easily detected by gel electrophoresis, there has been an explosion of information about genic polymorphism in natural populations. Many of the investigations of genetic variabil- ity of aquatic macroinvertebrates have been conducted with snails (Wium-Anderson, 1973; Coker & Kuma, 1974; Narang, 1974; Narang & Narang, 1974, 1976a, 1976b; Ukoli, 1974; Michelson & Dubois, 1975; Nickerson, 1975; Nyman, 1975; Wu & Burch, 1975; Ishay et al., 1976; Logvinenko et al., 1976; Monteiro & Narang, 1976; Chambers, 1977, 1978, 1980; Nyman & Skoog, 1977; Selander et al., 1978; Te, 1978; Davis, 1979; Wurzinger, 1979; Wurzinger & Saliba, 1979; Dillon & Davis, 1980). In most studies utilizing aquatic snails prior to 1970, allozymes were sepa- rated with paper, cellulose acetate, or poly- acrylamide media. In these earlier studies, banding patterns were recorded and com- pared with no attempts to ascertain the ge- netic control of the gene products observed. Most of the recent studies have dealt with species of the genera Bulinus or Biompha- laria as they serve as intermediate hosts for schistosomes and electrophoretic methods can aid in identifying strain and species within these genera. The Physidae are a speciose family of freshwater snails that have been used in vari- ous areas of environmental research such as model ecosystem studies of pesticide de- gradation and accumulation (Metcalf et al., 1971), toxicity testing (Patrick et al., 1968), and water quality assessment (Tucker & Ettinger, 1975). The chaotic taxonomy of the Physidae has recently been treated by Te (1975, 1978) and his work included some electrophoretic analyses. However, several aspects of the population genetics of this im- portant family have yet to be investigated. The major objectives of this study are to: (1) adapt known techniques of vertical starch gel elec- trophoresis for use on Physa, (2) examine gene expression in several enzyme systems in representative species of Physa, and (3) examine genetic divergence among popula- tions of Physa at the intraspecific and inter- specific levels. MATERIALS AND METHODS Snails were collected with dipnets and by hand. Live specimens were transported to the laboratory at ambient temperature, then fro- zen and stored at —20°C. Electrophoretic ex- aminations were completed within three weeks of the initial freezing. Voucher speci- mens from each locality were deposited in the collection of the Illinois Natural History Survey 1Department of Biology, University of California (UCLA), Los Angeles, CA 90024, U.S.A. 2 Aquatic Biology Division, Illinois Natural History Survey, Urbana, IL 61801, U.S.A. (351) 352 (INHS—JS series). The following listing of populations sampled includes the number of individuals examined from each site in paren- theses. Physa anatina Lea. |. Busey Woods Pond, 0.16 km NE of Urbana, Champaign Co., IL, T19N, ROE, NW Ya, NE Ya, NW Ya, Sect. 8 of Thomasboro 72 min. quad., (14) [JS 1002]; |. Pond No. 16 of the Illinois Natural History Survey, 0.8 km S of Champaign at the Natural Resources Annex, Champaign Co., IL, T19N, R8E, SE 14, NE Ya, SE Ya, Sect. 24 of Urbana 7Y2 min. quad (10) [JS 1003]. Physa gyrina Say. |. Busey Woods Pond, same locality as for P. anatina |, (26) [JS 1001]; Il. Boneyard Creek, in Champaign at Second Street, Champaign Co., IL, T19N, R9E, SW Ya, SW Ya, SW Ya, Sect. 7 of Urbana 7¥2 min. quad., (28) [JS 1004]. All soft body parts were dissected from each specimen, mixed with an equal volume of 0.1M Tris-HCI at pH 7.0, mechanically homogenized, and centrifuged at approxi- mately 480 g at 4°C for 15 mins. The super- natant fractions of the whole body extracts were subjected to electrophoresis at 4°C us- ing 14% starch gels (lot #303; Electrostarch Co., Madison, WI 53701). Enzyme systems examined and their histochemical staining conditions are listed in Table 1. Electrophore- tic buffers used include a 0.25M sodium borate buffer at pH 8.6 (Sackler, 1966), a 0.41 M sodium citrate buffer at pH 7.0 or 8.0 (Brewer, 1970), a Tris-citrate buffer at pH 7.0 (Whitt, 1970), and an EDTA-borate-Tris buffer at pH 8.6 (Whitt et al., 1973; Wilson et al., 1973). Electrophoretic conditions were those of Buth & Burr (1978) using vertical starch gel apparatus. BUTH AND SULOWAY Gene products were scored as those of autosomal loci with codominant alleles. Elec- tromorphs of common electrophoretic mobility are assumed to represent homologous allelic products of a given locus. A gene locus was designated as the “A-locus” for a particular enzyme in cases of presumptive single gene control of such enzymes. Multilocus systems were lettered (or numbered in the case of non-specific esterases) based on the mobility of their gene products from cathode to anode. Locus homologies with taxa outside the genus Physa are not reflected in our genetic nomenclature and should not be inferred with- out additional study. Allelic terminology is based on the electrophoretic mobility of the allelic products relative to the origin as used by Selander & Kaufman (1975) and Selander et al. (1978). The reference allele at each locus (=100) was chosen as the most com- mon allele in the population of Physa gyrina from the Busey Woods Pond locality. Cathod- ally migrating allelic products are assigned negative values. Heterozygosity estimates are based on a direct count of heterozygotes (Selander et al., 1978) and the calculation of average hetero- zygosity (Nei, 1978). RESULTS The electrophoretic examination of 11 en- zyme systems allowed the resolution of the gene products of 14 presumptive loci. Six of these loci were polymorphic within or among the populations of P. anatina and P. gyrina. The allele frequencies at these polymorphic loci are given in Table 2; the genotypic distri- TABLE 1. Enzyme systems examined and their electrophoretic requirements. See text for buffer references. Enzyme Electro- commission phoretic Staining Enzyme number Locus buffer reference Acid phosphatase 311132 Acph-A Tris-citrate Shaw & Prasad (1970) Alkaline phosphatase SAN Akph-A Tris-citrate Shaw 8 Prasad (1970) Aspartate aminotransferase 2.6.1.1 Aat-A Citrate, pH8 Shaw & Prasad (1970) Esterase — Est-1,2 EBT Brewer (1970) Glucosephosphate isomerase eS, Gpi-A EBT Buth & Murphy (1980) Glycerol-3-phosphate dehydrogenase ASES G-3-pdh-A,B Citrate, pH8 Shaw 8 Prasad (1970) Leucine aminopeptidase 3.4.1.1 Lap-A Citrate, pH7 Brewer (1970) L-iditol dehydrogenase RARE: Iddh-A Borate Shaw & Prasad (1970) Malate dehydrogenase 137 Mdh-A Tris-citrate Shaw & Prasad (1970) Phosphoglucomutase 2.7.5.1 Pgm-A Citrate, pH8 Buth & Murphy (1980) Superoxide dismutase 115411 Sod-A,B EBT Johnson et al. (1970) nn II I ER. PHYSA BIOCHEMICAL GENETICS 353 TABLE 2. Allele frequencies at six polymorphic loci in four populations of Physa. EEE EL EE EE SN P. anatina P. gyrina Locus Allele Busey Woods INHS Pond Busey Woods Boneyard Creek Aat-A 100 = = 1.00 1.00 156 1.00 1.00 — — Est-1 -100 0.11 0.25 1.00 1.00 -162 0.89 0.75 — — Est-2 100 — — 0.63 0.61 116 — — 0.37 0.25 132 0.54 0.15 — 0.14 154 0.46 0.80 — — 170 — 0.05 — — Gpi-A 89 == — — 0.09 100 = = 0.77 0.77 111 1.00 1.00 0.23 0.14 G-3-pdh-A 100 — — 1.00 1.00 156 1.00 1.00 — — Pgm-A 83 1.00 1.00 — — 100 — — 1.00 1.00 butions at the three intraspecifically poly- morphic loci are provided in Table 3. The small sample sizes in this preliminary study precludes accurate Chi-square tests of Hardy-Weinberg expectations. Given the dis- tributions resolved in this study and the po- tential hermaphroditic mode of reproduction in Physa, we recommend much larger sample sizes for populations in future studies. Genetic variability, as ascertained via sev- eral measures of heterozygosity, in these populations is compared in Table 4. Physa anatina and P. gyrina show no marked differ- ences in these measures with both species exhibiting approximately 5% heterozygosity. Genetic differentiation within and between the species was quantified by calculating Nei’s (1972) coefficients of genetic similarity (/) and genetic distance (D) between all pairs of populations (Table 5). Intraspecific genetic distances were D = 0.012 or less. The ge- netic distance between P. anatina and P. gyrina is D = 0.45. DISCUSSION The use of whole-body homogenates in the study of gastropod enzymes has yielded complex patterns in genera other than Physa, presumably due to the expression of multiple loci (Nyman & Skoog, 1977). Relatively few multi-locus systems were resolved in this study. Gene duplication is believed to have played a negligible role in the evolution of the genus Physa (White, 1978), although gene expression differences may be of great utility in ascertaining relationships among pulmo- nate genera as have been used in studies of vertebrate polyploids, e.g. Ferris & Whitt (1978) and Buth (1979). The enzyme systems examined in this study were divided into four groups based on gene expression and allelic variability. The groups are discussed below. Single locus—Monomorphic: The five en- zymes in this category include acid and alka- line phosphatases, leucine aminopeptidase, L-iditol (sorbitol) dehydrogenase (Fig. 1A), and malate dehydrogenase (Fig. 1B). A con- siderable mobility difference observed be- tween the gene products of the Acph-A and Akph-A loci supports the genetic recognition of these as separate systems although they share a common substrate. Several other areas of enzymatic activity appeared with the LAP stain although these may be esterases rather than additional LAP loci. Multiple LAP loci have been reported in other gastropods, e.g. Rumina decollata (Selander & Hudson, 1976), although the zymograms for these species are unavailable for comparison. The single locus expression of MDH in Physa may be unusual. In vertebrates, MDH is at least a two-locus system in a mitochondrial-cytosol BUTH AND SULOWAY 354 PL ES 0 8 91 € | | euuAb “y L OL Sl 0 0 | euuAb “y OL 0 0 0 0 || euneue “y vl 0 0 0 0 | euneue “y LLL/LLL LLL/OOL 001/001 001/68 68/68 v-ıdg “€ 0 0 0 0 й | 9 | el | euuAb “y 0 0 0 0 0 0 S 6 21 | euuAb “y | д г 0 0 0 0 0 0 | euneue “y 0 € Z y 0 0 0 0 0 | euneue “y 041/Z€L ySL/pS! pSL/Zel ZEL/ZEL 251/001 ZEL/OLL 9LL/9LL 911/001 001/001 2-183 ‘2 0 0 8c Н euuAb y 0 0 92 | euuAË “y 9 € | | euneue “y LE € 0 | Buneue “y 291-/291- 291-/001- 001-/001- L-1S3 1 enn NP EP EP EE OP "с elqel ul pasn asou} эле suoneuBbisap энэ|е ay] “eSAY_ JO suonendod JNO} эц} и! 190] o1yd1owÁjod Ajjesıoadsenqui Besy} eu} je зиоцпащер 91dAjou89 ‘€ 3719v1 PHYSA BIOCHEMICAL GENETICS 355 TABLE 4. Genetic variability in four populations of Physa. Calculations are based on the gene products of 14 loci. Average Proportion of Proportion of | heterozygosity Effective Sample heterozygotes heterozygotes (Nei, 1978) number of Taxon Locality size observed expected” ONE: alleles P. апайпа Busey Woods 14 0.051 0.050 0.051 + 0.039 1.088 P. anatina INHS Pond 10 0.043 0.051 0.053 + 0.036 1.079 P. gyrina Busey Woods 26 0.052 0.059 0.060 + 0.041 1.102 P. gyrina Boneyard Creek 28 0.051 0.066 0.067 + 0.047 1.130 *“Hardy-Weinberg heterozygosity,” sample size not considered. TABLE 5. Comparison of Nei’s (1972) genetic similarity coefficients (/; above diagonal) and genetic distance coefficients (D; below diagonal) between populations. Standard errors of the distance coefficients are given in parentheses. Calculations are based on the gene products of 14 loci. P. anatina | P. апайпа || P. gyrina | P. gyrina Il P. anatina | — (Busey Woods) 0.988 0.630 0.631 P. anatina II 0.012 — 0.641 0.638 (INHS Pond) (+0.022) P. gyrina | 0.462 0.445 = 0.998 (Busey Woods) (+0.157) (+0.153) P. gyrina Il 0.460 0.449 0.002 — (Boneyard Creek) (+0.156) (+0.154) (+0.009) Sy O ес N N 9 9 А À © N 9 9 © © 4. Q. Q: Q. N © © N N 9 А S 9 9 100 2 < —Idh А —Mdh-A'99 Origin —> — twin ne ‘in Origin — O | 2 а © | 2 (A) L-Iditol Malate dehydrogenase dehydrogenase FIG. 1. Zymograms of monomorphic L-iditol (A) and malate dehydrogenase (B) expression in P. anatina and P. gyrina. 356 BUTH AND SULOWAY relationship (e.g. Rainboth & Whitt, 1974). Two MDH loci have been reported in other gastropods, e.g. Campeloma decisa (Selan- der et al., 1978). The absence of a second MDH system in Physa may be due to its weaker expression and/or restriction in ex- pression to a proportionately small organ. Single locus—Polymorphic: The aspartate aminotransferase, glucosephosphate iso- merase (Fig. 2A), and phosphoglucomutase (Fig. 2B) systems are included in this group. Species-specific allelic differences in the AAT system have been of taxonomic utility in Goniobasis (Chambers, 1978) and serve to distinguish P. anatina and P. gyrina. AAT, formerly “GOT” (glutamate oxalacetate trans- aminase), might be expected to be a multi- locus system in a mitochondrial-cytosol rela- 9 6 & & N N ©) ME © N dO o о Y RR Y © бр!-А'" — Gpi-A'99 Origin 2 | (А) Glucosephosphate isomerase tionship comparable to MDH as is the case in vertebrates. In Physa, faint cathodally-migrat- ing gene products of what may be a second AAT locus were inconsistently resolved. Multi- ple AAT (“СОТ”) loci have been reported from other gastropods, e.g. Rumina decollata (Selander & Hudson, 1976). The GPI poly- morphism in P. gyrina yields three-electro- morph presumptive heterozygotes. The for- mation of a single heteropolymer in this situa- tion suggests a dimeric structure of GPI in Physa as is the case in vertebrates. We have resolved a single PGM system in Physa al- though as many as five loci have been re- ported in other gastropods (Selander & Hud- son, 1976). Our resolution of PGM products is less than optimal with the production of nu- merous equidistant anodal subbands appear- ing in all specimens (Fig. 2B). 2 3 4 © | 2 > 4 Phosphoglucomutase FIG. 2. Zymograms of polymorphic glucosephosphate isomerase (A), and phosphoglucomutase expression (B) in P. anatina and P. gyrina. All individuals in Fig. 2 are in the homozygous state. PHYSA BIOCHEMICAL GENETICS Multiple locus—Monomorphic: The gene products of two superoxide dismutase loci (SOD) plus a presumptive interlocus hetero- dimer were resolved from all specimens. Mul- tiple SOD loci, formerly “IPO” (indophenol oxidase), have been reported from other gas- tropods, e.g. Rumina decollata (Selander & Hudson, 1976). Multiple locus—Polymorphic: Two G-3-PDH loci are expressed in Physa. The G-3-pdh-B locus is monomorphic in both P. anatina and P. gyrina whereas a species-specific allelic difference is observed at the G-3-pdh-A locus. : N N À 9 © Ô 0 357 An interlocus heteropolymer is formed (Fig. 3), suggesting a dimeric structure for G-3- PDH as is the case in vertebrates. The G-3- PDH system has received little attention in previous gastropod electrophoretic studies. The low staining cost and clarity of its resolu- tion make this enzyme system particularly de- sirable for use in future studies. Two very polymorphic esterase systems (one cathodal and one anodal) were clearly resolved in Physa. A third apparently polymorphic ester- ase locus, with extremely rapid anodal migra- tion of its products (Est-3), was not consistent- ly resolved. Esterases have been extensively _/S3-pdn-B2" _G-3-pdh-A/°68/90 — 6-3-pdh-A!00g!90 F6 3 pd A? - = \G-3-pan-A1°0 4 Glycerol-3-phosphate dehydrogenase FIG. 3. Zymogram of the multilocus glycerol-3-phosphate dehydrogenase system in P. anatina and P. gyrina. The subunit composition of each electromorph is indicated. 358 BUTH AND SULOWAY studied in gastropods. Their genetic control has not always been easily elucidated (e.g. Nyman & Skogg, 1977), however, some in- vestigators have resolved a considerable number of esterase loci, e.g. twelve in Rumina decollata (Selander & Hudson, 1976). All esterase heterozygotes in this study have two-electromorph expression, suggesting the monomeric composition of the enzyme. Our examination of geographically proxi- mate populations of Physa has provided no evidence for substantial local intraspecific dif- ferentiation in these species. This, however, does not preclude regional differentiation elsewhere in the ranges of these widespread forms. The substantial levels of polymorphism observed in Physa should allow the resolution of such regional restrictions in gene flow if they exist. The genetic differentiation between these subgenerically distinct species of Physa is substantial yet not absolute. Thus, electro- phoretic characteristics may be of taxonomic utility at several levels within the genus: popu- lations, species, subgenera. ACKNOWLEDGMENTS We thank M. Nei and E. Zimmerman for supplying the computer programs used for the calculation of genetic distances/similarities and heterozygosity measures, and G. Te for his verification of specific identifications. We appreciate the assistance of G. M. Davis and the comments of an anonymous reviewer. REFERENCES CITED BREWER, С. J., 1970, An Introduction to Isozyme Techniques. Academic Press, New York, 186 p. BUTH, D. G., 1979, Duplicate gene expression in tetraploid fishes of the tribe Moxostomatini (Cypriniformes, Catostomidae). Comparative Biochemistry and Physiology, 63B: 7-12. BUTH, D. G. & BURR, B. M., 1978, Isozyme varia- bility in the cyprinid genus Campostoma. Copeia, 1978: 298-311. BUTH, D. G. & MURPHY, R. W., 1980, Use of nico- tinamide adenine dinucleotide (NAD)-dependent glucose-6-phosphate dehydrogenase in enzyme staining procedures. Stain Technology, 55: 173- 176. 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” (Mesogastropoda: Pleuroceridae). Malacologia, 17: 157-162. CHAMBERS, S. M., 1980, Genetic divergence be- tween populations of Goniobasis (Pleuroceri- dae) occupying different drainage systems. Malacologia, 20: 63-81. COKER, W. Z. & KUMBA, E., 1974, A study of the genetic heterogeneity in Bulinus globosus popu- lations in southern Ghana, electrophoretic pat- terns of esterase enzymes in snail extracts. Ghana Journal of Science, 14: 19-22. DAVIS, G. M., 1979, Experimental methods in mol- luscan systematics, p. 99-169, In Pulmonates, Vol. 2A, FRETTER, V. & PEAKE, J. E, eds: Academic Press, London. DILLON, R. T., Jr. & DAVIS, G. M., 1980, The Goniobasis of southern Virginia and northwest- ern North Carolina: genetic and shell morpho- metric relationships. Malacologia, 20: 83-90. FERRIS, S. D. & WHITT, G. S., 1978, Phylogeny of tetraploid catostomid fishes based on the loss of duplicate gene expression. Systematic Zoology, 27: 189-206. HARRIS, H., 1966, Enzyme polymorphisms in man. Proceedings of the Royal Society of London, ser. B, Biological Sciences, 164: 298-310. HUBBY, J. L. & LEWONTIN, R. C., 1966, A mole- cular approach to the study of genic hetero- zygosity in natural populations. |. The number of alleles at different loci in Drosophila pseudoob- scura. Genetics, 54: 577-594. ISHAY, J., FISCHL, J. & TALMOR, N., 1976, Gyraulus piscinarum (Burguignat [sic], 1852) (water snail) and Xeropicta vestalis (Pfeiffer, 1841) (land snail): comparative biochemical data and saccharidase activity. Journal of Molluscan Studies, 42: 243-250. JOHNSON, A. G., UTTER, F. M. & HODGINS, H. O., 1970, Interspecific variation of tetrazolium oxidase in Sebastodes (rockfish). Comparative Biochemistry and Physiology, 37: 281-285. LOGVINENKO, B. M., GERMAN, S. M., KODO- LOVA, O. P. & MAKEEVA, V. M., 1976, On the system of muscle esterases in Lymnaea stag- nalis. Zoologicheskii Zhurnal, 55: 1138-1142. In Russian, English summary. METCALF, R. L., SANGHA, G. K. & KAPOOR; |. P., 1971, Model ecosystem for the evaluation of pesticide biodegradability and ecological magnification. Environmental Science and Technology, 5: 709-713. MICHELSON, E. H. & DUBOIS, L., 1975, Intraspe- cific variation in the hemolymph of Biomphalaria glabrata, a snail host of Schistosoma mansoni. Malacologia, 15: 105-112. MONTEIRO, W. & NARANG, S., 1976, Genetic variation in two natural populations of Biompha- laria tenagophila: esterases and alcohol de- hydrogenase. Anais da Academia Brasileira de Ciéncias, 48: 341-346. NARANG, S., 1974, Lactate dehydrogenase E.C. 1.1.1.27. of Biomphalaria glabrata. Part |. PHYSA BIOCHEMICAL GENETICS Physiochemical characterization of lactate de- hydrogenase E.C. 1.1.1.27. of various tissues. Comparative Biochemistry and Physiology, 47B: 641-655. NARANG, S. & NARANG, N., 1974, Characteriza- tion of malate dehydrogenase E.C. 1.1.1.37. electrophoresis isoelectric focusing thermosta- bility inhibition and activity studies on homo- genates of various organs of Biomphalaria glabrata (Mollusca: Pulmonata). Comparative Biochemistry and Physiology, 49B: 477-490. NARANG, S. 8 NARANG, N., 1976a, Alcohol de- hydrogenase of Biomphalaria glabrata (Mol- lusca: Pulmonata). Polymorphism genetic analy- sis and interspecific variations. Biochemical Genetics, 14: 75-86. NARANG, S. & NARANG, N., 1976b, Glucose-6- phosphate dehydrogenase of Biomphalaria glabrata (Mollusca: Pulmonata). Characteriza- tion and inheritance pattern. Anais da Academia Brasileira de Ciéncias, 48: 125-131. NEI, M., 1972, Genetic distance between popula- tions. American Naturalist, 106: 283-292. NEI, M., 1978, Estimation of average heterozygos- ity and genetic distance from a small number of individuals. Genetics, 89: 583-590. NICKERSON, R. P., 1975, A survey of enzyme and shell variation in 16 populations of the stream limpet Ferrissia rivularis (Say). Ph.D. thesis, Syracuse University. NYMAN, L., 1975, Allelic selection in a fish (Cymnodephalus cernau [L.]) subjected to hot- water effluents. /nstitute of Freshwater Research Drottningholm Report, 54: 75-82. NYMAN, L. & SKOOG, G., 1977. Species specific enzymes in some snail species (Mollusca, Gastropoda) from fresh and brackish water in Sweden. Zoologica Scripta, 6: 327-330. PATRICK, R., CAIRNS, J. & SCHEIER, A., 1968, The relative sensitivity of diatoms, snails, and fish to twenty common constituents of industrial wastes. Progressive Fish Culturist, 30: 137-140. RAINBOTH, W. J. & WHITT, G. S., 1974, Analysis of evolutionary relationships among shiners of the subgenus Luxilus (Teleostei, Cypriniformes, Notropis) with the lactate dehydrogenase and malate dehydrogenase isozyme systems. Com- parative Biochemistry and Physiology, 49B: 241-252. SACKLER, M. L., 1966, Xanthine oxidase from liver and duodenum of the rat: histochemical localiza- tion and electrophoretic heterogeneity. Journal of Histochemistry and Cytochemistry, 14: 326- 333. SELANDER, R. K. & HUDSON, R. O., 1976, Ani- mal population structure under close inbreeding: the land snail Rumina in southern France. Amer- ican Naturalist, 110: 695-718. 359 SELANDER, R. K. & KAUFMAN, D. W., 1975, Genetic structure of populations of the brown snail (Helix aspersa). |. Microgeographic varia- tion. Evolution, 29: 385—401. SELANDER, R. K., PARKER, E. E., Jr. 4 BROWNE, R. A., 1978, Clonal variation in the parthenogenetic snail Campeloma decisa (Vivi- paridae). Veliger, 20: 349-351. SHAW, C. R. & PRASAD, R., 1970, Starch gel electrophoresis of enzymes—A compilation of recipes. Biochemical Genetics, 4: 297-320. TE, G. A., 1975, Michigan Physidae, with system- atic notes on Physella and Physodon (Basom- matophora: Pulmonata). Malacological Review, 8: 7-30. TE, G. A., 1978, Systematics of the Physidae (Gastropoda: Pulmonata). Ph.D. thesis, Univer- sity of Michigan. TUCKER, W. J. & ETTINGER, W. H., 1975, A bio- logical investigation of the South Fork, Sanga- mon River and tributaries. Unpublished report by the Illinois Environmental Protection Agency, Di- vision of Water Pollution Control, 57 p. UKOLI, F. M. A., 1974, Electrophoretic studies on foot muscle esterases of some African Biom- phalaria species. Malacological Review, 7: 15— 24. WHITE, M. J. D., 1978, Modes of Speciation. Free- man, San Francisco, 455 p. WHITT, G. S., 1970, Developmental genetics of the lactate dehydrogenase isozymes of fish. Journal of Experimental Zoology, 175: 1-36. WHITT, G. S., MILLER, E. T. 4 SHAKLEE, J. B., 1973, Developmental and biochemical genetics of lactate dehydrogenase isozymes in fishes, p. 243-276, In: SCHRODER, J. H., ed., Genetics and Mutagenesis of Fish. Springer, New York. WILSON, Е. R., WHITT, С. $. & PROSSER, С. L., 1973, Lactate dehydrogenase and malate de- hydrogenase isozyme patterns in tissues of temperature acclimated goldfish (Carassius auratus). Comparative Biochemistry and Physi- ology, 46B: 105-116. WIUM-ANDERSON, G., 1973, Electrophoretic studies on esterases of some African Biompha- laria species (Planorbidae). Malacologia, 12: 115-122. WU, S.-K. & BURCH, J. B., 1975, Bulinus serinus (Gastropoda: Planorbidae) from Ethiopia. Mala- cological Review, 89: 31-46. WURZINGER, K.-H., 1979, Allozymes of Ethiopian Bulinus serinus and Egyptian Bulinus truncatus. Malacological Review, 12: 51-58. WURZINGER, K.-H. & SALIBA, E. K., 1979, A cyto- logical and electrophoretic comparison of Jordanian Bulinus with three other tetraploid Bulinus populations. Malacological Review, 12: 59-65. MALACOLOGIA, 1983, 23(2): 361-374 GENETIC AND MORPHOLOGICAL DIVERGENCE AMONG NOMINAL SPECIES OF NORTH AMERICAN ANODONTA (BIVALVIA: UNIONIDAE) Pieter W. Kat Department of Earth and Planetary Sciences, The Johns Hopkins University, Baltimore, Maryland 21218, U.S.A. ABSTRACT Stomach morphology of six species of the unionid genus Anodonta is illustrated, and a phenogram based on percent similarity of stomach morphological character-states is presented for these and two additional species. The eight species can be divided into four groups based on stomach structure. The first and second groups seem closely related, and contain species in the subgenus Utterbackia (A. imbecilis, A. couperiana, A. peggyae) and the subgenus Anodonta, s.s. (A. cygnea); the third group is composed of some of the members of the subgenus Pyganodon (A. cataracta cataracta, A. grandis, A. gibbosa), and a fourth group is formed by a single species, A. implicata. Support for the separation of A. implicata from the rest of the subgenus Pyganodon comes from electrophoretic data. This proposed separation, the close relationship in stomach structure between the American subgenus Utterbackia and the Euro- pean subgenus Anodonta, s.s. and the proposed European affinity of A. c. fragilis from Nova Scotia are all contrary to present taxonomic classifications. These data indicate that an inte- grated study of the entire genus based on shell, soft-part, and electrophoretic characters is in order. Key words: Anodonta; stomach morphology; electrophoresis; taxonomy; Unionidae. INTRODUCTION Bivalves of the family Unionidae are char- acterized by a high degree of phenotypic plas- ticity in shell shape and soft-part morphology. Early taxonomists relied almost exclusively on conchological characters to describe species, with the result that geographically wide- spread, variable species are burdened with extensive synonymies. Characters of the soft anatomy used in modern classification in- clude marsupial structure and morphology of the gills and characteristics of the siphons, mantle margin, and mode of reproduction (i.e. hermaphroditic, dioecious). However, these soft-part characters still are used almost en- tirely at taxonomic levels above that of the species, whereas conchological characters such as umbonal sculpture, dentition type, nacre color, and shell shape are used to dif- ferentiate among species within a genus. Allozyme electrophoretic methods, as well as immunoelectrophoresis, are important taxo- nomic tools for discriminating among species (Davis & Fuller, 1981; Davis et al., 1981), but even these methods pale in the face of the high genetic similarity that can occur among some conchologically defined species of the apparently recently radiating genus Elliptio (Davis et al., 1981). It is evident that an inte- grated approach that uses soft-part and conchological characteristics as well as electrophoretic techniques should be used to resolve the prevailing taxonomic confusion about the Unionidae. Details of stomach structure have been used successfully in the past for determina- tion of taxonomic affinities within the Bivalvia (Purchon, 1956, 1957, 1958, 1960, 1968; Graham, 1949; Dinamani, 1967). Relation- ships between and patterns of variation among various structural components of the stomach, however, have generally been used to differentiate among rather large taxonomic groupings, such as orders or families. To my knowledge, bivalve stomach structure has not been examined in order to determine rela- tionships within a genus, probably in part be- cause Graham (1949) and Purchon (1956 et seq.) indicated a fundamental similarity of structure throughout the class, so that con- sistent differences would be expected to oc- cur only among higher taxa. For example, with respect to the Unionacea, Purchon (1958) mentioned that “. . . the internal struc- ture of the stomach is of high stability . . .” and (361) 362 found little difference between Anodonta cygnea and Velesunio ambiguus (Philippi, 1847), taxa from different families (Unionidae and Hyriidae) within the Unionacea. Species of the circumboreal genus Anodonta were chosen in this study for sev- eral reasons. The genus seems to have un- dergone a moderate radiation yielding 14 species on the North American subcontinent (see Johnson, 1970; Burch, 1975). While conchologically defined species are not beset by extensive synonymies indicative of the confusion prevalent among some other genera, it has become increasingly apparent that the Anodonta complex needs careful revision. As the name implies, Anodonta lacks dentition, and taxonomic characters used to discriminate among species include umbonal sculpture, glochidial shell structure, and reproductive mode. The inadequacy of this system became clear because of recent debate about the taxonomic validity of A. peggyae, which had been included in A. imbecilis; uncertainty as to the status of sev- KAT eral nominal species such as A. hallenbeckii Lea, 1858, A. henryana Lea, 1857, and A. corpulenta Cooper, 1834, and proliferation of subspecies of the phenotypically diverse A. grandis. Six species of Anodonta that occur in the eastern United States (A. cataracta cata- racta, A. couperiana, A. gibbosa, A. imbe- cilis, A. implicata, and A. peggyae) were chosen for this study in an attempt to deter- mine whether stomach characteristics could be used to differentiate among unionid spe- cies. In addition, stomach morphology of A. grandis and A. cygnea was studied for com- parative purposes. METHODS The classification and collection localities of the taxa studied are presented in Table 1. Voucher specimens have been deposited at the Academy of Natural Sciences of Phila- delphia (ANSP). TABLE 1. Classification and collection localities of the species of Anodonta included in this study. Family Unionidae (Fleming, 1828) Ortmann, 1911 Subfamily Anodontinae (Swainson, 1840) Ortmann, 1910 Genus Anodonta Lamarck, 1799 Subgenus Anodonta, s.s. A. cygnea (Linnaeus, 1758) ANSP 355545 Subgenus Pyganodon Crosse & Fischer, 1894 A. cataracta Say, 1817 ANSP 355544 A. cataracta fragilis (sensu Clarke & Rick, Clarke & Rick, 1963) No vouchers A. grandis Say, 1829 No vouchers (shells broken) A. gibbosa Say, 1824 ANSP 355542 A. implicata Say, 1829 ANSP 355543 Subgenus Utterbackia F. C. Baker, 1927 A. couperiana Lea, 1840 ANSP 355539 A. peggyae Johnson, 1965 ANSP 355541 A. imbecilis Say, 1829 ANSP 355540 Waterwingebied 1 km E of Loenen, Noord Holland, The Netherlands Norwich Creek, 8 km E of Wye Mills, Talbot Co., Maryland First Lake O' Law, ca. 25 km NW of Baddeck, Vic- toria Co., Nova Scotia, Canada Lake Champlain, 5 km W of Colchester, Chittenden Co., Vermont Ocmulgee River, 10 km SW of Jacksonville, Ben Hill Co., Georgia Norwich Creek, 8 km E of Wye Mills, Talbot Co., Maryland Lake Osborne, Lantana, Palm Beach Co., Florida Withlacoochee River, Pasco Co., Florida 1.5km N of Lacoochee, Lake Osborne, Lantana, Palm Beach Co., Florida ANODONTA STOMACHS AND GENETICS 363 Stomach structure was studied in pre- served specimens that had been relaxed with sodium nembutal, fixed in 10% formalin, and preserved in 70% ethyl alcohol. The stomach was exposed by removing the surrounding muscle fibers and digestive diverticula. The stomach then was opened by first removing the entire dorsal region with a circular cut be- ginning and ending at the oesophagus, and then making a second cut around the base of the stomach to expose the stomach floor. The gastric shield was removed. This method has the advantage of imparting clarity to the rather complex anatomy of the stomach, a clarity commonly lacking in previous studies that often distorted the shape of the stomach for purposes of illustration. The dorsal region of the stomach exhibited only minor variation among the species studied, and consequently only the stomach floor has been illustrated. A minimum of four individuals of each species was examined in order to determine the ex- tent of intraspecific variation, and the illustra- tions represent a compilation of features from these individuals. Intraspecific variation did occur, but it was never sufficient to cause con- fusion among species. The terminology of stomach structures used by Owen (1955), Purchon (1958), and Dinamani (1967) has been followed in this paper. Degree of resemblance among species was determined on the basis of levels of simi- larity of nine structural components of the stomach, including shape of the minor typhlo- sole, shapes of the style sac and midgut, and the relationship between the minor typhlosole and the right sorting pouch. Each character was assigned a set of character states (Table 2) according to which each species was clas- sified (Table 3). It is important to point out, however, that the level of resemblance be- TABLE 2. Stomach characters analysed to determine relationships among species of Anodonta. 1. Shape and relationship of the minor typhlosole to the midgut A. Hooded B. Open 2. Shape of the style sac A. Rounded, with continuous raised rim B. Rounded, with partially raised rim C. Elongate, with partially raised rim 3. Shape and relationship of the major typhlosole fold A. Regularly curved B. Regularly curved, terminal concavity C. Sinusoidal 4. Curvature of the major typhlosole fold A. Relatively straight B. Curved C. Wavy 5. Shape of the sorting pouch A. Slightly curved B. Highly curved C. Hooked 6. Position of the junction between the sorting pouch and the minor typhlosole fold A. Dorsal B. Medial C. Distal 7. Curvature of the minor typhlosole fold A. Uniformly curved B. Non-uniformly curved C. Straight sections 8. Distance separating the minor typhlosole fold and the anterior part of the minor typhlosole A. Considerable B. Slight 9. Curvature of the terminal portion of the minor typhlosole A. Slightly curved B. Highly curved C. Hooked 364 TABLE 3. Character states (see Table 2) for the various species of Anodonta examined. Levels of similarity between species pairs presented in Table 4 are based on these character states. oe ame ee а A A A nn A AE SE NES Character and character state Species 12 3 4 5 6,7 8 9 c. cataracta A A c. fragilis A . grandis A . implicata B gibbosa A . соирепапа В . реддуае В . imbecilis В . судпеа А ъьььььььь mww>>O>wx»> PPOPO>>>U OWW>>000> >www>O>>> >wmwwm>O>wx> DDOOWOPD P>>>u>u>u >00>wWO>>}> tween species pairs is not calculated only from the number of characters in common be- tween each species pair. Rather, each char- acter was examined separately among repre- sentatives of each species pair, assigned a rating of “identical,” “similar,” or “different,” KAT and given a value of 2, 1, or 0, respectively. The values of all nine features were then added and divided by 18 (9 x 2) in order to ascertain level of resemblance between spe- cies pairs. By this method, species could share no characters in common in Table 3, but still exhibit some degree of similarity. As examples, Anodonta c. cataracta and A. c. fragilis share three identical characters, four similar characters, and two dissimilar charac- ters =3*24+4x1+2x0= 10/18 = 55 А. imbecilis and А. cygnea share three identi- cal characters, four similar and two different characters = 3 x 2 + 4 x 1 + 2 x = 10/18 = .55. This particular approach was adopted because it accounts for intraspecific variability better than does a method that assigns dis- crete character states to each feature. There nevertheless exists overall accordance be- tween the number of character states in com- mon and levels of similarity, especially among closely related species. The relationships among species and spe- cies groups were calculated from the percent FIG. 1. Stomach floor of Anodonta cataracta. C—conical mound, F—minor typhlosole fold, LD—left duct to the digestive diverticula, MG—midgut, OES—oesophagus, P—right sorting area pouch, RD—right duct to the digestive diverticula, SA1—sorting area immediately inside the oesophageal opening, SA2—right sorting area, SS—style sac, T—major typhlosole, TM—minor typhlosole. ANODONTA STOMACHS AND GENETICS similarity of stomach features between spe- cies pairs. Several “rounds” of comparison were set up in which individual species or species groups were compared to each other. In cases where species groups were involved, the features were averaged. This averaging technique results in the level of similarity of 0.51 in comparisons between Anodonta im- plicata and the A. cygnea-couperiana- imbecilis-peggyae “group,” for instance, while comparisons between A. implicata and individual members of this group can result in higher or lower levels of resemblance. Horizontal starch gel electrophoresis was performed on three populations of Anodonta implicata, three populations of A. cataracta fragilis, two populations of A. c. cataracta, and one population of A. gibbosa according to the methods described by Ayala et al. (1973) which were modified for unionids by Davis et al. (1981). Fourteen loci were ex- amined and scored according to the methods of Ayala et al. (1973). Nei’s (1972) genetic 365 distances were computer generated; these distances then were used in a cluster analysis (with unweighted arithmetic averages) routine available in the multivariate statistical pro- gram NT-SYS (Rohlf et al., 1974). This rou- tine was used to generate a phenogram de- picting the relationships among the taxa ex- amined. RESULTS The general anodontine stomach ground- plan can be described as follows. The short, wide oesophagus (OES; all structures labeled in Fig. 1) is longitudinally grooved and enters the anterodorsal region of the stomach. Im- mediately inside the oesophageal opening lies a small sorting area (SA1), also described by Purchon (1958) for Anodonta cygnea, which is considerably smaller than that indi- cated for Lamellidens corianus (Lea, 1836) by Dinamani (1967). The posterior section of the FIG. 2. Stomach floor of Anodonta gibbosa. For explanation of lettering see caption to Fig. 1. 366 OES FIG. 3. Stomach floor of Anodonta imbecilis. For explanation of lettering see caption to Fig. 1. stomach is folded inward so that a broad, fea- tureless shelf (not illustrated) overlies the style sac and the right sorting pouch (P); the posterior section of this shelf connects to the roof of the stomach. The major typhlosole (T) connects the com- mon aperture of the style sac (SS) and the midgut (MG) to the ducts of the digestive di- verticula (LD, RD), which lie in the antero- ventral region of the stomach. As already described by Purchon (1958) and Dinamani (1967), all specimens of Anodonta dissected in this study displayed a prominent “conical mound” (C) on the stomach floor between the style sac and the digestive diverticula open- ings. Purchon (1958) proposed that the coni- cal mound forms a division between the ante- rior and posterior section of the stomach; it could also function to keep the gastric shield in place. The minor typhlosole (TM) runs adjacent to the major typhlosole for a short distance, but then turns sharply to the right and terminates in the right sorting area (SA2). This sorting area is composed of two sections, one above the other, the lowermost connected to the up- permost by a prominent pouch (P). To the left of this pouch a fold (F) defines the border of the sorting area. The general structure of the anodontine stomach corresponds to Type IV of Purchon (1958) and Type IIIC of Dinamani (1967). Figs. 1-6 illustrate the stomachs of Anodonta cataracta cataracta, A. gibbosa, A. imbecilis, A. couperiana, A. peggyae, and A. implicata. Table 4 contains levels of resem- blance among these species, as well as A. cygnea, A. grandis and A. c. fragilis. Fig. 7 depicts relationships among eight of the spe- cies examined: A. c. fragilis was not included because of remaining uncertainty about its taxonomic status, which will probably remain unclear until topotypical A. fragilis from New- foundland can be examined. ANODONTA STOMACHS AND GENETICS 367 OES FIG. 4. Stomach floor of Anodonta couperiana. For explanation of lettering see caption to Fig. 1. TABLE 4. Levels of similarity of stomach morphology among species of Anodonta examined. © x 3 L S y x © 2 NS р NL a о D Se Q =) o Q Q y = o) Q Е 8 о Е > 5 y © < я & > y —— Say Pyganodon 9 76 FIG. 7. Phenogram (based on Table 4) depicting levels of similarity (1.0 = most similar) in stomach morphology between Anodonta implicata, A. cygnea, A.peggyae, À. imbecilis, A. couperiana, A. grandis, A. c. cataracta, and A. gibbosa. and the American Utterbackia: similarities in reproductive mode (i.e. hermaphroditism) and umbonal characteristics were considered un- reliable means of classification because con- vergence (rather than parallelism) could have given rise to the similarities. Similarity in stomach structure, however, revives the pos- sibility of a common ancestry for A. cygnea and Utterbackia. Anodonta cataracta cataracta, A. grandis and A. gibbosa are all rather closely related. It is important to point out that the specimens of A. grandis examined were collected in Vermont and thus could have hybridized to a certain extent with A. c. cataracta; Ortmann (1914) pointed out that conchological inter- grades between the species exist where their geographic ranges merge. While hybridiza- tion could have resulted in some blending of stomach characteristics to create a high de- gree of similarity, the fact that these species hybridize in the first place points to a close genetic relationship. Also interesting is the close relationship between A. c. cataracta and A. gibbosa, which was first pointed out by Frierson (1912), who thought A. gibbosa to be a subspecies of A. c. cataracta. Because of substantial differences in stom- ach structure between the groups discussed above and that of A. implicata, this species was placed in a separate group. Of the 372 Dar: FIT Ai NS Ai NS Ai NB AccNJ Acc NJ Acf NS Acf NS Acf NS Agb GE 6 4 .2 FIG. 8. Phenogram (based on Nei distances) depicting levels of similarity (0 = most similar) between Anodonta implicata (Ai), A. c. cataracta (Acc), A. c. fragilis (Acf) and A. gibbosa (Agb), Collection locations: NS = Nova Scotia, NB = New Brunswick, NJ = New Jersey, GE = Georgia. anodontine stomachs here examined, that of A. implicata is the most complex. In addition, A. implicata stomachs have three unique character states (Table 3): an elongate, rela- tively complex style sac and midgut opening, unique placement of the sorting pouch in rela- tion to the minor typhlosole, and a hook- shaped sorting pouch (Fig. 6). Electrophoretic data support separation of A. implicata from Pyganodon (Table 5). The level of genetic identity between A. implicata and A. c. cata- racta (0.370) is much lower than that ob- served between A. c. cataracta and A. gib- bosa (0.608), and the distribution of alleles among loci of A. implicata differs radically from that of A. c. cataracta and A. gibbosa (Table 7). Also, recent immunoelectrophoretic studies (Davis & Fuller, 1981) show consider- able differences between A. implicata and A. c. cataracta; in fact, A. implicata seems more closely related to Alasmidonta undulata (Say, 1816) and Lasmigona costata (Rafinesque, 1820) by multivariate analysis of immunoelec- trophoretic distances than to all congeners. These differences separate A. implicata from the subgenus Pyganodon, and stomach morphology places the species close to, but not within, the subgenera Anodonta and Utterbackia (Fig. 7). An interesting correlate to this study in- volves comparison between Anodonta cata- racta cataracta and A. c. fragilis from Nova Scotia. Clarke & Rick (1963) mentioned that A. fragilis was described from Newfoundland and that this species intergrades with A. c. cataracta on Nova Scotia, establishing the concept A. c. fragilis. Examination of A. c. fragilis stomachs reveals similarities to those of A. cygnea and A. imbecilis in the region of the right sorting area pouch and the midgut, and similarities to that of A. c. cataracta in the region of the minor typhlosole. This Nova Scotian anodontine thus could represent a taxon with European affinities which survived Wisconsinan glaciation of eastern Canada; Clarke (1966) found a close conchological relationship between A. kennerlyi Lea, 1860 (California to British Columbia and Alberta) and both A. cygnea and A. c. fragilis, indicat- ing that such anodontines with European af- finities could be more widespread than previ- ously supposed. Electrophoretic results re- veal that A. c. fragilis resembles A. c. cata- racta at a level similar to that of the species relationship between A. c. cataracta and A. gibbosa (| = 0.611), and exhibits both higher heterozygosity and polymorphism than A. c. cataracta (Table 6). An electrophoretic ex- amination of A. fragilis from the type locality and a careful study of the possible phenome- non of hybridization among species in the subgenus Pyganodon (A. c. cataracta x A. fragilis, A. c. cataracta x A. grandis) are clearly indicated. ANODONTA STOMACHS AND GENETICS 373 The species groupings presented above are based on an examination of a single organ and some initial electrophoretic data, and, while this permits a certain amount of speculation about taxonomic affinities, it does not permit taxonomic revisions. Nevertheless, while most relationships fall within accepted boundaries, the hypothesized relationships between American Utterbackia and A. cygnea and the placement of A. implicata in a group separate from the rest of Pyganodon should constitute sufficient encouragement for a synthetic study of the entire genus based on shell, soft-part, and electrophoretic charac- ters. Furthermore, these results should en- courage use of stomach morphology at the species level to determine relationships among other problematical groups within the Unionidae. ACKNOWLEDGMENTS My interest in the genus Anodonta was kindled by discussions with Arthur H. Clarke, Jr.; Derek S. Davis identified collection loca- tions in Nova Scotia; Eugene R. Meyer aided in the collection of Georgian populations, and George M. Davis and Caryl Hesterman pro- vided facilities and help during electrophore- sis. The study was funded, in part, by grants from the National Science Foundation (DEB 78-01550 to G. M. Davis), the National Insti- tutes of Health (Biomedical Research Support Grant 5-S07-RR07041-14) and the Consoli- dated Gifts Fund of the Department of Earth and Planetary Sciences of the Johns Hopkins University. REFERENCES CITED AYALA, F. J., HEDGECOCK, D., ZUMWALT, G. S. & VALENTINE, J. W., 1973, Genetic variation of Tridacna maxima, an ecological analog of some unsuccessful evolutionary lineages. Evolution, 27: 117-191. BURCH, J. B., 1975, Freshwater Unionacean Clams (Mollusca: Pelecypoda) of North Amer- ica. Rev. ed. Malacological Publications, Ham- burg, Michigan, 204 p. CLARKE, A. H., Jr., 1966, Genetic and ecopheno- typic relationships in northern Anodonta popula- tions. Annual Report of the American Malaco- logical Union for 1966, p. 24-26. CLARKE, A. H., Jr. & RICK, A. M., 1963, Supple- mentary records of Unionacea from Nova Scotia with a discussion of the identity of Anodonta fragilis Lamarck. National Museum of Canada Bulletin 199: 15-27. DAVIS, G. M. & FULLER, S. L. H., 1981, Genetic relationships among Recent Unionacea (Bi- valvia) of North America. Malacologia, 20: 217- 253. DAVIS, С. M., HEARD, W. H., FULLER, $. L. H. & 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, 15: 131-150. DINAMANI, P., 1967, Variation in the stomach structure of the Bivalvia. Malacologia, 5: 225- 268. FRIERSON, L. S., 1912, Notes on Anodonta couperiana and A. gibbosa. Nautilus, 25: 129- 130. GRAHAM, A., 1949, The molluscan stomach. Transactions of the Royal Society of Edinburgh, 61: 737-776. HEARD, W. H., 1975, Sexuality and other aspects of reproduction in Anodonta (Pelecypoda: Unionidae). Malacologia, 15: 81-103. JOHNSON, В. 1., 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- 449. NEI, M., 1972, Genetic distance between popula- tions. American Naturalist, 106: 283-292. ORTMANN, A. E., 1914, Studies in najades. Nautilus, 28: 41-47. OWEN, G., 1955, Observations on the stomach and digestive diverticula of the Lamellibranchia. |. Anisomyaria and Eulamellibranchia. Quarterly Journal of Microscopical Science, 96: 517-537. PURCHON, R. D., 1956, The stomach in the Proto- branchia and Septibranchia. Proceedings of the Zoological Society of London, 127: 511-525. PURCHON, R. D., 1957, The stomach in the Fili- branchia and the Pseudolamellibranchia. Pro- ceedings of the Zoological Society of London, 129: 27-60. PURCHON, R. D., 1958, The stomach in the Eulamellibranchia: Stomach type IV. Proceed- ings of the Zoological Society of London, 131: 487-525. PURCHON, R. D., 1960, The stomach in the Eulamellibranchia: Stomach types IV and V. Proceedings of the Zoological Society of Lon- don, 135: 431-489. PURCHON, R. D., 1968, The Biology of the Mol- lusca. Pergamon Press, Oxford, 560 p. ROHLF, E. J., KISHPAUGH, J. & KIRK, D., 1974, NT-SYS: Numerical Taxonomy System of Multi- variate Statistical Programs. State University of New York, Stony Brook, New York. 374 KAT APPENDIX Collection localities of the populations and species examined electrophoretically. A. Anodonta implicata Population 1. French Lake at Sunbury-Oromocto State Park, ca.10 km S of Oromocto, Sun- bury Co., New Brunswick, Canada. Population 2. Sydney River, Blachette Lake, ca. 6 km SW of Sydney, Cape Breton Co. Nova Scotia, Canada. Population 3. Shubenacadie Grand Lake, Grand Lake, Halifax Co., Nova Scotia, Canada. Anodonta cataracta fragilis Population 1. First Lake O' Law, ca. 25 km NW of Baddeck, Victoria Co., Nova Scotia, Canada. Population 2. Shaw Lake, ca. 6 km NNE of Arichat, Isle Madame, Richmond Co., Nova Scotia, Canada. Population 3. Newville Lake, Halfway River East, Cumberland Co., Nova Scotia, Canada. Anodonta cataracta cataracta Population 1. Swartswood Lake, Delaware River Drainage, Sussex Co., New Jersey. Population 2. Concord Pond, Nanticoke River Drainage, Sussex Co., Delaware. D. Anodonta gibbosa Population 1. Ocmulgee River at Ben Hill/Coffee Counties Public Boat Ramp, Ben Hill Co., Georgia. MALACOLOGIA, 1983, 23(2): 375-396 THE BIOLOGY AND FUNCTIONAL MORPHOLOGY OF THE TWISTED ARK TRISIDOS SEMITORTA (BIVALVIA: ARCACEA) WITH A DISCUSSION ON SHELL “TORSION” IN THE GENUS Brian Morton Department of Zoology, University of Hong Kong, Pokfulam Road, Hong Kong ABSTRACT The biology and functional morphology of the twisted ark Trisidos semitorta (Lamarck) are described. The adult occupies clean sands, with the anterior end buried and the anterior sagittal plane vertical to the surface. Because of posterior shell twisting, the posterior face of the right valve lies beneath that of the left which projects above the sand surface. The posterior sagittal plane of the shell thus lies parallel with the sand-water interface. The mode of life of the juvenile is also described for the first time. It lies within the inner surface of empty bivalve shells, firmly attached by a byssus. Juveniles of both 7. semitorta and T. tortuosa (Linnaeus) are less twisted than adults, the two species being similar when young. The adult Trisidos can be derived from an ancestor in which anatomical modifications adapting it to a byssally attached, nestling mode of life have been retained, indeed enhanced, in the transition from juvenile to adult to permit colonization of sands with fast-flowing currents above. Only the posterior region of the shell is twisted about the dorsoventral axis of ligament and byssus. In the free-living adult the byssus is absent, but the growth processes begun in the juvenile are continued into adult life. Shell twisting results from the contraction of asymmetrically developed posterior pedal retractor muscles, the left being larger and lying behind the diminutive right. This is analogous to the phenomenon of torsion in the Gastropoda, but the end result is different and the term twisting, and not torsion as used by McGhee (1978), more aptly defines the process of Trisidos. Key words: Trisidos; anatomy; twisting; posterior pedal retractors; evolution. INTRODUCTION Members of the arcid genus Trisidos are unusual in being twisted. More particularly, it would seem as though the anterior end of the shell had been held stable and the posterior end twisted through approximately 90°. This phenomenon has been studied most recently by McGhee (1978) and Tevesz & Carter (1979) in species, e.g. T. tortuosa, in which twisting is most obvious. The first study is a largely theoretical account of the pattern of torsion (the term being used synonymously with twisting by the former author) within the Arcidae and extinct Bakevelliidae. The latter account derives Trisidos from a Barbatia-like ancestor but is of insufficient detail to confirm or deny this hypothesis. The anatomical de- scription is superficial. Most authors describe twisting in Trisidos, but neither demonstrates how it occurs in this unusual group of animals. An understanding of Trisidos may cast light on the mechanism of twisting in the Bakevelliidae, the only other group of twisted bivalves (McGhee, 1978). Little is known of the biology of Trisidos. The adult is known to inhabit soft deposits into which it can reburrow, but the mode of life of the juvenile is unknown. This study was initiated when juveniles of 7. semitorta were found. Juvenile adaptations cast light on how shell twisting in the adult is achieved and the claim by Tevesz & Carter (1979) that the Trisidos ancestor was a Barbatia-like nestler. MATERIALS AND METHODS Adult and juvenile specimens of Trisidos semitorta were obtained with an Agassiz trawl from about 8 m of water in the NE territorial waters of Hong Kong (Mirs Bay), some 3 km to the ESE of the village of Sha Tau in the People’s Republic of China (grid reference 225,994 from 1:100,000.L681, 1970). The sea bed comprises a coarse sand and is covered by empty shells (occupied by the juvenile T. semitorta) of the bivalves Anadara antiquata (Linnaeus) and Tapes dorsatus (Lamarck). Also present is the burrowing ark Cucullaea concamerata (Martini) (Morton, (375) 376 MORTON 1981) and the shell boring gastrochaenid Cucurbitula cymbium (Spengler) (Morton, 1982). The samples were hosed down in a 5 mm sieve and returned to the departmental sea water aquarium. Large specimens were separately placed in shallow tanks with a thick bed of sand and circulating sea water. The ciliary currents of the organs of the mantle cavity and the stomach were elucidat- ed using carmine in sea water. Two small specimens were fixed in Bouin's fluid and serially sectioned at 6 um. The sections were stained in either Ehrlich's haematoxylin and eosin or Mallory's triple stain. Specimens of Trisidos tortuosa in the col- lections of the British Museum (Natural His- tory) have been examined as follows: 11 adult (dried valves) (BM (N.H.) reg. no. 1953.1.23.377) from Singapore. R. Winck- worth collection. Acc. No. 1838. 6 juveniles (dried valves) (no reg. no.) from Karachi, India. F. W. Townsend collection. Acc. No. 1831. 1 adult (alcohol preserved) (Reg. no. 81.11010) from Port Collis. 11 fathoms 'Alert' collection. The latter specimen was opened to confirm details of the posterior pedal retractor muscles previously noted for 7. semitorta. ABBREVIATIONS USED IN THE FIGURES A A cell layer of the style sac AA anterior adductor muscle or scar AA(Q) “Quick” component of the anterior adductor muscle “Slow” component of the anterior adductor muscle AN anus AP anal papilla APP anterior pedal protractor muscle APR anterior pedal retractor muscle ASO abdominal sense organ AU auricle B B cell layer of the style sac (the major typhlosole) В, B, се! layer of the stye sac (the mi- nor typhlosole) BG byssal gland BY byssus BYG byssal groove C C cell layer of the style sac CA ctenidial axis CFC coarse frontal cilia CR “chitinous” rod CS crystalline style CSM conjoined style sac and mid-gut CSS crystalline style sac D D cell layer of the style sac DD digestive diverticula DDD duct to digestive diverticula DH dorsal hood EA exhalant aperture Е foot FC food sorting caecum FFC fine frontal cilia FGC filament gland cell FO fold in the stomach wall FS fragmentation spherules G gonad GO gonoduct GP gonopore GS gastric shield H heart HG hind-gut IA inhalant aperture ID inner demibranch IG intestinal groove ILP inner labial palp IME inner mantle epithelium IMF inner mantle fold K kidney LC lateral cilia LEC left ctenidium LFC latero-frontal cilia LP left pouch MG mid-gut MI mantle isthmus MMF middle mantle fold MT minor typhlosole N nerve NF nerve fibers NU nucleus O oesophagus OC Oulastrea crispata OD outer demibranch OFM _ oblique fibres of mantle OLP outer labial palp OME outer mantle epithelium OMF outer mantle fold P periostracum PA posterior adductor muscle or scar PA(Q) “quick” component of the posterior adductor muscle PA(S) “Slow” component of the posterior adductor muscle PE pericardium PEG pedal gland PG pallial glands PO Polydora sp. PPR(L) left posterior pedal retractor muscle PPR(R) right posterior pedal retractor muscle TWISTING IN TRISIDOS PRM pallial retractor muscle R ridge entering the dorsal hood RA renal aperture RE rectum RPA reno-pericardial aperture SA sorting area of the stomach SC secretory cell SEC sensory cell nucleus dE major typhlosole TE transverse fibres TFM transverse fibres of mantle V ventricle TAXONOMY Trisidos Róding, 1798 is a genus of the family Arcidae (see Newell, 1969 for taxo- nomic details), the latter being divided into two subfamilies—the Arcinae Lamarck, 1809 and the Anadarinae Reinhart, 1935. The former are generally considered to be either power- fully attached nestlers, e.g. Barbatia Gray, 1842 and Arca Linnaeus, 1758, or borers, e.g. Litharca Gray, 1842, while the latter are typically either burrowing and abyssate, e.g. Anadara Gray, 1847, Scapharca Gray, 1847 or but weakly byssally attached, e.g. Bathy- arca Kobelt, 1891 and Bentharca Verrill & Bush, 1898. Despite this disparity in habitat, Newell (1969) places the burrowing, abyssate Trisidos in the Arcinae—a suggestion that will be discussed here. The genus Trisidos is relatively modern (Eocene) and has an Indo-Pacific distribution. The type-species is 7. semitorta Lamarck, 1819. According to Oyama (1974), there are three other species of the genus: 7. torta (Mörch), 7. kiyonoi (Makiyama) and 7. tortu- osa (Linnaeus). Oyama considers T. yongei Iredale to be a junior synonym of T. tortuosa. The form of Trisidos is very variable and some of these species may be in doubt, the genus warranting careful taxonomic revision. BIOLOGY Juvenile specimens of Trisidos semitorta are byssally attached and in Hong Kong waters occupy the inner surfaces of empty, large bivalve shells (Fig. 1). The byssus is relatively large and the animal is securely attached. It has been recorded from Tapes dorsatus, Anadara antiquata and adults of its own species. This habit undoubtedly ensures protection both from predators and from the 377 rapid water movement that must occur over the well-aerated sands the adult inhabits. As the bivalve grows, the host shell must become more restrictive until a time is reached when the juvenile detaches. At this time the byssus is lost and the adult assumes a burrowing mode of life. Cucullaea concamerata occurs in the same habitat and is also adapted for fast current speeds (Morton, 1981). Observa- tions on adult 7. semitorta have shown that, as with 7. yongei (Tevesz & Carter, 1979), reburrowing can occur though this process is slow, taking many days to complete. Typically, the shell of T. semitorta is eroded posteriorly and colonized by other organisms, mostly on the left valve (Figs. 2, 3A). Coloniz- ing species include the scleractinian coral Oulastrea crispata, the boring polychaete worm Polydora sp. and small Lithophaga malaccana (Reeve). Thus, the often exten- sively damaged posterior surface of the left valve projects above the sand surface while the remainder of the shell is buried. Makiyama (1931) has figured Arca (= Trisidos) kiyonoi in its natural position in the sand. In Hong Kong occasional adult specimens of T. semitorta have been collected intertidally from sand flats in Hoi Sing Wan (Starfish Bay), Tolo Harbour and more frequently by Agassiz trawl off this and other beaches. Tevesz & Carter (1979) report a similar habitat for T. yongei, i.e. a muddy, fine to medium sand subject to the effects of tidal currents and wave action and containing abundant fragmental shell material. FUNCTIONAL MORPHOLOGY The shell The aragonitic shell of Trisidos tortuosa, as in other members of the Arcacea (Taylor, Kennedy & Hall, 1969), comprises a crossed lamellar outer layer and a complex crossed lamellar inner layer though, unusually, the pallial myostraca is prismatic in the umbonal regions only. Both 7. semitorta and T. tortuosa are antero-posteriorly elongate, with a long multi- vincular ligament. The hinge plate is narrow with a continuous row of taxodont teeth. These are represented only by minute projec- tions under the central part of the hinge. Lat- erally, however, the teeth are relatively large and function in valve alignment and in pre- venting shear. MORTON 378 (av ) ‘wo | = ajeos ejenbnyue esepeuy (q ‘)) pue емоиша$ ‘| ynpe JO SSAJBA |эц$ JO ээвулп$ ¡eusaju! ay] о payoeye suawroads эпиэлпГ ом} jo syde1Bojoud Jamod (q ‘g) чбщ pue (9 “y) MO 'емоиша$ SOPISI/ ‘| "Ol TWISTING IN TRISIDOS 379 it appears that the posterior end has been twisted clockwise, the posterior part of the Sagittal plane turned through approximately 45° in T. semitorta and almost 90° in T. tortuosa. The animal, though lying approxi- mately vertically disposed in the sand with the anterior dorsoventral axis of the shell at right angles to the sediment-water interface, has its posterior margin lying approximately flush with the sand. The right valve, located under- neath the left, is buried. In both, but more noticeably in T. tortuosa, the posterior edge of the left valve projects beyond the margin of the right valve. The animal is slightly tilted in the sand, anterior end down, so that only the posterior face of the left valve is seen in life. The anterior end is not (except coinciden- tally) involved in the twisting process and thus the anterior region of the hinge plate remains vertically aligned whereas the hinge plate is twisted posteriorly and the posterior teeth in- terlock in a different plane to those anterior. Because of the twisting, the posterior ad- ductor acts at a different angle to that of the anterior and may augment the function of the hinge teeth in preventing shear (McGhee, 1978). The dorso-ventral axis of the shell, through the ligament and byssus, constitutes the fixed pivot point around which posterior twisting occurs. Tevesz & Carter (1979) con- sidered the hinge axis alone to be the pivotal point while McGhee (1978) considered it the byssus. The ligament is straight, though much larger in T. semitorta than T. tortuosa. T. tortuosa is more delicate than T. semitorta. In T. semitorta the shell is delicately ribbed, though this is often masked by the thick, fib- rous periostracum and by heavy erosion of the left valve posteriorly. 7. semitorta also possesses a weak postero-ventral sulcus ex- tending from the umbo. This is most notice- able in the left valve. In 7. tortuosa, the sulcus is much more sharply defined and angles the left valve so that the posterior face lies at right angles both to the remainder of the valve and to the sediment-water interface. As a result of the twisting process, the ventral margin of both species is sinusoidally curved. There is no trace of a byssal notch or ventral indenta- tion, though Newell (1969) considers this characteristic of the subfamily. In T. tortuosa, but not 7. semitorta, both posterior and ante- rior halves of the shell are laterally com- pressed, emphasizing the angularity of the shell. Juveniles of T. semitorta are byssally at- tached, though whether this is also true of 7. tortuosa is unknown but probable. Fig. 5 lem FIG. 2. Trisidos semitorta. A posterior view of the animal in a natural position in the sand with the inhalant and exhalant apertures open. For abbreviations see p. 376. MORTON 380 "WO | = ajeos ‘syoedse jessop ay) ‘а pue jeuuan ay] ‘9 ‘Ubu eu} ‘а ‘YO! ayı y шод pamela пецз BY ‘вроишез SOPISUL ‘€ “DI3 TWISTING IN TRISIDOS shows dorsal and ventral views of a juvenile specimen of 7. tortuosa 11 mm long and an adult 73 mm long. It can be seen that the juvenile is less twisted than the adult, the twisting process progressively influencing shell form with age and growth. Although no newly settled individuals have been seen, growth probably proceeds from an equilateral body plan. Neither juvenile 7. semitorta nor T. tortuosa possess a byssal notch, though the former at least is byssally attached when young and possesses a ventral byssal inden- tation (McGhee, 1978). This results in a slight heteromyarian form with an inflated posterior region relative to the anterior, the inequilat- erality being enhanced by the byssal indenta- tion. Figs. 1 and 6A of a young specimen of T. semitorta, 13 mm long, demonstrate the low degree of twisting. The shell, somewhat in- equivalve, has the characteristics of a nestling B E Gis LEFT [RIGHT \ с ple F N >= y RIGHT /LEFT < ] EN D LEFT 25cm RIGHT FIG. 4. Trisidos semitorta. The adult shell viewed from various aspects. A, the right valve; B, the left valve; C, dorsal aspect; D, ventral aspect; E, poste- rior aspect; F, anterior aspect. 381 bivalve with relatively reduced anterior and in- flated posterior shell slopes. The greatest shell width is dorsal to the dorso-ventral axis of the shell so that contraction of the byssal retractor muscles effectively serves to pull the animal down into the inner surface of large bivalve shells. The inflated posterior region of the shell increases the size of the apertures to the water above, thereby enhancing ex- change and is typical of nestling species, e.g. Philobrya (Limopsacea), Neogaimardia (Cyamiacea) and Trapezium (Arcticacea) (Morton, 1978; 1979a, b). The similarity be- tween juveniles of 7. tortuosa and T. semitorta is evidence for similar life styles (Fig. 6). Con- versely, the adults are dissimilar, though both exhibit twisting. Clearly, the constraints of the juvenile niche are responsible for a uniformity of body form that subsequently, in adult free- dom, achieves individual character. LEFT | RIGHT | \ x | \ e | 25 mm JN J ANTERIOR ANTERIOR | ] | RIGHT LEFT LEFT) RIGHT \ / х | | | lem ANTERIOR ANTERIOR FIG. 5. Trisidos tortuosa. A and B, dorsal and ven- tral views of a juvenile; C and D, dorsal and ventral views of an adult. 382 MORTON FIG. 6. The left shell valves of juvenile specimens of A, Trisidos semitorta and B, T. tortuosa. Scale = 5 mm. The musculature Anterior and posterior adductor muscles are present (Figs. 11 and 17), the former be- ing smaller and more dorsal than the latter. Inequality of the adult adductors represents a continuation of the juvenile condition. Both adductors are divided into slow (AA(S); PA(S)) and quick (AA(Q); PA(Q)) compo- nents of approximately equal size. Ventral to the anterior adductor, a pair of anterior pedal protractor muscles (APP) ex- tend into the visceral mass. Similarly, from the postero-dorsal edge of the anterior adductor is attached a pair of anterior pedal retractor muscles (APR). Left and right anterior pedal retractors and protractors are of equal size. How twisting in Trisidos is achieved has never been elucidated. Anterior to the poste- rior adductor muscle of 7. semitorta is a pair of posterior pedal retractor muscles (PPR). Figs. 7A and 8 show that the left muscle (PPR(L)) is large, with a wide area of attachment and passes into the visceral mass and foot to largely assume responsibility for posterior re- traction of the foot both on the left and right sides. The right posterior pedal retractor (PPR(R)) is small, with a small attachment area and is located anterior to the left retractor and its muscle blocks extend only a small way into the visceral mass. The situation in 7. tortuosa is similar (Fig. 7B) but exaggerated, i.e. the right posterior pedal retractor is minute in comparison with the left and as such is unusual, seen only in Trisidos (as far as is known), and results in the posterior twisting of the shell. The mantle The mantle is very thick and fleshy, tinted brown, with no mantle fusion. Posteriorly, the left and right mantle lobes are apposed so that inhalant and exhalant apertures are formed. When the animal is lying in sand (Fig. 2, lA, EA), these are clearly visible. The large foot can protrude mid-ventrally from between the mantle lobes to effect digging. In transverse section (Fig. 9), the mantle epithelia are widely separate and the en- closed haemocoel divisible into two compo- nents. Beneath the inner epithelium (IME) is an extensive haemocoel crossed by a few obliquely oriented muscle fibres (OFM) that presumably maintain the turgidity of the haemocoel, perhaps ensuring it is not over- filled with blood. This cavity contains numer- ous amoebocytes. Similar oblique muscle fibres occur in the spacious haemocoel in the mantle of the anomalodesmatan Pholadomya candida (Morton, 1980). Beneath this region is a further haemocoel abutting the outer mantle epithelium (OME) and crossed by many transverse fibres (TFM). Clearly, this haemocoel can expand very little. Throughout the mantle occur large num- bers of cells (SC), termed secretory cells, each apparently discharging at an epithelium and containing granules staining bright red in Ehrlich's haematoxylin and eosin and either red or green in Mallory's triple stain. These cells, a maximum of some 25 ¡um in diameter, are also found throughout the body, being particularly common in those epithelia in con- tact with the mantle cavity and in the gut epi- thelia. The mantle margin (Fig. 10) comprises three folds (Yonge, 1957), the inner (IMF) be- ing the largest and the middle fold (MMF) ex- tremely reduced. Discharging onto the gen- eral outer surface of the inner fold are many TWISTING IN TRISIDOS 383 _PA PPRIL) FIG. 7. Dorsal views of the pericardium of A, Trisidos semitorta; B, T. tortuosa. For abbreviations see p. 376. subepithelial gland cells (PG), staining blue/ green in Mallory's triple stain and pale red in Ehrlich's haematoxylin and eosin. A mucin probably is produced. Most of the branches of the pallial retractor muscle (PRM) penetrate the inner fold. Beneath the surfaces of the outer fold (OMF), which is subdivided into two sub-folds, occur large numbers of the subepi- thelial secretory cells (SC) noted above. The periostracum (P) is thin. The ciliary currents of the mantle Waste material landing on the surface of the mantle is rejected posteriorly. On each lobe (Fig. 11), a major rejection tract com- mences ventral to the anterior adductor mus- cle, extends postero-ventrally and then turns in a postero-dorsal direction so that unwanted material is eventually discharged via the ex- halant aperture. This is achieved largely by ciliary means, as in other arcids (Lim, 1966) and mytilids (Morton, 1973). This major rejec- tion tract is fed from the dorsal and ventral areas of the mantle by, respectively, down- —n FIG. 8. Trisidos semitorta. Transverse section through the posterior pedal retractor muscle. For abbreviations see p. 376. MI PPRIL) PPRIR) A AN © [Lt оао A Perreo, . , Porro. N %, CLS S e. 3 000000 7 00900000 iF aN 1mm BYG 384 MORTON IME SC Dim FIG. 9. Trisidos semitorta. Section through the general mantle. For abbreviations see p. 376. FIG. 10. Trisidos semitorta. Transverse section through the ventral mantle margin. For abbreviations see p. 376. TWISTING IN TRISIDOS 385 FIG. 11. Trisidos semitorta. The ciliary rejection currents of the left mantle lobe. For abbreviations see p. 376. AP ASO A DR EE O ot tests Y AAN AA FIG. 12. Trisidos semitorta. Surface view of the anus and abdominal sense organs located beneath the posterior adductor muscle. For abbreviations see p. 376. wardly and upwardly beating cilia. The ciliary currents are extremely strong. The abdominal sense organs Members of the Arcacea typically possess a pair of large abdominal sense organs (Heath, 1941) (Fig. 12, ASO) located close to the anus with its anal papilla (AP) and in close proximity to the postero-ventral face of the posterior adductor muscle. Heath (1941) has shown that in 7. tortuosa, along with the very great asymmetry of the valves, the right ab- dominal sense organ is markedly larger than that of the left. This is not so in the less twisted T. semitorta. In section (Fig. 13) large numbers of secre- tory cells (SC) (earlier described) are present in the sense organs, apparently being dis- charged from the epithelium. The epithelium comprises a very regular row of vertically aligned cells with long (8 um) nuclei (SEC), located just beneath the outer cell mem- brane. Beneath occur large numbers of round nuclei (NU), some 4 um in diameter and form- ing a layer 16 um thick, and beneath this again is a zone 4 um thick comprising verti- cally aligned, fine fibrils (NF) overlying a layer (4 um) of horizontally aligned nervous tissue (N). It would seem that the fibrils arise from the nerve and extend upwards towards the vertically aligned apical nuclei, but this con- 386 MORTON 100 um SEC NU B RSR NF SE 20 um FIG. 13. Trisidos semitorta. Low (A) and high (B) power sections through the abdominal sense organ(s). For abbreviations see p. 376. nection is obscured by the mass of interven- ing nuclei. The ctenidia The ctenidia of 7. semitorta (Fig. 14) com- prise two equal demibranchs: left and right ctenidia are similarly equal, ¡.e. valve inequali- ty does not affect gill dimensions. Gill ciliation is of Type B(1a) (Atkins, 1937b), typical of the Arcacea and Limopsacea (Atkins, 1937a; Lim, 1966; Morton, 1978). Acceptance tracts are located in the ctenidial axis and in the junctions between the ascending lamella of the inner (ID) and outer (OD) demibranchs with the visceral mass and mantle, respec- tively. The ventral marginal grooves pass large particles posteriorly to be rejected from between the posterior borders of the mantle along with pseudofaeces collected by the visceral mass and mantle. The posterior ex- tremities of the ctenidia are supported by a thick, muscular, suspensory membrane which gives this region great mobility. The ctenidia, with acceptance and rejection tracts sepa- rately located, act as primary sorting mecha- nisms, facilitated by the ciliation of the gill fila- ment. In transverse section (Fig. 15), each fila- ment has an apical crown of some six cells, TWISTING IN TRISIDOS 387 PA AN Е 2 и BEC E: OD 1D FIG. 14. Trisidos semitorta. The organs of the mantle cavity as seen from the right side after removal of the right shell valve and mantle lobe. For abbreviations see p. 376. ‘ each ciliated. The cilia are arranged in three vertical rows. A central row of coarse frontal cilia (CFC) 5-6 um long, is flanked by rows of fine frontal cilia (FFC) 3-4 ¡um long. Lateral to the fine frontal ciliated cells is another cell with a long (8-10 um), stiff cilium designated the latero-frontal cilium (LFC) by Atkins (1937a). Lateral again is a secretory cell (FGC), prob- ably producing mucus, and a series of, typi- cally, three cells possessing long (12 um) lateral cilia (LC) responsible for creating the flow of water through the ctenidium and an- other large secretory cell (FGC), again prob- ably producing mucus. The apex of the fila- ment is supported by “chitinous” rods (CR); and the base of the filament enclosing the fila- ment blood vessel is long, thin and composed of narrow cells, the two sides cross-linked by transverse fibres (TF). The labial palps The labial palps (Fig. 14, ILP, OLP) are lo- cated on the postero-ventral face of the ante- rior adductor muscle. Only the tips of the demibranchs extend between the palps. The ctenidial-labial palp junction is of category 3 (Stasek, 1963), typical of the Arcacea. The palps of T. semitorta have a parallel series of ridges and grooves on their inner surfaces oriented at approximately right angles to the oral groove. Very fine particles quickly pass over the crests of the palps towards the mouth (Fig. 16). Large particles pass into the depths of the troughs between ridges and are trans- ported outwards towards the palp margin where they fall off onto either the visceral mass or the mantle and are thence removed. On the oral surfaces of the crests, the ciliary currents beat downwards whereas on the aboral surface they generally beat upwards, out of the troughs. On this surface are a num- ber of laterally directed resorting currents. Resorting currents also exist on the crests of each ridge and these re-subject particles of intermediate size to either the acceptance or the rejection currents. In this process, apposi- tion or parting of the crests ensures that virtu- ally all or very little material is accepted or rejected. The foot and byssus The foot of the adult (Figs. 14 and 17, F) is antero-posteriorly elongate with a rather small digging “toe.” There is a long, ventral byssal groove (BYG) but no byssus, though adult T. 388 MORTON LFC CFC FFC \ | ECC EC. | ae N Foc A 5 E H ou E @ 10 um ey o e d | à 8 E als ¡ae TE FIG. 15. Trisidos semitorta. A transverse section through a single ctenidial filament. For abbrevia- tions see p. 376. tortuosa possesses a long, thin byssus (McGhee, 1978; Tevesz & Carter, 1979). Sections of juvenile T. semitorta (Fig. 18) having a stout byssus (BY) show the byssal roots radiating deeply into the visceral mass. The epithelium of the byssal groove is sur- rounded by dense numbers of subepithelial cells of the byssal gland (BG) which stain bright red in both Ehrlich's haematoxylin and eosin and Mallory's triple stain. Ventrally the foot contains another exten- sive sub-epithelial gland (PEG) that is not in- volved in secretion of the byssus but which may be responsible for the mucus copiously produced here. The cells are basophilic and stain red in Mallory's triple stain. The ciliary currents of the visceral mass The ciliary currents of the visceral mass (Fig. 14) complement those of the mantle. Thus a major rejection tract on each side of the body commences ventral to the anterior adductor muscle and extends to the postero- ventral edge of the visceral mass. Cilia on the visceral mass beat towards it. There are few currents supplying it from the foot. Waste ma- terial arriving at the posterior edge of the vis- ceral mass falls off, largely onto the right man- tle lobe. The alimentary canal The mouth, located on the ventral face of the anterior adductor muscle, opens to the oesophagus which passes dorsally to merge with the stomach. In transverse section (Fig. 19A) the oesophagus of a small juvenile spec- imen comprises a tube some 160 um in di- ameter composed of a columnar epithelium, approximately 60-80 um tall with cilia 10 um long. It is thrown into four longitudinal folds, though this number may increase in the adult, as in T. tortuosa (Heath, 1941, pl. 10, fig. 4). From the postero-ventral wall of the stom- ach the conjoined style sac and mid-gut ex- tends vertically down into the visceral mass. Transverse sections (Fig. 19B) show that the greater part of the style sac is lined by a col- umnar epithelium termed the A cell layer (A) consisting of cells 30 um tall with a nucleus 8 um in diameter and a thick border of cilia 10 ит long. In the Arcacea major and minor typhlosoles largely serve to separate the mid- gut and style sac. In section they comprise thin, elongate cells some 100 um tall, pos- sessing a centrally located, similarly elongate (6 ит) nucleus and with a fringe of cilia 8 um long. Internal to each typhlosole (the B (major) and B, (minor) cell layers) is a C cell layer (C), forming part of the epithelial lining of the mid-gut. The cells are approximately 65 um tall each with a fringe of stiff, bristle- like cilia, 4 wm long, that characterize this re- gion (Henschen, 1904; Kato & Kubomura, 1954; Morton, 1969). The remainder of the mid-gut epithelium comprises the D cell layer (D) of cuboid cells 16 um tall and with cilia 8 um long. TWISTING IN TRISIDOS 389 ABORAL FIG. 16. Trisidos semitorta. The ciliary currents of two ridges and a groove of the labial palps. CSM BYG FIG. 17. Trisidos semitorta. The structure and ciliary currents of the visceral mass as seen from the right side after removal of the right shell valve, mantle lobe and ctenidium. For abbreviations see p. 376. 390 MORTON MI OD "rn | PAL ALLER IT e BG PEG \ BYG / PEG FIG. 18. Trisidos semitorta. A transverse section through the visceral mass of a juvenile in the region of the byssus. For abbreviations see p. 376. In the ventral region of the visceral mass (Fig. 17), the mid-gut (MG) is separate from the style sac and coils before passing dorsally as the hind-gut (HG). In transverse section (Fig. 19C) the mid-gut is a circular tube 200 um in diameter with a single typhlosole compris- ing cells 80 um tall with cilia 14 ¡ym long. The remainder of the mid-gut epithelium com- prises a columnar epithelium 40 ¡um tall simi- larly ciliated. The hind-gut gives rise to the rectum. In section (Fig. 19D) the rectum com- prises a tube 140 um in diameter and com- prising cells 20 ¡um tall with cilia 6 um long. In the rectum the typhlosole divides into two ventral, longitudinal ridges. The rectum (Fig. 17, RE) penetrates the ventricle of the heart (H), passes between the posterior pedal re- tractor muscles (PPR) and thus also between the kidneys (K) and over the posterior adduc- tor muscle (PA) to terminate on the postero- ventral face of this muscle at an anus (AN) with a distinctive anal papilla. The stomach The large stomach (Fig. 20) lies ventral to the anterior region of the hinge plate and is of Type Ш (Purchon, 1957). The terminology TWISTING IN TRISIDOS 391 50 ут FIG. 19. Trisidos semitorta. Transverse sections through (A), the oesophagus; (B), the conjoined style sac and mid gut; (C), the mid gut; (D), the rectum. For abbreviations see p. 376. used in this description follows that of Purchon. The minor typhlosole terminates soon after emerging into the stomach. The major typhlosole extends across the floor of the stomach from right to left to terminate in a capacious food sorting caecum (FC). The food sorting caecum extends dorsal to the entrance of the oesophagus (O). The major typhlosole has on its right side the intestinal groove (IG) which transports waste material into the mid-gut. In the food sorting caecum is a very large sorting area (SA). Between each of the adjacent ridges of the sorting area is an aperture which leads into a component part of the digestive diverticula (DDD). A ridge dorsal to the row of sorting ridges carries material into the caecum. There is, in addition to the intestinal groove, a further ridge (R) carrying recycled material to the dorsal hood (DH) from where it is probably returned to the head of the style (CS) rotating against the gastric shield (GS). The latter is very small in relation to the size of the stomach and is located on the postero-dorsal wall. It sends spurs into the dorsal hood and the left pouch (LP). From the left pouch a series of ducts opens into the digestive diverticula. Particles settling on the surface of the major typhlosole are swept towards the left pouch and the ridge leading to the dorsal hood. The pericardium and associated organs The heart (Figs. 7 and 21) comprises a single ventricle (V) and paired lateral auricles (AU). From the posterior wall of the pericardi- um (Fig. 21, PE), paired reno-pericardial apertures (RPA) open into the paired kidneys 392 MORTON FO CSM T IG FIG. 20. Trisidos semitorta. The internal structure and ciliary currents of the stomach after opening by an incision in the right wall. For abbreviations see p. 376. PPRU Е у РРК(К) РЕ AU FIG. 21. Trisidos semitorta. The organs of the pericardium as seen from the right side. For abbreviations see p. 376. TWISTING IN TRISIDOS 393 (K). The kidneys lie posterior to the pericardi- um and lateral to the posterior pedal retractor muscles (Fig. 8). In section the kidneys com- prise a ventral series of tubules, but dorsally there are few tubule cells as described for 7. tortuosa (Health, 1941, pl. 11, fig. 3). The renal apertures (RA) discharge into the supra- branchial chamber. Close to the renal aperture is the gonopore (GP), which has thick fleshy lips. In T. tortuosa gonopore and renal aperture are joined (Heath, 1941). DISCUSSION The shell of Trisidos, especially 7. tortuosa, has been discussed previously (Makiyama, 1931; McGhee, 1978; Tevesz & Carter, 1979). Trisidos is unusual in being the only known extant twisted bivalve genus. In most morphological respects, Trisidos semitorta is a typical ark. Thus, the ctenidia and labial palps are little modified, the general disposi- tion of the mantle and organs of the mantle cavity and visceral mass are typically arcacean. The ciliary rejection currents of the mantle cavity, whilst also arcacean, are modi- fied to enable the adult, partially buried in soft deposits, to remove large amounts of sedi- ment that may enter the largely open mantle cavity. The shell is eroded posteriorly on the left valve only, the posterior end lying flush with the sediment-water interface and the right valve being wholly buried. The anterior end lies vertical to the surface. The structure and position of the posterior abdominal sense organs suggests a photoreceptive function, there being no sensory cilia suitable for moni- toring water flow. T. kiyonoi and T. tortuosa live in the same way as T. semitorta (Maki- yama, 1931; Tevesz & Carter, 1979). Twisting does not seem to have any effect upon the distribution or size of the organs of the mantle cavity. Thus in 7. semitorta left and right ctenidia, mantle lobes and labial palps are of approximately the same size in marked con- trast to the left and right inequality seen in the tangentially coiled Chamidae and Cleidothae- ridae (Yonge, 1967; Morton, 1974) and to a lesser extent in the markedly inequivalve Claudiconcha japonica (Morton, 1977). In T. tortuosa, however, left and right abdominal organs are of notably different size (Heath, 1941)—this is not so in T. semitorta. Also, the left and right posterior pedal retractor muscles are of different sizes; this influences slightly the size of left and right kidneys. Tevesz & Carter (1979) suggest that Trisidos is more probably evolved from a morphologically less specialized representa- tive of the Arcinae similar to the modern Barbatia. They argue that because of its rela- tively efficient ligament (Thomas, 1976) and streamlined shape a Barbatia-like ancestor was preadapted for the evolution of a shallow burrowing life habit. Superficially it would seem more reasonable to derive the burrow- ing, abyssate Trisidos from a shallow burrow- ing limopsacean (Limopsidae and Glycymer- ididae), but Thomas (1976) has shown that the duplivincular ligament of these is inherent- ly weak, arguing for morphological conserva- tism and not conducive to evolutionary diver- sification. Purchon (1957) describes stomachs of rep- resentatives of the Arcidae and Glycymer- ididae, and it is clear that a Barbatia-Trisidos link is Supported. The stomach of T. semitorta is similar to that of both Anadara and Arca and different from that of representatives of the Glycymerididae. It is, however, difficult, at first, looking only at the adult, to understand why Trisidos could not have evolved from a burrowing arcoid lineage, by posterior elonga- tion, twisting and, in 7. tortuosa, some degree of lateral flattening. There appears no reason for not deriving Trisidos from, say, an anadarine ancestor, especially as T. semitorta more closely fits the definition of the Anadarinae (Newell, 1969) than the Arcinae. However, the views of Newell (1969) and Tevesz & Carter (1979) are borne out by this research. Juvenile 7. semitorta are byssally attached to the inside of empty bivalve shells; subsequently, attachment and the byssus are lost. The byssus functions in the classical manner as a means of securing post-larvae in a position suitable for the growth of the juve- nile. Juvenile 7. tortuosa and T. semitorta are not so twisted when young, twisting being a progressive condition. The terms “torsion” and “twisted” require consideration. Traditionally Trisidos is re- ferred to as the “twisted” ark, but McGhee (1978) replaced this with “torted.” Generally, gastropods are torted, the mantle cavity mov- ing from a posterior to an anterior position in larval development which is related to an “asymmetry in the development of the retrac- tor muscles” (Garstang, 1929). In primitive prosobranchs (Crofts, 1955) and tectibranchs (Saunders & Poole, 1910), torsion is initiated by the contraction of a single asymmetrical, precociously developed larval cephalopedal retractor muscle. The term “twisted” has been 394 MORTON i F Ro. AE Die ai > в. в’ | à FIG. 22. A, Transverse section through the shell of an isomyarian, equilateral bivalve. B, Transverse section through the shell of a heteromyarian, equilateral bivalve, with byssus. C, Transverse section through an equilateral, partly heteromyarian, nestling bivalve. D, Transverse sections through the anterior (dotted lines) and posterior (solid lines) shell of a juvenile 7. semitorta. E, The same through an adult T. semitorta. F, The same through an adult 7. tortuosa. (Large open arrows indicate lateral compression in F). (A-Ay, the dorso-ventral axis of the shell; B-B;, the dorso-ventral axis through the posterior region of the shell of Trisidos; X-X+, the region of shell exhibiting the greatest shell width). applied to Trisidos though, hitherto, it was not known how this was achieved. The process of twisting in Trisidos has obvious similarities and major differences with torsion of the larval gastropod. Thus, twisting results from the contraction of unequal, asymmetrically aligned posterior pedal retractor muscles. These can be regarded as “cephalopedal” re- tractors because, though posteriorly located, they would in the primitive bivalve have with- drawn the head-foot. The difference is that twisting occurs /aterally about the sagittal plane of the mantle shell and not by altering the mantle cavity from a posterior to an ante- rior position. Moreover, it would seem that twisting in Trisidos is a post-larval and not a TWISTING IN TRISIDOS 395 larval feature. It is contended that the term torsion should be used only with reference to that process characteristic of the larval gas- tropod. In Trisidos the term twisted should be used, as it more appropriately defines the situation and distinguishes a process that is interesting though of restricted phylogenetic importance. From an isomyarian, infaunal, abyssate (except in the larva) ancestor (Fig. 22A), evo- lution in the byssally attached adult bivalves has proceeded in a number of directions. In the various heteromyarian bivalve lineages, e.g. Mytilacea and Dreissenacea (Yonge & Campbell, 1968) (Fig. 22B), the byssus acts as the point about which the reorganization of the body occurs. In epibyssate heteromyarian bivalves, the greatest shell width lies ventral to the mid-point of the dorsoventral axis of the shell. This ensures stability on wave-tossed, exposed beaches or in fast-flowing fresh waters (Morton, 1969). There are also hetero- myarian byssally nestling species. In the nest- ling bivalve (Fig. 22C), the greatest shell width usually lies dorsal to the mid-point of the dorso-ventral axis of the shell. The narrow ventral region of the shell enables the shell to tightly fit into crevices. From Fig. 22D it can be seen that the juvenile 7. semitorta is funda- mentally a nestler, its shell form matching that of others (Fig. 22C). In the adult (Fig. 22E and F), the anterior end retains its typical nestling form. There is a morphological compromise which, perfectly suitable for neither juvenile nor adult, permits maintenance of both. Both forms are essential if juvenile and adult are to Survive in fast-moving waters, but twisting optimises success for both. It should be noted that adult 7. semitorta possesses no byssus and that the twisting process seen in the free adult is a continuation of a processs begun in the post-larva. The twisted shell allows Trisidos to lie in the sediment in a way that minimizes the effect of fast currents and scour. Tevesz & Carter (1979) suggested that in very twisted 7. yongei lateral compression may reduce the ventilation efficiency of the mantle cavity. While this cannot be as acute in the more rounded 7. semitorta, species of Trisidos occupy well-aerated sands, possibly facing into the current, so that the problems of ventilation are reduced. Even in 7. tortuosa, the anterior region of the mantle cavity is still relatively unaffected by twisting and the mantle cavity functions in the usual manner. ACKNOWLEDGEMENTS | am grateful to Solene Morris of the British Museum (Natural History) for the loan of specimens of Trisidos tortuosa, to Dr. В. $. 5. Wu of the Agriculture and Fisheries Research Station, Aberdeen, Hong Kong for the provi- sion of boat facilities to collect T. semitorta and to Mr. H. C. Leung of the University of Hong Kong for histological assistance. REFERENCES CITED ATKINS, D., 1937a, On the ciliary mechanisms and interrelationships of lamellibranchs. Part I. New observations on sorting mechanisms. Quarterly Journal of Microscopical Science, 79: 181-308. ATKINS, D., 1937b, On the ciliary mechanisms and interrelationships of lamellibranchs. Part Ill. Types of lamellibranch gills and their food cur- rents. Quarterly Journal of Microscopical Sci- ence, 79: 375421. CROFTS, D. R., 1955, Muscle morphogenesis in primitive gastropods and its relation to torsion. Proceedings of the Zoological Society of Lon- don, 125: 711-750. GARSTANG, W., 1929, The origin and evolution of larval forms. Report of the British Association 1928 (Glasgow), Sect. D, p. 77-98. HEATH, H., 1941, The anatomy of the pelecypod family Arcidae. Transactions of the American Philosophical Society, 31: 287-319. HENSCHEN, F., 1904, Zur Kenntnis der blasen- formigen Sekretion. Arbeiten aus anatomischen Instituten Wiesbaden, 26: 573-594. KATO, K. & KUBOMURA, K., 1954, On the origins of the crystalline style of lamellibranchs. Scien- tific Reports of Saitama University, 3(B): 135- 152 LIM, C. F., 1966, A comparative study of the ciliary feeding mechanisms of Anadara species from different habitats. Biological Bulletin, 130: 105— 117: MAKIYAMA, J., 1931, On Arca kiyonoi n. sp. Venus, 2: 269-277. MCGHEE, G. R., 1978, Analysis of the shell torsion phenomenon in the Bivalvia. Lethaia, 11: 315- 329. MORTON, B. S., 1969, Studies on the biology of Dreissena polymorpha Pall. 1. General anatomy and morphology. Proceedings of the Malaco- logical Society of London, 38: 301-321. MORTON, B. S., 1973, Some aspects of the biol- ogy and functional morphology of the organs of feeding and digestion of Limnoperna fortunei (Dunker) (Bivalvia: Mytilacea). Malacologia, 12: 265-281. MORTON, B. S., 1974, Some aspects of the biol- 396 MORTON ogy and functional morphology of Cleidothaerus maorianus Finlay (Bivalvia: Anomalodesmata: Pandoracea). Proceedings of the Malacological Society of London, 41: 201-222. MORTON, B. S., 1977, The biology and functional morphology of Claudiconcha japonica (Bivalvia: Veneracea). Journal of Zoology, 184: 35-52. MORTON, B. S., 1978, The biology and functional morphology of Philobrya munita (Bivalvia: Philo- bryidae). Journal of Zoology, 185: 173-196. MORTON, B. S., 1979a, The biology, functional morphology and taxonomic status of Gaimardia (Neogaimardia) finlayi (Bivalvia: Gaimardiidae). Journal of Zoology, 188: 123-142. MORTON, B. S., 1979b, Some aspects of the biol- ogy and functional morphology of Trapezium (Neotrapezium) sublaevigatum (Lamarck) (Bi- valvia: Arcticacea). Pacific Science, 33: 177- 194. MORTON, B. S., 1980, The anatomy of the “living fossil” Pholadomya candida Sowerby 1823 (Mollusca: Bivalvia: Anomalodesmata). Videnskabelige Meddelelser fra Dansk natur- historisk Forening i Kébenhavn, 142: 7-101. MORTON, B. S., 1981, The mode of life and func- tion of the shell buttress in Cucullaea con- camerata (Martini) (Bivalvia: Arcacea). Journal of Conchology, 30: 295-301. MORTON, B. S., 1982, Pallial specializations in Gastrochaena (Cucurbitula) cymbium Spengler 1783 (Bivalvia: Gastrochaenacea). In: Proceed- ings of the First International Marine Biological Workshop: the Marine Flora and Fauna of Hong Kong and southern China, Hong Kong, 1980. Eds. B. S. MORTON & C. K. TSENG. Hong Kong Univerity Press, p. 859-873. NEWELL, N. D., 1969, Superfamily Arcacea Lamarck, 1809, p. N250-264. In: MOORE, R. C. (ed.). Treatise on Invertebrate Paleontology. Part N, Vol. 1 (of 3). Mollusca 6. Bivalvia. Geo- logical Society of America and University of Kan- sas Press. OYAMA, K., 1974, Revision of the genus Trisidos (Bivalvia: Arcidae). Venus, 32: 125-127. PURCHON, R. D., 1957, The stomach in the Fili- branchia and Pseudolamellibranchia. Proceed- ings of the Zoological Society of London, 135: 431-489. SAUNDERS, A. M. C. & POOLE, M., 1910, The development of Aplysia punctata. Quarterly Journal of Microscopical Science, 55: 497-539. STASEK, C. R., 1963, Synopsis and discussion of the association of ctenidia and labial palps in the bivalved Mollusca. Veliger, 6: 91-97. TAYLOR, J. D., KENNEDY, W. J. & HALL, A. 1969, The shell structure and mineralogy of the Bivalvia. Introduction. Nuculacea-Trigonacea. Bulletin of the British Museum (Natural History), Zoology, Supplement 3: 1-125, 29 pls. TEVESZ, M.J. S. & CARTER, J. G., 1979, Form and function in Trisidos (Bivalvia) and a com- parison with other burrowing arcoids. Malaco- logia, 19: 77-85. THOMAS, R. D. K., 1976, Constraints of ligament growth, form and function on evolution in the Arcoida (Mollusca: Bivalvia). Paleobiology, 2: 64-83. YONGE, C. M., 1957, Mantle fusion in the Lamelli- branchia. Pubblicazioni della Stazione Zoologi- ca di Napoli, 29: 151-171. YONGE, C. M., 1967, Form, habit and evolution in the Chamidae (Bivalvia) with reference to condi- tions in the Rudists (Hippuritacea). Philosophical Transactions of the Royal Society of London, ser. B, 252: 49-105. YONGE, С. M. & CAMPBELL, J. I., 1968, On the heteromyarian condition in the Bivalvia with special reference to Dreissena polymorpha and certain Mytilacea. Transactions of the Royal So- ciety of Edinburgh, ser. B, 68: 21-43. MALACOLOGIA, 1983, 23(2): 397-426 PHYLOGENETIC RELATIONSHIPS IN THE CEPHALOPOD FAMILY CRANCHIIDAE (OEGOPSIDA) Nancy A. Voss! and Robert S. Voss? ABSTRACT Fourteen qualitative morphological characters of squids of the oegopsid family Cranchiidae are described, and the distribution of their states among thirteen genera tabulated. Primitive and derived conditions for each character are inferred on the basis of outgroup comparisons and analyses of ontogenetic transformation series. Application of a Wagner Tree algorithm and Character Compatibility Analysis to the resulting data matrix yields nearly identical reconstruc- tions of cranchiid phylogeny. Hypotheses of monophyly for the traditional subfamilies Cranchi- inae and Taoniinae as well as for two of the three groupings of taoniin genera proposed by N. Voss (1980) are shown to be well corroborated, and refinements of previous ideas about cranchiid relationships are also proposed. Little homoplasy is evident in most of the characters of the study, but the anatomical position of digestive duct appendages appears to possess con- siderable evolutionary lability. Sources of new data for phylogenetic tests are suggested, and the need for additional research on teuthoid comparative morphology is emphasized. Key words: Mollusca; Cephalopoda; Cranchiidae; phylogenetic inference; Wagner Tree; Character Compatibility. INTRODUCTION Phylogenetic studies of living Cephalopoda are long overdue, but until recently have hard- ly been possible because of the uncertain tax- onomy of many groups and the absence of sufficient comparative anatomical data on which to base reliable estimates of evolution- ary relationships (G. Voss, 1977a). Past tax- onomic studies have usually emphasized ex- ternal morphology with only incidental treat- ment, if any, of internal structures, and the systematic potential of many organ systems has, therefore, seldom been explored. This is unfortunate because it seems desirable that classifications be based on as broad a suite of biological attributes as possible. Additionally, the fossil record of cephalopods, as it relates to the genealogy of most contemporary taxa, is inadequate (Donovan, 1977). The large and morphologically diverse pelagic squid family Cranchiidae was recently revised by N. Voss (1980). Thirteen valid genera were recognized and were arranged into two subfamilies, the Cranchiinae with three constituent genera (Cranchia, Lio- cranchia and Leachia), and the Taoniinae with ten (Helicocranchia, Bathothauma, Sandalops, Liguriella, Taonius, Galiteuthis, Mesonychoteuthis, Egea, Megalocranchia and Teuthowenia). Hypotheses of natural generic groupings within the Taoniinae were presented and the taxonomic distribution of a large number of morphological characters was tabulated. The present paper subjects data gathered by N. Voss (1980) on cranchiid comparative morphology to quantitative phylogenetic analysis in order to derive maxi- mally-corroborated hypotheses of relation- ships for these squids. It is our intention by so doing to test ideas about cranchiid classifica- tion presented in the 1980 paper, to argue the utility of much broader surveys of teuthoid morphology than have hitherto been under- taken, and to demonstrate the application of explicitly phylogenetic procedures to system- atic studies of contemporary cephalopods. MATERIALS AND METHODS The material examined during this study is from the extensive cranchiid collection amassed at the University of Miami over a period of several years from numerous loan- ing institutions and from the general cephalo- pod collection of Miami’s invertebrate mu- seum, supplemented by the collections of the TRosenstiel School of Marine and Atmospheric Science, University of Miami, Miami, Florida 33149, U.S.A. 2Museum of Zoology and Division of Biological Sciences, University of Michigan, Ann Arbor, Michigan 48109, U.S.A. (397) 398 U.S. National Museum. Specimens illustrated in the text belong to the following institutions: Australian Museum (AM), Dominion Museum, New Zealand (DMNZ), Institut Oceano- graphique, Monaco (IOM), Institut fur See- fischerei und Zoologisches Museum der Uni- versitat Hamburg (ZMH), Newfoundland Bio- logical Station (NBS), Scripps Institution of Oceanography (SIO), South African Museum (SAM), United States National Museum of Natural History (USNM) and the University of Miami Rosenstiel School of Marine and Atmospheric Science (UMML). 1. Character Analysis This study analyzes the historical informa- tion content of 14 qualitative morphological characters and samples variation in anatomi- cal features associated with reproduction, locomotion, feeding, digestion, excretion, structural support and concealment from predators; aspects of both larval and adult morphology are represented. Characters here treated were selected from among those dis- cussed by N. Voss (1980) on the basis of their within-genus constancy and because the vari- ants of the morphological expressions they represent could be coded as discrete states with minimal ambiguity. Character constancy within taxa of low rank seems desirable be- cause such constancy may reflect evolution- ary conservatism (Farris, 1966), and choosing characters with easily-described states mini- mizes the possibility of misclassifying the ob- jects of study. Less rigorous criteria for char- acter choice would have admitted a larger number of characters for phylogenetic analy- sis but probably at the risk of introducing more homoplasy to the data. The single exception, in this study, to our requirement that character expressions be constant within genera is dis- cussed in the analysis of Character 11, below. Our arguments for determinations of cran- chiid character polarities are presented indi- vidually, character by character, but fall broad- ly into two categories based on the criteria that we used to estimate relative primitiveness. Recent reviews and discussions of methods of polarity estimation are provided by Stevens (1980) and Watrous & Wheeler (1981). Outgroup comparisons: Of two or more al- ternative morphological conditions observed among cranchiid squids, the one that also oc- curs among other teuthoid cephalopods is here hypothesized to be primitive. When two or more states of a cranchiid character whose N. A. VOSS AND R. S. VOSS polarity was in question were encountered among other teuthoids, however, then polarity estimation required comparisons within yet narrower limits of cranchiid relationships. Choice of a more restricted outgroup was also dictated by the impracticality of tabulating character state distributions for all of the 24 families and 66 genera of living, non-cranchiid teuthoids (see G. Voss, 1977b) when original dissections were necessary to determine ana- tomical features rarely described or figured in the literature. In the absence of any well-cor- roborated estimate of teuthoid phylogeny from which an appropriate cranchiid sister group might have been chosen, we restricted our attention, when necessary, to compari- sons with only seven other oegopsid families: Thysanoteuthidae, Cycloteuthidae, Chiro- teuthidae, Grimalditeuthidae, Mastigoteuthi- dae, Joubiniteuthidae, and Promachoteuthi- dae. These families share, with cranchiids, 1) a funnel-locking apparatus other than a sim- ple ridge-and-groove, and 2) ventral connec- tives between the buccal membranes and arms IV (Young & Roper, 1969a, b; Roper et al, 1969). Whether these traits are really synapomorphies that would indicate a close relationship of the seven families to the Cranchiidae, however, is not yet known. Young & Roper (1968) believed that a simple ridge-and-groove funnel locking apparatus is the primitive oegopsid condition, but they also thought that the other types of locking (or fu- sion) arrangements may have been derived independently. Furthermore, because a ven- tral attachment of the buccal connectives to arms IV also occurs in the myopsids as well as the majority of the oegopsids, this charac- ter likewise provides but weak justification for our choice of cranchiid outgroups. Neverthe- less, some manageable basis for compari- sons had to be established in order to imple- ment our analyses, and the two characters cited above are among the few available on which to base such a selection. Ontogenetic precedence: In the absence of unambiguous results from the outgroup comparisons, relative primitiveness is esti- mated on the basis of developmental data; the ontogenetically antecedent character state is hypothesized to be primitive while ontogenetically subsequent alternative ex- pressions are hypothesized to be derived. In effect, we assume that evoluticnary novelties are, more often than not, cevelopmental modifications of phylogenetically ancestral conditions. Examples of neoteny and paedo- CRANCHIID PHYLOGENY 399 Character Number(s) Tree с А 1 Ь <а»а 23 d«»b>»>c ASF 8.12.13 a>b 6 bb>c 10 bc <, d 14 Ба ен! FIG. 1. Tree diagrams illustrating the estimates of polarity of states of characters described in the text. Hypotheses of polarity are presented in the right- hand column, and the characters whose states are believed to have evolved in the sequences illus- trated are listed to the left. genesis would provide obvious exceptions to this generalization; two cases of apparent paedomorphosis in cranchiid phylogeny are discussed below. The (two or more) derived conditions of multistate characters were arranged, when- ever possible, as geometrical or topological series (e.g., Character 6, Fig. 1) that could reasonably be expected to represent the se- quential order of appearance of advanced states from the plesiomorph under a gradual- istic model of phyletic change. Where no such series could be discerned (e.g., Characters 1 and 14, Fig. 1), all non-primitive states were regarded as independently derived by default. Multistate characters were then subjected to an additive binary recoding procedure (see Farris et al., 1970) that reduces transforma- tion series (morphoclines) with t states to t-1 binary (two-state) factors while preserving all of the phyletic information contained in the original character state tree topology. Table 1 presents the distribution of states of the origi- nal, unfactored characters among the cran- chiid genera, Fig. 1 provides diagrams of char- acter state trees, and Table 2 is the data ma- trix that results from application of binary re- coding to the character state distribution of Table 1 given the character state tree topolo- gies of Fig. 1. Binary factors of multistate characters are labelled with the name of the character state that is the apomorph for the transformation represented by the factor. Thus, binary factor 6d of Table 2 represents the character state transition (c>d) of Char- acter 6; cranchiid genera with a score of (1) for binary factor 6d exhibit either state (d) of Char- acter 6 or one of the two other states derived from (d) in the state tree for Character 6 (Fig. 1). 2. Inferring Tree Topologies Numerous quantitative methods have been proposed to construct hierarchical arrange- ments of organisms, but only a few are direct- ly pertinent to the problem of deriving well- corroborated hypotheses of phylogeny. Phenetic clustering algorithms, typically ap- plied to matrices of overall similarity meas- ures, do not address phylogenetic inference per se and are not employed here; Colless (1970) has argued that phenograms some- times provide reasonable estimates of phylo- geny, but the set of assumptions under which they may be presumed to do so seems to us unnecessarily onerous. Of explicitly phylo- genetic methods we have chosen two that operationalize, at least in part, the analytic procedures of Hennig (1966). The Wagner Tree method (Kluge & Farris, 1969; Farris, 1970) implements a heuristic procedure for discovering the most parsimon- ious hypothesis of phylogeny for a study col- lection of organisms and a set of cladistic characters. A most parsimonious phylogeny is defined to be that tree topology that re- quires the least number of convergent or re- versed evolutionary events in order to derive the character state distributions observed among the extant organisms of the study from the hypothesized morphology of the common ancestor. Unlike earlier parsimony ap- proaches (for example, Camin & Sokal, 1965), the Wagner method does not assume that evolution is irreversible, and for this rea- son we regard it as the more biologically rea- sonable. Caveats regarding uncritical use of the Wagner Tree method, however, have re- cently been offered by researchers (e.g., 400 N. A. VOSS AND R. S. VOSS TABLE 1. Primary data matrix. Columns represent cranchiid genera; rows represent characters as num- bered and described in the text. The entry for a given column x row is the character state label appropriate to the corresponding genus and character. Abbreviations of taxa for this and the subsequent tables and figures: Cra, Cranchia; Lio, Liocranchia; Lea, Leachia; Hel, Helicocranchia; Bat, Bathothauma; San, Sandalops; Lig, Liguriella; Tao, Taonius; Gal, Galiteuthis; Mes, Mesonychoteuthis; Ege, Egea; Meg, Megalocranchia; Teu, Teuthowenia. EE ——_——— TS Taxa Character number Cra Lio Lea Hel Bat San Lig Tao Gal Mes Ege Meg Teu 1 d d d C € С а а а а b b a 2 a b a a d a b с (© c (e (> с 5 a a a d d a a b b b с C с 4 b b b a a a a a a a a a a 5 a a b b b b b b b b a a b 6 b b b © © d d e e e e e f ih a a a b b b b b b b b b b 8 b b b a a a a a a a b b b 9 a a a a a a a b [© € a a a 10 a a a a b b b a a a a c a 11 С С а С © 6 С b b/c С b a b 12 b b b a a a a a a a a a a 13 a a b b b b b b b b b b b 14 b b d e f a e e e e (e C e Colless, 1980, but see also Mickevich & ble characters there exists at least one phylo- Farris, 1981) who report that applications of geny that all member characters can support, the algorithm to some data yield phylogenetic and that tree supported by the largest clique is reconstructions that are not uniquely most sometimes chosen as the best estimate of parsimonious; other, equally or more parsi- true evolutionary history. In this study, Char- monious hypotheses of evolutionary relation- acter Compatibility Analysis was used to de- ships may exist, and consideration of plausi- velop alternative hypotheses of phylogeny to ble alternative methods of phylogenetic in- be tested against the results of Wagner analy- ference are therefore of interest. ses. The method of Character Compatibility Minimum tree lengths and estimates of Analysis (Estabrook et al., 1977; Meacham, hypothetical ancestral phenotypes were cal- 1980) rests on the concept of the compatibility culated using the parsimony-optimizing proc- of cladistic characters (see also Estabrook, edure proposed by Farris (1970) subject to 1972). Two cladistic characters are said to be the constraint that the most recent common compatible if there exists at least one hy- cranchiid ancestor exhibit the primitive mor- pothesis of phylogeny for the organisms of the phology determined a priori by the methods of study collection that both can support. Thus, if individual character analysis described two characters are incompatible, then both above. cannot support historical truth; at least one The computer programs we used to exe- (and perhaps both) has undergone homo- cute the Wagner analyses were written by plasy in the course of the evolution of the J. S. Farris; the program for Character Com- study collection. All characters that support patibility Analysis was written by K. L. Fiala true statements of phylogenetic relationships, and С. Е. Estabrook. Analyses were per- however, must be mutually compatible, while formed on the Michigan Terminal System at characters that do not support historical truth the University of Michigan. may or may not be pairwise compatible with each other and/or with true characters. Given a study collection of organisms and a set of RESULTS cladistic characters, the compatibility of all character pairs can be analyzed and groups 1. Character Descriptions and Analyses (cliques) of mutually compatible characters identified. For any clique of mutually compati- Character 1. Funnel-mantle fusion cartilages: CRANCHIID PHYLOGENY 401 TABLE 2. Factored data matrix. Columns are labelled as in Table 1. For an explanation of factor labels and table entries, see the Methods section of the text. Cra = S Lea Hel Bat D ooOooo0-0---0000-00000-0-000000-00 oooo-0---0000-00000-0-00000--00 oo-00--000000-00000---000000-00 o-000-0--00000-000-0-0-000000-0 --000-0--0-000-000-0-0-00-000-0 San ooooo-0--0-000-00--0-00000000-0 Taxa Tao Gal Teu Mes Lig Ege — ooo-o+-o0o0o0o+}000+}-0++-0000;+++0+;+++00-+ 5 © o-000-0--0-000-00--0-000000-000 o-000-00-000-0-0---0-000-0--000 So 200: 1010110000: 1000 o-000-0--00--0-0---0-000-0--000 ooo-0-00-0000--0---0000--0--00- o-000-00-0000------0-00--0--000 stout, roughly oval, subtriangular or spindle-shaped; elongate, triangular; narrow, straight; fused into ventral cartilaginous strips. In the Cranchiidae, the mantle is fused to the funnel at its posterolateral corners along two acutely diverging lines. These lines of fu- sion, found only in the cranchiids, replace the diverse forms of funnel-mantle locking carti- lages present in all other teuthoids. In all members of the Cranchiinae, external carti- laginous strips, located on the ventral surface of the mantle, extend along one (the dorsal- most: Leachia) or both (Cranchia, Liocran- chia) of the paired internal lines of funnel- mantle fusion and probably serve to strength- en the attachments (Fig. 2.4). The strips may run for partial or full length of, or beyond, the lines of fusion. In Cranchia, these external strips are short, of coequal length and smooth except for a multipointed apical tubercle, while in Liocranchia they are relatively long, of co- equal or unequal length, and tuberculate for their full extent. In Leachia, the strips are al- ways tuberculate and vary, among the spe- cies of that genus, from about 10 to 50 per- cent of the mantle length. Outside of the cranchiids, fusion of the fun- nel and mantle occurs in only two other teuthoid genera: Symplectoteuthis (Ommas- trephidae) and Grimalditeuthis of the mono- typic family Grimalditeuthidae. In Symplecto- teuthis, only the posterior portion of the funnel-mantle locking cartilages are fused, while in Grimalditeuthis there is complete fu- sion of the cartilages. In the Taoniinae, only remnants, termed funnel-mantle fusion carti- lages, remain of the typical, separate locking 402 N. A. VOSS AND R. S. VOSS FIG. 2. Funnel-mantle fusion cartilage, left: (1) Galiteuthis glacialis, Elt 1112 (USNM), adult, 395 mm mantle length; (2) Egea inermis, WH 467-71 (ZMH), adult, 260 mm mantle length; (3) Bathothauma lyromma, AD 329-79 (ZMH), subadult, 190 mm mantle length; (4) Liocranchia reinhardtii, CI 71-98 (USNM), subadult, 160 mm mantle length. Dotted lines trace internal lines of funnel-mantle fusion; full extent of fusion lines not shown (see Character 1). cartilages of non-cranchiid teuthoids. Funnel- mantle fusion cartilages cannot be identified as separate elements in the Cranchiinae, but it seems probable that such cartilages were the points of origin from which the ventral cartilaginous strips evolved. The funnel-mantle fusion cartilages and the derived strips are the only known instances of cartilaginous elements present externally at the two funnel-mantle junctions. Among other teuthoids, with the exception of Symplecto- teuthis and Grimalditeuthis, the locking ap- paratus consists of two separate, compli- mentary, internal elements—one on the man- tle and the other on the funnel. In the majority of oegopsid families, the locking apparatus is a simple, straight groove-and-ridge arrange- ment that was hypothesized to be primitive for the order by Young & Roper (1968). In two families, Ommastrephidae and Thysanoteu- thidae, the locking apparatus is L- or —- shaped; in the remaining six families, which together with the Thysanoteuthidae comprise the outgroup, the funnel locking apparatus is round, oval, or subtriangular in shape. The form of the funnel-mantle fusion carti- lages varies within the Taoniinae. In Helico- cranchia and Bathothauma, where the ex- ternal funnel-mantle fusion area is markedly broad, the cartilage is straight, very slender and barely discernible (Fig. 2.3). In Sanda- lops, where the external fusion area is not broad but narrow as in the other taoniins, the cartilage is also straight, but distinct, and is shorter and wider than in Helicocranchia and Bathothauma. In all three genera, the carti- lage (coded as “narrow, straight,” above) fol- lows the dorsalmost of the two internal lines of funnel-mantle fusion. The cartilage in Egea and Megalocranchia is elongate and triangu- lar, with the longest side following the dorsal- most internal line of fusion (Fig. 2.2). In the remaining five taoniin genera, Liguriella, Taonius, Galiteuthis, Mesonychoteuthis and CRANCHIID PHYLOGENY 403 Teuthowenia, the cartilage is stouter, varies considerably in shape, and is positioned more apically with respect to the internal lines of fusion, the long axis of the cartilage sometimes tending to follow the ventralmost line (Fig. 2.1). In this group of five genera, the cartilage also bears tubercles on the anterior end. These are present only in the young of Mesonychoteuthis and of some species of Taonius, but are pres- ent in both young and adults of Liguriella, Teuthowenia and most species of Galiteuthis. Because the more apical position and gener- ally stouter outline of the fusion cartilage in these five genera most closely approaches the orientation and shape of the funnel-mantle locking cartilages among members of the out- group, the character state to which we have assigned these squids is judged to be primitive for the extant cranchiids, and the remaining four states are hypothesized to have been in- dependently derived. Character 2. Posterior end of gladius: (a) conus present in larva and adult; (b) conus present in larva, lost or obscured in pseudoconus of adult; (c) conus lacking, pseudoconus present in larva and adult; (d) conus and pseudoconus absent. Re-evaluation of the shape of the posterior end of the gladius of adults and a careful ex- amination of the gladius in the larvae of all of the cranchiid genera have revealed anatomi- cal differences in addition to those described earlier (N. Voss, 1980: Table 1). Voss de- scribed the character states: a) short conus; b) medium to long conus; and c) conus lack- ing, but did not distinguish between the two types of “conus” that occur within the family: 1) a “true” conus that exhibits no evidence of fusion or convergence of the edges of the gladial vanes along the ventral midline, and 2) a pseudoconus formed by the infolding of the posterolateral margins of the gladial vanes that converge along the midventral line, with or without subsequent fusion (Fig. 3). The definition of pseudoconus has been expand- ed here from that of McSweeny (1978) in order to include the instances of ontogeneti- cally subsequent fusion that follow infolding in some cranchiids (N. Voss, 1980), and in some families of the outgroup and of other FIG. 3. Posterior end of gladius, ventral view, showing: (1) conus with enlarged cross-section and detail of Sandalops sp. B, WH 443-71 (ZMH), adult, 144 mm gladial length; (2) pseudoconus with enlarged cross- section and detail of Teuthowenia megalops, B 6 (NBS), adult, 254 mm gladial length (see Character 2). 404 oegopsids. In all observed cases, a fusion line is distinguishable. Among extant teuthoids, the conus portion of the gladius, presumably a vestige of the ancestral phragmacone, may be very small, or is sometimes found only in the young, or may be entirely absent (Naef, 1921/1923). All three conditions, in addition to the formation of a pseudoconus, are en- countered among cranchiids. Among members of the outgroup, a gladius with a narrow, usually elongate pseudoconus is typical of the Chiroteuthidae, Grimalditeuth- idae, Mastigoteuthidae and Joubiniteuthidae. Specimens of three of the four nominal cycloteuthid species were examined; the gladius has what appears to be a true conus in one species, a pseudoconus in a second species and neither a conus nor a pseudo- conus in a third species. In the little known Promachoteuthidae, a gladius with what ap- pears to be a weakly-formed conus is found in an unnamed species (R. Toll, personal com- munication). Thysanoteuthids have a weakly formed conus in the young stages (R. Toll, personal communication) but lack both conus and pseudoconus in the adults. A pseudo- conus is also found in three other oegopsid N. A. VOSS AND R. S. VOSS families that do not belong to the outgroup: Lepidoteuthidae, Brachioteuthidae and Bato- teuthidae. Even though a pseudoconus is the com- moner structure in the outgroup, we believe that the presence, in larva and adult, of a small conus displaying no evidence of mid- ventral fusion or convergence of the lateral margins of the vanes is primitive for cranchi- ids. We would support this judgment by the observation that, where both conus and pseudoconus are sequentially exhibited in the ontogeny of extant cranchiids, it is the conus that is invariably precedent and the pseudo- conus that is developmentally subsequent. The absence of both conus and pseudo- conus, a condition found only in Batho- thauma, we judge to have been derived inde- pendently from the primitive state. This judg- ment is based on the unique modification of the posterior end of the gladius in Batho- thauma in which the vanes are transformed into a transverse bar that gradually expands laterally to shovel-shaped ends on which the fins insert (Fig. 4.4). An elongation of the pos- terior end of the conus in Leachia and Heli- cocranchia serves to extend support for the FIG. 4. Posterior end of mantle, dorsal view, showing variation in shape of fins: (1) Sandalops sp. C, Cl 71-6-26 (USNM), adult, 102mm mantle length; (2) Galiteuthis glacialis, Elt 1323 (USNM), adult, 333 mm mantle length; (3) Teuthowenia megalops, B 17 (NBS), adult, 352 mm mantle length; (4) Batho- thauma lyromma, O 4713 (UMML), subadult, 165 mm mantle length (1-3 from N. Voss, 1980) (see Char- acter 3). CRANCHIID PHYLOGENY fins; it is a solid structure and does not appear to be homologous with the hollow pseudo- conus. Character 3. Shape of fins: elliptical, oval or circular, terminal; lanceolate or stout, ovate, terminal; lanceolate or long-narrow, terminal-lat- eral; small, paddle-shaped, subterminal. (d) Ontogenetically, the fins develop from the shell fold (Naef, 1921/1923). In the early larva of all cranchiids, the fins are small, separate and paddle-shaped. They then typically be- come longer and rounded with growth, later become contiguous, and finally elongate to varying degrees to accompany the elongation of the posterior end of the gladius. While growth of the oegopsid fin is typically anterior, fin growth among cranchiids is typi- cally posterior. The fins are terminal, elliptical, oval or circular in all genera in which the conus is present in the larvae (Fig. 4.1), with the exception of Helicocranchia. With onto- genetic disappearance of the conus, the fins are extended posteriorly on the developing pseudoconus and assume a generally lance- olate form. They remain terminal in Taonius, Galiteuthis and Mesonychoteuthis (Fig. 4.2), while in Egea, Megalocranchia and Teutho- wenia, they simultaneously grow anteriorly on the mantle to become terminal-lateral (Fig. 4.3). The musculature of the fins is usually poorly developed except in Mesonycho- teuthis in which the fins become stout and ovate in shape, and very muscular medially. The form of the fins in the outgroup varies considerably. The typical shape is elliptical, oval or circular, with the marked exception of thysanoteuthids in which the fins are rhom- boid. The fins may be subterminal, terminal, terminal-lateral or extended to nearly the full length of the body. The pseudoconus often projects beyond the posterior margins of the fins as a slender to needle-like tail of varying length; this structure may bear a supplemen- tary, or auxiliary finlike structure. On the basis of its ontogenetic precedence, the state of “elliptical, oval or circular, termi- nal” fins is considered to be primitive for ex- tant adult cranchiids. The common develop- mental trend in the family toward posterior elongation of the fins with support afforded by a lengthening pseudoconus, together with the subsequent occurrence (in three genera) of anterior growth to form terminal-lateral fins, is 405 interpreted to reflect the evolutionary se- quence of appearance of these conditions in cranchiid phylogeny. The retention, into adult- hood, of the larval state of small, paddle- shaped fins in Helicocranchia and Batho- thauma (Fig. 4.4) is interpreted to represent a neotenous condition independently derived from the primitive state. Character 4. Funnel-head fusion: (a) (b) In the Cranchiidae, lateral fusion of the fun- nel to the head occurs only in Cranchia, Liocranchia and Leachia. The funnel is free laterally in all members of the Taoniinae. Among other teuthoids, additional instances of the fused state are found in the Bathyteuthi- dae and in the sole member of the Joubini- teuthidae; the Ommastrephidae, and some members of the Chiroteuthidae and Mastigo- teuthidae also display varying degrees of lat- eral fusion. In the majority of teuthoids, how- ever, including members of the remaining four outgroup families, the funnel is free laterally. The free, unfused condition is tentatively in- terpreted as primitive for the family Cranchi- idae, and the fused condition as derived. Though the functional significance of the varying degrees of lateral fusion is not known, it presumably relates to the role that the funnel plays in locomotion in the different groups. funnel not fused to head laterally; funnel fused to head laterally. Character 5. Funnel valve: (a) (b) A valve with a free anterior margin is found on the inner, dorsal surface of the anterior part of the funnel in all Cephalopoda except for the Octopoda and some genera of the Teuthoidea: Valbyteuthis (Chiroteuthidae) and nine of the thirteen cranchiid genera. Among cranchiids, a funnel valve is found only in Cranchia, Liocranchia, Egea and Megalocranchia. Considering its near univer- sal occurrence among all other oegopsids, it is inferred that a funnel valve was likely found in the most recent common ancestor of the extant cranchiids and, therefore, would best be considered primitive for the family. Though it is commonly believed (Naef, 1921/1923) that the valve functions to pre- vent water from entering the funnel when the mantle is being expanded, Zuev (1967) as- present; absent. 406 sociated the absence of a valve with the loss of the ability to swim headfirst (forward move- ment). Character 6. Ocular photophores: (a) unknown, extinct; (b) four or more, small, simple photo- phores; (c) one, large, complex photophore; (d) one large plus one small, contiguous, complex photophore; (e) one large plus one small, non-contigu- ous, complex photophore; (Г one large plus two small, non-contigu- ous complex photophores. Ocular photophores, found also in many other families of teuthoids, occur in all mem- bers of the Cranchiidae (Fig. 5). Three changes have here been made in the charac- ter state coding employed by N. Voss (1980: Table 1). Firstly, newly-acquired taoniin speci- mens show that the first small, non-contigu- ous photophore grades from “short” to “long- narrow,” without the distinct break that was formerly thought to occur in the group. As a result, the states originally described as “one large plus one small, short non-contiguous photophore” and “one large plus one long, narrow non-contiguous photophore” have been united to read “one large plus one small, non-contiguous, complex photophore.” Sec- ondly, an additinal new state “one large plus two small, non-contiguous, complex photo- phores” is coded for the unique condition ex- hibited by Teuthowenia (described in a foot- note in the original table). Thirdly, more de- tailed study of photophore morphology has resulted in the insertion of “simple” and “com- plex” to express important differences sub- sequently observed. Between the Cranchiinae and the Taoni- inae, there are differences in the appearance, structural morphology, photogenic material and ontogeny of the ocular photophores. In the Cranchiinae, the organs are small, round to oval in shape, and relatively simple in struc- ture (Fig. 5.5), comprised of apparently ecto- dermal invaginations that retain their connec- tions with the ectodermal epithelium (Chun, 1910); consequently, the cup of photogenic tissue has direct contact to the exterior. By contrast, taoniin photophores are markedly dissimilar in size, one of them is usually cres- cent- or sickle-shaped, and all are more com- plex in structure than the corresponding or- gans among cranchiins. The photogenic tis- N. A. VOSS AND R. S. VOSS sue in taoniins is embedded below the sur- face of the photophore in a narrow band along one margin, with the emitted light spread over the wide surface of the organ by means of a thick layer of light guides (Dilly & Herring, 1974; Dilly & Nixon, 1976; Herring, 1977). Studying the ocular photophore in Batho- thauma, Dilly & Herring (1974) found that the photogenic tissue contained paracrystalline material. Herring (1977) later reported the same material in the ocular organs of Egea and Megalocranchia (correct generic identifi- cations for Herring's Phasmatopsis lucifer and P. oceanica respectively) and considered that it probably occurs in all taoniins, in con- trast to the cranchiins in which it does not oc- Cur. Larvae of the majority of cranchiid species, including representatives of every genus, were examined. In all members of the Cranchiinae, the ocular photophores first ap- pear as separate organs in their approximate final adult position. They make their appear- ance in the developing young in groups or singly over varying periods of time until the definitive adult pattern is attained. This is not the case in the Taoniinae. In all of the taoni- ins, a single, poorly-defined patch first ap- pears on the narrow, posteroventral end of the oval, stalked eye of the larva. With growth, the photophore becomes better defined and enlarges to conform approximately to the ventral surface of the eye. This is the only photophore that develops in Helicocranchia and Bathothauma (Fig. 5.1), but in the other genera (Fig. 5.2-5.3) a second and, in Teuthowenia (Fig. 5.4), a third small organ forms as the eye enlarges and gradually be- comes sessile and near-hemispherical in shape. In the larvae of Taonius, Galiteuthis, Mesonychoteuthis and Teuthowenia (Fig. 5.6-5.9), the initial photophore patch extends from the undersurface to along the edge of the narrow, posteroventral end of the eye. Along this edge a thickening and a break oc- cur in the patch to form the second organ which then gradually separates and assumes the final position. The third organ in Teutho- wenia splits off from the inner end of the sec- ond organ as it, in turn, separates from the first. In Sandalops, Liguriella, Egea and Megalo- cranchia, the narrow, posteroventral end of the oval larval eye is extended by a pro- nounced cone-shaped rostrum, or ocular ap- pendage, that J. Young (1970) found (in Bathothauma) to be filled with loose connec- CRANCHIID PHYLOGENY 407 FIG. 5. Eye, left, showing variation in shape of ocular photophores: (1) Bathothauma lyromma, O 4713 (UMML), subadult, 165 mm mantle length; (2) Liguriella podophtalma, WH 417-1-71 (ZMH), subadult, 243 mm mantle length; (3) Galiteuthis glacialis, Elt 1323 (USNM), adult, 333mm mantle length; (4) Teuthowenia sp. B, WH 417-71 (ZMH), subadult, 154 mm mantle length; (5) Leachia atlantica, IOM 1880, adult, 105 mm mantle length. (6-9) Teuthowenia sp. B, ontogenetic series showing development of ocular photophores, anterolateral and ventral views of left eye: (6) ЕЁ 1776 (USNM), 7 mm mantle length; (7) Elt 2270 (USNM), 27 mm mantle length; (8) DMNZ, 60 mm mantle length; (9) SAM A31421, 81 mm mantle length. (10-13) Sandalops sp. C, ontogenetic series showing development of ocular photophores, anterior and ventrolateral views of left eye: (10) Е II (USNM), 27 mm mantle length; (11) ЕЁ 31-24A (USNM), 26 mm mantle length; (12) F IV (USNM), 33 mm mantle length; (13) F IV (USNM), 38 mm mantle length (2 from Voss, 1980) (see Character 6). 408 tive tissue. R. Young (1975b) described a somewhat different development of the photo- phores in Sandalops (Fig. 5.10-5.13). The first, large organ appeared as a patch on the underside of the rostrum, but did not extend to the edge of the apex. With growth, the ros- trum progressively shortened until only a rem- nant remained, on top of which a second ocu- lar photophore appeared. With the disap- pearance of the rostrum, the second organ assumed a contiguous position with the first. An examination of the larvae of Liguriella, Egea and Megalocranchia demonstrated a similar type of development, but in the latter two genera, the second organ subsequently separates from the first as the eye becomes sessile, while in Liguriella and Sandalops the two photophores remain contiguous. N. Voss (1974) did not have an adequate series of larvae with intact eyes to show the details of development of the ocular photophores in Egea. It appears that the separate develop- ment of the two photophores in the above four taoniin genera is a result of the penetration and subsequent division of the tissue of the photophore patch at an early stage by the de- velopment of a pronounced ocular rostrum. The majority of the ocular photophores found among the other teuthoids are of the complex type, with the surface layer of light guides diffusing the light from a photogenic core, similar to that found in the taoniins (Herring, 1977). This is supported by our in- vestigations. In the outgroup, ocular photo- phores are found in the Chiroteuthidae (ma- jority of species), Mastigoteuthidae (one out of numerous species) and the Cycloteuthidae (two out of four nominal species). They are not found in the Promachoteuthidae. In Gri- malditeuthidae, Joubiniteuthidae, the juvenile state of Thysanoteuthidae and a few mem- bers of the Chiroteuthidae, there is a broad, usually thick, highly reflective, gold band sur- rounding the lens and often extending around the ventral surface of the eye. Whether this band contains luminous tissue in any of the groups was not determined. In the Chiroteu- thidae, ocular photophores may occur as long bands, small round organs arranged in rows, or bands of what appear to be incompletely separated round organs. The situation sug- gested to Herring (1977) that the separate round organs coalesce to form the long bands. Instead, the opposite may occur as in the taoniins, where several small organs are derived, at least ontogenetically, from a single large one. Naef (1921/1923) suggested that N. A. VOSS AND R. S. VOSS the primitive form for ocular photophores in cephalopods “may be a diffuse luminescence of the whole skin of the eyeball.” The marked differences in the structure and ontogeny of the ocular photophores between the cranchiins and the taoniins suggest sepa- rate lines of development. A common ances- tral state cannot be confidently identified from among the conditions exhibited by extant cranchiids, and is therefore presumed to be extinct. The ontogenetic findings reported here suggest the existence of a single evolu- tionary trend towards photophore fragmenta- tion in the Taoniinae; the character state “one large, complex photophore” is therefore judged to be the most primitive for this mor- phocline. The ocular photophores of cranchiids ap- pear to function, at least in part, as a ventral camouflage mechanism of advantage to the animal in avoiding predators (R. Young, 1975b). Character 7. Hectocotylus: (a) present; (b) absent. One or more arms of the males of many cephalopods are modified for courtship and copulation. The modification, commonly termed hectocotylization, may be symmetri- cal, equally affecting both arms of a pair, or asymmetrical, affecting only one arm or a pair of arms unequally. There is great diversity in the modification. It may involve the whole arm or only part of the arm and affect any or all of its features—suckers, sucker pedestals, pro- tective membranes, general surface and overall shape and size. The word “hectocotylus” was originally used for the autotomous third (right or left) arm found in certain families of pelagic incir- rate octopods—Tremoctopodidae, Ocythoi- dae, Argonautidae and Alloposidae. The arm is used for insemination. During mating it de- taches from the male and remains within the mantle cavity of the female, carrying with it the spermatophore of the male. In most other incirrate octopods, a lesser modification for insemination is found in which the terminal portion of the same third (right or left) arm is transformed into a discrete organ, called a ligula, which remains attached. This organ is not found in cirrate octopods (G. Voss, per- sonal communication). Steenstrup (1857) considered that the autotomous structure found in Tremoctopus and the other pelagic CRANCHIID PHYLOGENY 409 incirrate octopods mentioned above is but an elaborate modification of the sessile structure found in most of the remaining incirrates. This is Supported by the ontogeny of the structure in Tremoctopus described by Thomas (1977). Instances of lesser symmetrical modification, such as enlarged suckers, are scattered throughout the octopods. The Sepioidea and Teuthoidea display a wider diversity of both asymmetrical and sym- metrical modifications of the male arms. The arms most strongly affected in these two groups are the first and fourth pairs, and, when the modification is asymmetrical, it al- ways involves one of these pairs. In the Sepioidea, the modification is primarily asym- metrical and is found in most of the member families. In the Teuthoidea, asymmetrical modification occurs in both families of myopsids but is only known to occur in six of the twenty-three families of oegopsids— Enoploteuthidae, Lycoteuthidae, Architeuthi- dae, Ommastrephidae, Thysanoteuthidae and Cranchiidae. Various types of symmetri- cal modification of one or more arm pairs fre- quently occurs. The nature of the symmetrical modifications suggests a holding and caress- ing function. The asymmetrical modification occurs earlier in the ontogeny of the animal than do the symmetrical modifications which occur at varying later periods, some appear- ing just prior to maturity. The word “hectocotylus” is commonly used for the single asymmetrically modified arm in the sepioids and teuthoids, which is known in some (and presumed in the remainder) to be used to transfer the spermatophores to the female either by directly grasping the sperma- tophores or by acting as a bridge. Robson (1926) doubted that the modified arm was used in the same way in the octopods as it is in the sepioids and teuthoids, and suggested that the so-called hectocotylus of the latter two orders be termed the nuptial arm. In the sepioids and teuthoids, there is no structure formed that may be termed an “organ” that is common to all members similar to that found in the octopods. Indeed, in some groups many of the details of the asymmetrically modified arm are so bizarre that it is difficult to imagine their function in handling spermato- phores, and some appear to have developed for holding or tactile purposes, perhaps giving the arm a dual purpose. This is not inconsis- tent with the observed use of the modified arm in Octopus (Robson, 1926). In sepioids and teuthoids it is difficult to observe the exact use of the arm because copulation occurs so rapid- ly. Spermatophores, however, have been ob- served on the modified arm during courtship in Sepioteuthis (Arnold, 1965), and transferred to the female by the modified arm during copula- tion in several species of Loligo (Drew, 1911; McGowan, 1954; Hamabe & Shimizu, 1957; Arnold, 1962). Thus it appears that the primary function of the asymmetrically modified arm of the male in these two orders is similar to that in octopods and therefore can be correctly called the hectocotylus, and asymmetrical modifica- tion of both arms of a pair can be referred to as hectocotylization. The occurrence and diversi- ty of structure of the hectocotylus in the sepi- oids and teuthoids suggests that it is polyphy- letically derived (Naef, 1921/1923). N. Voss (1980) called the symmetrical modification of arm pairs or all of the arms of the male “secondary sexual modification” to distin- guish their Known or presumed use of holding or caressing from the primary use of the hectocotylus. The occurrence, position and general form of the hectocotylus are generally correlated with taxon membership (Steenstrup, 1857), and are usually constant within families. In the females of a number of groups, there are dif- ferent structures for the reception of the spermatophores, sperm reservoirs or sperm that correspond to the particular arrangement of the hectocotylus and method of transfer of the spermatophores (Hoyle, 1907). On the grounds of its usually constant occurrence within a family, we have judged the presence of a hectocotylus to be primitive, and its ab- sence to be derived in the Cranchiidae. In the cranchiids, a hectocotylus is only found in the three genera of the Cranchiinae (N. Voss, 1980; Figs. 1b, 2c, 3b). It occurs on the fourth (right or left) arm and is similar in appearance in all species. There is no special structure in the females of either subfamily for the recep- tion of the spermatophores. Throughout the family, spermatophores appear to be trans- ferred directly to the exterior dorsal surface of the mantle; sperm reservoirs have been found embedded in the mantle walls (occasionally in head and arms) and in various stages of emergence into the mantle cavity of mature females in Liocranchia, Leachia, Helico- cranchia, Bathothauma, Sandalops, Galiteu- this, Megalocranchia and Teuthowenia (N. Voss, unpublished notes). The symmetrical or secondary modifica- tions of the arms of the males, which are com- pared in Table 2 of N. Voss (1980), are too 410 variable for use in this study. The modifica- tions, however, are usually similar within a genus. They are more numerous in occur- rence and variable in form in the Taoniinae than in the Cranchiinae. Character 8. Brachial end-organs: (a) absent; (b) present. The brachial end-organ is a leaf or spoon- shaped organ found on the distal ends of arm pairs in near-mature and mature females of some cranchiid genera (N. Voss, 1980; Fig. Зе). It occurs in all species of the Cranchiinae, and in all species of the taoniin genera Egea, Megalocranchia and Teuthowenia. Typically the end-organ appears when the female squid nears maturity and has descended into the deeper waters; at that time, the trabecu- late protective membrane on both sides of the affected arms expands and becomes darkly pigmented. This process is accompanied by a reduction and eventual loss of the suckers, and the oral surface of the affected portion usually becomes rugose or spongy; the ped- estals of the affected suckers may be lost or greatly modified. The end-organ varies in pro- portional size and extent of occurrence on the arms in the different species. Among the cranchiins, the organ occurs only on arms Ill in Liocranchia and in all of the species of Leachia, except for L. danae, and occurs on all of the arms in L. danae and in Cranchia; among the taoniins, it appears on arms Ill in Egea and in some species of Megalocranchia (rarely on arms II), on arms |, ll and III in the remaining species of Megalocranchia, and on all of the arms in Teuthowenia. The organ ranges in size from about 5 to 30% of the arm length, and is approximately the same size on the different arms in a species, except in Cranchia where it is markedly disproportion- ately developed. It tends to be proportionally larger in the cranchiins than in the taoniins. The brachial end-organ is reported to occur only in the Cranchiidae; the collections of the U.S. National Museum, however, contain two large teuthoids, a male and a female as yet unidentified to family, which both display long, similar-appearing organs on the ends of arms IV. In these specimens (kindly shown us by C. F. E. Roper), only the dorsal protective membrane of the arm is modified to form the organ, not the entire oral surface as among cranchiids. The mature stage is not known for many members of the outgroup, and it is pos- sible that some will eventually be found to N. A. VOSS AND R. S. VOSS have brachial end-organs. At this time, how- ever, the presence of these peculiar struc- tures seems best regarded as a derived con- dition for extant cranchiids since its absence is conspicuously more widespread among other teuthoids. The brachial end-organ appears to be a photophore of unique structure that probably functions as a sexual attractant (R. Young, 1975a). Character 9. Clubs: (a) without hooks; (b) with hooklike teeth on large suckers; (c) with hooks. Hooks are found on the tentacular clubs in only four of the twenty-five families of teu- thoids—the Cranchiidae, and three other fam- ilies that are not presently thought to share recent common ancestry with cranchiids, Gonatidae, Enoploteuthidae and Onycho- teuthidae. Hooks are not present on the clubs in any of the members of the outgroup. Among gonatids, hooks occur on the clubs in only one of the two genera where the adult morphologies of the clubs are known. In the large family Enoploteuthidae, hooks are found on the clubs of all species except for those of the genus Pterygioteuthis. In all onychoteu- thid species, the clubs, where known, bear hooks. Of the thirteen genera belonging to the Cranchiidae, only two, Galiteuthis and Mesonychoteuthis, have hook-bearing clubs. In all of the families in which they occur, the hooks are absent in the larvae and first ap- pear in the early or midjuvenile stages. They are formed from typical suckers (Fig. 6.1) of the median one or two rows on the manus, and their appearance is often accompanied by a reduction or loss of the suckers of the marginal rows. The hooks develop by gradual enlargement of a median tooth on the distal margin of the sucker ring (Figs. 6.3-6.8). As the median tooth enlarges, incorporating the lateral teeth, the ring aperture is greatly re- duced; in the process, the outer margin of the sucker is transformed into a hood for the hook. Among cranchiids, an intermediate stage between sucker and hook is found in the members of the genus Taonius (Fig. 6.2); in the postlarval animal, the suckers of the two median rows of the manus elongate and be- come greaty enlarged, with the distal margin of the sucker ring drawn out into one or two large, central, hooklike teeth. The aperture of the ring, however, is not reduced, and the CRANCHIID PHYLOGENY 411 2 FIG. 6. Largest sucker from tentacular club of: (1) Teuthowenia megalops, WH 712-73 (ZMH), subadult, 187 mm mantle length; (2) Taonius pavo, O 4812 (UMML), adult, 540 mm mantle length. (3-8) Galiteuthis glacialis, ontogenetic series showing modification of ring from largest tentacular sucker to form hook: (3) Elt 697 (USNM), 29 mm mantle length; (4) SC 24-62 (USNM), 38 mm mantle length; (5) Elt 935 (USNM), 55 mm mantle length; (6) Elt 949, (USNM), 58 mm mantle length; (7) Elt 943 (USNM), 73 mm mantle length; (8) Elt H371 (USNM), 297 mm mantle length (3-8 redrawn from McSweeny, 1978) (see Character 9). structure presumably can still function as a sucker in these species. As the morphological sequence from suck- er, to hooklike sucker, to hook appears to re- flect increasing functional specialization, so also do the clubs on which the different struc- tures are found; there is a progressive defini- tion of a carpal sucker cluster, reduction of the suckers of the marginal rows of the manus, and reduction of the dactylus and the dorsal keel; the end result of this transformation series is a simpler and more efficient club for capturing and holding soft-bodied animals (Naef, 1921/1923). Ontogenetic evidence and out-group comparisons combine to sug- gest that the ancestral state for the cranchiids is the club without hooks (i.e., solely with typi- cal suckers); clubs with hooklike teeth on the suckers, as in Taonius, and clubs with well developed hooks, as in Galiteuthis and Mesonychoteuthis, appear to be successively derived conditions. Character 10. Digestive gland: (a) stout, spindle-shaped; (b) elongate, spindle-shaped; (c) rounded, with a large photophore. The digestive gland in oegopsids is usually stout and spindle- or ovoid-shaped, and lies at 412 N. A. VOSS AND R. S. VOSS an acute angle to, or parallel with, the longi- tudinal axis of the body. In the Cranchiidae, the digestive gland is spindle-shaped in all genera with the exception of the later growth stages of Megalocranchia species, and is suspended at a right angle to the longitudinal body axis. From dissections and from the liter- ature it appears that this unusual position of the digestive gland, while common in oegop- sid larvae, is found in the adults of only two other teuthoid families, both members of the outgroup—the Grimalditeuthidae, and in some species of the Chiroteuthidae. The elongation of the spindle shape of the gland, as found in the young and subadult of Ligu- riella (adult unknown) and in all growth stages of Bathothauma and Sandalops, appears de- rived from the stout, spindle shape that we hypothesize to be the primitive state for the cranchiids. A large, rounded digestive gland with an associated compound photophore overlying the ink sac characterizes all members of the genus Megalocranchia. In the larva of Megalocranchia, the gland is typically stout and spindle-shaped, with the photophore first appearing in the late larva. With growth, the gland gradually becomes rounded and the photophore proportionally enlarges to cover the entire ventral surface. In the outgroup, a photophore is also found on the ink sac ina number of species of the Chiroteuthidae and in two of the four nominal species of the Cycloteuthidae. Nevertheless, the ontoge- netic derivation of the condition found in Megalocranchia from the commoner photo- phore-less condition of the digestive gland seen among all other cranchiids would ap- pear to argue that the presence of a photo- phore on the gland is an independently de- rived condition. R. Young (1975b, 1977) suggests that the spindle shape and vertical orientation of the opaque digestive gland, by reducing the ven- tral countershading problem of the animal, and the photophore on the large, rounded di- gestive gland in Megalocranchia, by its coun- tershading luminescence, are devices for ventral camouflage. Character 11. Digestive duct appendages: (a) on ducts; (b) on ducts and digestive gland; (c) on digestive gland. From the researches of Bidder (1966, 1976) and Schipp & von Boletzky (1975, 1976), among others, it appears that the structure and function of the digestive duct appendages, which are formed from the di- gestive ducts, differ between the Octopoda, Sepioidea and Teuthoidea, and can be re- lated to the different position of the organ in each group. In octopods, the appendages are found on the posteroventral surface of the di- gestive gland and lie within its connective tis- sue envelope, while in the sepioids and teuthoids, the appendages are found outside of the envelope of the digestive gland and are covered by renal epithelium. Among sepioids, the digestive duct appendages always occur as grapelike follicles on the digestive ducts and are in close topical relationship with “renal” epithelium; by contrast, the position and gross morphology of the appendages are variable in the teuthoids, often within families and sometimes within genera. Greater varia- tion is found among cranchiids than among members of the outgroup. In the Cranchiidae, digestive duct append- ages may occur on the ducts, on the ducts and digestive gland, or on the digestive gland alone. The state “on the ducts” is a correction of the state mistakenly described in Table 1 of N. Voss (1980) as “on posterior end of ducts or on caecum.” The generic definitions given in the text and a re-examination of the speci- mens support this correction. The append- ages appear in the form of two large, com- pound lobes on the posterior portion of the united duct in Leachia, and in the form of small clusters of follicles on the posterior por- tion of the separate ducts in Megalocranchia. In Taonius, Egea, Teuthowenia and two spe- cies of Galiteuthis the appendages are in the form of two large, compound lobes on the posterodorsal surface of the digestive gland at the exit of the digestive ducts and in the form of small clusters of follicles on the entire length of the ducts. In the remaining seven genera, Cranchia, Liocranchia, Helico- cranchia, Bathothauma, Sandalops, Liguri- ella, Mesonychoteuthis and four species of Galiteuthis, the appendages occur as two large, compound lobes in the same position on the digestive gland as in the preceding group. In the outgroup, the appendages appear as small clusters of follicles on the entire length of the ducts in Thysanoteuthidae, Cycloteu- thidae (one species), Mastigoteuthidae (two species), Promachoteuthidae and Joubiniteu- thidae. They appear as a thick coating of spongy tissue on the major or entire length of CRANCHIID PHYLOGENY 413 the ducts in Chiroteuthidae (one species), Grimalditeuthidae and Cycloteuthidae (one species). In the remaining five species of chiroteuthids and five species of mastigoteu- thids examined, the appendages occur as medium to large, compound lobes on portions of the ducts. Thus, it would appear that the position of the digestive duct appendages on the ducts is most parsimoniously regarded as ancestral for extant cranchiids, and that the presence of these appendages on the diges- tive gland is likely a derived condition. The investigations of Schipp & von Boletzky (1975, 1976) suggest that the digestive duct appendages in the sepioids play a role in ex- cretion and nutrient absorption as well as osmoregulation and urine formation. Our present lack of knowledge of the fine structure of the appendages in the teuthoids, however, precludes meaningful speculation on the functional significance of the differences in their position and gross morphology. Character 12. Caecum: (a) smaller than stomach; (b) larger than stomach. The relative size of the caecum and the stomach varies within the oegopsids. In the Cranchiidae, the caecum is larger than the stomach in the three genera of the Cranchi- inae and is smaller than the stomach in the ten genera of the Taoniinae. An examination of as many members as possible of the out- group revealed that the caecum is larger than the stomach in the Cycloteuthidae and the Promachoteuthidae, larger than or approxi- mately the same size as the stomach in the Mastigoteuthidae, and is smaller than the stomach in the Thysanoteuthidae, Chiroteu- thidae, Grimalditeuthidae and Joubiniteu- thidae. Considering that the caecum is smaller than the stomach in the majority of the outgroup members, we are inclined to regard that state as primitive for the extant cranchi- ids. The functional significance of the relative size differences of the caecum and the stom- ach is not known but might reflect differences in feeding habits (see Bidder, 1966). Character 13. Eyes of larvae: (a) sessile; (b) stalked. Stalked eyes are found in the larvae of all cranchiids with the exception of Cranchia and Liocranchia. The length of the larval eye stalks and the period of their persistence varies considerably among the ontogenies of the different genera. Though markedly pro- truding eyes are found in the larvae of some teuthoids, for example, in the Octopodoteu- thidae, Thysanoteuthidae and at least one of the four nominal species of Cycloteuthidae, there is no known occurrence of stalked eyes in cephalopod larvae outside of the Cranchi- idae. The absence of stalked eyes in the larvae of all other known cephalopods (and of two genera of the Cranchiidae, and the varia- bility of the character within the remaining members of the family) would indicate that sessile eyes may be considered primitive for cranchiids. N. Voss (1980), however, referred to the loss of the character of stalked eyes in Cranchia and Liocranchia. The present, broader analysis of this character, suggests that the contrary is true, i.e. that sessile eyes are retained in these two genera as an un- modified ancestral state and that the pres- ence of stalked eyes in the remaining cranchi- id genera is likely derived. Clarke et al. (1979) support the suggestion made by J. Young (1970) that the eye stalks of cranchiids may contain ammonium to pro- vide buoyancy, but eye stalks may be of addi- tional advantage to the larva by providing greater mobility to the eyes, thereby affording broader vision (J. Young, 1970; R. Young, 1975b; Weihs & Moser, 1981). The loss of the eye stalks with growth can be related to the vertical distribution of the animal (R. Young, 1975a,b). The length of time that the larvae spend in the shallower waters appears to cor- respond with the varying persistence of eye stalks in the different species, but does not necessarily correspond with the degree of de- velopment of the stalks. Character 14. Dorsal pad of funnel organ: (a) one median papilla plus two lateral flaps; (b) one median flap plus two lateral flaps; (c) two lateral flaps; (d) one median papilla plus two to six markedly flattened, lateral papillae; (e) one median papilla plus two round or elliptical, lateral papillae; (f) two lateral papillae. The funnel organ, comprised of one or more pads of mucus-secreting epithelium, is found on the inner surface of the funnel in all cephalopods. Usually located in the middle 414 N. A. VOSS AND R. S. VOSS part of the funnel, the organ is sometimes confined to the dorsal surface, as in nauti- loids, some octopods, and Vampyroteuthis, or may be found on both the dorsal and ventral surfaces, as in the majority of cephalopods. In octopods, the organ is generally W-shaped, but the lateralmost of the vertical bars are sometimes separate, or the organ may take the form of two modified V-shaped pads. In sepioids and teuthoids, the funnel organ is typically three parted—an inverted V- or U- shaped dorsal pad and two paired, usually oval and elliptical-shaped, ventral pads. Vari- ations in the size, outline and surface sculp- ture of these two basic forms of pads is con- siderable, especially among oegopsids. Vari- ation is greater in some families than in others, and is displayed to the highest degree among cranchiids. The sculpture of the dorsal member of the funnel organ has received the most taxonom- ic attention. In the Cranchiidae, the dorsal pad always has sculpture on the lateral arms. The pad may exhibit one median papilla plus two lateral flaps (Sandalops), one median flap plus two lateral flaps (Cranchia and Lio- cranchia), two lateral flaps (Egea and Megalocranchia), one median papilla plus two to six markedly flattened, lateral papillae (Leachia), one median papilla plus two round or elliptical, lateral papillae (Helicocranchia, Liguriella, Taonius, Galiteuthis, Mesonycho- teuthis and Teuthowenia), or two lateral papillae (Bathothauma) (Fig. 7). The lateral flaps that occur in Sandalops, Cranchia, Liocranchia, Egea and Megalo- cranchia are all longitudinally (i.e. antero- posteriorly) oriented. In Leachia, the flat- tened, lateral papillae are also longitudinally oriented, and when the lateral papillae are multiple on a side, they form along a single anteroposterior line and are sometimes con- nected by a basal ridge, all suggesting that the papillae have developed from a longitudi- nal flap. There is a trend in Leachia toward multiple lateral papillae; the number of papil- lae may vary within a species or an individual, where sometimes there is a single papilla on one side and two on the other. In the majority of the members of the outgrop, the dorsal pad is unsculptured except for a median papilla. Several members have a longitudinal ridge, FIG. 7. Dorsal pad of funnel organ: (1) Sandalops sp. C, Cl 71-6-26 (USNM), adult, 102 mm mantle length; (2) Cranchia scabra, WH 439-11-71 (ZMH), subadult, 122 mm mantle length; (3) Egea inermis, WH 471-11-71 (ZMH), subadult, 207 mm mantle length; (4) Leachia danae, MV 65-1-53 (SIO), subadult, 167 mm mantle length; (5) Galiteuthis glacialis, Elt 1323 (USNM), subadult, 333 mm mantle length; (6) Bathothauma sp. D, C 108662 (AM), subadult, 187 mm mantle length (1, 3 from N. Voss, 1980; 5 redrawn from McSweeny, 1978) (see Character 14). CRANCHIID PHYLOGENY 415 developed to various extents, either on the lateral or median sections of the pad or on both. The occurrence of a median papilla in the majority of the cranchiids, and of a longitudi- nally oriented flap, or its apparent modifica- tions, on each lateral arm in nearly half of the family, is similar to the occurrence in the out- group of a median papilla in most of the mem- bers, and of a longitudinal ridge on the lateral arm when lateral sculpture is present. This suggests that a dorsal pad with one median papilla and two lateral flaps as found now only in Sandalops, might be considered primitive to the cranchiids. The remaining character states appear to be independently derived except for “(f) two lateral papillae,” which is hypothesized to have been derived from “(e) one median papilla plus two round or ellipti- cal, lateral papillae.” Nesis (1974), in his analysis of the sculpture of the dorsal pad in the Taoniinae, concluded that the state “one median papilla and two lateral papillae” was basic to the subfamily and that the flaps were derived. His conclusions resulted from analy- sis of the distribution of character states only in the taoniins, however; he did not study the family as a whole nor compare the character as it occurs in other families considered to be allied. Our conclusions appear supported by a broader comparative approach. Though the function of the funnel organ and the significance of its variations are not known, it has been considered that the mucus produced is used for keeping the funnel and perhaps mantle cavity clean of debris. An alternate or additional function is suggested by the observations of Hall (1956), Nicol (1964) and M. R. Clarke (as reported by Dilly & Nixon, 1976) that the mucus might serve as a carrier for the ink produced by the ink sac and expelled through the funnel by the animal when irritated. 2. A Phylogenetic Hypothesis Application of the Wagner method to the binary data matrix of Table 2 yields the recon- struction of cranchiid evolutionary history il- lustrated in Fig. 8. In Fig. 8, internal nodes (branching points) represent hypothetical an- cestors and are labelled with capital letters; external nodes (branch tips) represent extant cranchiids and are labelled with the first three L FIG. 8. Wagner reconstruction of cranchiid phylogeny. See text for explanation. 416 N. A. VOSS AND R. S. VOSS letters of the generic name; lines connecting the nodes represent phyletic lineages and are drawn proportional to the estimated amounts of morphological evolution (number of charac- ter state transitions) that separate extant cranchiids from their hypothetical ancestors or hypothetical ancestors from one another. Some extant cranchiids are indistinguishable from their most recent shared ancestors with respect to the characters employed in this study; the external nodes representing these forms (e.g., Liguriella) are drawn as open cir- cles and have been removed an arbitrary one branch length unit from their most recent an- cestors. Phenotypes of hypothetical ances- tors are provided in Table 3. The Wagner Tree hypothesizes a basal separation of the Cranchiidae into two phy- letic lineages that correspond in membership to the traditional subfamilies Cranchiinae and Taoniinae, the Cranchiinae containing Cranchia, Liocranchia and Leachia, and the Taoniinae comprised by the remaining ten genera. Within the taoniin clade, three major generic assemblages can be discerned, two of which, the group Megalocranchia + Egea + Teuthowenia and the group Taonius + Galiteuthis + Mesonychoteuthis, are further hypothesized to have shared a common an- cestor more recently than either did with members of a third group consisting of Sandalops + Liguriella + Helicocranchia + Bathothauma. All relationships are fully re- solved in this estimate of cranchiid phylogeny, and the topology of the tree requires a mini- mum of 45 character state transitions in order to derive observed phenotypes of extant cranchiids from the morphology of the com- mon ancestor estimated in the preceding sec- tion; the consistency index (Kluge & Farris, 1969) for the Wagner Tree is .69, indicating a remarkably good fit of hypothesis to data. Because it cannot be known with certainty, however, that the Wagner Tree is actually the most parsimonious of all possible reconstruc- tions of cranchiid relationships, the binary data matrix of Table 2 was subjected to Com- patibility Analysis in order to develop testable alternatives. The 31 binary factors of our data form 21 cliques of mutually compatible mem- bers, and each of these cliques supports one (Or more) estimate(s) of cranchiid evolution that is (are) not supported by any other clique. The compatibility matrix for the binary factors is presented in Table 4, character member- ships of the ten largest cliques are provided in TABLE 3. Reconstructed phenotypes of hypotheti- cal cranchiid ancestors. Columns represent the hypothetical ancestors labelled with capital letters in Fig. 8. Character numbers and character state labels are the same as those in Table 1 and de- scribed in the text. Ancestors Character number АВС ЕЕ ЗФ НТК 1 Баааасс аа чеаа 2 сссоссаа ъь ааа 3 c cb b b d a. a avarana 4 aa a a aaa a ава 5 a bb b b b bb в маа 6 e-e e e e ec d. dd: bibka 7£ bb b b b bb ааа 8 bb a a aaa a a bebea 9 аа с bp aa a “a ‘agamama 10 aa a. a ab b b аммыаа 11 Bb ib Бес с Decaama 12 aaa aa ara аа 5 ыа 13 bb b bi bb br be Бабада 14 ceeeeeeeebaa Table 5, and the cladograms supported by the two largest cliques are drawn in Fig. 9. Cliques | and II, whose trees are drawn in Fig. 9, share 21 binary factors (1b, 1d, 2c, 2d, 3b, 3c, 3d, 4b, 6b, 6c, 6e, 6f, 7b, 9b, 9c, 10c, 12b, 14b, 14c, 14d, 14f), and this large set of characters determines those cladistic pat- terns common to both compatibility trees and to the results of Wagner analysis. Disagree- ment between the two trees of Fig. 9 reflects underlying differences in clique memberships and concerns only the relationships of Sandalops and Liguriella within the Taoni- inae. Clique | differs from clique II by the in- clusion of binary factor 1c which asserts that Bathothauma, Helicocranchia and Sanda- lops comprise a monophyletic group, but leaves the relationships of Liguriella unre- solved. Clique Il omits factor 1c but includes factor 6d whose effect is to remove Sanda- lops and Liguriella, but not Helicocranchia or Bathothauma, to a monophyletic group with the remaining six taoniin genera; the relation- ships of Sandalops and Liguriella within the latter group are unresolved, however. The trees supported by cliques | and Il both include trichotomies because binary charac- ters that might fully resolve the relationships of Liguriella and/or Sandalops do not support other aspects of the cladograms drawn in Fig. 9. In order to test the compatibility results CRANCHIID PHYLOGENY 417 TABLE 4. Compatibility matrix for the binary factors whose distributions are provided in Table 2. Because the matrix is symmetrical, only the lower half is illustrated. Rows and columns are binary factors; an entry of (0) for a given row and column signifies that the corresponding pair of binary factors is not compatible, an entry of (1) that the pair of factors is compatible. оао 2102 dam -“2-2-2-20 -00 2-22 mm dl ll ll ll dl dl mn Om = = = = = © 101100000101 200212000002 ua HO 1002-10 dl ll dl dl LAO) = = LL — aaa do a a a a a a a a do do dd a a dd dd HO 1001-1102 = = = = © = — — HO mOn mm mm mn = nn = = O- + iii dl dl 2 2 2020-0 dl ld dl dl dl dl dl = = — aa y On OO A Ll dl dl ld dl dl dd =O AA mt E MSM MS) AA e 0101010 aa a 4 2O 200» dl Ll a a = a — — Ad dl dl dl dl nr OO = OA — on “O1 100-010 —- — — 2Oa a 1 1 OO = 2 AO EC Cee ann mm mn ni à dl dd à OO = = = O) 1 O = = = O = OO = = —= — XA A a 2 ld ld à O = on mm dd dl dl + O» = — Oh ld ll ann E A = O Aa 2 НТ о ADO — Oh = OO —b — + + + CE D = = = À ее Е, mb mb os o's ait 9c 10b 10c 11b 11c 12b 13b 14b 14c 14d ide Binary factor label TABLE 5. Memberships for the ten largest cliques of mutually compatible binary factors. Clique number Membership | 1b, 1c, 1d, 2c, 2d, 3b, 3c, 3d, 4b, 6b, 6c, 6e, 6+, 7b, 9b, 9c, 10c, 125, 145, 14c, 144, 144. Il 1b, 1d, 2c, 2d, 3b, 3c, 3d, 4b, 6b, 6c, 6d, 6e, 6f, 7b, 9b, 9c, 10c, 12b, 14b, 14c, 14d, 14f. Ш 16, 1d, 2c, 2d, 3b, 3c, 4b, 6b, 6c, 6e, 64, 7b, 9b, 9c, 10, 10c, 125, 145, 14c, 14d, 144. IV 1b, 1c, 2c, 2d, 3b, 3c, 3d, 6c, 6e, 6f, 7b, 9b, 9c, 10c, 13b, 14b, 14c, 14d, 14f. V 1b, 2c, 2d, 3b, 3c, 3d, 6c, 6d, 6e, 6f, 7b, 9b, 9c, 10c, 13b, 14b, 14c, 14d, 144. VI 1b, 1d, 2d, 3d, 4b, 6b, 6c, 6f, 7b, 9b, 9c, 10c, 12b, 14b, 14c, 14d, 14e, 14f. VII 1b, 1c, 1d, 2d, 3c, 3d, 4b, 6b, 6f, 8b, 9b, 9c, 10c, 12b, 14b, 14c, 14d, 14f. VIII 1b, 2c, 2d, 3b, 3c, 6c, 6e, 6f, 7b, 9b, 9c, 10b, 10c, 13b, 14b, 14c, 14d, 14f. IX 1b, 1d, 2d, 3c, 4b, 6b, 6f, 8b, 9b, 9c, 10b, 10c, 12b, 14b, 14c, 14d, 14f. X 1b, 1c, 2b, 2c, 2d, 3b, 3c, 3d, 6e, 6f, 9b, 9c, 10c, 14c, 14d, 14f. 418 O O — = E = = o 5 © O =! OU = ca 5 E n — an [= 9 = E O © O ar OU at a AG 179) © = Lig N. A. VOSS AND R. S. VOSS > — o 2 © Ф 5 lo] © © © > ©) r = = wu clique | n == © Oo 2 © 4 o O © = o 2 ©) = + = wu clique || FIG. 9. Cladograms corresponding to the estimates of cranchiid evolution supported by cliques | and II (see Table 5). Branch lengths are arbitrary and do not represent any estimated parameter of phylogeny. against those of the Wagner analysis, how- ever, it is convenient to resolve such tricho- tomies fully. To each trichotomy in a clado- gram there correspond three completely bi- furcating alternative interpretations (Nelson & Platnick, 1980), and the six alternatives that result from so interpreting the ambiguities of Fig. 9 are shown in Fig. 10. Only that por- tion of the taoniin lineage descended from the cranchiid ancestor but ancestral to the monophyletic group Megalocranchia + Egea + Teuthowenia + Taonius + Galiteuthis + Mesonychoteuthis is depicted for each vari- ant; the unillustrated portions of the clado- grams of Fig. 10 are identical to those elicited in all preceding analyses. We note that one of CRANCHIID PHYLOGENY 419 5756 5 Do o 5 [= а SE (Tp) = Ber) со ГА IB a EC о 5 са JE (Tp) “sj со ЗЕ НА НВ Hel Lig с => — (5 с с Ф le] о WY co as WY = ГС с — (= 5 5 © 5 о n со I: (Tp) er ИС FIG. 10. Three bifurcating interpretations for each of the trichotomous branchings in the cladograms of Fig. 9. See text for explanation. the bifurcating interpretations of the tree de- termined by clique | is identical with the Wagner Tree. A modified version (see Materials and Methods) of the parsimony-optimizing pro- cedure of Farris (1970) was used to fit ob- served character state distributions (Table 1) to the six variants in order to determine which of them provides the most parsimonious in- terpretation of cranchiid phylogeny. The Wagner Tree (Fig. 8 and 1A of Fig. 10) with 45 required character state transitions proved most parsimonious, followed by trees IC, ПА and ИС with 46 necessary transitions apiece, and trees IB and IIB with 47 transitions each. The character state transition in which two of 420 N. A. VOSS AND R. S. VOSS these hypotheses differ are illustrated in Fig. 11. As can be seen, while both hypotheses ‘explain’ the same observed phenotypes for the four genera diagrammed, they differ in the simplicity with which they do so. DISCUSSION In the absence of fossil cranchiids and of a priori knowledge of character conservatism among these squids, the principle of parsi- mony appears to us the only defensible cri- terion with which to test alternative hypotheses of cranchiid phylogeny. However, because more than 300 billion different bifurcating tree diagrams could be drawn to unite our 13 ter- minal taxa (Felsenstein, 1978) exhaustive testing by any criterion is clearly impractical. The purpose of applying operational phylo- genetic techniques, as those employed here, is simply to reduce this vast array of possibles to a much smaller set of well-corroborated alternatives among which the true phylogeny has a reasonably high likelihood of being in- cluded. The close congruence revealed above between the results of Wagner and of Character Compatibility analyses lends cre- dence to the possibility that the tree diagram Bat Hel San Lig of Fig. 8 represents, if not historical truth ex- actly, then at least an estimate sufficiently close that an examination of the details of the reconstruction will not be far wrong. Although Pfeffer (1912) and Nesis (1974) previously discussed phylogenetic relationships among cranchiids, the inadequate materials available to them resulted in such taxonomic confusion as to effectively preclude meaningful com- parisons of their conclusions with our results; the reader is referred to N. Voss (1980) for a discussion of their generic assignments. A basal division of the Cranchiidae into the traditional subfamilies Cranchiinae and Taoniinae is supported by five characters (1, 4, 6, 7, 12), and character state transitions separating the cranchiin and taoniin ances- tors (Table 3) account for nearly 30% of all of the morphological evolution estimated to have occurred in the course of cranchiid phylogeny. Apparently unique synapomorphies uniting the three cranchiin genera are the cartilagi- nous strengthenings along one or both of the paired ventral lines of funnel-mantle fusion (1d), the lateral fusion of the funnel to the head (4b), the possession of four or more small and simple ocular photophores (6b) and of a caecum larger than the stomach (12b). Unique synapomorphies to support a hy- Bat Hel San o = FIG. 11. Two alternative estimates of relationships for four cranchiid genera. The labelled slashes across branches of the tree signify the evolution of the corresponding character state from a locally more plesio- morphic condition. Only those character state transitions in which the two hypotheses differ are illustrated. CRANCHIID PHYLOGENY 421 pothesis of taoniin monophyly are the loss of the hectocotylus (7b) and the possession of ocular photophores of complex construction (6c-f). Other derived conditions that may have characterized the cranchiin or taoniin ancestor are the results of non-unique character state transitions (i.e. those replicated or reversed elsewhere) that do not, therefore, support the cranchiin-taoniin dichotomy per se. Within the Cranchiinae, the genera Cranchia and Liocranchia form a monophy- letic group that is supported by a derived morphology of the funnel organ (14b) as well as by the position of the digestive duct ap- pendages on the digestive gland alone (1 1c); the latter state, however, is shared with some taoniin genera as well (see below). Leachia appears well separated from the other two cranchiins in the characters discussed above and in the morphology of funnel-mantle fusion (see discussion of Character 1) and shares with the Taoniinae the derived absence of a funnel valve (5b; except Egea and Megalo- cranchia) and the presence of stalked larval eyes (13b). These last two anomalous traits may either represent convergent evolution of the derived conditions in question, or we may be mistaken in assuming sessile larval eyes and the presence of a funnel valve to be primi- tive for the family. If it is our estimation of polarities that is in error, then a funnel valve and sessile larval eyes are synapomorphies for Cranchia and Liocranchia. The major monophyletic clusters identifia- ble within the Taoniinae include the generic groups recognized by N. Voss (1980), but substantial refinements of earlier hypotheses of taoniin interrelationships are also repre- sented in Fig. 8. The Megalocranchia group (Megalocranchia + Egea + Teuthowenia) and the Taonius group (Taonius + Galiteuthis + Mesonychoteuthis) together comprise a monophyletic unit defined by shared, derived aspects of gladius morphology (2c), fin shape (3b,c) and ocular photophore arrangement (6e,f) that appears well separated from its putative (see below) sister group, the Sanda- lops assemblage (Sandalops + Liguriella + Bathothauma + Helicocranchia). Monophyly of the Megalocranchia group is supported by shared possession of elongat- ed, terminal-lateral fins (3c) and by the pres- ence of brachial end-organs on the arms of mature females (8b). Brachial end-organs, however, also occur among cranchiin squids and seem best regarded as another instance of convergent evolution. To argue otherwise, for example that cranchiins and the Megalo- cranchia group form a monophyletic assem- blage by virtue of a unique derivation of brachial end-organs, would necessarily in- voke homoplasy in so many other characters (e.g. 2, 3, 6, 7) as to be extravagantly un- parsimonious. Within the Megalocranchia group the genera Megalocranchia and Egea form a morphologically distinctive pair as noted by N. Voss (1980: 406). Members of the Taonius group uniquely share the derived presence of hooks or of hooklike teeth on the larger suckers of the clubs (9b,c). Taonius, Galiteuthis and Mesonychoteuthis are also united by having lanceolate or stout, ovate, terminal fins (3b), a character state derived for the Taoniinae as a whole but a plesiomorph within the mono- phyletic assemblage that includes Megalo- cranchia and its allies. Mesonychoteuthis and Galiteuthis both exhibit clubs with well-devel- oped hooks (9c), a uniquely derived condition not shared with Taonius. Four species of Galiteuthis share, with Mesonychoteuthis, the (not uniquely) derived position of digestive duct appendages on the digestive gland alone (11c), but two other species of Galiteu- this share with Taonius the more plesiomor- phic position of appendages on both the gland and the digestive ducts (11b). All of the analyses reported here were repeated, using either 11b or 11c to characterize Galiteuthis, with identical results: the cladistic position of the genus was unaffected by the substitution. The distinctiveness of neither Mesonychoteu- this nor Galiteuthis is compromised by the interspecific variation in Character 11 ob- served within the latter; the two genera are well defined with respect to other morpho- logical features discussed by N. Voss (1980: 392-396). All of the relationships discussed above are common to the results of both Wagner and Character Compatibility analyses and appear adequately supported by the comparative morphological evidence at our disposal. Re- grettably, the same cannot be said of any ar- rangement of the genera Bathothauma, Helicocranchia, Sandalops and Liguriella along the phyletic line descended from the cranchiid ancestor but ancestral to the Taonius and Megalocranchia groups. While the hypothesis that the four genera of the Sandalops assemblage form a monophyletic cluster is slightly more sparing of character state transitions than any of the other five al- ternatives treated here (Fig. 10), we would 422 N. A. VOSS AND R. S. VOSS point out that no unique synapomorphy can be adduced in support of this arrangement. Instead, members of the Sandalops group are united by character states that are shared by other cranchiids as well (4a, 5b, 7b, 8a, 9a, 11c, 12a, 13b) and evidence for their near affinity is therefore largely by phenetic simi- larity. By contrast, derived resemblances in ocu- lar photophores (6d and derivatives) argue that Sandalops and Liguriella form a mono- phyletic unit with the Taonius and Megalo- cranchia groups that does not include Batho- thauma or Helicocranchia (tree Il of Fig. 9) while the shared, derived possession of nar- row, Straight funnel-mantle fusion cartilages supports the inclusion of Bathothauma, Helicocranchia and Sandalops, but not Liguriella, in a different monophyletic arrange- ment (tree | of Fig. 9). The most parsimonious arrangement (Fig. 8 and IA of Fig. 10) is sup- ported weakly by assuming the presence of digestive duct appendages on the digestive gland alone (11c) to be a local synapo- morphy; the condition is shared with Mesony- choteuthis, one group of Galiteuthis species, Cranchia and Liocranchia, however, and the hypothesis that the Sandalops assemblage is monophyletic should be regarded as a best guess among alternatives but slightly less parsimonious; only the discovery of new char- acters seems likely to satisfactorily resolve the phyletic structure of this problem group. Bathothauma and Helicocranchia share morphologies of the fins (3d) and ocular photophores (6c) that are unique among adult cranchiids though widespread in the larval stages of other genera; as adult features of Bathothauma and Helicocranchia, the traits appear to represent derived, neotenous con- ditions. Cranchiid adaptive radiation appears to have involved, to a significant degree, the evolution of differing schedules of ontogenetic descent in the water column (N. Voss, unpub- lished notes), and several of the monophyletic groups discussed in the preceding para- graphs may be characterized by the ecologi- cal distribution of the growth stages of their member taxa. Thus, the three cranchiin genera on the one hand and the genera of the Megalocranchia group on the other represent apparently independent clades whose larvae (with the single known exception of Lio- cranchia valdiviae) nevertheless resemble one another ecologically by remaining in the upper waters for longer periods in their devel- opment than do larvae of other cranchiid groups. This ecological resemblance, either convergently evolved or inherited unmodified from the cranchiid ancestor, might account for the peculiar similarities between the groups in possession of brachial end-organs and (be- tween Liocranchia-Cranchia and Едеа- Megalocranchia) in the presence of a valve in the funnel. However, as we know so little of the adaptive significance of either anatomical trait, their causal relationships (if any) to onto- genetic lingering in the upper waters are unclear. The Taonius group, in contrast, consists of cranchiids that typically descend to mid and deep water at a much earlier immature stage than do members of the Cranchiinae or of the Megalocranchia group. It may be noted in passing that the phenomena of early onto- genetic descent displayed by the Taonius group might be related to the extended geo- graphic distribution of the member genera. Though circumglobal distribution in tropical and subtropical waters is typical for most of the cranchiid genera, the ranges of the Taonius assemblage extend into subpolar and polar regions, with Mesonychoteuthis restricted primarily to Antarctic waters. Geo- graphic range extension into cold waters is not unique to this group, however, for Teuthowenia, and, to a lesser extent, Liguri- ella and Bathothauma are also found in sub- polar waters, but in a more limited pattern of distribution. The Sandalops group is not easily defined in ecological terms and the lack of morphological cohesiveness remarked earlier for these squids may be a reflection of the apparent absence of any ecological distinctiveness. Functional correlations appear to have con- tributed little, if at all, to the hierarchic pattern of character state distributions revealed in the preceding analyses: the phylogeny is sup- ported by morphological features associated with such a diversity of biological activities (see discussions of Characters 1, 3, 4, 7, 9, 12) that we think it unlikely that our estimate of relationships reflects divergence in only a single co-adapted anatomical complex or functional role. Of the characters treated here, only two seem obviously associated in a close functional sense: the form of the posterior end of the gladius (Character 2) and the shape of the fins (Character 3). As discussed previously in the analyses of these characters, changes in the shape of the posterior end of the gladius are usually accompanied by changes of fin CRANCHIID PHYLOGENY 423 Character 2 Character 8 FIG. 12. Enlarged tree diagrams for Characters 2 and 3. Lower case letters label the states of Characters 2 and 3, drawn as boxes containing the cranchiid genera that exhibit the appropriate morphological condition; arrows indicate polarities hypothesized in the text. The two characters are seen to constitute different (but compatible) partial estimates of cranchiid relationships. shape, perhaps for reasons of structural sup- port. Nevertheless, as can be seen from the enlarged character state trees in Fig. 12, the relationship between fins and gladius is evi- dently not wholly deterministic, and the two characters each contribute some phyletic in- formation not contained in the other. We ob- serve that Character 2 does not, in fact, sup- port our hypothesis of phylogeny while Char- acter 3 does. The fact that not all of the characters we studied are pairwise compatible (Table 4) is sufficient demonstration of the existence of homoplasy in the course of cranchiid evolu- tion. If the tree topology of Fig. 8 and the re- constructed ancestral phenotypes of Table 3 be accepted as reasonable estimations, then the minimal amount of homoplasy in each character commensurable with those esti- mates is easily determined and may provide an approximate measure of conservatism that might inform the choice and weighting of char- acters in subsequent systematic studies (see also Farris, 1969). Characters 1, 3, 4, 7, 9, and 12 support the estimate of Fig. 8; if that estimate is taken to be correct, then these characters, in addition to being mutually compatible, are also true characters: they have undergone no homo- plasy in the course of cranchiid evolution. Characters 2, 5, 6, 8, 10, 11, 13 and 14 have all undergone one or more instances of con- vergence or reversal, of which those involving brachial end-organs (Character 8), larval eye position (13) and a funnel valve (5) have al- ready been discussed as examples above. Most of these latter characters have under- gone but one or two instances of homoplasy, and we would hesitate, based on this obser- vation alone, to enjoin caution in their use in future phylogenetic investigations, but Char- acter 11 is an exception. Over the course of cranchiid phylogeny, digestive duct append- ages appear to have migrated on and off the digestive duct and gland with abandon. Primi- tively situated on the digestive duct (see analysis for Character 11, above), append- ages are here interpreted to have moved onto the digestive gland in the common ancestor of Cranchia and Liocranchia, and in the taoniin ancestor, to have reverted to the ancestral state in Megalocranchia, and to have van- ished from the ducts entirely in Cranchia and Liocranchia, in the ancestor of the Sandalops group and in Mesonychoteuthis and some species of Galiteuthis. Evidently, the digestive duct appendages are evolutionarily labile structures, and it would be interesting to know what adaptive significance accruing to their anatomical positions makes them so. CONCLUDING REMARKS The form of a phylogenetic hypothesis, the topology of a tree diagram, results both from the analysis of individual characters and from the procedures subsequently employed to re- 424 N. A. VOSS AND R. S. VOSS solve character conflicts. We have endeav- ored to be as explicit as possible about each step that led us to adopt the hypothesis pre- sented here so that would-be critics can dis- cover exactly where we may have gone wrong and set about directly to correct the error. Because errors in phylogeny recon- struction, when they exist, usually consist of mistakes in determining homologies or in esti- mating polarities, the greater part of this paper is devoted to individual character dis- cussions, and the future, critical tests of our phylogeny that we hope to have provoked will perforce consist either of discovering new characters or of more detailed analyses of the characters treated here. In neither case will materials be found wanting. As sources of new characters, for example, the myology of teuthoids remains little explored; the mor- phology of the cranial cartilages and the spermotophores likewise invites attention as does the comparative anatomy of the nerv- Ous, reproductive and circulatory systems. Of characters treated here, the histology of the ocular photophores and of the brachial end- organs is in need of study, and careful ob- servations of courtship and mating behavior may confidently be expected to permit more informed treatment of the hectocotylus and of other male sexual modifications of the arms. Additionally, we know little or nothing of the functional significance of variations in the form of the funnel organ, of the larval eye Stalks, or of the relative size of the caecum and stomach to name but three of many enigmatic aspects of cranchiid morphological variation. References provided in the individ- ual character discussions will provide intro- ductions to these and other promising areas of teuthoid morphological research; we know of few animal groups in which the potential for innovative and phylogenetically rewarding comparative studies appears so great. ACKNOWLEDGMENTS G. K. Creighton, G. F. Estabrook, A. G. Kluge, G. L. Voss and R. E. Young critically read the draft manuscript; their comments and suggestions are sincerely appreciated. We thank R. Toll for reading the early version of our discussion on the gladius and for pro- viding helpful criticism and new information. Thanks are extended to the curators and staffs of the numerous institutions that have continued their loans of specimens for the Ongoing cranchiid studies. In particular, we thank C. F. E. Roper and M. J. Sweeney for hospitality extended during a visit to the col- lections of the U.S. National Museum and for making available many non-cranchiid teu- thoids not available at Miami for the compara- tive studies. The illustrations were done by L. Sartucci and C. S. McSweeny; we acknowledge their contributions. This study was supported by National Sci- ence Foundation grants DEB-7713945 and DEB-8105193. It is a contribution from the Rosenstiel School of Marine and Atmospheric Science, University of Miami. LITERATURE CITED ARNOLD, J. M., 1962, Mating behavior and social structure in Loligo pealii. Biological Bulletin, 123: 53-57. ARNOLD, J. M., 1965, Observations on the mating behavior of the squid Sepioteuthis sepioidea. Bulletin of Marine Science, 15: 216-222. BIDDER, A. M., 1966, Feeding and digestion in cephalopods. In: WILBUR, K. M. & YONGE, C. M. (eds.), Physiology of Mollusca, Vol. 2, Academic Press, New York, p. 97-124. BIDDER, A. M., 1976, New names for old: the cephalopod “mid-gut gland.” Journal of Zoology, 180: 441-443. CAMIN, J. H. & SOKAL, R. R., 1965, A method for deducing branching sequences in phylogeny. Evolution, 19: 311-326. CHUN, C., 1910, Die Cephalopoden, |: Oegopsida. Wissenschaftliche Ergebnisse der Deutschen Tiefsee-Expedition, Valdivia, 18: 1-410, 61 pl. (English translation: Israel Program for Scientific Translations, Jerusalem, 1975.) CLARKE, M. R., DENTON, E. J. & GILPIN- BROWN, J. B., 1979, On the use of ammonium for buoyancy in squids. Journal of the Marine Biological Association of the United Kingdom, 59: 259-276. COLLESS, D. H., 1970, The phenogram as an esti- mate of phylogeny. Systematic Zoology, 19: 352-362. COLLESS, D. H., 1980, Congruence between mor- phometric and allozyme data for Menidia spe- cies: a reappraisal. Systematic Zoology, 29: 288-299. DILLY, P. N. & HERRING, P. J., 1974, The ocular light organ of Bathothauma lyromma (Mollusca: Cephalopoda). Journal of Zoology, 172:81-100. DILLY, P. N. & NIXON, M., 1976, Growth and de- velopment of Taonius megalops (Mollusca: Cephalopoda), and some phases of its life cycle. Journal of Zoology, 179: 19-83. DONOVAN, D. T., 1977, Evolution of the dibranchi- ate Cephalopoda. Symposia of the Zoological Society of London, 38: 15-48. CRANCHIID PHYLOGENY 425 DREW, G. A., 1911, Sexual activities of the squid Loligo pealii (Les.) 1. Copulation, egg-laying and fertilization. Journal of Morphology, 22: 327- 359. ESTABROOK, G. F., 1972, Cladistic methodology: a discussion of the theoretical basis for the in- duction of evolutionary history. Annual Review of Ecology and Systematics, 3: 427-456. ESTABROOK, G. F., STRAUCH, J. G. & FIALA, K. L., 1977, An application of compatibility analysis to the Blacklith's data on orthopteroid insects. Systematic Zoology, 26: 269-276. FARRIS, J. S., 1966, Estimation of conservatism in characters by constancy within biological popu- lations. Evolution, 20: 587-591. FARRIS, J. S., 1969, A successive approximations approach to character weighting. Systematic Zoology, 18: 374-385. FARRIS, J. S., 1970, Methods for computing Wag- ner Trees. Systematic Zoology, 19: 83-92. FARRIS, J. S., KLUGE, A. G. & ECKARD, M. J., 1970, A numerical approach to phylogenetic sys- tematics. Systematic Zoology, 19: 172-189. FELSENSTEIN, J., 1978, The number of evolution- ary trees. Systematic Zoology, 27: 27-33. HALL, D. N. F., 1956, Ink ejection by Cephalopoda. Nature, 177: 663. HAMABE, M. & SHIMIZU, T., 1957, The copulation behavior of Yariika, Loligo bleekeri K. Annual Report of the Japan Sea Regional Fisheries Research Laboratory, 3: 131-136 (In Japanese with English summary). HENNIG, W., 1966, Phylogenetic systematics. Uni- versity of Illinois Press, Urbana, 263 p. HERRING, P. J., 1977, Luminescence in cephalo- pods and fish. Symposia of the Zoological Soc- iety of London, 38: 127-159. HOYLE, W. E., 1907, Opening address, the British Association, Section D, Zoology. Nature, 76: 452-457. KLUGE, A. G. & FARRIS, J. S., 1969, Quantitative phyletics and the evolution of anurans. System- atic Zoology, 18: 1-32. MCGOWAN, J. A., 1954, Observations on the sex- ual behavior and spawning of the squid, Loligo opalescens, at La Jolla, California. California Fish and Game, 40: 47-54. MCSWEENY, E. S., 1978, Systematics and mor- phology of the Antarctic cranchiid squid Galiteu- this glacialis (Chun). Biology of the Antarctic Seas VII, Antarctic Research Series, 27: 1-39. MEACHAM, C. A., 1980, Phylogeny of the Berberi- daceae with an evaluation of classifications. Systematic Botany, 5: 149-172. MICKEVICH, M. F. & FARRIS, J. S., 1981, The implications of congruence in Menidia. System- atic Zoology, 30: 351-370. NAEF, A., 1921/1923, Die Cephalopoden. Fauna und Flora Neapel, Monograph 35, 1(part 1: num- bers 1, 2): 1-863. (English translation: Israel Program for Scientific Translations, Jerusalem, 1972). NELSON, G. 8 PLATNICK, N. 1., 1980, Multiple branching in cladograms: two interpretations. Systematic Zoology, 29: 86-90. NESIS, K. N., 1974, A revision of the squid genera Corynomma, Megalocranchia, Sandalops, and Liguriella (Oegopsida, Cranchiidae). Transac- tions of the Institute of Oceanology, Academy of Sciences, USSR, 96: 6-22. (English translation: Translation Bureau, Department of the Secretary of State of Canada, Translation Series No. 3522.) NICOL, J. A. C., 1964, Special effectors: luminous organs, chromatophores, pigments, and poison glands. In: WILBUR, K. M. & YONGE, C. M. (eds.), Physiology of Mollusca, Vol. 1, Academic Press, New York, p. 353-381. PFEFFER, G., 1912, Die Cephalopoden der Plank- ton-Expedition. Ergebnisse der Plankton-Expe- dition der Humboldt-Stiftung, 2: i-xxi, 1-815, 48 pl. ROBSON, G. C., 1926, On the hectocotylus of the Cephalopoda—a reconsideration. Proceedings of the Malacological Society of London, 17: 117-122. ROPER, С. FE YOUNG, В: Е. VOSS Gaile 1969, An illustrated key to the families of the order Teuthoidea (Cephalopoda). Smithsonian Contributions to Zoology, 13: 1-32. SCHIPP, R. & V. BOLETZKY, S., 1975, Morphol- ogy and function of the excretory organs in di- branchiate cephalopods. Fortschritte der Zoologie, 23: 89-111. SCHIPP, R. & V. BOLETZKY, S., 1976, The pan- creatic appendages of dibranchiate cephalo- pods. 1. The fine structure of the “pancreas” in Sepioidea. Zoomorphologie, 86: 81-98. STEENSTRUP, J., 1857, Hectocotylus-formation in Argonauta and Tremoctopus explained by ob- servations on similar formations in the Cephalo- poda in general. Annals and Magazine of Natu- ral History, ser. 2, 20: 81-114, 2 pl. STEVENS, P. F., 1980, Evolutionary polarity of character states. Annual Review of Ecology and Systematics, 11: 333-358. THOMAS, R. F., 1977, Systematics, distribution, and biology of cephalopods of the genus Tremoctopus (Octopoda: Tremoctopodidae). Bulletin of Marine Science, 27: 353-392. VOSS, G. L., 1977a, Present status and new trends in cephalopod systematics. Symposia of the Zoological Society of London, 38: 49-60. VOSS, G. L., 1977b, Classification of recent cephalopods. Symposia of the Zoological So- ciety of London, 38: 575-579. VOSS, N. A., 1974, Studies on the cephalopod family Cranchiidae. A redescription of Egea inermis, Joubin, 1933. Bulletin of Marine Sci- ence, 24: 939-956. VOSS, N. A., 1980, A generic revision of the Cranchiidae (Cephalopoda; Oegopsida). Bulletin of Marine Science, 30: 365-412. WATROUS, L. E. & WHEELER, Q. D., 1981, The out-group comparison method of character analysis. Systematic Zoology, 30: 1-11. 426 N. A. VOSS AND R. S. VOSS WEIHS, D. & MOSER, H. G., 1981, Stalked eyes as an adaption towards more efficient foraging in marine fish larvae. Bulletin of Marine Science, 31: 31-36. YOUNG, J. Z., 1970, The stalked eyes of Batho- thauma (Mollusca, Cephalopoda). Journal of Zoology, 162: 437-447. YOUNG, R. E., 1975a, Leachia pacifica (Cephalo- poda, Teuthoidea): spawning habitat and func- tion of the brachial photophores. Pacific Sci- ence, 29: 19-25. YOUNG, R. E., 1975b, Transitory eye shapes and the vertical distribution of two midwater squids. Pacific Science, 29: 243-255. YOUNG, R. E., 1977, Ventral bioluminescent countershading in midwater cephalopods. Sym- posia of the Zoological Society of London, 38: 161-190. YOUNG, R. E. & ROPER, C. F. E., 1968, The Batoteuthidae, a new family of squid (Cephalo- poda; Oegopsida) from Antarctic waters. Antarc- tic Research Series, 11: 185-202. YOUNG, R. E. & ROPER, C. F. E., 1969a, A mono- graph of the Cephalopoda of the North Atlantic: The family Cycloteuthidae. Smithsonian Contri- butions to Zoology, 5: 1-24. YOUNG, R. E. & ROPER, C. F. E., 1969b, A mono- graph of the Cephalopoda of the North Atlantic: The family Joubiniteuthidae. Smithsonian Contri- butions to Zoology, 15: 1-10. ZUEV, G. V., 1967, Functional importance of the funnel valve in cephalopods. Zoological Journal, Academy of Sciences, USSR, 46: 1099-1101. (English translation: Translation Bureau Depart- ment of the Secretary of State of Canada, Trans- lation Series No. 968.) MALACOLOGIA, 1983, 23(2): 427-428 LEMER TO) THE EDIMORS ON SOME APLACOPHORAN HOMOLOGIES AND DIETS (1) The assertion by Salvini-Plawen (1981) that the subradular membrane of mollusks is a direct continuation of pharyngeal cuticle is not supported by Markel (1958, fig. 35, p. 281), Runham (1963), or Peters (1978, fig. 3, p. 287), who demonstrated that the subradu- lar membrane is distinct from, and at its ante- rior end overrides, the pharyngeal cuticle. The subradular membrane is secreted by the distal inferior epithelium of the radular sac and is firmly attached to both the inferior epitheli- um by apical processes and the radula mem- brane (Kerth, 1976, fig. 3, p. 275). The sub- radular membrane develops ontogenetically later than the radula membrane (Kerth et al., 1981). The radula membrane, which bears the radula teeth, is secreted by the most ante- rior odontoblasts at the posterior end of the sac (Kerth & Krause, 1969, fig. 11, p. 66; Wiesel & Peters, 1978, fig. 8, p. 84). According to Salvini-Plawen, the “basal cuticle” upon which the teeth of neomenio- morphs (= Solenogastres) usually are born “is... not independently formed at the... blind end of the radula sheath,” and on p. 378 he states that the “radula itself” is formed “as usual in a separate sheath by odontoblasts.” Thus we are left to conclude that radula teeth, but not a radula membrane, are formed by odontoblasts; these teeth somehow become attached to a membrane presumably continu- ous with pharyngeal cuticle. In my own experience with isolated entire radulae those of the Neomeniomorpha as well as the Chaetodermomorpha consist of teeth on a discrete membrane that is not continu- ous with buccal cuticle and looks just like any other radula ribbon. The concept of a “basal cuticle” seems to rest on no evidence. (2) Gymnomenia is one of the primitive genera of neomeniomorphs but the lack of ventral salivary glands is a derived state. As the absence or presence of foregut glands does not define the primitive order Pholi- doskepia, | apologize for attributing to Salvini- Plawen the idea that either state is primitive. (3) My “non-homology” of the oral shield and the molluscan foot seems altogether cor- rect to me. The oral shield is clearly derived from gut epithelium (Scheltema, 1981, fig. 2, p. 364); the continuous cuticle of the gut and oral shield is of course secondary to a non- cuticularized state. Since the oral shield is not a vestigial foot, then there is no cladistic basis for splitting the Aplacophora into two classes (Salvini-Plawen, 1972). This split rests (a) on the assumption that the Chaetodermomorpha (Caudofoveata) and Neomeniomorpha (Sol- enogastres sensu Salvini-Plawen) evolved in- to a worm shape in two separate evolutionary events from an ancestor whose mouth opened through a gliding foot-sole and (b) on the homology between the molluscan creep- ing foot and the oral shield of the Chaetoder- momorpha. The latter homology originated with Hoffman (1949), who compared the evo- lutionarily advanced Chaetoderma nitidulum with Dorymenia hoffmani (= Proneomenia antarctica). He wrote (p. 382, herein translat- ed): “The oral shield integument consists . . . of elements which we are acquainted with in the integument of the ventral furrow: [a] cells which show signs that they were ciliary cells and which have gone through the same change as the most common cells in the inner surface of the lateral fold (cuticularization of cilia) [i.e., outside wall of the furrow]; [b] true sensory cells, and finally [c] gland cells which have the same form (deeply sunk, pyriform), the same arrangement to one another (joined in lobes), and the same secretion (strongly stained with Ehrlich’s haemotoxylin) as the gland cells of the ventral furrow.” (cf. Hoff- man, 1949, figs. 16, 17, 18, 21, 30, 31, 34.) Of this homology Salvini-Plawen wrote (1972: 295, herein translated): “The extreme- ly detailed likeness of the foot-shield [oral shield] epithelium and ventral-furrow epitheli- um (Solenogastres) including the glands in themselves leaves no doubt as to the mutual homology.” He then went on to illustrate and describe carefully for several species the di- rect connections between the cerebral gan- glion and precerebral ganglia which give rise to the innervation of the oral shield. Hoffman had interpreted the precerebral ganglia as arising from fused cerebrolateral-cerebro- ventral connectives. According to Salvini- Plawen, in both Chaetodermomorpha and Neomeniomorpha the area around the mouth is innervated from the cerebral ganglion (1972, fig. 45). In the English summary (1972, (427) 428 p. 376, paragraph 16) he wrote that the “foot- shield” as shown by histologic detail “repre- sents a portion of a previously overall ventral gliding surface—merely distinguished from the recent molluscan foot by innervation. Con- sequently, the foot-shield represents the cerebrally innervated fragment of the overall ventral gliding-sole of the Archimollusca.” From my own observations (1978, 1981), | have concluded that: (a) The oral shield is a specialized (derived) structure formed from foregut epithelium; (b) The gland cells of the oral shield are diffuse in the primitive genus Scutopus and thus do not agree with Hoff- man's homology. | do not follow the reasoning by which Salvini-Plawen on the one hand re- tains Hoffman's homology, which is based on the “deeply sunk” mucus glands “joined in lobes” in the more highly evolved genus Chaetoderma, while on the other hand he cor- rects the homology to include scattered, dif- fuse cells (i.e., of Scutopus). After all, diffuse mucous cells are ubiquitous among mollusks as an aid to all sorts of ciliary and muscular movement and are therefore not reliable markers of homologous structures. The only sure evidence for a vestigial foot would be innervation from the ventral ganglia or from a branch from the lateroventral con- nectives, which Hoffman endeavored to show. The cerebral innervation of the oral shield detailed by Salvini-Plawen is precisely the evidence that determines that the oral shield is not a vestige of the archimolluscan foot. (4) Finally, | do not believe that my paper “corroborates” Salvini-Plawen's wherein it is concerned with the diets of the Chaetodermo- morpha. Organic debris, pieces of sponge spicules, bits of radiolaria, diatoms, and occa- sional inorganic sediment particles usually occur in the guts of chaetoderms. Salvini- Plawen himself states (p. 375) that because these animals are burrowers, “findings of other particles and/or stated organisms, therefore, may be an accidental by-product” of selective feeding. It is regrettable that Salvini-Plawen's Table 1 (pp. 376-77) does not distinguish between likely diets (recogniz- able crustacean eggs, crustacean parts, Foraminifera) and organic debris that may have been accidentally ingested. In my experience there is no indication from LETTERS TO THE EDITORS gut contents that any species of Chaetoder- momorpha is a deposit-feeder, i.e. “seize[s] sediment particles without specific selection” (p 375): REFERENCES CITED HOFFMAN, S., 1949, Studien úber das Integument der Solenogastren ... Zoologiska Bidrag fran Uppsala, 27: 293-427. KERTH, K., 1976, Licht- und electronenoptische Befunde zum Radulatransport bei der Lungen- schnecke Limax flavus L. (Gastropoda, Stylom- matophora). Zoomorphologie, 83: 271-281. KERTH, K. 8 KRAUSE, G., 1969, Untersuchungen mittels Rôntgenbestrahlung über den Radula- Ersatz der Nacktschnecke Limax flavus L. Wil- helm Roux’ Archiv, 164: 48-82. KERTH, K., REDER, |. & ZIMMERMANN, R., 1981, Der Radulaabbau beim Embryo der Sumpf- deckelschnecke Viviparus fasciatus Müll. (Gastropoda, Prosobranchia). Zoologische Jahrbücher, Abteilung für Anatomie, 106: 104- Le MARKEL, K., 1958, Bau und Funktion der Pulmo- naten-Radula. Zeitschrift für wissenschaftliche Zoologie, 160: 213-289. PETERS, W., 1978, Degradation of the radula in the snails Biomphalaria glabrata Say and Limnaea stagnalis L. (Gastropoda, Pulmonata). Cell and Tissue Research, 193: 283-295. RUNHAM, N., 1963, A study of the replacement mechanism of the pulmonate radula. Quarterly Journal of Microscopical Science, 104: 271- 277. SALVINI-PLAWEN, L. V., 1972, Zur Morphologie und Phylogenie der Mollusken. Zeitschrift für wissenschaftliche Zoologie, 184: 205-394. SALVINI-PLAWEN, L. V., 1981, The molluscan di- gestive system in evolution. Malacología, 21: 371-401. SCHELTEMA, A. H., 1978, Position of the class Aplacophora in the phylum Mollusca. Malaco- logia, 17: 99-109. SCHELTEMA, A.H., 1981, Comparative morphol- ogy of the radulae and alimentary tracts in the Aplacophora. Malacologia, 20: 361-383. WIESEL, R. & PETERS, W., 1978, Licht- und elecktronen-mikroskopische Untersuchungen am Radulakomplex und zur Radulabildung von Biomphalaria glabrata Say (= Australorbis gl.) (Gastropoda, Basommatophora). Zoomorph- ologie, 89: 73-92. Amelie H. Scheltema Woods Hole Oceanographic Institution Woods Hole, MA 02543, U.S.A. INDEX TO TAXA IN VOLUME 23 An asterisk (*) denotes a new taxon Abralia, 122, 135-163, 205 atlantica, Leachia, 407 Abraliopsis, 122, 155, 159 aurantiaca, Berthella, 221, 235, 247, 248 Acmaea, 37 aurantiaca, Bouvieria, 235, 243, 245, 247 Acoela, 225 aurantiacus, Bouvieria, 243 adamsi, Seila, 37 aurantiacus, Pleurobranchus, 235, 243 adelphus, Bulimulus, 219 australis, Robsonella, 123 adelphus, Naesiotus, 219 Australorbis, 328 affinis, Partula, 24, 26, 27, 31, 32 avara, Anachis, 37 affinis, Terebra, 3-7, 10 Axiothella, 291, 293-295 africana, Berthellina, 228 baileyi, Neoagardhiella, 282 Agaronia, 12 Bakevelliidae, 375 agassizi, Pleurobanchaea, 254 Bankia, 319 akamatus, Bulimulus, 216, 219 Barbatia, 375, 377, 393 akamatus, Naesiotus, 216, 219 Basommatophora, 333 Alasmidonta, 372 Bathothauma, 122, 157, 397-426 alderi, Natica, 43 Bathyarca, 377 alethorhytidus, Bulimulus, 219 *Bathyberthella, 221, 222, 226, 227, *248-253, 263 alethorhytidus, Naesiotus, 219 Bathypolypus, 122 algoensis, Pleurobranchaea, 254 Bathyteuthidae, 405 Alloposidae, 408 Bathyteuthis, 122 Alloteuthis,121 Batoteuthidae, 404 alternatum, Bittium, 37-46 Bentharca, 377 amabilis, Partula, 26, 28 Bentheledone, 122 amarillius, Oscanius, 228 Benthoctopus, 122 amarillius, Pleurobranchus, 228 Berryteuthis, 122 ambiguus, Velesunio, 362 Berthella, 221-270 ambiseta, Mediomastus, 293-295 Berthellina, 221-224, 226-236, 239, 247, 253, 256, americana, Berthella, 227 263 americanus, Pleurobranchus, 228, 236 Berthellinae, 226, 227 ampulla, Bulla, 11 Berthellinops, 236 Ampullariidae, 13-21 bicallosus, Nassarius, 6, 11 Anachis, 37 bicanaliculata, Platynereis, 291, 293, 294 Anadara, 375, 377, 378, 393 bifasciatum, Clypeomorus, 3 Anadarinae, 377, 393 bimaculoides, Octopus, 196 Anaspidea, 222 Biomphalaria, 351 anatina, Physa, 351-359 Bithynia, 333-349 Anhfeltia, 282 Bittium, 37-46 anilis, Terebra, 6, 11 Bivalvia, 47-54, 361-396 Anodontinae, 362 bleekeri, Loligo, 117 Anomalocardia, 313 blomberghi, Bulimulus, 219 Anodonta, 361-374 blomberghi, Naesiotus, 219 antarctica, Proneomenia, 427 borealis, Hemipodus, 293 antiquata, Anadara, 375, 377, 378 borneensis, Berthella, 228 Aplacophora, 427-428 Bouvieria, 228, 235, 236, 243 Aplysiomorpha, 225 Brachiopoda, 319 Apostomea, 126, 127 Brachioteuthidae, 404 Arca, 377, 393 brasiliana, Anomalacardia, 313 Arcacea, 375-396 brevicephaloides, Dicyemennea, 129 Architeuthidae, 409 brevis, Armandia, 293 Arcidae, 375, 377, 383, 393 brevis, Lolliguncula, 90, 93 Arcinae, 377, 393 Buccinidae, 8 Arcticacea, 381 Bulimulidae, 209-219 arenatus, Conus, 6, 11 Bulimulus, 209-219 Argonauta, 122 Bulinus, 328, 351 Argonautidae, 408 Bulla, 11, 228, 235 Armandia, 293 Bullia, 318 aspera, Rhinoclavis 3-7, 10, 313-315, 317, 319 Bullomorpha, 225 “Aspergillus,” 122 burchi, Samoana, 31 aterrima, Pilsbryspira, 6, 12 Bursidae, 8 athearni, Goniobasis, 81, 83, 84 caecoides, Nephthys, 293, 294 (429) 430 MALACOLOGIA Caecum, 37 Calappa, 1-12, 319 Calcichordata, 319 caliendrum, Polycirrus, 295 californica, Opuntiella, 289 californica, Pleurobranchaea, 254, 258, 261, 263 callospira, Nassarius, 6, 11 Campeloma, 356 canaliculata, Thais, 63-73 canarium, Strombus, 10 Cancellaria, 6, 12 cancellata, Ophiodermella, 281-312 Cancer, 281, 287, 298, 299, 309 candida, Pholadomya, 382 capensis, Pleurobranchaea, 254 capitata, Capitella, 293-295 Capitella, 293-295 Capitellidae, 291 Carcinus, 7, 8 Cassidae, 8 cataracta, Anodonta, 361, 362, 364-374 catenulatus, Modulus, 6, 11 Caudofoveata, 427 cavagnaroi, Bulimulus, 211, 216, 219 cavagnaroi, Naesiotus, 211, 216, 219 Cavoliniidae, 225 Cepaea, 23 Cephalaspidea, 222, 225 Cephalopoda, 87, 89-134, 165-175, 177-201, 203- 208, 397-426 Cerithiacea, 313 Cerithiidae, 8, 313, 319 Cerithiopsis, 37-46 Cerithium, 3, 6, 11, 313, 318 Cestodea, 123 Chaetoderma, 417, 428 Chaetodermomorpha, 427, 428 Chaetozone, 295 Chamidae, 393 chemnitzi, Natica, 6, 11 Chiroteuthidae, 398, 404, 405, 408, 412, 413 Chiroteuthis, 122 Chromidina, 121-134 Chromidinidae, 126, 128-132 Ciliata, 125-127 Cirrata, 174, 408 cirrata, Laonice, 295 cirrata, Prionospio, 295 Cirratus, 295 cirratus, Cirratus, 295 cirrhosa, Eledone, 126 citrina, Berthella, 253 citrina, Berthellina, 221-224, 226, 228-236, 239, 247, 248, 263 clara, Partula, 33 Clathromangelia, 300 Claudiconcha, 393 clavulus, Strombina, 3, 6, 11 Cleanthus, 235, 236 Cleidothaeridae, 393 clenchi, Goniobasis, 81, 84, 85 Clionidae, 225 Clionopsidae, 225 Clostridium, 179 Clypeomorus, 3 Coleoida, 121 Columbellidae, 8 columna, Cerithium, 3 concamerata, Cucullaea, 375, 377 concinna, Anhfeltia, 282 Conidae, 281 Conocyema, 123 Conocyemidae, 123 contortus, Bulinus, 328 Conus, 6, 7, 10-12, 281 convexa, Calappa, 3 convexa, Crepidula, 43 coralium, Cerithium, 6, 11 Coretus, 333 corianus, Lamellidens, 365 corneus, Coretus, 333 corneus, Planorbarius, 333, 338, 345, 346 coronata, Chromidina, 126 coronatus, Conus, 6, 11 corpulenta, Anodonta, 362 costata, Lasmigona, 372 couperiana, Anodonta, 361, 362, 364-367, 369, 371 cracilenta, Terebra, 12 Cranchia, 122, 397-426 Cranchiidae, 397-426 Cranchiinae, 397-426 crassa, Partula, 27 Crassostrea 271-279 crenulata, Terebra, 318 Crepidula, 37-46, 277 crispata, Oulastrea, 376, 377 cristata, Valvata, 333, 335-337, 340 Crustacea, 1, 126, 319, 428 Ctenopteryx, 122 Cucullaea, 375, 377 Cucurbitula, 376 cuvieri, Berthellina, 228 cuvieri, Pleurobranchus, 228 Cyamiacea, 381 Cycloteuthidae, 398, 404, 408, 412, 413 cygnea, Anodonta, 361, 362, 364-367, 369, 371- 373 cymatias, Bulimulus, 219 cymatias, Naesiotus, 219 Cymatiidae, 8 cymbium, Cucurbitula, 376 Cymbuliidae, 225 Cypraeacea, 8 Cyrenoida, 47-54 Cyrenoididae, 47-54 Cystophora, 268 danae, Leachia, 410, 414 Decapoda, 123, 128, 129, 165, 173 decisa, Campeloma, 356 decollata, Rumina, 353, 356-358 Dendraster, 282 dentifer, Donax, 313 dentifer, Nassarius, 11 Desmarestia, 282 Dicyemidae, 123, 124, 126, 128-132 Diastoma, 40 dickinsoni, Goniobasis, 81, 85 Dicyema, 123, 131 Dicyemennea, 123, 125, 129 Dicyemodeca, 123 Didymozoa, 123 Digenea, 123 dimidiata, Terebra, 313-316, 318, 319 Diodontidae, 3 dislocata, Terebra, 11 distortus, Nassarius, 6, 11 Divaricella, 313 dolabrata, Pyramidella, 6, 10 Donax, 313 Doridacea, 222, 225 dorsalis, Pleurobranchaea, 255 dorsatus, Tapes, 375, 377 Dorymenia, 427 Doryssa, 81 Doryteuthis, 90 Dosidicus, 122 Dreissenacea, 395 Drilonereis, 295 Drosophila, 81, 82, 85 Dunaliella, 262, 288 duncanus, Bulimulus, 219 duncanus, Naesiotus, 219 duryi, Helisoma, 346 Echinodermata, 319 edulis, Loligo, 117 edulis, Mytilus, 78 Egea, 397-426 elata, Terebra, 6, 12 Eledone, 89, 122, 129 elegans, Chromidina, 126 elegans, Oenopota, 300 elegans, Pomatias, 337 elegans, Pygospio, 291, 293 elegans, Sepia, 126 ellipticus, Lanistes,14 Elliptio, 361, 371 emarginata, Thais, 63-73 emersoni, Cerithiopsis, 37-46 engeli, Berthellina, 228, 233, 235 Enoploteuthidae, 409, 410 Enoploteuthis, 122, 157 Enteromorpha, 282 Entodesma, 36, 37 eos, Bulimulus, 214, 216, 219 eos, Naesiotus, 214, 216, 219 Eteone, 293, 294 eudiscus, Helisoma, 346 Euphausiacea, 126, 127 Euprymna, 121, 177-192 Euselenops, 226, 227 exasperatum, Vexillum, 6, 11 excurvata, Oenopota, 300 excentricus, Dendraster, 282 falcata, Drilonereis, 295 fasciata, Rhinoclavis, 3-7, 10 Ferrissia, 346 ferruginea, Nephthys, 294, 295 fidicuia, Oenopota, 300 flavus, Limax, 329 INDEX TO VOL. 23 431 floridana, Cyrenoida, 47-54 floridensis, Goniobasis 81-85 Foettingeriidae, 126, 128 Foraminifera, 428 forbesi, Loligo, 91, 117 fornicata, Crepidula, 37, 43, 44, 277 fragilis, Anodonta cataracta, 361, 362, 364-368, 370-374 fusiformis, Owenia, 281-312 Galiteuthis, 122, 397-426 Gastrochaenidae, 376 Gastropoda, 1-12, 37-46, 55-73, 209-270, 281- 349, 375, 393, 395 gela, Pleurobranchaea, 254 gemini, Pleurobranchaea, 254 Gemmula, 6, 11 gibberulus, Strombus, 1, 3-7, 10, 11 gibbosa, Anodonta, 361, 362, 364-372, 374 Gigartina, 289 gigas, Crassostrea, 271-279 gilderoyi, Bulimulus, 211 gilderoyi, Naesiotus, 211 glabratus, Australorbis, 328 glacialis, Galiteuthis, 402, 404, 407, 411, 414 globosus, Nassarius, 6, 11 glutaeus, Malacoceros, 291, 293, 294 Glycinde, 293, 294 Glycymerididae, 393 Glycymeris, 267 Gonatidae, 410 Gonatopsis, 122 Gonatus, 122 Goniobasis, 19, 81-86 Gonodactylidae, 1 gouldi, Terebra, 314 gracilis, Cancer, 281, 287, 298, 299, 309 graeffei, Gemmula, 6, 11 grandis, Anodonta, 361, 362, 364, 366, 367, 369, Silt 373 Graneledone, 122 granifera, Tarebia, 83 granulosa, Pleurobranchaea, 221, 255, 262 Grapsus, 319 grapsus, Grapsus, 319 greeni, Cerithiopsis, 37 Grimalditeuthidae, 398, 401, 404, 408, 412, 413 Grimalditeuthis, 401, 402 Grimpoteuthis, 122 gurneyi, Ptilosarcus, 282 Gymnodinioides, 130 Gymnomenia, 427 Gymnosomata, 225 Gymnotoplax, 221, 228, 235, 236 gyrina, Physa, 351-359 hallenbeckii, Anodonta, 362 Halopsychidae, 225 hamva, Pleurobranchaea, 254 haraldi, Pleurehdera, 227, 253 hedgpethi, Pleurobranchaea, 254 Hedophyllum, 289 Helicocranchia, 122, 397-426 Helisoma, 346 Helix, 33, 333 432 MALACOLOGIA Helminthes, 123 Hemipodus, 293 Hemisinus, 81 henryana, Anodonta, 362 hepatica, Calappa, 14 heros, Lunatia, 37-46 Heteroteuthis, 121, 173 hirsutus, Bulimulus, 219 hirsutus, Naesiotus, 219 Histioteuthis, 122 hoffmani, Dorymenia, 427 hyalina, Partula, 33 Hyalophysa, 130 Hyriidae, 362 illecebrosus, Illex, 123 Illex, 122, 123 ilisima, Berthellina, 228 llyanassa, 37 imbecilis, Anodonta, 361, 362, 364, 367, 369, 371, 372 Imbricaria, 6, 11 implicata, Anodonta, 361, 362, 364-374 Incirrata, 174, 408, 409 incisa, Ophiodermella, 282 inermis, Egea, 402, 414 inermis, Ophiodermella, 281-312 inhacae, Pleurobranchus, 224 interfossa, Clathromangelia, 300 intermedia, Natica, 43 Isochrysis, 38, 288 Japetella, 122 jacobi, Bulimulus, 219 jacobi, Naesiotus, 219 Japonica, Claudiconcha, 393 Japonica, Pleurobranchaea, 254 jayana, Cancellaria, 6, 12 Joubiniteuthidae, 398, 404, 405, 408, 412, 413 Juga, 81 kaniae, Berthella, 242 kennerlyi, Anodonata, 372 kiyonoi, Trisidos, 377, 393 Koonsia, 254 Kurtziella, 295, 300 labiatus, Strombus, 6, 10, 11 Lacuna, 37 Lamellidens, 365 lamellosa, Nucella, 299 lamellosa, Thais, 63-73 Lanistes, 13-21 Laonice, 295 lapillus, Nucella, 69 Lasmigona, 372 Laurencia, 268 Leachia, 122, 397-426 Lepidoteuthidae, 404 lessoniana, Sepioteuthis, 166, 170 levidensis, Oenopota, 295, 300 lewisii, Polinices, 299 lignaria, Partula, 27 ligulata, Desmarestia, 282 Liguriella, 397-426 Limacinidae, 225 Limax, 329 Limopsacea, 381, 386, 393 Limopsidae, 393 Liocranchia, 122, 397-426 Litharca, 377 Lithasiopsis, 81 Lithophaga, 377 littorea, Littorina, 37 Littorina, 8, 37 livescens, Goniobasis, 19 Loligo, 89-208, 409 Loliolopsis, 121 Lolliguncula, 90, 122, 193 longa, Eteone, 293, 294 Lophocercidae, 225 lucifer, Phasmatopsis, 406 luhuanus, Strombus, 10 Lumbrineris, 294 lunata, Mitrella, 37 Lunatia, 37-46 luniceps, Euselenops, 227 luridus, Nassarius, 6, 11 luteostoma, Nassarius, 3, 6, 12 lycodus, Bulimulus, 219 lycodus, Naesiotus, 219 Lycoteuthidae, 409 Lymnaea, 55-62, 321-349 Lymnaeoidea, 333 lyromma, Bathothauma, 157, 402, 404, 407 macrosoma, Rossia, 166, 171-174 maculata, Pleurobranchaea, 221, 222, 224, 229, 254-258, 260-263 maculatum, Pleurobranchidium, 255 maenas, Carcinus, 8 malaccana, Lithophaga, 377 Malacocerus, 291, 293, 294 maorum, Octopus, 123 maorum, Plagioporus, 122 marina, Zostera, 282 Mastigoteuthidae, 398, 404, 405, 408, 412, 413 Mastigoteuthis, 122 meckeli, Pleurobranchaea, 254 meckelii, Pleurobranchaea, 254, 261-263 meckelii, Pleurobranchidium, 254 mediata, Berthella, 243 mediatas, Berthella, 221-270 Mediomastus, 293-295 Megalocranchia, 122, 397-426 megalops, Teuthowenia, 403, 404, 411 Melanoides, 81, 83 membranaceus, Pleurobranchus, 224 Mercenaria, 75-79 mercenaria, Mercenaria, 75-79 Mesogastropoda, 313, 333 Mesonychoteuthis, 397-426 Mesozoa, 123, 128, 131, 132 Microcyema, 123 microlampas, Pterygioteuthis, 158 minor, Berthella, 228 minor, Berthellina, 229 minuta, Pholoe, 295 Mitra, 313-320 mitralis, Otopleura, 6, 10 Mitrella, 37 INDEX TO VOL. 23 433 Mitridae, 8, 313, 319 Modulus, 6, 11 morosus, Pleurobranchillus, 254 Moroteuthis, 122 multifilis, Tharyx, 294, 295 Muricidae, 8 muriculatus, Conus, 6, 11 mutabilis, Strombus, 3 Myopsida, 89-119, 398, 409 Myriochele, 281, 286, 295, 304, 305, 308 Mytilacea, 395 Mytilidae, 383 Mytilus, 78 Naesiotus, 209-219 Nassariidae, 8 Nassarius, 3, 5-7, 11, 12, 37 Natica, 4, 6, 11, 43 Nautiloidea, 121, 414 Nautilus, 121 Nematoda, 123 Nematoscelis, 127 nemoralis, Cepaea, 23 Neoagardhiella, 282 Neogaimardia, 381 Neomeniomorpha, 427 Nephthys, 293-295 nesioticus, Bulimulus, 219 nesioticus, Naesiotus, 219 nigrocincta, Triphora, 37-46 nitida, Natica, 43 nitidulum, Chaetoderma, 427 nodicincta, Otopleura, 10 nodulosum, Certhium, 318 Notaspidea, 221-270 notata, Mercenaria, 75-79 Northia, 11 Notobranchaeidae, 225 Notomastus, 293, 294 novaezelandiae, Pleurobranchaea, 255, 262 novaezealandiae, Pleurobranchaea, 221, 236, 254, 255, 262 Nucella, 69, 299 Nudibranchia, 225 nyassanus, Lanistes, 13-21 obesa, Koonsia, 254 oblongata, Berthellina, 235 obsoletus, Nassarius, 37 oceanica, Phasmatopsis, 406 ocellata, Berthella, 242 ochsneri, Bulimulus, 219 ochsneri, Naesiotus, 219 Octopoda, 122, 405, 409, 412, 414 Octopodoteuthidae, 413 Octopoteuthis, 122 Octopus, 87-201, 409 oculata, Myriochele, 281, 286, 295, 304, 305, 308 Ocythoe, 122 Ocythoidae, 408 Oegopsida, 397-426 Oenopota, 295, 300 officinalis, Sepia, 122, 166, 169, 170 Olivella, 6, 12 olla, Bulimulus, 219 olla, Naesiotus, 219 Ommastrephes, 122, 178 Ommastrephidae, 401, 402, 405, 409 Onychoteuthidae, 410 Onykia, 122 opalescens, Loligo, 92, 115, 117, 123, 178 Opalinopsidae, 126 Opalinopsis, 122, 125 Ophiodermella, 281-312 Opisthobranchia, 221-270 Opisthoteuthis, 122 Opuntiella, 289 orbigniana, Sepia, 126 ornata, Berthella, 221, 222, 227, 235-245, 263, 269 ornata, Berthellina, 230, 231, 239 ornata, Bouvieria, 236 ornatus, Pleurobranchus, 236 Oscanius, 227 Ostrea, 276, 278 otaheitana, Partula, 23-35 Otopleura, 6, 10 oualaniensis, Symplectoteuthis, 157 Oulastrea, 376, 377 ovalis, Pleurobranchus, 227, 242 ovum, Lanistes, 14 Owenia, 281-312 Oweniidae, 310 Oxynoeidae, 225 Oxyperas, 267 Pachychilus, 81 pacifica, Rossia, 129 pacificus, Ommastrephes, 178 pagodus, Nassarius, 6, 12 Palinuridae, 2 pallidus, Bulimulus, 213, 219 pallidus, Naesiotus, 213, 219 papillata, Gigartina, 289 Pareledone, 122 Partula, 23-35 Partulidae, 23-35 patricius, Conus, 12 pavo, Taonius, 411 pealei, Loligo, 90, 92, 115, 116, 203, 204, 206, 208 Pecten, 267 peggyae, Anodonta, 361, 362, 364-369, 371 pellucida, Berthella, 242 peregra, Radix, 333, 342-346 perfringens, Clostridium, 179 peroni, Pleurobranchus, 224 perversa, Triphora, 40 Phaeodactylum, 38 Phasmatopsis, 122, 406 Philobrya, 381 Pholadomya, 382 Pholidoskepia, 427 Pholoe, 295 Phyllochaetopterus, 306 Physa, 327, 351-359 Physalia, 203, 205 Physidae, 351-359 picta, Glycinde, 293, 294 Pila 14, 20 Pilidae, 13 434 MALACOLOGIA Pilsbryspira, 6, 12 Plagioporus, 122 plana, Crepidula, 37-46 Planorbarius, 333-349 Planorbidae, 346 Planorboidea, 333 planospira, Bulimulus, 213, 219 planospira, Naesiotus, 213, 219 Platynereis, 291, 293, 294 plei, Loligo, 89-119, 193, 196, 200, 205, 208 Pleodicyema, 123 Pleurehdera, 226, 227, 253 Pleurobranchacea, 225, 226 Pleurobranchaea, 221-270 Pleurobranchaeinae, 222, 226, 253 Pleurobranchella, 221, 226, 227, 253 Pleurobranchidae, 221-270 Pleurobranchidium, 254, 255 Pleurobranchillus, 254 Pleurobranchinae, 222, 226, 227, 253 Pleurobranchomorpha, 225 Pleurobranchopsis, 221 Pleurobranchus, 224, 226, 227, 233, 236, 237, 242, 243, 253 Pleuroceridae, 81 plumbea, Kurtziella, 295, 300 plumula, Berthella, 228 plumula, Bulla, 228, 235 Pneumodermatidae, 225 podophtalma, Liguriella, 407 poliana, Natica, 43 Polinices, 4-6, 10, 11, 299 Polycirrus, 295 Polydora, 294, 297, 376, 377 Polyspira, 132 pomatia, Helix, 33 Pomatias, 14, 333, 337 pomilia, Physa, 327 porosa, Berthella, 228, 235 Potamididae, 8 Potamogeton, 15-17 Potodoma, 81 primolecta, Dunaliella, 262 Prionospio, 295 pristis, Northia, 11 procerus, Lanistes, 14 productus, Cancer, 281, 287, 298, 299, 309 Promachoteuthidae, 398, 404, 408, 412, 413 Proneomenia, 427 Prosobranchia, 13-21, 37-46, 346, 347, 393 Prunum, 12 Pseudicyema, 123 Pteroctopus, 122 Pterygioteuthis, 122, 127, 131, 158, 410 Ptilosarcus, 282 pugettensis, Scoloplos, 293, 294 pulchella, Natica, 43 pulchellum, Caecum, 37 pulicarius, Conus, 10 pullus, Nassarius, 6, 11 Pulmonata, 55-62, 209-219, 321-331, 333-349 punctatus, Pleurobranchus, 243 Pyganodon, 361, 362, 371-373 Pygospio, 291, 293 pyramidalis, Oenopota, 300 Pyramidella, 6, 10 Pyrene, 6, 11 quadrasi, Nassarius, 6, 11 quadridens, Berthellina, 228 Radiolaria, 428 Radix, 333 reibischi, Bulimulus, 219 reibischi, Naesiotus, 219 reinhardtii, Liocranchia, 402 reticulatum, Bittium, 38 retroflexa, Cystophora, 268 Rhinoclavis, 1, 3-7, 10, 11, 313-320 Rissooidea, 333 Robsonella, 122 rondeleti, Sepiola, 126 Rondeletiola, 121 Rossia, 121, 129, 166, 171-175 rubescens, Octopus, 193, 194 rubescens, Partula, 24, 26-28, 31, 32 rubrocincta, Axiothella, 291, 293-295 rudis, Littorina, 8 Rumina, 353, 356-358 Saccostrea, 278 Sacoglossa, 225 saeronius, Bulimulus, 219 saeronius, Naesiotus, 219 saidensis, Berthellina, 235 salutii, Octopus, 126 Samoana, 23, 31 Sandalops, 122, 397-426 sapotilla, Prunum, 12 scabra, Cranchia, 414 Scaeurgus, 122, 126 Scalesia, 216 scalesiana, Bulimulus, 214, 219 scalesiana, Naesiotus, 214, 219 Scapharca, 377 scintillans, Watasenia, 157 Scleractinia, 377 scolopes, Euprymna, 177-192 Scoloplos, 293, 294 scutata, Berthella, 242 Scutopus, 428 Seila, 37 Semisulcospira, 81 semitorta, Trisidos, 375-396 Sepia, 87-194 Sepietta, 121 Sepioidea, 121, 123, 129, 165, 409, 412-414 Sepiola, 121, 126 Sepiolidae, 123, 129, 177-192 Sepioteuthis, 89, 122, 166, 170, 175, 409 serenitas, Berthellinops, 236 sessile, Hedophyllum, 289 setacea, Bankia, 319 setosa, Chaetozone, 295 sinicum, Umbraculum, 221, 263 sinistralis, Partula, 27 sinistrorsa, Partula, 27 Sinonovacula, 313 sloani, Ommastrephes, 178 INDEX TO VOL. 23 socialis, Polydora, 294 Solenogastres, 427 solidus, Lanistes, 14 Spirula, 121 sponsalis, Conus, 7 stagnalis, Lymnaea, 55-62, 321-331, 343, 346 steenstrupi, Prionospio, 295 stellata, Berthella, 242 Stomatopoda, 1 Strigata, Terebra, 12 Strigilla, 313 stroemi, Terebellides, 295 Strombidae, 8 Strombina, 3, 6, 11 Strombus, 1, 3-7, 10, 11 subspinosus, Nassarius, 6, 11 subulatum, Cerithiopsis, 40 Susania, 227 suturalis, Partula, 32, 33 Symplectoteuthis, 122, 157, 401, 402 taeniata, Partula, 33 tanneri, Bulimulus, 216, 219 tanneri, Naesiotus, 216 Taoniinae, 397-426 Taonius, 397-426 Tapes, 375, 377 Tarebia, 81, 83 Tawera, 267 Tectibranchia, 393 tentaculata, Bithynia, 333, 337-343 tenuis, Notomastus 293, 294 Terebra, 313-320 Terebridae, 281, 313, 319 Terebellides, 295 Terebra, 1, 3-7, 10-12 tessellatus, Pleurobranchus, 242 testacea, Agaronia, 12 testudinalis, Acmaea, 37 Teuthoidea, 121, 129, 165, 397-426 Teuthowenia, 397-426 Thais, 63-73 Tharyx, 294, 295 Thaumeledone, 122 Thecosomata, 225 Thelidioteuthis, 122 Thiaridae, 81 Thiptodontidae, 225 Thysanoteuthidae, 398, 402, 404, 405, 408, 409, 412, 413 Todarodes, 122 Tonnacea, 8 torta, Trisidos, 377 tortuosa, Trisidos, 375-377, 379, 381-383, 385, 388, 393, 394 translirata, Anachis, 37 Trapezium, 381 Tremoctopodidae, 408 Tremoctopus, 408, 409 435 trigonura, Abralia, 135-163, 205 Trilobita, 319 Triphora, 37-46 Triphoridae, 40 Trisidos, 375-396 trivittatus, Nassarius, 37 truncatus, Bulinus, 328 tuberculata, Melanoides, 83 tumidus, Polinices, 4-6, 10, 11 tupala, Berthella, 242 turricula, Oenopota, 300 Turridae, 281-312 Tylodina, 225, 226 Tylodinella, 225, 226 Tylodinidae, 226 uber, Polinices, 6, 11 Umbraculacea, 225 Umbraculidae, 221, 224-226 Umbraculum, 221, 225, 226, 263 Umbrellidae, 225 undulata, Alasmidonta, 372 unicirrhus, Scaeurgus, 126 Unionacea, 361, 362 Unionidae, 361-374 urceus, Strombus, 11 Utterbackia, 361, 362, 369, 371-373 Valbyteuthis, 405 valdiviae, Liocranchia, 422 Vallisneria, 15-17 Valvata, 333-349 Valvatoidea, 333 Vampyromorpha, 122, 173 Vampyroteuthis, 122, 173, 414 varium, Bittium, 40 Velesunio, 362 veranyi, Abralia, 135 versicolor, Nassarius, 6, 12 versicolor, Pyrene, 6, 11 vertagus, Rhinoclavis, 5, 6, 11 Vexillum, 6, 10, 11 vibex, Nassarius, 37 vincta, Lacuna, 37 virginica, Crassostrea, 271-279 Viviparus, 329, 333-349 viviparus, Viviparus, 329, 333-335, 338, 339 volutella, Olivella, 6, 12 vulgaris, Loligo, 111, 116, 166-168 vulgaris, Octopus, 111, 114, 122, 126, 193, 195, 196, 198 Watasenia, 157 wautieri, Ferrissia, 346 willistoni, Drosophila, 81, 82, 85 ximenes, Conus, 6, 12 yongei, Trisidos, 377, 395 *zelandiae, Bathyberthella, 221, 222, “248-253, 263 Zostera, 282 WHY NOT SUBSCRIBE TO MALACOLOGIA? ORDER FORM Your name and address Send U.S. $17.00 for a personal subscription (one volume) or U.S. $27.00 for an institutional subscription. Make checks payable to “MALACOLOGIA.” Address: Malacologia, Academy of Natural Sciences Nineteenth and the Parkway, Philadelphia PA 19103, U.S.A. AWARDS FOR STUDY AT The Academy of Natural Sciences of Philadelphia The Academy of Natural Sciences of Philadelphia, through its Jessup and McHenry funds, makes available each year a limited number of awards to support students pursuing natural history studies at the Academy. These awards are primarily intended to assist predoctoral and immediate postdoctoral students. Awards usually include a stipend to help defray living expenses, and support for travel to and from the Academy. Application deadlines are 1 April and 1 October each year. Further infor- mation may be obtained by writing to: Chairman, Jessup-McHenry Award Committee, Academy of Natural Sciences of Philadelphia, 19th and the Parkway, Philadelphia, Pennsylvania 19103, U.S.A. UNITAS MALACOLOGICA Eighth International Malacological Congress Organizing Committee: Laszlo PINTER (president); Endre KROLOPP (secretary); Karoly DATE: PLACE: TOPICS: LANGUAGES: EXCURSIONS: BABA; Tamas DOMOKOS; Levente FUKOH; Gyula KOVACS; Laszlo MERENYI; Ede PETRO; Andor RICHNOVSZKY; Andras VARGA First Circular August 29—September 3, 1983. Budapest, Hungary. The Congress is open to any topic of malacology (recent and fossil). Further- more, in case of sufficient number of people, we organize Colloquia in the fol- lowing subjects: 1. Quaternary malacology and fauna genesis 2. Non-marine mollusc biogeography 3. Sphaeriidae We organize meetings for the EIS and for museum curators, also the General Assembly of the UM. English, French and German. Preceding, during and after the Congress we shall organize collecting tours and excursions. Further details shall be issued in the Second Circular. Laszlo Pinter Hungarian Natural History Museum Budapest Baross u. 13. H-1088 Hungary 23(2) INSTRUCTIONS FOR AUTHORS 1. MALACOLOGIA publishes original re- search on the Mollusca that is of high quality and of broad international interest. Papers combining synthesis with innovation are par- ticularly desired. Papers of local geographical interest should be submitted elsewhere, as should papers whose primary thrust is physi- ology or biochemistry. Nearly all branches of malacology are represented on the pages of MALACOLOGIA. 2. Manuscripts submitted for publication are received with the tacit understanding that they have not been submitted or published elsewhere in whole or in part. 3. Manuscripts may be in English, French, German or Spanish. Papers in lan- guages other than English must include a translation of the Abstract in English. Authors desiring to have their abstracts published in other languages must provide the translations (complete with main titles). Include all foreign accents. Both American and British spellings are allowed. 4. Unless indicated otherwise below, con- tributors should follow the recommendations in the Council of Biology Editors (CBE) Style Manual (ed. 4, 1978), available for U.S. $16.00 from CBE, 9650 Rockville Pike, Bethesda, MD 20014, U.S.A. 5. Be brief. 6. Manuscripts must be typed on one side of good quality white paper, double-spaced throughout (including the references, tables and figure captions), and with ample margins. Tables and figure captions should be typed on separate pages and put at the end of the manuscript. Make the hierarchy of headings within the text simple and consistent. Avoid internal page references (which have to be added in page proof). 7. Choose a running title (a shortened version of the main title) of fewer than 50 let- ters and spaces. MALACOLOGIA 1983 8. Provide a concise and informative Ab- stract summarizing not only contents but re- sults. A separate summary generally is super- fluous. 9. Supply between five and eight key (topic) words to go at the end of the Abstract. 10. Use the metric system throughout. Mi- cron should be abbreviated um. 11. Illustrations are printed either in one column or the full width of a page of the jour- nal, so plan accordingly. The maximum size of a printed figure is 13.5 x 20.0 cm (prefera- bly not as tall as this so that the caption does not have to be on the opposite page). 12. Drawings and lettering must be dark black on white, blue tracing, or blue-lined paper. Lines, stippling, letters and numbers should be thick enough to allow reduction by V2 or Уз. Letters and numbers should be at least 3mm high after reduction. Several drawings or photographs may be grouped to- gether to fit a page. Photographs are to be high contrast. High contrast is especially im- portant for histological photographs. 13. All illustrations are to be numbered se- quentially as figures (not grouped as plates or as lettered subseries), and are to be arranged as closely as possible to the order in which they are first cited in the text. Each figure must be cited in the text. 14. Scale lines are required for all non- diagrammatic figures, and should be con- venient lengths (eg. “200 um,” not “163 ит”). Magnification in captions are not acceptable. 15. All illustrations should be mounted, numbered, labeled or lettered, i.e. ready for the printer. 16. A caption should summarize what is shown in an illustration, and should not dupli- cate additional information given in the text. Each lettered abbreviation labeling an indi- vidual feature in a figure must either be ex- plained in each caption (listed alphabetically), or be grouped in one alphabetic sequence after the Methods section. Use the latter method if many abbreviations are repeated on different figures. 17. Tables are to be used sparingly, and vertical lines not at all. 18. References cited in the text must be in the Literature Cited section and vice versa. Follow a recent issue of MALACOLOGIA for bibliographic style, noting especially that serials are cited unabbreviated. Supply pagi- nation for books. Supply information on plates, etc., only if they are not included in the pagination. 19. п systematic papers, synonymies should not give complete citations but should relate by author, date and page to the Litera- ture Cited section. 20. For systematic papers, all new type- specimens must be deposited in museums where they may be studied by other scien- tists. Likewise MALACOLOGIA requires that voucher specimens upon which a paper is based be deposited in a museum where they may eventually be reidentified. 21. Submit each manuscript in triplicate. The second and third copies can be reproduc- tions. REPRINTS AND PAGE COSTS 22. When 50 or more reprints are ordered, an author receives 25 additional copies free. Reprints must be ordered at the time proof is returned to the Editorial Office. Later orders _ cannot be considered. For each authors’ change in page proof, the cost is U.S. $3.00 or more. 23. When an article is 10 or more printed pages long, MALACOLOGIA requests that an author pay part of the publication costs. SUBSCRIPTION COSTS 24. For Vol. 24, personal subscriptions are U.S. $17.00 and institutional subscriptions are U.S. $27.00. For information on Vol. 25, ad- dress inquiries to the Subscription Office. VOL. 23, NO. 2 MALACOLOGIA 1983 CONTENTS G. COPPOIS & C. GLOWACKI Bulimulid land snails from the Galapagos: 1. Factor analysis of Santa Grue Island SDbGIOS {gar 2.2 2 Sted Ve MAT a ste eee ARE 209 В. С. WILLAN New Zealand side-gilled sea slugs (Opisthobranchia: Notaspidea: Pletrobrarichidae) 7%. Oy МК ль Fe ack ee acre ET 221 N. E. BUROKER Sexuality with respect to shell length and group size in the ut Oyster Crassosived GIDAS ое en о LR M NRC RES 271 R. L. SHIMEK Biology of the northeastern Pacific Turridae. |. Ophiodermella .............. 281 P. W. SIGNOR Ill Burrowing and the functional significance of ratchet sculpture in tur- oO e pa o eo {аа ед Re AR A 313 J. SEUGÉ & R. BLUZAT Effets des conditions d'éclairement sur de potentiel reproducteur de Lymnaea stagnalis (Gastéropode Pulmoné) ............................. . 321 M. MARTOJA & M. TRUCHET Données analytiques sur les concretions du tissu conjonctif de quel- ques gastéropodes d'eau dOUCE 2.2.4 Oe Nie ee 333 D. G. BUTH & J. J. SULOWAY Biochemical genetics of the snail genus Physa: a comparison of pop- A nations BF WO SECIS ZU... a LV E RNA O 351 wa P. W. KAT Genetic and morphological divergence among nominal species of SuM North American Anodonta (Bivalvia: Unionidae) ................. Е 3611188 B. MORTON The biology and functional morphology of the twisted ark Trisidos N semitorta (Bivalvia: Arcacea) with a discussion on shell “torsion” in the 2% O О И RIN а NE 375.008 N. A. VOSS & R. S. VOSS Phylogenetic relationships in the cephalopod family Cranchiidae "CAS arar To] ERA IU O OA he tae ta PN 397. м LETTER TO THE EDITORS 000 A. H. Scheltema. On some Aplacophoran homologies and diets ................. 427 30 INDEX TO VOL 23 No. 1-2. RR о ‚429 № i A ki So ACME = SEP 9 1983 |100 CAMBRIDGE STREET | “CHARLESTOWN, MASS. | | QU 3 2044 072 160 484 ЗА ео ir 5 L 3 г ARI A RE У СЯ к ТУВЕ ba . AS a ER TE RES TOUTE vee HER 2 + 2 1 > Fi =. A RENE AER ó RR: és u no m а " Ne A da we “ng CRE Van, a Er a с, < r pt riet “ » x 3 : iota : pron PR pére x 2 y ere Edo eet se eo PE ere A da ARA SAA AD ra Ad Mn not > ROT rn Krane, | a be nn x y o AA A tate : Ben A 7 > x ae wen perte vee A Ne re eu NP tt Am du Re de. TS clerk M ee - RAMA AA ron EN м ль ae tn Sli mika AA eo Tae a eters EE Laits о ns on nn o уж и NE net LT ee een A RE ae ape gi 4 Я 2 E or ee Wr im > > a Pa a te See ve. gr Débat Coe e 2. 0 nom A » NN AU. qee > ee a rw NPS WEDER ve Is re One nee DS heh an il ых = en an on gg Clear Lea RS APS GPO End > 2 er nina II ot Reems о ; 2 nn de en U ren un or ER - a = - nen nn m ir = ra RA artes oe IR CT Meee Peale eee ee ee ки pees urn e MA on ¡A A e Nearer nan RMS DCE EE ва у LD prev pen TA RE PES PETRER € ha ene de. En ETRE DAL IED COR IAL TER ин ten aon gr ie te AA LA er A dl AREAS ; > x Sw on ep de rs MANE VE RENE < . Er фи. a Sew pes POET qe Dep hin, A e es мл ù Con mar LT ce y Lane x ur \ у че речи ee o pm ne ENT er wi y № . An A ' CALE NES nenn RARE D АЗЛК