мо а inate’ Be ane at say MES ir OT FRE TH . nett ces HI PERTE TEEN ua о ame: an QUE РЕНО an Ka Pe pee TES win A id EE eds | Y TE MA Lee A y ” ener vera 795% pas 5, OCR PETITE aie Rat va we yee ws revere © 4 ты TFT PPT) Я Fee HARVARD UNIVERSITY e Library of the Museum of Comparative Zoology военные — ———— 1988 ALACOLOGIA 25th Anniversary MGZ LIBRARY JAN 26 1988 HARVARD UNIVERSITY tional Journal of Malacolog y Revi sta ола] de Malacologia LAN a irnal International de Malacologie be N Международный Журнал Малакологии р Internationale ЕЯ Zeitschrift MALACOLOGIA Editors-in-Chief: GEORGE М. DAVIS ROBERT ROBERTSON Editorial and Subscription Offices: Department of Malacology The Academy of Natural Sciences of Philadelphia Nineteenth Street and the Parkway Philadelphia, Pennsylvania 19103, U.S.A. Associate Editors: JOHN B. BURCH University of Michigan, Ann Arbor ANNE GISMANN Maadi, Egypt Assistant Managing Editor: CARYL HESTERMAN MALACOLOGIA is published by the INSTITUTE OF MALACOLOGY, the Sponsor Members of which (also serving as editors) are: KENNETH J. BOSS, President Museum of Comparative Zoology Cambridge, Massachusetts JOHN B. BURCH, Vice-President MELBOURNE R. CARRIKER University of Delaware, Lewes GEORGE M. DAVIS Secretary and Treasurer CAROLE S. HICKMAN University of California, Berkeley PETER JUNG, Participating Member Naturhistorisches Museum, Basel, Switzerland JAMES NYBAKKEN, President-Elect Moss Landing Marine Laboratory, California ROBERT ROBERTSON GLYDE:F: E: ROPER Smithsonian Institution Washington, D.C. W. D. RUSSELL-HUNTER Syracuse University, New York SHI-KUEI WU University of Colorado Museum, Boulder J FRANCIS ALLEN, Emerita Environmental Protection Agency Washington, D.C. ELMER G. BERRY, Emeritus Germantown, Maryland Copyright © 1988 by the Institute of Malacology - % ERRATA, EMBERTON, 1988 Throughout, change: (1) (Say, 1816)” to "(Say, 1817)" (2) "(Webb, 1954а)” to "(Webb, 1954)” (3) "Neohelix_albolabris hubrichti” to “Neghelix albolabris bogani” р. 161, right column, paragraph 3, line 8: Change “Kimura, 1982” to “Kimura, 1983” р. 162, left column, paragraph 1, lines 5,6: Change “Woodruff & Gould 1978” to “Woodruff, 1978” р. 163, left column, paragraph 2, line 2: Change “Babrakzai & Miller, 1975” to “Babrakzai et al., 1975" р. 164, right column, paragraph 1, lines 11, 12: Change “Tiller, 1986” to "Tillier, 1985" р. 167, right column, paragraph 1, lines 8-10: Omit parentheses from around dates of Tryon, Binney & Bland, and Von Martens р. 182, left column, paragraph 2: Change "Maryck" to “Mazyck" р. 225, То "Triodopsis”. add “Rafinesque, 1819” To "Webbhelix”, add “Emberton, new genus” Te "Xolotrema”, add "Rafinesque, 1819” р. 226, right column bottom line: Change "Grimm (1976)” to "Grimm (1975)" р. 237, right column, paragraph 1. lines 11, 12 from bottom: Put "N, alleni allenj into italics р. 241, right column, lines 3,4: Change "Eberhard (1986)” to "Eberhard (1985)" р. 247, left column, top line: Change "Gould, 1985” to “Gould, 1984” р. 249, left column, paragraph 2, line 7: Change "chadwjcki” to "chadwicki (Ferris, 1907)" р. 249, left column, paragraph 2, line 15: Change "yaversensig” to "waversensis ("Leach” Pilsbry, 1894)" р. 257, left column, paragraph 6, line 1: Change “Studied material” to “Studied material (Syntypes)” р. 260, right column, paragraph 5, line 1: Change “(holotype and paratypes)” to “(Syntypes)” р. 260, right column, paragraph 5, line 18: Delete "(HOLOTYPES)" Literature Cited: Insert the following: Ayala, FJ., Medgecok, D., Zumwalt, GS. & Valentine, J.W. 1973. Genetic variation in Tridacna maxima, an ecological analog of some unsuccessful evolutionary lineages. Evolution 27: 177-191. Bookstein, F., Chernoff, R., Elder, R., Humphries, J., Smith С. & Strauss, К. 1985. Morphometrics in Evolutionary Biology: the Geometry of Size and Shape Change, with Examples from Fishes. Academy of Natural Sciences, Special Publication 15 Gould, SJ, 1984. Covariance sets and ordered geographic variation in Cerion from Aruba, Вопайс and Curagao: A way of studying nonadaptation. Systematic Zoology. 33:217 237 Hubricht, L.. 1971. The land snails of South Carolina. Sterkiana, 41:41-44 Maze, RJ. & Johnstone, С.. 1986. Gastropod intermediate hosts of the meningeal worm Parelaphosyongylus tenuis in Pennsylvania: observations on their ecology. Canadian Joumal of Zoology. 64: 185-188. Miles, C.D., 1983. Land snails (Polygridae) as a source of anti-A agglutinin for typing human blood. Bulletin of the American Malscological Union. 1:97-98. Nei, M., Tajima, Е. & Tateno, Y. 1983. Accuracy of estimated phylogenetic trees from molecular data П. Gene frequency datas. Joumal of Molecular Evolution. 19: 153- 170 Nichols, E.A., Chapman, V.M. & Ruddle, Е.Н. 1973. Polymorphism and linkage for mannosephosphate isomerase in Mus musculus. Biochemical Genetics, 8:47-53. Literature Cited: Delete the following, which are not mentioned in the text: Carson, 1982; Dixon & Brown, 1979; Patterson & Burch, 1978; Pilsbry, 1895: Pilsbry, 1905; Pilsbry, 1946: Pilsbry, 1948; Poulick, 1957; Randles, 1900; Reeder & Rogers, 1979; Rogers el al. 1980; Shaffer, 1984; Simpson, 1944; Solem, 1972; Solem, 1975. m! De р $ р x wif О e e Y = > d >, Ben) $ qe o ~~ $s gj o E q ¿y | @ pa че pura As yA prt i sm dam en oo ENT REE о ” A AA А. assy) dios ‚Iren Id An a u й (> Aa! ‚te ¿rio ol Ми ой Tun y yr es parties А. 4 À ur ö № dt À artos muta Ne az RR A ag cree tery a Mad i N LES st ni ines tas $ es my Amel 5 Gal st 7 vi ye hh Me tir vy Ут. ит 3% 4 ‘ i er hy! i peel Wy var N 4 | A | e va A, A” т > = f nr nal emt Mi u Ed 4 > ni? 1 / у Ane stele щи" $ 4 h i us } } ñ esit f i 1! фоне hap у ‘> A m Maty où RON car it gee cow en thet SE rm } ral | wer м A | Le) р IT it pages Ее sm | ing Mal { à y . 34 ee ul 1108 6 L LE Tiel | яя & ПЕ u ade » etgh wy Toe ar veils A Aena yor] пей AT er u ee ee oo e wh, N A m Far x ra 4 myst) AT tar; j iad re OR NP A. DENT “vee bah te se Mein ane ben ро de ee) er U Усы; ud aqi Si! po Сие lid т wei. Г «0 al om АЙ 65 outs, ди. a Mis LE AR a dl Ан > 7 y coûte © UA Er ИК» then, Amr pt Fete o do (ee rt a+ Pr к 4 pe , A u re ps ый FR k y ae 4 1988 EDITORIAL BOARD J. A. 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. Portland, Texas, U.S.A. B. C. CLARKE University of Nottingham United Kingdom R. DILLON College of Charleston SC, U.S.A. C. J. DUNCAN University of Liverpool United Kingdom E. FISCHER-PIETTE Muséum National d'Histoire Naturelle Paris, France V. FRETTER University of Reading United Kingdom E. GITTENBERGER Riiksmuseum van Natuurlijke Historie Leiden, Netherlands F. GIUSTI Universita di Siena, Italy A. N. GOLIKOV Zoological Institute Leningrad, U.S.S.R. 5. J. 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 J. A. HENDRICKSON, Jr. Academy of Natural Sciences Philadelphia, PA, U.S.A. K. E. HOAGLAND Association of Systematics Collections Washington, DC, U.S.A. B. HUBENDICK Naturhistoriska Museet Göteborg, Sweden S. HUNT University of Lancaster United Kingdom R. JANSSEN Forschungsinstitut Senckenberg, Frankfurt am Main, Germany (FederalRepublic) 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 А. 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 Faculte des Sciences Brest, France C. MEIER-BROOK Tropenmedizinisches Institut Tübingen, Germany (Federal Republic) H. K. MIENIS Hebrew University of Jerusalem Israel J. E. MORTON The University Auckland, New Zealand J. J. MURRAY, Jr. University of Virginia Charlottesville, U.S.A. R. NATARAJAN Marine Biological Station Porto Novo, India J. OKLAND University of Oslo Norway T. OKUTANI University of Fisheries 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 MHZ: Academia Sinica Qingdao, People's Republic of China N. W. RUNHAM University College of North Wales Bangor, United Kingdom S. G. SEGERSTRÄLE Institute of Marine Research Helsinki, Finland G. A. SOLEM Field Museum of Natural History Chicago, U.S.A. Е. STARMUHLNER Zoologisches Institut der Universität Wien, Austria У. 1. STAROBOGATOV Zoological Institute Leningrad, U.S.S.R. М. ЭТВЕЕЕ Universite de Саеп France J. STUARDO Universidad de Chile Valparaiso T. E. THOMPSON University of Bristol United Kingdom S-’TIEHIER Museum National d’Histoire Naturelle Paris, France R. D. TURNER Harvard University Cambridge, Mass., U.S.A. W. S. S. van BENTHEM JUTTING Domburg, Netherlands J. A. van EEDEN Potchefstroom University South Africa N. H. VERDONK Rijksuniversiteit Utrecht, Netherlands B. R. WILSON Nedlands, Western Australia H. ZEISSLER Leipzig, Germany (Democratic Republic) A. ZILCH Natur-museum und Forschungs-Institut Senckenberg Frankfurt-am-Main, Germany (Federal Republic) MALACOLOGIA, 1988, 28(1-2): 1-15 THE SCORING OF POLYMORPHIC COLOUR AND PATTERN VARIATION AND ITS GENETIC BASIS IN MOLLUSCAN SHELLS A. J. Cain Department of Zoology, University of Liverpool, P.O. Box 147 Liverpool L69 3BX, England ABSTRACT Present modes of scoring the phenotypic and genetic variations in molluscan shell polymorph- isms vary widely. Some disregard accepted conventions for distinguishing phenotypic from genetic variation, others misuse morph to mean any variant whether discontinuous or not. The requirements for the recognition and description of a polymorphism are discussed, and for acceptable notations for its phenotypic manifestation and genetic basis. Key words: Shell polymorphism; discontinuous variation; morph; description; notation; symbolization. INTRODUCTION The methods used by Cain & Sheppard (1957), Cain, King & Sheppard (1960) and Cain, Sheppard & King (1968) for scoring colour and banding polymorphism in the snail Cepaea have been explicitly followed by later workers (Pettitt, 1973; Roth, 1981; Roth 8 Bogan, 1984) for other gastropods, both ter- restrial and marine. However, as there have been considerable departures in these pa- pers from Cain and Sheppard's methods, and as incorrectly applied methods can actually conceal the nature of the variation in a spe- cies, it is necessary to review the descriptive and notational procedures used by these and other authors, to determine which are the most suitable for the ends in view. Continuous variation in any one character requires only a definition of that character and simple measurement or ranking. The problems discussed in the present paper are of discontinuous variation. Phenotypic varia- tion can be either continuous or discontinu- ous; genetic variation, by the nature of the genes, can only be discontinuous, and if it occurs within a population (except as a rare mutation) it is necessarily a polymorphism as defined by Ford (1940, 1945). Phenotypic polymorphism is not a necessary conse- quence of genetic polymorphism, which may produce continuous phenotypic variation; and phenotypic polymorphism may be produced without genetic mediation as in the solitary and gregarious forms of locusts, and workers as against queens in social Hymenoptera. (1) When mediated genetically, phenotypic poly- morphism normally occurs in relation to sex, mimicry (e.g. Clarke & Sheppard 1960a, b), apostatic selection (Clarke, 1964; Clarke & O'Donald, 1964), industrial and other melan- isms, and other phenomena of particular ev- olutionary interest. It is important, therefore, that it should be recognised, described, sym- bolized, and separated clearly from continu- ous variation. PROTOCOL FOR A POLYMORPHISM For a complete description and notation for a polymorphism, the following elements are required: 0.1. Descriptions of the different morphs (to use the term introduced by Huxley, 1955) either in absolute or relative terms and pref- erably both, e.g. for colour variation by refer- ence to a standard colour atlas, and ex- pressed as a difference from other forms. 0.2. Statements, with supporting evidence, of discontinuity. 0.2.1. Segregations within bred material are the best evidence (but many species cannot be bred in the laboratory). 0.2.2. Clear segregation in random sam- ples, taken from an area small in comparison with the normal dispersal distances of the species, is also evidence that the forms con- cerned are morphs, not individual variants picked out of a continuum of variation. Some forms may segregate wherever they are 2 CAIN found, but one still needs random samples to determine their separation; museum samples are seldom random. The importance of obser- vational evidence for discontinuous variation within populations was specially emphasized by Diver (1939). In very large samples, or when many sam- ples are looked at together, the full range of individual variation within morphs can be ob- served, and individuals apparently intermedi- ate between otherwise distinct forms may appear. This can happen in two ways: 0.2.2.1. The segregation is clear at any one locality, but an accumulation of modifiers in a particu- lar population, or environmental influences, may shift the expression of the alleles so that, for example, a segregation genetically of dark pink and pale pink shell colour may become in expression one of medium pink and faint pink. If all the samples are considered together, there may be a complete chain of forms connecting dark pink to pale pink. Breeding experiments will give the true explanation, but if the shift in average expression is confined to one or a few populations out of many, one may suspect the right answer merely from the samples. If a segregation appears at all, then there is a major gene difference, and the fact that it is obscured elsewhere does not abolish that finding. It does suggest that the discrete differences may be strongly affected by environmental or other genetic influences. 0.2.2.2. The expression of a morph is highly variable everywhere, and a few individuals appear intermediate between it and other morphs in any large sample. Breeding exper- iments are almost essential here, as in the studies of polymorphism in Theba pisana by Cowie (1984) and Cain (1984). Observational data can suggest the true explanation, e.g. for mantle pigmentation in Monacha cantiana (Cain, 1971). Here the difficulty was com- pounded both by the coarseness of the pig- mentation, which meant that only ranking could be used, not scoring against a colour atlas, and because one form was un- pigmented (except for an anal blotch) and might have been merely the extreme of vari- ation of the pigmented form. Nevertheless the frequency distribution of the ranked forms suggested two peaks, a broad one in the pigmented class and a (necessarily) narrow one in the unpigmented. Evidence that a genetic polymorphism was indeed involved was obtained by breeding. 0.3. A notation for the morphs defined by means of 0.1 and 0.2. This may be nominal(using names) or symbolic (using let- ters, numbers, etc.). Names may be descrip- tive (e.g. rosea, lutea, quinquefasciata) or ascriptive (e.g. baudonia). Symbols can be combined more readily than names and can give a partial analysis of particular forms; for example DY00300 al in Cepaea has the dark yellow colour morph, the banding represented only by the middle band, and the lip of the adult shell white, without the usual pigment. Nevertheless, there are situations in which names are of use (pp. 5, 9 below). 0.4. Statements, with supporting evidence, about the genetic control of the morphs de- fined by 0.1 and 0.2. The evidence may be 0.4.1. segregations within random samples as in 0.2.2, merely suggesting that the poly- morphism has a genetic basis; 0.4.2. observations of segregation within a sample known to be a single brood, e.g. Pilsbry (1912) on apex colour in Liguus; Mayer (1902) on uterine young of Partula; or 0.4.3. full genetic data from numerous matings or whole lineages. Since type 0.4.3 includes 0.4.2 and is of greater evidential force than 0.4.1, there is no need for 0.4.1 and 0.4.2 if 0.4.3 is obtainable. Type 0.4.1 is also open to the objection that it does not exclude the case of a polymorphism medi- ated purely phenotypically, aphenomorphism (for types of polymorphism see Cain, 1977). There is the possibility even in molluscs of such a phenomenon (e.g. in the bivalve Corbicula, Prezant & Chalermwat, 1984). Further, on occasion evidence of type 0.4.3 may help to clear up confusing or misleading evidence of type 0.4.1, as in the case of the snail Theba pisana (Cain, 1984; Cowie, 1984). 0.5. Statements of the genetic relations between the morphs, e.g. multiple allelo- morphism, dominance, epistasy, linkage, com- plementarity, also from evidence of type 0.4.3 above. Dominance relationships are espe- cially needed since they are not expressed in the genetic symbolization of a polymorphism (see below, p. 3). 0.6. A notation for the genetic basis as ascertained from type 0.4.3 evidence as in- terpreted in 0.5, conforming to normal genetic practice. As exemplified below, this notation cannot be the same as that in 0.3 since it refers not to the phenotypes but to the under- lying genetic basis. Ford (1955) pointed out a similar confusion in the nomenclature then accepted for the blood groups (. . . “in the current literature it is often impossible to de- SCORING MOLLUSCAN SHELL POLYMORPHISMS 3 termine whether a given symbol refers to a gene or to an antigen”). Where morphs and alleles approximately correspond, it is useful to have a corresponding similarity of refer- ence, but not, however, an identity which could lead to confusion. На species can be bred easily, all these elements (0.1-0.6) can be produced, but much useful, indeed essential, work can be done directly from shell samples provided that the exact status of the specimens (see 0.2.2) is fully understood. They must form random samples of sufficient size from a restricted area (Diver, 1939, p. 114 and earlier authors referred to therein). If not, as in many highly biassed museum collections, they give no basis even for separating continuous from discontinuous variation. SYSTEMS OF NOTATION 1. Genetic notation The usual conventions for symbolizing genes are given in most textbooks of genetics (e.g. Avers, 1984). The principles are well set out in a proposal for a uniform nomenclature for bacterial genetics (Demerec, Adelberg, Clark & Hartman, 1968, developing a system by Demerec, 1958), which stresses the im- portance of distinguishing between “symbols representing the genotype [of a bacterial strain] and abbreviations of words which de- scribe phenotypic properties”. “Each locus of a given wild-type strain is designated by a three-letter, lower-case, italicized symbol”, the letters being chosen to recall the pheno- typic change produced by mutants, e.g. ага is that locus which affects the response of the cell to arabinose. A recognition that a locus so designated is composite is shown by italicized capitals, e.g. ara A, ara B, ara D. They rec- ommend that all mutants should be desig- nated by serial numbers only, since different mutants at the same locus may affect it in different ways. Their system is based on the probability that the exact sequence of nu- cleotides for each allele can be determined, so that the allele can be recognised as such irrespective of its effects, and the fact that the actual phenotypic effect of a given allele “may be readily altered by mutations at other loci or by changes in the environment”. In the comparatively primitive state of mollus- can genetics, the symbolization is more meaningful if the alleles are designated by something obviously referring to their pheno- typic effects, since this is all that is known about them. Moreover, apart from dominance and epistasy and rare complementarity there seems as yet comparatively little genic inter- action in visible polymorphisms of molluscs. Phenotypic effects, they recommend, should be either stated in words or abbreviated from descriptive words or phrases, the abbrevia- tions being clearly defined the first time they appear (in a given paper). Phenotypic abbre- viations should never consist of three-letter, lower-case italicized abbreviations, which are reserved for genetic loci. Similarly in the ‘Rules for nomenclature of genes, chromosome anomalies and inbred strains’ put forward by an international com- mittee for workers on mice in the ‘Mouse News Letter’ no. 72 (officially not a publication) it is recommended that phenotype symbols should be the same as genotype symbols “except that symbols for phenotypes should be in capitals, not italicized, and with superscripts lowered to the line”. Careful provision is made for priority; the standardization of nomenclature between species to show homology is strongly recom- mended; and various rules are given for the symbolization of subunit structure etc., mouse genetics being a good deal more advanced than molluscan genetics. There is now so general an agreement on the use of italicized letters for loci with itali- cized superscripts for alleles, in all sorts of plants and animals, that no discussion is necessary. In polymorphisms, no one allele at a locus can be singled out as wild-type, so the former use of across or plus sign for wild-type and letters for mutant alleles is precluded (Cain & Currey, 1963). Moreover, as there may be many alleles at a locus, the simple use of a capital letter for the dominant and a lower case of the same letter for the recessive is also precluded. Several authors have used letters that have some meaning in relation to the descriptions of the morphs. Thus in the notation for Cepaea given by Cain, Sheppard & King (1968), for shell colour the locus is C, with alleles C® for brown, CP? for dark pink, and CP’ for pale yellow; В is used for presence or absence of banding, S for the presence or absence of spread bands, a form in which the banding pigment is diffused over the whole extent of the shell normally occupied by the black bands, and so on. For some morphs, the initials of old varietal names were adopted, e.g. / for punctate bands (var. inter- 4 CAIN rupta), since P was in use for degrees of pigmentation of the bands and lip. Such con- nection between the symbols and the morph names or descriptions is not only usual, but convenient in memorizing the symbols. Nev- ertheless Cain and Sheppard were careful to symbolize morphs differently from genes, by using roman upper-case letters, even if they were the same letters. Thus DP would be the phenotype of homozygous СР, but also of heterozygotes of this allele with alleles lower in the dominance hierarchy, e.g. CPP CPP, and CPP CP”. The genetic basis of a pheno- type and the characteristics of that phenotype must be distinguished, as can be seen from the following considerations. 1.1. The same phenotype may have a different genetic basis in different individuals, because of dominance (just exemplified), but also because of epistasy. An unbanded white-lipped shell carries either one or two unbanding alleles B°, but at the locus P it may have either white lip, Р^ (albolabiate) or РТ (hyalozonate) with pigment in neither lip nor bands; since in this morph the bands are suppressed by BO, these alleles cannot be distinguished. 1.2. Complementarity also can cause con- fusion. A shell with darkly pigmented bands may be, and most usually is, РМ, but rarely it may be PT PT (hyalozonate, with no band pigment) combined with O° O° (orange-ban- ded, with dilute band pigment); see Cain, Sheppard & King (1968). 1.3. Without any interaction, the same phe- notype may be produced by genes at different loci. An orange-banded form is produced in Cepaea nemoralis by O°, but also by the allele P* which is homologous with the /urida orange-banded form in C. hortensis (Murray, 1963; Cook & Murray, 1966; Cook, 1967). 1.4. Incomplete dominance can also pro- duce the same phenotype by different means genetically, e.g. Wolda (1969) on Cepaea banding; Cowie (1984) and Cain (1984) on shell pattern in Theba pisana. 1.5. The same phenotype can be produced by a major gene difference in some individu- als, but by an accumulation of polygenic modifiers in others. The banding form 00345, with the two upper bands missing, segregates clearly in some samples and broods of Cepaea nemoralis, especially on the conti- nent of Europe, and is at a locus T (trifasciata) unlinked to that for presence or absence of bands, B, or to that, U, for reducing the five-banded form to one with only the middle band, 00300. In many British samples 00345 is connected phenotypically to 12345 by all degrees of intermediate expression of the bands, such 0:345, 10345, ::345 etc. (: mark- ing an incomplete band) and is probably polygenically controlled. Wolda (1969), how- ever, has evidence that some 00045 at least may be only an environmentally induced vari- ant of 00345. 1.6. The recognition of segregants in a random sample does not always allow us to assign them to loci even when they appear to be alternatives. Thus three very common alternative states of banding in Cepaea nemoralis are unbanded, midbanded and five-banded, 00000, 00300, and 12345 (or some minor variant of the last). In southern England it is possible to find populations containing only one of these, or any two, or all three. It would be easy to conclude, as Diver (1932) appears to have done, that 00300 shows close linkage with the colour locus, as do 00000 and 12345. A population with only dark yellow midbandeds and dark pink unbandeds would suggest this. Yet what it really contains are the supergenes (groups of tightly linked loci) for dark yellow five-banded, СОУ ВВ, and dark pink unbanded, CPP B®, but it is saturated with the wholly unlinked modi- fier U? which converts a five-banded into a midbanded pattern. So far from being an allele of unbanded and five-banded, mid- banded is not even linked to them. These examples, and they are not exhaus- tive, show the necessity of distinguishing be- tween the morphs and their genetic bases. Furthermore, since the genetic architecture of a polymorphism (or other forms of variation) may differ in different species or even popu- lations, a study of it and its relationships to the polymorphism is of considerable evolutionary interest. It is recommended, therefore, that the gen- erally accepted practice of italicized capitals for loci and italicized superscripts for alleles should be used only for genetic notation. 2. Morph notation When Cain and Sheppard began the work on Cepaea nemoralis, they attempted to de- scribe the colour and some other variation by means of the numerous varietal names listed in Taylor's Monograph (1914) and, like Taylor, they used von Martens’s (1832) nu- merical system of scoring banding. The ex- cessive bestowing of varietal names on SCORING MOLLUSCAN SHELL POLYMORPHISMS 5 nemoralis shells had led to the ridiculous situation in which the same shell could be var. castanea because chestnut brown, quin- quefasciata because five-banded, and bris- sonia because both brown and five-banded. That much of the variation in Cepaea resulted from different combinations of the same char- acters had been pointed out several times before, e.g. by Diver (1939), and by Adams (1896). Adams advocated combinations of varietal names (p. 17), e.g. Helix nemoralis var. rubella-minor-albolabiata (plus the band- ing formula) but remarked (p. 68) after an example with 6 names, “lt is perhaps fortu- nate that there are certain practical limits to an infinite series.” A symbolism showing the combinations was obviously much more infor- mative than a series of ascriptive names, each unintelligible without its definition, as Diver showed (1939, p. 113). Moreover, since much of the variation could be referred to four types of character, namely shell colour, band- ing pattern, states of the banding (e.g. normal or punctate, fully-pigmented or unpigmented), and colour of the lip, the same simple formula, easily pronounceable and printable, could be used in tables, text, and conversation. It could be extended easily, both by addition of new types of variation as they became known (e.g. spread-banded) and by subdivision of, or addition to, existing types, e.g. dark pink, pale pink, faint pink instead of just pink. Shell colour is placed first in the formula and sym- bolized by roman capitals, as few as possible, e.g. Y for yellow, DY for dark yellow, FY for faint yellow, YW for yellow-white (with yellow periostracum and white calcareous layers of the shell). Superscripts are not used. This is followed by the banding formula in as much detail as required, but often reduced to unbanded 00000, midbanded 00300, and five-banded 12345, the last standing for both truly five-banded shells and all the minor variants which are probably polygenic modifi- ers of the same allele, B®. 0 indicates the absence of a band. When necessary, the fusion of adjacent bands is shown by paren- theses; for example a shell with formula (12)3(45) would have effectively only three bands produced by the fusion of 1 and 2 and of 4 and 5. The formulae are often abbrevi- ated to 0, 3 and 5 as in the tables in Cain, Sheppard & King (1968). Where further detail is required, the formula can be expanded accordingly; for example Cook (1967) showed that the form 00:45 segregates from and is dominant to 00345; in such formulae a colon stands for an incomplete band (but Wolda (1969) uses a semi-colon, and the colon has also been used for a punctate band, broken up into dots). The symbol fa (fascialbate) in position 3 signifies a middle band with a white or pale stripe along one or both sides. A t instead of a number indicates a band shown only by a trace of pigment near the mouth of the adult shell. The states of the banding and lip are shown by roman letters after the band formula, for ex- ample pb for punctate bands, al for white Пр (albolabiate), hz (hyalozonate) for unpig- mented lip and bands. Lower-case letters are used inconsistently for states both recessive (al, hz) and dominant (pb) to the unmodified bands, and, also inconsistently, a capital S for the spread-banded form (dominant to un- modified). Occasionally an old varietal name is used, e.g. punctata instead of pb, also inconsistently. Murray (1963) has used the same system very effectively, with additional symbols, for the polymorphism of Cepaea hortensis. For some reason that | cannot now recall, the shell colour symbols in Cain & Sheppard (1957) but not the banding formulae were printed in italics, which was certainly wrong. This may or may not have been journal us- age. This notation was originally designed to mark phenotypic segregants, and is useful, therefore, for those not yet fully described genetically as well. Thus Cain, Sheppard & King (1968) refer to the segregants PB pale brown, FP faint pink and YW yellow-white as almost certainly belonging to the colour locus C; their retention in roman upper-case indi- cates clearly that, so far, what is known of them is only their segregation. When, as with the formula 00:45, a some- what complex set of characters is found to constitute a morph (in this case bands 1 and 2 absent, band 3 incomplete, bands 4 and 5 normal) there is some reason for using a varietal name, and Cook (1967) gives the varieties 00345 as listeria and 00:45 as donovania. Indeed, when a very complex pattern is inherited as a whole, and is not reducible to such a series of components as is much of the variation in Cepaea, the vari- etal name is often the shortest and simplest designation. This is the case, for example, with the three morphs of wing-pattern, dominula, medionigra (the heterozygote) and bimacula of the scarlet tiger moth, Panaxia dominula (see Fisher 8 Ford, 1947, for a 6 CAIN coloured plate). While the excesses of vari- etal naming in Cepaea described above have no real use in notation, description or symbol- ization, the use of some descriptive names (e.g. hyalozonata, roseozonata) and some such as donovania can be recommended. DEFINITIONS AND DESCRIPTIONS 3. Descriptions and figures Taylor (1914) frequently laments the im- possibility of understanding earlier varietal names published with neither figure nor de- scription. Sometimes they may have been thought to be self-evident, being of descrip- tive words, e.g. rubra, lutea, but this is not usually good enough. Cain & Sheppard (1954) were fortunate in being able to refer to the excellent coloured figures in Taylor's Monograph (1914). Descriptions should always be given of each morph at the time of its designation, to serve as definitions. Coloured figures are highly desirable but expensive, and not often satisfactory for slight but constant differences of hue, clear enough on specimens but easily obscured in printing a plate. A plate was published by Goodhart (1962) which shows several forms adequately, but the figure of a yellow with five bands fused looks more like a pink. In general, coloured plates should be printed first and scanned carefully, and a comment on their deficiencies included in the text. More usually, a reference to a standard colour atlas is all that can be given. 4. Statements of discontinuity Cain & Sheppard (1954, p. 90) pointed out what in their samples showed clear segrega- tion, e.g. 00000, 00300, 12345, and what was connected by frequent intermediates. (Colton (1922) did the same for varieties of the dog- whelk Thais lapillus.) In their purely genetical work, of course, this was obligatory, and they devoted much space (e.g. Cain, Sheppard 8 King, 1968) to discussion of apparent inter- mediates. In both breeding work and the examination of random samples it is essential to say what has been found to segregate from what. If form A segregates from В and from С, it does not necessarily follow that B segre- gates from C; compare the discussion of bred material of Theba pisana in Cain (1984). USAGES BASED ON CAIN AND SHEPPARD'S 5. Pettitt (1973) on Littorina Pettitt noticed the same inconvenience in scoring winkle shells as had caused Cain and Sheppard to propose their formulae for scor- ing Cepaea. He therefore proposed a system of notation for variation in Littorina saxatilis “based on that used for Cepaea as set out by Cain, Sheppard and King (1968) . . .”. In his paper, however, no definitions of morphs were given, although the word morph is fre- quently used. In my experience, scoring large samples of this species begins easily with a number of distinct forms, but further scoring produces more and more intermediates, until many apparently obvious morphs have to be abandoned. (This was also the experience of Reimchen (1979) in L. mariae.) Pettitt, there- fore, has at least in part confounded continu- ous and discontinuous variation. Breeding was not possible, since only recently have adequate techniques been produced (Atkinson & Warwick, 1983), but unfortunately Pettitt used not the morph notation but the gene notation of Cain and Sheppard, with italicized capitals and superscripts, as though the genetic basis was known. To assume, however reasonably, that the colour and banding polymorphism in Littorina is genetic does not warrant the use of a symbolism for loci and alleles. Furthermore, Pettitt proposed a notation for banding with B° for unbanded, B’ for one-banded, B? for 2-banded and so on. Unbanded corresponds phenotypically, of course, to unbanded in Cepaea, but in this genus one-banded can be any of the formu- |ае 10000, 02000, 00300, 00040, and 00005. Most of these are excessively rare in Cepaea and their genetics unknown; only 00300 is known genetically to segregate as a distinct form, and, as noted above, although it may be an alternative phenotype to 00000 and 12345, it is not an allele at their locus. In Cepaea, the banding morphs segregate not on the basis of the number of bands, but on the pattern. Pettitt remarks (1973, p. 532) that he had decided “to attempt a re-description of the phenotypes of L. saxatilis on a 'genetic' basis” (his quote marks for 'genetic') but he did not take into account the complexity of relationship between genotype and pheno- type. Pettitt does give a set of references to SCORING MOLLUSCAN SHELL POLYMORPHISMS 7% various colour standards in defining his colour forms, and sufficient indications (including figures) of other variation. In scoring sepa- rately the ground-colour of the shell, banding, the colour of the bands, tessellation, inter- rupted lines etc. he was providing a partly analytical notation, superior to amere naming of varieties. With no clear indication, however, of what is continuous and what is discontinu- ous variation, and with a tendentious interpre- tation of the variation by loci, the result is principally useful as а source of symbols for a properly-based classification. All his symbols should be converted to roman capitals and all superscripts demoted to the common line. His notation has then one possible advantage over Cain and Sheppard's, in that a prefixed capital (the former locus symbol) is now com- mon to those forms that are alternatives. For ground colour of the shell, for example, his symbols for white, fawn, grey, brown and orange now become GW, GF, GG, GB and GO. A similar arrangement for colour in Cepaea would give CB, CDB, CDY, CPY etc., and may be useful in scoring other complex polymorphisms. Atkinson & Warwick (1983) have used Pettitt's symbolization but rightly converted it to roman capitals, and by simplifying it they have removed most of the objectionable fea- tures. The necessity for marking their pattern symbols with an asterisk is not obvious, and although they refer to morphs, they give no statements about the continuity or discontinu- ity of the variation. 6. Roth (1981) on Monadenia Roth's investigation of shell colour and banding variation in the helminthoglyptid snail Monadenia fidelis is explicitly based on ran- dom samples, with emphasis on the continu- ity or otherwise of the variation. Previous workers on this species had bestowed both varietal and geographical (subspecific) names. Roth states that his notation “is modelled after the systems of Cain, Shep- pard, and King (1968) for Cepaea and Pettit [sic] (1973) for Littorina”. In fact, his notation uses roman capitals, in agreement with Cain and Sheppard's scoring of phenotypes, but in format is of a single capital for a series of exclusive phenotypes, with superscripts for each state, thereby agreeing with Pettitt's scheme. Roth gives a table of the exact composition of his random samples, scored according to his scheme, and extensive de- scriptions of the different forms he rec- ognises, with colour-atlas references, and good black-and-white photographs of some morphs. Roth's system is incomplete in that super- scripts are provided only for the ground colour of the shell, banding, and the presence or ab- sence of a basal patch; this last appears to be based on named varieties, not on his samples. The presence of a green tinge to the basal patch is noted and symbolized, but it is not made clear whether this, like pink/not pink, is a clear-cut segregation. The banding notation proposed is of two symbols, A for the periph- eral band absent, B for the shoulder band light or absent, medium, or present. Roth remarks that if the shoulder band is light or medium, “the center of the band may lack pigment; that is, the band may be rendered as two parallel lines.” The illustrations suggest that the fullest banding corresponds to what in Cepaea would be called (12)3(45), with band 1 extending very close to the suture, and band 5, unlike in helicids, extending right to the umbilical re- gion. To avoid prejudicing the question of ho- mology between helminthoglyptid and helicid bands, if the band plus the basal patch are simply numbered from above downwards, the formulae forthe conditions described would be (12)34 (all present); 1234 and ::34; and (if the shoulder band is absent when the peripheral band is absent, which seems to be the case in var. semialba Henderson) 0004. Shoulder banding is included in his statement (p. 41) that variation in his random collection is mark- edly discontinuous; and his descriptions indi- cate full pigmentation with fusion; dilute pig- mentation with or without fusion; traces of banding, again with or without fusion; and total absence. In Cepaea the occurrence of even traces of banding indicates that the allele for bands, B® is present; total absence of bands in our breeding stocks is given by B° dominant to B®, but an absence of bands could also be produced by delaying their appearance until not even traces were produced. It would be interesting to know whether the numbers of Monadenia without bands and with only traces in Roth's samples suggest two classes here also. Furthermore, if fusion is independent of pigmentation, except that heavily pigmented bands (dark band) are always fused, separate symbols should be used for fusion and band intensity. Roth's carefully-based work gives us the first analytical notation for polymorphism in a helminthoglyptid. 8 CAIN 7. Roth & Bogan (1984) on Liguus This paper is a particularly welcome contri- bution to the literature on molluscan variation. Surely no snail has ever suffered so badly from its devotees as this unfortunate animal. As Roth and Bogan point out, a plethora of varietal, and subspecific but often not geo- graphical names has served as the basis of remarkable theories of the species’ origin and spread, which are hardly tenable if the nota- tion for the variability is changed. They rightly remark that these epithets “tend to obscure rather than illuminate the relationship of one form to another” and “their use has canalized systematic and zoogeographic thought re- garding the genus Liguus”. The notation they propose, of twelve cate- gories, is of the same character as that of Roth (1981), with roman capitals as in a phenotypic notation but superscripts (for 6 categcries only) as in a genetic notation. They provide a table giving their formulae for many of the varietal names already proposed, and some sketches of particular conditions (to- gether with maps of the distributions of par- ticular character states, and the variation of diversity in Florida). No scoring of random samples is given, however, and the problem of the use of museum material is passed over in two sentences. “The characters used here are ones in which the alternative states can be seen to segregate in randomly selected material. Most museum lots were sorted by earlier workers to conform to the standard nomenclature and cannot be used to deter- mine whether a particular variation is discon- tinuous or not”. Much museum material is indeed useless for working on polymorphisms, since it is very far from being collected at random. While it is no doubt true that every one of the alternative character-states they define can be seen seg- regating in one or other museum lot, it is es- sential that full details of these lots, their scores, and why they are regarded as random should be published, both to validate the seg- regations proposed, and to allow other work- ers to consult them as standards. In the mean- time, since a definite statement is made that the notation is based on segregants, the work provides a valuable basis for further studies. One situation requires special care, namely an apparent segregation of the presence or absence of a particular character. When, as in various random samples of Cepaea, all the shells are clearly either unbanded or very heavily five-banded, no doubts need arise. When, however, there is considerable varia- tion, down to near-absence, in the category character present, there is a serious question as to whether a continuous variation is being artificially split into an apparent polymorphism of presence/absence merely because the lan- guage does not have single-word terms for very nearly absent, nearly absent, very slightly present, etc. This problem, which is particularly acute when the variation is such as to require ranking, and direct measure- ment is not possible, was considered by Cain (1971) in the case of mantle pattern in the snail Monacha cantiana. In that case, the frequencies in the different ranks in large samples suggested a bimodality of variation probably mediated by two major alleles, plus much polygenic background variation. Here again, it is necessary to give the full data for the basis of any conclusion about segregation versus continuous variation. OTHER USAGES 8. Dogwhelks The complex variation in shell colour, band- ing and sculpture in dogwhelks (Thais and Nucella) has been studied by several authors; Thais emarginata is the only gastropod in which sculpture appears to be (at least in part) polymorphic (Palmer, 1984). Colton (1922) in a paper on Thais (or Nucella) lapil- lus gave definitions of 8 color morphs, with references to the Ridgway colour chart, and a clear statement that all were quite distinct. He also produced a formula for the banding by counting the maximum number of bands, and used letters, W and D to indicate a white and dark stripe. Thus a particular combination is given as: 1. ADS 7432455) POL OO, W DW WwW D W DW DW Narrow stripes are defined as those оп а single ridge of the sculpture; wide ones in- clude 2 or 3. In labelling both white and dark stripes, Colton has produced a very descrip- tive formula, but one too cumbrous to express easily the nature of the banding variation. It is as though one should describe both the bands and the interspaces of a yellow five- banded Cepaea as SCORING MOLLUSCAN SHELL POLYMORPHISMS 3 instead of Y 12345. Palmer (1984), like many other researchers on genetics, ecogenetics and evolutionary genetics, used only categories indicated by his breeding data, and has not so far pro- duced any symbolism for alleles or loci. He gave a simple roman notation for shell colour (e.g. OR = orange, WH = white, GR-OR = grey-orange) stating explicitly which colour morphs can be recognised as discrete, and which shades of colour could not be scored reliably. In the brood in which it segregates clearly, he scored sculpture as SM (smooth) and STR (strong spirals) with continuous cat- egories of WK (weak) and MOD (moderate) for other broods lacking a clear segregation. Such combinations of letters are more remi- niscent of the words they stand for than single letters. Banding was scored only for presence or absence, corresponding genetically to two alleles at a single locus. More recently, Palmer (personal communi- cation) has achieved considerable genetic analysis of this very complex variation. He uses roman capitals and superscripts for his loci, and since there is now evidence that the colour in the outer part of the shell may be inherited independently of that in the inner part, e.g. on the columella, he prefixes O to the alleles and loci affecting the outer layers. Thus banding is mediated by a single auto- somal locus with two alleles, banded OB®, and unbanded OB”, with banded dominant. Outer shell colour is symbolized as OC with variable dominance, e.g. OC® for black, OCT for orange, and is independent of OB. A further locus, for pigment intensity, is sug- gested with Ol having no effect, OIF reducing pigment intensity partially in heterozygotes, completely in homozygotes, this last resem- bling the usage of a superscript dash for no visible effect by Cain, Sheppard & King (1968) for some alleles in Cepaea. Palmer uses the < symbol for ‘dominant to’ (e.g. BEN HE 0565) This seems a highly convenient symbolism, allowing for the repetition of symbols with different prefixes, so that if a locus for internal shell colour becomes necessary it can be symbolized as IC, as against ОС. Since it is based on actual breeding, one might suggest that it should be printed in italics. It is worth noting that Palmer describes the banding in Thais emarginata as formed, not as in pulmonates by the imposition of bands of a different pigment upon various shell ground colours, but by the regularly spaced suppres- sion of outer shell pigment. Berry 8 Crothers (1974), working on large numbers of random samples of Thais lapillus, while giving careful descriptions of colour types, point out particular difficulties in scor- ing (1974, p. 125). For banding, they also find too much variation to use as yet more than presence or absence, but they illustrate pat- terns characteristic of particular localities. 9. Partula One of the most extensive breeding programmes in land snails is that of Clarke & Murray (1969, 1971; Murray 8 Clarke 1966, 1976a, b) on the Pacific islands genus Partula. The notation of the results contrasts with that produced by Cain and Sheppard for Cepaea, since only varietal names have been used, even when, as in P. taeniata, they have dissected the variations into component loci. There are three reasons for this. 9.1. Much of the variation falls into well- defined banding patterns not easily character- ized by a single descriptive word (unlike un- banded, mid-banded and five-banded in Cepaea). As these patterns are inherited as well-defined wholes, a simple varietal naming gives a practicable system. 9.2. The exact relationship between the component bands in different varieties is not easy to make out. While the presence or absence of a band just below the suture, another at the umbilical region, and some others in between them can be recognised with little difficulty from morph to morph, the exact number of the bands around the middle of the whorl is not easy to determine. This means that a simple numbering from above downward runs into uncertainty, and different authors might number the same lower bands differently. 9.3 Clarke 8 Murray rightly wished to main- tain continuity with the pioneering work of Crampton (1917, 1925, 1932), who gave in- valuable data on the distribution of species and of many of these distinct forms. This is an excellent example of the virtues of varietal names, which should not be lost sight of because of the excessive use of them in Cepaea and Liguus. A black body-whorl divided by a single white band near the mid- 10 CAIN dle is bisecta, a wholly black shell atra, a white body-whorl with two black bands, ap- proximately positioned so that they would frame the single white band of bisecta, is frenata. But it is not easy to say what the single broad black band of cestata corre- sponds to in the other forms. Professor J. J. Murray kindly tells me (letter of 19 July 1985) that they did consider a banding system, with 5 bands above the umbilical blot, but also discovered difficulties in homologizing the middle bands from morph to morph. As vari- etal names provide a good system of refer- ence and are not tendentious since they require no homologization of bands, they al- low ease and rapidity of reference while leav- ing the question of homology to be settled by further work. Obviously a banding homology is desirable (and Professor Murray remarks that on their system Crampton used the name zonata for 10305 and 1(234)5 as well as 0(234)0). The development of a numerical system and its comparison with that in helicid or helminthoglyptid snails can be considered elsewhere; here it is sufficient to point out the advantage of varietal names as labelling morphs with complex patterns, without impos- ing a theoretical structure of homology. 10. Theba pisana Several workers on this extremely variable snail have used only broad categories; e.g. Johnson (1980, 1981) used a classification into unbanded, effectively unbanded (with the upper bands missing) and fully banded, based on a similar classification used by Cain & Sheppard (1954) when considering varia- tion in Cepaea in relation to habitat. Heller (1981) used a somewhat more elaborate classification. The first genetic analyses were presented by Cowie (1984) and Cain (1984) and neces- sitated a far more elaborate symbolism since good segregants are found to be character- ized by highly particular banding formulae (Cain, 1984), e.g. (for the 3 upper bands) 00у, :3, пу, in which y indicates a yellow-buff band, not one with black pigmentation. Sacchi (1952) was the first to provide a symbolization for the extraordinarily complex patterns into which the black pigmentation of a band (when present) can be distributed, but on the basis of 4 bands, not 5, on the completely banded shell. This question is discussed by Cowie (1984). As a result of their observations and breed- ing experiments, both Cain and Cowie identify a thin line almost at the upper edge of the shell whorl as band 1, so that the banding in Theba is basically five-banded as in Cepaea and many other helicid snails. Sacchi (1952 and now, personal communication) does not recognise this line as a separate band, and the four bands he recognises, numbered from above as 1, 2, 3, 4 correspond on Cain's & Cowie's interpretation to (12)345 or 02345 in Cepaea. Heller (1981) also recognises only 4 bands. A four-banded phenotype is extremely common in the Common Snail Helix aspersa which is described as four-banded by Germain (1930). But comparison of its bands with those of other helicids shows that this phenotype corresponds to 1(23)45 in Cepaea and other helicids. Although all present work- ers agree in numbering the bands from above downwards, the numbers used by different workers are therefore not homologous. Num- bering is too convenient not to be used as a symbolism for repeated elements, recognis- able from one shell to another. Cain's sym- bolism is much simpler than Sacchi's, but no doubt further breeding will produce finer dis- crimination of forms, as happened with 00:45 in Cepaea (see above). A full description will need to use something as complex as Sac- chi's scheme if not more so. Theba pisana is particularly interesting be- cause although several morphs can be recognised, the breeding data prove that the expression of a particular allele may be some- what variable, and occasionally shells may be produced that are indistinguishable from forms with a different genetic basis. Such a blurring will account for the fact that it is often difficult to separate all the shells of a random sample into clearcut morphs, and the varia- tion appears to be continuous (as some of it undoubtedly is). This may be the type of variation found also in Littorina, and perhaps to some extent in Liguus. Its evolutionary significance is discussed briefly by Cain (1984). It must not be confused with the usually clearcut variation (except in minor banding varieties) found in Cepaea by de- scribing it with an inappropriate symbolism. 11. Helix aspersa The paper by Chevallier (1977) gives a general account of all variation in this species. Polymorphism is used simply to mean varia- tion, but variations in size, colour and band- SCORING MOLLUSCAN SHELL POLYMORPHISMS dl ing, shape, thickness and sculpture are stated to be morphs. The word morphotype appears to be used as a synonym for morph; some morphs are thought to be subspecies. New varietal names are bestowed in the traditional style; even the commonest form is newly named var. typica. On the other hand, the author utilizes the banding formula as em- ployed for Cepaea, and discusses the breed- ing experiments known to him, noting possi- ble cases of direct influence of the environment. Albuquerque de Matos (1985), however, distinguishes carefully between forms, varie- ties and morphs, recommends that the gen- eral usage of polymorphism for variation should be avoided, and gives a set of itali- cized symbols for the genetic variation in this species, which accord well with the sugges- tions in the present paper. 12. Cochlicella acuta An account of the polymorphism in this species, based on breeding experiments and random sampling, is given by Lewis (1975, 1977). The same banding formulae as in Cepaea are shown to be applicable. In addi- tion the states CO (continuous ostracum— opaque, usually white, shell) and DO (discon- tinuous ostracum—opaque shell interrupted by transparent glassy areas) in Cochlicella are delimited for the first time. Ground colours of the shell (amber and pale amber) have also been bred out as morphs, but are often inde- terminable on particular banding and CO forms. In 1975, Lewis, like Palmer (1984) in Thais, gives a notation for phenotypes, but indicates the genetic basis only verbally; in his 1977 paper he gives a properly italicized notation for the supergenes, but leaves his gene nomenclature in roman (p. 426). Else- where in this paper (e.g., p. 449) the roman/ italics convention is used fully. Lewis (1975) places an asterisk in his tables when the character is not determinable—a useful con- vention. DISCUSSION Variation has been a subject of close study ever since the publication of the Origin of Species (before which time it was usually thought to have no bearing on the nature of species), and, with the rise of genetics, the nature of different types of variation has been clarified considerably. One of the more re- markable differences in type of variation is that between continuous variation of the phe- notype, so very common in nature and produceable both genetically and by environ- mental influence, and discontinuous variation. Where this latter refers only to a few very rare mutants, it is merely a necessary conse- quence of genetic mutation. Where, as in polymorphism as defined by Ford (1940), it involves the maintenance of high proportions of different clearcut phenotypes in their pop- ulations, it is clearly of great evolutionary interest. It may serve many different func- tions, as for example in apostatic polymorph- ism (e.g. Clarke, 1964), in polymorphic mim- icry (e.g. Clarke 8 Sheppard 1960a, b) or as a source of genetic variation in the most familiar sort, sex. lt may be genetic or not (see references and discussion in Cain, 1977). Any work on variation, therefore, should distinguish clearly continuous and discon- tinuous modes, in genetically controlled vari- ation. For discontinuous variation, Ford's definition of polymorphism and Huxley's of morph, provide a simple terminology, unfortu- nately often grossly misapplied. (In French, polymorphisme more often than not still means no more than variation, and morph in American (and other languages!) is usually used for anything whatever that someone wishes to distinguish.) As some critics of an earlier draft of this paper have found difficulty with Ford's defini- tion (1940) of polymorphism, or have felt that it has been superseded, a brief examination of it is necessary. He divides genetic variabil- ity (p. 493) into four types, “(1) disadvanta- geous varieties eliminated by selection and maintained at a low level by recurrent muta- tion of the genes controlling them; (2) varia- tions due to the effects of genes approxi- mately neutral as regards survival value; (3) those dependent upon genes maintained by a balance of selective agencies; and (4) advantageous varieties controlled by genes spreading through the population and displac- ing their allelomorphs”. He points out explic- itly that “The third and fourth types constitute polymorphism. Here two or more well-marked forms, capable of appearing among the off- spring of a single female, occur with frequen- cies high enough to exclude the maintenance of the rarest of them by recurrent mutation”. The expression “genetic variability” in the first sentence quoted meant in 1940 (when the exact basis of not a single polymorphism was 12 CAIN known) variability in the phenotype believed on good evidence to be mediated genetically, not phenotypically. While the word ‘genetic’ could be inserted with advantage before ‘polymorphism’ in the last sentence but one quoted above, the meaning of the sentence is clear, since only genetic variability is being discussed. Those who believe they have good exam- ples of neutral polymorphisms will presum- ably bring Ford's type (2) variation also under the definition of polymorphism, adding some qualification as to the frequency of the rarest form being too high to be due to immediate mutation. Treatments subsequent to Ford's vary somewhat. For example, Mayr, Linsley 4 Usinger (1953, p. 96) merely equate discon- tinuous variation within a population with polymorphism, though the heading of the paragraph makes it clear that genetic polymorphism is meant. No mention is made of frequency. Hartl (1980, р. 77) goes straight to the genetic locus. “А POLYMORPHIC LOCUS is a locus at which the most common allele has a frequency of less than .99. Conversely a MONOMORPHIC LOCUS is one that is not polymorphic. The cut off at .99 in the definition of polymorphism is arbitrary, but it serves to focus attention on those loci with common allelic variation . .. RARE ALLELES are alleles with frequencies of less than .005.. .” Later (р. 79) he explains “The definition of polymorphism is an attempt to focus on loci having alleles with frequencies too high to be explained solely by recurrent mutation.” | prefer Ford's treatment as emphasizing dis- continuity in the phenotype not mediated merely by recurrent mutation. Albuquerque de Matos (1985) has specially emphasized the distinction between phenotypic and geno- typic variation, and has gone so far as to propose populational pluralism (“pluralismo (genetico) populacional”) for the presence in a population of different alleles and one, or more generally several, loci. In symbolizing phenotypes, apart from the general use of roman letters, there seems to be considerable variation in practice. Demerec et al. (1968) merely recommend words or abbreviations of them, with the re- quirement that the abbreviations should never be of three italicized lower case letters (as for loci). The ‘Mouse News Letter’ rules recom- mend two-, three-, or four-letter abbreviations of the name of the gene locus concerned in capitals, the name itself being “chosen so as to convey as accurately as possible the char- acter by which the gene is usually rec- ognised”. Arabic numbers can be included but always following a letter. Cain and Shep- pard used both lower-case letters and capi- tals, together with arabic numerals for the banding formula, which can stand by them- selves when only it is in question. Their usage with regard to lower-case letters and capitals is not fully consistent. It would be simpler to elevate all letters to capitals, but a case could be made for retaining lower-case letters as symbols for qualifying words (adjectives etc.) and using capitals for substantives, or in compound symbols for using capitals for the initial letter only. In view of the complexity of variation in many molluscan shell patterns (e.g. in the prosobranch Clithon oualaniensis, see Grüneberg, 1976, 1978, 1979) it is thought better to make no recommendation on this point, and to wait until we Know better, by experience, what flexibility is needed. Phe- notype symbols can be printed on the line; there is no need to elevate their qualifying (adjectival) parts to superscripts. In the present state of molluscan genetics, there is no molecular evidence for the homol- ogy of loci, yet it seems unnecessary to believe that the shell colour locus which pro- duces virtually identical phenotypic effects with the same dominance relationships be- tween the effects, and the same linkage rela- tionships with other loci in the sibling species Cepaea nemoralis and C. hortensis is not constitutionally the same in both. Whether the red-brown and yellow-brown segregants in Helix aspersa, a species of a certainly very closely related genus, are genetically the same as the pink-shelled and yellow-shelled forms in Cepaea, but with their expression shifted towards brown, is more dubious (Cain, 1971). Nevertheless, the use by Albuquerque de Matos (1985) of C for the shell colour locus in Helix aspersa, the same symbol as used for that purpose by Cain and Sheppard in Cepaea, seems justified at present, in that it draws attention to the similarity of the gene expressions in these species; when we know that the loci are different (if they are) it will be time to replace or qualify the symbols. In the meanwhile, one should be very cautious in speaking of genetic homologies. Instability of the symbolization is highly undesirable, and should be avoided as far as possible. Devel- opment of it, e.g. the later distinction within the pink class P (of shell colour in Cepaea) of deep pink, medium pink, and pale pink, DP, SCORING MOLLUSCAN SHELL POLYMORPHISMS 13 MP, PP is inevitable, but as in this example should utilize the earlier symbols and build on them. It is unfortunate that even giving a symbol may be taken to imply homology where there is none, and should therefore be explicitly disclaimed when there is no inten- tion of asserting it. For example, a numbering of the bands from above downwards in helicid snails as 1, 2, 3, 4, 5 probably does corre- spond to a genuine homology, at least within the Helicinae, but its use for Partula cannot imply homology between this genus and the very distantly related Helicidae. Indeed, in Partula, since the exact nature of the banding (number of bands, relative positioning, modes of individual band variation) has not yet been worked out, names, e.g. subsutural, umbilical, are probably better than numbers. Palmer's suggestion that banding in Thais is by sup- pression of pigment, not by pigmentation of particular areas as in pulmonates points to a mere analogy between banding-patterns in some molluscs. The examples given in the present paper are discussed both to show the interest of a comparative study of polymorphisms, and to examine the characteristics of the nomencla- tures and other notations proposed so far. (Other examples of molluscan polymorphism are mentioned by Murray, 1975). From a consideration of them it is clear that (i) discontinuous and continuous variation should be distinguished; (ii) the nomenclature and symbolization for the morphs should be clearly separated from that for their genetic basis; (iii) a nomenclature, using varietal names, has advantages over a symbolization when complex patterns, inherited as units are to be referred to and the homologies of their com- ponents are uncertain; (iv) a symbolization, being analytical, has advantages over a nomenclature when it can be applied with certainty; (v) any nomenclature or symbolization should allow easy augmentation as further information becomes available. There can be no objection to the use of a notation such as that proposed by Roth & Bogan (1984) in place of the excessive vari- etal names bestowed on Liguus fasciatus. It does not distinguish between continuous and discontinuous variation, but neither did that of Cain and Sheppard for variation in Cepaea, which was therefore supplemented by explicit statements of what segregrates from what, especially since, as already described, the same phenotype may be produced both by polygenic variation and by segregation of alleles. When, as with Roth and Bogan’s, а notation is proposed for variation in general, it would be preferable to print it in a different type-face; perhaps the best plan would be to print it in ordinary roman, and transfer the phenotypic notation for known morphs to bold-face. But this would involve a consider- able departure from present practice, and in view of the conservatism of editors and the usual incompetence of proof-readers, is prob- ably impracticable. At least it should be pos- sible to restrict the word morph to Huxley’s very useful definition, and to use the word polymorphism only for variation composed of morphs. The principles given in this paper seem equally applicable to the scoring of pheno- typic polymorphisms in other organisms be- sides molluscs. ACKNOWLEDGEMENTS | am very grateful to R. J. Berry, A. E. Bogan, A. F. Brown, B. C. Clarke, L. M. Cook, R. H. Cowie, J. Heller, J. J. Murray, A. R. Palmer and B. Roth who criticised, often in great detail, an earlier draft. | also thank A. R. Palmer and B. Roth for advance notice of work in hand, and R. J. Berry and L. M. Cook for the loan of reprints and other documents. M. F. Lyon kindly gave permission to quote from the rules and guidelines for gene no- menclature in the Mouse News Letter. REFERENCES CITED ADAMS, L. E., 1896, The collectors manual of British land and freshwater shells. . . . Ed. 2. Leeds; Taylor Brothers. ATKINSON, W. & WARWICK, T., 1983, The role of selection in the colour polymorphism of Littorina rudis Maton and Littorina arcana Hannaford-Ellis (Prosobranchia:Littorinidae). Biological Journal of the Linnean Society of London, 20: 137-151. AVERS, C. J., 1984, Genetics. Ed. 2. Boston, U.S.A., Willard Grant Press (PWS Publishers). BERRY, R. J. & CAROTHERS, J. H., 1974, Visible variation in the dog-whelk, Nucella lapillus. Jour- nal of Zoology, London, 174: 123-148. CAIN, A. J., 1971, Undescribed polymorphisms in two British snails. Journal of Conchology, Lon- don, 26: 410-416. CAIN, A. J., 1977, The efficacy of natural selection in wild populations. In GOULDEN, C., (ed.), The 14 CAIN changing scenes in natural Sciences, PP. 111-133. Academy of Natural Sciences of Phil- adelphia, Special publication no. 12. CAIN, A. J., 1984, Genetics of some morphs in the land snail Theba pisana. Malacologia, 25: 381411. CAIN, А. J. & CURREY, J. D., 1963. Area effects in Cepaea. Philosophical Transactions of the Royal Society of London, ser. B, 246: 1-81. CAIN, А. J., KING, J. М. В. & SHEPPARD, P. M., 1960, New data on the genetics of polymorphism in the snail Cepaea nemoralis L. Genetics, Princeton, 45: 393—411. CAIN, A J. & SHEPPARD, P. M., 1954, Natural selection in Cepaea. Genetics, Princeton, 39: 89-116. CAIN, A. J. & SHEPPARD, P. M., 1957, Some breeding experiments with Cepaea nemoralis (L.). Journal of Genetics, 55: 195-199. CAIN, A. J., SHEPPARD, Р. М. & KING, J. М. B., 1968, Studies on Cepaea |. The genetics of some morphs and varieties of Cepaea nemoralis (L.). Philosophical Transactions of the Royal Society of London, ser. B, 253: 383-396. CHEVALLIER, H., 1977, La variabilité de l'escargot petit-gris Helix aspersa Muller. Bulletin du Muséum national d'Histoire naturelle, série 3 no. 448, Zoologie 311: 425—442. CLARKE, B. C., 1964, Frequency-dependent se- lection for the dominance of rare polymorphic genes. Evolution, Lancaster, Pa., 18: 364-369. CLARKE, В. [С.] 8 MURRAY, J. [J.]., 1969, Ecolog- ical genetics and speciation in land snails of the genus Partula. Biological Journal of the Linnean Society of London, 1: 31-42. CLARKE, В. [C.] 8 MURRAY, J. [4.]., 1971, Polymorphism in a Polynesian land snail Partula suturalis vexillum. In CREED, E. R., ed., 1971. Ecological genetics and evolution. Essays in honour of E. B. Ford, pp. 51-64. Oxford, Blackwell Scientific Publications. CLARKE, В. С. & O'DONALD, P., 1964, Fre- quency-dependent selection. Heredity, London, 19: 201-206. CLARKE, C. A. & SHEPPARD, P. M., 1960a, The evolution of dominance under disruptive selec- tion. Heredity, London, 14: 73-87. CLARKE, C. A. & SHEPPARD, P. M., 1960b, The evolution of mimicry in the butterfly Papilio dardanus. Heredity, London, 14: 175-185. COLTON, H. S., 1922, Variations in the dog-whelk Thais (Purpura) lapillus. Ecology, 3: 146-157. COOK, L. M., 1967, The genetics of Cepaea nemoralis. Heredity, London, 22: 397—410. COOK, L. M. & MURRAY, J. J., 1966, New infor- mation on the inheritance of polymorphic char- acters in Cepaea hortensis. Journal of Heredity, 57: 245-247. COWIE, В. H., 1984, Ecogenetics of Theba pisana (Pulmonata : Helicidae) at the northern edge of its range. Malacologia, 25: 361-380. CRAMPTON, Н. E., 1917 ["1916”], Studies on the variation, distribution, and evolution of the genus Partula. The species inhabiting Tahiti. Carnegie Institution of Washington, Publication 228. CRAMPTON, H. E., 1925, Studies on the variation, distribution and evolution of the genus Partula. The species inhabiting the Mariana Islands, Guam and Saipan. Carnegie Institution of Wash- ington Publication 228a. CRAMPTON, H. E., 1932, Studies on the variation, distribution and evolution of the genus Partula. The species inhabiting Moorea. Carnegie Institu- tion of Washington Publication 410. ОЕМЕВЕС, M., 1958, Proposal for a uniform no- menclature in bacterial genetics. Microbial Ge- netics Bulletin, 16: 38-41. ОЕМЕВЕС, M., ADELBERG, Е. A., CLARK, А. J. & HARTMAN, P. E., 1968, A proposal for a uniform nomenclature in bacterial genetics. Journal of General Microbiology, 50: 1-14. DIVER, C., 1932, Mollusca. Genetics. Proceedings of the VI International Congress on Genetics, Menastu, Wisconsin, 2: 236-239. DIVER, C., 1939, Aspects of the study of variation in snails. Journal of Conchology, London, 21: 91-141. FISHER, R. A. & FORD, E. B., 1947, The spread of a gene in natural conditions in a colony of the moth Panaxia dominula L. Heredity, London, 1: 143-174. FORD, E. B., 1940, Polymorphism, parallel varia- tion, and taxonomy. In HUXLEY, J. S., ed., The new systematics, pp. 493-513. Oxford: Oxford University Press. FORD, E. B., 1945, Polymorphism. Biological Re- views, 20: 72-88. FORD, E. B., 1955, A uniform notation for the human blood groups. Heredity 9: 135-142. GERMAIN, L., 1930, Faune de France 21. Mol- lusques terrestres et fluviatiles. Paris: Leche- valier. GOODHART, C. B., 1962, Variation in a colony of the snail Cepaea nemoralis (L.). Journal of ani- mal Ecology, 31: 207-237. GRUNEBERG, H., 1976, Population studies on a polymorphic prosobranch snail (Clithon (Pic- toneritina) oualaniensis Lesson). Philosophical Transactions of the royal Society of London, ser. B, 275: 385-437. GRUNEBERG, H., 1978, Micro-evolution in a poly- morphic prosobranch snail (Clithon oualaniensis (Lesson)). Proceedings of the Royal Society of London, ser. B, 200: 419-440. GRÜNEBERG, H., 1979, A search for causes of polymorphism in Clithon oualaniensis (Lesson) (Gastropoda; Prosobranchia). Proceedings of the Royal Society of London, ser. B, 203: 379-386. HARTL, D. L., 1980, Principles of population genet- ics. Sunderland, Mass.: Sinauer Associates Inc. HELLER, J., 1981, Visual versus climatic selection of shell banding in the landsnail Theba pisana in Israel. Journal of Zoology, London, 194: 85-101. SCORING MOLLUSCAN SHELL POLYMORPHISMS 15 HUXLEY, J. S., 1955, Morphism and evolution. Heredity, London, 9: 1-52. JOHNSON, M. S., 1980, Association of shell band- ing and habitat in a colony of the land snail Theba pisana. Heredity, London, 45: 7-14. JOHNSON, M. S., 1981, Effects of migration and habitat choice on shell banding frequencies in Theba pisana at a habitat boundary. Heredity, London, 47: 121-133. LEWIS, G., 1975, Shell polymorphism in the snail Cochlicella acuta (Muller) and some data on its genetics. Biological Journal of the Linnean Soci- ety of London, 7: 147-160. LEWIS, G., 1977, Polymorphism and selection in Cochlicella acuta. Philosophical Transactions of the Royal Society of London, ser. B, 276: 399-451. Е MARTENS, G. Von, 1832, Uber die Ordnung der Bänder an den Schalen mehrerer Land- schnecken. Nova Acta Physico-medica, 16: 177-216. (Abhandlungen der deutsche Akade- mie der Naturforscher. Nova Acta Leopoldina.) MATOS, В. М. ALBUQUERQUE DE, 1985, Variacao intraespecífica, morfos e sua determinacáo genética em Helix aspersa Müller 1774. Publicacöes ocasionais da Sociedade Ропи- guesa de Malacologia, по. 5: 15-30. MAYER, A. G., 1902, Some species of Partula from Tahiti. A study in variation. Memoirs of the Mu- seum of Comparative Zoology at Harvard, 26: 117-135. MAYBZESEINSEENTETGTRSTUSINGER, В: LE, 1953, Methods and principles of sytematic zool- ogy. New York etc.: McGraw-Hill. MURRAY, J. [J.], 1963, The inheritance of some characters in Cepaea hortensis and Cepaea nemoralis (Gastropoda). Genetics, Princeton, 48: 605-615. MURRAY, J. [J.], 1975, The genetics of the Mollusca. In KING, R. G., ed. Handbook of genetics, 3: 3-31. New York: Plenum. MURRAY, J. [J.] & CLARKE, В. [C.], 1966, The inheritance of polymorphic shell characters in Partula (Gastropoda). Genetics, 54: 1261-1277. MURRAY, J. [J.] & CLARKE, В. [C.], 1976a, Supergenes in polymorphic land snails. |. Partula taeniata. Heredity, 37: 253-269. MURRAY, J. [J.] & CLARKE, B. [C.], 1976b, Supergenes in polymorphic land snails. И. Partula suturalis. Heredity, 37: 271-282. PALMER, A. R., 1984, Species cohesiveness and genetic control of shell colour and form in Thais emarginata (Prosobranchia, Muricacea): prelim- inary results. Malacologia, 25: 477-491. PETTITT, C., 1973, A proposed new method of scoring the colour morphs of Littorina saxatilis (Olivi, 1792) (Gastropoda: Prosobranchia). Pro- ceedings of the Malacological Society of London, 40: 531-538. PILSBRY, H. A., 1912, A study of the variation and zoogeography of Liguus in Florida. Journal of the Academy of Natural Sciences of Philadelphia, 15: 429-471. PREZANT, R. S. & CHALERMWAT, K., 1984, Induction of color forms in Corbicula [meeting abstract]. American Malacological Bulletin, 2: 87. REIMCHEN, T., 1979, Substratum heterogeneity, crypsis, and color polymorphism in an intertidal snail (Littorina mariae). Canadian Journal of Zo- ology, 57: 1070-1085. ROTH, B., 1981, Shell color and banding variation in two coastal colonies of Monadenia fidelis (Gray) (Gastropoda: Pulmonata). Wasmann Journal of Biology, 38: 39-51. ROTH, B. & BOGAN, A. E., 1984, Shell color and banding parameters of the Liguus fasciatus phe- notype (Mollusca: Pulmonata). American Mala- cological Bulletin, 3: 1-10. SACCHI, C., 1952, Ricerche sulla variabilita geografica in popolazioni Italiane di Euparypha pisana Müll. (Stylommatophora Helicidae). An- nali del Museo civico di Storia naturale di Genova, 65: 211-258. TAYLOR, J. W., 1914, Monograph of the land and freshwater Mollusca of the British Isles, 3. Leeds, Taylor. WOLDA, H., 1969, Genetics of polymorphism in the land snail, Cepaea nemoralis. Genetica, 40: 475-502. Revised Ms. accepted 30 September 1986. MALACOLOGIA, 1988, 28(1-2): 17-27 THE INCIDENCE AND VARIETY OF LEHMANNIA VALENTIANA CONJOINED TWINS: RELATED BREEDING EXPERIMENTS (GASTROPODA, PULMONATA) Jeanine Mason! & Jonathan Copeland? ABSTRACT The major objective of this study was to record the incidence and types of terata occurring in Lehmannia valentiana, a terrestrial mollusk, and to determine if the occurrence was inheritable. A series of breeding experiments was done comparing various groups of L. valentiana, which seems to be a hybrid, and its proposed parents Limax maximus and Deroceras reticulatum. L. valentiana produces conjoined twins naturally, which are frequently viable, at a higher rate than previously recorded in terrestrial mollusks. When they are raised to maturity and mated, other conjoined twins will be included among their offspring. Self-fertilizing (isolated from birth) L. valentiana will sometimes produce offspring, including conjoined twins. The related limacid species, Limax maximus and Deroceras reticulatum, produced no conjoined twins during the study period. Conjoined twins can be tentatively identified on the basis of two close zygotes (doublets) in one egg capsule immediately after oviposition or even in the capsules contained in the laying animal's oviduct. The number of doublets and conjoined twins in a clutch of eggs can be increased by mating animals that were themselves close doublets. The occurrence of doublets can be reduced by mating paired singlets (animals originating as one zygote in a capsule) and/or maintaining a colony of animals which all originated as singlets. Key words: Lehmannia valentiana; Limax maximus; Deroceras reticulatum; conjoined twins; terata; malformations; self-fertilization; hybridization. INTRODUCTION Conjoined twinning in the Mollusca, as in most animals, has been uncommon. Since no references to twins or “double monstrosities” in Mollusca had been cited, Newman (1923) initially concluded that twinning did not occur due to the characteristic determinate cleav- age of the molluscan zygote. Refutation followed with reports of twinning in Serpuloides vermicularis (a sessile tubic- olous mollusc) (Hall 1925). Crabb 8 Crabb (1927), however, questioned whether the occurrence of more than one embryo in a single egg capsule of some pulmonates was true twinning. Newman (1923) said true twins must arise from a single cell. In a later work, Crabb (1931) found con- joined twin embryos in several fresh-water snails. He concluded, after extensive study and some unsuccessful experimentation, that the conjoined twins arose as separate ova that fused before cleavage or during cleavage up to the early blastula stages. He claimed that the occurrence of two or more ova per capsule was not hereditary. Bigus (1981) found an average incidence of 0.02% conjoined twins in all eggs collected from the pulmonate Physa acuta. When the egg capsules contained more than one ovum, the average rate was 2.78%. Concurring with Crabb (1931), Bigus suggested that con- joined twinning only occurred in egg capsules containing more than one ovum and that the trait was not inheritable. Experimentally, molluscan separate and/or conjoined twins have been produced by com- pression of a zygote (Guerrier, 1970). George (1958) obtained three conjoined twins by chemically treating 200 egg capsules contain- ing two or more zygotes and then centrifuging them. Whether experimentally induced or oc- curring naturally, none of the conjoined twins previously studied have reached the hatching stage. While we were studying the progeny of 30 Lehmannia valentiana (Férussac) that we had obtained from the egg capsules of "W297 №3020 Oakwood Grove Road, Pewaukee, WI 53072, U.S.A. “Department of Biology, Swarthmore College, Swarthmore, PA 19081, U.S.A. 18 MASON & COPELAND Limax maximus Linnaeus [L. valentiana may be a hybrid obtained by crossing L. maximus and Deroceras reticulatum Müller (Mason, 1985)], we observed a healthy, two-headed, one-tailed animal among the other newly- hatched slugs. Microscopic examination of unhatched capsules revealed other abnormal animals. By maintaining a laboratory-raised popula- tion of L. valentiana for several generations, we were able to obtain, rear, and mate a considerable number of viable conjoined twins and other anomalies, such as fused tentacles, supernumerary eyes, etc. The data collected included the rates of conjoined twinning in L. valentiana and its putative parents, L. maximus and D. reticu- latum, and the occurrence rate of two or more ova per egg capsule and whether this rate could be increased or decreased by selective breeding. We also catalogued the kinds of conjoined twins and other anomalies. MATERIAL AND METHODS D. reticulatum animals were collected lo- cally in Wisconsin. L. maximus animals were laboratory-raised from an original group of animals from New Jersey. Thirty founder L. valentiana were obtained from eggs collected from a L. maximus colony. We concluded that L. valentiana, identified by L. F. Chichester (personal communication, 1981), must be a hybrid. Individuals from a natural population of L. valentiana were collected in Lexington, Kentucky, by David Prior. All animals were maintained using meth- ods similar to Reingold & Gelperin (1980). In addition to the lab chow, they were given fresh lettuce and Gerber baby food green beans. L. valentiana fed only lab chow did not produce fertile eggs, but animals fed any two of these three foods did so. Tegosept M, a mold inhibitor, used by Reingold & Gelperin (1980), was not added to the food. All animals were maintained with an artifi- cial photoperiod consisting of long days (LD 16:8) in an attempt to maximize reproduction (Sokolove & McCrone, 1978). Eggs were collected from all groups at least once weekly during periods of reproduc- tion. The eggs were washed in a strainer under tap water and then placed in labeled petri dishes lined with filter paper cut to fit in the dish. The eggs were placed within a large central hole cut in the filter paper. This procedure facilitated viewing the embryos under the microscope. The paper was kept uniformly moist throughout the embryonic growth period. Newly hatched slugs were transferred to 12 cm diameter by 7 cm deep plastic dishes lined with filter paper. Micropore tape over holes in the cover provided ventilation while preventing escape. Abnormal and/or con- joined twins were separated from normal an- imals. The animals were transferred to larger cages as they matured. Percentages of eggs with conjoined twins and other defects in embryos which com- pleted embryonic development were deter- mined for the laboratory populations of L. maximus and D. reticulatum, the original group and three further generations of L. valentiana and the field-collected L. valentiana from Kentucky. The percentages of conjoined twins and other defective em- bryos were also determined for two L. valentiana we obtained by crossing L. maximus and D. reticulatum, and for a self-fertilizing animal. Three L. valentiana were individually isolated to determine if L. valentiana could self-fertilize and produce conjoined twins. Parthenogenesis, while un- known in pulmonates, is unlikely but not excluded (McCracken 4 Selander, 1980). Determination of self-fertilization and the production of conjoined twins could demon- strate the ability of a single hybrid to found a population containing the trait. These data are presented in Table 1. Several abnormally developed groups of L. valentiana were maintained, as explained be- low, and eggs from these animals were ex- amined to determine the percentage of con- joined twins and other defects in the embryos which completed embryonic development (Table 2). We wanted to determine if animals with specific abnormalities such as “fused tentacles” or “two-head, one-tail” were fertile and whether they would produce progeny with similar abnormalities. All matched anom- alies were kept together but isolated from other categories of animals from the time they hatched. The percentage of egg capsules with mul- tiple embryos was determined for the follow- ing groups: D. reticulatum, L. maximus, and the L. valentiana mixed colony, Kentucky colony, singlet colony, doublet colony, paired close doublets, paired singlets and one self- fertilizing animal (Table 3). No selection was utilized in the mixed col- LEHMANNIA VALENTIANA CONJOINED TWINS 19 onies. In the singlet colony, animals were obtained by examination of egg capsules after oviposition and selection of capsules that contained one ovum. These animals were raised as a colony but isolated from other categories of animals. Doublet colony animals were obtained by selecting egg capsules which contained two ova. These animals were then raised together in colonies, i.e. all animals originated as dou- blets. Four pairs of paired close doublets were obtained by microscopic examination of cap- sules after oviposition. Capsules containing two ova close enough together that it was impossible to measure the distance separat- ing them at x 100 magnification were desig- nated as “close doublets.” Just prior to hatch- ing, the capsules were opened manually with forceps to ensure that the animals would not hatch by themselves and mix with other hatching animals. These paired close dou- blets were then raised to sexual maturity completely isolated from other categories of animals. Three pairs of paired singlet animals were obtained by taking animals which had devel- oped singly in an egg capsule and pairing them, isolated from other animal categories, throughout their lives. The self-fertilizing animal was isolated from the time of hatching. RESULTS Morphology The many variations of conjoined twinning and other defects in L. valentiana are shown in Fig. 1. All animals shown with the characteristic mantle line of L. valentiana reached the hatching stage (approx. 20-22 days after oviposition). Numbers 48 and 49 are only two examples of forms of conjoining where the animals did not reach full maturity during a normal development period. Abnor- mal embryos arrested in development (Fig. 2) frequently remained responsive to tactile stimuli, such as tapping the capsule with a forceps, for a month after the hatching due date. Many would eventually die, but in some, growth continued at a slow rate until the animal reached hatch status. Death was determined by tissue opaqueness and/or cell disintegration. We found that tentacle and eye abnormal- ities (Fig. 1) were often associated with the twinning process. Opening capsules manu- ally with forceps revealed that many abnor- malities occurred in an animal that had developed as one of a doublet (two ova per egg capsule). To determine if there was consistency in this observation, we opened 150 egg capsules containing doublets, or a singlet and an amorphous mass, or a developed doublet with an arrested develop- ment sibling. In 17 cases, we found that, whereas one of the doublet pair might be normal, the other had various deformities, e.g. зирегпитегагу or missing eyes; missing, fused or additional tentacles; mantle deformi- ties and/or mouth deformities. Abnormal slugs found with amorphous tissue (numbers 13 and 14) in which eyes or other body parts were identifiable provided a clue that many of these anomalies might also be forms of conjoined twinning. Newman (1923) noted that additional appendages or organs could indicate an initial case of conjoined twinning. Tentacle and/or eye abnormalities (includ- ing the absence of eyes) were numerous both in combination with other morphological doubling and separately (numbers 1 to 6, 44). Although extra eyes under the mantle oc- curred occasionally (number 4), they were not always in the position denoted by the arrow and did not always seem to be located within the ocular tentacle. Seven was the maximum number of eyes on one tentacle (number 8). Number 22 had at least six eyes on the third ocular tentacle. Therefore, the maximum number of eyes observed per animal was eight (this could be higher because some eyes seem to be fused). Arrows at numbers 20, 24, and 33 illustrate what appeared to be cyclopia, or the fusion of one or more eyes. Fused tentacle slugs [Fig. 1 (No. 3), Fig. 3] were initially identified in offspring of the orig- inal colony. The defect was frequently found in doublet capsules in which one embryo had ceased development. In most occurrences, the conjoined twins or parts thereof are aligned anterior to anterior. Numbers 27-30 illustrate animals where there is some deviation from this alignment. Number 27 was the only animal observed that was fused with the heads orthogonal to the tails. The animal lived several months, func- tioned well and moved with no apparent diffi- culty (Fig. 7). The posterior vestige of an incorporated MASON & COPELAND rat DR u =" — —— >= N = = > DIT > 7 я ea = eg o ee ans Ам — Fn E = со < = ри en er ее: ES — A > => = = So $ = D II 55 5 CA ES > => © LEHMANNIA VALENTIANA CONJOINED TWINS 21 -——200um A is _—_- 10mm +4 : FIG. 2. Representative sample of unusual morphology in L. valentiana with an embryo showing arrested development (arrow indicates posterior sac). FIG. 3. Adult L. valentiana with fused tentacles. FIG. 4. Two-headed one-tailed conjoined twin. FIG. 5. Mantle hump, short tail adult animal. FIG. 6. Example of double ova representative of close doublet, some of which develop into conjoined twins. FIG. 7. Unusual conjoined twin oriented at 180° (Photos 3, 4 and 5 taken by J. Coggins.) 22 MASON & COPELAND TABLE 1. The incidence of conjoined twins and other defects at the time of hatching recorded in Limax maximus, Deroceras reticulatum, and various Lehmannia valentiana groups. Complete embryonic Conjoined Other Total Number times development* twins defects abnormalities Group eggs collected (no.) (%) (%) (%) Deroceras reticulatum 1 1562 0.0 0.1 0.1 Limax maximus 14 1791 0.0 0.7 0.7 Lehmannia valentiana (orig. group) 18 371 6.7 3.2 9.9 L. valentiana (1st generation) 50 4201 1.4 2.4 3.8 L. valentiana (2nd generation) 25 7198 94 0.7 3.8 L. valentiana (3rd generation) 4 633 17 0.3 PA L. valentiana (Kentucky-wild) 6 22 4.5 4.5 9.1 L. valentiana** (hybrid)(N = 2) 5 256 1.6 07 223 L. valentiana (isolate)*** (N = 1) 3 7 14.3 0.0 14.3 "Complete embryonic development was determined by the appearance of pigmentation and the disappearance of embryonic features, such as the pedal lobe. **These two animals were obtained from L. maximus eggs in a breeding experiment between L. maximus and D. reticulatum (Mason, 1985). ***This was the only one of three animals that were isolated from birth that produced any eggs. twin, a partially developed foot and body, is shown in number 45. The lateral views (numbers 43 and 47) show mantle deformities that may not be directly associated with twinning in all cases. An epidermal invagination, forming a pouch, separates the mantle and viscera from the foot (number 43). In number 47, the animals were abbreviated in the antero-posterior axis. The viscera and mantle were elongated dorsoventrally as if torsion had only partially occurred or had occurred at an abnormal angle. The abbreviated bodies (numbers 7 and 37) are both the result of conjoining as evi- denced by morphological duplication of parts. Fig. 1 is fairly thorough in the presentation of anomalies with these exceptions: 1. The numerous variations of dorsolateral joining are not shown (all antero-posterior joining variations are shown. 2. The morphology of animals that did not reach hatch status, other than numbers 48 and 49, were not recorded. 3. No animals with mouth deformities, a frequent but fatal occurrence, are shown. Capsule doubling Conjoined twinning has been attributed to fusion of two or more zygotes in the same capsule (Crabb, 1931; George, 1958, Bigus, 1981). The distance between two L. valenti- ana zygotes in the same quadrant of a cap- sule was measured using a light microscope micrometer. The average distance between the two zygotes (43 capsules) was 16.2 um, with a range of 0.0 um to 68.8 um at x 400 magnification. Fig. 6 is an example of the 0.0 um distance, where no separation is dis- cernible. If doubled, the two zygotes are fre- quently in the same quadrant of a capsule in both D. reticulatum and L. valentiana, but seldom in L. maximus. By segregating L. valentiana capsules con- taining close doublets (0.0 um separation), we determined that conjoined twinning did not occur unless the zygotes were extremely close, a finding substantiated by others (George, 1958; Crabb, 1931). When we dis- sected slugs while they were laying eggs, we found the closeness of the double zygotes was present throughout the unlaid capsules LEHMANNIA VALENTIANA CONJOINED TWINS 23 TABLE 2. The incidence of conjoined twins and other defects at the time of hatching recorded for various L. valentiana abnormal groups. Complete embryonic Conjoined Other Total Number times development” twins defects abnormalities Group eggs collected (no.) (%) (%) (%) L. valentiana fused tentacles (N = 6) (Fig. 3) 26 577 4.8 2.1 6.9 Two-head, one-tail (N = 4) (Fig. 4) 10 118 ПЕЙ 2.5 4.2 Mantle hump-short tail (N = 4) (Fig. 5) 13 381 0.3 0.5 0.8 Eye/tentacle defect (N = 3) 15 239 0.4 1%) st Paired close doublets** (N = 8) 27 959 6.1 Sal 9.2 “Complete embryonic development was determined by the appearance of pigmentation and the disappearance of embryonic features, such as the pedal lobe. **Paired close doublets were animals obtained by raising animals which had developed as two animals in one egg capsule. Closeness was determined by whether there was a measurable distance between the fertilized ova prior to first cleavage. If the doublets were “close,” the distance was not measureable at х 100 magnification. Those close doublets which resulted in conjoined twins were not used in this breeding experiment. enclosed in the oviduct. Fusion or whatever mechanism was responsible for conjoined twinning was occurring prior to cleavage and might be occurring prior to oviposition. Conjoined twinning in L. valentiana is not the result of incomplete fission of the first or subsequent cleavage stages. This was deter- mined by segregating all capsules containing double or multiple zygotes from singlets at the time of oviposition. No conjoined twins were recovered from the singlet capsules. The incidence of conjoined twinning and other anomalies was recorded for L. maximus, D. reticulatum and various L. valentiana groups (Table 1). No conjoined twins were observed for L. maximus and D. reticulatum. The “other defects” were generally mouth de- formities, i.e. extrusion о the buccal cavity. No duplication of parts was observed. The L. valentiana original colony’s percentage of con- joined twins is higher than that of the following generations, but similar to the Kentucky pop- ulation, and the paired close doublets. Of the three animals isolated from birth, only one produced eggs. This animal laid 22 eggs when it was 8.3 months old (a non- isolated or colony animal frequently lays 100 or more eggs at a time beginning at 4.5 months of age). Of those 22 eggs, 6 produced normal animals and one a conjoined twin, while the capsules containing multiple ova did not develop. The late onset of egglaying and the high percentage (63.6%) of multiple ova in the capsules (Table 3) may represent a “last ditch” effort to reproduce. One of the three animals lived 15.5 months without laying eggs. That is the longest a L. valentiana has lived in our laboratory. The results of mating several groups of animals possessing similar abnormalities (Fig. 3, 4, and 5) are recorded in Table 2. Although conjoined twins and/or other defects occurred in all groups, the morphology of the offspring did not match that of the parents, i.e. neither of the two conjoined twins produced by the “two-head, one-tail” parents was sim- ilarly joined. Considerable disparity exists in the percent- age of conjoined twins produced by the five groups of abnormal parents. The “mantle hump-short tail” and “eye/tentacle defect’ groups’ percentage of conjoined twins was less than a third that of any of the other groups. The “paired close doublets” produced the most conjoined twins, but not higher than the original L. valentiana colony (Table 1). The total number of doublets and multiple egg capsules in various groups was tested using Chi-square contingency tables at the Роэ significance level (Table 3). No signifi- cant difference in the proportion of total dou- blets and multiples was found between the D. 24 MASON & COPELAND TABLE 3. The incidence of egg capsules containing more than one ovum recorded in Deroceras reticulatum, Limax maximus and other Lehmannia valentiana groups. Singlet = animal originated in capsule containing one ovum. Doublet = animal originated in capsule containing two ova. Number times Group eggs collected Deroceras reticulatum Limax maximus 16 Lehmannia valentiana (mixed lab colony)* 26 L. valentiana (Kentucky-wild) 6 L. valentiana (singlet colony)** 7 L. valentiana (doublet colony)*** 7 L. valentiana (paired close doublets) (4 pair, N = 8) 27 L. valentiana (paired singlet)***** (3 pair, N = 6) 5 L. valentiana (selfing [singlet] individual) (N = 1) 3 ek “Colony consisted of both singlets and doublets. Capsules Doublets Multi Total (no.) (%) (%) (%) 1142 6.9 1.0 7.9 1716 5.1 2.2 723 5330 9.6 ES 10.9 139 10.1 0.7 10.8 1565 4.0 08 4.3 1608 9.8 153 10.4 2593 24.5 213 26.8 287 3.8 1 4.9 22 9.1 63.6 WET, **Colony consisted only of animals which had originated as single ova in one capsule. ***Colony consisted only of animals which had originated as double ova in one capsule. ***Two animals which had originated as double ova in one capsule were paired with each other and isolated from other animals. Four pair of close doublets were used. ss animals. Three pair of singlets were used. reticulatum and L. maximus colonies (df — 1), between the L. valentiana singlet colony and paired singlets (df — 1), or be- tween the L. valentiana mixed, Kentucky, and doublet colonies (df — 2). A significant differ- ence was found between the D. reticulatum, L. maximus and L. valentiana mixed colonies (df — 2), the L. valentiana paired singlets and the mixed colony (df — 1), and the L. valentiana paired close doublets and mixed colony (df = 1). If the L. valentiana paired close doublets are excluded, the proportion of multiple (more than two per capsule) ova was not signifi- cantly different in the other L. valentiana groups (df — 4). The variation between these groups seems to be dependent on the num- ber of doublets produced. L. valentiana can self-fertilize (Table 1), but not consistently. The rate of doubling in the selfing individual is similar to that in the mixed colony, i.e. 9.1%. The percentage of multiple Two animals which had originated as single ova in capsules were paired with each other and isolated from other ova capsules is 63.6%, additional evidence of the independence of doubling and multiples. The selfing animal was not statistically tested with the other groups due to the small sample number. DISCUSSION Duplication of parts Eye and tentacle defects without evidence of other morphological duplication do not ini- tially imply conjoined twinning as a causal agent. Extra eyes, extra heads and other duplications have been chemically induced in insect embryos (Walton et al., 1983). Separa- tion of molluscan embryos after first or sec- ond cleavages can result in the absence of eyes (and tentacles) in one of the halves or both, or both halves may each have two eyes (Cather et al, 1976). The anterior end of an LEHMANNIA VALENTIANA CONJOINED TWINS 25 animal is the most susceptible to agents that inhibit development (Newman, 1917). The frequent occurrence of supernumerary or absent eyes in L. valentiana animals sug- gests that two initially conjoined zygotes may separate after the eye anlage has differenti- ated. This event might occur after the first quartette of micromeres is produced as these cells give rise to the cerebral ganglia, cephalic eyes and tentacle (Verdonk, 1979). This event, however, would seem to account only for the presence of four eyes, i.e. two from each zygote. Tweedell (1953) noted that dis- symmetries might arise if a structure is formed from a fraction of the total mass. Such a fraction might be embryonic cells detached from one twin and incorporated into the other. If random fusion were the causal mecha- nism, one would expect conjoined twins to resemble those shown in numbers 28, 30 or 31 of Fig. 1, where the twins are equal and complete. However, Hess (1971) claims that in experimentally produced gastropod single- egg twins, the material originally included in only one egg is capable of developing three or four tentacles or eyes instead of the normal single pair. The presence of eight eyes on animals 8 and 22 (Fig. 1) parallel Hess’ observation. Although L. valentiana conjoined twins could arise by fusion of two zygotes, an incomplete and unequal fission of a fertilized ovum is not completely ruled out. However, in the latter case, if Hess is correct, animals with eight eyes are theoretically unlikely. Body alignment The frequency of anterior-to-anterior align- ment seems unusual in L. valentiana con- joined twins if orientation occurred randomly. Animal-vegetative polarity is established dur- ing oogenesis (Verdonk, 1979). The animal pole protrudes into the lumen of the gonad in spiralian oogenesis (Huebner & Anderson, 1976). Raven (1967) argued that even the symmetry and dorsoventrality of the future embryo is imprinted on the egg cortex by the surrounding gonadal follicle cells. We posit several mechanisms to account for the preferential anterior-to-anterior con- joining observed: (1), the oocytes are fused in the gonad; (2), anterior-posterior fusion re- duces the viability of the embryo, ¡.e. death occurs before we could determine alignment; (3), fusion induces a polarity change; (4), conjoined twins do not arise as a result of fusion, but by some other mechanism. Fusion of individual cells, whether germ or somatic, is not easily achieved despite the normal occurrence of early embryonic junc- tions between blastomeres. In separated blastomeres, “... very tight coupling resumes only if the cells are brought back together quickly and in the original orientation” (Pow- ers & Tupper, 1977). Most centrifuge experi- ments aimed at fusing close zygotes or oocytes yield less than satisfactory results (including our own). N. H. Verdonk (personal communication, 1982) noted that whereas multiple ova egg capsules were found in nearly all mollusk groups, “Spontaneous fusion of eggs or em- bryos is very exceptional even when many eggs are stored in the same cap- sules. The reason is that most eggs are surrounded by a vitelline-membrane and as soon as the embryo comes out of this membrane it starts turning around.” Verdonk suggested that the mechanism un- derlying the relatively high rate of “germ fu- sions” in P. acuta (Bigus, 1981) is a missing or defective vitelline membrane, which would allow fusion to occur. The Lymnaea stagnalis embryo leaves the vitelline membrane approximately 32 hr. after the first cleavage and begins turning (Verdonk, personal communication, 1983). L. valentiana embryos are turning around at the time the first polar body is extruded (usually within one hour after oviposition). Rotation is easily observed by watching the polar body seem to appear and disappear. If the L. valentiana zygote turns around after leaving the vitelline membrane, as Verdonk observed in L. stagnalis, then L. valentiana zygotes have left the vitelline membrane prior to ex- trusion of the first polar body, or the vitelline membrane may be absent or defective. The actual number of L. valentiana con- joined twins may be under-represented in Table 1 since a number of the terata listed under “other defects” may originate as dou- blets or conjoined twins. Therefore, especially in the early data, a problem of “fuzzy-sets” (Root-Bernstein, 1983) exists between the L. valentiana categories “conjoined twins” and “other defects.” The number of conjoined twins and conse- quently total abnormalities is higher for the original and Kentucky colonies and the paired close doublets than in the other groups. Since the paired close doublets were selected with 26 MASON & COPELAND the intent of increasing the doubling and consequently the conjoined twinning, this higher rate was anticipated. However, other explanations must account for the higher rate in the original and Kentucky colonies. In the original L. valentiana colony, con- joined twinning data were not taken for the first four months of egg laying since we did not know it was occurring. The Kentucky colony slugs oviposited in late autumn and died soon thereafter. Data, in both cases, were, therefore, from the later stage of fertil- ity. If one does not have data from the first two-thirds of fertility, the last third may appear inflated. Bigus (1981) noted that both dou- bling and zygote fusion in P. acuta occurred only during the last third of fertility. This is not true in L. valentiana since many early clutches contain both conjoined twins and doublets. The original L. valentiana colony differs from the Kentucky colony in that whereas the total abnormalities produced is similar, the original colony produced more conjoined twins while the Kentucky colony produced more “other defects.” Since the original colony was a first gener- ation hybrid, the trait may have been attenu- ated in subsequent generations. However, the two L. valentiana later obtained by hybrid- ization did not produce a higher rate of con- joined twins. Minimally, the L. valentiana conjoined twin- ning rate was greater than 1.0% in all eggs collected with the exception of the “mantle hump-short tail” and “eye/tentacle defect” matings (Table 2). This 1.0 % rate is 50 times greater than the rate Bigus (1981) observed in P. acuta and the other mollusks cited. More importantly, L. valentiana data only include animals that developed to the hatch stage and were viable. We conclude that the initial rate of conjoined twinning is even higher. None of the conjoined twins observed by Bigus or others reached hatch status. The production of conjoined twins by an isolated L. valentiana animal demonstrates the possibility of a founder animal producing a population containing the trait, an important consideration in a hybrid animal. The low rate of abnormalities, especially conjoined twins produced by the “mantle hump-short tail” and “eye/tentacle defect” pairings is difficult to explain. Since most “eye/tentacle defect” animals originated as doublets, one would expect the trait to be expressed with a frequency equivalent to that of the other colonies. These pairings, how- ever, do demonstrate the fertility of animals with several types of abnormalities. The results of selective breeding of various L. valentiana groups indicates that the occur- rence of doublets in egg capsules is inherita- ble (Table 3). This finding contradicts both Crabb (1931) and Bigus (1981). Using the L. valentiana mixed laboratory colony as stan- dard, the rate of doubling is 9.6 %. The singlet colony and paired singlets rate of doubling is less than half that rate (4.0 and 3.8 % respec- tively). The colony consisting only of animals which originated as doublets had a rate of doubling consistent with that of a mixed col- ony, possibly an indication that mating among these animals was random, ¡.e. twins did not mate with their capsule siblings. However, one would expect the doublet colony animals to produce more doublets than the unselected colonies if there were complete penetrance of the trait. The ability to inherit the doubling trait is also demonstrated by the results of the paired close doublet matings, ¡.e., 24.5 % of their offspring were also doublets. The difference in doubling rate between the paired close doublets (24.5%) and the doublet colony (9.8 %) may be attributable to the selection process. Colony doublets were obtained from capsules containing two ova but not neces- sarily closely apposed ova as was the case in the paired close doublets. More than one mechanism may exist for doubling, one based on an anatomical defect such as described by C. P. Raven (personal communication, 1981) and one unknown. The cause of doubling and/or conjoined twinning may be inherent in the zygote or a result of the reproductive environment. As a hybrid, L. valentiana may be exhibiting a mixture of the developmental pathways of the putative parental species D. reticulatum and L. maximus, neither of which produced conjoined twins. Rachootin & Thomson (1981) noted that *. .. a mixture of related but distinct developmental pathways might pro- duce adaptively interesting novelties, which on occasion are assimilated.” LITERATURE CITED BIGUS, L. von, 1981, Polyvitellinity and fusion of germs in Physa acuta (Pulmonata, Basom- matophora). Zoologische Jahrbucher; Abteilung LEHMANNIA VALENTIANA CONJOINED TWINS 27 für Anatomie und Отодете der Tiere, 105: 526-550. CATHER, J. N., VERDONK, N. H. & RENE DOH- MEN, M., 1976, Role of the vegetal body in the regulation of development in Bithynia tentaculata (Prosobranchia, Gastropoda). American Zoolo- gist, 16: 455-468. CRABB, E. D., 1931, The origin of independent and of conjoined twins in fresh-water snails. Wilhelm Roux Archiv für Entwicklungsmechanik der Organismen, 24: 332-356. CRABB, E. D. & CRABB, R. M., 1927, Polyvitelliny in pond snails. Biological Bulletin, 53: 318-327. GEORGE, J.D., 1958, Experimental fusion of em- bryos of Limnaea stagnalis L. Proceedings Akademie van Wetenschappen, Amsterdam, Biological and Medical Sciences, Afdeling Natuurkunde, 61: 595-597. GUERRIER, P., 1970, Les caracteres de la ségmentation et la détermination de la polarité dorsoventrale dans le développement de quelques Spiralia. Ш. Pholas dactylus et Spisula subtruncata (Mollusques, Lamellibranches). Journal of Embryology and Experimental Mor- phology, 23: 667-692. HALL, R. P., 1925, Twinning in a mollusc, Ser- puloides vermicularis. Science, 61: 658. HESS, О., 1971, Fresh water Gastropoda. In: REVERBERI, G., ed., Experimental embryology of marine and fresh-water invertebrates, pp. 215-247. American Elsevier, New York. HUEBNER, E. & ANDERSON, E., 1976, Compar- ative spiralian oogenesis—structural aspects: an overview. American Zoologist, 16: 315-343. MASON, J., 1985, The taxonomic status of Lehmannia valentiana. In: Conjoined twinning in Lehmannia valentiana (Gastropoda, Pulmonata), pp. 140-168. Ph.D dissertation, University of Wisconsin, Milwaukee, WI. McCRACKEN, С. Е. & SELANDER, В. K., 1980, Self-fertilization and monogenic strains in natural populations of terrestrial slugs. Proceedings of the National Academy of Science USA, 77: 684-688. NEWMAN, H. H., 1917, On the production of mon- sters by hybridization. Biological Bulletin, 32: 306-320. NEWMAN, Н. H., 1923, The physiology of twinning. University of Chicago Press, Chicago. POWERS, В. D. & TUPPER, J. T., 1977, Intercel- lular Communication in the early embryo. In DE MELLO, W.D., ed., Intercellular communica- tion, pp. 231-251. Plenum, New York. RACHOOTIN, $. P. & THOMSON, К. S., 1981, Epigenetics, paleontology, and evolution. In: SCUDDER, G. G. E. & REVEAL, J. L., eds., Evolution today, Proceedings of the Second In- ternational Congress of Systematic and Evolu- tionary Biology, pp. 181-193. RAVEN, C. P., 1967, The distribution of special cytoplasmic differentiations of the egg during early cleavage in Limnaea stagnalis. Develop- mental Biology, 16: 407-437. REINGOLD, S. C. & GELPERIN, A., 1980, Feeding motor programme in Limax. Il. Modulation by sensory inputs in intact animals and isolated central nervous systems. Journal of Experimen- tal Biology, 85: 1-19. ROOT-BERNSTEIN, R. S., 1982, Mendel and methodology. History of Science, 21: 275-295. SOKOLOVE, P. G. & MCCRONE, Е. J., 1978, Reproductive maturation in the slug, Limax maximus, and the effects of artificial photoperiod. Journal of Comparative Physiology A, 125: 317-325. TWEEDELL, K. S., 1953, Identical twinning and the information content of zygotes. In: QUASTLER, H., ed., Essays on the information theory in biology, pp. 215-250. University Press, Urbana, IL VERDONK, М. H., 1979, Il Symmetry and asymme- try in the embryonic development of molluscs. In: VAN DER SPOEL, S., VAN BRUGGEN, А. D., LEVER, J. eds., Pathways in malacology, pp. 25-45. Junk, Utrecht, The Надие. WALTON, В. T., O'NEILL, Е. G., KAO, G. L., 1983, Benzoquinolinediones: activity as insect tera- togens. Science, 222: 422-423. Revised Ms. accepted 24 November 1986 MALACOLOGIA, 1988, 28(1-2): 29-39 THE FEEDING OF TERRESTRIAL SLUGS IN RELATION TO FOOD CHARACTERISTICS, STARVATION, MATURATION AND LIFE HISTORY C. David Rollo Department of Biology, McMaster University 1280 Main St. W., Hamilton, Ontario, Canada, L8S 4K1 ABSTRACT Food consumption by adult Deroceras reticulatum was measured as dry weight, wet weight and volume of food eaten. Consumption varied least among diets when wet weights were compared. Slugs ingested one to four meals daily. The number, duration and size of meals and the interval between them varied widely among individuals, with the type of food and with the duration of starvation. Starved adults showed no compensation following starvation when either consumption or growth was considered, although the number and frequency of meals increased with increasing deprivation. Similar results were found for the much larger, longer-lived species Limax maximus, suggesting that differences in life history tactics were not involved. Immature D. reticulatum, however, showed strong compensatory growth (and presumably feeding) following starvation. Degrowth was a key response to starvation which may explain why gastropods do not accumulate appreciable reserves of lipids. Slugs apparently are capable of long-term regulation during the growth phase of their life cycle. During reproduction, however, adults lost weight even when fed, and compensation was lacking. Key words: slugs; Deroceras reticulatum; Limax maximus; feeding regulation; starvation. INTRODUCTION Terrestrial molluscs are major consumers and decomposers in natural, agricultural and horticultural communities (Godan, 1983). Con- sequently, there is considerable literature con- cerned with their food consumption (see Rollo, 1987). Understanding feeding is particularly important since control is mainly by poisoned baits (Wright 8 Williams, 1980). Senseman (1978) and Reingold & Gelperin (1980) exam- ined control of ingestion for particular meals, but there is almost nothing known about the daily frequency and duration of meals as in- fluenced by food characteristics. Many invertebrates respond to starvation or malnutrition by compensatory mechanisms such as increased feeding rates (Waldbauer, 1968; Gelperin, 1971; Barton-Browne, 1975; Slansky & Scriber, 1985). Some gastropods, however, may lack such compensatory abili- ties (Susswein & Kupfermann, 1975a, 1975b; Senseman, 1977). Food quantity is rarely limiting for general herbivores such as snails or slugs, but unfavourable weather may re- strict foraging (Richter, 1976; Rollo, 1982), or require aestivation (Schmidt-Nielsen et al., 1971; Jaremovic & Rollo, 1979; Rollo, 1982). Whether molluscs compensate to offset such (29) setbacks has important implications for their growth, reproduction and population dynam- ics. Calow (1975a, 1975b) found post-star- vation compensation in aquatic snails, but terrestrial slugs have not been studied in this regard. The present study characterized the daily feeding pattern of the terrestrial slugs Deroceras reticulatum (Muller) and Limax maximus L. on various diets. Post-starvation responses were also examined to determine if long-term regulation occurred. METHODS Daily feeding pattern and the starvation response Adult D. reticulatum were collected from fields in Hamilton, Ontario in October and November (part of their normal reproductive period). Slugs were housed individually in glass jars 6 cm deep and 5 cm in diameter. Each jar contained 2 cm of moistened vermic- ulite which maintained high humidity but was never eaten. To ensure entrainment of the animals’ circadian rhythms (Rollo, 1982), slugs were housed in environmental cham- 30 ROLLO bers with a light:dark cycle of 16:8h, and a temperature of 18°C for two weeks prior to the experiments. During entrainment the slugs were fed lettuce. Jars were cleaned and the vermiculite was replaced weekly. Slugs were weighed and placed in clean jars prior to experiments. Most faecal strings were depos- ited during the light period, and these were removed prior to the dark period to prevent coprophagy. Adults were assigned so that their mean weight was similar among treat- ments (see Table 1). For calculating dry weight, all animals were assumed to be 89% water, based on a sample of 10 individuals. Cucumber (Cucumis sativus L. var. angli- cus) was mainly used to examine the influ- ence of starvation on feeding. Ingestion is also influenced by physical characteristics of food such as hardness (Senseman, 1978). Therefore a harder food, carrot root (Daucus carota L.), was also studied. Both foods were highly palatable to D. reticulatum. There is little standardization in the way that consump- tion rates or food characteristics are mea- sured which makes comparative studies diffi- cult. Therefore, wet weight, dry weight and volume eaten were all evaluated. Feeding by D. reticulatum on cucumber was monitored following starvation for 16, 40, 100 and 220 h. Similar observations were made when carrot was fed to slugs starved 16 or 100 В. For each experiment, 20 1 cm? cubes of carrot or cucumber were dried at 60°C to constant weight. This sample provided esti- mates of hydration and dry weight-to-volume relationships. Each slug was given a pre- weighed cube of appropriate food just before the dark period, and the behaviour of each individual was recorded at 0.5 h intervals until the next light period. Whenever a slug com- pleted a meal, the remaining food was re- moved and a new pre-weighed food cube was supplied. The food remaining following a meal was dried to constant weight at 60°C. Consumption was calculated by subtracting the dry weight of the remainder from the estimated dry weight of the original food. The volume and wet weight eaten were calculated by multiply- ing the dry weight ingested by the appropriate conversion factors. Slugs that oviposited were excluded from the analysis because this activity had a longer duration and higher priority than feeding. Slugs ate distinct meals separated by several hours. Consequently a 0.5 п observation interval was sufficient to distinguish meals. Snails may reduce their metabolism or aestivate during starvation or dehydration (Heeg, 1977). If slugs respond similarly, they could require time to become fully active and so their initial feeding could be relatively low. Consequently, feeding of slugs starved 16 or 100 h was observed for two consecutive days after being fed cucumber. Life cycle considerations Initial experiments did not detect compen- satory feeding following starvation. Conse- quently, several additional hypotheses were tested. Other preliminary experiments sug- gested that starved adult D. reticulatum con- tinued ovipositing which suggested that they were irreversibly committed to reproduction. Most adults lost weight even when fed, sug- gesting that regulatory mechanisms such as compensatory feeding may be absent and so- matic support reduced. Compensatory mech- anisms might still occur in juveniles, but it was difficult to accurately measure ingestion by small slugs. If compensation occurs, however, it should be reflected in growth rate. Young D. reticulatum (mean wet weight of only 66 mg), were collected from burdock plants (Arctium minus (Hill) Bernh.) in June. The animals were starved for 2 days to evacuate their guts, and then weighed. Eighteen slugs were fed fresh burdock leaves at 18°C, 100% R.H. and with a light:dark cycle of 16:8 h. Another 18 slugs were starved for 36 days in identical conditions (to 50% mortality) and then fed burdock. An- imals in both treatments were weighed every 3to 4 days and slugs that died were omitted from the analysis. The growth rates of fed and starved adult D. reticulatum were also examined to see if their growth response was consistent with their feeding behaviour. Adult slugs were collected from burdock rossettes in the fall and were treated identically to the juveniles (32 starved adults, 18 fed adults). This experiment was terminated after 20 days and any slugs that died were omitted from the analysis. Life history considerations D. reticulatum is an annual and may sacri- fice the parent to augment reproduction. Con- sequently compensatory feeding could be abandoned in mature animals. This hypothe- sis was tested by examining the response of a related species with greater parental invest- ment. L. maximus lives 2 to 3 years and TERRESTRIAL SLUG FEEDING 31 TABLE 1. Consumption of adult Deroceras reticulatum on carrot or cucumber following various periods of starvation in a photoperiod of light:dark 16:8 h and temperature of 18°C. FD = The percentage of dry matter in the food. SL = The mean dry pre-starvation weight (mg) of slugs in the treatment. n = The number of slugs on which the means are based. Tot. = The mean daily feeding by an individual slug during the 8 h dark period. | Consumption (means and S.E.) Starvation period g wet food/ cm?) g dry food/ Percent Food (h) Meal g wet slug g dry slug g dry slug feeding Carrot 16 1 0.215 0.029 1.536 0.214 0.234 0.033 100.0 FD = 11.96 2 0.263 0.045 1£87240:322 0.286 0.049 81.3 SL = 0.0740 3 0.242 0.043 1.723 0.304 0.263 0.046 25.0 п = 16 ot: 0.489 3.488 0.532 Carrot 100 1 0.224 0.036 1.596 0.254 0.243 0.039 100.0 РО) = 11:96 2 0.225 0.037 1.606 0.267 0.245 0.041 62.5 SL = 0.0861 3 0.434 0.0 3.093 0.0 0.472 0.0 6.3 п = 16 Tot. 0.392 2.793 0.426 Cucumber 16 1 0.465 0.077 3.366 0.560 0.126 0.021 100.00 FD = 2.98 2 0.379 0.059 2.742 0.543 0.103 0.020 63.2 SL = 0.0574 3 0.391 0.059 2.831 0.434 0.106 0.016 21-1 п = 19 4 0.218 0.0 1.581 0.0 0.059 0.0 5.3 Tot. 0.798 DT 0.216 Cucumber 40 1 0.303 0.048 2.805 0.449 0.104 0.017 100.0 FD = 3.79 2 0.258 0.038 2.387 0.356 0.089 0.013 78.6 SL = 0.0613 3 0.405 0.192 37511782 0.139 0.066 21.4 n= 14 Tot. 0.593 5.484 0.204 Cucumber 100 1 0.393 0.021 2.843 0.154 0.107 0.006 100.0 FD = 2.98 2 0.238 0.022 1.726 0.159 0.065 0.006 76.5 SL = 0.0742 3 0.228 0.039 1.652 0.281 0.062 0.011 29.4 п = 17 4 0.261 0.0 1.888 0.0 0.071 0.0 5.9 Tot. 0.657 4.759 0.179 Cucumber 220 1 0.319 0.046 2.399 0.347 0.075 0.011 100.0 FD = 3.17 2 0.210 0.032 1.580 0.239 0.049 0.007 63.6 SL = 0.0432 3 04131770!022 0.988 0.162 0.031 0.005 36.4 n = 11 4 0.145 0.045 1.092 0.341 0.034 0.011 18.2 Tot. 0.527 3.962 0.124 grows to 5 to 20 g compared to only 0.5 to 1.5 9 for D. reticulatum. If life history tactics are important, adults of L. maximus should strongly compensate for starvation. L. maximus were collected from a local deciduous woodland in Hamilton, Ontario (a new locality record) in October. The slugs were starved for 24 h to clear their guts and then weighed (range = 3.9 to 15.6 д live weight, n = 11). The experiment was carried out identically to that for D. reticulatum, ex- cept that cubes of potato tuber (Solanum tuberosum L.) were used as food, and only dry weight consumption was considered. Po- tato was chosen because it was highly palat- able to this species, whereas the acceptability of carrot or cucumber varied among individu- als. Nocturnal feeding was monitored follow- ing 24 h starvation and, using the same ani- mals, after 288 h starvation. The potato used following 24 h of starvation was 78.98% water and that for the 288 h starvation period was 78.19% water. RESULTS Daily feeding pattern and the starvation response To calculate mean meal sizes, daily con- sumption and the likelihood of successive meals, only animals that fed were considered (Table 1). Slugs starved 16 or 40 h all ate but 32 ROLLO TABLE 2. Feeding of Deroceras reticulatum on the second night following 16 h or 100 h of starvation. FD = The percentage of dry matter in the food. SL = The mean dry weight of slugs (mg). п = The number of slugs in the treatment. Initial Consumption (g wet starvation food/g wet slug) period Meal un Percent Food (h) number Mean SS; feeding Cucumber 16 1 0.336 0.098 100 FD = 3.34 2 0.322 0.123 29.4 SL = .0574 Total 0.373 n = 17 Cucumber 100 1 0.181 0.031 100 FD = 3.34 2 0.321 0.168 38.5 SL = .0861 3 0.166 0.0 TEL. n = 13 Total 0.317 TABLE 3. Feeding of Limax maximus on potato tubers following starvation for 24 h or 288 h (18° С, light:dark, 16:8, R.H. 100%). Feeding (mean +/- S.E.) Mean dry food/g dry sl weight Number EISEN) of slugs of Daily Treatment (dry 9) animals Meal 1 Meal 2 Consumption Starved 0.908 11 0.405, 0.057 0.237, 0.074 0.583, 0.096 24 h п! ni 5 n = 11 Starved 0.859 11 0.247, 0.050 0.081, 0.021 0.321, 0.051 288 h п = 11 п = 10 n = 11 only 90% of slugs starved 100 h and 85% of slugs starved 220h ate. Slugs starved longer than 40 h also responded slowly to the presence of food. Starvation had little influ- ence on the percentage taking a second meal of cucumber (63%-79%), but the probability of eating a third or fourth meal increased markedly with prolonged starvation (Table 1). Despite this, there was a progressive decline in daily consumption and meal size with in- creasing deprivation, no matter how con- sumption was measured. D. reticulatum ate 1 to 4 meals of cucumber per night. The first meals were the largest (.303 to .465g wet food/g live slug), and fourth meals were relatively small (.145 to .261 g wet food/g live slug) (Table 1). D. reticulatum starved 220 h ate 57% as much dry food/day as individuals starved 16 h (66% on a wet weight basis) but they were still 88.7% of their original wet body weight. Thus, feeding was reduced more than expected from loss of body mass. D. reticulatum starved 100 h also ate less carrot than those starved 16h, although meal size was not reduced. No more than three meals of carrot were eaten per night and meal frequency declined with longer deprivation. Nearly twice as much dry carrot as cucumber was eaten per day but at least 1.6 times more cucumber than carrot was ingested in terms of wet weight or volume. For individual meals, the least variation between the diets was ob- tained when wet weight was measured. Most of the difference in daily consumption was related to the number of meals. Although some meals were shorter than the 30 min observation period, it was possible to discern patterns in the duration and frequency of meals. Nearly all meals of cucumber took less than 30-45 min. Initial meals were the longest whereas most fourth meals were com- pleted in less than 30 min. Whereas slugs usually fed and then returned to their resting position, those starved 100 or 220 h spent 60 to 90 min resting on the food. Slugs took con- TERRESTRIAL SLUG FEEDING 33 6 ® ig "а оу? « & м | N Y 5 q or O С, 2 ¿e N ¿RAY < = Е E EB a aie x Dm |] O E an = O = = ETE =4 SL uJ Vn = S = = 15 7} се Y > а’ $ © 3 О 20 40 60 80 100 TIME (days) FIG. 1. Growth of juvenile Deroceras reticulatum at 18°C when fed leaves of burdock, or starved for 36 days prior to feeding. The regression equation for fed controls (м) was: Log(Y) = 4.175 + 0.01805(X) г? = 0.93, п = 18, p < 0.001. The best-fitting equation for degrowth of starved slugs (У) was: Log(Y) = 4.187 — 0.0291(X) г’ 099 п = Эр < 0001. During the compensatory period (®) the regression equation was: Log(Y) = 0.7517 + 0.0615(X) г — 101972 — 6, p < 0002: The following final growth trajectory (©) had the equation: Log(Y) = 3.8051 + 0.01066(X) where Y = log wet weight (mg) and X = days. siderably longer to eat carrot. Average dura- tions ranged from less than 30 up to 90 min, but first meals usually required nearly 1 h. The interval between the first and second meals of cucumber was the longest, but de- creased with longer starvation. Slugs starved 16h ate their second meal after about 4h, whereas Slugs starved 220 h ate again in only fi— 0:82, — 5, р = 0410; 2 h. The interval between meals 2 and 3 was usually 2.5 h. Slugs starved 220 h, however, ate again in about 1.5 h. In all cases meals 3 and 4 were only 1 or 2h apart. Animals deprived 16h and then fed carrot ate again about 2.5 h following their first feeding. Those starved 100 h, however, took 3.5 to 4h to initiate a second meal. In both treatments the 34 ROLLO 6:6 u a oo 6-4 e = o > E = T Pe a E z 6-2 2 ш = 6-0 e SE O 2 - 6 8 10 12 14 16 18 20 TIME (days) FIG. 2. Growth of adult Deroceras reticulatum when fed leaves of burdock or starved at 18°C. The best-fitting equation for the fed animals (m) was: Log(Y) = 6.5277 — 0.009446(X) Log(Y) = 6.4646 — 0.021795(X) where Y = log wet weight (mg) and X = days. interval between the second and third meals was about 2 h. Next day feeding Feeding on the second day following rees- tablishment of food was greatly reduced, re- gardless of the deprivation period. Ten per- cent of the slugs did not feed, individual meals were smaller, and the likelihood of taking consecutive meals was markedly reduced (Table 2). Slugs initially starved 16h never took more than two meals of cucumber on the second day of feeding. Daily wet weight con- sumption was only 46.7% that of the first day for slugs deprived 16 h and 48.2% of first-day feeding for those starved 100 h (Table 2). г — 0.510 — 6.0 =.0.50: For starved animals (@) the regression equation was: rm = 0.95, п =6,-p < 0.005, Life cycle considerations Juvenile D. reticulatum fed burdock grew rapidly, but after 36 days of starvation they were 37% of their original weight and only 22% the weight of controls (Fig. 1). There was no mortality in controls but starvation was continued until 50% of the slugs died. When fed again the starved animals grew three times faster than controls (р <0.001) for about 20 days (Fig. 1). Following this their growth rate was similar to that of fed controls (Fig. 1). Despite compensatory growth, slugs that had been starved were only 38% the weight of fed controls after 100 days. There was remarkable variation in individual growth rates, and this was accentuated by starvation. TERRESTRIAL SLUG FEEDING 35 Some individuals lost weight slowly compared to others, and when re-fed some slugs grew slowly. Other individuals had compensatory growth rates much greater than the mean values illustrated in Fig. 1. Adult D. reticulatum had negative growth rates but starved adults lost weight twice as rapidly as controls (Fig. 2). After 20 days fed and starved animals were 90.5% and 66% of their original weights, respectively. Although fed animals oviposited at a constant rate of about 1.2 eggs/slug/d over the 20 days, starved slugs produced 0.5 eggs/slug/day for the first 15 days and then produced none. Mortality increased steadily with time in both treatments. There was no mortality for the first 5 days, but after 20 days 80% of the starved slugs, and 60% of the fed animals had died. In a preliminary experiment with 82 adults col- lected in late fall, none withstood more than 27 days starvation. Life history considerations L. maximus showed a similar response to starvation as D. reticulatum. Slugs starved 288 h ate only 55% as much as those de- prived 24h (Table 3). This was a conse- quence of meal size since 91% of those starved 288 h ate a second meal compared to only 45% of those starved 24 h. The second meal was much smaller than the first. Thus, despite more frequent feeding, food con- sumption was reduced by longer starvation. The pattern of daily feeding was strongly bimodal. A large early peak of feeding oc- curred during the first 1.5 h of the dark period and a second, smaller peak occurred during h 3 to 4.5. Large amounts of pink saliva were secreted during feeding. One animal that weighed 5.4 g, for example, left 500 mg of saliva on the food. DISCUSSION D. reticulatum ate 2 to 4 meals per night but meal number varied strongly among individu- als and diets. Meal frequency may be related to food hydration. There were at most, two meals of potato (hydration = 78%) (by L. maximus), three of carrot (hydration = 88%), and four of cucumber (hydration = 97%) (by D. reticulatum). In unpublished studies with L. maximus, cucumber was digested faster than carrot. Slugs ate more meals of moister foods and thus ate more on a wet-weight basis (Table 1) but greater amounts of drier foods were ingested on a dry-weight basis (0.583 of potato, 0.532 of carrot and 0.216 of cucum- ber (g dry food/g dry slug/day)). Rollo (1987) provides a more extensive analysis support- ing this result. Despite differences in process- ing rates among diets, all meals were depos- ited as faecal strings within 24 h in agreement with observations by Pallant (1970) and Walker (1972). Wet weight and volume were better com- parative measures of meal size than dry weight (Table 1). Even small variations in the hydration of particular cucumbers pro- duced noticeable differences among dry weight meal sizes (Table 1). Thus, although dry weight may be more important for produc- tivity, wet weight or volume appear to be more important for regulation of ingestion. Volume was slightly more variable than wet weight, possibly due to compression of food in the crop, or water exchange between the food and body. Langer (1975) even suggested that instant mashed potatoes can kill slugs by swelling in their guts. Meal size is partially controlled by inhibitory feedback from the crop, presumably by stretch receptors (Sus- swein & Kupfermann, 1975a, 1975b; Sense- man, 1978; Reingold 8 Gelperin, 1980), sug- gesting that volume should regulate intake. If these receptors were arranged to detect load, however, it could explain why wet weight was more consistent. The behaviour of D. reticulatum suggested that compensation for starvation was occur- ring since the number and frequency of meals increased with increasing deprivation. Starved animals ate more meals and had shorter in- tervals between them. Due to decreased meal size, however, daily consumption was pro- gressively reduced by increasing starvation (Table 1). The fact that starved animals fed more actively, and that their feeding the next day did not increase, suggests that a reduction in general metabolism cannot itself explain these results. Snails may arouse from dor- mancy in a matter of minutes and can rapidly alter their metabolic rate (Vorhaben et al., 1984). Alternatively, slugs starved for long pe- riods were slow to respond to the presence of food, and more of them did not feed which suggests that metabolism may have been de- pressed. Slugs starved 100 and 220 h often remained in contact with the food long after eating whereas slugs starved less always returned to their resting sites. Food intake is regulated by 36 ROLLO antagonism between signals that stimulate feeding (food palatability, empty crop, de- pleted reserves) and those that inhibit eating (feeding deterrents, adaptation of gustatory sense organs, full crop, high metabolic re- serves) (Gelperin, 1971; Senseman, 1978; Reingold & Gelperin, 1980). Hunger may not be well represented by amounts consumed if reserves are important. For example, starved rats eat about the same sized meals as fed ones, but starved rats will ingest foods con- taining greater amounts of repellents (Miller, 1955). The extended association of starved slugs with their food may be related to con- tinued demands from depleted reserves even after the crop is full. If so, it also means that sensory adaptation of the gustatory receptors takes considerably longer than the time re- quired to feed, and so inhibitory feedback from the gut would be a more important regulatory mechanism. Susswein & Kupfermann (1975b) found that starved sea slugs (Aplysia) ate most on the first day of re-feeding. Consumption on subsequent days was reduced by 37%—48%, similar to the results with D. reticulatum (Ta- ble 2). The reduced feeding in Aplysia was due to the presence of food in the anterior gut and the sole mechanism was inhibitory feed- back from bulk. This feeding bottleneck may be exaggerated if passage of food through the gut is slowed to enhance assimilation efficiency (Calow, 1975b). There may be a switch to more rapid processing in the contin- ued presence of food. The fact that adult L. maximus did not exhibit compensatory feeding (Table 3), sug- gests that this response was not related to life history tactics (i.e., long-lived large adults versus short-lived small adults). Alternatively, immature D. reticulatum showed strong com- pensatory growth (and presumably feeding) following starvation (Fig. 1). Adults continued to reproduce and lost weight whether they were starved or not (Fig. 2). L. maximus also has a period of rapid growth during June and July, followed by reproduction and weight loss after mid-August (Rollo, 1983). Feeding was very high during the growth phase, but unex- pectedly declined during reproduction (Rollo, 1983). Growth and reproduction in molluscs are interdependent and largely antagonistic. Hormones that stimulate maturation and re- production simultaneously retard body growth (see Geraerts & Joosse, 1984). These results suggest that one phase of the life cycle is devoted to accumulation of resources (with compensatory control) and another phase is associated with output and lack of homeo- stasis. It was surprising that adult slugs did not compensate for starvation, particularly since reproduction would presumably be enhanced by increased feeding (Sota, 1985). Cock- roaches starved for two weeks had feeding rates four times greater than normal and ele- vated feeding persisted for more than two weeks (Rollo, 1984). Barton-Browne (1975) suggested that compensatory feeding follow- ing starvation might be universal in insects. What differences between insects and mol- luscs might explain these observations? One major difference is that insects store large quantities of lipid in localized sites (i.e. their fat bodies), whereas gastropods have relatively diffuse reserves of carbohydrates. Glycogen is stored in special connective tissue cells con- centrated in the mantle, digestive gland and ovotestis of gastropods. The muscles also store glycogen and the albumen gland con- tains large quantities of galactogen (Veld- huijzen 8 Dogterom, 1975; Veldhuijzen 4 Cuperus, 1976; Widjenes & Runham, 1977; Hemminga et al., 1985a, 1985b). Lipids act mainly as structural elements (1.36% of live weight in the terrestrial snail Cepaea nemo- ralis), not as energy stores (van der Horst, 1970; Horne, 1977). The glycogen reserves from various body compartments are mobi- lized during starvation to maintain concentra- tions of blood sugar (Veldhuijzen, 1975), but the galactogen from the albumen gland is not utilized (Veldhuijzen & van Beek, 1976). A marked response of slugs to starvation was degrowth. After 40 d of starvation, imma- ture D. reticulatum were 37% of their original body weight and they were only 22% the weight of controls (Fig. 1). Despite this, they were completely normal in appearance which is unlikely unless there was de-differentiation. Other soft-bodied invertebrates adjust their size to food conditions. Triclads, for example, may lose 90% to 97% of their body mass dur- ing starvation (Calow, 1977). Degrowth has been documented in aquatic snails in terms of protein loss (Russell-Hunter 8 Eversole, 1976) and de-differentiation (de Jong-Brink, 1973). A degrowth interpretation is also consistent with the results of Horne (1977) who showed that Limax flavus utilized protein as a major sub- strate during starvation. Degrowth cannot completely explain the lack of compensatory feeding following starvation in D. reticulatum, since feeding was reduced much more than TERRESTRIAL SLUG FEEDING 37 body size. After 220 h of starvation, feeding was only 57% of consumption by slugs starved 16 h whereas body size was 88.7% of original wet weight. Similarly L. maximus were 91% of their original wet weight after 288 h starvation but feeding was reduced to 55% of those starved 24 h (Table 3). Some aquatic snails do show compensa- tory responses like those of insects. Calow (1975a) observed increased ingestion rates following starvation or on low quality food. Vianey-Liaud (1984) showed that immatures of the snails Biomphalaria glabrata and B. pfeifferi had compensatory growth rates fol- lowing starvation. Like the present results with D. reticulatum, the compensatory period lasted about two weeks and starved snails never attained the size of fed controls. In another study (Hawryluk & Rollo, unpub- lished), we found that the aquatic snails Stagnicola elodes and Physella gyrina in- creased their daily consumption of food by 3.97 and 1.75 times, respectively, when their diets were diluted by 75% with cellulose. Reingold & Gelperin (1980) did not see such an increase with L. maximus when the diet was diluted with agar (but this also increased meal hardness) and no compensatory in- crease in feeding was observed in the sea hare, Aplysia, fed low quality food (Susswein 8 Kupfermann, 1975a, 1975b). Insects, having extensive hard parts, may be better able to employ “set point” homeo- static control whereas slugs and snails may rely more on flexibility. Slugs in particular may scale their size in response to environ- mental constraints and opportunities. Thus, although molluscs are capable of long-term regulation of feeding (i.e., feeding responsive to reserve depletion or deviations from potential growth), degrowth may be more important, particularly after maturation and reallocation of reserves into reproduction. The ability of molluscs to rapidly alter their metabolic rate and degrow may make reli- ance on large reserves of concentrated energy (i.e., lipids) unnecessary. More data are required before a model of feeding regulation can be constructed for gastropods. There appear to be major differ- ences among species and with maturation. It would probably be worthwhile to examine changes in water content, and relative degrowth of particular organ systems during starvation. Changes in relative gut size might be especially important (see Sibly, 1981). Data addressing a greater number of species with divergent diets, morphologies and life histories are required. ACKNOWLEDGEMENTS | thank Diana Rollo for help with the exper- iments. This project was funded by a grant from the Natural Sciences and Engineering Research Council of Canada. | was sup- ported as a N.S.E.R.C. research fellow during this time. LITERATURE CITED BARTON-BROWNE, L., 1975, Regulatory mecha- nisms in insect feeding. Advances in Insect Physiology, 11: 1-116. CALOW, P., 1975a, The feeding strategies of two freshwater gastropods Ancylus fluviatilis Müll. and Planorbis contortus Linn. (Pulmonata), in terms of ingestion rates and absorption efficien- cies. Oecologia (Berlin), 20: 33-49. CALOW, P., 1975b, Defecation strategies of two freshwater gastropods, Ancylus fluviatilis Müll. and Planorbis contortus Linn. (Pulmonata), with a comparison of field and laboratory estimates of food absorption rate. Oecologia (Berlin), 20: 51-63. CALOW, P., 1977, The joint effect of temperature and starvation on the metabolism of triclads. Oikos, 29: 87-92. GERAERTS, W. P.M. & JOOSSE, J., 1984, Fresh- water snails (Basommatophora). In: TOMPA, A.S., VERDONK, N. H. & VAN DEN BIG- GELAAR, J. A. M., eds., The Mollusca: Repro- duction, Vol. 7, pp. 141-207. Academic Press, New York and London. GELPERIN, A., 1971, Regulation of feeding. An- nual Review of Entomology, 16: 365-368. GODAN, D., 1983, Pest slugs and snails: biology and control. Springer Verlag, Berlin, pp. 1—445. НЕЕС, J., 1977, Oxygen consumption and the use of metabolic reserves during starvation and aestivation in Bulinus (Physopsis) africanus (Pulmonata: Planorbidae). Malacologia, 6: 549— 560. HEMMINGA, M. A., KOOMEN, W., MAASKANT, J.J. & JOOSSE, J., 1985a, Effects of photo- period and temperature on the glycogen stores in the mantle and the head-foot muscles of the freshwater pulmonate snail Lymnaea stagnalis. Comparative Biochemistry and Physiology, 80B: 139-143. HEMMINGA, M. A., MAASKANT, J. J., KOOMEN, W. 8 JOOSSE, J., 1985b, Neuro-endocrine con- trol of glycogen mobilization in the freshwater snail Lymnaea stagnalis. General and Compar- ative Endocrinology, 57: 117-123. HORNE, F. R., 1977, Regulation of urea bio- 38 ROLLO synthesis in the slug Limax flavus Linne. Com- parative Biochemistry and Physiology, 56B: 63-69. HORST, van DER, D. J., 1970, Investigation of the synthesis and distribution of fatty acids in the lipids of the snail Cepaea nemoralis (L.). Neth- erlands Journal of Zoology, 20: 1-79. JAREMOVIC, R. & ROLLO, C. D., 1979, Tree climbing by the snail Cepaea nemoralis (L.): a possible method for regulating temperature and hydration. Canadian Journal of Zoology, 57: 1010-1014. JONG-BRINK, М. DE, 1973, The effects of desicca- tion and starvation upon the weight, histology and ultrastructure of the reproductive tract of Biomphalaria glabrata, intermediate host of Schistosoma mansoni. Zeitschrift für Zell- forschung und Mikroskopische Anatomie, 136: 229-262. LANGER, R. W., 1975, Grow it indoors. Saturday Review Press, E. P. Dutton and Company, New York. MILLER, N. E., 1955, Shortcomings of food con- sumption as a measure of hunger: results from other behavioural techniques. Annals of the New Jersey Academy of Science, 63: 141-143. PALLANT, D., 1970, A quantitative study of feeding in woodland by the grey field slug (Agriolimax reticulatus (Múller)). Proceedings of the Mala- cological Society of London, 39: 83-87. REINGOLD, $. С. & GELPERIN, A., 1980, Feeding motor programme in Limax. Il. Modulation of sensory inputs in intact animals and isolated central nervous systems. Journal of Experimen- tal Biology, 85: 1-19. RICHTER, К. O., 1976, The foraging ecology of the banana slug Ariolimax columbianus Gould (Arionidae). Ph.D. Thesis, University of Wash- ington, Seattle, Washington. ROLLO, C. D., 1982, The regulation of activity in populations of the terrestrial slug тах maximus (Gastropoda: Limacidae). Researches on Popu- lation Ecology, 24: 1-32. ROLLO, C. D., 1983, Consequences of competition on the time budgets, growth and distributions of three species of terrestrial slugs. Researches on Population Ecology, 25: 44-68. ROLLO, C. D., 1984, Resource allocation and time budgeting in adults of the cockroach, Periplaneta americana: the interaction of behaviour and met- abolic reserves. Researches on Population Ecol- ogy, 26: 150-187. ROLLO, C. D., 1987, A quantitative analysis of food consumption for the terrestrial Mollusca. Malacologia, 28: RUSSELL-HUNTER, W. D. 8 EVERSOLE, А. G., 1976, Evidence for tissue degrowth in starved freshwater pulmonate snails (Helisoma trivolvis) from tissue, carbon and nitrogen analyses. Com- parative Biochemistry and Physiology, 54A: 447-453. SCHMIDT-NIELSEN, K., TAYLOR, С. В. & SHKOLNIK, A., 1971, Desert snails: problems of heat, water and food. Journal of Experimental Biology, 55: 385-398. SENSEMAN, D. M., 1977, Gastropod molluscs as model systems for the study of integrative mech- anisms controlling feeding behavior. In: KARE, M. & MALLER, O., eds., The chemical senses and nutrition, Academic Press, New York and London, 2: 3-23. SENSEMAN, D. M., 1978, Short-term control of food intake by the terrestrial slug Ariolimax. Journal of Comparative Physiology, 124: 37-48. SIBLY, R. M., 1981, Strategies of digestion and defecation. In: TOWNSEND, C. R. & CALOW, P., eds., Physiological ecology: an evolutionary approach to resource allocation and resource use. Sinauer Associates, Sunderland, Massa- chusetts, pp. 109-139. SLANSKY, F. JR. & SCRIBER, J. M., 1985, Food consumption and utilization. In: KERKUT, G. A. 8 GILBERT, L. I. eds., Comprehensive insect physiology, biochemistry and pharmacology, Pergamon Press, Oxford, New York, Paris, 4: 88-163. SOTA, T., 1985, Limitation of reproduction by feed- ing condition in a carabid beetle, Carabus vaconinus. Researches on Population Ecology, 27: 171-184. SUSSWEIN, А. J. & KUPFERMANN, I., 1975a, Bulk as a stimulus for satiation in Aplysia. Be- havioral Biology, 13: 203-209. SUSSWEIN, А. J. & KUPFERMANN, I., 1975b, Localizations of bulk stimuli underlying satiation in Aplysia. Journal of Comparative Physiology, 101: 309-328. VELDHUIJZEN, J. P., 1975, Effects of different kinds of food, starvation and restart of feeding on the haemolymph-glucose of the pond snail Lymnaea stagnalis. Netherlands Journal of Zo- ology, 25: 89-102. VELDHUIJZEN, J. P. & CUPERUS, R., 1976, Ef- fects of starvation, low temperature and the dorsal body hormone on the in vitro synthesis of galactogen and glycogen in the albumen gland and the mantle of the pond snail Lymnaea stagnalis. Netherlands Journal of Zoology, 26: 119-135. VELDHUIJZEN, J. Р. & DOGTEROM, С. E., 1975, Incorporation of '*C-glucose in the polysac- charides of various body parts of the pond snail Lymnaea stagnalis as affected by starvation. Netherlands Journal of Zoology, 25: 247-260. VELDHUIJZEN, J. P. & BEEK, G. van, 1976, The influence of starvation and of increased carbohy- drate intake on the polysaccharide content of various body parts of the pond snail Lymnaea stagnalis. Netherlands Journal of Zoology, 26: 106-118. VIANEY-LIAUD, M., 1984, Effects of starvation on growth and reproductive apparatus of two im- mature freshwater snails Biomphalaria pfeifferi and Biomphalaria glabrata (Gastropoda: Planorbidae). Hydrobiologia, 109: 165-172. VORHABEN, J. E., KLOTZ, A. V. & CAMPBELL, TERRESTRIAL SLUG FEEDING 39 J. W., 1984, Activity and oxidative metabolism of WIDJENES, J. & RUNHAM, N. W., 1977, the land snail Helix aspersa. Physiological Zool- Studies on the control of growth in Agrioli- ogy, 57: 357-365. max reticulatus (Mollusca, Pulmonata). Gen- WALDBAUER, G. P., 1968, The consumption and eral and Comparative Endocrinology, 31: 154— utilization of food by insects. Advances in Insect 156. Physiology, 5: 229-288. WRIGHT, A. A. & WILLIAMS, R., 1980, The effect WALKER, G., 1972, The digestive system of the of molluscicides on the consumption of bait by slug, Agriolimax reticulatus (Muller): Experi- slugs. Journal of Molluscan Studies, 46: ments on phagocytosis and nutrient absorption. 265-281. Proceedings of the Malacological Society of Lon- don, 40: 33—43. Revised Ms. accepted 26 Мау 1986 MALACOLOGIA, 1988, 28(1-2): 41-51 A QUANTITATIVE ANALYSIS OF FOOD CONSUMPTION FOR THE TERRESTRIAL MOLLUSCA: ALLOMETRY, FOOD HYDRATION AND TEMPERATURE C. David Rollo Department of Biology, McMaster University, 1280 Main St. W., Hamilton, Ontario, Canada, L8S 4K1 ABSTRACT Regression analysis was used to relate the dry-mass consumption of terrestrial slugs and snails to their body weight both intra- and interspecifically. The mean mass exponent for significant intraspecific analyses was 0.784 (7 species, 13 analyses). Multiple regression was performed for the interspecific analysis (686 cases, 18 gastropod species, 35 foods) so that variation associated with temperature and food hydration could be accounted for. The analysis explained 77% of the variation in the pooled data. Feeding was nearly in direct proportion to body weight (mass exponent of 0.919). There was a strong negative relationship between ingestion and the hydration of the food. The О.о for consumption was 1.7. The general equation provided may be used as a standard for palatability studies. Key words: allometry; feeding; terrestrial gastropods; temperature; palatability. INTRODUCTION There is a burgeoning literature concerned with application of allometric equations of the form Y = aMP that relate various physiologi- cal, behavioural and ecological attributes (Y) of organisms to their body mass (M) (re- viewed by Peters, 1983; Calder, 1984; and Schmidt-Nielsen, 1984). The fact that so many aspects of organism design scale sim- ply to an exponent of body mass (usually a power (b) of 0.75), suggests that there may be underlying rules or constraints that will allow us eventually to develop a unified theory of organism design. Most allometric studies have dealt with vertebrates, and there are particularly few treatments of molluscs available. Von Bertalanffy (1951) included some data on the respiration of freshwater snails (Planorbis spp.), and Innes & Houlihan (1981) provided a more comprehensive investigation for res- piration of intertidal gastropods. These ap- pear to be the only studies addressing the Mollusca with any generality. The present paper provides a general anal- ysis of food consumption for the terrestrial Mollusca. Besides its theoretical interest, feeding by terrestrial gastropods is of major ecological and economic concern. These an- imals constitute an important component of most natural communities, acting as both primary herbivores and decomposers. Their impact on particular plant species (Pallant, 1969; Cates, 1975; Phillipson, 1983; Phil- lipson & Abel, 1983), and their role in nutrient cycling and energy flow requires knowledge of feeding rates (Mason, 1970a, 1970b; Pal- lant, 1974; Jennings & Barkham, 1975, 1976; Richter, 1979; Seifert & Shutov, 1981). Terrestrial molluscs are also important pests of agriculture, sylviculture and floricul- ture (Runham & Hunter, 1970; Godan, 1983). Poisoned bait isthe major method for slug and snail control (Wright & Williams, 1980). Con- sequently, knowledge of consumption is use- ful both for predicting the impact of slugs and snails on plant populations, and for projecting the effectiveness of control programmes. METHODS The daily food consumption of terrestrial molluscs was analysed with respect to their body mass, ambient environmental tempera- ture and the water content of their food. The data consisted of 686 cases for 9 species of slugs and 9 species of snails. Suitable data (640 cases) were mainly obtained from pub- lished text and tables or were interpolated from figures (Mason, 1970a; Pallant, 1970; Stern, 1970; Gelperin, 1975; Jensen, 1975; Richardson, 1975; Williamson, 1975a, 1975b; Davidson, 1976; Jennings 8 Barkham, 1976; Richter, 1976, 1979; Williamson & Cameron, 42 ROLLO 1976; Morton, 1979; Senseman, 1978; Reingold & Gelperin, 1980; Wright & Wil- liams, 1980; Bailey, 1981; Seifert & Shutov, 1981). These data were supplemented with 46 original observations. Additional informa- tion on temperature, hydration of foods, ani- mal size and food consumption was obtained directly through correspondence with authors. The amount of dry food consumed was used as the dependent variable. If authors reported wet consumption and did not provide food hydration, this was obtained from other studies or it was determined directly in the laboratory. Some types of leaves could not be obtained, and these cases were assigned a hydration of 82.08%, the mean hydration of other leaves (range, 75%-85%). Dry tissue weight of the animals was used as an independent factor. Where authors used wet body weight and did not provide hydration, the value reported in other studies, or original observations were used. If no information was available it was assumed that the body was 89% water. This is typical of fully-hydrated terrestrial molluscs (Rollo et al., 1983). No correction was made for the small inter- nal shells possessed by some species of slugs. However, only dry tissues were consid- ered for snails. Where authors reported live snail weight, the percentage of the body mass attributed to the shell was obtained from the literature or original observations. Where no data were available, the animals were as- sumed to have shells similar to adult Cepaea nemoralis |. (i.e., shell = 15.46% of live weight). In preliminary regression analyses using body weight as the independent variable, more of the variation in feeding was explained when the wet weight of the food was used instead of the dry weight. Consequently, the hydration of the food was included as an independent variable in the current analysis of dry weight consumption. Temperature influences rates of physiolog- ical and behavioural processes and conse- quently is usually provided by authors. Where temperature was not reported it was some- times obtained by correspondence or was assumed to be the same as in other studies conducted in the same laboratory. Where no information was available, a value of 15°C was assumed. The optimal temperature range for activity of most temperate terrestrial molluscs falls between 10°C and 20°C. Some species do not tolerate prolonged tempera- tures above 20°C, and so 15°C-18°C is the most common range selected by authors who are not specifically addressing the influence of temperature. In addition to literature values, the analysis included original observations on daily feed- ing for the slugs Deroceras reticulatum (Müller), and Limax maximus L. These were obtained in the course of another study ex- amining the regulation of intake of individual meals. L. maximus (n = 11) were fed cubes of potato tuber (Solanum tuberosum L.) in a light to dark cycle of 16:8h at 18°C. D. reticulatum (n = 16) were fed cubes of carrot root (Daucus carota L.) or English cucumber (Cucumis sativus L., var. anglicus Bailey, п = 19). The animals were housed individually in jars with approximately 2 cm of moistened vermiculite on the bottom to maintain humidity (slugs do not ingest vermiculite). Slugs were maintained in the experimental conditions for several weeks prior to the observations. The animals were starved for 24 h prior to each experiment to clear their guts of food and then they were weighed. Faecal strings were removed to prevent coprophagy. For each experiment, 20 cubes of the appropriate diet were weighed, dried to constant weight at 60°C, and re-weighed. This sample provided an estimate of the water content of the mate- rial. Each slug was given pre-weighed cubes of fresh food at the beginning of each dark cycle. The food remaining after 24h was dried to constant weight at 60°C. The amount eaten was then calculated by subtracting the dry weight of the remainder from the esti- mated dry weight of the original food. Intraspecific analyses were conducted us- ing simple linear regression. For the interspecific analysis stepwise multiple re- gression and partial correlation analysis were performed using the Statistical Package for the Social Sciences (SPSS). In many in- stances authors reported the mean values of studies conducted with groups of animals in- stead of data for individuals. Clearly, a case based on 20 animals should have more weight than one based on a single specimen. Con- sequently, the analysis was performed with each case weighted (i.e., duplicated) for the number of animals that it represented. This resulted in the final analysis being based on 1,725 cases. The statistical model that ac- counted for the greatest amount of variation in the data was obtained by exploratory analysis (regressions and scatter diagrams) to find ap- propriate transformations to obtain linearity. LAND MOLLUSC FOOD CONSUMPTION 43 TABLE 1. Intraspecific regression analyses of food consumption (Log dry mg/d) for various gastropod species with respect to their body mass (Log dry mg shell-free tissue). Species marked with a “*" are freshwater and absorption rate was measured rather than consumption. Units for these species were ug/12 h for absorption and mg dry tissue x 100 for body mass. Con- stant Slope Species (a) (b) r n Ancylus fluviatilis* 1.26 al — 30 1.89 67 — 30 2.15 70 — 30 Planorbis contortus* 1.20 72 — 3 1.74 il — 30 2.80 72 — 30 Helix aspersa 79 ‚349 .468 8 1.25 ‚331.239, 09 —0.05 .514 .914 13 62722 029001512 Arion ater Ariolimax columbianus 1.49 ‚491 .849 33 — 2.78 Deroceras reticulatum 2.24 .624 .189 19 .48 1.394 698 9 .32 ‚464 .099 22 6 .759 .493 19 2.67 18272027716 Deroceras laeve 132 .619 .864 26 Cepaea nemoralis 35265053) Onli 19 Milax budapestensis 121888247827 3.65 —.069 .001 17 Limax maximus 1033 "617107 3.39 1185 2078) 116 Lehmannia marginata 1.37 113 057-23 RESUETS Table 1 is a compilation of intraspecific allometric relationships for consumption and body mass for terrestrial molluscs for which there were enough observations. Some data on absorption rate from aquatic species are also provided. There was considerable varia- tion among studies with respect to the mass exponent (range of —0.617 to 1.394). The mean (0.486), was exceptionally low. If only Temp- Prob- erature ability °C Food Author <.05 4.0 Algae Calow, 1975 <.05 10.0 005 18.0 <.05 4.0 Detritus and Calow, 1975 <.05 10.0 Bacteria <.05 18.0 <.10 10.0 Lettuce Mason, 1970a >.20 15.0 <.001 15.0 Lettuce Stern, 1970 >.50 16.5 Agar diet Jobin & Rollo, unpubl. <.001 15.5 Oplopanax horridum Richter, 1976 <.02 16.5 Agar diet Jobin & Rollo, unpubl. <.10 5.0 Flour bait Wright & <.01 10.0 Williams, 1980 = 10 18.0 Ranunculus repens Pallant, 1970 <.001 18.0 Cucumber This study >.50 18.0 Carrot This study <.001 16.5 Agar diet Jobin & Rollo, unpubl. >50 20.0 Lettuce Richardson, 1975 <.01 15.0 Flour bait Wright & >.50 5.0 Williams, 1980 >.20 18.0 Potato This study, >.20 16.5 Agar diet Jobin & Rollo, unpubl. >.20 16.5 Agar diet Jobin & Rollo, unpubl. those values that were significant (p < 0.05) are considered, however, the mean value was 0.784. The best interspecific analysis using stepwise multiple regression is presented in Table 2. This was obtained when both the dry body mass and dry daily consumption were converted to natural logarithms. A combina- tion of log body mass, temperature and hy- dration resulted in a highly significant regres- sion with г? = 0.7742. This means that 77% 44 ROLLO TABLE 2. Multiple regression analysis of food consumption (Log dry mg/d) of terrestrial Mollusca with respect to their body mass, environmental temperature and food hydration. Bracketed values resulted when the weighting procedure was not employed (see text). г? = 0.77417, п = 1,724, р < 0.00001; (г? = 0.67287, n = 685, p < 0.0001) B Variable (Slope) Body mass 0.9191051 (Log dry mg) (0.8593249) Temperature 0.0539130 (°C) (0.0353888) Food hydration —0.0374423 (%) (—0.0385831) Constant 0.2374659 (0.8768701) of the observed variation in daily consumption was accounted for by the regression with these three variables (Zar, 1974). Interaction effects were also examined (e.g., tempera- ture x body mass), but none were significant. Use of the weighting procedure resulted in statistical calculations based on a larger sam- ple than was originally available. For those readers who may hesitate to accept the cal- culated probabilities in this case, the values obtained when no weighting was employed are also presented (Table 2). The weighted coefficients and constant are certainly the most appropriate. The coefficients calculated without weighting were slightly different and so they are also provided in Table 2. Because consumption is influenced by all three independent factors, as well as other variables, the trend associated with a single factor is not always readily apparent unless it accounts for a large amount of the variation. Partial correlation analysis was performed to examine the association of consumption with particular variables while controlling for the influence of the others. When temperature and food hydration were controlled, body mass accounted for 75% of the residual vari- ation in feeding. When body mass and tem- perature were controlled, food hydration ac- counted for 37% of the remaining variation. When body mass and food hydration were controlled, temperature accounted for 6% of the residual variability. It is possible to illustrate these relationships by computing the appropriate residuals from the raw data. The variation associated with a particular factor can be removed by appropri- Standard error Probability 0.0126404 < 0.00001 (0.0250956) (< 0.0001) 0.0050549 < 0.0001 (0.0100655) (< 0.001) 0.0011767 < 0.0001 (0.0016202) (< 0.00001) 0.1139649 < 0.037 (0.1700361) (< 0.0001) ate transformation of the dependent variable. For example, if feeding were directly propor- tional to body mass, and the influence of hydration was to be examined, the influence of weight could be removed by simply dividing consumption by the animal’s body mass to obtain an independent variable with the units mg eaten/mg body mass. This is common practice. Since the present analysis found that feeding was proportional to an exponent of body weight, the coefficient derived from the multiple regression analysis was em- ployed to this effect (Table 2). A mass expo- nent of 0.75 is sometimes employed in the literature in this manner (e.g., Sibly, 1981). Thus the transformation to remove the varia- tion associated with body weight is: УТ = YO = 0:91910509(X) where YO = log dry consumption (mg), X = log dry body mass (mg), and the coefficient was obtained from Table 2. Similar transformations can be employed to remove the variation associated with temper- ature or hydration of the food. The effect of such transformation is illustrated in Figs. 1 and 2. Fig. 1 shows the relationship between dry consumption and dry body weight using sim- ple linear regression with logarithmic transfor- mations of both variables. Many of the high values of consumption for smaller gastropods were associated with baits that had low water content. Temperature effects also obscure the relationship. Fig. 2 illustrates the relationship between consumption and body mass when the dependent factor has been transformed to LAND MOLLUSC FOOD CONSUMPTION 45 . . . . CONSUMPTION 2 3 4 BODY WEIGHT FIG. 1. The relationship between food consumption by terrestrial gastropods and their body weight. Simple linear regression of (C) consumption (log dry mg) with (W) body weight (log dry mg). The equation for the best fitting line was: С = 0.87409(W) — 1.75540; г? = 0.64000, р < 0.00001, п = 1,725. remove the variation associated with both tem- perature and food hydration. It was very difficult to discern the relation- ship between consumption and temperature or between consumption and food hydration in the original data because of the large influence of weight and other factors. The trends are apparent, however, following ap- propriate transformations of the dependent variable. Fig. 3 illustrates the relationship between food consumption and temperature when variation due to body weight and hydra- tion have been removed. Similarly, Fig. 4 illustrates the relationship between consump- tion and food hydration when the variation associated with body weight and temperature have been removed. The relationship of consumption to temper- ature appears linear over the range of tem- perature examined (Fig. 3). Given the coeffi- cient for temperature from the multiple regression analysis (Table 2), it is possible to obtain the Оз for consumption using the relationship: Log(Q;5) = 10(B) where B = the coefficient for temperature from Table 1 (Innes & Houlihan, 1981). This was checked by using the standard equation for calculating Q;o: Log(Q;0) = (logRate 2 — logRate 1) (10/23.5-5.0) where the rates were calculated by sub- stituting the mean animal mass and food hydration into the multiple regression equa- tion (Table 2). The calculated Оо for consumption of terrestrial Mollusca over a temperature range of 5°C to 23.5°C was 1.7145 using either formula. DISCUSSION The intraspecific mass exponent for metab- olism generally has a value of 0.66 (Heusner, 1982; Feldman & McMahon, 1983; Wieser, 1984). Table 1 showed high variation in the value of b for consumption, with an exception- ally low mean of 0.486. Very few of these 46 ROLLO CONSUMPTION n o 2 3 4 BODY WEIGHT FIG. 2. Simple linear regression of consumption with body weight when the dependent factor (CT) was transformed to remove variation associated with (T) temperature (°C) and (H) hydration of the food (%). The transformation was: CT = С - 0.053913044(T) + 0.037442321(H). The equation for the best fitting line was: СТ = 0.91911(W) + 0.23747; г? = 0.75807, р < 0.00001, п = 1,725. Repeated values are not indicated in the figures. studies were conducted with deriving al- lometric relationships in mind. Consequently, the animals were often matched for size or developmental status. In addition, feeding may be relatively high in young growing stages but may decline in adult or senescent individuals (Wieser, 1984). The mass expo- nent will be strongly affected by the range of body sizes used and the age structure of the sample. This probably explains much of the variation in the calculated values (Table 1). If only studies which were significant are con- sidered, the mean mass exponent was 0.784, close to the value predicted for interspecific comparisons. Further data will be required to determine if terrestrial molluscs in fact obey the surface rule (i.e., b = 0.66) intraspecifi- cally. Heusner (1982) showed that the value of the mass coefficient for metabolism (a) in- creases with body mass in mammals. This indicates that metabolic power or basal me- tabolism increases in larger species (Wieser, 1984). There was no clear trend between the mass coefficients for feeding and body mass in the present analysis (Table 1), although such a trend might emerge in a more con- trolled comparison. Calow's (1975) data show a clear influence of temperature on the mass coefficient but they also suggest that the mass exponent is not affected by tempera- ture. For the interspecific analysis, the regres- sion model that best described the data con- formed to the standard allometric relationship found to apply in studies of other animals, except for the addition of further variables (see Peters, 1983). Considering that the raw data are based on 18 species feeding on 35 different diets, it is remarkable that body weight, temperature and food hydration ac- count for 77% of the observed variation (Ta- ble 2). This leaves only 23% of the variation to be accounted for by species differences, in- dividual variation, acclimation, age, matura- tion state, and activity levels of the gastro- pods or the palatability, physical properties, nutritional value and energy content of the LAND MOLLUSC FOOD CONSUMPTION 47 4 3 . 2 A: s г о Sf a = =) nw г со . o . . . . e . . . . . $ 0000000 0000. ........... = RA RR EEE EEE 5 o 15 20 23.5 TEMPERATURE FIG. 3. The relationship between food consumption by terrestrial gastropods and (T) temperature (°C) (log dry mg) was transformed to remove the variation associated with (W) body weight (log dry mg) and (H) food hydration (%). The transformation was: СТ = С — 0.91910509(W) + 0.037442321(H). The equation for the best fitting line was: СТ = 0.05391(T) + 0.23747; г? = 0.06584, р < 0.00001, п = 1,725. food. This is all the more remarkable since errors arising from assumptions concerning incomplete data, linearity of the relationships, and errors associated with measurement must introduce considerable variation in this kind of analysis. Von Bertalanffy (1951) identified three classes of animals: those whose respiration was directly proportional to their body weight, those whose respiration was proportional to their surface area, and those whose respira- tion was intermediate between these ex- tremes (i.e., mass exponents of 1.00, 0.66 and 0.75 respectively). For aquatic snails he found a mass exponent of 0.75. Subsequent research across wide phylogenetic bound- aries has shown that most behavioural, phys- iological and ecological features scale to this mass exponent interspecifically (Peters, 1983; Calder, 1984; Schmidt-Nielsen, 1984). Innes & Houlihan (1981) found a mass expo- nent of 0.724 for the respiration of intertidal gastropods in air, and 0.701 for those in water (range in equations, 0.63—0.81). This is con- sistent with Von Bertalanffy's (1951) results. Von Bertalanffy (1951) also observed, how- ever, that terrestrial snails of the family Helicidae exhibited respiration rates directly proportional to their body weight. The present results show that daily feeding of terrestrial molluscs also has a mass exponent (0.91) much closer to direct proportionality than to 0.75 (see Figs. 1 and 2). In most other organisms examined, ingestion has a mass exponent close to 0.75 (Peters, 1983). There is no accepted explanation of why particular organisms exhibit a particular mass expo- nent. If the apparent difference between aquatic and terrestrial molluscs is real, how- ever, an answer may emerge from compara- tive studies. One problem with the current data is that various species are not equally represented and there is probably a bias towards adult animals. This could influence the mass expo- nent and may explain the difference between exponents when weighting was applied or not (Table 2). What is required is a systematic survey in which a range of species is com- pared across developmental stages and un- 48 ROLLO ' ul CONSUMPTION | > 1 a 0 25 . ...eoboss | ....... .00. 50 75 100 HYDRATION OF FOOD FIG. 4. The relationship between food consumption by terrestrial gastropods and (H) food hydration (%). Consumption (C) (log dry mg) was transformed to remove variation associated with (W) body weight (log dry mg) and (T) temperature (°C). The transformation was: CT = C — 0.91910509(W) + 0.053913044(T). The equation for the best fitting regression line was: CT = 0.23747 — 0.03744(H); г? = 0.38847, р < 0.00001, n = 1,725. Repeated values are not indicated in the figures. der standard environmental and nutritive con- ditions. We are currently carrying out such a project using an artificial diet embedded in agar. A preliminary analysis using 5 species of slugs reared in identical conditions pro- vided an interspecific mass exponent of 0.697 (Jobin and Rollo, unpublished). Nevertheless, the present analysis indicates that there is a very clear trend in the interspecific data and the line with a slope of 0.919 appears to describe it very well (Fig. 2). The range of species of vastly different sizes and the large number of observations should provide a fairly accurate assessment of interspecific trends. In addition, this analysis allows the variation associated with temperature and food hydration to be addressed. The interspecific Q:, for consumption of terrestrial gastropods was 1.715. This is very close to the values obtained for intraspecific absorption rate in freshwater snails by Calow (1975) (range of 1.38 to 1.82). Terrestrial molluscs are generally adapted to relatively cool temperatures which are in the range of experimentaltemperaturesencountered (5°C— 24°C). Even with size and food hydration controlled, however, temperature only ac- counted for about 7% of the residual variation (Fig. 3). Although most species show maxi- mum activity between 15°C and 20°C, there are differences among species. For example, gastropods such as D. reticulatum, Deroceras laeve (Müller) and Arion hortensis Ferussac remain active at lower temperatures than species such as L. maximus or C. nemoralis. Thus different species have different pre- ferred temperature ranges and this may ob- scure the influence of temperature in the general analysis. In addition, the relationship between tem- perature and behavioural or physiological processes is not linear within a species. Typ- ically, rates are zero at lower and upper temperature thresholds. Within this range processes tend to accelerate gradually with increasing temperatures to an optimum, fol- lowing which they decline precipitously (e.g., Rollo, 1982). Acclimation can shift this tem- LAND MOLLUSC FOOD CONSUMPTION 49 perature response curve (Rising & Armitage, 1969), and this is probably an additional source of variation. The data show a simple linear increase with temperature (Fig. 3), instead of the parabolic response curve that would be expected for a single species. Zero values would undoubt- edly be obtained, however, if observations at even higher and lower temperatures were obtained. Although the pooled data show the general temperature response of terrestrial gastropods with respect to feeding, they may overestimate the temperature range for any single species. Moisture is a major limiting factor for terres- trial molluscs because they have unrestricted evaporation from their epidermis and must secrete a ribbon of mucus to move (Machin, 1975). lt might be expected that foods with higher water content would be preferred. A large component of most diets is water, how- ever, and this can contribute substantially to the volume of a meal. Since feedback from stretch receptors in the crop is a major factor controlling meal size (Senseman, 1978; Reingold & Gelperin, 1980), the dry weight of material ingested may be physically limited where there is a large structural component of water. Meal intake is also influenced by the concentration of gustatory stimulants in the food (Senseman, 1978; Reingold & Gelperin, 1980), and this would be higher in drier diets. This combination of effects probably explains the strong negative relationship between daily feeding and the hydration of the food (Fig. 4). Most of the studies considered for the present analysis ensured that the animals were fully hydrated. The ability of slugs and snails to become active and forage is influ- enced by their hydration (Rollo et al., 1983). Thus, dehydrated animals may exhibit behavioural changes in preference associ- ated with the water content of their food. They may also be physically limited by their inability to prevent losses of body water to osmotically concentrated food in the digestive tract. Quantitative studies are needed in this area. Although there is a large literature con- cerned with feeding by terrestrial molluscs, probably less than 10% of it was sufficiently complete and quantitative, or presented in a form that allowed an analytical synthesis. There are a number of key problems limiting the generality of the literature that can be corrected. For example, the most common measurements of food consumption are ei- ther surface area removed (Judge, 1972; Cates, 1975; Jennings & Barkham, 1975; Reader & Southwood, 1981; Rathcke, 1985), weight (wet or dry) or energy. Area consumed may be a valuable measurement if the impact of the gastropod on a plant's photosynthetic capacity is of interest. Due to variation in leaf thickness, consistency and morphology, how- ever, a given amount of area removed may represent a different amount of absolute con- sumption depending on the plant species considered. Authors should provide a surface-area-to-weight conversion factor for the material they are using. Similarly, if ener- getic units are employed, an energy-to-mass conversion factor should be provided. The water content of the diet is also an important factor (Table 2, Fig. 4). When the size of snails is given, various linear measurements of shell morphology are often used. This may be an advantage when the animals must be left alive since weight may vary with consumption and hydration. Shell morphology and thickness varies with species, age and environment, however, and authors often measure different aspects (e.g., aperture width or shell length). For generality, authors should provide the relationship be- tween the measures employed and the weight of dry tissues. A major focus of feeding studies is the palatability of gastropod foods. Molluscs have been used as model “general herbivores” to test foraging hypotheses (Cates 8 Опапз, 1975; Rathcke, 1985). The use of palatability or acceptability indices to rank the relative attractiveness of diets is standard practice. These indices are usually calculated by com- paring the amount of a particular food eaten to a standard that is highly preferred by a gastropod. Because the resulting index is a relative measure it has no generality, espe- cially since different studies use different standards (see Grime et al., 1968, 1970; Cates 8 Orians, 1975; Richter, 1976; Dirzo, 1980; Richardson & Whittaker, 1982; Whelan, 1982; Rathcke, 1985). Furthermore, the rank- ing of the diet may change according to which plant is used as the reference (even within a single experiment), and depending on whether dry weight or leaf area consumed is measured (Richardson & Whittaker, 1982). These problems of standardization and gen- erality may be overcome if authors use the statistical model developed here as a refer- ence. The degree of deviation from the pre- dicted consumption (positive or negative) could serve as a measure of preference and 50 ROLLO would ensure that all studies use the same units and consider all the relevant factors. ACKNOWLEDGEMENTS | thank all the authors who provided addi- tional information to augment the data. Dr. B. Richardson and Dr. J. B. Whittaker, in partic- ular, supplied considerable data on leaf con- sumption by D. reticulatum. This research was supported by a grant from the Natural Sciences and Engineering Research Council of Canada. | was also supported as an NSERC fellow. LITERATURE CITED BAILEY, S. E. R., 1981, Circannual and circadian rhythms in the snail Helix aspersa Múller and the photoperiodic control of annual activity and re- production. Journal of Comparative Physiology, 142A: 89-94. CALDER, W. A., 1984, Size, function and life history. Harvard University Press, Cambridge, Massachusetts, 431 pp. CALOW, P., 1975, The feeding strategies of two freshwater gastropods, Ancylus fluviatilis Müll. and Planorbis contortus Linn. (Pulmonata), in terms of ingestion rates and absorption efficien- cies. Oecologia (Berlin), 20: 33-49. CATES, R. G., 1975, The interface between slugs and wild ginger: some evolutionary aspects. Ecology, 56: 391-400. CATES, R. G. 8 ORIANS, G. H., 1975, Succes- sional status and the palatability of plants to generalized herbivores. Ecology, 56: 410-418. DAVIDSON, D. H., 1976, Assimilation efficiencies of slugs on different food materials. Oecologia (Berlin), 26: 267-273. DIRZO, R., 1980, Experimental studies on slug- plant interactions |. The acceptability of thirty plant species to the slug Agriolimax caruanae. Journal of Ecology, 68: 981-998. FELDMAN, H. A. & MCMAHON, T. A., 1983, The 3 mass exponent for energy metabolism is not a statistical artifact. Respiration Physiology, 52: 149—163. GELPERIN, A., 1975, Rapid food-aversion learning by a terrestrial mollusc. Science, 189: 567-570. GODAN, D., 1983, Pest slugs and snails: biology and control. Springer Verlag, Berlin, 445 pp. GRIME, J. P., BLYTHE, С. М. & THORNTON, J. D., 1970, Food selection by the snail Cepaea nemoralis L. In: WATSON, A., ed., Animal pop- ulations in relation to their food resources. British Ecological Society Symposium 10, Blackwell, London, pp. 73-99. GRIME, J. P., MACPHERSON-STEWART, $. F. & DEARMAN, R. S., 1968, An investigation of leaf palatability using the snail Cepaea nemoralis. Journal of Ecology, 56: 405-420. HEUSNER, A. A., 1982, Energy metabolism and body size. |. Is the 0.75 mass exponent of Kleiber's equation a statistical artifact? Respira- tion Physiology, 48: 1-12. INNES, A. J. & HOULIHAN, D. F., 1981, A review of the effects of temperature on the oxygen con- sumption of intertidal gastropods. Journal of Thermal Biology, 6: 249-256. JENNINGS, T. J. & BARKHAM, J. P., 1975, Food of slugs in mixed deciduous woodland. Oikos, 26: 211-221. JENNINGS, T. J. & BARKHAM, J. P., 1976, Quan- titative study of feeding in woodland by the slug Arion ater. Oikos, 27: 168-173. JENSEN, T. F., 1975, A tentative energy budget for a summer population of Arion ater (L.). Natura Jutlandica, 18: 10-20. JUDGE, F. D., 1972, Aspects of the biology of the gray garden slug (Deroceras reticulatum Müller). Search, 2: 1-18. MACHIN, J., 1975, Water relationships. In: FRET- TER, V. & PEAKE, J., eds., Pulmonata vol. |. Functional anatomy and physiology. Academic Press, London, pp. 105-163. MASON, C. F., 1970a, Food, feeding rates and assimilation in woodland snails. Oecologia (Ber- lin), 4: 358-373. MASON, C. F., 1970b, Snail populations, beech litter production, and the role of snails in litter decomposition. Oecologia (Berlin), 5: 215-239. MORTON, B., 1979, The diurnal rhythm and the cycle of feeding and digestion in the slug Deroceras caruanae. Journal of Zoology (Lon- don), 187: 135-152. PALLANT, D., 1969, The food of the grey field slug Agriolimax reticulatus (Müller) in woodland. Jour- nal of Animal Ecology, 38: 391-397. PALLANT, D., 1970, A quantitative study of feeding in woodland by the grey field slug (Agriolimax reticulatus (Müller)). Proceedings of the Malacological Society of London, 39: 83-87. PALLANT, D., 1974, An energy budget for Agriolimax reticulatus (Müller) on grassland. Journal of Conchology, 28: 243-245. PETERS, В. H., 1983, The ecological implications of body size. Cambridge University Press, Cambridge, 329 pp. PHILLIPSON, J., 1983, Slug numbers, biomass and respiratory metabolism in a beech wood- land—Wytham Woods, Oxford. Oecologia (Ber- lin), 60: 38-45. PHILLIPSON, J. & ABEL, R., 1983, Snail numbers, biomass and respiratory metabolism in a beech woodland—Wytham Woods, Oxford. Oecologia (Berlin), 57: 333-338. RATHCKE, B., 1985, Slugs as generalist herbi- vores: tests of three hypotheses on plant choices. Ecology, 66: 828-836. READER, P. М. & SOUTHWOOD, Т. В. E., 1981, The relationship between palatability to inverte- LAND MOLLUSC FOOD CONSUMPTION 51 brates and the successional status of a plant. Oecologia (Berlin), 51: 271-275. REINGOLD, 5. С. & GELPERIN, A., 1980, Feeding motor programme in Limax. Il. Modulation of sensory inputs in intact animals and isolated central nervous systems. Journal of Experimen- tal Biology, 85: 1-19. RICHARDSON, A. M. M., 1975, Food, feeding rates and assimilation in the land snail Cepaea nemoralis L. Oecologia (Berlin), 19: 59-70. RICHARDSON, B. & WHITTAKER, J. B., 1982, The effect of varying the reference material on ranking of acceptability indices of plant species to a polyphagous herbivore. Agriolimax reticu- latus. Oikos, 39: 237-240. RICHTER, К. O., 1976, The foraging ecology of the banana slug Ariolimax columbianus, Gould (Arionidae). Ph.D. Thesis, University of Wash- ington. RICHTER, K. O., 1979, Aspects of nutrient cycling by Ariolimax columbianus (Mollusca:Arionidae) in Pacific Northwest coniferous forests. Pedobiology, 19: 60-74. RISING, Т. L. 4 ААМПАСЕ, К. B., 1969, Acclima- tion to temperature by the terrestrial gastropods, Limax maximus and Philomycus carolinianus: oxygen consumption and temperature prefer- ence. Comparative Biochemistry and Physiol- ogy, 30: 1091-1114. ROLLO, C. D., 1982, The regulation of activity in populations of the terrestrial slug Limax maximus (Gastropoda: Limacidae). Researches on Popu- lation Ecology, 24: 1-32. ROLLO, С. D., VERTINSKY, I. B., WELLINGTON, W. G., THOMPSON, W. A. 8 KWAN, Y., 1983, Description and testing of a comprehensive sim- ulation model of the ecology of terrestrial gastro- pods in unstable environments. Researches on Population Ecology, 25: 150-179. RUNHAM, N. W. & HUNTER, P. J., 1970, Terres- trial slugs. Hutchinson, London, 184 pp. SCHMIDT-NIELSEN, K., 1984, Scaling: why is animal size so important? Cambridge University Press, Cambridge, 241 pp. SEIFERT, 0. \. 8 SHUTOV, S. V., 1981, The consumption of leaf litter by land molluscs. Pedobiology, 21: 159-165. SENSEMAN, D. M., 1978, Short-term control of food intake by the terrestrial slug Ariolimax. Journal of Comparative Physiology, 124A: 37-48. SIBLY, R. M., 1981, Strategies of digestion and defecation. In: TOWNSEND, C. R. & CALOW, P., eds., Physiological ecology: an evolutionary approach to resource allocation and resource use. Sinauer Associates, Sunderland, Massa- chusetts, pp. 109-139. STERN, G., 1970, Production et bilan energéiique chez la limace rouge. Terre et la Vie, 117: 403-424. VON BERTALANFFY, L., 1951, Metabolic types and growth types. American Naturalist, 85: 111-117. WHELAN, R. J., 1982, Response of slugs to unac- ceptable food items. Journal of Applied Ecology, 19: 79-87. WIESER, W., 1984, А distinction must be made between the ontogeny and phylogeny of metab- olism in order to understand the mass exponent of energy metabolism. Respiration Physiology, 55: 1-9. WILLIAMSON, P., 1975a, The feeding ecology and energetics of a grassland population of Cepaea nemoralis L. Ph.D. thesis, Portsmouth Polytech- nic, Portsmouth. WILLIAMSON, P., 1975b, Use of **Zn to determine the field metabolism of the snail Cepaea nemoralis L. Ecology, 56: 1185-1192. WILLIAMSON, P. & CAMERON, R. A. D., 1976, Natural diet of the landsnail Cepaea nemoralis. Oikos, 27: 493-500. WRIGHT, A. A. & WILLIAMS, R., 1980, The effect of molluscicides on the consumption of bait by slugs. Journal of Molluscan Studies, 46: 265-281. ZAR, J. H., 1974, Biostatistical analysis. Prentice Hall, Englewood Cliffs, New Jersey, 620 pp. Revised Ms. accepted 26 June 1986 MALACOLOGIA, 1988, 28(1-2): 53-64 MATING BEHAVIOUR OF LYMNAEA STAGNALIS Y.A. van Duivenboden and А. ter Maat! Department of Biology, Free University, P.O. Box 7161, 1007 MC Amsterdam, The Netherlands ABSTRACT Mating behaviour of the pond snail Lymnaea stagnalis was analysed in pairs of snails, reunited after a period of isolation. Apparent female behaviour is absent in this situation, whereas male behaviour is characterized by a series of consecutive acts: mounting, turning, eversion of preputium, intromission, withdrawal of preputium and moving off. Several variations on this basic pattern occur, of which sham-copulation (false coupling) and reciprocation (reversal of roles after completion of copulation) are the most remarkable ones. The same behavioural sequence appeared to be present in spontaneous matings. Prior experience is not needed for the performance of mating behaviour. The duration of the behaviour is variable. This is mainly due to the latency of intromission. The duration of intromission is fairly constant (36 + 4 min). Inspection of the vagina of female copulants for the presence of semen revealed that 90-100% of the copulations is successful. Surgical removal of the part of the vas deferens that runs through the body wall eliminates all male copulation behaviour without affecting the ability of copulation as a female. Key words: Lymnaea stagnalis, mating, male and female behaviour. INTRODUCTION The reproductive biology of the simulta- neous hermaphrodite freshwater snail Lymnaea stagnalis (Linnaeus) has received much attention in recent years. The first sys- tematic study of the relation between copula- tion and egg laying was conducted by Horstmann (1955). Since then, laboratory studies have been done on endocrine and neurophysiological control of ovulation and oviposition, the structure and maturation of the reproductive system, external factors af- fecting fecundity and related issues (for re- views see Joosse & Geraerts 1983, Geraerts & Joosse 1984). An experimental study on the fecundity of L. stagnalis in the field was presented by Brown (1979). Mating in freshwater pulmonate snails— mostly simultaneous hermaphrodites—is not a necessary condition for egg-laying in many species: under conditions of isolation they may reproduce by internally self-fertilized eggs (Duncan, 1975, Geraerts & Joosse, 1984). Yet mating must be considered a major element in the reproduction tactics of these hermaphrodite animals for several rea- sons: (1) Studies using genetic markers dem- 'To whom all correspondence should be addressed. onstrated that after the snails have been allowed to mate, cross-fertilized offspring is produced for some weeks by animals that have copulated as females (Biomphalaria glabrata: Paraense, 1956, 1959, Richards, 1970; L. stagnalis: Cain, 1956). (2) In L. stagnalis, isolated before the appearance of sexual behaviour, the onset of egg-laying is delayed by two or more weeks due to the absence of foreign semen (Van Duiven- boden, 1983). (3) Once egg-laying has started, mating reduces fecundity in those lymnaeids studied (De Witt & Sloan, 1958; Van Duivenboden et al., 1985). Under laboratory conditions L. stagnalis commences mating activity at the age of 7-8 weeks (shell height about 18 mm). Egg-laying starts 2-3 weeks later. Once egg-laying has started, the snails show continuous mating and oviposition activity. A general description of the copulation behaviour of L. stagnalis was given by Noland & Carriker (1946) and a more detailed one by Barraud (1957). The latter concluded that copulation was seldom successful. This seems to contradict the studies mentioned above, because it implies that the complex and time-consuming mating behaviour hardly 54 VAN DUIVENBODEN & TER MAAT serves any purpose in reproduction. More- over, Barraud suggested that a systematic study of mating in this snail is hardly possible. In recent years we studied masculinity and receptivity in L. stagnalis (Van Duivenboden & Ter Maat, 1985) and effects of mating on egg-laying (Van Duivenboden, 1983, Van Duivenboden et al., 1985). From these stud- ies a detailed description of the mating behaviour of L. stagnalis was compiled, which is presented here. Copulation behaviour can be induced readily by reunion of snails after a period of isolation (Noland & Carriker, 1946, Rudolph, 1979a). This method was used to describe mating behaviour and the results were com- pared with spontaneous matings as well. The success of copulation was assessed through dissection of female copulants, immediately after mating. Finally a simple operation was performed, which completely eliminates male mating activity. METHODS Laboratory-bred specimens of L. stagnalis were used. They were raised and kept in tanks with continuous water change at a temperature of 20 + 1°C (Van der Steen et al., 1969). A 12/12 light/dark cycle was main- tained with overhead fluorescent lighting. Isolation was performed by placing the an- imals individually in perforated polyethene jars in the tank. They were fed lettuce leaves ad libitum (cf. Scheerboom, 1978). After six or more days of isolation, mating was induced by housing the animals in pairs in clean jars (one pair per jar) filled with 250 ml of fresh aerated tap water in a temperature controlled room (20°C). The snails were marked with nail polish at the tip of the shells to simplify identification during observation. At first the behaviour was observed continuously to de- velop criteria for the analysis of behaviour. Mating behaviour in snails is very slow so two persons can observe adequately 20-25 pairs at a time by brief observations at 1 min inter- vals. Spontaneous mating behaviour was ob- served in the 800 liter breeding tanks (with up to 700 snails per tank) in the laboratory. The animals were fed three times a week, alter- natively lettuce leaves and fish food (Tetraphyll, Tetrawerke A.G.), in restricted amounts. Female copulants were dissected immedi- ately after copulation. The vagina normally looks flaccid and transparant and it is difficult to recognize. A copulation was considered as successful when the vagina had a white, swollen appearance (3-5 times its normal size). In those cases the otherwise transpar- ant duct of the bursa copulatrix was also filled with white material. The animals were anesthetized with MgCl> (Van Duivenboden, 1982). A small cut was made in the body wall to remove some mm of the vas deferens. The vas deferens was interrupted in either one of two places: 1) where it runs freely in the rear sinus, or 2) the part that runs through the body wall (see Fig. 5 for anatomical details). Sham-operated an- imals were treated like the operated ones, except for the cutting of the vas deferens. Recovery required up to 4 hours. Frequency data were tested by means of the G-test after Williams' correction. Analyses were carried out according to Sokal 8 Rohlf (1981). Data were tested for normality with the method of Shapiro & Wilk (1965, 1968) and for homogeneity of variances with the F-max procedure. The terms “male” and “female” refer to the male and female copulant, respectively. RESULTS Isolation-induced mating Mating behaviour of about 750 pairs of snails was observed, in a series of fifty exper- iments. After an isolation period of six days or longer, 75-100% of the pairs showed mating behaviour. Male mating behaviour stood out clearly from all other behaviours exhibited by L. stagnalis. A number of consecutive be- havioural acts could be distinguished in all successful copulations in males, but not in females. Females seemed to behave indiffer- ently for the greater part of the time, moving about, air breathing and feeding during all phases of copulation. Although we will there- fore focus on the male, some reactions of the female will be discussed. Firstly, a description of a straightforward mating sequence, will be given. This will serve as the basic framework for the treatment of all other behaviour ac- companying mating. When the isolated snails are paired, they crawl about in a seemingly random fashion. When they meet, the prospective male POND SNAIL MATING 55 FIG. 1.A. Mounting, male at right. B, C, D. Turning, male at left. E, F. Partial eversion of preputium (arrow), male at left. mounts the prospective female (mounting) and starts rounding the apex of her shell (turning), always in a (characteristic) counter- clockwise manner (seen from above). The male opening, near the base of the right tentacle, becomes visible as a white dot, indicating that the eversion of the preputium had started (partial eversion of preputium) (Fig. 1A-F). When the male reaches the ap- erture of the shell of its partner—at the right hand side—the male comes to an almost complete stop. At this point the head/foot part shortens and gets a swollen appearance. The tentacles are shortened and drooped and the 56 VAN DUIVENBODEN & TER MAAT apex of the shell is kept down. This posture is specific for mating snails. Progress is now very slow, as the right side of the foot of the male travels along the margin of the shell of the female. Meanwhile swellings and contrac- tions occur in the partially everted preputium. Locomotion ceases altogether when the lips and tentacles of the male have passed the pneumostome of the partner. The preputium is then totally everted and the tip makes searching movements under the female's shell. A totally everted preputium is unmistak- able because of its large size (length 10-20 mm, width 5-10 mm in adult snails). The actual eversion of the penis can only occasionally be observed. However, once the female gonopore of the female copulant is occupied by the tip of the preputium of the male, intromission is highly probable (see below: semen transfer). During intromission, undulations of the vas deferens are visible through the transparant wall of the preputium. The preputium is subsequently withdrawn and the male moves away. During mating the male is very firmly attached to the shell of the female: copulating animals can only forcibly be separated. Fig. 2 summarizes the mating behavior. Thick lines in the diagram indicate the basic features of the behavioural sequence as de- scribed above. Thin lines refer to events which may complicate mating. Several mountings, whether or not followed by turning and rarely even by partial eversion of the preputium, may occur, after which the ani- mals separate and start again. The snail which was the first to mount will generally, but not always, become the male. Occasionally reversal of roles occurs during the mounting or turning phase. During the phase of partial eversion of preputium, role reversal is very rare and during the phase of total eversion of the preputium role reversal never occurs. Another complication occurs when the preputium is totally everted before the male is in the right position. Two things may happen then: (1) the preputium makes some search- ing movements under the female's shell, is then partially withdrawn, and the male makes one or more turns followed by a second attempt, or (2) a “sham”-copulation takes place, i.e. the preputium is put under the shell of the female without subsequent intromission and ejaculation. A sham-copulation is generally character- ized by strong withdrawal of the forepart of the female, after which she relaxes again and may resume locomotion or floating, while the preputium remains in place (Fig. 3A-C). This situation continues for 15-60 min, or even for some hours. In most sham-copulations the preputium is placed between the tentacles of the female (frontally) but every position at the margin of the shell is possible. The preputium may even be inserted in the pneumostome of the partner. Sham-copulation comes to an end by partial withdrawal of the preputium. One or more turns are then made, followed by a second attempt. When the male was al- ready in the correct position, the turns may be omitted. Up to two successive sham- copulations may precede intromission. Sham- copulations occur frequently (=50% of the pairs, see Table 1). TABLE 1. Latency from pairing and duration of intromission and the occurrence of sham-copulation in isolation-induced matings. A.: Snails with shell heights of 31-35 mm, aged 13-17 weeks. B: Snails with shell heights of 21-24 mm, aged 9 weeks. —————— — Intromission Latenc Durati Period of у ce isolation Mean + SD Mean + SD (days) N (min) CV? (min) CV? Sham-copulation A. 6 20 158 + 46 0.29 36.4 + 3.2 0.09 50% 8 5 143 = 68 0.47 34/6227 0.08 80% 11 16 135 = 51 0.38 365 +41 0.11 94% 16 13 89 + 31 0.35 36.232 0.09 77% В. 16° 31 95 + 29 0.30 Эт 0.10 62% CV = Coefficient of Variation. “Isolated before the appearance of sexual activity. POND SNAIL MATING 57 snails separated locomotion? ©) (+) mounting? © (+) Turn? ЕС (+) partial eversion © of preputium? total eversion O of preputium? total eversion of preputium furning sham-copulation partial eversion of preputium withdrawal of preputium | | reversal of ES ee ae ee, FIG. 2. Diagram of the mating behaviour of L. stagnalis. Thick lines refer to the basic features of male mating behaviour. Thin lines refer to behaviours which often accompany mating. See text for details. 58 VAN DUIVENBODEN & TER MAAT FIG: 3. A. Contorted retraction of the female (at left). B. Relaxation of the female (at left). C. Fully relaxed female resuming locomotion (lower snail), carrying the male upon her shell. Arrow: unoccu- pied female gonopore of the female, indicating sham-copulation. During intromission the female may mount the shell of the male, while in copulo. This results in an extremely complex posture of both snails (Fig. 4A-E). After intromission is completed in this situation, the male loses foothold and a second mating sequence with the roles reversed takes place. Reciprocal copulation contains all the basic behaviours but the introductory behaviour is shortened. When both snails have acted as male and female in turn (Fig. 4F), no mating attempt occurs for at least 4 hrs. Once total eversion of the preputium is observed, completion of the mating sequence will be the rule, with a few exceptions. Occasionally a female half- way through the copulation sequence climbs some cm above water level and falls down. Then the copulants may become separated. The other exception may occur when the female start ovipositing. Then in some cases the preputium is totally withdrawn and the male moves away. When the deposition of the egg mass is finished, the whole mating sequence may restart, sometimes with the former female in the male role. In other cases the male waits with its preputium partially everted, while attached to the female, until the egg mass is deposited. Then a second, gen- erally successful, attempt follows. During the entire mating sequence, the partners may have mutual mouth contact from time to time. The longer the periods of isolation, the more frequently this behaviour occurs. The time relationships between the main male behaviours—turning, partial and total eversion of preputium and intromission—ap- peared to be variable. In a given experiment, e.g. in some pairs the sequence of mounting, turning and partial eversion of the preputium may take only a few minutes and the total eversion of preputium may occur as much as 90 min later. In other pairs the time between turning and partial eversion of the preputium may take more than an hour whereas it is followed immediately by total eversion of preputium and intromission. In Table | the analysis of the latency from pairing to intromission and the duration of intromission is Summarized for four experiments. The co- efficient of variation (CV) of the latency of intromission (0.29-0.47) appeared to be much higher than that of its duration (0.08-0.11). A clear trend towards decreasing mean latencies with increasing periods of isolation occurs (Р< 0.001, linear regression). No such trend could be found in the mean duration of intromission, which is fairly con- stant (36 + 4 min). The latency and the duration of intromis- POND SNAIL MATING 59 FIG. 4. A. Intromission, male at left. B, С, D. Female (at right) climbing upon the shell of the male while in copulo. Е. Complex posture of the two snails. From left to right: shell of {пе male, head/foot of the female, head/foot of the male, shell of the female. Arrow: preputium of the male. Е. Separation of the snails after both have acted as male and female in turn. 60 VAN DUIVENBODEN & TER MAAT TABLE 2. Presence of semen in the vagina of female copulants after isolation-induced and spontaneous mating. Shell height (mm) Isolation-induced >26 Spontaneous 18-22 23-27 >28 “Pooled observations from four experiments. N semen/ N observed Percentage 59/608 98% 28/32 88% 23/25 92% 23/24 96% TABLE 3. Intromission in isolation-induced mating of Operated (partial removal of vas deferens, body wall) and Sham-operated snails. Reciprocal intromission was excluded. Intromission snails 2e = Test on homogeneity G df P Op x Op 0 6 Overall 9.889 2 0.005 < Р < 0.01 Op x SH 3 3 Op x Op/Op x Sh,Sh x Sh 8.604 1 0.001 < Р < 0.005 Sh x Sh 5 1 Op x Sh x Sh 1.354 1 > 0.4 NS sion of inexperienced snails (Table 1B) agrees very well with that of experienced snails after 16 days of isolation (Table 1A, fourth row). Intromission occurred in 13 out of 18 pairs in the experienced snails (72%) and in 31 out of 40 pairs in the inexperienced snails (77.5%). These data indicate that cop- ulation ability does not depend on prior expe- rience. Spontaneous mating For comparison spontaneous matings in groups of snails were observed. The behaviours described for the isolation- induced matings were also present in spon- taneous ones. Reversal of roles after copula- tion was not observed in these groups, but sometimes a copulating female mounted a third snail and started male behaviour, with the copulating male passively on its shell. Occasionally chains of three snails in copulo were encountered, the upper one acting as a male, the middle one acting as its female partner and as a male copulant for the under- most snail. The undermost female sometimes mounted a fourth snail, but chains of more than three animals in copulo were not ob- served. Sham-copulations were frequently encoun- tered in grouped snails. As in the isolation- induced matings a sham-copulation was gen- erally followed by intromission. This fact was used to determine the duration of intromission in spontaneous copulations. Twelve sham- copulating pairs (shell height 28-33 mm) were followed until the completion of copula- tion. The mean duration of intromission was 35.8 + 3.5 min. This value corresponds with that of the isolation-induced copulations (Table 1). Semen transfer The data in Table 2 show that in most cases transfer of semen took place. Appar- ently, larger snails have more success but the relevant differences are not significant (P > 0.5, G-test for homogeneity). Whether the presence of ejaculate in the vagina prevents insemination by a second male, as in some insects (Parker, 1970), was investigated in the following experiment. Twenty-four snails were paired (12 pairs) after 8 days of isolation. In 10 pairs intromis- sion took place. Immediately after the first mating, the snails were separated to prevent reciprocal copulation. Subsequently 5 pairs consisting of former females were formed and behaviour was observed during the next 150 min. In four pairs intromission took place with durations of 33, 34, 36 and 39 min, respectively. Afterwards the vagina of the ten snails was inspected for semen. In all of them semen was present and in three of the four snails, that had acted twice as a female, the POND SNAIL MATING 61 FIG. 5. Dorsal view of the internal organization of the head/foot part of L. stagnalis. Arrows refer to the parts of the vas deferens that were surgically removed. AP = artery of the penis, BM = buccal mass, CNS = central nervous system, FG = female gonopore, М = mantle, ОЕ = oesophagus, P = preputium, PG = prostate gland, PM = protractor muscles, PN = penis nerve, PS = penis sheath, RM = retractor muscles, VD = vas deferens. amount was much larger than ever observed before, indicating that insemination had oc- curred by the second male. Thus the pres- ence of semen in the vagina does not prevent intromission. Moreover insemination by a second male is highly probable. Elimination of male mating activity In pilot studies attempts were made to block semen transfer by surgical removal of different parts of the vas deferens (see Fig. 5 for anatomical details). Of 16 animals with part of the vas deferens in the head sinus removed, 8 animals sur- vived (50%). Eight days after the operation the animals showed male behaviour (i.e. par- tial eversion of preputium), but total eversion of preputium and intromission were impaired. Dissection revealed that the loose ends of the vas deferens had grown into the body wall, making total eversion of preputium and intromission impossible. Ten animals with the part of the vas deferens that runs through the body wall removed, all survived (100%). Animals oper- ated this way not only lacked the possibility of semen transfer, but also failed to show mating activity, even when they were paired after three weeks of isolation. Therefore a more thorough study was made of the effects of this lesion. The operation was done in 18 snails (Op) and 18 snails were sham-operated (Sh). Af- terwards the snails were kept in isolation during a period of 8 days. They were divided in three experimental groups: 6 pairs of oper- ated snails (Op x Op), 6 pairs consisting of an operated and a sham-operated snail (Op x Sh) and 6 pairs of sham-operated snails (Sh x Sh). The behaviour of the pairs was observed during 330 min. In all pairs mounting was observed. No differences in the latency of mounting be- tween the groups were observed (one-way ANOVA, on log transformed data, 0.10 < P< 0.25, NS). In the Op x Sh-group all first mountings were made by the sham-operated snail. In the Op x Op-group no copulation behaviour followed, except rarely an incom- plete turn. In all pairs of the other groups male behaviour was exhibited, but only by the sham-operated snails. The number of intromissions in the groups with at least one sham-operated snail in the pairs was signifi- cantly higher than that in the Op x Op-group (Table 3). The operated snails did not initiate mating behaviour, but it is conceivable that copula- tion as a female could induce male behaviour in these snails (cf. Stagnicola elodes, Rudolph, 1979a). This hypothesis was re- jected by the total absence of any sign of reciprocal behaviour in the operated snails after copulation as a female, whereas the sham-operated snails all exhibited male behaviour after copulation as a female (G-test P < 0.05). In all females—operated or sham-operated—semen was present in the vagina. Removal of part of the vas deferens that runs through the body wall clearly eliminates all male behaviour, but it does not impair the ability to copulate as a female. 62 VAN DUIVENBODEN & TER MAAT DISCUSSION The mating behaviour of L. stagnalis is unilateral: one snail acts as the male and the other as the female. This is the general way of mating in lymnaeid snails (L. peregra: Diver et al., 1925; L. tomentosa: Boray, 1964; $. elodes: Rudolph, 1979a; L. truncatula: Smith, 1981). Characteristic female mating behaviour is absent, whereas male mating behaviour is clear and unmistakable. The basical features are: mounting, turning, eversion of preputium, intromission, withdrawal of preputium and moving away. Several variations are present of which sham-copulation and reciprocation are the most remarkable ones. In many respects our description of the mating behaviour of L. stagnalis is in agree- ment with the results of Barraud (1957), but there are some contradictions. Firstly, Bar- raud suggested that, in addition to the copula- tory position that we described (position 1, Barraud), intromission is possible from a fron- tal position (position 2, Barraud). Secondly, at the moment of actual penetration, we did not observe any reaction of the female, whereas he described a contorted retraction of the whole forepart of the female's body at intromission, like we observed in sham-copulations. Thirdiy the duration of intromission was 36 + 4 min in our experi- ments, whereas Barraud reported durations of a few min to twelve hours. A fourth differ- ence relates to the success of copulation, which was nearly 100% in our study and low in his. All these differences share a common cause: Barraud did not make a clear distinc- tion between sham-copulation and real copu- lation. A sham-copulation does resemble a real copulation, but an experienced observer is able to distinguish the two by the charac- teristic female reaction in sham-copulation. Moreover, occupation of the vagina can in most cases—e.g. with the aid of a mirror—be observed. The high incidence of sham-copulation probably explains all contra- dictions between our observations and those of Barraud. Reciprocal copulation behaviour was de- scribed extensively for S. elodes (Rudolph, 1979a). The readiness to exhibit male behaviour, induced in female copulants, lasts 30-60 min in this snail. When stimulated females are transferred to a third snail during this period, they behave as males. In groups the induced male behaviour of female copulants is probably directed towards a third snail rather than to the partner, since chain copulations but no reversal of roles were observed in groups of snails. Mouth contact was sometimes encoun- tered during the mating sequence of L. stagnalis. lt is a common feature of mating behaviour in snails. In helicids (terrestrial pulmonates) courtship commences with mouth to mouth contact (Lind, 1976) and it is a characteristic part of the mating behaviour of the opisthobranch Aeolidia papillosa (Longley & Longley, 1984). In L. stagnalis it is not an integrated part of mating behaviour. The mating behaviour of L. stagnalis can be broken off in the first stages of the sequence (mounting, turning and occasionally partial eversion of preputium). Once the preputium is totally everted, the sequence will come to completion, although this may take hours. Mating capability in L. stagnalis depends on maturation only, not on prior experience (Ta- ble 1). This has been found earlier in Biomphalaria globosus (Rudolph, 1983). The duration of intromission was found to be in- dependent of the experience or the period of isolation of the snails. Probably the duration is determined by neuronal timing circuitry, as is assumed to be the case in A. papillosa (Longley & Longley, 1984). After copulation as a female, the readiness to mate as a male as well as a female remains the same. The readiness to mate disappears when both snails have acted as male and female in turn. These observations as well as the decrease in the latency of intromission with increasing period of isolation is in accor- dance with our model of masculinity and receptivity (Van Duivenboden & Ter Maat, 1985). Extirpation of the part of the vas deferens that runs through the body wall eliminates all male behaviour in L. stagnalis. As yet it is not clear whether this is due to neurological, endocrinological or mechanical blockade. Lit- tle is known of mechanisms controlling mating behaviour in other snails. Jeppesen (1976) extirpated various parts of the reproductive system of Helix pomatia, but the initiation and the sequence of mating behaviour were not affected. He concluded that mating behaviour is controlled by the central nervous system. The cycle of the mating behaviour in helicids (Lind, 1976, Jeppesen, 1976) seems to de- pend partly on copulation itself, as in L. stagnalis (Van Duivenboden & Ter Maat, 1985) and on mechanical effects of dart- POND SNAIL MATING 63 shooting, a behaviour not present in L. stagnalis. The organization of the male and the female reproductive system of L. stagnalis (diaulic) is different from that in Helix (monaulic) (Visser, 1977, 1981, Geraerts & Joosse, 1984, Tompa, 1984). Therefore a comparison of the extirpations carried out by Jeppesen in Helix pomatia with the lesion carried out in this study in L. stagnalis is not in order. In almost all cases, copulation appeared to be successful, 1.е. semen could be observed in the vagina of the female. Similar results were found for S. elodes (Rudolph, 1979a) and for B. globosus (Rudolph, 1979b, 1983). As in Bulinus, the presence of semen in the female tract does not prevent intromission and insemination by a second male. ACKNOWLEDGEMENTS The authors thank Dr. W.J. van der Steen and Prof. Dr. T.A. de Vlieger for critical read- ing of the manuscript, Anton Pieneman for technical assistance, photography and pre- paring of the figures, and Thea Laan for typing the manuscript. REFERENCES CITED BARRAUD, E. M., 1957, The copulatory behaviour of the freshwater snail (Lymnaea stagnalis L.). British Journal of Animal Behaviour, 5: 55-59. BORAY, J. C., 1964, Studies on the ecology of Lymnaea tomentosa, the intermediate host of Fasciola hepatica. Australian Journal of Zoology, 12: 231-237. BROWN, K. M., 1979, Effects of experimental manipulations on the life history pattern of Lymnaea stagnalis appressa Say (Pulmonata: Lymnaeidae). Hydrobiologia, 65: 165-176. CAIN, G. L., 1956, Studies on cross-fertilization and self-fertilization in Lymnaea stagnalis appressa Say. Biological Bulletin (Marine Biological Labo- ratory, Woods Hole), 111: 45-52. DIVER, C., BOYCOTT, А. E. 8 GARSTANG, S., 1925, The inheritance of inverse symmetry in Limnaea peregra. Journal of Genetics, 15: 113-200. DUIVENBODEN, Y. A. van, 1982, Non-ocular photoreceptors and photo-orientation т the pond snail Lymnaea stagnalis (L.). Journal of Compar- ative Physiology, 149: 363-368. DUIVENBODEN, Y. A. van, 1983, Transfer of se- men accelerates the onset of egg-laying in fe- male copulants of the hermaphrodite freshwater snail, Lynmaea stagnalis. International Journal of Invertebrate Reproduction, 6: 249-257. DUIVENBODEN, Y. A. van & TER MAAT, A., 1985, Masculinity and receptivity in the hermaphrodite pond snail Lymnaea stagnalis. Animal Behaviour, 33: 885-891. DUIVENBODEN, Y. A. van, PIENEMAN, A. W. & TER MAAT, A., 1985, Multiple mating suppresses fecundity in the hermaphrodite freshwater snail Lymnaea stagnalis: a laboratory study. Animal Behaviour, 33: 1184-1191. DUNCAN, С. J., 1975, Reproduction. In Pulmon- ates, FRETTER, V. & РЕАКЕ, J., eds., vol. 1, Academic Press, London, pp. 309-365. GERAERTS, W. P. M. & JOOSSE, J., 1984, Fresh- water snails (Basommatophora). In The Mollusca, WILBUR, K. M., ed., vol. 7, Reproduc- tion, Academic Press, New York, pp. 142-208. HORSTMANN, H. J., 1955, Untersuchungen zur Physiologie der Begattung und Befruchtung der Schlammschnecke Lymnaea stagnalis L. Zeitschrift für Morphologie und Ökologie der Tiere, 44: 222-268. JEPPESEN, L. L., 1976, The control of mating behaviour in Helix pomatia L. (Gastropoda: Pulmonata). Animal Behaviour, 24: 275-290. JOOSSE, J. & GERAERTS, W. P. M., 1983, Endo- crinology. In The Mollusca, WILBUR, К. М. 4 SALEUDDIN, A. S. M., eds., vol. 4, part 1, Physiology, Academic Press, New York, pp. 317—406. LIND, H., 1976, Causal and functional organization of the mating behaviour sequence in Helix pomatia (Pulmonata, Gastropoda). Behaviour, 59: 162-202. LONGLEY, А. J. & LONGLEY, В. D., 1984, Mating in the gastropod mollusc Aeolidia papillosa: behaviour and anatomy. Canadian Journal of Zoology, 62: 8-14. NOLAND, L. E. 4 CARRIKER, M. R., 1946, Obser- vations on the biology of the snail Lymnaea stagnalis during twenty generations in laboratory culture. American Midland Naturalist, 36: 467-493. PARAENSE, W. L., 1956, A genetic approach to the systematics of planorbid molluscs. Evolution, 10: 403-407. PARAENSE, W. L., 1959, One-sided reproductive isolation between geographically remote popula- tions of a planorbid snail. American Naturalist, 93: 93-101. PARKER, G. A., 1970, Sperm competition and its evolutionary consequences in the insects. Bio- logical Review, 45: 525-567. RICHARDS, C.S., 1970, Genetics of a molluscan vector of Schistosomiasis. Nature, 227: 806-810. RUDOLPH, P. H., 1979a, The strategy of copula- tion of Stagnicola elodes (Say) (Basom- matophora: Lymnaeidae). Malacologia, 18: 381-389. RUDOLPH, P.H., 1979b, An analysis of copulation 64 VAN DUIVENBODEN & TER MAAT in Bulinus (Physopsis) globosus (Gastropoda: Planorbidae). Malacologia, 19: 147-155. RUDOLPH, P.H., 1983, Copulatory activity and sperm production in Bulinus (Physopsis) globosus (Gastropoda: Planorbidae). Journal of Molluscan Studies, 49: 125-132. SCHEERBOOM, J. Е. M., 1978, The influence of food quantity and food quality on assimilation, body growth and egg-production in the pond snail Lymnaea stagnalis (L.) with particular ref- erence to the haemolymph-glucose concentra- tion. Proceedings of the Koninklijke Nederlandse Akademie van Wetenschappen (С), 81: 184-187. SHAPIRO, S. 5. & WILK, М. B., 1965, An analysis of variances test for normality. Biometrica, 62: 591-611. SHAPIRO, S. S. & WILK, M. B., 1968, Approxima- tions for the null distribution of the W statistic. Technometrica, 10: 861-866. SMITH, G., 1981, Copulation and oviposition in Lymnaea truncatula (Müller). Journal of Mollus- can Studies, 47: 108-111. SOKAL, В. В. & ROHLF, Е. J., 1981, Вютейу. Freeman, San Francisco. STEEN, W. J. van DER, HOVEN, М.Р. VAN DEN & JAGER, J. C., 1969, A method for breeding and studying freshwater snails under continuous wa- ter change, with some remarks on growth and reproduction in Lymnaea stagnalis (L.). Nether- lands Journal of Zoology, 19: 403—468. TOMPA, A. S., 1984, Land snails (Stylom- matophora) |. In The Mollusca, WILBUR, К. M., ed., vol. 7, Reproduction. Academic Press, Lon- don, pp. 47-140. VISSER, M., 1977, The morphology and signifi- cance of the spermoviduct and prostate in the evolution of the reproductive system of the Pulmonata. Zoologica Scripta, 6: 43-54. VISSER, M., 1981, Monauly versus diauly as the original condition of the reproductive system of Pulmonata and its bearing on the interpretation of the terminal ducts. Zeitschrift für zoologische Systematik und Evolutionsforschung, 19: 59-68. WITT, В. M. DE 8 SLOAN, W. C., 1958, The innate capacity for increase in numbers in the pulmon- ate snail Lymnaea columella. Transactions of American Microscopical Society, 77: 290-294. Revised Ms. accepted 26 June 1986 MALACOLOGIA, 1988, 28(1-2): 65-79 REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCH GASTROPODS FROM THE WESTERN NORTH ATLANTIC Faigel K. Vale & Michael A. Rex' Department of Biology University of Massachusetts at Boston Boston, MA 02125, U.S.A. ABSTRACT We estimated the frequency of repaired shell damage in prosobranch assemblages collected from the continental shelf, upper and lower continental slopes, continental rise and abyssal plain south of New England, U.S.A. Damage was classified as either major (conspicuous breaks generally resulting in displacement of subsequent growth patterns and interruption of sculpture), or minor (fine discontinuities that do not disrupt growth). There is significant variation among regions in the incidence of major and minor damage to prosobranch individuals, but no clear trend with depth. The incidence of species that contain damaged individuals appears to be uniform throughout the 5 regions. Frequencies of major repaired damage in the deep sea (0.08-0.48, median 0.15 for samples > 200 m in depth) fall within the range of values reported for a variety of shallow-water marine habitats. Known predators of deep-sea snails include fishes, decapods and echinoderms. Their stomach contents indicate very general feeding habits and broad diets. Deep-sea snails and their predators show less evidence of coevolved adaptations than do their shallow-water counterparts. Key words: deep sea; prosobranchs; predation; repaired shell damage; coevolution. INTRODUCTION Biological disturbance by predation has been proposed as a cause of community structure in the deep-sea benthos (e.g. see reviews in Rex, 1981, 1983; Jumars & Gal- lagher, 1982; Jumars & Eckman, 1983), but its actual importance has been very difficult to determine. One useful approach to studying the effects and geographic patterns of preda- tion in shallow-water faunas has been to measure the incidence of repaired shell dam- age in gastropods (Vermeij, 1978, 1982a; Vermeij et al., 1982; Bertness & Cunningham, 1981). Gastropods are preyed upon by fishes and decapod crustaceans that are special- ized to break open shells to consume the soft parts (Zipser & Vermeij, 1978; Palmer, 1979; Bertness et al., 1981). Unsuccessful preda- tion attempts can result in shell breakage that is repaired by the snail, leaving a distinctive scar. The record of sublethal shell damage in a population is not correlated in a simple direct way with either the intensity or effectiveness Please correspond with М. A. Rex. (65) of predation (Schindel et al., 1982). Establish- ing the exact relationship between predation pressure and the incidence of shell repair requires information on the age structure, reproductive pattern and survivorship of prey, the contribution of predation to overall mortal- ity (Schoener, 1979), the relative abundance of predator and prey, the strength of preda- tors, and the ability of prey to resist or avoid predation (Vermeij, 1982a, 1983). Few such data exist for deep-sea species. Without di- rect evidence on predator-prey interactions, a low incidence of repair is especially difficult to interpret: predators could be scarce, or, con- versely, abundant and extremely efficient at killing prey leaving few scarred individuals, or rarely able to break shells (Schindel et al., 1982; Vermeij, 1982a). In coastal environ- ments, high frequencies of repair generally have been associated with a coevolved predator-prey system in which snails are ex- periencing potentially lethal predation and have evolved shell architecture to deter it (Schindel et al., 1982; Vermeij, 1982a, 1982b, 1983). Based upon the scanty evidence available, 66 VALE & REX TABLE 1. Station data for samples used in the analysis of repaired shell damage in deep-sea gastropods. N is the number of individuals, and S the number of species examined in each station. Frequencies of repaired shells, calculated as the percentage of repaired shells, are given for individuals (%N that show repair) and species (%S, some individual of which shows repair). Region Station Latitude Longitude Continental shelf 89 40°1.6'N 70°40.7'W Upper slope 88 39°54.1'N 70°37'W 105 3956.6'N 71°03.6 W Lower slope 73 39°46.5'N 70°43.3’W 103 39°43.6'N 70°37.4'W 131 39°38.8'N 70°36.8'W Continental rise 76 39°38.3'N 67°57.8'W 126 39°37.3'N 66°45.5'W TAU 38°0.7'N 69°16.0'W 85 37°59.2'N 69°26.2'W Abyssal plain 123 37°29.0'N 64°14.0’'W 124 37°25.5'N 63°58.8'W Frequency of repaired shells %individuals %species Depth SS (m) N S Major Minor Major Minor 196 46 5 004 037 0.40 0.80 478 158 18 0.08 022 028 0.72 530 480 10 0.14 026 0.80 0.70 1400 162 14 015 045 043 0:79 2022 188 17 0.48 0:72 0538806 2178 38 5 024 0.42 0801080 2862 63.15 0:21 0.35 0.47 0.60 3806 66 17 0.20 0.29 047 (053 3806 109 12 0.13 0.39 0.58 0.92 3834 239 25 0.13 0.35 0.44 0.64 4853 177 9 0.19 0.34 067 078 4862 112 7 0.14 0.41 0:29) (0:57 Vermeij (1978) cautiously suggested that pre- dation by crushing was unlikely to be impor- tant in the deep sea because deep-sea mol- lusks are small and have poorly developed antipredator armor, and their potential preda- tors apparently lack specialized shell- crushing adaptations. However, in this first analysis of repaired shell damage in deep- sea gastropods, we show that repaired shells are common at bathyal and abyssal depths in the western North Atlantic and occur in most prosobranch species. This suggests that bio- logical and/or physical sources of shell dam- age are common features of the deep-sea environment. MATERIALS AND METHODS The gastropod material We analyzed the gastropod fraction of 12 epibenthic sled samples (Hessler & Sanders, 1967) collected from the Gay Head-Bermuda Transect south of New England (Sanders et al., 1965). We included only prosobranch species because their early whorls are ex- posed and can be examined for repaired damage. In most deep-sea opisthobranchs (e.g. Scaphander, Cylichna, Retusa) the early whorls are completely obscured by the body whorl, so that sublethal damage in early life cannot be observed. Station data and sample sizes are given in Table 1; maps of sampling localities can be found in Rex (1973, 1976). One sample is from the outer continental shelf, and 11 are from the deep sea (>200 m). Species lists can be found in Rex & Warén (1982). A small number of speci- mens have been used for other purposes and some shells were too corroded to assess repaired shell damage: we report on 88% of the individuals in the original collections. In total, the material comprised 1838 individuals distributed among 79 species. Only live- collected specimens were used. The com- plete raw data used in the analysis are pro- vided in the Appendix. Scoring of repaired shell damage Identification of predator-induced shell breakage has been necessarily subjective, and has focused on conspicuous damage with clear effects on subsequent growth (see especially Schindel et al., 1982). Vermeij (1982a: 565) distinguished a scar from a normal “'growth line’ (interruption of shell growth) by its irregular, usually jagged trajec- tory”; similar definitions are found throughout the literature on repaired shell damage. Schindel et al. (1982) counted repairs in or- REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 67 FIG. 1. Examples of major and minor repaired shell damage in prosobranch snails collected from the deep sea of the western North Atlantic. See text for explanation of the injuries. Major damage: a. Frigidoalvania brychia (Verrill, 1884), station 105, 3.5 mm; b. Oenopota bergensis (Friele, 1886), station 103, 10.3 mm; c. Mitrella pura (Verrill, 1882), station 105, 4.0 mm; d. Aclis walleri (Jeffreys, 1884), station 103, 3.7 mm. Minor damage: e. Brookula capensis Clarke, 1961, station 85, 2.2 mm; f and g. Mitrella pura (Verrill, 1882), station 88, 4.0 mm and station 105, 3.5 mm respectively. Sizes represent shell height. Station data are provided in Table 1. namented species if the sculptural pattern was distorted. Reimchen (1982: 688) catego- rized breaks in Littorina by “extent of injury, from minor disruption of shell growth through major breakages”. We compiled data on the frequency of both major and minor breaks. Some examples of the kinds of repaired damage that we encountered are shown in Fig. 1. We classified as major breaks any clear discontinuities in the normal growth pattern of the shell that resulted in temporary post-break displacement of sculptural fea- tures (where existing), and large fractures showing evidence of breaking back from the aperture. Fig. la shows a displacement of the shoulder ridge by a break at the end of the second whorl. The break is oriented diagonally to the generating curve of the aperture (sensu Raup, 1966) and to normal growth lines. Fig. 1b shows a deep break at the end of the third adult whorl that has altered the spacing and shape of post-break axial ribs. Figs. 1c and 1d show jagged irregular breaks that deviate markedly from the shape of normal growth lines, indicating that the lip of the aperture had been broken 68 VALE & REX back. In most major breaks (e.g. Figs. 1b, с, d) the broken edge is imbricated over the area of resumed growth. Major breaks correspond to the same type of substantial injury that is known to be predator-induced in shallow-water species. Minor breaks included less extreme dam- age that caused no obvious distortion of post- break growth. Fig. 1e exhibits a discontinuity of growth, but very little disruption of spacing in either axial or spiral sculpture. Fig. 1f shows a fine irregular fracture without notice- able imbrication. Fig. 1g shows a small break that extends only about one-third of the way across the body whorl. We did not count breakage to the lip of the present aperture (see e.g. Fig. 1g) because this sometimes results from damage incurred during either dredging or sorting and, moreover, is not “repaired”. Minor breaks have been noted, but not included, in analyses by other investigators (e.g. Reimchen, 1982; Vermeij et al., 1980; Vermeij, 1982b; Schindel et al., 1982). They could result from ineffective handling of prey by predators, but could also be due to a variety of physical disturbances, especially in high-energy coastal environments. We felt that the minor breaks were especially inter- esting in deep-sea species. The deep milieu was long assumed to be very physically and biologically stable, and this view has had important implications for theories of commu- nity structure of the deep-sea benthos (Rex, 1981, 1983). However, recent evidence sug- gests that the deep sea of the western North Atlantic is much more physically and biologi- cally dynamic than once supposed (e.g. Deuser & Ross, 1980; Richardson et al., 1981; Gardner & Sullivan, 1981; Bulfinch et al., 1982; Thistle et al., 1985; Deuser, 1986). Any evidence on variation in the lives of individual organisms bears on the issue of stability in the deep sea. The Appendix gives the number of shells that showed either major or minor damage for each species. Although there is a continuum in the sever- ity of repaired damage, we had surprisingly little difficulty scoring breaks as either major or minor. These categories appear to repre- sent two different modal tendencies. Incidence of repair The most common measure to quantify the incidence of shell repair has been the average number of scars per shell, calculated as the number of repaired injuries divided by the total number of individuals examined (Vermeij, 1982b). This measure has often been standardized to either size classes of individuals or subsets of whorls, depending on the objectives of the study and limitations of shell form (cf. Currey & Kohn, 1976; Vermeij et al., 1980; Vermeij et al., 1981; Vermeij, 1982a,b; Schindel et al., 1982; Vermeij et al., 1982; Shimek, 1983, 1984). We defined frequency of repair somewhat differently as the percentage of repaired shells, which is computed as the number of individuals having at least one repair divided by the total number of individuals in the sample. This measure has been used by Raffaelli (1978), Elner & Raffaelli (1980), Geller (1983), Bergman et al. (1983) and others. It is a more conservative estimate of the frequency of repaired damage than the average number of scars per shell because shells can survive injury more than once (see especially Currey 8 Kohn, 1976; and Shimek, 1983, 1984). We felt that it was a more appropriate measure for our study because of the tremendous variation in shell form and numerical abundance among the 79 species studied, and because our aim was to get a preliminary overview of the frequency of repaired damage in the deep-sea environ- ment. We computed frequencies for major breaks, minor breaks and combined breaks (either major or minor with redundancies eliminated), for individuals and species. The twelve stations in Table 1 were grouped for analysis into the five biogeo- graphic assemblages identified by Rex (1977) in a multivariate study of gastropod species composition with depth. These as- semblages correspond to five bathymetric regions: the continental shelf (<200 m), upper (200-1000 m) and lower (1000-2000 m) continental slopes, continen- tal rise (2000-4000 m), and abyssal plain (>4000 m). We used a chi-square test with raw data (Siegel, 1956) to test for significant differences between regions in the number of repaired individuals. Fisher-Yates exact prob- abilities (using the tables of Finney et al, 1963) were used to test for significant differences between regions on the species level, because numbers of species were too low in some regions for valid use of chi-square testing. We also regressed the incidence of breakage in individual samples against depth. REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 69 2.57 00 nin SHELF au U. SLOPE L. SLOPE RISE MAJOR a o 20 PERCENTAGE OF INDIVIDUALS > eee 108.46 is he [28 | MINOR COMBINED 388 FIG. 2. The percentage of prosobranch individuals that exhibit major, minor and combined (either major or minor with redundancies removed) repaired shell damage in five bathymetric regions of the western North Atlantic. Station data for samples in the regions are given in Table 1. Numbers above the bars represent the number of individuals examined in each region. The tables above the histograms contain the chi-square values (for analyses using raw data) for all possible comparisons of the incidence of damaged individuals among the regions. Columns in the table correspond to the histogram bars directly below them. One, two and three asterisks indicate chi-square values that are significant at the Р < .05, Р < .01 and P < .001 levels respectively. RESULTS AND DISCUSSION Bathymetric patterns in the incidence of repair Frequency distributions of individuals with major, minor and combined breaks for the five depth regions are shown in Fig. 2. The lower continental slope shows а conspicuously and significantly (P<.001) higher incidence (32%) of major breaks among individuals than the other four regions (4-17%). However, this is largely attributable to just one of the lower slope stations (sta. 103), which by itself has a frequency of 48%. Without station 103, the lower slope frequency is 17%, and the only remaining significant difference is between the abyss and the shelf (P<.05). When the samples are considered individually, rather than being grouped into regions, there is no significant relationship between the frequency of major breaks and depth of the samples by using either linear regression (Y = 15.198106 + 0.000993Х, г = 0.1518, df. = 10, P>.10) or parabolic regression (Y = 4.057192 + 0.015960X — 0.000003X?, r = 0.5983, d.f. = 9, P>.05). Minor breaks among individuals are roughly twice as common as major breaks from the upper slope to the abyss and about ten times more common on the shelf (Fig. 2). The lower slope shows a higher frequency of minor breaks (58%) than the other four re- gions (25-37%). Again this is partly due to station 103 which has a frequency of 72%; although when station 103 is removed from the analysis, significant differences persist between the lower and upper slope (P<.001) 70 VALE & REX 7 MAJOR | MINOR COMBINED 5 80- 80 X. un 0177 44 Lo | a 11 a 60- AT п 607 u Su er 44 | © 443 40- ч = = _ LJ oO в 20- 20- а. SSS & SSS EE soe 2 Soe. FIG. 3. The percentage of prosobranch species that exhibit major, minor and combined (either major or minor with redundancies removed) repaired shell damage in five bathymetric regions of the western North Atlantic. Station data for samples in the regions are given in Table 1. Numbers above the bars represent the number of species examined in each region. The incidence of species showing damage was compared among the regions by using Fisher-Yates exact probabilities (tables of Finney et al., 1963). There were no significant differences among regions for any of the three categories of repaired damage. and between the lower slope and continental rise (P<.05). The upper slope has a signifi- cantly lower frequency than either the rise or the abyss (P<.001). Neither linear nor para- bolic regressions reveal any pattern in the frequency of minor breaks with depth when individual samples are analyzed (Y = 36.837998 + 0.000485Х, г = 0.0648, d.f. = 10, P>.10; and; Y = 27.330179 + 0.013258X — 0.000003Х2, г = 0.4368, d.f. = 9, P>.10 respectively). Clearly, the combined distribution is influenced most by the inci- dence of minor breaks. The percentage of species showing re- paired shell damage is shown in Fig. 3. There are no significant differences among any of the regions for either major, minor or com- bined breaks. Similarly, linear and parabolic regression models show no pattern in the frequency of either major or minor breaks with depth when individual samples are analyzed (highest r = 0.2935, P>.10). Most species in all of the regions contain individuals with some degree of repaired damage, which is especially remarkable since 54% of the spe- cies in the collections studied are represented by five or fewer individuals, and 26% of the species are represented by only a single individual. In addition to the previously-mentioned problems in assessing and measuring ге- paired damage, there are other difficulties with interpreting its incidence along depth gradients. One is that physiological activity rates may decline exponentially with depth. For example, Smith (1978) showed that benthic community respiration drops three orders of magnitude from the continental shelf to the abyss in the western North Atlantic. If growth rates are correspondingly slower and longevity higher at greater depths (e.g. Turek- ian et al., 1975), then snails from deeper regions have longer to experience shell dam- age. The effect of this, using our method, would be to overestimate progressively the actual frequency of repair with increasing depth. It is unknown whether rates of inflicting damage and growth rates of snails vary in some proportional way with depth. Another complication is that faunal density and biomass vary with depth and differ among taxa and ecological assemblages (Rex, 1983). It is impossible to say with any preci- sion how variation in community structure is REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 71 related to intensity of predation on snails by various potential predators. Since there is no direct observational evidence on predator- prey interactions, we cannot address critically the behavioral implications of “unsuccessful predation” (see Vermeij, 1982c, 1985; Sih, 1985). The complexity and uncertainty of the situation limit us to fairly general conclusions. Interregional comparison of repaired damage How do frequencies of shell repair found in deep-sea gastropods compare with frequen- cies observed in shallow-water environ- ments? There are many difficulties with mak- ing such comparisons in a critical way, including differences in sample sizes, num- bers of species studied, habitat type, shell architecture among the species and methods used to assess the frequency of repair. Only major breaks can be considered, since minor breaks have been excluded from other stud- ies. Regional variation in frequency of repair can be estimated by comparing medians and ranges of frequencies among sampling sites. Table 2 provides these data for living temper- ate and tropical coastal faunas. The deep-sea (>200 m) samples have a median frequency of 0.15 and a range of 0.08-0.48 (Tables 1 and 2). If the outer shelf sample (sta. 89) is included, the median is 0.145. Much of what is known about shell repair in the temperate North Atlantic comes from extensive studies on species of Littorina, a rocky intertidal group that is subject to predation by durophagous crabs (especially Carcinus maenas), and to crushing by boul- ders. Medians for repair frequency in littorinids are =0.09. Median frequencies in Pacific temperate prosobranchs appear to be higher, but these populations have not been sampled as extensively and the medians are well within the ranges of values reported for Atlantic littorinids. Temperate terebrids show a median of 0.31. Terebrids have many- whorled, tall-spired shells. Populations can have frequencies of repair that are an order of magnitude higher than less-turreted species. Terebrid shell shape is apparently an adapta- tion to thwart predation; snails can withdraw up into the shell beyond the crabs’ ability to peel back from the aperture (Vermeij et al., 1980). The median for all of these temperate populations, including the terebrids, is 0.09 (range 0-0.96). Although data from Vermeij's (1982a) extensive analysis of Littorina littorea make up most of the sample (and therefore might be expected to strongly affect the cal- culated median), when this study is omitted the median remains very similar at 0.10. The median without L. /ittorea and the terebrids is 0.08 (range = 0-0.50). Frequencies in tropical populations tend to be higher. This can be seen for the thaidids (Vermeij, 1978), for which data were not presented to enable us to calculate medians, but which show a Clear shift in range to higher values at tropical sites. The tropical study most comparable to the deep-sea data pre- sented here and the temperate data dis- cussed above is Vermeij’s (19825) analysis of snail faunas from 14 localities in the Pacific. The median frequency is 0.28. The overall pattern to emerge is an increase in median frequency of repair from temperate coastal environments (0.10) to the deep sea (0.15) to tropical environments (0.28). A more conser- vative conclusion is that the deep-sea gastro- pod fauna shows frequencies of repaired shell damage that fall within the range of those found in coastal faunas. Predation on deep-sea gastropods Gastropods have been found in the stom- ach contents of many deep-sea fishes including clupeoids (Mauchline & Gordon, 1983), chimaeriformids, halosaurids, gadids, zoarchids (Sedberry & Musick, 1978) and macrourids (Haedrich & Polloni, 1976; McLel- lan, 1977; Mauchline & Gordon, 1984). Deep- sea demersal and benthopelagic fishes that rely on benthic prey tend to be highly euryphagous (McLellan, 1977; Sedberry & Musick, 1978). Gastropods are not common prey items, usually making up about one percent or less of prey individuals when they are found at all in fish stomach contents. Bright (1970), in his study of the stomach contents of 36 species of deep-sea fishes, expressed surprise that gastropods com- prised such a small proportion (one percent) of prey items. The explanation is that deep- sea Snails live at low density. In a quantitative sampling study of the deep-sea benthos south of New England (>200 m), snails were encountered in 56% of the samples taken, and made up only 0.1-2.0% (median 0.4%) of the macrobenthos in samples where they occurred (Sanders et al., 1965). Their inci- dence in stomach contents of deep-sea fishes corresponds roughly to their availability. The presence of shell fragments in stomach con- 72 VALE & REX TABLE 2. Comparison of frequencies of repaired shell damage among deep-sea, temperate coastal and tropical coastal gastropod faunas. PRS and ANS indicate percentage of repaired shells and average number of scars respectively; see text for explanation. The median frequency of repair is calculated among collecting sites for each study. Frequency of repair # of Reference Region Habitat Fauna sites Method Median Range This study Deep-sea Soft bottom 79 Species 11 РАЗ 0.15 0.08-0.48 Western North Atlantic Geller Temperate Rocky Tegula funebralis 3 PRS 0.29 0.04—0.50 (1983) Eastern North Intertidal Nucella emarginata 4 PRS 0.10 0.06—0.20 Pacific Bergman Temperate Seagrass Alia carinata 2 PRS 0.25 0.15—0.36 et al. Eastern North Beds (1983) Pacific Raffaelli Temperate Rocky Littorina rudis 24 PRS 0.06 0-0.48 (1978) Eastern North intertidal Atlantic Elner & Temperate Rocky Littorina rudis SEE RRS 0.06 0.03-0.32 Raffaelli Eastern North intertidal Littorina nigrolineata SPAS 0.08 0.07-0.11 (1980) Atlantic Reimchen Temperate Rocky Littorina mariae 5 PRS 0.09 0.05-0.46 (1982) Eastern North intertidal Littorina obtusata BARS 0.02 0.01-0.44 Atlantic Vermeij Temperate Rocky Littorina littorea 186 ANS 0.08 0-0.59 (1982а) North Atlantic intertidal Vermeij Tropical Pacific Soft bottom 53 species 14 ANS 0.28 0-0.82 (1982b) Vermeij Tropical Indo- Soft bottom Terebridae (61 spp.) 144 ANS 0.72 0-9.19 et al. Pacific and (1980) Atlantic Temperate Soft bottom Terebridae (14 spp.) 20 ANS 0:31 0-0.96 Pacific and Atlantic Vermeij Tropical Rocky Thaidid snails (4 spp.) 17 ANS — 0.09-0.63 (1978) Eastern intertidal Pacific Temperate Rocky Thaidid snails (8 spp.) 21 ANS = 0-0.29 Eastern intertidal Pacific tents (Sedberry & Musick, 1978) indicates that fish predation can cause shell damage. Decapods are poorly represented in the deep sea, compared to coastal waters (Hessler & Wilson, 1983), but Lagardere (1977a,b) has shown that mollusks, including snails, are common (0-41%, median = 11%) prey items in the stomach contents of many deep-sea decapods. Snails are eaten by deep-sea lobsters, and members of three families of deep-sea shrimps (Penaeidae, Pandalidae and Crangonidae). There are brachyuran crabs living on the continental slope south of New England which could crush snail shells in the same way that Carcinus maenas crushes Littorina (e.g. Vermeij, 1982a). However, except for the red crab Geryon quinquidens, whose bathymetric range extends to 1670 m, most species occur at depths less than 600 m (Wenner & Boesch, REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 73 1979). Lagardere (1977a,b) found bivalves in the stomachs of deep-sea Geryon and Pagurus, suggesting that other hard-shelled prey could be consumed. As with deep-sea fishes, the diets of deep-sea decapods ap- pear to be very generalized (Lagardere, 1977a,b). Deep-sea echinoderms also prey on snails. Most deep-dwelling ophiuroids are unselec- tive omnivores (Pearson & Gage, 1984). Their stomach contents frequently include whole snails (0-2% of diet items) and shell fragments (Litvinova & Sokolova, 1971; Pearson & Gage, 1984). Carey (1972) found snails among stomach contents in 6 out of 26 species of deep-sea asteroids; snails were a dominant food source for 4 species. He pointed out that, in contrast to shallow-water asteroids, deep-sea species tend to have highly generalized feeding habits. The inci- dence of omnivorous species increases from 0% in the sublittoral zone to 71% in the abyssal zone. Vermeij (1978) was correct in saying that deep-sea predators are not highly adapted to crush hard-shell prey. There appear to be no potential predators in deep water comparable to tropical brachyuran crabs like Calappa that peel open snail shells with their massive specialized chelae (Shoup, 1968), or to the spiny puffer fish Diodon that uses reinforced jaws to crush snails (Palmer, 1979). However, deep-sea snails are consumed by a wide variety of more generalized fishes, decapods and echinoderms which are capable of caus- ing shell damage, and major breaks in deep- sea snails resemble those caused by crush- ing from fishes and crabs in coastal en- vironments. Whether what we have termed minor breaks result from a special and different set of causes is purely conjectural. Since most predators that are known to consume snails appear to be unspecialized megabenthic croppers (sensu Dayton & Hessler, 1972), it is easy to imagine both major and minor breaks resulting from ineffective prey handling. Bio- logical activities like the “mud-grubbing” for- aging behavior of macrourid fishes (McLellan, 1977) and burrowing of red crabs (Hecker, 1982) could also inflict damage, although these are probably most prevalent from upper to mid-bathyal depths. A possible source of minor damage at lower bathyal and abyssal depths in the western North Atlantic is strong near-bottom currents which resuspend and transport sediments (Richardson et al., 1981; Bulfinch et al., 1982) and could probably tumble snails, chipping the outer lips of their apertures. Turbidity currents and sediment slumps (Bulfinch et al., 1982) may have а similar effect at bathyal depths. It is also conceivable that annual and long-term varia- tion in trophic input from the surface (Deuser, 1986) and resuspension and transport of sed- iments in benthic storms (Gardner & Sullivan, 1981) sometimes result in nutrient depletion of sediments that is especially severe even by deep-sea standards. Weakened calcification of the outer lips of shell apertures during such periods might make them more subject to damage, or result in distinct growth checks that are more pronounced than normal growth lines on the shell. Implications for coevolution Vermeij (1978, 1983) has reviewed aspects of shell form that serve as antipredator adap- tations. Thick shells and bold sculpture of forms like Frigidoalvania brychia (Fig. 1a) have been shown experimentally to be effec- tive antipredator devices (Palmer, 1979). However, this type of shell armor is uncom- mon in deep-sea snails. In general, they are not heavily calcified and have more delicate sculpture. Many archaeogastropods are umbilicate, which would make them more vulnerable to predation (Vermeij, 1978). Deep-sea snails are also quite small, gener- ally less than one centimeter and frequently only a few millimeters in length. К is likely that energy limitations in the deep sea have scaled down the physical strength of oredators and the resistance of hard- shelled prey. Another limitation to both mol- lusks and some of their invertebrate predators is that calcium carbonate becomes more sol- uble with decreased temperature and in- creased pressure. Dissolution of calcium car- bonate selects for more efficient use of calcite and aragonite in shells which is often mani- fested by thinner shells and constraints on shell form (Graus, 1974). Our future research plans include inter- and intraspecific analyses of Vermeij's antipredator morphologies and application of Graus” calcification index to determine the relative importance of preda- tion and calcium carbonate availability for shell form in deep-sea snails. The highly coevolved predator-prey sys- tems found in shallow water, where reciprocal selection has led to very powerful and spe- cialized shell-crushing structures in predators 74 VALE & REX and resistant antipredator architecture in snail shells, appear not to have developed as ex- tensively in the deep sea. The predators of deep-sea snails appear to be unspecialized consumers with very general diets. Similarly, snails, for the most part, have not evolved elaborate and specific defense armor. They are subjected to unsuccessful attacks from potentially lethal predators. Predation and, or, physical disturbance strong enough to break shells are common features of the deep-sea environment. ACKNOWLEDGEMENTS We thank John Ebersole, Ron Etter, Jeremy Hatch and Andrea Rex for reading the manuscript. Mary Smith photographed the specimens in Fig. 1. The gastropods were collected by vessels of the Woods Hole Oceanographic Institution and were made available to us by Howard Sanders. REFERENCES CITED BERGMAN, J., GELLER, J. B. & CHOW, V., 1983, Morphological divergence and predator-induced shell repair in Alia carinata (Gastropoda: Prosobranchia). Veliger, 26: 116-118. BERTNESS, M. D. & CUNNINGHAM, C., 1981, Crab shell-crushing predation and gastropod ar- chitectural defense. Journal of Experimental Ma- rine Biology and Ecology, 50: 213-230. BERTNESS, M. D., GARRITY, S. D. & LEVINGS, S. C., 1981, Predation pressure and gastropod foraging: a tropical-temperate comparison. Evo- lution, 35: 995-1007. BRIGHT, T. J., 1970, Food of deep-sea bottom fishes. Texas A. & M. University Oceanographic Studies, 1: 245-252. BULFINCH, D. L., LEDBETTER, М. T., ELLWOOD, В. В. & BALSAM, W. L., 1982, The high-velocity core of the western boundary undercurrent atthe base of the U.S. continental rise. Science, 215: 970-973. CAREY, A. G., JR., 1972, Food sources of sublit- toral, bathyal and abyssal asteroids in the north- east Pacific Ocean. Ophelia, 10: 35-47. CURREY, J. D. & KOHN, A. J., 1976, Fracture in the crossed-lamellar structure of Conus shells. Journal of Materials Science, 11: 1615-1623. DAYTON, P. K. & HESSLER, R. R., 1972, Role of biological disturbance in maintaining diversity in the deep sea. Deep-Sea Research, 19: 199-208. DEUSER, W. G., 1986, Seasonal and interannual variations in deep-water particle fluxes in the Sargasso Sea and their relation to surface hydrography. Deep-Sea Research, 33: 225-246. DEUSER, W. G. & ROSS, E. H., 1980, Seasonal change in the flux of organic carbon to the deep Sargasso Sea. Nature, 283: 364-365. ELNER, R. W. & RAFFAELLI, D. G., 1980, Interac- tions between two marine snails, Littorina rudis Maton and Littorina nigrolineata Gray, a preda- tor, Carcinus maenas (L.), and a parasite Microphallus similis Jägerskiold. Journal of Ex- perimental Marine Biology and Ecology, 43: 151-160. FINNEY, D. J., LATSCHA, R., BENNETT, B.M. & HSU, P., 1963, Tables for testing significance in a2 x 2 contingency table. University Press, Cambridge, 103 pp. GARDNER, W. D. & SULLIVAN, L. G., 1981, Benthic storms: temporal variability in a deep- ocean nepheloid layer. Science, 213: 329-331. GELLER, J. B., 1983, Shell repair frequencies of two intertidal gastropods from northern Califor- па: microhabitat differences. Veliger, 26: 113-115. GRAUS, В. R., 1974, Latitudinal trends in the shell characteristics of marine gastropods. Lethaia, 7: 303-314. HAEDRICH, R. L. & POLLONI, P. T., 1976, A contribution to the life history of a small rattail fish, Coryphaenoides carapinus. Bulletin of the Southern California Academy of Sciences, 75: 203-211. HECKER, B., 1982, Possible benthic fauna and slope instability relationships. In: SAXOV, S. 4 NIEUWENHUIS, J. K., eds., Marine slides and other mass movements. Plenum, New York, pp. 335-347. HESSLER, В. В. 8 SANDERS, H. L., 1967, Faunal diversity in the deep-sea. Deep-Sea Research, 14: 65-78. HESSLER, R. R. & WILSON, G. D. F., 1983, The origin and biogeography of malacostracan crus- taceans in the deep sea. In: SIMS, R. W,, PRICE JL He & WHALLEY AP ЧЕ АЕ Systematics Association Special Volume No. 23, “Evolution, Time and Space: The Emergence of the Biosphere.” Academic Press, London and New York, pp. 227-254. JUMARS, P. A. & ECKMAN, J. A., 1983, Spatial structure within deep-sea benthic communities. In: ROWE, G. T., ed., Deep-Sea Biology, The Sea. Wiley, New York, 8: 399-452. JUMARS, P. A. £ GALLAGHER, E. D., 1982, Deep-sea community structure: three plays on the benthic proscenium. In: ERNST, W. G. 4 MORIN, J. G., eds., The Environment of the Deep Sea. Prentice-Hall, New Jersey, pp. 217-255. LAGARDERE, J. P., 1977a, Recherches sur la distribution verticale et sur l'alimentation des Crustaces Décapodes benthiques de la pente continentale du golfe de Gascogne—analyse des groupements carcinologiques. Bulletin du REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 75 Centre d'Études et de Recherches Scientifiques Biarritz, 11: 367—440. LAGARDERE, J. P., 1977b, Recherches sur le régime alimentaire et le comportement prédateur des Décapodes benthiques de la pente continentale de l'Atlantique norde oriental (Golfe de Gascogne et Maroc). European Symposium on Marine Biology, 11: 397-408. LITVINOVA, М. М. & SOKOLOVA, М. N., 1971, Feeding of deep-sea ophiuroids of the genus Amphiophiura. Okeanologija, Moscow, 11: 240-247. МАЧСНИМЕ, J. & GORDON, J. D. M., 1983, Diets of clupeoid, stomiatoid and salmonoid fish of the Rockall Trough, northeastern Atlantic Ocean. Marine Biology, 77: 67-78. MAUCHLINE, J. & GORDON, J. D. M., 1984, Diets and bathymetric distributions of the macrourid fish of the Rockall Trough, northeastern Atlantic Ocean. Marine Biology, 81: 107-121. McLELLAN, T., 1977, Feeding strategies of the macrourids. Deep-Sea Research, 24: 1019-1036. PALMER, A. R., 1979, Fish predation and the evolution of gastropod shell sculpture: experi- mental and geographic evidence. Evolution, 33: 697-713. PEARSON, M. & GAGE, J. D., 1984, Diets of some deep-sea brittle stars in the Rockall Trough. Marine Biology, 82: 247-258. RAFFAELLI, D. G., 1978, The relationship between shell injuries, shell thickness and habitat charac- teristics of the intertidal snail Littorina rudis Maton. Journal of Molluscan Studies, 44: 166-170. RAUP, D. M., 1966, Geometric analysis of shell coiling: general problems. Journal of Paleontol- ogy, 40: 1178-1190. REIMCHEN, T. E., 1982, Shell size divergence in Littorina mariae and L. obtusata and predation by crabs. Canadian Journal of Zoology, 60: 687-695. REX, M. A., 1973, Deep-sea species diversity: decreased gastropod diversity at abyssal depths. Science, 181: 1051-1053. REX, М. A., 1976, Biological accommodation in the deep-sea benthos: comparative evidence on the importance of predation and productivity. Deep- Sea Research, 23: 975-987. REX, M. A., 1977, Zonation in deep-sea gastro- pods: the importance of biological interactions to rates of zonation. European Symposium on Ma- rine Biology, 11: 521-530. REX, M. A., 1981, Community structure in the deep-sea benthos. Annual Review of Ecology and Systematics, 12: 331-353. REX, M. A., 1983, Geographic patterns of species diversity in the deep-sea benthos. In: ROWE, G. T., ed., Deep-Sea Biology, The Sea, Wiley, New York, 8: 453-472. REX, M. A. & WAREN, A., 1982, Planktotrophic development in deep-sea prosobranch snails from the western North Atlantic. Deep-Sea Re- search, 29: 171-184. RICHARDSON, M. J., WIMBUSH, M. & MAYER, L., 1981, Exceptionally strong near-bottom flows on the continental rise of Nova Scotia. Science, 213: 887-888. SANDERS, H. L., HESSLER, R. R. & HAMPSON, G.R., 1965, An introduction to the study of deep- sea benthic faunal assemblages along the Gay Head-Bermuda transect. Deep-Sea Research, 12: 845-867. SCHINDEL, D. E., VERMEIJ, С. J. & ZIPSER, E., 1982, Frequencies of repaired shell fractures among the Pennsylvanian gastropods of north- central Texas. Journal of Paleontology, 56: 729-740. SCHOENER, T. W., 1979, Inferring the properties of predation and other injury-producing agents from injury frequencies. Ecology, 60: 1110-1115. SEDBERRY, G. R. 4 MUSICK, J. A., 1978, Feed- ing strategies of some demersal fishes of the continental slope and rise off the mid-Atlantic coast of the USA. Marine Biology, 44: 357-375. SHIMEK, R. L., 1983, Biology of the northeastern Pacific Turridae. |. Ophiodermella. Malacologia, 23: 281-312. SHIMEK, R. L., 1984, The biology of the northeast- ern Pacific Turridae. IV. Shell morphology and sexual dimorphism in Aforia circinata (Dall, 1873). Veliger, 26: 258-263. SHOUP, J. B., 1968, Shell opening by crabs of the genus Calappa. Science, 10: 887-888. SIEGEL, S., 1956, Nonparametric statistics for the behavioral sciences. McGraw-Hill, New York, 312 pp. SIH, A., 1985, Evolution, predator avoidance, and unsuccessful predation. American Naturalist, 125: 153-157. SMITH, К. L., JR., 1978, Benthic community respi- ration in the N.W. Atlantic Ocean: in situ mea- surements from 40 to 5200 m. Marine Biology, 47: 337-347. THISTLE, D., YINGST, J. Y. & FAUCHALD, K., 1985, A deep-sea benthic community exposed to strong near-bottom currents on the Scotian rise (Western Atlantic). Marine Geology, 66: 91-112. TUREKIAN, К. K., COCHRAN, J. K., KHARKAR, D. P., CERRATO, R. M., VAISNYS, J. R., SAND- ERS, H. L., GRASSLE, J. F. & ALLEN, J. A., 1975, Slow growth rate of a deep-sea clam determined by 2?8Ва chronology. Proceedings of the National Academy of Sciences U.S.A., 72: 2829-2832. VERMEN, С. J., 1978, Biogeography and adapta- tion. Harvard University Press, Cambridge, MA, 332 pp. VERMEIJ, С. J., 1982a, Environmental change and the evolutionary history of the periwinkle (Littorina littorea) in North America. Evolution, 36: 561-580. VERMEIJ, С. J., 1982b, Gastropod shell form, 76 VALE & REX breakage and repair in relation to predation by the crab Calappa. Malacologia, 23: 1-12. VERMEIJ, С. J., 1982c, Unsuccessful predation and evolution. American Naturalist, 120: 701-720. VERMEIJ, G. J., 1983, Shell-breaking predation through time. In: TEVESZ, М. J. $. & MCCALL, P.L., eds., Biotic interactions in Recent and fossil benthic communities. Plenum, New York, pp. 649-669. VERMEIJ, G. J., 1985, Aptations, effects, and fortuitous survival: comment on a paper by Sih. American Naturalist, 125: 470-472. VERMEIJ, С. J., ZIPSER, Е. & DUDLEY, E. C., 1980, Predation in time and space: peeling and drilling terebrid gastropods. Paleobiology, 6: 352-364. VERMEIJ, G. J., SCHINDEL, D. E. 8 ZIPSER, E., 1981, Predation through geological time: evi- dence from gastropod shell repair. Science, 214: 1024—1026. VERMEIJ, С. J., ZIPSER, Е. & ZARDINI, R., 1982, Breakage-induced shell repair in some gastro- pods from the upper Triassic of Italy. Journal of Paleontology, 56: 233-235. WENNER, Е. |. & BOESCH, D. F., 1979, Distribu- tion patterns of epibenthic decapod Crustacea along the shelf-slope coenocline, middle Atlantic Bight, USA. Bulletin of the Biological Society of Washington, 3: 106-133. ZIPSER, E. & VERMEIJ, G. J., 1978, Crushing behavior of tropical and temperate crabs. Journal of Experimental Marine Biology and Ecology, 31: 155-172. REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 77 APPENDIX Total number Number with major Number with minor Locality and species examined repaired damage repaired damage Station 89, 196 m Onoba pelagica 37 1 13 Mitrella pura 5 1 1 Eulimella unifasciata 2 0 2 Aclis tenuis 1 0 1 Aclididae sp. 1 0 0 Station 88, 478 m Mitrella pura 31 4 7 Anachis haliaeeti 30 4 8 Solariella obscura 19 0 3 Frigidoalvania brychia 17 2 3 Pusillina harpa 11 0 1 Oenopota ovalis 8 0 1 Lepetella tubicola 10 0 0 Pusillina pseudoareolata 6 0 2 Admete contabulata 5 2 1 Lissospira sp. A 5 0 3 Cocculinidae sp. A 5 0 0 Colus pygmaeus 3 0 2 Onoba pelagica 3 0 0 Turridae sp. A 1 0 1 Cerithiella whiteavesii 1 0 1 Aporrhais occidentalis 1 0 0 Aclis tenuis 1 1 1 Calliotropis sp. A 1 0 0 Station 105, 530 m Frigidoalvania brychia 155 33 50 Mitrella pura 163 22 36 Pusillina harpa 89 3 16 Onoba pelagica 44 5 15 Solariella obscura 21 1 6 Aclis walleri 3 2 2 Admete contabulata 2 0 0 Colus pygmaeus 1 0 1 Anachis haliaeeti 1 1 0 Taranis morchi 1 1 0 Station 73, 1400 m Oenopota ovalis 64 0 23 Aclis walleri 30 19 29 Висситаае sp. A 23 1 4 Cyclostrematidae sp. A 14 0 2 Gymnobela sp. A 9 0 Y Natica sp. A 4 0 0 Oenopota graphica 4 0 2 Pleurotomella packardi 4 1 1 Cerithiella whiteavesii 2 0 2 Admete contabulata 2 1 1 Lissospira sp. A 1 0 0 Benthonella sp. A 1 1 1 Gymnobela sp. D 1 1 1 Bathysciadium costellatum 3 0 0 Station 103, 2022 m Aclis walleri 101 69 94 Oenopota ovalis 20 ih 6 Oenopota graphica 16 3 13 Boreotrophon abyssorum 10 3 7 78 VALE & REX Appendix Continued Locality and species Station 103, 2022 m (cont.) Theta chariessa Buccinidae sp. A Pleurotomella packardi Cerithiella whiteavesii Lora harpularia Pleurotomella sandersoni Rissoidae sp. A Cancellariidae sp. A Gymnobela brevis Gymnobela sp. C Aclididae sp. A Gymnobela sp. A Pleurotomella sp. A Station 131, 2178 m Lissospira sp. A Rissoidae sp. A Aclis walleri Cyclostrema smithi Boreotrophon abyssorum Station 76, 2862 m Benthomangelia antonia Gymnobela frielei Boreotrophon abyssorum Pleurotomella packardi Gymnobela sp. E Gymnobela sp. C Tacita sp. A Pleurotomella sandersoni Gymnobela tincta Theta chariessa Pleurotomella sp. A Gymnobela bairdii Gymnobela sp. B Gymnobela sp. F Benthomangelia sp. A Station 126, 3806 m Benthomangelia antonia Solariella sp. A Benthonella tenella Tacita sp. A Pleurotomella sandersoni Lissospira sp. C Gymnobela tincta Theta lyronuclea Omalogyra sp. A Epitonium nitidum Cocculinidae sp. B Leucosyrinx sp. A Gymnobela sp. F Gymnobela sp. G Buccinidae sp. A Omalogyra sp. B Benthomangelia sp. A Station 77, 3806 m Benthomangelia antonia Benthonella tenella Total number examined bh Nat | po GQ 40g0 kh 20400 ZA < < — ld — — NN — CO O1 ON — mb AA A — = — ANY A 2409000000 39 23 Number with major repaired damage — © © © ND O00000O0ON 0-0 oOo-0---000-O0000-.N oooo/_-o0o00-00+ ON Number with minor repaired damage — ONRO BR OO OO = OO = = = —= OO — — 0 O = 2 O = = = OO O©O = WON + OGo0-.00-000=MNO0O0 hy 12 REPAIRED SHELL DAMAGE IN DEEP-SEA PROSOBRANCHS 79 Appendix Continued Total number Number with major Number with minor Locality and species examined repaired damage repaired damage Station 77, 3806 m (cont.) Boreotrophon abyssorum 11 1 5 Pleurotomella sandersoni 12 5 2 Benthobia tryoni 8 1 3 Pleurotomella lottae 5 1 1 Brookula sp. A 3 0 0 Gymnobela sp. F 3 0 2 Typhlomangelia sp. A 1 0 1 Leucosyrinx sp. B 1 1 1 Tacita sp. A 1 0 1 Gymnobela sp. E 2 0 1 Station 85, 3834 m Benthonella tenella 65 Benthomangelia antonia 44 Lissospira sp. D 24 Adeorbis umbilicatus 20 Pleurotomella sandersoni 15 Pleurotomella lottae 13 Benthobia tryoni 12 Brookula sp. A 7. Pleurotomella sp. B 2 Theta Iyronuclea 2 4 2 5 2 2 2 2 2 1 1 1 1 5 4 1 № a Drilliola sp. A Gymnobela sp. B Benthomangelia sp. A Turridae, sp. B Boreotrophon abyssorum Gymnobela sp. F Epitonium sp. B Tacita sp. A Xanthodaphne sigmoidea Belomitra sp. A Turridae sp. C Tharsiella sp. A Lissospira sp. B Cocculinidae sp. B Gymnobela curta Station 123, 4853 m oOOOOOOODOOOOOD-OD—-DPRD—-MPMO — OO 100 10 - © © © ND © © B D D B ND © 0101 D Benthonella tenella 116 24 50 Adeorbis umbilicatus 21 1 2 Lissospira sp. D 20 3 4 Pleurotomella sp. B 6 1 0 Gymnobela sp. F 6 4 1 Theta lyronuclea 4 0 1 Drilliola sp. A 1 0 1 Cocculinidae sp. B 2 0 0 Belomitra sp. A 1 1 1 Station 124, 4862 m Benthonella tenella 82 15 41 Adeorbis umbilicatus 23 1 2 Drilliola sp. A 2 0 2 Lissospira sp. D 2 0 0 Gymnobela sp. F 1 0 1 Benthobia tryoni 1 0 0 Xanthodaphne sp. B 1 0 0 re Revised Ms. accepted 18 February 1987 MALACOLOGIA, 1988, 28(1-2): 81-94 SPERMATOGENESIS IN ONCOMELANIA HUPENSIS QUADRASI, A MOLLUSCAN HOST OF SCHISTOSOMA JAPONICUM Е. а. Claveria & F. J. Etges Department of Biological Sciences, University of Cincinnati Cincinnati, Ohio 45221, U.S.A. ABSTRACT Hepatotestes of laboratory reared male Oncomelania hupensis quadrasi (25 wk old) were processed for transmission electron microscopy, and spermatogenesis was studied. Within testicular acini are spermatogonia, sperm, numerous spermatocytes and spermatids in various stages of differentiation. Mature sperm are filiform and contain a homogeneous mass of nucleoprotein, spiralled around the head shaft. The Nebenkern consists of seven giant mitochondria which share a common, outer mitochondrial membrane, with their inner mem- branes remaining intact. Head shaft and flagellar axoneme show the typical cartwheel pattern of 9+2 microtubules. A single Golgi-complex was noted consistently in developing cells, whose granular secretions contribute to acrosome formation. Proximal and distal centrioles, seen only in early stages of spermatid differentiation, apparently contribute to the formation of the intranuclear and flagellar axoneme. A row of microtubules (= manchette) surrounds the non-helicoidal, homogeneous mass of nucleoprotein and developing Nebenkern of elongate spermatids. While microtubules around the nucleus are located away from the nuclear membrane, in the Nebenkern, these microtubules closely appose the outer mitochondrial membrane. Microtubules are either absent or widely scattered in mature sperm. Whether these microtubules participate in determining the final corkscrew form of the sperm is not clear. While a few biflagellate sperm were noted, atypical forms reported in other prosobranch snails such as apyrene and oligopyrene sperm, were not observed. Sertoli cells with many electron dense bodies and prominent nuclei are confined to the acinar wall. Key words: spermatogenesis, prosobranchia, Oncomelania hupensis quadrasi. INTRODUCTION Studies on the process of spermatogenesis at the ultrastructural level have been reported in various prosobranch snails such as Viviparus spp. (Hanson et al., 1952; Gall, 1961), Cipangopaludina spp. (Yasuzumi & Tanaka, 1958; Yamasaki, 1966), Epitonium tinctum (Bulnheim, 1968), Nucella lapillus (Walker, 1970), Littorina sitkana (Buckland- Nicks & Chia, 1976), Ocenebra erinacea (Féral, 1977), Colus stimpsoni (West, 1978), Bithynia tentaculata (Kohnert, 1980), and Lambis lambis and Conomurex luhuanus (Koike & Nishiwaki, 1980). While working on Oncomelania hupensis quadrasi infected with Schistosoma japonicum, we observed total or partial loss of testicular tissue. Although de- struction of the gonads has been reported before in some mollusks infected with trematode parasites (Rees, 1934, 1936; Pratt & Barton, 1941; Sullivan et al., 1985), there are no published reports of such damage in oncomelanian snails. To date, the process of spermatogenesis has not been critically de- scribed in any strains of Oncomelania hupensis. We anticipate that the present de- scription in mature, uninfected O. h. quadrasi may eventually be used for comparison of spermatogenesis in infected male snails which exhibit parasitic castration. Further- more, these data will contribute to our general knowledge of sperm formation in prosobranch snails. MATERIALS AND METHODS Snail cultivation Oncomelania hupensis quadrasi snails were reared in the laboratory following the cultivation techniques of Van der Schalie & Davis (1968) and French (1974) with some modifications. Plastic lined aluminum trays (17 x 24"), half filled with sterile muddy soil and water, were exposed to two-40W cool white fluorescent lights at least 6 hr/day, to stimulate growth of blue-green algae as snail 82 CLAVERIA & ETGES FIG. 1. Section of testis showing testicular acini with spermatocytes (Sc), various stages of spermatids (St) and sperm (S). Acinar wall (arrow). Bar = 40 um food. Filaments of Nostoc sp. were added to the culture trays to supplement snail diet. Aerated tap water was added when neces- sary. Electron microscopy Six mature male O. h. quadrasi (25 wk old) were cleaned of soil particles with a fine brush, lightly crushed and their shells care- fully removed under a stereoscope. Hepatotestes were cut from the rest of the snail body, fixed in 3% glutaraldehyde in 0.15 М sodium cacodylate buffer (pH 7.4) overnight at 4°C. Tissues were washed in 0.2 M cacodylate buffer four times at 30 min intervals, post-fixed in 1% buffered osmium tetroxide for 1 hr at 10°C, and then stained with 2% uranyl acetate in 10% ETOH for 45 min. Specimens were dehydrated in seri- ally graded ethanol (30, 50, 70, 80, 95%) for 10 min each wash, followed by 100% ethanol and propylene exode (2 changes each) for 10 min. Tissue infiltration employing 50:50 parts propylene oxide and Spurr resin for 6 hr was followed by embedding in 100% Spurr resin in plastic capsules, polymerized at 60°C for 48 hr. Sections 8-10 nm thick were cut with a diamond knife using ultramicrotomes (Reic- hert OM U3 and Sorvall MT 2-B), then stained with lead acetate for 3 min. Sections were observed using a 9S-2 Zeiss and a Phillips 300 electron microscopes. RESULTS Testes of normal, mature male O. h. quadrasi show all the developmental stages of spermatogenesis, including spermato- gonia, spermatocytes, spermatids and sperm within testicular acini (Fig. 1). The acinar wall has an inner layer of Sertoli cells, which is delimited from the outer germinal epithelium by a narrow, less electron dense layer. Sertoli ONCOMELANIA SPERMATOGENESIS 83 cells have prominent nuclei and highly gran- ular cytoplasm containing numerous electron dense bodies of varied shapes and sizes. The germinal epithelium has a layer of flattened cells, with large nuclei and scanty cytoplasm (Figs. 25, 26). Very few spermatogonia were observed near the acinar wall. They are relatively smaller than spermatocytes, measuring 4.5-6.0 um diam, (nuclear diam from 3.5-5.4 um). Their scanty cytoplasm contains few cytoplasmic organelles such as endo- plasmic reticulum, mitochondria and Golgi material. A single nucleolus was noted in some cells (Fig. 2). Spermatocytes are spherical to irregular in shape, with nuclei containing patchy chroma- tin with or without nucleoli (Fig. 3). Primary spermatocytes are generally larger than spermatogonia, measuring 6.2-8.7 um diam with nuclei ranging from 4.0-5.8 um across. They contain increased numbers of mito- chondria with early signs of clustering; both the Golgi material and endoplasmic reticulum are conspicuous as well (Fig. 3). Primary and secondary spermatocytes are often difficult to differentiate. However, sec- ondary spermatocytes can be recognized by the presence of well-formed Golgi body and larger mitochondria developing from the fusion of smaller ones, forming a cluster at one end of the cell (Figs. 4, 5). Cytoplasmic bridges were seen between some late secondary spermatocytes, indicating delayed cytokinesis, and nuclei of some cells have basal and apical thickenings (Fig. 5). Spermiogenesis is divided arbitrarily into four stages (Buckland-Nicks & Chia, 1976; Eckelbarger & Eyster, 1981), based on gen- eral nuclear shape and chromatin condensa- tion, development of giant mitochondria into a Nebenkern, formation of Golgi-complex and acrosome and the axonemal complex. Stage A (pre-cup stage) spermatids are irregularly shaped, measuring 2.4-7.3 um diam, with subspherical nuclei, 2.4-3.1 um diam in either central or eccentric position (Fig. 6). The antero-posterior axis of the cell is established early, with the formation of a basal and apical thickening of electron opaque material at opposite ends of the nu- cleus (Figs. 6, 7). The basal plate invaginates, forming a small cavity or indentation (Figs. 7, 8). Meanwhile, the nuclear material begins to condense into granular chromatin and aggre- gates to form lateral patches on the inner nuclear envelope, leaving a somewhat less electron dense space in the basal or apical area (Figs. 6, 7). The cytoplasm has numer- ous cisternae of endoplasmic reticulum, large mitochondria, and a well-developed Golgi body with stacked saccular membrane and granular secretions (Figs. 8, 9). During the pre-cup stage, Golgi-complex and mito- chondria are not necessarily positioned ac- cording to where they occur in later develop- mental stages along the antero-posterior axis of the cell. The change in cell size from stage A to stage B (cup-shaped) is slight, and many spermatids have features common to both stages. The small shallow indentation in the center of the basal end of the nucleus, normally observed during the pre-cup stage, grows deeper with the insertion of a cap-like terminal end of the developing intranuclear axoneme (Fig. 10). Following this insertion, the seven prominent giant mitochondria begin to aggregate at the base of the nucleus and eventually become closely associated with the developing flagellar axoneme (Fig. 11). At various points, numerous distinct electron dense granules appear between inner mitochondrial membranes; many cristae are present as well. Cisternae of endoplasmic reticulum are scattered in the cytoplasm with numerous free and attached ribosomes. The nucleus is flattened basally and somewhat rounded apically. Laterally or apically, a single prominent Golgi-complex forms transfer vesicles, presumably contain- ing pro-acrosomal material. The presence of cisternae of endoplasmic reticulum adjacent to transfer vesicles, suggests their participa- tion in the synthesis of pro-acrosomal secre- tions (Figs. 9,11). Several transfer vesicles apparently aggregate and then fuse together to form the pro-acrosome. At least two pro-acrosomes were noted in several of the differentiating spermatids (cup to post-cup stage). Presumably these pro-acrosomes form a larger pro-acrosome, which initially exhibits an electron dense central core (Fig. 12) and finally gives rise to the acrosomal component of the sperm. The residual Golgi body moves toward the developing mid-piece and continues to produce transfer vesicles, possibly to aid in the removal of superfluous cytoplasm from the mid-piece, during the elongation phase of spermiogenesis. A proximal centriole was noted in some cells on the tip of developing intranuclear axonemes; while the distal centriole was seen posterior to the basal nuclear plate, and is 84 CLAVERIA & ETGES FIGS. 2-5. 2. Spermatogonium with large nucleus (N) and prominent nucleolus (arrow). Note scanty cytoplasm (С) with few cytoplasmic organelles. Bar = 4 рт. Figs. 3-5. Spermatocytes. 3. Spermatocytes showing patchy chromatin with or without nucleoli (long arrow). Golgi body (short arrow), mitochondria (Mi), acinar wall (Aw). Bar = 10 um. 4. Secondary spermatocyte with well-formed Golgi body (Gb), cluster of mitochondria (Mi), nuclear membrane (arrow). Bar = 1.5 um. 5. Late secondary spermatocytes joined by a cytoplasmic bridge (long arrow). Note basal-apical nuclear thickenings (short arrows). Nucleus (N). Bar = 2.5 pm. closely associated with the cluster of giant mitochondria (Fig. 10). Stage C (post-cup stage) spermatids are characterized by condensation of granular chromatin, fusion of giant mitochondria to form the Nebenkern around the flagellar axoneme, and further differentiation of the pro-acrosome into the acrosomal component of the sperm head. Depending on the stage of transformation, the nuclei are spherical with a flattened basal plate, ovoid or some- what elongate. Along the antero-posterior axis of the nucleus, granular chromatin material condenses into fibrous strands, which fuse and form lamellar chromatin. Formation of lamellae commences peripher- ONCOMELANIA SPERMATOGENESIS 85 FIGS. 6-9. Stage A (Pre-cup) spermatids. 6. Basal thickenings of spermatids, showing the start of invagination (long arrow). Note chromatin condensation and less electron dense area basally. Nucleus (N), endoplasmic reticulum (short arrow). Bar = 2 рт. 7. Spermatid with apical and База! thickenings (short arrows). Developing intranuclear canal (long arrow). Bar = 1 рт. 8. Large mitochondria (Mi) with electron dense granules (long arrow), Golgi body (Gb), and numerous cisternae of endoplasmic reticulum (Er). Note chromatin condensation within nucleus and basal plate invagination (short arrow). Bar = 1.5 um. 9. Stacked saccular membranes (M) of Golgi body (Gb), and endoplasmic reticulum (long arrow), with transfer vesicles (short arrows). Bar = 0.5 шт. ally (Figs. 13, 14, 15, 16), and they appear common outer mitochondrial membrane in cross sections to adhere to the inner (Fig. 16). surface of the nuclear envelope (Fig. 15). Stage D (elongate stage) spermatids fur- Also, clustered giant mitochondria fuse, elon- ther increase in length until they reach a gate, and form the Nebenkern (= typical filiform shape. The lamellar chromatin mitochondrial sheath) enclosed within a forms a single homogeneous mass of 86 CLAVERIA & ETGES FIG. 10. Stage B (Cup-shaped) spermatid. Nucleus (N) showing developing intranuclear axoneme (short arrow). Distal centriole (long arrow), with microtubules posterior to intranuclear canal. Note granular chromatin material in nucleus. Mitochondria (Mi), endoplasmic reticulum with ribosomes (Er). Bar = 1.0 um. nucleoprotein that winds around the nuclear axoneme, giving the sperm head a helicoidal or corkscrew form (Figs. 17, 18). Likewise, the Nebenkern spirals around the flagellar axoneme (Figs. 17, 19), and has numerous prominent cristae (Fig. 20). Extrusion of residual cytoplasm in the head takes place anteriorly, as shown by the presence of cytoplasmic fragments lying close to the apex of sperm heads (Fig. 18). Extrusion of more superfluous cytoplasm also apparently takes place in the posterior end of the developing sperm, judging from the large cytoplasmic accumulation in the tail region (Figs. 18, 19). The sperm head and mid-piece are sup- ported by an axoneme of typical cartwheel pattern of 9 + 2 microtubules (Figs. 13, 20). Stage D spermatids, characterized by homo- geneous nucleoprotein and a Nebenkern undergoing elongation, also have a row of microtubules around the nucleus and mitochondrial sheath. Microtubules around the nucleus do not appose with the nuclear membrane (Fig. 21), while those around the Nebenkern lie side by side with the outer mitochondrial membrane (Fig. 22). These microtubules were not seen in earlier stages. In mature sperm, however, either very few scattered microtubules remained or they were absent (Figs. 23, 24).) There are Sertoli cells closely associated with developing spermatocytes and differen- tiating spermatids and are confined to the acinar wall (Figs. 25, 26). Sertoli cells showed mid-pieces and nuclei of elongate spermatids embedded in their cytoplasm, as well as junctional complexes (= gap junctions) with spermatocytes (Fig. 26). Although there are a few biflagellated sperm, atypical forms such as the oligopyrene and apyrene types re- ported in many other prosobranch snails were not observed in O. h. quadrasi. Mature sperm are long, with the mid-piece and tail comprising about 90% of their entire length. They possess a cone-shaped acro- some and a spirally twisted nucleus measur- ing 0.5-1.2 um wide and 5.0-7.3 um long (Figs. 25, 27). ONCOMELANIA SPERMATOGENESIS 87 FIGS. 11, 12. Stage B (Cup-shaped) spermatids. 11. Spermatid with seven giant mitochondria (Mi), undergoing fusion, surrounding flagellar axoneme (long arrow). Note common, outer mitochondrial membrane (short arrow). Golgi body (Gb) with secretory granules. Bar = 1.0 um. 12. Golgi body (Gb) adjacent to a pro-acrosome with electron dense central core (arrow). Bar = 1.0 um. DISCUSSION Among prosobranch snails, condensation of chromatin materials during spermiogenesis transforms the nucleus into a homogeneous mass of nucleoprotein. Nuclear aggregation of fibrous strands in Littorina sitkana (Buckland-Nicks & Chia, 1976) and Ocenebra erinacea (Féral, 1977) begins both at the center and on the periphery. In Oncomelania h. quadrasi, however, lamellar chromatin for- mation commences peripherally and moves inward to the center of the nucleus, in a pattern similar to that of Cipangopaludina malleata (Yasuzumi & Tanaka, 1958) and Colus stimpsoni (West, 1978). The helical form of the sperm head of O. h. quadrasi resembles that of Viviparus spp. (Hanson et al., 1952; Kaye, 1958; Gall, 1961), C. malleata (Yasuzumi & Tanaka, 1958), and Truncatella subcylindrica (Giusti & Mazzini, 1973). Interestingly, in Nucella lapillus, the head shaft initially forms a gentle spiral of 5-7 turns clockwise, with no corresponding twist- ing in the flagellar axoneme and the nucleus. As the sperm nucleus condenses and elon- gates to its final length, the head shaft is pulled out straight (Walker, 1970). Franzén (1970) noted a cytoplasmic spiral keel around the spirally-twisted nucleus and mitochondrial sheath, which enhances the corkscrew con- figuration of the nucleus of Partulida spiralis, a pyramidellid snail. In O. В. quadrasi, a single row of microtubules (= manchette) was found around the nucleus and Nebenkern, similar to the arrangement in Bithynia tentaculata (Kohnert, 1980). In С. stimpsoni (West, 1978), the nucleus is helically wound with 3-5 rows of microtubules, biradially arranged and lying perpendicular to the two central axonemal fibers. Among snails with cylindri- cal sperm heads, Buckland-Nicks & Chia (1976), and Walker (1970) observed micro- tubules during later stages of condensation and suggested that these microtubules may assist in nuclear elongation and provide a sufficiently rigid support to sustain the shap- ing of the sperm head. Microtubules during later stages of nuclear condensation also were noted in O. h. quadrasi and apparently participate in the elongation process. How- ever, the role these microtubules play in de- termining the helicoidal shape of the sperm head is doubtful, judging from the location of the microtubules relative to the nuclear mem- 88 CLAVERIA & ETGES FIGS. 13-16. Stage C (Post-cup) spermatids. 13. Cross section of spermatid nucleus showing intranuclear axoneme (A) and partially condensed chromatin. Note fibrous (long arrow) and lamellar chromatin (short arrow) peripherally. Bar = 0.5 рт. 14. Longitudinal section of spermatids, showing fibrous strands of chromatin (long arrow) chromatin lamellae (short arrow). Nebenkern (Ne) around flagellar axoneme (A). Bar = 1.0 um. 15. Cross sections of nuclei. Advanced stage of chromatin condensation with less fibrous chromatin. Note thick chromatin lamellae on the inner surface of nuclear envelope (short arrows). Bar = 0.5 um. 16. Longitudinal section of early stage C spermatid, with granular chromatin forming fibrous strands (long arrow). Note development and elongation of Nebenkern (Ne) and distinct cristae (short arrow). Endoplasmic recitulum with ribosomes (double arrows). Bar = 1.0 рт. ONCOMELANIA SPERMATOGENESIS 89 FIGS. 17-20. Stage D (elongate) spermatids. 17. Tangential sections of corkscrew-shaped nuclei (N) and Nebenkern (Ne) spiralled around axoneme (arrow). Bar = 1.0 „m. 18. Early stage elongate spermatids with extruded cytoplasm (Ec) apically. Note superfluous cytoplasm (long arrow) in sections of mid-piece (short arrows). Bar = 6.0 um. 19. Tangential section of spermatid with fragments of residual cytoplasm (arrows) attached to plasma membrane. Note spiral Nebenkern (Ne) and nucleus supported by axoneme (A). Bar = 0.5 pm. 20. Cross sections of mid-piece. Flagellar axoneme (long arrow). Seven mitochondria composing the Nebenkern (Ne) are evident. Note common outer mitochondrial membrane (short arrow) and numerous cristae. Bar = 1.0 pm. 90 CLAVERIA & ETGES FIGS. 21-24. 21. Cross section of spermatid with microtubules (arrows) around nucleus (N). Note distance of microtubules from nuclear membrane. Bar = 0.5 рт. 22. Cross sections of mid-piece of spermatids with row of microtubules (arrows) beside outer membrane of Nebenkern (Ne). Bar = 0.5 рт. 23. Cross section of sperm head with helicoidal nucleus (N). Plasma membrane (long arrow), sections of end-piece of sperm tail (short arrow). Bar = 0.5 um. 24. Cross sections of mid-piece of elongate spermatids. Note widely scattered microtubules (long arrows) and shape of Nebenkerne. Bar = 0.5 um. ONCOMELANIA SPERMATOGENESIS 91 brane (Fig. 21), and therefore requires further investigation. Fawcett et al. (1971) found that in avian finch sperm, which have a rather complex corkscrew-shaped sperm head, microtubules are absent during the elongation process. During the intermediate and late stages of differentiation, when the helical form of the nucleus is already evident, 6-8 rows of microtubules are associated with the nucleus. They postulated that microtubules are proba- Ыу not essential in initiating the helical shape of the nucleus. Most of their evidence favors the view that the configuration of the finch sperm head is determined by intrinsic nuclear factors, and not by the helical microtubules of the manchette. They further argued that microtubules, extending posteriorly and spi- ralling around the Nebenkern, form a tempo- гагу organelle homologous to the manchette, which induces spermatid elongation and may even determine the form of the mitochon- drial sheath. In O. h. quadrasi, microtubules around the Nebenkern are closely apposed to the outer mitochondrial membrane (Fig. 22), suggesting a similar function. While the Golgi body makes no apparent contribution to the formation of acrosome in Nerita senegalensis (Garreau de Loubresse, 1971), т О. h. quadrasi, a single Golgi body is involved in acrosome formation. Morphologically and physiologically, giant mitochondria that compose the Nebenkern vary extensively among different phyla and species (Anderson & Personne, 1976; Franzén, 1970). The pattern of Nebenkern formation even varies to a certain degree among prosobranch species. Cipango- paludina spp. (Yasuzumi & Tanaka, 1958; Yamasaki, 1966), O. erinacea (Féral, 1977) and C. stimpsoni (West, 1978) have two giant mitochondria. In Cipangopaludina spp. these mitochondria are tightly spiralled around the flagellar shaft as separate bodies. There are 4-5 giant mitochondria in Littorina sitkana (Buckland-Nicks & Chia, 1976). N. lapillus (Walker, 1970) and Viviparus contectoides (Kaye, 1958), 7-9 in L. lambis (Koike & Nishiwaki, 1980) and 9 in B. tentaculata (Kohnert, 1980). In O. h. quadrasi there are usually 7 giant mitochondria enclosed within a common outer mitochondrial membrane, with their inner membranes intact. It seems rather unlikely that the Sertoli cells of O. h. quadrasi play an active role in the transportation of spermatogenic cells. Such conjecture is based on the location of Sertoli cells, which is on the periphery of the acinar FIGS. 25, 27. 25. Section of testicular acinus packed with sperm (S). Note Sertoli cells (long arrows) and germinal epithelium (two long arrows), with a less electron dense homogeneous layer (short arrow). Mid-piece partly embedded in a Sertoli cell (two short arrows). Nucleus (N), electron dense bodies (Db). Bar = 10 jm. 27. Longitudinal sections of sperm heads showing cone-shaped acrosome (arrow) and helicoid nucleus (N). Bar = 1.0 pm. 92 CLAVERIA & ETGES . у ‚ LW dh. FIG. 26. Portion of acinar wall of two juxtaposed testicular acini. Each acinar wall has an inner layer of Sertoli cells (Sc), a middle homogeneous layer (H) and an outer germinal epithelium (long arrows). Note electron dense bodies (Db) and other cytoplasmic inclusions within Sertoli cells, several sections of mid-piece and nucleus of elongate spermatids embedded in the cytoplasm (two long arrows). Also, note desmosome-like processes (= gap-junctions) (short arrows) between a spermatocyte and a Sertoli cell. Nucleus (N). Bar = 3.0 pm. ONCOMELANIA SPERMATOGENESIS 93 wall. Also, the absence of defined cytoplas- mic microfilaments in association with sperm are not present. There are indications, how- ever, that Sertoli cells are involved in the nutrition of developing spermatocytes and differentiating sperm, evidenced by the pres- ence of some junctional complexes with spermatocytes and elongate spermatids, em- bedded in their cytoplasm (Fig. 26). Some electron dense cytoplasmic inclusions in Sertoli cells resemble residual bodies re- ported in Biomphalaria glabrata (De Jong- Brink et al., 1977), which suggest that cells possibly function in phagocytosis of superflu- ous, extruded cytoplasm. Sperm dimorphism, that is the production of typical and atypical sperm has been reported in many prosobranch snails (Nishiwaki, 1964; Tochimoto, 1967; Koike & Nishiwaki, 1980). In О. В. quadrasi, although a few biflagellated sperm were observed, they do not resemble the atypical apyrene and oligopyrene sperm reported in other prosobranchs. Physiological dimorphism is possible in О. h. quadrasi, a form of dimorphism suggested to occur in C. stimpsoni, which have only typical eupyrene sperm. ACKNOWLEDGMENTS We thank Feliza Thompson and Dr. E. Rivera for their excellent technical assistance in transmission electron microscopy. This work was carried out, in part, with the support of a Fulbright-Hays Predoctoral Fellowship awarded to Florencia G. Claveria. REFERENCES CITED ANDERSON, W. A. & PERSONNE, P., 1976, The molluscan spermatozoon: Dynamic aspects of its structure and function. American Zoologist, 16: 293-313. BUCKLAND-NICKS, J. A. & CHIA, F. S., 1976, Spermatogenesis of marine snail, Littorina sitkana. Cell and Tissue Research, 170: 455-475. BULNHEIM, H. P., 1968, Atypische Sperma- tozoenbildung bei Epitonium tinctum: Ein beitrag zum problem des spermatozoen-dimorphismus der Prosobranchia. Helgoländer wissenschaft- liche Meeresuntersuchungen, 18: 232-253 (with English abstract). DE JONG-BRINK, M., BOER, H. H., HOMMES, T. G. & KODDE, A., 1977, Spermatogenesis and the role of Sertoli cells in the freshwater snail Biomphalaria glabrata. Cell and Tissue Re- search, 181: 37-58. ECKELBARGER, К. J. & EYSTER, L. S., 1981, An ultrastructural study of spermatogenesis in the nudibranch mollusc Spurilla neapolitana. Journal of Morphology, 170: 283-299. FAWCETT, D. W., ANDERSON, W. A. 8 PHIL- LIPS, D. M., 1971, Morphogenetic factors influ- encing the shape of the sperm head. Develop- ‚mental Biology, 26: 220-251. FERAL, C., 1977, Etude de la spermatogenese typique chez Ocenebra erinacea, Mollusque Gastéropode, Prosobranche. Bulletin de la Societe Zoologique de France, 102: 25-30. FRANZEN, Ä., 1970, Phylogenetic aspects of the morphology of spermatozoa and spermi- ogenesis. In: Comparative Spermatology, ed. BACCETTI, B., Academic Press, London, pp. 29-46. FRENCH Ill, J. R., 1974, Improved methods for culturing the subspecies of Oncomelania hupensis: The snail hosts of Schistosoma japonicum, the oriental human blood fluke. Sterkiana, 56: 1-20. GALL, J. G., 1961, Centriole replication: A study of spermatogenesis in the snail, Viviparus. Journal of Biophysical and Biochemical Cytology, 10: 163-193. GARREAU DE LOUBRESSE, N., 1971, Spermi- ogenèse d'un Gasteropode Prosobranche Nerita senegalensis: Evolution du canal intranucleare. Journal de Microscopie Paris, 12: 425-440 (with English abstract). GIUSTI, F. & MAZZINI, M., 1973, The sperma- tozoon of Truncatella (s. str.) subcylindrica (L.) (Gastropoda Prosobranchia). Monitore Zoo- logico Italiano, 7: 181-210. HANSON, J., RANDALL, J. T. & BAYLEY, $. Т., 1952, The microstructure of the spermatozoon of the snail, Viviparus. Experimental Cell Research, 3: 65-78. KAYE, J. S., 1958, Changes in the fine structure of mitochondria during spermatogenesis. Journal of Microscopy, 102: 347-369. KOHNERT, V. В. 1980, Zum spermi- endimorphismus der Prosobranchier: Spermi- ogenese und ultrastruktureller Aufbau der Spermien von Bithynia tentaculata (|). Zoologischer Anzeiger, 205: 145-161 (with En- glish abstract). KOIKE, K. & NISHIWAKI, S., 1980, The ultra- structure of dimorphic spermatozoa in two spe- cies of the Strombidae (Gastropoda: Prosobranchia). Japanese Journal of Malacology, 38: 195-201. NISHIWAKI, S., 1964, Phylogenetical study on the types of the dimorphic spermatozoa in Prosobranchia. Science Reports of the Tokyo Kyoiku Daisaku, 11: 237-275. PRATT, I. & BARTON, С. D., 1941, The effects of four species of larval trematodes upon the liver and ovotestis ofthe snail, Stagnicola emarginata 94 CLAVERIA & ETGES angulata (Sowerby). Journal of Parasitology, 27: 283-288. REES, Е. G., 1934, Cercaria patellae Lebour, 1911, and its effect on the digestive gland and gonads of Patella vulgata. Proceedings of the Zoological Society of London, 45-53. REES, F. G., 1936, The effects of parasitism by larval trematodes on the tissue of Littorina littorea (Linné). Proceedings of the Zoological Society of London, 357-368. SULLIVAN, J. T., CHENG, Т. С. & HOWLAND, К. H., 1985, Studies on parasitic castration: Castra- tion of llyanassa obsoleta (Mollusca: Gastro- poda) by several marine trematodes. Transac- tions of the American Microscopical Society, 104: 154-171. TOCHIMOTO, T., 1967, Comparative histochemi- cal study on the dimorphic spermatozoa of the Prosobranchia with special reference to polysac- charides. Science Reports of the Tokyo Kyoiku Daigaku, 13: 75-109. Van DER SCHALIE, H. & DAVIS, G. M., 1968, Culturing Oncomelania snails (Prosobranchia: Hydrobiidae) for studies of oriental schistosomiasis. Malacologia, 6: 321-367. WALKER, M. H., 1970, Unusual features of the sperm of Nucella lapillus (L.). In Comparative Spermatology, ed. Baccetti, B., Academic Press, London, pp. 383-392. WEST, D. L., 1978, Reproductive biology of Colus stimpsoni (Prosobranchia: Buccinidae) Il. Spermiogenesis. Veliger, 21: 1-9. YAMASAKI, M., 1966, On the mitochondria and Golgi apparatus in spermiogenesis of the pond snail, Cipangopaludina japonica \wakawa (Pilsbry). Science Reports of the Tohoku Univer- sity, 32: 237-249. YASUZUMI, G. & TANAKA, H., 1958, Sperma- togenesis in animals as revealed by electron microscopy: VI. Researches on the sperma- tozoon-dimorphism in a pond snail, Cipan- gopaludina malleata. Journal of Biophysical and Biochemical Cytology, 4: 621-632. Revised Ms. accepted 14 November 1986 MALACOLOGIA, 1988, 28(1-2): 95-103 THE INITIAL STAGES OF RADULAR DEVELOPMENT IN CHITONS (MOLLUSCA: POLYPLACOPHORA) D. J. Eernisse' & K. Kerth? ABSTRACT The initial stages of development of the chiton radula were examined in Mopalia lignosa (Gould, 1846), M. muscosa (Gould, 1846), Lepidochitona fernaldi Eernisse, 1986, and L. caverna Eernisse, 1986. It starts in postmetamorphic juveniles with the secretion of the 2nd, 5th and 8th pairs of laterals, which are the main functional teeth of adult chitons. Moreover, it appears that juveniles are equipped with an efficient feeding instrument nearly as soon as radula formation begins, and certainly before the chitons have their complete set of teeth. This is evident from the mineralization of the 2nd laterals (“magnetite” teeth) from the start, the indications of normal degradation of “used” radula teeth in young juveniles, and observations of feeding in juveniles. As juveniles mature, new laterals are added between existing ones. The 1st laterals and the central tooth originate by fragmentation of a medial “precursor” plate. The phylogenetic implications of the polyplacophoran mode of tooth pattern formation are discussed and related to inferences concerning a primitive ancestral molluscan radula. Key words: radula; Polyplacophora; chiton; morphogenesis; ontogeny; phylogeny. INTRODUCTION Comparative ontogenetic investigations of the molluscan radula have potential to reveal shared patterns of radular formation or diver- gent patterns that distinguish between partic- ular lineages of mollusks. For all mollusks, only polyplacophorans (Minichev and Sirenko, 1974; French translation by Sirenko & Minichev, 1975), Solenogastres (“aplacoph- orans”) (Salvini-Plawen, 1972, 1978), and pulmonates (Kerth, 1979) have been thor- oughly investigated. Minichev and Sirenko (1974) describe the radula in several genera of “larval” polyplacophorans as having a broad, monostichous form. They conclude from this observation that the radula of primitive mol- lusks is derived from a monostichous ances- tral state. Salvini-Plawen (1981; 1985) has reached similar conclusions for two species of Solenogastres (or Neomeniomorpha), based on Pruvot's famous larva (Pruvot, 1890) and his own observations (Salvini-Plawen, 1972, 1978) of Simrothiella, although he shows only aslender connection between two already well shaped halves in Simrothiella. In contrast, Kerth (1979) showed that the radulae of em- bryos of several pulmonate families pass through a distichous stage. If the general scheme proposed by Mini- chev and Sirenko (1974) and Salvini-Plawen (1985) is correct, then polyplacophorans (also referred to as chitons hereafter) and “apla- cophorans” would appear to have a funda- mentally different ontogenetic sequence of radular development from pulmonates, sug- gesting a possible phylogenetic discontinuity. This, and the availability of chiton larvae, led us to reexamine the process of radular forma- tion in chitons. Here we reexamine the ontogeny of radular development in four chiton species: Mopalia lignosa, M. muscosa, Lepidochitona fernaldi, and L. caverna. The successful culturing of chitons through meta- morphosis has been difficult for most workers, and often published descriptions have been based on cultures with a low percentage of metamorphosing juveniles. These four spe- cies were selected because of the fortuitous availability of healthy larvae and juveniles. In retrospect, this selection also permitted com- parisons between two families, between closely related species of two genera, and between free spawners (both Mopalia spp.) and brooders (both Lepidochitona spp.). Fi- nally, we infer a more general view of the basic polyplacophoran radula from our com- parisons of these four species, and compare this view to the one proposed by Minichev 8 Sirenko (1974). ‘Friday Harbor Laboratories, University of Washington, Friday Harbor, Washington 98250 USA. Current address: Museum of Zoology and Department of Biology, University of Michigan, Ann Arbor, Michigan 48109 USA. ?Zoologisches Institut, Universität Würzburg, Róntgenring 10, 8700 Würzburg, Е.В. Germany. 96 EERNISSE & KERTH MATERIALS AND METHODS Larvae and juveniles of four chiton species were obtained from adult chitons spawning at Friday Harbor Laboratories. All adults except for Lepidochitona caverna were collected on San Juan Island, Washington U.S.A. Adults of L. caverna were descended from a breeding population first maintained at Santa Cruz, California U.S.A. (site of original collection, Eernisse, 1984; 1986) and later at Friday Harbor Laboratories on San Juan Island. Mopalia lignosa and M. muscosa free spawned their gametes, and embryos hatched as swimming larvae in less than two days after fertilization for M. lignosa, and less than four days after fertilization for M. mus- cosa. These swimming larvae were main- tained in beakers for approximately one week with daily changes of filtered sea water and kept at ambient sea water temperatures (12 to 14° C). For each of the following samples of known age, the radulae of 4 to 14 specimens were examined, all fixed in 70% ethanol. The first series of larvae and juveniles was obtained from a single spawning M. muscosa female and several lightly spawning males, on June 10, 1985. Fixations were made at 7, 8, 9, 10, 11, 13, 15, 17, and 21 days after fertilization. Less than 10% of several hun- dred larvae had metamorphosed by 13 days, when a selection of representative larvae and all the benthic survivors of one of three cul- tures were fixed. At 17 days, about 40% of the remaining larvae in the cultures had meta- morphosed and the benthic survivors of the better of the two remaining cultures were fixed. Finally, at 21 days, nine metamor- phosed juveniles were fixed. The best series of larvae and juveniles was obtained from two M. lignosa spawning within hours of collection on August 1, 1985. An isolated male spawned first, and his sperm were introduced to an isolated female, prompting her to spawn copiously. The result- ing larvae were observed at least daily until they were near to metamorphosis. The first fixation was made at eight days, when about 40% of the larvae in all cultures (and in the fixed subsample) were metamorphosed. The next fixation was made at 12 days, approxi- mately one day after more than 95% of the larvae had completed metamorphosis. Sub- sequent fixations for this study were made at 18, 22, 29, 36, 51, 66, and 105 days after fertilization (length of juveniles examined: 0.35 to 1.6 mm), and other animals from this cohort were kept alive including 19 that were still alive at 14 months (mean length = st. dev. = 21.4mm + 3.26; max. length = 27.6 mm; min. length = 13.3 mm). In contrast to the free spawners M. muscosa and M. lignosa, the brooders L. fernaldi and L. caverna care for their embryos until the emerging larvae are capable of crawling and are within one or two days of metamorphosis (Eernisse, 1984). A large se- lection of adult brooders of these species were kept in the lab, and juveniles were collected near adults shortly before or after they had metamorphosed. For L. caverna, these juveniles ranged from recently meta- morphosed, about 0.5 mm length, to consid- erably older juveniles, to a maximum of 1.8 mm length. For L. fernaldi, we examined a series of juveniles ranging from 1.1 mm to 1.6 mm length. The exact age of juveniles collected in this way could not be determined. However, for L. fernaldi, additional broods were removed from three adults and cultured as for M. lignosa and M. muscosa. The age of each of these three broods (i.e. days since fertilization) was estimated with a high degree of confidence based on the appearance of previously timed developmental features in the embryos (Eernisse, 1984). Their age in relation to metamorphosis could be deter- mined by direct observation. All 45 larvae and juveniles from one brooder were fixed on July 11, 1985, when about 70% of the larvae had metamorphosed (approx. 13 days after fertil- ization), including 1 of 45 metamorphosed on July 8 (at 10 days), and 15 of 45 on July 10 (at 12 days). Juveniles from two other brooders were fixed at about 19, 25, and 28 days after fertilization. The length of the 13 to 28 day old L. fernaldi juveniles ranged from 0.35 mm to 0.5 mm. In addition to the above species, we exam- ined premetamorphic larvae of Lepidochitona cinerea (Linnaeus, 1767) (a kind gift from Prof. Dr. W. Haas, Bonn, Fed. Rep. of Ger- many). In preparation for phase contrast and Nomarski-interference contrast light micros- copy, specimens were first rehydrated, then the calcareous dorsal plates and girdle spicules were dissolved with 1N HCl. Next, specimens were macerated in cold 5-10% KOH (1 to 2h). Finally, the radulae were prepared by pressing the macerated tissue under a cover glass in hot glycerine gelatine. For SEM observations, juveniles and adults were macerated in warm 5% KOH only until INITIAL DEVELOPMENT OF THE CHITON RADULA 97 FIG. 1: Radulae of adult Mopalia lignosa in dorsal (slightly lateral) views with SEM. A. Adult (length 43 mm) from San Juan Is., Washington, USA; total radula length 8.6 mm with 33 transverse rows of mineralized teeth. Complete transverse rows 6-11, scale bar: 190 рт. В. Small adult (length 16 mm) from Año Nuevo Pt., California, USA. Left central portion of transverse rows (approx. Ys distance from first anterior row) with teeth spread in preparation, scale bar: 80 um. L, to Lg = laterals, М = medial (central) tooth. the radulae were clean and could be teased away from other tissue. After a distilled water rinse, radulae were stored in 70% ethanol, then transferred to 100% ethanol before mounting on a specimen stub. The radulae were sputtercoated with gold for two to six minutes and scanned at 40 kV on a JEM 1200 EX" STEM (in conjunction with energy dis- persive Xray microanalysis of juvenile and adult radulae as reported in a subsequent study, Eernisse and Fontaine, in prep.) or at 15kV on a JEOL SM-35 SEM. RESULTS The chiton radula The chiton radula is remarkably uniform in tooth number and type, bearing transverse rows of 17 teeth of predictable shapes (Figs. 1A,B) except 11 or 13 teeth per row in Juvenichiton (Sirenko, 1975). Each row is “stepped,” or v-shaped, with each tooth an- terior (at its base) to the next most distal tooth. The eight lateral teeth (L, to Lg) on each side of the medial or “central” tooth (“M”) are attached to the elastic radular mem- brane. The Lo and Ls pairs are the most elongate teeth. The L, pair are the main working teeth and bear highly magnetized dark caps (Lowenstam, 1962; Carefoot, 1965; Towe and Lowenstam, 1967), usually each with one to three sharp cusps. Each 15 tooth has the general appearance of a sickle, usu- ally with a flattened distal tip, and lies in close association over the mineralized portion of the 15 tooth from one row posterior. The relationship of the L; and Ls cusps suggests that they cooperate in scraping and collecting food or, alternatively, the L; cusps protect other soft parts from the highly mineralized L> cusps as these teeth roll back into their nor- mal tube-like orientation. Finally, the margins of the radula are stabilized by the plate- like Le 7,8. The development of the juvenile radula We found no radular structure in “trochophore” larvae (those larvae that still had a prototroch); the radula first appears after metamorphosis. Even the specimens of L. cinerea with conspicuous valve rudiments (plate-anlagen) lacked radulae. The first radulae were recognized in M. lignosa 8 days after fertilization (3 to 6 longitudinally re- peated, transverse rows of teeth); in M. muscosa in the course of the first week after metamorphosis (up to 10 transverse rows); 98 EERNISSE & KERTH FIG. 2: Radula morphogenesis in juvenile chitons. A. Foremost transverse row of the larval radula with 3 pairs of teeth (composite reconstruction with teeth slightly separated from camera lucida drawings of Mopalia lignosa, M. muscosa, Lepidochitona fernaldi). H = hump. B. Earliest larval radula of L. fernaldi 13 days after fertilization, phase contrast, scale bar: 20 um. C. Transverse row with 9 teeth (composite reconstruction as in Fig. 2A of M. lignosa, M. muscosa, L. fernaldi) corresponding with Fig. 2F. D. Transverse row with 13 teeth (camera lucida drawing as in Figs. 2A,C of L. fernaldi). E. L,-pair and medial tooth (L. fernaldi oldest series). Compare with the medial “precursor” plate (MP) in younger juveniles (Figs. 2 C,F), phase contrast, scale bar: 20 рт. Е. Bending plane of the radula, anteroventral SEM view (М. lignosa, 19 days after fertilization). Note the shape of the medial plate. Scale bar: 10 jm, A = alar membrane (subradular membrane) of the radula. INITIAL DEVELOPMENT OF THE CHITON RADULA 99 FIG. 3: Light micrograph dorsal view of radula (L. fernaldi 25 days after fertilization) with prominent dark-capped cusps (“magnetite teeth”) of L,-pairs. Left to right: anterior to posterior. Scale bar: 20 um. FIG. 4: Anterior end of a juvenile (1.1 to 1.6 mm length) radula with degradation. The foremost inner laterals and the medial tooth are shed (arrows). (L. fernaldi, phase contrast). Scale bar: 20 um. and in L. fernaldi 13 days after fertilization (4 to 7 transverse rows). The tooth shape and pattern were identical in the youngest specimens of M. lignosa, M. muscosa, and L. fernaldi: Radular develop- ment consistently starts simultaneously with the paired formation of the “magnetite teeth” (Lz), the sickle-shaped L; teeth, and the out- ermost marginal plates (Lg). Therefore the first transverse row of the newly formed radula usually consists of 6 teeth (pairs of Е 58). More rarely 4 teeth (pairs of Lo 5) or, much more rarely, a single pair (Lo) was noted. Each tooth was easily identified from the start by its characteristic shape (Figs. 2A,B,F). There is a bilaterally-symmetrical tooth pattern from the outset. The cusps of the anterior-most L, pair were dark-capped even in the earliest cases, suggesting miner- alization of these cusps from the onset of radular formation. Moreover, juveniles of all species considered here began active forag- ing movements within a week of metamor- phosis, moving from side to side and leaving a trail of corresponding rasp markings in the substrate covering of diatoms. A medial tooth could not be detected in early stages of radular development. This finding is contra- dictory to observations of Schizoplax by Sirenko 8 Minichev (1975: figs. 2b,c). In a few cases amorphous humps (“Н” in Fig. 2A) were observed in front of the foremost plates. Further radular development was docu- mented in M. lignosa, L. fernaldi and L. caverna. The radula elongates and the num- ber of transverse rows increases up to 30 to 40 in the oldest series. Several of the 18 to 28 day old juveniles showed radula degradation which is characteristic of all older animals. For juveniles of all ages as in adults, the medial part of the radula's anterior end was shed first (Fig. 4). New longitudinal rows of teeth appear with the increasing age of the chitons. A large 100 EERNISSE & KERTH medial plate (“MP,” Figs. 2C,D) and the Ls teeth are next to be added between the existing L> 5 g teeth. The radular development is identical in the three species until each has nine teeth within the transverse row, but dif- ferences were observed after this stage. In each transverse row, additional teeth form in the order L;, L4 and L; in L. fernaldi and M. lignosa, but L4, 17 and L, in L. caverna. Almost all of the oldest specimens of Lepidochitona appeared to exhibit a complete radula with 17 teeth in the transverse row, although the small L¿ could not be identified unequivocally. New teeth are added to the radula of M. lignosa very slowly by compari- son. The oldest juveniles have at most 11 longitudinal rows of teeth in their radula. Minichev & Sirenko (1974) state that nearly all new teeth in the transverse row originate from a fragmentation of “precursor”-plates. Such a process can be excluded at least for the 1545658 because we never observed these tooth pairs in an intermediate stage of fragmentation. On the other hand the medial plate apparently splits up to form the L, pair and the definitive medial tooth (Fig. 2D,H). This process was evident by comparing juve- niles that had only a broad medial plate with juveniles at a slightly more advanced stage that had a small medial tooth flanked by the L, pair. The medial tooth could be identified as a small medial ridge on the “precursor”- plate before its apparent separation. DISCUSSION In the three species examined at early stages, radula formation starts soon after metamorphosis. First a symmetrical tooth pat- tern arises, consisting normally of several transverse rows of “teeth,” each with three “tooth” pairs. A medial plate for each trans- verse row is added later. This sequence is basically the same as has been observed in radulae of many gastropods (Kerth, 1979; 1983a, b) including seven families of pulmo- nates and in two genera of opisthobranchs, Polycera and Adalaria, which pass through a stage with one to three pairs of laterals in each transverse row before a central tooth is added. The radulae of the chitons we examined, however, differ considerably from those of the gastropods examined in their later develop- ment. In gastropods, new longitudinal rows of laterals are added only on the outermost margins of the radula (Kerth and Hansch, 1977). In chitons, new longitudinal rows of laterals are inserted (i.e. erupt) between ex- isting laterals. We found that the first teeth or plates to appear in a chiton radula are, appro- priately enough, the main adult working Las teeth and Lg plates, the latter previously sug- gested to serve as margin stabilizers. Evi- dence presented here would indicate that particular radular teeth are formed with char- acteristic shapes making them functional al- most from the start, and certainly before all 17 teeth per row are present. Judging from their dark color, we concluded that the cusps of the initial Lo pairs were apparently mineralized from the start. This result has more recently been confirmed with energy dispersive Xray microanalysis of M. lignosa juveniles only 16 days post-fertilization (Eernisse and Fon- taine, in prep.). Finally, our observations of feeding behavior in newly metamorphosed juveniles provide strong evidence of the func- tionality of the newly formed radula. Minichev and Sirenko (1974) described the radular development in four chiton genera and in some unidentified chiton “trocho- phores.” Our results differ from theirs in sev- eral ways: (1) We observed no radulae earlier than postmetamorphic stages. (2) These au- thors describe a primordial radula in the uni- dentified trochophores with only one longitu- dinal row of broad plates. We didn't observe any comparable structure, although there oc- casionally were a few amorphous humps in front of the foremost transverse row (Fig. 2A). (3) Minichev and Sirenko (1974) depicted pri- mary central teeth in the foremost parts of the youngest radula, but secretion of these teeth stopped very early. We were not able to find any comparable structure even with phase contrast or Nomarski-interference optics. (4) According to these authors, almost all of the laterals originate by fragmentation of a pair of “precursor”-plates on either side of the Ls pair. Although the order of fragmentation is never explicitly stated in Minichev and Sirenko (1974: 1136), Sirenko and Minichev (1975: fig. 2b,c,d) clearly indicate that they believe the first fragmentation of the “precur- sor”-plates will lead to the adult Lz and L48 pairs, the next fragmentation to the L, and L;.s pairs, and so on until finally the L7 and Lg fragment. We can rule out such a fragmenta- tion process in the species we examined for all laterals with exception of the L,-pair. These and the medial tooth originate by frag- mentation of the medial plate (Fig. 2C,E,F). (5) We observed a different order that new INITIAL DEVELOPMENT OF THE CHITON RADULA 101 teeth are added in each transverse row. Sirenko and Minichev (1975: figs. 2b,c,d) depicted very exactly the shape of the laterals in the earliest radula and because their draw- ings are completely in accordance with the shape of laterals in our investigated species, it is clear they have misinterpreted several teeth or plates. For example, the teeth or plates Sirenko and Minichev (1975: fig. 2d) have labeled 15, L4, and Ls.g should instead be labeled Ls, L, and Lg, respectively. (6) Minichev and Sirenko (1974: 1136, fig. 1:4) argued that the dark portions of the Lo pairs are a secondary feature, with the first few Lo pairs lacking mineralization altogether. We observed mineralization concurrent with the start of radular formation. There is a possibility that our results differ from those of Minichev and Sirenko (1974) because they studied different chiton species, or because some of their species differ be- cause they are brooders (e.g. Schizoplax brandtii and Hanleyella asiatica). However, our consistent results for members of two chiton families, and for both free spawners and brooders, suggests to us that the patterns we have observed are general for chitons. If we are correct then our results are important not only in documenting a previously un- known pattern of tooth formation but are also important to recent discussions of molluscan evolution. This is true because the ontogeny ofthe chiton radula has been used as a prime case in favor of a bilateral yet monostichous ancestral condition. In order to appreciate both the underlying assumptions and previ- ously stated support for this idea, some re- view 15 necessary. The discovery of living monoplacophorans and descriptions of their anatomy (Lemche and Wingstrand, 1959; Wingstrand, 1985) has again brought to prominence the often suggested hypothesis that metamerism is a basic feature of mollusks, perhaps a primitive condition shared with other metameric pro- tostomium ancestors. Organs are also re- peated in polyplacophorans (chitons) and in the cephalopod genus Nautilus as was discussed in depth by Naef (1926) and previous authors (for review see Wingstrand, 1985). Particularly striking are the repetition of kidneys, atria and gills in monoplac- ophorans, polyplacophorans, and in Nautilus. Other authors regarded the metameric condi- tion as a convergence (Hoffmann, 1937; Boettger, 1959; Yonge, 1960; Salvini- Plawen, 1985) and argued that single paired systems were present in a hypothetical mol- luscan ancestor. Wingstrand (1985) supports grouping the sister groups Polyplacophora and Conchifera (the latter group including monoplacophorans) as a monophyletic unit, the “Testaria” (Salvini- Plawen, 1972; 1980; Lauterbach, 1983), itself a sister group to the either mono- or biphyletic aplacophoran mollusks (i.e. the Caudovo- veata and the Solenogastres). In support of this view, Wingstrand describes many testar- ian synapomorphic features including the radula and radular apparatus, the velum, the subradular organ, the large pharyngeal diver- ticula, the large digestive gland, the coiled in- testine, the eight pairs of pedal retractor groups, and the already mentioned similarities of the heart complex. As Wingstrand (1985) has noted, even if as he has concluded, metamerism is primitive for testarians, it is dif- ficult to determine whether a basic metameric organization is a plesiomorphic condition for testarians, present also in a protostomian an- cestor or, alternatively, if metamerism is a synapomorphy for testarians. Only the “Apla- cophora” are available for outgroup compar- ison and their nonmetameric condition could either be a primitive molluscan feature or at- tributed to convergent evolution due to a vermiform habit or neotenic reductions result- ing from small adult body size. The serial na- ture of all known molluscan radulae might pro- vide insight on the issue of metamerism but, not surprisingly, there is little general agree- ment on the features that are primitive to a radula. First, there are obviously two issues concerning the presumed serial or nonserial nature of the primitive radula, differing in whether the “metamerism” is bilateral (left and right) or longitudinal (serially repeated rows). Nierstrasz (1905) and Boettger (1955, 1959) proposed that the basic ancestral radula was bipartite (i.e. in two parts, symmetrical left and right) and distichous (i.e. arranged with two matched teeth in each longitudinally repeated row), while Salvini-Plawen (1972, 1978, 1981, 1985) contended it had a broad monostichous form. Meanwhile, Minichev and Sirenko (1974) and Ivanov € Tzetlin (1981) attributed to the Aplacophora and Polyplacophora a pri- marily monstichous radula and to the Conchifera a polystichous (Minichev and Sirenko, 1974) or a distichous (Ivanov & Tzetlin, 1981) radula. Because aplacophorans have been re- garded as the one or two earliest diverging of extant molluscan lineage(s) (i.e. WingStrand, 102 EERNISSE & KERTH 1985) and because they have the simplest adult radula, the aplacophoran radula might be especially appropriate to consider. However, the highly specialized and diverse modes of feeding of many aplacophorans, especially those that are interstitial, could confound this conclusion. For example, about 25% of known members of the Solenogastres (Neomenio- morpha) lack a radula, using enzymatic se- cretions of a protrusible foregut to dissolve cnidarian tissue (Salvini-Plawen, 1985). The Caudofoveata (Chaetodermomorpha or Chae- todermatida) include several genera with a distichous radula and several that display a specialized feeding apparatus of disputed construction, whereas the distichous tooth pattern prevails unequivocally in members of the Solenogastres that have a radula (Salvini- Plawen, 1978). Moreover the radula of both aplacophoran groups clearly shows features of a bipartite construction: The teeth attach to a radular membrane which is often split medi- ally, perforated by a series of slits, or is fused together from two ribbons (Heath, 1905; Salvini-Plawen, 1978; Scheltema, 1981; pers. comm. 1984; contrast with Hyman, 1967). We have presented evidence that from the start the juvenile chiton radula is in most cases polystichous, not monostichous as believed by Minichev and Sirenko (1974). However, we believe the issue is much more fundamental than this distinction. Even if it could be shown that certain mollusks (i.e. two species of Solenogastres as claimed by Salvini-Plawen, 1985) pass through a monostichous stage in their radular formation, this would not neces- sarily indicate the primitive radular condition of a presumed early molluscan ancestor. Com- parative ontogenetic studies might reveal the initial ancestral state (i.e. “Von Baer's laws”) but this assumes that early ontogenetic stages are less prone to modification than later stages or, stated differently, the initial expression of a morphological trait reflects more accurately than later expressions an ancestral condition (Kluge and Strauss, 1985). In practice, testing this assumption requires outgroup compari- son (Kluge, 1985) which in the present case is difficult because it would require comparisons of radular ontogeny with the ontogeny of a structure presumed to be homologous to the molluscan radula in a non-molluscan out- group. Moreover, it would be mistaken to as- sume that the initial state of a juvenile chiton radula must correspond to the adult radula of a hypothetical ancestral mollusk. Instead, the juvenile condition of chitons is better com- pared to the ancestral juvenile condition, and simplicity (i.e. few teeth or even a monostich- ous condition) attributed to the inherently small size of juveniles. Thus, there is no longer any reason to pos- tulate that the presumed ancestor of early mol- lusks was equipped with a monostichous radula as suggested by Salvini-Plawen (1985), Minichev and Sirenko (1974), and Ivanov and Tzetlin (1981). In addition to the uncertainties inherent in using early on- togenetic stages to infer a primitive condition, two facts are incompatible with the suggestion of an ancestral monostichous condition. First, the predominant radula type of the Apla- cophora is distichous and basically bipartite, even if there is an initial connection in the juvenile radula as claimed by Salvini-Plawen (1985). Second, none of six genera of chitons hitherto examined (this paper and Minichev and Sirenko, 1974) reveal any sign of a monostichous stage in their radular develop- ment, except for the “monostichous” radula of the unidentified “trochophore” depicted in Minichev and Sirenko (1974). We would reinterpret this latter case as distichous, con- sisting of two longitudinal rows of incomplete L, teeth. Both the prevailing aplacophoran radula type and the ontogenetic sequence of the chiton radula lead us to propose a rather different basic feeding instrument in early mol- lusks. We conclude that it was bilateral or even bipartite with one or more pairs of longitudinal rows of teeth. It would be tempting to assume that such a radula represents the ancestral type for all mollusks, but this extrapolation needs to be tested with additional comparative studies. ACKNOWLEDGEMENTS DJE acknowledges support by NSF Grant OCE-8415258 to R.R. Strathmann and DJE, and thanks Dr. A.O.D. Willows, Director, Fri- day Harbor Laboratories, University of Wash- ington, and Dr. A. Fontaine, Electron Micro- scope Facility, University of British Columbia, Victoria, B.C., for making those facilities avail- able, and Dr. E.N. Kozloff for help with trans- lations. KK thanks U. Holzöder (Würzburg) for excellent cooperation. LITERATURE CITED BOETTGER, C., 1955, Beiträge zur Systematik der Urmollusken (Amphineura). Zoologischer An- zeiger, Supplement, 19: 223-256. INITIAL DEVELOPMENT OF THE CHITON RADULA 103 BOETTGER, C., 1959, Discussion in: Protostomian interrelationships in the light of Neopilina by H. LEMCHE. Proceedings XV International Con- gress of Zoology, London, Section 4, 386-389. САВЕРООТ, T. H., 1965, Magnetite in the radula of the Polyplacophora. Proceedings of the Malacological Society of London, 36: 203-212. EERNISSE, D. J., 1984, Lepidochitona Gray, 1821 (Mollusca: Polyplacophora) from the Pacific coast of the United States. Ph.D. Dissertation. Univ. of California, Santa Cruz, 1984: 1-358. EERNISSE, D. J., 1986, The genus Lepidochitona Gray, 1821 (Mollusca: Polyplacophora) in the northeastern Pacific Ocean (Oregonian and Cal- ifornian Provinces). Zoologische Verhandelingen (Leiden), 228: 3-52. HEATH, H., 1905, The morphology of a Solenogastre. Zoologische Jahrbücher, Ab- teilung Anatomie, 21: 703-773. HOFFMANN, Н., 1937, Uber die Stam- mesgeschichte der Weichtiere. Zoologischer Anzeiger, Supplement, 10: 33-69. НУМАМ, L. H., 1967, The Invertebrates, Vol. 6, Mollusca 1. McGraw Hill Book Co., N.Y., London, 792 pp. IVANOV, D. L. 8 TZETLIN, А. B., 1981, О. _:. and evolution of the cuticular pharyngeal armauure in trochophore animals having the ventrai pharynx. Zoologicheskii Zhurnal, 60: 1445-1454 [in Rus- sian]. KERTH, K., 1979, Phylogenetische Aspekte der Radulamorphogenese von Gastropoden. Mal- acologia, 19: 103-108. KERTH, К. 1983a, Radulaapparat und Radulabildung der Mollusken. |. Vergleichende Morphologie und Ultrastruktur. Zoologische Jahr- bucher, Abteilung Anatomie, 110: 239-269. KERTH, К. 19836, Radulaapparat und Radulabildung der Mollusken. Il. Zahnbildung, Abbau und Radulawachstum. Zoologische Jahrbücher, Abteilung Anatomie, 110: 239-269. KERTH, K. & HÄNSCH, D., 1977, Zellmuster und Wachstum des Odontoblastengürtels der Wein- bergschnecke Helix pomatia. Zoologische Jahrbücher, Abteilung Anatomie, 98: 14-28. KLUGE, A. G., 1985, Ontogeny and phylogenetic systematics. Cladistics, 1:13-27. KLUGE, A. G. & STRAUSS, R. E., 1985, Ontogeny and systematics. Annual Review of Ecology and Systematics, 16: 247-268. LAUTERBACH, K.-E., 1983, Erórterungen zur Stammesgeschichte der Mollusca, insbesondere der Conchifera. Zeitschrift für Zoologische Systematik und Evolutionsforschung, 21: 201-216. LEMCHE, H. & K. G. WINGSTRAND, 1959, The anatomy of Neopilina galatheae Lemche, 1957. Galathea Report, 3:9-71. LOWENSTAM, H. A., 1962, Magnetite in denticle capping in recent chitons (Polyplacophora). Bul- letin of the Geological Society of America, 73: 435-438. MINICHEV, Y. $5. & В. I. SIRENKO, В. 1., 1974, Development and evolution of radula in Polyplacophora. Zoologicheskii Zhurnal, 53: 1133-1139 [in Russian]. NAEF, A., 1926, Studien zur generellen Mor- phologie der Mollusken. 3. Theil. Ergebnisse und Fortschritte der Zoologie, 6: 27-124. NIERSTRASZ, H. F., 1905, Kruppomenia minima und die Radula der Solenogastren. Zoologische Jahrbücher, Abteilung Anatomie, 21: 655-702. PRUVOT, G., 1890, Sur le developpment d’un solenogastre. Comptes Rendus hebdomadaires des Séances de l'Académie des Sciences [Paris], 111: 689-692. SALVINI-PLAWEN, L.v., 1972, Zur Morphologie und Phylogenie der Mollusken: Die Beziehungen der Caudofoveata und der Solenogastres als Aculifera, als Mollusca und als Spiralia. Zeit- schrift für wissenschaftliche Zoologie, 184: 205-394. SALVINI-PLAWEN, L.v., 1978, Antarktische und subantarktische Solenogastres (Eine Mono- graphie: 1898-1974). Zoologica (Stuttgart), 44: 1-305. SALVINI-PLAWEN, L.v., 1980, A reconsideration of systematics in the Mollusca (phylogeny and higher classification). Malacologia, 19: 249-278. SALVINI-PLAWEN, L.v., 1981, The molluscan di- gestive system in evolution. Malacologia, 21: 371-401. SALVINI-PLAWEN, L.v., 1985, Early evolution and the primitive groups. In TRUEMAN, Е. В. & CLARKE, М. R., ed., The Mollusca, “Evolution,” 10: 59-150. Academic Press, London. SCHELTEMA, A., 1981, Comparative morphology of the radulae and alimentary tract in the Aplacophora. Malacologia, 20: 361-383. SIRENKO, В. I., 1975, A new subfamily of chitons Juvenichitoninae (Ischnochitonidae) from the north-west Pacific. Zoologicheskii Zhurnal, 54: 1442-1451 [in Russian]. SIRENKO, B., & MINICHEV, Y., 1975, Développe- ment ontogénétique de la radula chez les Polyplacophores. Cahiers de Biologie Marine, 16: 425-433. TOWE, K. M. & LOWENSTAM, H. A., 1967, Ultrastructure and development of iron mineral- ization in the radular teeth of Cryptochiton stelleri (Mollusca). Journal of Ultrastructure Research, Wie: WINGSTRAND, K. G., 1985, On the anatomy and relationships of Recent Monoplacophora. Galathea Report, 16: 7-94, 12 pl. YONGE, C. M., 1960, General characteristics of Mollusca. п MOORE, В. C., ed., Treatise on Invertebrate Paleontology, Part 1 (Mollusca 1): 3-36. Geological Society of America and Univer- sity of Kansas, Lawrence, Kansas. Revised Ms. accepted 17 March, 1987 MALACOLOGIA, 1988, 28(1-2): 105-117 LA VARIABILIDAD DE HEMICYCLA BIDENTALIS (GASTROPODA, HELICIDAE)' М. Ibanez, J. Barquín, E. Cavero & М. В. Alonso Departamento de Zoologia, Facultad de Biologia. Universidad de La Laguna, Tenerife, Islas Canarias, Espana RESUMEN Se realiza un estudio biométrico de Hemicycla bidentalis, endémica de la isla de Tenerife (Archipiélago canario). Es una especie extraordinariamente variable, tanto con respecto a la concha como al aparato reproductor, llegando a ser tan grandes las diferencias conquiológicas entre algunas poblaciones que éstas parecen pertenecer a especies diferentes. Sin embargo, al ser graduales estas variaciones y al no existir separación geográfica entre ellas, se concluye que el flujo genético no ha sido interrumpido. Solo existe una poblacion aislada geográficamente del resto en época reciente, pero no muestra indicios de un proceso de especiacion. La variabilidad de H. bidentalis está relacionada con los tipos de vegetación sobre los que vive (íntimamente relacionados a su vez con las correspondientes características climáticas у altitudinales), de los que los principales son la laurisilva, la zona de transición y la zona basal; esta variación se debe a la extraordinaria capacidad de adaptación de la especie al biotopo, al igual que occure con /berus gualtierianus en la península ibérica, por lo que las poblaciones más diferenciadas son, simplemente, ecotipos de ella. Key words: Helicidae; Hemicycla; variability; biometry. INTRODUCIÓN La variabilidad en los gasterópodos pul- monados constituye un fenómeno cuya e- xistencia prácticamente era desconocida hasta el siglo 19, en el que ya comenzaron a publicarse algunos datos interesantes, como los de Kobelt (1881) sobre las poblaciones de Murella en Sicilia. A lo largo de este siglo, en cambio, son bastante numerosos los trabajos sobre variabilidad. Boettger (1913) mostró una seriacion de conchas de diferentes taxones de /berus de la Península Ibérica, entre el globoso /. alonensis y el aquillado I. gual- tierianus; Pfeiffer (1931) y Rensch (1937) trataron mas de 30 variedades de Murella, indicando Rensch que las formas globosas y aquilladas podrían estar relacionadas con los climas secos y cálidos; Biggs (1959) concatena varias formas de Егетта; Heller (1979) también estudia en este sentido el género Levantina; Alonso & Ibanez (1978) muestran una seriación entre el aquillado Iberus rositai, que vive en una zona kárstica, y el globoso-subdeprimido /. loxanus, que se encuentra en los alrededores de esta zona; y Bartolomé (1982) revisa la literatura sobre este tema, añadiendo varios ejemplos de los géneros ya reseñados y algunos otros (Tyrrheniberus, Rossmaessleria, Macularia, etc.); López-Alcántara & cols. (1983, 1985) y Alonso & cols. (1985) realizan un estudio estadístico y biogeográfico de la variabilidad en el género /berus, con I. alonensis e I. gualtierianus, concluyendo que ambas formas son ecotipos de la misma especie; y el caso más espectacular es el tratado por Woodruff (1978) y Woodruff & Gould (1980), que estudian la exuberante di- versidad morfológica de las conchas de Cerion en las islas del Caribe y en Florida, indicando que el género está constituído por una serie de semiespecies variables y politípicas. En Canarias, en la isla de Tenerife (Fig. 1), existe otro notable caso de variabilidad relacionada con el biotopo en Hemicycla bidentalis (Lamark, 1821) (syn. = malleata Férussac, 1821), que habita fundamental- mente en la zona montañosa de Anaga, al NE de la isla, en 3 tipos básicos de vegetación ‘Notes on the Malacofauna of the Canary Island, Nr. 10; Nr. 9: Revision of the genus Hemicycla Swainson 1840 (Mollusca: Helicidae) from Tenerife: 1 n. subgen. and description of 3 new taxa. Bull. Mus. Paris (in press). Work supported by project 1692/82 of the “Comisión Asesora de Investigación Científica y Técnica” of Spain (CAICYT). (105) 106 IBANEZ, BARQUIN, CAVERO & ALONSO "20 30 en 240 50 Hemicycla bidentalis (LAMARCK 1821) FIG. 1. Hemicycla bidentalis. Distribución geográfica. В: Benijo; I: Igueste de San Andrés; P: Palo Blanco; e: localidades de procedencia del material estudiado. (Fig. 2): el primero en las zonas altas, la laurisilva, bosque subtropical termófilo muy húmedo, relíctico del terciario; el segundo en las zonas bajas, el piso basal, formado por arbustos y matorrales xéricos de influencia africana, con muchas especies crasas del género Euphorbia, realizándose el paso de una a otra a través del tercero: el piso de transición. Dentro de esta especie hay un conjunto amplio de poblaciones con características conquiológicas que a veces difieren de tal forma que a primera vista algunas de ellas parecen pertenecer a especies distintas (Fig. 3); esto ocurre al comparar la forma típica, de la laurisilva del macizo de Anaga, con la forma extrema del piso basal de Igueste de San Andrés y con un taxón fósil del Cuaternario parecido al de Igueste, H. colla- rifera Boettger, 1908, cuya localidad típica es Tejina, en la vertiente Norte del macizo de Anaga (Boettger, 1908). Pero entre ellas no hay aislamiento geográfico y, además, hemos observado un cambio gradual entre las 2 primeras a través de poblaciones intermedias, por lo que pensamos que no ha cesado el flujo genético y no pueden, por tanto, considerarse como especies distintas. Con menor espectacularidad, se diferencian también otras poblaciones, destacando la de Palo Blanco, que en la actualidad está aislada del macizo de Anaga por la acción humana (agrícola y urbanística), por una franja de unos 25 km de ancho (Fig. 1). MATERIAL Y MÉTODOS Para certificar su identificación, hemos comparado nuestro material (depositado en LA VARIABILIDAD DE HEMICYCLA BIDENTALIS 107 2 3 a 5 6 ZA FIG. 2. Corte esquemático del macizo de Anaga, entre las localidades de Benijo e Igueste de San Andrés, mostrando la altitud a la que se encuentran los 3 tipos básicos de vegetación. LAU: laurisilva; TRA: piso de transición; PBS: piso basal; N: vertiente Norte; S: vertiente Sur. el Departamento de Zoología de la Uni- versidad de la Laguna, DZUL) con el de algunos museos: fotografías de los sintipos, del Museum National d'Histoire Naturelle, Paris, enviadas por Mr. K. Groh (SMF); 1 concha de Santa Cruz, del Naturmuseum Senckenberg, Frankfurt/Main (SMF 33.635); 1 concha de La Paz (La Orotava; RNHL 50804) y 3 de Agua Garcia (RNHL 50805), del Rijkmuseum van Natuurlijke Historie, Leiden; y 22 conchas de Taganana, 10 del Barranco de San Antonio (La Orotava), 22 de las Mercedes y 20 de las cumbres de Anaga, del Museo Insular de Ciencias Naturales de Tenerife. Para el estudio estadistico, realizado con el ordenador Digital VAX/VMS de la Uni- versidad de la Laguna, se han recolectado 1698 ejemplares adultos (1550 conchas y 148 vivos), procedentes de diversas lo- calidades (Fig. 1), de los que se extrajo el aparato reproductor en buen estado a 100 individuos. Las variables analizadas fueron: — De la concha: diámetro (D), altura (H), altura de la última vuelta (HU) y los índices D/H, D/HU y H/HU. — Del aparato reproductor (longitudes): pene (PE), epifalo (E), flagelo (F), conducto común (CC), conducto de la bolsa copulatriz (BC), divertículo (DI) y los índices PE/E, F/PE, CG/PE, BG/PE; DI/PE, F/E,€C/E, BC/ E DI/E; F/GGC,E/BG, EiDI, ВС/СС, DICC y BC/DI (eligiendo siempre en el numerador la variable de media más alta para el conjunto de la población). Se realizaron 5 análisis estadísticos: 1. Un análisis bivariante de la correlación entre todos los pares de variables posibles, en los casos en que se tenía información de todas ellas para cada ejemplar adulto (100 en total: Fig. 11), estudiándolos por separado según los 3 tipos básicos de vegetación en que se encontraron los poblaciones (LAU, laurisilva; TRA, piso de transición; y PBS, piso basal), obteniendo gráficos con las correspondientes nubes de puntos y los coeficientes de correlación producto-mo- mento (r) entre ellas. 2. Otro análisis bivariante similar comparando sólo los datos conquiológicos de conchas adultas (1237 casos: Fig. 12), obteniendo además sus curvas de regresión respectivas, segun la expresión у = ax®. IBANEZ, BARQUIN, CAVERO & ALONSO FIG. 3. Hemicycla bidentalis. A) Forma tipica, de las cumbres de Anaga. B) Poblaciön de Igueste de San Andres. C) Hemicycla collarifera, fösil de Bajamar (escala, 5 mm). LA VARIABILIDAD DE HEMICYCLA BIDENTALIS 109 FIG. 4. Hemicycla bidentalis. Serie de conchas en las que se puede apreciar la transiciön desde la forma típica de la laurisilva de Anaga hasta la población extrema de Igueste de San Andrés. El cambio gradual se aprecia tanto en la escultura como en las denticulaciones (escala, 5 mm). A: Cumbres de Anaga (laurisilva, 800 m); В: ljuana (laurisilva, 700 m); С: Bco. Roque Bermejo (piso de transición, 450 т); D: Benijo (piso basal, 200 m); E: Bco. de Anosma (piso basal, 200 m); Е: Igueste de San Andrés (piso basal, 100 m); G: Igueste de San Andrés (piso basal, 80 m); Н: Igueste de San Andrés (piso basal, 200 m). NOTA: Una serie similar a la fotografiada en esta lámina está depositada en las colecciones de la Academia de Ciencias Naturales de Philadelphia (ANSP 361423-361427). 110 IBANEZ, BARQUIN, CAVERO & ALONSO FIG. 5-9. Hemicycla bidentalis. 5) Detalle de la protoconcha de la forma tipica. 6-8) Detalle de la penultima y última vueltas de espira. 6) Población de las cumbres de Anaga. 7) Población de Benijo. 8) Población de Igueste de San Andrés. 9) Rádula; detalles de los dientes central, laterales y marginales. Escala: 5-8) 600 um; 9) 25 pm. LA VARIABILIDAD DE HEMICYCLA BIDENTALIS dell FIG. 10. Hemicycle bidentalis. Aparato reproductor (escala, 5 mm). A) Forma típica. В) Población de Igueste de San Andrés (el atrio y la porción anterior del pene están evaginados, mostrando la papila accessoria del pene). C) Población de Palo Blanco. 3. Un análisis discriminante por etapas (stepwise discriminant analysis, BMDP7M) entre las variables e índices conquiológicos (1237 casos) para los ejemplares procedentes de las poblaciones anteriores (LAU, TRA y PBS), pero dividiendo la primera en 2: LA1 (LAU de Anaga) y LA2 (LAU de Palo Blanco), para estudiar la posible segregación de esta última por su aislamiento geográfico actual con respecto a la primera. En estos mismos ejemplares se midió el grado de or- namentación (GO) en una escala arbitraria del 1 al 5, siendo mínimo (GO = 1) en conchas lisas y máximo (GO = 5) en las más rugosas y costuladas. 4. Otro análisis discriminante entre las variables e índices del aparato reproductor (100 casos) para comprobar si existen diferencias significativas entre las mismas poblaciones del análisis anterior. 5. Un test de t-student para igualdad de medias entre las variables e índices con- quiológicos (incluyendo GO) y del aparato reproductor, para comprobar si existen diferencias entre LA1 y LA2. RESULTADOS A) Descripción de la forma típica y de sus modificaciones 1. Forma típica: Se encuentra en la laurisilva de Anaga y de Palo Blanco. La concha es gruesa, imperforada, globosa- cónica, con 4: vueltas de espira, con suturas marcadas (Fig. 3A); su color es verdoso o amarillento claro en algunos ejemplares, pero sobre él se situan generalmente 5 bandas más oscuras, que a veces se fusionan entre sí, dándole un color oscuro uniforme; es ligeramente brillante, sobre todo en su superficie basal, que es más lisa, y en las 2 últimas vueltas tiene una maleación fina y uniforme (Fig. 6), mientras que las primeras poseen una débil costulación que se cruza con una leve estriación espiral existene en 112 IBANEZ, BARQUIN, CAVERO & ALONSO TABLA 1. Matriz de correlaciön entre las variables. Los coeficientes mayores que 0.25 son sig- nificativos al nivel de P = 99% y todos los son al nivel de P = 95%, para n = 100. Concha; D: diámetro; H: altura; HU: altura de la última vuelta. Reproductor: PE: pene; E: epifalo; Е: flagelo; CC: conducto común; BC: conducto de la bolsa copulatriz; DI: divertículo. toda la concha, dando lugar en algunas zonas a la formación de gránulos. La protoconcha es más oscura y ligeramente rugosa (Fig. 5). La ultima vuelta es muy globosa, ligeramente angulosa en su origen, pero sin quilla; presenta una estrangulación en las proximidades del peristoma, originando una pequeña gibosidad. El peristoma es blanco y presenta 2 fuertes calosidades, una en el punto de inserción del margen superior y la otra en el margen externo, estando el margen columelar engrosado por dentro. 2. Modificaciones hacia la costa: Al descender desde las cumbres de Anaga hacia la costa, se modifican la forma y la ornamentación de la concha: el color se hace más oscuro y uniforme y desaparece el brillo al aumentar la granulación y marcarse más la estriación espiral (Fig. 7); el tamaño también aumenta, al aumentar en Ya el número de vueltas de espira, sin que aumente la altura; y las callosidades del peristoma se van haciendo más tenues (Fig. 4). En los alrededores de Igueste de San Andrés (en el Sur de Anaga) se produce un cambio hacia una forma extrema, de aspecto completamente diferente (Fig. 3B): de forma gradual y en una distancia aproximada de Y km (en la misma curva de nivel) desparecen las maleaciones, se mantiene muy fuerte la estriación espiral y se marca mucho más la costulación, originando fuertes costillas (Fig. 8); la forma se hace más deprimida y aqui- llada y la abertura es más redondeada, al desaparecer la callosidad del margen supe- rior y reducirse a un vestigio la del margen TABLA 2. Dimensiones extremas y medias (en mm) de las variables estudiadas para el conjunto de las poblaciones. CV: coeficiente de variación; n: número de casos. Los demás símbolos utilizados son los mismos que en la Tabla 1. MAX ММ MEDIA CV(%) n О 25.8 15.9 21.43 6 1237 H 17.9 la 14.45 8 1237 HU 13.2 8.5 10.84 7. 1237 PE 13.5 4.5 9.50 16 100 E 8.0 2.0 3.91 25 100 F 24.0 8.5 ZZ 18 100 CC 17.0 4.0 10.08 23 100 BC 16.0 6.5 11.42 iS 100 DI 21.0 6.0 1178 25 100 externo (Fig. 3B). Esta forma se parece en su ornamentación y microescultura al taxón fósil H. collarifera, de la vertiente Norte de Anaga (Fig. 3C). 3. Anatomía interna: La rádula (Fig. 9) tiene en todas las poblaciones la estructura típica del género, con la siguiente fórmula: (C + 10-12L + 28-37M) x 110-130. El aparato reproductor (Fig. 10) varia extra- ordinariamente, tanto de unas poblaciones a otras como dentro de una poblaciön cualquiera. B) Resultados de los analisis estadisticos En la Tabla 1 se muestran los coeficientes de correlación entre todas las variables medidas; aun siendo significativas en su mayor parte, destaca el bajo valor de la mayoría de las correlaciones, sobre todo de las que presentan las variables del genital entre sí y con las demás. En la Tabla 2 se resumen las dimensiones extremas y medias de todas las variables para el conjunto de las poblaciones y los coeficientes de variación. De los análisis bivariantes por poblaciones se deduce que no se pueden establecer zonas de discontinuidad entre ellas (Figs. 11 y 12), siendo destacable la enorme va- riabilidad de las diversas partes del aparato reproductor, tanto en conjunto como dentro de cada población. Los análisis discriminantes muestran que las variables e índices conquiológicos que mejor caracterizan a cada subpoblación son D, H, y HU para la concha, y CC, BC/PE, F/E y F/DI para el reproductor, en este orden; las demás lo hacen con valores no significativos. Los coeficientes para las variables canónicas LA VARIABILIDAD DE HEMICYCLA BIDENTALIS 113 21- 18- =. Aree >- | 19CC 21 BC FIG. 11. Hemicycla bidentalis. Contornos de las nubes de puntos obtenidas al comparar las variables del reproductor 2 a 2. Linea de trazo continuo: laurisilva; linea de trazo discontinuo: piso de transiciön; linea de puntos: piso basal (n = 100 casos). 114 IBANEZ, BARQUIN, CAVERO & ALONSO FIG. 12. Hemicycla bidentalis. Contornos de las nubes de puntos y lineas de regresiön obtenidas para las variables conquiológicas. Los símbolos son los mismos que en la Fig. 11; DH: пасе D/H (п = 1237 casos). LA VARIABILIDAD DE HEMICYCLA BIDENTALIS IS TABLA 3. Coeficientes de los análisis discriminantes de la concha y del aparato reproductor para las variables canónicas |, Пу Ill; las demás variables no se indican, al no ser suficientemente discriminantes. PA: porcentaje de la dispersión total explicado por cada variable canónica. Los demás símbolos utilizados son los mismos que en las Tablas 1 y 2. CONCHA (n = 1237) REPRODUCTOR (n = 100) | Il Ш | Il Ш РА 88.8 10.9 0.3 67.9 28.0 4.1 D 1.12 0.24 0.17 CC —0.40 0.13 0.14 H —0.64 —0.18 1252 BC/PE 0.18 2.73 4.95 HU —0.62 —1.42 2102 F/E 0.30 OA —0:57 F/DI | Si) 0.56 0.31 TABLA 4. Porcentajes de clasificación correcta obtenidos en los análisis discriminantes para las variables canónicas de la tabla 3. LA1: laurisilva de Anaga; LA2: laurisilva de Palo Blanco; TRA: piso de transición; PBS: piso basal; PCOR: porcentaje correcto; n: número de casos. CONCHA REPRODUCTOR PCOR LA1 LA2 TRA PBS TOTAL PCOR LA1 LA2 TRA PBS TOTAL n 318 202 316 401 1237 14 48 12 26 100 TOTALES 561.5 25.7 163 256 324 100 69.0 14.0 48.0 12.0 26.0 100 LA1 663 663 16.3 14.1 3.3 100 78.8 78.8 7.8 3.8 9.6 100 LA2 64.7 12:9) 164.7 20.0 2.4 100 85.7 9.5 85:7 0.0 4.8 100 ТАА 46.1 135 18.0 46.1 22.4 100 38.5 231 15.3 38:5: 23.1 100 РВ$ 67.7 4.3 6.4 21.6 67.7 100 35.7 14.3 14.3 35.7 35.7 100 I, II y Ill se exponen en la Tabla 3 y los porcentajes de clasificación correcta según las funciones de clasificación se exponen en la Tabla 4. Se observa que para la concha el 61.5% de los casos se clasifican en su grupo correspondiente, existiendo por tanto algo mas de una tercera parte que puede ser clasificada en un grupo diferente al suyo; para el reproductor, en cambio, lo hace el 69%, por lo que las variables del reproductor son ligeramente más discriminantes que las conquiológicas. En la Fig. 13 se representa el contorno de las nubes de puntos en el plano definido por las 2 primeras variables canónicas, en donde está representado el 99.7% y el 95.9%—respectivamente—de la dispersión total, y los centroides de cada grupo para la concha (Fig. 13A) y para el aparato reproductor (Fig. 13B). En ambos casos las 4 poblaciones se solapan entre sí, correspondiéndose los solapamientos entre cada nube de puntos y las demás con los porcentajes de la Tabla 4. Prácticamente no existe ninguna frontera o separación entre los 4 grupos, aun empleando las 2 combinaciones lineales (variables canónicas) de las variables que más separan los 4 grupos entre sí; para ambas figuras, el centroide de LA2 se separa de los demás tanto como el de LA1; también es destacable la proximidad de los centroides de TRA y PBS en el caso del reproductor (Fig. 13B). En la Tabla 5 se muestra la variación del grado de ornamentación según el tipo de vegetación, cuyo valor aumenta desde la laurisilva hacia la costa; esta variable, junto con el desarrollo de las callosidades del peristoma, son las más conspicuas en la diferenciación de las poblaciones, aunque son difíciles de cuantificar para su estudio estadístico; también se indican las medias por poblaciones y el resultado del test de t-student para comparar las 2 poblaciones de laurisilva (LA1 y LA2), que muestra que las diferencias entre ambas poblaciones son significativas salvo para las variables BC y DI, siendo mayores los valores correspondientes а LA2. DISCUSIÓN Tras el examen de los coeficientes de correlación mostrados en la Tabla 1, se observa que éstos son bastante bajos 116 IBANEZ, BARQUÍN, CAVERO & ALONSO 11 LA en TR A L 4 . PBS LA1 . = = ¡ES | eTRA A ы | | 11 | 4 4 A © | Y | R ТА | Т « PBS | A dE iz — | . LA2 A Pe FIG. 13. Hemicycla bidentalis. Resultados de los análisis discriminantes para la concha (A; n = 1237) y para el aparato reproductor (B; n = 100); se muestra el contorno de las nubes de puntos en el plano definido por las 2 primeras variables canónicas y los centroides de cada grupo. LA1: laurisilva de Anaga (trazo continuo); LA2: laurisilva de Palo Blanco (trazo de raya-punto-raya); TRA: piso de transición (trazo discontinuo); PBS: piso balsa (línea de puntos). excepto entre CC-F y BC-CC y entre las medidas de la concha lo que indica que, salvo estas excepciones, la variabilidad de cada uno de los parámetros medidos no se relaciona con la variabilidad de los demás lo suficiente como para poder afirmar que son taxones diferentes. A esta misma conclusión se llega tras el exámen de las nubes de puntos obtenidas en los análisis bivariantes (Figs. 11 y 12), que poseen contornos generalmente redondeados sin que se TABLA 5. Valores medios de las variables estudiadas, por poblaciones. *: diferencias significativas entre las medias de las dos poblaciones LA1 y LA2 al nivel de Р = 95%, segun el test de t-student para igualdad de medias. GO = grado de ornamentación; los demás símbolos utilizados son los mismos que en las Tablas 1 y 4. LA1 LA2 TRA PBS n 368 85 317 467 D 19.89 21.30 21:92 22.34 * Н 14.06 15.04 14.87 14.36 * HU 10.42 11.36 AS 10.86 + GO 1.1 1.0 2.5 3.3 п 52 21 13 14 * РЕ 9.01 10.93 9.58 9.09 * Е 3.49 5.36 3.38 3.80 * Е 14.87 19.93 19.26 19.60 * СС 8.34 13.14 11.31 10.78 ВС 161215 11.19 11.69 12.48 DI 11.04 12.57 11.81 13.07 aprecie un alargamiento notable еп ninguna dirección, observándose también un amplio solapamiento de las 3 poblaciones. El mismo resultado se obtiene con los 2 análisis discriminates. El solapamiento de las 3 poblaciones de Anaga, junto con la inexistencia de aislamiento geográfico entre ellas, son evidencias de que no se ha inter- rumpido el correspondiente flujo genético. También se produce solapamiento con la poblacion LA2 de Palo Blanco, aislada recientemente. Por lo que respecta a la concha (Tabla 3), se observa una tendencia al aumento de D desde las poblaciones de laurisilva a las de los pisos de transición y basal (Figs. 4 y 12); H y HU no experimentan, en cambio, una variación tan fuerte, presentándose la media más alta en las poblaciones del piso de transición. En cuanto al aparato reproductor, se puede apreciar la enorme variación que ex- perimentan sus conductos (Tabla 2), así como una tendencia al aumento en las poblaciones costeras con respecto a las de laurisilva de Anaga, tendencia que cu- riosamente está más acentuada en la población de laurisilva de Palo Blanco (Fig. 13B), de lo que podría deducirse que en esta última se está esbozando un proceso de especiación. Sin embargo, aunque las diferencias entre las medias de las 2 poblaciones de LAU son significativas LA VARIABILIDAD DE HEMICYCLA BIDENTALIS 117 estadisticamente (como muestra el test de t-student; Tabla 5), son similares a las exis- tentes entre las poblaciones de LAU, TRA y PBS de Anaga excepto en el grado de ornamentaciön, en que son mucho menores, por lo que las 4 coresponden sin duda a la misma especie; por otro lado, la variaciön del grado de ornamentación que exhibe H. bidentalis con respecto a los 3 tipos de vegetación puede interpretarse como una adaptación a cada biotopo. Al no haberse interrumpido el flujo genético entre las poblaciones del macizo de Anaga, y al haberse diferenciado formas similares en el Cuaternario, todos estos cambios sólo pueden considerarse como el resultado de un proceso iterativo, iniciado a partir de una morfología común y debido a la extraordinaria capacidad de adaptación al medio ambiente de la especie, por lo que las poblaciones más diferenciades (como H. collarifera y la de Igueste de San Andrés) deben considerarse, simplemente, como ecotipos de H. bidentalis. REFERENCIAS CITADAS ALONSO, М. В. & IBÁNEZ, M., 1978, El género Iberus Montfort 1810 (Pulmonata: Helicidae). 1. Iberus rositai Fez 1950. Archiv für Mollus- kenkunde, 108: 185-192. _ ALONSO, M.R., LOPEZ-ALCANTARA, A., RIVAS, P. & IBANEZ, M., 1985, A biogeographical study of Iberus qualtierianus (L.) (Pulmonata: Helicidae). Soosiana (8th International Mal- acological Congress), 13: 1-10. BARTOLOME, J. F. M. de, 1982, Comments on some Mediterranean rockdwelling helicids. Jour- nal of Conchology, 31: 1-6. BOETTGER, O., 1908. Liste der Mollusken aus einem Sande im Barranco von Tegina auf Tenerife (Canaren). Zeitschrift der deutschen geologischen Gesellschaft, 60: 246-249. BOETTGER, C. R., 1913, Aus der Schausam- mlung. Die Veránderlickheit der Schale von Iberus gualtierianus L. Bericht sencken- bergischen naturforschenden Gesellschaft, 44: 183-197. HELLER, J., 1979, Distribution, hybridization and variation in the Israeli landsnail Levantina. Zoo- logical Journal of the Linnean Society of London, 67: 115-148. KOBELT, W., 1881. Exkursionen in Súditalien. Die sizilianischen /berus. Jahrbücher deutschen malakozoologischen Gesellschaft, 8: 50-67. LÓPEZ-ALCÁNTARA, A., RIVAS, P., ALONSO, М. В. 8 IBANEZ, M., 1983, Origen de /berus gualtierianus. Modelo evolutivo. Haliotis, 13: 145-154. LÓPEZ-ALCÁNTARA, A., RIVAS, P., ALONSO, M. В. 8 IBANEZ, M., 1985, Variabilidad de /berus gualtierianus (Linneo, 1758) (Pulmonata: Helicidae). /berus, 5: 83-112. PFEIFFER, K. L., 1931, Die Murellen Sardiniens. Abhandlungen senckenbergischen naturfosch- enden Gesellschaft, 472: 1-32. RENSCH, B., 1937. Untersuchungen über Ras- senbildung und Erblichkeit von Rassenmerk- malen bei sizilischen Landschnecken. Zeitschrift für induktive Abstammlungs- und Vererbung- slehre, 72: 564-588. WOODRUFF, D. S., 1978, Evolution and adaptive radiation of Cerion: a remarkably diverse group of West Indian land snails. Malacologia, 17: 223-239. WOODRUFF, D. $. & GOULD, S. J., 1980, Geo- graphic differentiation and speciation in Cerion— a prelimianry discussion of patterns and pro- cesses. Biological Journal of the Linnean Soci- ety, 14: 389-416. ABSTRACT THE VARIABILITY OF HEMICYCLA BIDENTALIS (GASTROPODA, HELICIDAE) М. Ibanez, J. Barquin, E. Cavero & М. В. Alonso Hemicycla bidentalis is studied biometrically. It is a very variable species: the shell differences between the various populations on the Anaga massif are so great that at first sight some of them seem to be different species. However, the overlapping sets of data points from the different populations, and considering that they are not geographically separated, gives evidence that gene flow has not been interrupted. Only one population has become geograph- ically isolated in recent times, but it shows no evidence of speciation. The variability of this species is linked with the main types of vegetation among which it lives: the very wet laurisilva (evergreen laurel forest community), the transition zone and the arid lowland zone. It is due to its extraordinary adaptative capacity to the biotope, similar to the case of реги; gualtierianus on the Iberian peninsula, that the most differentiated populations are in fact ecotypes of the same species Revised Ms. accepted 24 March 1987 MALACOLOGIA, 1988, 28(1-2): 119-130 ANATOMY AND HISTOLOGY OF THE ALIMENTARY TRACT OF THE SNAIL THEBA PISANA (GASTROPODA: PULMONATA) Carmen Roldan! & Pedro Garcia-Corrales? ABSTRACT A morphological and histological description of the digestive tube of Theba pisana is given in this paper. Light-microscope observations demonstrated that the alimentary tract is divisible into six morphologically distinct regions: buccal mass, oesophagus, crop, stomach, intestine and rectum. The alimentary canal is lined by an epithelium of the columnar monostratified type, which shows three predominant cell types: unciliated, ciliated and glandular cells. Both the unciliated and ciliated cells possess a dense brush border of microvilli. The presence of glycogen and lipid droplets in their cytoplasm suggest they have an absorptive function. The ciliated cells are also related with the continued movement of food particles in the intestinal lumen. The mucous gland cells are likely responsible for lubrication of the luminal surface on the digestive tube and their secretions help to compact the faeces and cover the faecal pellets. The major difference between the various regions of the alimentary canal is the relative number of the three epithelial-cell types. On the basis of our histological observations, there is evidence for the functional division of the T. pisana gut. The oesophagus appears to be specialized for movement of food particles. The crop serves as the storage organ. Notwithstand- ing, these regions of the digestive tube are most likely to be concerned with absorption. The stomach is very simple and lacks a gastric shield and style. The proximal ciliated and the secretory mid-intestine participate in digestion and absorption. The distal absorptive intestine and rectum are generally most important in faeces formation. The alimentary tract is surrounded by a thin, continuous layer of loose connective tissue in the middle of which are few muscle fibres, obliquely and longitudinally arranged. This musculature is responsible for the peristalsis of the digestive tube. Key words: anatomy; histology; alimentary tract; Theba pisana. INTRODUCTION Gastropods have been the subject of nu- merous studies. These studies have included gross anatomy observations and light micro- scopic investigations on their alimentary tracts. Yet surprisingly little is Known about the histology. The digestive canal has been mainly studied in the prosobranch gastropods (Wu, 1965; Brown, 1969; Demian & Michelson, 1971; Мапоа & Thiriot- Quiévreux, 1975; Sheridan et al., 1978; Bolognani-Fantin et al., 1982). In the pulmonate gastropods studies on the gut are mainly concerned with the annexed glands, such as the digestive gland. Only scattered observations have been made on the histology of the alimentary tract in pulmonates. Studies on the anatomy and histology of the pulmonate digestive canal have been carried out by Carriker (1946a) in Lymnaea stagnalis appressa Say, Ghose (1963) in Achatina fulica Bowdich, Rigby (1963, 1965) in Oxychilus cellarius (Muller) and Succinea putris (Linné), Walker (1972) in Agriolimax reticulatus (Muller). Fragmentary results have been published on Helix pomatia Linné by Ferreri (1958a, 1958b, 1961) and Sumner (1965), and on Arion ater (Linne) by Bowen (1970). In addition, comparisons have been made of the gross anatomy of the alimentary tract of helicarionid, succineid and athoracophorid snails and slugs (Tillier, 1984). To gain an appreciation of alimentary canal diversity in the gastropods, a stylomato- phoran pulmonate, Theba pisana (Muller), was examined by light and scanning electron microscopy in this work. T. pisana is a herbiv- orous snail which crops bits of plants. Preliminary observations on the anatomy and histology of the buccal mass of 7. pisana have been carried out by Roldan & Diaz Cosin (1975). We have examined the struc- ture of the epithelium in the digestive tube of T. pisana. ‘Department of Zoology, Complutense University, 28040 Madrid, Spain. 2Department of Zoology, University of Alcalá de Henares, Alcala de Henares, Madrid, Spain. (119) 120 ROLDAN & GARCIA-CORRALES The present study is aimed to reveal the regional differences in the alimentary canal of this species, as part of an ongoing study on its digestive system. Such a study is an impor- tant prelude to our attempt to correlate diges- tive activity with ultrastructural changes to the different cells of the digestive epithelium. MATERIALS AND METHODS Specimens of T. pisana were collected from Santander and Pontevedra (Spain). In- dividuals were transported to Department of Zoology, Complutense University, where they were maintained in a terrarium. The animals were fed lettuce. To show the gross morphology of the di- gestive tract, the snails were anaesthetized in a 0.1% solution of chloral hydrate for 12 hr. The shell was then removed and whole ani- mals were fixed in cold 10% neutral formalin for at least 24 hr. The snail was progressively dissected with the aid of a binocular micro- scope. When the digestive tract was uncov- ered, it was drawn under a camera lucida. Finally, the alimentary canal was opened lon- gitudinally to observe its internal morphology. A reconstruction of the stomach was made from serial sections. It was not possible to observe ciliary cur- rents and food transport within the gut of live specimens. Tissue preparation for light microscopy. The snails were directly immersed in the cold fixative fluids, and the alimentary tract was then rapidly removed from several animals. The fixation of small tissue samples, repre- senting different regions of the digestive ca- nal, were completed in the correspondent fixative fluid: Bouin's and Zenker media. 10% neutral buffered formalin and Baker's formol- calcium fluid. Fixed samples were washed, dehydrated in graded ethanols, cleared in xylene and finally embedded in paraffin (52° C). Serial sections, 3-7 jm thick, were cut on a Yung microtome, mounted on glass slides and stained with either hematoxylin- phloxin-light green, or with the methods of Heidenhains azan and Mann-Dominici (Gabe, 1968). In order to detect lipids, the tissue were fixed in cold Baker's formol-calcium fluid, then quickly dehydrated, embedded in paraffin, and sectioned at 6 рт. Staining was Бу Sudan black B (Gabe, 1968). For evidencing glycogen, the tissue were fixed in cold absolute ethanol, and the sec- tions were stained with periodic acid-Schiff's reagent (PAS) with maltase digestion as con- trol (Gabe, 1968). Tissue preparation for scanning electron microscopy (SEM). The alimentary tract was removed from several snails and divided into little segments, these were opened longitudi- nally to expose their luminal surface, and washed rapidly in three changes of cold ster- ile Locke's solution. Samples representing the different regions of the gut were then fixed immediately in 10% buffered (pH 7.4) for- malin, dehydrated through a graded series of ethanol, transferred into acetone, and dried in a Polaron model E 3000 Series II critical point drying apparatus using liquid carbon dioxide. Dried samples were mounted on metal spec- imen stubs, then sputter coated with pure gold in a Polaron model E 5000 vacuum evaporator, and viewed with an ISI SX-25 scanning electron microscope operating at 25 KV. RESULTS The digestive system of Theba pisana con- sists of a buccal mass, two salivary glands, a slender oesophagus, the large thin-walled crop, a rounded stomach, the large digestive gland, a long twisted intestinal tract, with a proximal, mid and distal portion, the rectum, and the anus which opens on the right side of the body close to the pneumostome (Fig. 1). The spheroid buccal mass is attached to the walls of the buccal cavity by numerous tensor muscles that insert onto its entire sur- face. This organ was previously studied by Roldan 8 Diaz Cosin (1975). The salivary glands originate from both sides of the dorso- posterior portion of the buccal mass, just above the pharyngo-oesophageal connec- tion. These glands are relatively wide but elongate organs that extend back over the top of the oesophagus. They join the posterior end of the buccal mass by narrow ducts near the beginning of the oesophagus (Fig. 1). The digestive gland is the largest organ in the animal and comprises a substantial portion of the posterior region of the visceral mass. This gland surrounds the stomach and intestine, its two ducts opening into the stomach. The alimentary canal is lined by a columnar monostratified epithelium which shows three predominant cell types at the light microscope level. It is composed mainly of columnar cells, HISTOLOGY OF THE ALIMENTARY TRACT OF THEBA PISANA 121 FIG. 1. Theba pisana. Dorsal view of digestive system. Salivary glands and digestive gland re- moved. bm, buccal mass; с, crop; 99а, duct of digestive gland; i, intestine; oe, oesophagus; r, rectum; s, stomach; sgd, duct of salivary gland. Scale bar 5 mm. ciliated cells and gland cells (Fig. 2). The columnar and ciliated cells typically possess a brush border of microvilli. The whole digestive epithelium rests upon a thin, continuous layer of loose connective tissue interspersed with a few muscle fibres, which are obliquely and longitudinally arranged (Fig. 2). Two major differences are observed be- tween the various regions of the digestive tract. The first is the cell types present in the digestive epithelium, and the second the ar- rangement of their interior folds. Oesophagus. The oesophagus (3.5- 4.5 mm long) begins at the posterior upper aspect of the buccal mass, and merges with the crop (Fig. 1). The oesophagus forms no definitive, easily distinguishable pharyngo- oesophageal junction with the buccal mass. Rather, the oesophagus is formed by the FIG. 2. Semi-schematic drawing of transverse sec- tion of oesophageal wall. cc, ciliated cell; ct, con- nective tissue; oel, oesophageal lumen; mf, muscle fibre; mgc, mucous gland cell; uc, unciliated cell. Scale bar 20 um. gradual tapering of the buccal cavity, and posteriorly it narrows to merge with the crop. The transition from the oesophagus to the crop is not abrupt (Fig. 1). The oesophagus is round to oval in cross section, and its wall has internal longitudinal ridges (Fig. 3). These thick ridges are straight and extend into the anterior region of the crop. The cell types seen in the oesophageal digestive epithelium consist of columnar, cili- ated and glandular cells (Fig. 2). The unciliated columnar cells are tall, nar- row (20-30 um high and 6-8 um wide) and tightly packed (Fig. 4). These cells are abun- dant throughout the epithelium. Their apical regions have a prominent striated border about 2 um thick, which consists of many, uniformly-distributed microvilli (Fig. 4) and stains positively by the PAS reaction. With the Heidenhain’s azan method this brush border is stained blue. The cytoplasm of these cells is acidophilic and slightly granular in appear- ance (Figs. 2, 4). Their nuclei tend to occupy the mid to basal third of the cells, and they have an ovoid or spherical shape of about 5 um in average diameter. These nuclei have numerous and disperse granules of chroma- tin, and one or two prominent nucleoli (Figs. 2, 4). These cells possess in their supranuclear cytoplasm abundant to lower amounts of large granules whose size varies from 0.3-1.5 um. There are also lipid droplets as the Sudan black B stain reveals. When the sections are stained with the PAS technique 122 ROLDAN & GARCIA-CORRALES FIG. 3. Transverse section of the oesophagus. Note the prominent ridges (arrowhead). oel, oesophageal lumen. Scale bar 0.1 mm. FIG. 4. Digestive epithelium of oesophagus, showing predominant unciliated cells (uc) and a clump of ciliated cells (cc). Beneath the epithelium is connective tissue (ct). The basement membrane (arrowhead) and prominent brush border (b) can be seen. Note the granular texture of cytoplasm. Scale bar 20 рт. FIG. 5. Transverse section of ridge from oesophageal wall, showing abundant ciliated cells (cc), unciliated cells (uc) and mucous gland cells (arrowhead). Beneath the epithelium is well-vascularized connective tissue layer (ct). oel, oesophageal lumen. Scale bar 20 рт. FIG. 6. Scanning electron micrograph of the luminal surface on an oesophageal ridge. Note the distribution of cilia in groups (arrowhead). Scale bar 0.1 mm. HISTOLOGY OF THE ALIMENTARY TRACT OF THEBA PISANA 123 with maltase digestion as control, the cyto- plasm of these cells show the presence of glycogen which assumes different consisten- cies and location within the cytoplasm. The columnar cells are intermingled with ciliated cells. The apices of these cells have brush border with cilia extending beyond the limit of the microvilli (Figs. 2, 5). The ciliated cells are tall and columnar with centrally- located, oval nuclei, and with cytoplasm very similar to that of the unciliated cells described above (Figs. 2, 5). Ciliated cells appear in groups where they protrude slightly into the oesophageal lumen, and the cilia present at the luminal surface of the oesophagus are found in groups instead of being evenly dis- tributed throughout the surface (Fig. 6). In the anterior oesophagus, the ciliated cells are numerous, but over the folds they are more abundant. As the anterior oesophagus passes to the posterior one, the cilia decrease innumber and they are found only on the tops of the folds. The gland cells are goblet-shaped and lie between the remainder cells of the oe- sophagus digestive epithelium (Fig. 2). They are less numerous than the other cell types, and appear uniformly distributed throughout the epithelium. The nuclei in these cells are situated in the basal cytoplasm. They have an ovoid or spherical shape of about 3 um in average diameter (Fig. 5). These densely reticulate nuclei possess no nucleoli, and stain more heavily than those of the other cell types. Their cytoplasm contains large num- bers of granules, which are stained strongly by the PAS reaction. By Heidenhain's azan staining, the granule centre is light blue while the peripheral layer stains deep blue. These staining reactions of the granules denote their basophilic and acid mucopolysaccharide na- ture. There is a thin layer of connective tissue surrounding the oesophagus. This layer con- tains many small muscle fibres (Fig. 2). Crop. The oesophagus dilates and passes posteriorly to an inflated crop which is a sac- like enlargement of the digestive tract (Fig. 1). The transition from oesophagus to crop 1$ gradual. This is a long (12 mm in length) wide tube, which is parallel to the longitudinal axis of the foot and enters the stomach. The pos- terior crop is not differentiated from the stom- ach, and it ends without any morphological discontinuity, except a slight annular constric- tion, in this organ (Fig. 1). The crop in com- parison with the oesophagus has few longitu- dinal ridges, and these become less conspicuous so that the posterior half of the crop has no internal ridges. The crop is internally lined with a non- ciliated epithelium. Two cell types constitute this epithelium: columnar and gland cells. The columnar cells are the most abundant throughout most of this epithelium. Both cell types are similar in morphology to those de- scribed in the oesophagus. But the cells of the crop digestive epithelium are taller than those of the oesophagus. The crop is buried in a layer of loosely compacted connective tissue containing nu- merous scattered muscular fibres. Stomach. The crop opens into the stom- achal crop which continues to the stomach which is the most posterior region of the digestive tract, and is defined as the part of the alimentary canal which receives the two ducts of the digestive gland. The stomach extends farthest from the mouth, and forms a bend from which the intestine goes forward (Fig! 1): The stomach is a rather globose to some- what elongate curved organ; the most poste- rior part of this is the top of the stomachal pouch, whose posterior end is tapering to form a tube which becomes the intestine. The stomach lacks a gastric shield and a style. It possesses a large, non-cuticular smooth area in its internal anterior region lying adjacent to the crop opening (Fig. 7). In the stomachal pouch, near to its intestinal end, is a large area marked by many longitudinal ridges (Fig. 7) The digestive gland surrounds the reflexed stomach and intestine. The two ducts of the digestive gland are almost circular in section. The anterior duct opens backward into the angle formed by the stomach and the proxi- mal intestine, near to the entry of the latter (Figs. 1, 7). The posterior duct opens into the columellar side of the stomach. The duct openings of the digestive gland are round, and the gastric walls around them show radi- ally arranged ridges (Fig. 7). Two unequal typhlosoles extend from the duct openings of the digestive gland to the entry of the intestine (Fig. 8). The smallest one issue from the anterior duct, and the largest one from the posterior duct. The minor typhlosole ends at the beginning of the prox- imal intestine, and the major typhlosole con- tinues along the proximal intestine (Fig. 7). The epithelium that lines the stomachal crop is similar to that ofthe oesophageal crop, 124 ROLDAN & GARCIA-CORRALES FIG. 7. Graphical reconstruction of the stomach and intestine. c, crop; dad, duct of the digestive gland; i, intestine; s, stomach; r, ridge; t, typhlosole. Scale bar 2 mm. but there are mucous gland cells present and the unciliated columnar cells are the most prevalent cell type. The cells lining the stomachal pouch are ciliated columnar (Fig. 9). The cilia on the ridges of this region are more abundant and larger than those else- where on the epithelium. Some mucous gland cells, similar to those described above, are intermingled with the ciliated cells (Fig. 9). The connective tissue that surrounds the stomach is thicker and has more muscle fibres than that of the oesophagus and crop. Intestine. The intestine opens from the columellar lower side of the stomach, curves to form a U, and passes under the oe- sophagus describing a half circle around it, to reach the left side of the body (Fig. 1). The intestine goes backward over the dor- sal surface of the stomach and coils before going to the ventral surface of the stomach where it reflexes again. The prerectal intesti- nal bend coils around the stomach and passes to its dorsal surface (Fig. 1). For convenience the whole intestine is here divided into three regions, namely, the proxi- mal, mid and distal intestine, since they differ histologically. The major typhlosole, extending from the stomach, prolongs inside the proximal intes- tine (Fig. 7) where it ends gradually. This typhlosole has a median sheet of connective tissue with muscle fibres, and is tilted towards a side, delimiting a straight intestinal groove (Fig. 10). In addition to this, the proximal intestine has, at most, a few internal ridges which are slight (Fig. 10). The digestive epithelium of the proximal intestine is strongly ciliated. The ciliated co- lumnar cells are taller (30-40 am in height) than those of the oesophagus, but morpho- logically similar to them. The cilia in this region are longer (4 шт in length) and more numerous than those of other regions of the alimentary canal (Fig. 11). Numerous gland cells of two types are intermingled with the ciliated cells. The first type is identical to the mucous gland cells of the oesophagus de- scribed above (Figs. 11, 12). The second type is similar in size and nuclear features, but possesses a cytoplasm different in some way. The greatest number of its secretory granules possess a slightly basophilic centre sur- rounded by a thin halo of more basophilic material. Intermixed with these granules are other strongly basophilic secretory granules which stain uniformly and intensely deep blue with the Heidenhain's azan method (Figs. 11, 12). The number of these homogeneous secretory granules varies from one to another gland cell of this type. The digestive epithelium of the proximal intestine is invested by a thin layer of connec- tive tissue containing a few muscle fibres. The mid-intestine is distinguishable from the proximal intestine by the absence of inter- nal ridges. Sections of the mid-intestine stained with the Heidenhain's azan method reveal its digestive epithelium comprised of ciliated and unciliated columnar cells, in- tense-staining, granular secretory cells, and lighter, highly vacuolated mucous gland cells (Fig. 13). The unciliated and ciliated epithelium cells are identical to those described above. The latter decrease in number from the anterior to the posterior region of the mid-intestine, where they are found in clumps of two to three cells (Fig. 13). The granular secretory cells constitute the most remarkable feature of the mid-intestine digestive epithelium. They have large, elon- gate (9 рт in average diameter) nuclei which HISTOLOGY OF THE ALIMENTARY TRACT OF THEBA PISANA 125 FIG. 8. Scanning electron micrograph of luminal surface on posterior region of stomach and entry of intestine. Note ridges (arrowhead), minor typhlosole (small arrow) and major typhlosole (large arrow). Scale bar 0.3 mm. FIG. 9. Section of stomach posterior region. The epithelium displays columnar ciliated cells (arrowhead) and some mucous gland cells (arrow). ct, connective tissue; sl, stomachal lumen. Scale bar 20 рт. FIG. 10. Transverse section of the anterior intestine demonstrating the titled typhlosole (arrow) and slight internal ridges (arrowhead). il., intestinal lumen. Scale bar 0.1 mm. FIG. 11. Section of anterior intestine wall showing strongly ciliated epithelium (arrowhead) and mucous gland cells of two types (arrows). ct, connective tissue. Scale bar 30 um. are situated in the mid to basal third of the cells, and stain heavily. Numerous large gran- ules which range in size and shape, fill almost totally the cytoplasm (Fig. 14). Two or more granules fuse to form larger ones. The secretory granules accumulate in the apical region of these cells, where they have a spherical shape and a maximum diameter of 1.5 um. With the Heidenhain's azan tech- nique, these granules show an orange-red homogeneous content; with the Mann- Dominici method, the same granules stain red-purple. The granules of these gland cells show a different aspect and dye affinities 126 ROLDAN & GARCIA-CORRALES FIG. 12. Semi-schematic drawing of section of proximal intestine. cc, ciliated columnar cells; ct, connective tissue; dmgc, mucous gland cell with granules of two types; il, intestinal lumen; mf, muscle fibre; smgc, mucous gland cell with similar granules. Scale bar 10 um. depending on their location in the cell. It is observed that granular gland cells empty their secretory granules into the intestinal lumen. The mucous glandular cells in the mid intestine are similar to those of the proximal intestine (Figs. 12, 13). The distal intestine is not differentiated from the mid- intestine, and have no internal longi- tudinal ridges. At the end of the distal intestine, the digestive epithelium and underlying con- nective tissue are elevated to form a single, longitudinal lateral fold, which is very patent (Fig. 15) and continues along the rectum. The digestive epithelium of the distal intes- tine is mainly composed of unciliated colum- nar and gland cells, although isolated ciliated cells are also present. In contrast, the lateral ridge is densely covered by abundant ciliated cells. All these cells are similar to those described above. The connective tissue surrounding the mid and distal intestine is similar to that of the proximal intestine, although more muscle fi- bres are present. Rectum. The distal intestine continues to the rectum. The distal intestine and rectum go into the right side of the body. The rectum is morphologically similar to the distal intestine; the transition from one region to another is gradual. The rectum is identical in diameter with the distal intestine, and nei- ther show any internal morphological differ- ences. The inner surface of the rectum is smooth except for a longitudinal lateral ridge which rises in the distal intestine, prolongs inside the FIG. 13. Semi-schematic drawing of section of mid intestine. cc, ciliated columnar cells; ct, connective tissue; dmgc, mucous gland cell with granules of two types; gsc, granular secretory cell; il, intestinal lumen; mf, muscle fibre; smgc, mucous gland cell with similar granules; uc, unciliated absorptive cell. Scale bar 20 um. rectum and ends at the rectum posterior region. The digestive epithelium of the rectum is comprised of ciliated columnar cells and mu- cous gland cells identical to those of the proximal intestine. The rectum is surrounded by a thin layer of connective tissue having few muscle fibres. At the end of the rectum there is a sphincter around the anus. Tiny, ovoid, faecal pellets found in the intestine and rectum are held in a fine mucous strand. DISCUSSION The Theba pisana digestive tube is com- posed of: a buccal mass, two salivary glands, an oesophagus, the crop, the stomach, the digestive gland, the intestine, the rectum and the anus. lt is similar to that described by Rigby (1963; 1965) in Oxychilus cellarius and Succinea putris, and Walker (1972) in Agriolimax reticulatus. A great deal of variation between animals in the size and number of folds within approxi- mately similar regions of the alimentary canal was observed, but this may be correlated with how recently the animals had been feeding. We think that at least some of the folds may be temporary structures which disappear when the digestive tube wall is stretched, such as at times when the animal is taking in large quantities of food. The digestive epithelium in T. pisana is HISTOLOGY OF THE ALIMENTARY TRACT OF THEBA PISANA 127 FIG. 14. Section of mid intestine wall showing granular secretory cells (arrows) and mucous gland cells (arrowhead). ac, columnar absorptive cell; f, food. Scale bar 20 um. FIG. 15. Scanning electron micrograph of luminal surface on distal intestine, demonstrating single, longitudinal lateral ridge (arrowhead). Scale bar 0.2 mm. made up of columnar, ciliated and glandular cells. Striated borders have been described as being related to the function of absorbing and transporting relatively large volumes of water, salts and protein over short time periods (Palay and Karlin, 1959). The dense brush border of the ciliated and unciliated columnar cells of the 7. pisana digestive epithelium, in addition to the presence of glycogen and lipid droplets in their cytoplasm suggest they have an absorptive function. Carbohydrate and lipid absorption by the cells of the digestive epithelium has been demonstrated by Car- riker (1946b) in Lymnaea stagnalis Linne, in which the oesophagus also present ciliated cells containing glycogen and lipids. The presence of ciliated cells throughout almost the entire length of the 7. pisana alimentary tract is likely a reflection of their assistance in the movement of food particles in the intestinal lumen. The mucous gland cells of the 7. pisana digestive epithelium produce large amounts of mucoid substances in the form of secretory granules. The highly positive reaction to PAS staining in light microscopic preparations is indicative of the carbohydrate nature of this material (Pearse, 1968). The fact that these granules, probably acid mucopolysaccharide in nature, form a continuous column from the base to the apex of cell suggests a continu- ous production of granules. Mucous gland cells are likely responsible for lubrication of the luminal surface on the digestive tube and may be important as stem cells for the diges- tive epithelium. The composition of the mucous cell gran- ules varies in the different regions of the T. pisana digestive tract. Thus in the oesophagus the texture of the granules is loose. In the intestine two types of mucous cells are ob- served, the first is similar to that of the oesophagus while the second type has gran- ules with both loose and dense texture. It is conceivable that the texture of the mucous granules reflects differences in their gly- coprotein composition. Alternatively it is pos- sible that the differences in texture may rep- resent granules in various stages of maturation. The present state of our prelimi- nary ultrastructural investigations together with the limited number of histochemical tests does not allow us to decide between these alternatives. Notwithstanding the different texture of mu- cous granules may represent granules in var- ious stage of maturation, the present state of our preliminary ultrastructural investigations does not allow us to decide whether the dif- ference in dye-binding capacity shown by the granules in the mucous gland cells may be referred to different stages in secretory activ- ity, or rather to the existence of two different compounds. The changes in the number of the different epithelial cell types are evident in each region of the 7. pisana digestive tube. On the basis 128 ROLDAN & GARCIA-CORRALES of gross anatomy alone, there is evidence for the functional division of the alimentary tract in T. pisana. T. pisana is a herbivorous pulmonate, whose oesophagus is devoid of a ciliated dorsal food groove, and of oesophageal glands. Our findings on the oesophagus anat- omy in T. pisana agree with those described by Ghose (1963) for Achatina fulica. The digestive epithelium of the T. pisana oesophagus possesses mucous gland cells with numerous granules, ciliated cells and columnar-absorptive cells. On the basis of our morphological study, gland cells with enzy- matic secretions appear to be not present in the oesophagus of T. pisana. The primary function of the oesophagus ap- pears to be for the passage of food to the crop and this movement is no doubt aided by the release of mucous substances and the con- tinued movement of food particles by cilia along the oesophageal epithelium. The abun- dance of ciliated cells in the 7. pisana oesoph- ageal epithelium may constitute a feature cor- related with weak peristaltic action of the poorly-developed muscular coats of the oesophagus. The presence in such digestive epithelium of numerous cells with a well-deve- loped brush border suggest that absorption may be other main function of the oesophagus. T. pisana crop is devoid of chitinous plates or teeth; this fact suggests that it does not function as agizzard. The digestive epithelium of T. pisana crop and stomachal crop is com- posed of gland cells and interspersed, unciliated supporting cells, indicating that muscle fibres of the subjacent connective tis- sue rather than cilia are probably responsible for directing food particles towards the strong- ly-ciliated sorting area in the stomach pouch. The T. pisana crop, while serving as a storage and digestive organ, is also most likely to be concerned with absorption, as suggested by the presence of columnar- absorptive cells, the most abundant cell type throughout its epithelium. T. pisana has neither gizzard nor gastric shield, and its stomach lacks a style. The ef- ficient grinding structures of other snails could be substituted in T. pisana for the action of digestive gland enzymes, as it occurs in Helix pomatia in whose stomachal lumen Ferreri (1961) demonstrated proteolytic activity. The absence of ciliated cells in the diges- tive epithelium of the stomachal crop in Т. pisana is balanced by the well-developed muscular layer of its subjacent connective tissue. The T. pisana stomachal pouch has a pos- terior grooved sorting area around the intes- tine entry and the digestive gland duct open- ings. The folds of this area could act in the selection of the food particle size. The two gastric typhlosoles in T. pisana are similar to those described by Ghose (1963) in A. fulica, Walker (1972) in A. reticulatus and Tillier (1984) in some helicarionid species. The func- tion of cilia in the typhlosoles and sorting area of the T. pisana stomach appears to be to provide assistance in conveying food particles toward the intestine. There are few data about the gastric func- tions in pulmonate molluscs. The exact roles played by the stomach in the digestive events are still not understood. The T. pisana intestine is divided into the proximal ciliated, the mid secretory and the distal absorptive region. While the proximal and mid-intestine participate in digestion, the distal intestine is generally most important in absorption and faeces formation. This study shows that the T. pisana proxi- mal intestine possesses a digestive epithe- lium strongly ciliated and a musculature less- developed than that of the distal intestine which is devoid of ciliated cells. The proximal intestine typhlosole increases luminal surface and helps to move food particles. The presence of granular secretory cells in the digestive epithelium of the T. pisana mid- intestine may be indicative of a higher secretory and therefore digestive activity in this region. Brown (1969) located enzymatic activity in the intestinal lumen of Nassarius obsoletus (Say). In the oesophageal and rectal epithelium of Murex brandaris (Linne) are present granular gland cells whose secretions are of enzymatic nature (Bolognani-Fantin et al., 1982). These cells are morphologically similar to that described by us in T. pisana mid-intestine. This fact lends support to the proposal that this site is the another principal region for enzyme secretion, and both diges- tion and absorption of simple nutrients are car- ried out. The digestive epithelium of the T. pisana distal intestine is entirely comprised of unciliated, columnar-absorptive cells and mu- cous gland cells. Absorption of different mol- ecules by the intestine epithelial cells has been showed by Guardobassi and Ferreri (1953) and Sumner (1965) in H. pomatia, Rigby (1963) in O. cellarius and Brown (1969) HISTOLOGY OF THE ALIMENTARY TRACT OF THEBA PISANA 129 in N. obsoletus. On the basis of our findings and these data, it can be assumed that the primary function of the T. pisana distal intes- tine is one of absorption. Formed faeces are of considerable impor- tance for pulmonate gastropods because the anus is near the pneumostome, and firm faeces are less likely to foul this. Faeces formation generally involves the secretion of abundant mucus, the squeezing of the mucus and rejected material to form firm bodies and possibly some absorption of water. A pellet compressor for the faeces formation has been described by Carriker (1964b) in L. stagnalis appressa, Demian & Michelson (1971) in Marisa cornuarietis (Linné), and Richards (1973) in Biomphalaria glabrata (Say). T. pisana produces solid faecal matter, but it lacks a definitive pellet compresor. Thus lubrication in the distal intestine and rectum would seem to be important, and an increased lubrication requirement would seem practical. The digestive epithelium of the T. pisana distal intestine and rectum contains many mucous gland cells which secrete mucus, helping to compact the faecal material. This abundance of PAS positive secretion in the intestine and rectum suggests that it provides the mucous covering of the faecal pellets. The very ciliated, longitudinal fold of the T. pisana distal intestine and rectum is likely ho- mologous to that of Agriolimax reticulatus (Walker, 1972) and L. stagnalis (Carriker, 1946b). Although there is undoubtedly some movement created by the peristalsis produced by the thin muscle layer, cilia would be par- ticularly beneficial for the narrowed rectum ex- tending from the distal intestine to the anus in T. pisana. Not much is known of the details of rectal functions in gastropods. On the basis of the cell types predominant in the digestive epithe- lium of this region in T. pisana some food and water absorption appears to occur, but the main task is the condensation of the faeces which are well formed when released. The intestine and rectum in pulmonate snails have likely other functions, but more work is needed to define them. ACKNOWLEDGEMENTS We are grateful to Dr. F. Pardos for techni- cal assistance during this work. This work was supported by the “Comision Asesora para la Investigacion Cientifica y Técnica” (CAICYT, proyect n° 1156). REFERENCES CITED BOLOGNANI-FANTIN, A. M., BOLOGNANI, L., OTTAVIANI, Е. & FRANCHINI, A., 1982, The digestive apparatus of Murex brandaris (L.) and Trunculariopsis trunculus (L.). Zeitschrift für Zellforschung und mikroskopische Anatomie, 96, 4: 561-582. BOWEN, I. D., 1970, The fine structure localization of acid phosphate in the epithelial cells of the slug Arion ater L. Protoplasma, 70: 247-260. BROWN, S. C., 1969, The structure and function of the digestive system of the mud snail Nassarius obsoletus (Say). Malacologia, 9: 447-500. CARRIKER, M. R., 1946a, Morphology of the ali- mentary system of the snail Lymnaea stagnalis appressa Say. Transactions of the Wisconsin Academy of Science, Arts and Letters, 38: 1-88. CARRIKER, M. R., 1946b, Observations of the alimentary system of the snail Lymnaea stagnalis appressa Say, Biological Bulletin of the Marine Biological Laboratory, Woods Hole, 91: 88-111. DEMIAN, Е. $. & MICHELSON, E., 1971, Histochemistry of the epithelial mucins in the alimentary tract of the snail Marisa cornuarietis. Journal of Morphology, 135: 213-238. FERRERI, E., 1958a, Uattivitá lipasica dell’epitelio intestinale de Helix pomatia. Nota Il. Bolletin della Societa italiana di Biologia Sperimentale, 34: 379-382. FERRERI, E., 1958b, Richerche biochimiche ed histochimiche sull’attivita lipasica dell'epitelio intestinale di Helix pomatia. Zeitschrift für vergleichende Physiologie, 41: 373-389. FERRERI, E., 1961, L’attivita proteolitica del tubo digerente di Helix pomatia L. e di Murex trunculus L. (Mollusca, Gasteropoda). Richerche biochimiche. Bolletin della Societá italiana di Biologia Sperimentale, 1: 141-149. GABE, M., 1968, Techniques histologiques. Mas- son, Paris. GHOSE, K., 1963, The alimentary system of Achatina fulica. Transactions of the American Microscopical Society, 82: 149-167. GUARDOBASSI, A. & FERRERI, E., 1953, Istefisi- ologia dell aparato digerente di Helix pomatia. Archivo Zoologico Italiano, 38: 61-156. MARTOJA, M. 8 THIRIOT-QUIEVREUX, C., 1975, Donnes histologiques sur l'appareil digestif et la digestion des Atlantidae (Prosobranchia: Heteropoda). Malacologia, 15: 1-27. PALAY, S. L. & KARLIN, L. J., 1959, An electron microscopic study of the intestinal villus. И. The pathway of fat absorption. Journal of Biophysical and Biochemical Cytology, 5: 373-384. PEARSE, А. С. E., 1968, Histochemistry theoreti- cal and applied. Vol. I. Ed 3. London: Churchill. 130 ROLDAN & GARCIA-CORRALES RICHARDS, C. S., 1973, Fecal pellets of Biomphalaria glabrata. Malacological Review, 6:125-132. RIGBY, J. E., 1963, Alimentary and reproductive systems of Oxychilus cellarius (Müller) (Stylomm.). Proceedings of the Zoological Soci- ety of London, 141: 311-359. RIGBY, J. E., 1965, Succinea putris, a terrestrial opisthobranch mollusc. Proceedings of the Zoo- logical Society of London, 144: 445-486. ROLDAN, С. & DIAZ COSIN, D. J., 1975, Contribuciön a la histologia y anatomia de la masa bucal de Theba pisana. Boletin de la Real Sociedad Espanola de Historia Natural (Serie Biología), 73: 169-192. SHERIDAN, R., VAN MOL, J. & BOUILLON, J., 1978, Etude morphologique du tube digestif de quelques Turridae (Mollusca, Gastropoda, Prosobranchia, Toxoglossa) de la region de Roscoff. Cahiers de Biologie Marine, 14: 159-188. SUMNER, A. T., 1965, Experiments on pha- gocytosis and lipid absorption in the alimentary system of Helix. Journal of the Royal Microscop- ical Society, 84: 415-421. TILLIER, S., 1984, Patterns of digestive tract mor- phology in the limacisation of helicarionid. suc- cineid and athoracophorid snails and slugs (Mollusca: Pulmonata). Malacologia, 25: 343-348. WALKER, G., 1972, The digestive system of the slug Agriolimax reticulatus (Muller): experiments on phagocytosis and nutrient absorption. Pro- ceedings of the Malacological Society of London, 40: 33-43. WU, S. K., 1965, Comparative functional studies of the digestive system of the muricid gastropods Drupa ricina and Morula granulata. Malacologia, 3: 211-233. Revised Ms. accepted 24 February, 1987 MALACOLOGIA, 1988, 28(1-2): 131-146 THE COMPARATIVE ECOLOGY OF FOUR SYMPATRIC LIMACID SLUG SPECIES IN NORTHERN IRELAND Anthony Cook & D.J. Radford' Department of Biology, University of Ulster, Coleraine, Northern Ireland BT52 1SA, United Kingdom ABSTRACT The distribution, feeding ecology, population structures and temperature sensitivity of the embryos of four sympatric species of Limax were examined. L. pseudoflavus and L. flavus were found mainly on walls, L. maximus on the ground and L. marginatus on trees. The feeding ecology was assessed both by direct observation and by faecal analysis. L. maximus showed a high proportion of vascular plant material in its diet whereas the other species fed predominantly on lichens. Examination of population structures indicate that L. marginatus is a univoltine, iteroparous species whilst the other three are polyvoltine and semelparous. At sites where either L. pseudoflavus or L. marginatus were found alone there was а significantly higher proportion of small individuals in the population than at a site where all four species occurred together. During incubation, L. marginatus embryos were the most tolerant of low temperatures whilst L. maximus and L. flavus were the most tolerant of high temperatures. These results are discussed in the light of the behaviour and European distribution of the four species and it is concluded that there is substantial niche separation between L. maximus and the other species which is mainly based on its feeding preferences. The remaining species have a very similar feeding ecology but the substantial differences in life cycles and temperature sensitivity distinguishes L. marginatus from both L. flavus and L. pseudoflavus. Key words: Gastropoda; ecology; feeding; niche separation; Limax. INTRODUCTION Species with close taxonomic affinities are usually also similar in their physiology and ecology. Where closely related species are sympatric it is to be expected either that adverse conditions act to keep coexisting populations below the carrying capacity in the areas of niche overlap or that there is a substantial element of niche separation (den Boer, 1986). In cases where the diets and distribution of closely related sympatric spe- cies have been examined niche separation has often been demonstrated (Pontin, 1982) and in closely related marine gastropods such factors as habitat, food size and activity pat- terns are involved (Spight, 1981). Where niche separation has been examined in ter- restrial pulmonates however, such factors have been more difficult to identify (Cameron, 1978). The genus Limax is represented in Ireland by five indigenous species. L. cinereoniger Wolf, 1803 is rare, but the others (L. flavus L., 1758, L. maximus L., 1758, L. marginatus Múller, 1774 and L. pseudoflavus Evans, 1978) are widespread and often numerous. These slugs are relatively large, have daytime resting sites to which they home (Gelperin, 1974; Cook, 1979, 1980) and in which they lay their eggs. They are distinguishable on external characteristics alone (Kerney 4 Cameron, 1979). The general habitat types occupied by these Limax species have been described (e.g. Quick, 1960; Kerney 8 Cameron, 1979; Anderson, 1977; Evans, 1978). In summary, the habitats of L. pseudoflavus, L. mar- ginatus, and L. maximus extend from wood- land to gardens and walls. L. flavus on the other hand is more synanthropic and is rarely found away from buildings. The most detailed information on pulmonate life cycles is available for slugs (Duncan, 1975), but this is largely for small pest species such as Deroceras reticulatum Múller (Hunter, 1968; South, 1982). The little information which exists on the life cycles of Limax spp. is largely anecdotal and this suggests that they mate and lay their eggs late in the year “Present address: c/o Reserves Division, Royal Society for the Protection of Birds, The Lodge, Sandy, Beds., SG19, 2DL, United Kingdom, (131) 132 COOK & RADFORD UNPOINTED STONEWORK SHADED BARE GROUND TRODDEN PATH POINTED STONEWORK BRICKWORK UNSHADED DOMESTIC REFUSE FIG. 1. Details of the main part of the Cranagh site. Small lengths of wall at both ends of this wall were included in the site. The main wall faces south-east. (Quick, 1960). A more detailed study of L. maximus has shown that gonads undergo first male and then female maturation in response to increasing day lengths (Sokolove & Mc- Crone, 1978) and therefore would be expected to lay cross-fertilised eggs in the autumn. In Northern Ireland, slug communities often include three Limax species, and occasionally four. The objective of the present work was to compare aspects of the biology of these four species in an attempt to identify differences of ecological significance. METHODS General survey of habitats Sixty-seven areas within the triangle whose corners are marked by Coleraine, Portrush and Portstewart (approximately 16 km*) were inspected for Limax species in April and May, 1978. Surveys were undertaken at night when the animals were active and each area visited more than once. A quantitative study was not undertaken since activity is determined by the prevailing weather conditions (Ford, 1986) and only a small proportion of the areas could be adequately sampled on one night. Frequency and size distribution of field populations The mobility of slugs between and within sites was examined by collecting, freeze branding (Richter, 1976) and returning all the L. pseudoflavus (39) found on a small isolated group of stone walls. This site and neighbour- ing groups of stone walls were examined nightly for the following 22 nights until the brands became indistinct. The populations of slugs at two field sites were monitored over the course of three years (1978-1980). Both sites were isolated from other loca- tions inhabited by Limax spp. Each site had well defined boundaries to facilitate the re- peated sampling of the same area. The Cranagh site (Irish grid ref. C841346) (Fig. 1) was part of a complex of stone farm buildings on the Coleraine campus of the University of Ulster. The study area com- prised two exterior walls of an outhouse plus about 2 m of an adjoining garden wall. A strip of ground about 1 m wide in front of these walls and a small rubbish tip was included in the area. The tip was cleared in May 1979 and at the same time a wall of another stone building 10 m away was exposed. This wall supported a large population of L. marginatus and was also sampled after May 1979 on the same basis as the original site. It will be referred to as the uncovered wall. The Kiltinny site was 1km from the Cranagh site (Irish grid ref. C844354). It con- sisted of the shell of a small, derelict, stone building 3 x 3 m with walls 2 m in height, an adjacent wall 3 т long and 1.5 m high and a derelict field wall which was little more than a line of loose stones. The building had по roof and the site included both the outside and inside faces of the walls. The area sampled was roughly twice that of the original Cranagh site. The walls of all three sites had a substantial covering of saxicolous lichen. Fresh higher plant material was always available at the base of the walls. Rotting vegetable matter was plentiful only at the Cranagh site. Slugs were collected at night after emer- LIMACID ECOLOGY 133 gence from their day-time resting sites. For each sample the site was visited from dusk till dawn on three successive nights and all the slugs emerging on each night were removed. Each slug was stored individually in a plastic bag after its position and, if it was feeding, its feeding substrate had been noted. All slugs collected over the three day period were released on the fourth night at their points of collection. Estimates of the total population size were made using the rate at which the catches on successive nights declined (Zip- pin, 1956, 1958; Southwood, 1978). This method assumes that a constant proportion of the population is available for capture each night and that a large proportion of the total population is captured during the observation period. For the first collection (May 1978) each slug was weighed in the field immediately after collection and then reweighed in the labora- tory the following day. The two weights were not found to be significantly different (paired t-testt = 0.62, d.f. = 120, p > 0.05) and for subsequent collections all slugs were weighed in the laboratory except for a very few slugs found on the fourth night which were weighed in the field. Nineteen samples were taken between May 1978 and September 1980. The obser- vations on the life cycles of these species are derived from all these samples but those on the distribution within the site are only based on the 7 samples between May and Novem- ber 1978. Such a restriction was necessary because the area around the site was unex- pectedly cleared in May 1979. Faecal pellet analysis Faeces produced by slugs collected during the field sampling were stored in 70% alcohol. Faeces collected in November 1978, March 1979, May 1979 and June 1979 were analysed. Faecal pellets were suspended in water and sonicated until they broke up. Five aliquots of the resulting suspension were ex- amined on a haemocytometer slide, the per- centage cover of each food type estimated and the results averaged. Some pellets failed to break up on sonication and these were teased apart with fine needles before exami- nation. Trial experiments with L. pseudoflavus fed on specific food items allowed the classi- fication of faecal material into lichen, vascular plant material, Pleurococcus type algal cells, filamentous algal cells, fungal hyphae, and minerals. A very few pellets contained the remains of insects and earthworms but these were not frequent enough to be included in the classification. Egg production in laboratory cultures A culture of each species was set up in fibre glass bins 1 x 1 x 0.5m covered in clear polyethylene and containing a layer of soil and two plastic trays as homes. The bins were protected from extremes of temperature but open to ambient day lengths. Each bin was kept supplied with ‘Readybrek’ breakfast cereal and this diet was supplemented with leaf litter and fungi. The cultures were established in May 1979 with 20 mature animals of each species col- lected in the field away from the main study area. In August 1979 the number of slugs in each culture was reduced to 10 to avoid symptoms of overcrowding. Whenever possi- ble dead slugs were replaced immediately with individuals of similar size but this was difficult for L. maximus and L. marginatus during the winter months. The L. maximus culture was completely restarted twice (Octo- ber 1979, July 1980) and the L. marginatus culture completely restarted once (July 1980). Each bin was inspected almost daily and any eggs removed. In all species the eggs normally were laid in clutches in the homes. For the purposes of analysis a ‘clutch’ was considered to be a group of at least 5 eggs. Groups of less than five eggs were treated as belonging to clutches laid at about the same time. These small numbers of eggs were never added to clutches laid more than two days before or after their discovery. Occa- sionally two slugs laid at the same time in the same place, but unless two individuals were actually seen laying simultaneously all large groups of eggs were treated as one clutch. Influence of temperature on egg development The eggs were incubated in cooled incuba- tors at four temperatures: 5, 10, 15 and 20° C. Clutches of more than 16 eggs were split into batches and each incubated at a different temperature. In practice this resulted in 28 batches from L. pseudoflavus containing 22.8 + 1.9 (s.e.) eggs, 49 batches from L. flavus containing 17.7 + 1.0 eggs, 23 batches from L. maximus containing 39.1 + 4.6 eggs and 18 batches from L. marginatus containing 134 COOK & RADFORD TABLE 1. The number of areas at which the various species were found. Number of areas containing Site type (N) No Limax L. pseudoflavus L.flavus L. maximus L. marginatus Open rock face (7) 7. 0 0 0 0 Field wall (7) 4 8 0 0 0 Field wall and trees (17) 1 9 0 4 14 Isolated building (9) 6 S 0 1 1 Building and trees (10) 1 8 1 2 6 Town wall (no garden) (5) 0 5 2 2 1 Town walls with garden (2) 0 2 1 2 0 Wall in woodland (6) 0 6 0 1 5 Woodland (4) 0 4 1 3 4 All sites (67) 19 40 5 es 31 TABLE 2. The percentage of each species found in the areas of the Cranagh site between June and November 1978. Analysis of the frequency of occurrence of each species in each habitat in a 4 x 4 contingency table showed there to be significant differences in distribution (x? = 19.2, d.f. = 9, p < 0.001). Significance levels in the table refer to a posteriori binomial tests performed comparing the observed frequencies with those expected from a consideration of the whole table in order to partition the contingency between individual cells (Siegel, 1956; Stephenson & Poole, 1976). Where significant results were obtained it is also indicated whether the observed frequency of a species constituted a high or a low proportion of the slugs in that subdivision of the habitat. * — р < 0.05, ** — p < 0.01, *** — p < 0.001. Species Habitat feature L. pseudoflavus L. flavus L. maximus L. marginatus Ground 23.5 16 35.4** TEST high low Trees 20.6** 32 OR 30.9*** high low low high Wall 42.9 74.2 З25 + 46.4 high low Roof and wood 12.9* 657 Sil elise 15.5 low low high 19.3 + 3.0 eggs. Eggs were kept moist on were selected so that they all contained stone filter paper on a bed of plaster of Paris which faces and/or trees which were the habitat was continuously in contact with water. The features known to be associated with Limax temperatures of the ovens normally deviated species. Table 1 shows a classification of the from that set by less than one degree. areas examined and the numbers of those An egg was deemed viable if an active areas which were occupied by a particular embryo was visible approximately half way species. L. marginatus was significantly asso- through the incubation period. An embryo ciated with sites containing trees (chi squared was deemed viable if it hatched successfully. = 16.3; dif. = 1; р < 0:01) butcno-other significant associations were apparent. RESULTS Sites of activity General survey of habitats In a pilot study the distance between the sites of release and collection of freeze The areas examined for the presence of branded slugs was measured. 39 L. pseu- Limax species were not chosen at random but doflavus were marked with four distinctive LIMACID ECOLOGY 135 individuals of Number 50 JUN SEP DEC MAR JUN SEP DEC MAR JUN SEP JUN SEP DEC MAR JUN SEP DEC MAR JUN SEP 150 Uncovered wall 100 39 Original site JUN SEP DEC MAR JUN SEP DEC MAR JUN SEP FIG. 2. Estimates (mean + 95% confidence limits) of the total population sizes at the Cranagh site of LP) L. pseudoflavus, LF) L. flavus, LX) L. maximus, and LG) L. marginatus. The data for L. marginatus show the population on the original Cranagh site and the wall uncovered in June 1979 and those for L. pseudoflavus include both the Cranagh and Kiltinny sites. Sampling commenced in May 1978 and ended in September 1980. brands and released at four different release points close to their points of collection. On the 22 nights following release only two unmarked individuals were found and no marked slugs were found in the neighbouring Kiltinny site 30 m away. This indicates that immigration and emigration is minimal in the isolated sites being considered. On day 18 the highest number of recoveries was re- corded (14) and these were at a mean (+ s.e.) distance of 3.2 + 0.5 m from the release points. Only 2 slugs were found emerging from their original release points. L. pseudoflavus therefore does not remain sta- tionary within this habitat. Fig. 1 shows the details of part of the Cranagh site. The physical features of the area may be divided into 1, the ground in front of the wall, 2, the trees adjacent to the wall (sycamores, Acer pseudoplatanus), 3, the wall itself, and 4, the roof which was incom- plete and consisted of rotting timbers partially covered with roofing felt. Table 2 shows the frequency with which the four species were found in these sub-divisions of the habitat. It is clear that the species show significant differences in their occupancy of the area. L. pseudoflavus and L. marginatus constituted a larger proportion of the slugs on the trees than expected, a high proportion of the slugs on the wall were L. flavus, and L. maximus was found more frequently on the ground and on the roof than the other species. Within this single site therefore the species are not equally distributed over areas with different characteristics. The frequency and size distribution of field populations The estimated total populations of each species at both the Kiltinny and the Cranagh site are shown in Fig. 2. For most collections over 80% of the estimated total populations was removed from the site. L. pseudoflavus was clearly the most numerous and L. flavus the most scarce. L. maximus populations decreased from 44 in May to 11 in June 1979 and remained low. This coincides with the clearing of the ground in front of the wall. The distributions of body weight are shown in Fig. 3 for the Kiltinny site and Figs. 4 to 7 for the Cranagh site. All four species show sea- sonal changes in population structure, mak- 136 COOK & RADFORD 1978 | ı0065 “19 eno F <-15 — 1118 MAR Г АР CTN р р © | = O A ocr A NA eae DE Co 9 Body weight (g) —-NwWRUDHYD “it ze | | | MAR Sa Ea ee Ee ee ee ee ee SEP OCT IE NOV See) oe 10 Slugs MAR [APRO [er MAY JUNE JUL AUC SER ON NOV СЕНТ] FIG. 3. Size distribution of L. pseudoflavus at the Kiltinny site expressed in 1 g size classes. Those slugs of less than 0.15 д in the O to 1 g class are also shown separately as this size group reflects the presence of hatchlings in the population. ing it possible to follow the progress of a generation for most of the first year after hatching. Small L. pseudoflavus appeared in autumn and winter and grew through the spring and summer so that their size distribution merged with that of the smaller individuals of the pre- vious generation (Figs. 3-4). Slugs of less than 1 g formed a greater proportion of the total population at the Kiltinny site (92% in Decem- ber 1978, and 87% in November 1979) than at the Cranagh (55% in December 1978, and 73% in December 1979). This is a significant difference (e.g. for December 1978, chi squared = 25.33, d.f. = 1, p < 0.001). The L. flavus population showed seasonal fluctuations similar to those of L. pseu- doflavus (Fig. 5). The largest L. flavus only attained approximately three-quarters of the size of the largest L. pseudoflavus. This is in contrast to their growth in laboratory cultures where L. flavus was consistently larger (per- sonal observations DJR). Although there is a pronounced seasonal pattern in the population structure of L. maximus for the early samples the habitat changes at the original Cranagh site in May 1979 makes a full interpretation difficult (Fig. 6). The clearing of the rubbish tip and the general tidying up preceded a reduction in the numbers of L. maximus which was not seen in any of the other species, (Fig. 2) and it seems likely that this species had day time resting sites in the rubbish. There was a difference between the popu- lation structures for L. marginatus at the orig- inal Cranagh site and the uncovered wall (Fig. 7). At the original site the individuals attained a larger size but there were fewer young slugs. In December 1979, 94% of all individ- uals on the uncovered wall weighed less than 1 g, compared with only 15% in this size class at the original Cranagh site. This difference in size distribution is significant (chi squared = 28.07, d.f. = 1, р < 0.001). Data from both sites (Fig. 7) suggest that L. marginatus is an annual species, with the eggs hatching in the autumn and winter and very few adults sur- viving into the following spring. This is partic- ularly clear in 1978 when at the original LIMACID ECOLOGY 137 HN OEM vo 1978 | le <- 15 ill MAR ENS ОО ИСО Е A Y О | RS EPP SN ei DE CA] 1980 Lil | <-1 а AS A) могло © OO — 0 — 9 | al 1979 5 =i 6 | = 5 с 4 Fr 3 Ф 2 | dx > | ‹ 15 = == om = Gere 3 MAR [EA PRE AM AVE UN A A ULT AU GA | SEP | TO CTA ELN OA | Res DE G | a | 0 Stugs MAR FEN EZ BETTEN BETT BES Be Bro NO VE | AE DEC ae] FIG. 4. Size distribution of L. pseudoflavus at the Cranagh site. The presence of hatchlings is indicated as in Fig. 3. Cranagh site the 3, 4 & 5 g slugs decrease in number from a peak of 21 in August to 2 in November to be replaced by new recruitment in the following spring. Analysis of faecal pellets There were some significant differences between the months for some species and some food types. They occurred, however for those food items which constituted only small proportions of the faeces and showed no consistent pattern. For ease of analysis there- fore, and because too few data were available for some species in some months, further analysis was conducted without regard to the month in which the sample was collected. Table 3 shows the percentage composition of the faeces for each species. Analyses of variance on arc sine square root transformed data showed that there were significant differ- ences between species for all food types except filamentous algae (4). Student-New- man-Keuls a posteriori tests (Sokal & Rohlf, 1969) showed that for food types 1, 2 and 3 there were significant differences between L. flavus and L. maximus and all other species but that there were no significant differences between L. pseudoflavus and L. marginatus. Feeding observations For those individuals found feeding, the substrate over which they were grazing was classified into corticolous lichen (i.e. growing on bark), saxicolous lichen (i.e. growing on the wall), vascular plant, animal, and fungal material. Algae were mingled with the lichen and it is probable that they were consumed together. Table 4 shows the frequency with which the species were found feeding on these classes of substrate. The frequency with which slug species were found on differ- ent feeding substrates are as would be ex- pected from a consideration of the sites of activity (Table 2) and the results of faecal analysis (Table 3). Thus L. pseudoflavus and L. flavus favour saxicolous lichens, L. maximus was found feeding on vascular plant material and L. marginatus on corticolous lichens. The occasions on which slugs were found feeding on fungi or animal remains were too infrequent to allow further analysis. 138 COOK & RADFORD 1978 | a ND 0 E no œ Solos 1979 — 7 zfs | | | QT 3 | 6 v 2 $ 1 |] | | Es > | 15 1 = — = У О ООО О ZEN GR < OT юм Г bec] a 1980 -.nurumn„o L | r «15 | | ШИ 0 sus MAR [APR | М | JUN | м AUG | SEP | | nov [_ DEC | FIG. 5. Size distribution of L. flavus at the Cranagh site. The presence of hatchlings is indicated as in Fig. 3. Niche overlap Estimation of niche overlap between these species is potentially useful in indicating likely areas of competition although the precise identification of competition depends on the demonstration of resource limitation (Giller, 1984). A measure of Slobodchikoff & Schulz (1980) is concerned with the proportional utilisation of a resource and is therefore, appropriate for this type of data: overlap, = 1 — 0.5 x У (ру — PK |) where р, represents the proportional utilisa- tion of the jth partition of the resource by species i, and the differences between spe- cies are summed over all partitions. This measure of niche overlap based on the comparison of the faecal content (data from Table 3) is given in Table 5. It is clear that the least overlap occurs between L. maximus and the other species. Fig. 8 shows niche overlap measured in similar ways for the occupancy of the site (data from Table 2), faecal analysis (data from Table 3) and feed- ing sites (data from Table 4). These axes are not truly independent since the distribution in the site obviously dictates where the animal may be found feeding and the sites of feeding inevitably bear a relationship to the subse- quent faecal analysis. Nevertheless such a presentation serves to illustrate the extent of the differences between species pairs. L. maximus is separated from the other species mainly by its feeding habits but also by differ- ences in their occupancy of the site. L. flavus and L. marginatus are separated largely by their feeding sites. L. pseudoflavus is very similar to both L. marginatus and L. flavus in this particular habitat. Egg production in laboratory cultures The number of clutches of eggs and the number of eggs. slug ' month ‘are shown in Fig. 9. All species showed a seasonal pattern of egg production. There are differences be- tween the two years for L. pseudoflavus and L. marginatus. The same L. pseudoflavus had been in culture for 18 months by the time their second egg laying period started and both its delay and brevity could be an age effect. The L. marginatus died in the spring of the first year and were replaced in the July preceding the LIMACID ECOLOGY вы 1978 | A -~NWEUDRILBABOS 15 en 139 | НН | MARS APRO | en M À Ve UN И ЕО К STA Y GAR] SE Ps REA OCT | PINO Y E | ES DEC PEA] ) ЗГо | - 1979 | = 7 1 £ 6 Dm) 5 | = $ = | > 1 1 El = | «15 a m MAR | — Oe A OA GS RE us œvO + — [lo] © о «15 | EEE 0 sus MAR Ge LS ta UN E Y (O ERA Y GRA PEA SE PEE | O CT | NO VE | DE Gem] FIG. 6. Size distribution of L. maximus at the Cranagh site. The presence of hatchlings is indicated as in Fig. 3. second egg laying period, the onset of which was delayed. Influence of temperature on egg development Comparisons between the viabilities of both eggs and embryos at different temperatures are given in Table 6. Two-way analysis of variance of the transformed egg viability data showed significant main and interaction ef- fects of species and temperature. Student- Neuman-Keuls a posteriori testing (Sokal 8 Rohlf, 1969) of the differences between the means indicates that these effects are largely attributable to the poor survival of L. flavus and L. marginatus at 5° C and the increased survival of L. pseudoflavus at 15° C. Similarly comparisons between the species at the dif- ferent temperatures indicate that significant effects on embryo viability are attributable to the poor survival of L. marginatus at both 15 and 20°C and the higher survival of L. pseudoflavus at 5 and 10°C and of L. marginatus at 5°C. The time taken for the first egg to hatch in each batch is shown in Table 7. Again 2 way-analysis of уапапсе showed significant main and interaction effects and these are attributable to significantly shorter develop- ment times in L. maximus at both 10 and 15°C and significantly longer development times for L. flavus at 5” C, L. marginatus at 15°C and L. pseudoflavus at 20°C. L. marginatus failed to hatch at 20°C. The spread of hatching times from single batches of eggs (Table 7) is variable. L. maximus shows the greatest range, but since the spread is dependent on the number of survi- vors further analysis is inappropriate. DISCUSSION L. pseudoflavus was the most widespread of the species in the general survey and was the only occupant of 11 of the 67 areas. Experiments in laboratory culture with this species show that it can be kept at high densities without apparent ill effects and that its behaviour in the home is not disrupted by the presence of the other species (Cook, 140 COOK & RADFORD MN ttt р “nN D EU am | 5 MAR APR МАУ JUN A > a ae E a re р „вы me 2 1979 = 5 I 4 of; | Ф 2 3 1 Е || «-15 . ' >| D AR [ETAPA (ОНИ JUN Sie ЕЕ О AU | SERA | © CT ПОТ CO Ve BEE о a 4 2 UNCOVERED WALL 1980 1 pe 15 ' 1 = ' 5 5 4 3 da à | 0 505: 2 F <-15 О АРА МАУ] JUN JO JU a AUGE SEP OCT S| NOV es EC FIG. 7. Size distribution of L. marginatus at the Cranagh site and on the 'uncovered wall' (inset). The presence of hatchlings is indicated as in Fig. 3. TABLE 3. The percentage of each food type found in faecal pellets (mean + s.e.). n refers to the number of individuals collected from four monthly samples from which pellets were analysed. Analyses of variance followed by Student-Neuman-Keuls tests show that L. pseudoflavus and L. marginatus did not differ in the frequency in which they took any of the food types. For lichen, vascular plant and Pleurococcus all other comparisons showed significant differences. There are no differences between the species for the remaining food types. Food type Vascular Pleurococcus Filamentous Species Lichen plant algae algae Fungi Mineral L. pseudoflavus 62.1 += 2:4 141023285 ПУ. 2-Е. 6 0.6 = 0.3 10 = 92 52-Е 10 п = 90 L. flavus 52.1 + 6.8 30:7 = 8:0 ERES 041 = 051 05725208 5:6 = 10 п = 24 L. maximus 7.8 = 3.8 86.3 = 4.5 0.7 = 0.4 0.1 = 01 А. ЗЕЕ 210 0.9 = 0.3 п = 38 L. marginatus 60.5 + 4.7 SES 14.9 + 3.0 3025 3.3 = 0.8 3.3 0:8 п = 22 19815). Within the Cranagh site it was the most common species (Fig. 2), it was signifi- cantly associated with trees, but occurred less frequently than expected on the roof compared with other species (Table 2). L. flavus is very similar to L. pseudoflavus in its behaviour (Cook, 1977, 1981a, b). It is the least widespread species in the field, judged by the number of areas at which it is found (Table 1). Four of the five areas at which it was found in the present survey were associated with buildings. The fifth area was a LIMACID ECOLOGY 141 TABLE 4. The percentage of each species found feeding on different substrates. n refers to the total number of that species which was found feeding. Ignoring animal and fungal material which were present in too few observations to allow analysis, there was a significant deviation from a random expectation when the original frequencies were considered (x? = 62.1, 4.1. = 6, р < 0.001) Significance levels in the table refer to the results of binomial tests (* — р < 0.05, ** — p < 0.01, *** р < 0.001.) For each significant result an indication is given of whether the frequency is higher or lower than expected from a consideration of the whole table. Saxicolous Corticolous Higher plant Species lichen lichen material Animal Fungi L. pseudoflavus 67.5* 24.8 51877 1.8 0 n = 326 high — low L. flavus 74.6 15:57 8.5 1.4 0 n = 71 — low — L. maximus 18.4*** 82% БИ 1222 4.1 n = 49 low low high L. marginatus 46.1* 48.7*** 5.2, 0 0 n = 115 low high low TABLE 5. Indices of niche overlap based on food consumption inferred from faecal analysis. (Data from Table 3.) A value of 1 would indicate total overlap. L. flavus L. pseudoflavus 0.835 L. flavus E L. maximus L. maximus L. marginatus 0.266 0.949 0.428 0.844 — 0.290 garden and had also been considerably influ- enced by man. Within the Cranagh site it was the least numerous species prior to the dis- ruption in May 1979 (Fig. 2) and found more often than expected on the wall and less often on the trees and roof. Е. marginatus has a similar distribution to L. pseudoflavus. It is however, significantly as- sociated with trees, both in those areas in which it occurs (Table 1) and in its position within a single area. At the Cranagh it was rarely found on the ground (Table 2). L. maximus has a more restricted distribu- tion than either L. pseudoflavus or L. mar- ginatus. It occurs mostly at sites which also contain trees (Table 1), although at the Cranagh its frequency on the trees was sig- nificantly lower than expected. This was also apparent during the general survey in which L. maximus was normally found in the litter and among stones and fallen logs rather than actually on trees. There are significant differences in the sites of activity of these four species but these do not amount to a substantial stratification of the species within the habitat (Table 2). Previous work on the sites of activity of terrestrial molluscs also failed to demonstrate substan- tial spatial separation of closely related sympatric species (Cameron, 1978). The growth of young slugs can be followed in the size distributions (Figs. 3 to 7). As slugs become older the size boundaries between cohorts breaks down. This lack of distinction between the cohorts older than about six months is probably a consequence of the highly variable growth rates of slugs (Prior, 1983). Thus it is difficult to draw conclusions concerning the age structure of any but the smallest slugs. There are clearly at least two generations of L. pseudoflavus, L. flavus and L. maximus present throughout the year and most individ- uals of these species are capable of breeding in their first autumn (personal observation DJR). They are therefore probably polyvoltine and semelparous. Most large terrestrial gas- tropods have adopted this type of life cycle (Peake, 1978). Most slugs which have been studied breed in their first year (Runham 4 Hunter, 1970), though some of the larger polyvoltine shelled species take longer to 142 COOK & RADFORD PX(0:04) ХС (0:05) SITE OCCUPANCY о un ae 0:6 ANALYSIS 0-7 FX(0:07) PG (0:54) FG (0-22) РЕ (0-42) 0:5 0.4 FEEDING 0-3 «SITE FIG. 8. A three dimensional view of three measures of niche overlap. Each point represents the overlap between one pair of species (e.g. point PG represents the point for L. pseudoflavus and L. marginatus). The figures in brackets are the products of the three measures of overlap and indicate the proportion of the total space occupied in common by the species concerned. P- L. pseudoflavus, F- L. flavus, X- L. maximus, G- L. marginatus. mature (Cowie, 1984). Most of the large au- tumn individuals of L. marginatus die during the winter and spring to be replaced the following autumn by the previous years hatchlings. It therefore has a univoltine, iteroparous life cycle. Small L. pseudoflavus, L. flavus and L. maximus (0-0.15 g) were present in the field populations from September to May or June (Figs. 3 to 7). Those small slugs still present in late spring probably hatched a month or more earlier (Fig. 9). Hatching in the field therefore occurred from September to about April. This corresponds reasonably well to the laboratory egg-laying periods though there are some disparities for L. flavus. Small L. marginatus (0-0.15g) were present in the field populations in extremely low numbers throughout the year. This spe- cies is the smallest of the four and is an annual, but the prolonged presence of small individuals in the field (Fig. 7), together with the well defined egg-laying period (Fig. 9) have no obvious interpretation. Comparison of slugs from sites occupied by all four species with those from sites occupied by only one is potentially useful in identifying changes brought about by coexistence. Infor- LIMACID ECOLOGY 143 50 37256077131 3 5 LP MJJASONDJFMAM)JJASOND Month Egg production (eggs/slug/month) MJJASONDJFMAM)JJASOND Month 50 14060112 29532. 22 MJJASONDJFMAM)JJASOND Month 50 1 9722172 1 40 Ее MJJASONDJFMAMJJASOND Month FIG. 9. The egg production of slugs in culture (egg/slug/month). The numbers refer to the number of clutches of eggs laid per month. LP) L. pseudoflavus, LF) L. flavus, LX) L. maximus, LG) L. marginatus. mation has been presented for the size distri- bution for L. pseudoflavus and L. marginatus at such sites and for both species there was a significant increase in the proportion of small animals in populations where other species were not present. These differences in size distribution may be attributable to a variety of factors which may not necessarily be associ- ated with the presence or absence of other Limax species. Limax species are known to share day-time resting sites, at least under laboratory conditions (Cook, 1981b) despite exhibiting some interspecific aggression (Rollo & Wellington, 1979). The significance of disturbances in the home brought about by a second species with different requirements is unknown though it has been suggested that aggressive interactions between slugs may reduce the energy available for growth (South, 1982). The faecal string of slugs consists mostly of those items of food too large to be passed to the digestive gland for intracellular digestion and can therefore be used to indicate the sub- strate over which the animals had been feed- ing. Faecal analysis however does not give precise details of diet since only ingested items rejected by the stomach can be identi- fied. The analysis of the faeces shows that, unlike the other species L. maximus feeds pre- dominantly on vascular plant material. This correlates with the frequency with which it was found grazing on wood. The sites of activity of this species (Table 2) show a significant pref- erence for the roof and the ground both of which would provide rotting vascular plant ma- terial. The lack of mineral material in the faeces is further evidence that few L. maximus feed on the wall. About 27% of the L. maximus which were found feeding were grazing on lichens whereas only 8% of the faecal material was lichen. This discrepancy may arise be- cause the animals were collected whilst feed- ing on their way to normal grazing areas rather than actually at them and suggests that the animals on the roof had travelled there from resting sites on the ground. The sensitivity of the eggs of L. flavus to low temperatures, its comparative scarcity in field communities (Fig. 2 & Table 1) and the observation that Northern Ireland is near the northern limit of its European distribution (Kerney & Cameron, 1979) support the view of Bruijns et al. (1959) that this species is of Mediterranean origin. L. marginatus on the other hand has the most northerly distribution of the species under consideration, being found in Iceland and on the coast of northern Norway (Kerney & Cameron, 1979). The comparatively high viability of its embryos at 144 COOK & RADFORD TABLE 6. A comparison of the survival of eggs and embryos incubated at different temperatures. Egg viability (% of eggs producing a normal embryo) and embryo viability (% of normal embryos successfully hatching) are shown separately. 2 way analysis of уапапсе on агс sin square root transformed data for egg viability showed both main effects (Temperature, Е = 11 35; d.f. = 3,99; р < 0.001: Species, Е = 6.61; 4.1. = 3,99; р < 0.001) and the 2-way interaction (Species x temperature, Е = 3.21; d.f. 9,99; р < 0.002) to be significant. A similar analysis for embryo viability showed again that both main effects (Temperature, F = 17.94; d.f. = 3,90; р < 0.001: Species, Е = 7.49; d.f. = 3,90; р < 0.001) and the 2-way interaction (Species x temperature, Е = 4.23; d.f. = 9,90; р < 0.001) were significant. Incubation temperature 53С 10°С 150 20°C % viable % viable % viable % viable L. pseudoflavus egg 82.5 + 0.6 87.8 = 0.3 96.0 = 0.3 77.9 + 0.9 embryo ИВ. ЕЕ 5 99.5 = 0.2 99.4 = 0.3 82.2= 10/7 L. flavus egg 16.2 = 2.1 76.6 = 2.1 83.6 + 0.1 85er embryo 17.3 + 8.5 ЕВ 0:9 97.6 +05 92.4 + 0.6 L. maximus egg 79.4 + 2.1 90.5 = 1.1 83.6 + 0.7 88.6 + 0.2 embryo 16.4 - 6.5 hea 2.7 98.9 + 0.2 83:2 = 0:2 L. marginatus egg 54.4 + 3.3 86.8 + 15.4 71.9 + 1.6 63.9 embryo 82.7 + 9.2 67.0 17 62.9 + 1.6 0.0 =0 TABLE 7. The time in days to the hatching of the first egg in а batch (mean + s.e.) and the duration т days of the hatching period (mean + s.e.) A 2-way analysis of variance of the hatching times showed there to be significant main effects (Temperature — F = 1003.59; d.f. = 3,82; p < 0.001: Species — F = 6.96; d.f. = 3,82; p < 0.001) and 2-way interaction (Species x temperature — F = 7.92; d.f. = 8,82; p < 0.001). Temperature Species 52:6 10°C 15276 20°C L. pseudoflavus hatching time 146 = 47 63 = 32 0.3 20 salen duration 18:9873.4 4.0 + 1.2 2.9 + 0.4 4.3 = 2.9 L. flavus hatching time 18937 = 78 би ЕЕ]. 8 32 =03 23 +04 duration 18.0 + 18 13.8 3:4 4.0 = 1.0 JODER L. maximus hatching time 138 + 10.9 5622150 28705 207,07 duration 43.5 = 14.5 20.8 + 6.9 53:0 8.7 =37 L. marginatus hatching time 141 + 10.8 64 +13 45 +21 = duration 25.7 == 7.2 11.8'= 2.4 2.7 = 12 = 5°C and the low viability at 20°C (Table 6) time taken for the first egg of a batch to hatch. correlates well with this distribution. L. maximus showed the greatest duration of The period over which an egg batch the hatching period at all temperatures de- hatches clearly varies with temperature (Ta- spite having the shortest incubation time. This ble 7) but within a species there is little is surprising since it has been reported by difference when the duration of the hatching Prior (1983) that single clutches of eggs of L. period is considered as a proportion of the maximus all hatch on the same day. LIMACID ECOLOGY 145 TABLE 8. A summary of the differences between Limax species. Additional data from Cook (19816 *) and Kerney & Cameron (1979*). Factor L. pseudoflavus L. flavus L. maximus L. marginatus Predominant habitat type Wall/trees Walls ground trees Predominant food saxicolous saxicolous vascular corticolous type lichen lichen plant lichen Life cycle type polyvoltine polyvoltine polyvoltine univoltine semelparous semelparous semelparous iteroparous Temperature sensitivity of embryo Lowest mortality 15, 10? © 10,15,20° С 10,15,20° С SAC Highest mortality 5,20%C IAC 5° C 15,20° C Northern limit to distribution* ? Scotland/Denmark $. Norway/Sweden Iceland/N. Norway Peak egg-laying period Aug-Nov Jun-Dec Aug-Oct Nov-Feb Distribution in home * huddled huddled touching dispersed Whilst there are major differences in the sensitivity of these species to temperature which may be related to their geographical distribution, their broad egg-laying strategy is similar. Egg laying in autumn and early winter reduces the dangers from dehydration and predation by insects both to the eggs and the juveniles. Furthermore, egg-laying at this time allows slugs to overwinter as eggs, juveniles or adults giving a potentially flexible response to the varying conditions of winter. These four species of Limax are obviously different in morphology, but this and other work (Cook, 1981b) have shown substantial ecological and behavioural differences. These differences are summarised in Table 8. ACKNOWLEDGEMENTS We are grateful to Dr. Keith Day for criticis- ing an early draft of the manuscript and to N.E.R.C. (Grant No. GT4/77/RS/47) and the Garfield-Weston Trust for financial support to D.J.R. REFERENCES CITED ANDERSON, R., 1977, Mapping non-marine Mollusca in North-East Ireland. /rish Naturalist Journal, 19: 29-38. BOER, Р. J. DEN 1986, The present state of the competitive exclusion principle. Trends in Ecol- ogy and Evolution, 1: 25-28. BRUIJNS, М. F., АЕТЕМА, С. O. van R. 8 BUTOT, L. J. M., 1959, The Netherlands as an environ- ment for land Mollusca. Basteria, 23: 135-162. CAMERON, В. А. D., 1978, Differences in the sites of activity of co-existing species of land mollusc. Journal of Conchology, 29: 273-278. COOK, A., 1977, Mucus trail following by the pulmonate slug, Limax grossui. Animal Be- haviour, 25: 774-781. COOK, A., 1979, Homing by the slug Limax pseudoflavus Evans. Animal Behaviour, 27: 545-552. COOK, A., 1980, Field studies of homing in the pulmonate slug Limax pseudoflavus Evans. Journal of Molluscan Studies, 46: 100-105. COOK, A., 1981a, Huddling and the control of water loss by the slug Limax pseudoflavus Evans. Animal Behaviour, 29: 289-298. COOK, A., 1981b, A comparative study of aggre- gation in pulmonate slugs (Genus Limax). Jour- nal of Animal Ecology, 50: 703-713. COWIE, R. H., 1984, The life cycle and productivity of the land snail Theba pisana (Mollusca: Helicidae). Journal of Animal Ecology, 53: 311-326. DUNCAN, C. J., 1975, Reproduction. In Pulmon- ales, FRETTER, V. € PEAKE, J., eds., vol 1. Academic Press, London. pp. 309-365. EVANS, N. J., 1978, Limax pseudoflavus Evans: a critical description and comparison with related species. /rish Naturalist Journal, 19: 231-236. FORD, D. G. J., 1986, Rhythmic activity of the pulmonate slug Limax pseudoflavus Evans. D. Phil. Thesis, University of Ulster. 146 COOK & RADFORD GELPERIN, A., 1974, Olfactory basis of homing behaviour in the giant garden slug, Limax maximus. Proceedings of the National Academy of Sciences, 71: 966-970. GILLER, P. S., 1984, Community structure and the niche. Chapman & Hall, London. HUNTER, P. J., 1968, Studies of slugs on arable ground. II. Life cycles. Malacologia, 6: 391-399. KERNEY, M. P. & CAMERON, R. A. D., 1979, A field guide to the land snails of Britain and North-west Europe. Collins, London. PEAKE, J., 1978, Distribution and ecology of the Stylommatophora. In Pulmonates, FRETTER, V. & PEAKE, J., eds., vol. 2A. Academic Press, New York. pp. 420-526. PONTIN, A. J., 1982, Competition and coexistence of species. Pitman, London. PRIOR, D. J., 1983, The relationship between age and body size of individuals in isolated clutches of the terrestrial slug, Limax maximus (Linnaeus, 1758). Journal Experimental Zoology, 225: 321-324. QUICK, H. E., 1960, British slugs. Bulletin of the British Museum of Natural History, 6: 105-226. RICHTER, K., 1976, A method for individually marking slugs. Journal of Molluscan Studies, 42: 146-152. RUNHAM, N. W. & HUNTER, R. J., 1970, Terres- trial slugs. Hutchinson, London. SIEGEL, S., 1956, Non-parametric statistics for the behavioural sciences. McGraw Hill, London. SLOBODCHIKOFF, С. N. & SCHULZ, W. C., 1980, Measures of niche overlap. Ecology, 61: 1051-1055. SOKAL, R. R. & ROHLF, F. J., 1969, Biometry. Freeman, London. SOKOLOVE, P. G. & McCRONE, E. J., 1978, Reproductive maturation in the slug Limax maximus, and the effects of artificial photoperiod. Journal of Comparative Physiology, B, 125: 317-325. SOUTH, A., 1982, A comparison of the life cycles of Deroceras reticulatum (Müller) and Arion intermedius Normand (Pulmonata: Stylom- matophora) at different temperatures under lab- oratory conditions. Journal of Molluscan Studies, 48: 233-244. SOUTHWOOD, Т. В. E., 1978, Ecological meth- ods. Chapman & Hall, London. SPIGHT, Т. М. 1981. How three rocky shore snails coexist on a limited food resource. Researches in Population Ecology, 23: 245-261. STEPHENSON, M. F. & POOLE, T. B., 1976, An ethogram of the common marmoset (Calithrix jacchus jacchus): general behavioural repertoire. Animal Behaviour, 24: 428-451. ZIPPIN, C., 1956, An evaluation of the removal method of estimating animal populations. Biometrics, 12: 163-189. ZIPPIN, C., 1958, The removal method of popula- tion estimation. Journal of Wildlife Management, 22: 88-90. Revised Ms. accepted 18 February, 1987 MALACOLOGIA, 1988, 28(1-2): 147-157 INCOMPLETE CONVERGENCE OF SHELL SIZES AND SHAPES IN FOREST SNAIL FAUNAS FROM TWO CONTINENTS: A RELIC OF ENVIRONMENTAL HISTORY? R. A. D. Cameron Department of Extramural Studies, University of Birmingham, Birmingham, B15 2TT, United Kingdom ABSTRACT A comparison is made of the range of shell sizes and shapes in forest snail faunas from British Columbia and north-west Europe. While there are many species in each region which share characters of size and shape with species in the other, there are also differences. British Columbian faunas lack large tall-spired species, and have fewer flattened or globular species of medium or large size than those of Europe. Conversely, they have more very small species. While there are a few cases of close convergence, there are also species in each region which differ substantially from any found in the other. This non-convergence cannot be accounted for solely as a product of present environmental differences between the regions, either in climate or in vegetation. An explanation is offered in terms of the effects of on the Pleistocene and Holocene history of the two regions, and in particular on the opportunities for speciation and colonization presented. The absence of appropriately shaped ancestors in source areas for colonization after retreat of Pleistocene ice-sheets may be of particular importance. INTRODUCTION Ecological and evolutionary theories pre- dict that when similar ecological niches are occupied, in different regions, by different species, those species will show similarities in those aspects of their morphology that relate to the niche occupied. Where such similarities cannot be accounted for by common ances- try, they are examples of convergent evolu- tion (Cain, 1964). In particular cases, conver- gence may be remarkably exact, and the species concerned are referred to as ecolog- ical equivalents (e.g. Cox 4 Moore, 1973). Where such convergences are numerous, the situation provides strong evidence for the operation of natural selection (Cain, 1964). Where convergence does not occur, how- ever, a variety of explanations are possible, and are not mutually exclusive. Assumptions concerning the similarity of niches may be wrong; the morphological characters may not relate to the aspects of the niches which are similar; time elapsed since the occupation of the niche may be too short for convergence to be complete, or developmental constraints or adaptive troughs (Wright, 1932) may prevent convergence or delay its completion. This study compares certain features of shell morphology in terrestrial snail faunas from forests in coastal British Columbia and in north-west Europe. These regions experience very similar temperate and oceanic climates. Temperature regimes are very similar, Janu- ary means varying from 0-6” C and July means from 13-19” C—figures varying be- tween individual stations, but with complete overlap between regions (Anon., 1982a, 1984). The range of precipitation is great in both regions (700 mm-2000 mm + per year), but rainfall is more seasonal in coastal British Columbia, much more falling in winter than summer (Anon., 19826, 1984, Waring 4 Franklin, 1979). Nevertheless, many Euro- pean sites have summer rainfall of compara- ble magnitude to British Columbian sites. The range of forest soils and litter is also comparable, with podsolic soils and mor litter in poorer sites, and brown earths and mull litter in the richer ones (Klinka, Green, Trowbridge 8 Low, 1981). The close similarity of abiotic conditions generally in the two re- gions is confirmed by the practical experience of the British Forestry Commission; after much experimentation, plantings of non- native conifers in Britain are predominantly of species and stocks from the Pacific North- West, and particularly from the coastal region. (147) 148 CAMERON These perform better than stocks and species derived from more continental climates (Locke, 1970). Forests in coastal British Columbia are predominantly coniferous and evergreen. In N.W. Europe, such forests tend to occur in montane or high latitude zones climatically more extreme than the lowlands, which are dominated by deciduous broadleaves. Conif- erous forests tend to be associated with, and to produce acidic soils with mor litter, а сот- bination hostile to terrestrial snails. This as- sociation is not, however, complete; British Columbian conifer forests with brown earth soils and mull litter do exist. Furthermore, some British Columbian forests are domi- nated by broadleaf deciduous trees, and their snail faunas do not differ significantly from those of conifer forests with comparable soil and litter conditions (Cameron, 1986). Both in N. America and in Europe, conifer forests with appropriate soil and litter conditions have snail faunas comparable to those of decidu- ous forests, as demonstrated in Cameron (1986). Snail faunas from forests in both regions have been sampled in similar ways. They share some families, genera and species, but show sufficient taxonomic differences to make the comparison interesting. Features of morphology considered are those for which there is evidence for their adaptive signifi- cance. MATERIAL AND CHARACTERS USED Data on the composition of snail faunas from forests in coastal British Columbia are taken from Cameron (1986), and are based on 38 sample sites. One alien species, Val- lonia pulchella, found at one site, has been omitted from consideration. For European comparisons, the results of three studies are used. That of Cameron (1973), based on 44 sample sites in decidu- ous broadleaf forests on the South Downs, S. England, uses identical sampling techniques. That of Körnig (1966) covers a very wide range of habitats in central Germany; the rich Hangbuchenwalder series (20 sites), also from deciduous broadleaf forests is used here. British Columbian forests tend to be domi- nated by evergreen conifers, even on lime- stone and on ти! litter. The third European study used is that of Schmid (1966) from the Spitzberg, W. Germany, using 42 samples made in spruce (Picea) and fir (Abies) forests on favourable soils. Other quantitative Euro- pean studies in coniferous forests are not on soil and litter types comparable with the rich- est Canadian sites, and have impoverished faunas (see discussion in Cameron, 1986). The British Columbian, English and Ger- man Hangbuchenwalder samples were made by a combination of searching and litter sam- pling and sieving, and give comparable esti- mates of frequencies. Schmid's survey differs from these in one important respect; each sample represents the sieving and searching of 1 m? of litter, and frequency data are not comparable with those from the other studies, as larger species in particular will have lower frequencies per sample, and some may be missed altogether. The appendix lists the species recorded in each study, and their frequency of occurrence in the British Columbian, British and the Ger- man Hangbuchenwalder samples. Nomen- clature for the European sites follows Kerney 8 Cameron (1979), and for British Columbia Branson (1977) with a few exceptions noted in Cameron (1986). The principal characters used in this study are the height and maximum diameter of the adult shell. For European species, the data were obtained from Kerney and Cameron (1979), using mid-points where a range is given. For British Columbian species, the data come from measurements of adult shells collected by Cameron (1986), except in the cases of Euconulus fulvus and Zonitoides arboreus, where no shell measured ap- proached the dimensions given by Pilsbry (1939-1948). His data are used for these species; in all the rest, differences between his data and those obtained by measurement of Cameron's material are very slight. Data on number of whorls, and on the rate of whorl expansion come from the same sources. RESULTS Fig. 1 shows logarithmic plots of shell height and diameter for the species recorded in each study. Dashed diagonal lines indicate contours of approximate volume, derived from actual measurements of weight in Euro- pean species (Cameron, 1981), supple- mented by estimating volumes of cones with specified heights and basal diameters. The range of diameters, and of volume is SHELL CONVERGENCES IN FOREST SNAILS 149 20 HEIGHT mm 1 2 5 10 20 30 40 DIAMETER mm HEIGHT mm 1 2 5 10 20 30 40 DIAMETER mm HEIGHT mm 1 2 5 10 20 30 40 HEIGHT mm DIAMETER mm FIG. 1. Logarithmic plots of shell height and diameter for (A) species in British Columbian forests, (B) species in beechwoods in central Germany, (C) species in beechwoods in S. England, (D) species in west German conifer forests. Open circles represent species in genera common to both British Columbian and European sites. The bisector is the line of equal height and diameter. Dashed lines represent contours of volume: from left to right, 7 mm?, 100 mm?, 1400 mm*. similar in all scatters, which also shows the characteristic bimodality in height/diameter ratios described by Cain (1977). There are, however, differences between the regions in the proportion of species occur- ring in different parts of the scatters. In par- ticular, larger tall-spired species are missing in British Columbia, where only one species with a tall spire has a volume greater than 7 mm’. Other differences also occur (Table 1a). The British Columbian fauna has more very small flattened/globular species and fewer large species of the same shape than any of the European faunas. Species in genera common to the faunas in both continents are shown with open circles in figure 1. Since similarities between them might reflect common ancestry, table 1b shows the effect of removing them from the size/shape comparisons. Although their re- moval further reduces the already limited number of species involved, the trends noted above persist, and in some cases intensify. Two trivial phenomena could complicate 150 CAMERON TABLE 1. Numbers of species classified by size and shape of shell (a) for all species, (b) excluding species in genera common to both Europe and British Columbia. See text for sources of data. Tall-spired species Flattened and globular species Volume: <7 mm? 7-100mm? > 100 mm? <7mm? 7-100 mm? > 100 mm? Total a British Columbia 5 1 0 6 7 6 25 South Downs 3 6 2 3 7 13 34 C. Germany 4 9 6 4 8 15 46 W. Germany (conifers) 6 2 3 2 7 8 28 (b) British Columbia 0 0 0 4 3 6 13 South Downs 2 4 2 2 3 13 26 C. Germany 1 8 6 3 4 15 37 W. Germany (conifers) 1 1 =} 1 3 8 17 IAEA НЕЕ TABLE 2. Mean numbers of species per site classified by size and shape for 3 comparable studies. EE <7 mm 7-100 mm? > 100 mm? Tall-spired species British Columbia 2.6 0.2 0 South Downs 1.9 3.7 2.0 C. Germany 0.9 3.7 2.9 Flattened and globular species British Columbia 4.5 2.8 4.4 South Downs 1.8 3.9 6.1 C. Germany 1.8 3.6 9.0 e m — nn ————äää— the interpretation of these trends. Replace- ment of some species by others of similar size and shape in various samples from the same region would inflate {пе number of species in that size and shape class relative to a region in which the same species were present in all samples. A similar bias could occur if rare or accidental species, occurring in only one or two sites were commoner in one region than in another. The effects of these phenomena can be removed by weighting each species by the frequency of its occurrence, a procedure which gives, in effect, the number of species of any given size and shape that one would expect to find in a single sample. Schmid's (1966) data for German coniferous woods are not comparable with others and have not been used. In both the English and British Columbian surveys the sites studied include some with acid soils and impoverished fau- nas, whereas the Hangbuchenwalder series of Kórnig is ecologically uniform. To allow for this, frequencies of occurrence in England and British Columbia are based on the richest association-type found: the group C sites (n = 24) of Cameron (1973), and the mull series (n = 19) of Cameron (1986). Table 2 shows that differences between the British Columbian and European faunas are maintained when frequencies are used. : Other similarities and differences between the shells of species in the two regions are more subtle, and are best considered in com- parisons between species of similar size, shape and habit. Amongst very large species (volume 1400 mm? +), there are no very precise convergences (Table 3a). The European spe- cies (helicids and one bradybaenid) are all more globular, and have more rapidly ex- panding whorls relative to their size (Cameron, 1981) than British Columbian M. fidelis and A. townsendiana which also differ from the helicids in being umbilicate, and in having more whorls. The closest match, other SHELL CONVERGENCES IN FOREST SNAILS 191 TABLE 3. Shell characters of (a) species larger than 1400 mm, (b) species showing close inter-continental resemblances between 100 and 1400 mm?*. H/D = height to diameter ratio; Whorls = number of whorls to nearest 0.25 whorl; Umbilicus = width of umbilicus relative to diameter, Raup's W = rate of whorl expansion. See Cameron (1981) for details of measurement. (a) Very large species mm Diameter H/D Whorls Umbilicus Raup's W Europe Arianta arbustorum 21 0.76 5.5 0.02 1.66 Bradybaena fruticum 20 0.85 6 0.11 1.80 Cepaea hortensis 18 0.83 5.25 0 172 Cepaea nemoralis 24 0.76 5.5 0 172 Helix aspersa 36 0.89 5 0 2.09 Helix pomatia 45 0.95 5:5 0 2.16 British Columbia Monadenia fidelis 36 0.64 6.5 0.08 1.56 Allogona townsendiana 27 0.62 5.75 0.09 1.64 Haplotrema vancouverense 24 0.48 5.25 0.20 1.94 (b) Medium-to-large species—closest comparisons only mm Diameter H/D Whorls Umbilicus Raup’s W Hairs Barriers Europe Isognomostoma isognomostoma 9 0.61 5.5 0 1.38 yes yes Perforatella incarnata 14 0.69 6 0.09 1.40 no no Trichia hispida 8 0.63 6 0.20 1.41 yes no Trichia plebeia 8 0.69 5:5 0.13 1.45 yes no Trichia striolata 13 0.61 6 0.17 1.53 juv. no British Columbia Vespericola columbiana 13 0.67 5.25 0.05 1.44 yes no Triodopsis germana 7 0.65 5 0.05 132 yes yes than in size, is between Arianta arbustorum and Allogona townsendiana, a match which extends to shell pattern and colour. All these species have in common a relatively thick shell, and an everted or thickened peristome. The other large British Columbian snail, H. vancouverense, has no north- west European equivalent—it is very flattened, has a thin, horny shell, a large umbilicus and a simple peristome. In these features, as in its reput- edly carnivorous habits, it resembles the zonitids of north-west Europe, but is far larger than any of them. Amongst somewhat smaller species (100-1400 mm?) which are globular or flat- tened, the European fauna is more diverse both in taxa and in the range of shell charac- ters. Of the three British Columbian species, H. sportella has some resemblances to zonitids, but has athicker, sculptured shell and an everted peristome. The two polygyrids, T. germana and V. columbiana do show rather close resemblances to a number of helicids (Table 3b), but there are others which lack British Columbian equivalents, such as the sharply keeled Helicigona lapicida. There are fewer species in the smaller size classes which lack congeners on the other continent. European Vitrea species show some resemblances to North American Microphysula, and, to a lesser extent to Pristiloma, which, while shiny and tightly coiled, are brown, and have an appreciable spire. The minute British Columbian Plan- ogyra clappi and Striatura pugetensis have no obvious European equivalents; both expand their whorls more rapidly, and have larger umbilicuses than either European or N. Amer- ican Punctum species. All tall-spired British Columbian species 152 CAMERON have congeners in north-west Europe, and congeners resemble each other very closely. The larger tall-spired European snails belong to families not present in N. America. DISCUSSION Convergence or parallelism for the charac- ters studied in these British Columbian and European snail faunas is far from complete. There are some convincing convergences be- tween distantly-related species, and other similarities which could have derived from common ancestry. There are also differences between the faunas, both containing species for which there is no obvious equivalent in the other. The characters studied relate to mode of life in snails. Cain (1977, 1981) has shown that there is a near-universal bimodality in the dis- tribution of height-diameter ratios in terrestrial pulmonate faunas, and studies both in this field and in the laboratory have shown asso- ciations between this measure of shape and the preferred angle and type of substrate for activity (Cain 4 Cowie, 1978, Cameron, 1978, Cook & Jaffar, 1981). High-spired species tend to prefer hard vertical surfaces, such as rock, tree trunks and logs, while more flattened species resort more to horizontal surfaces. Globular shelled species may be better adapted to mobile surfaces (such as living or senescent herbaceous plants). Cain's work, cited above, makes it clear that these repeated patterns of size and shape have arisen inde- pendently in faunas of different taxonomic or- igin. Cameron (1981) discusses the functional significance of some other shell characters used, such as rate of whorl expansion. In terms of present environment in the two regions, the slight climatic differences seem inadequate as an explanation of these faunal differences. British Columbia does have a more seasonal pattern of precipitation than north-west Europe, but European forest Snails of shapes and sizes not found in British Columbia, especially in the families Helicidae, Clausilidae and Enidae, extend into the Mediterranean region which has hot dry sum- mers. Soil and litter conditions are compara- ble; all the sample series considered here include sites on limestone derived soils with high pH—the optimum for snail diversity. Differences in the nature of the predominant forest cover could be of more significance. Much of the literature on forest molluscs (re- viewed in Cameron 1986) stresses the com- parative poverty of coniferous forest on both continents. This poverty is not, however, sim- ply a product of conifer cover per se; it is pri- marily a consequence of the nutrient-poor and acidic soils, of the sometimes excessively dry and wet conditions, and of the adverse cli- mates on and in which conifer forests are com- monly found. Where these conditions are ameliorated, conifer forests can support di- verse snail faunas. Thus, in British Columbia, snail faunas from conifer stands in bottom- lands or on limestone derived soils do not differ significantly from those from broadleaf stands in the same situation. The example of the Spitzberg, used here, is not the only one of species-rich coniferous forest in Europe. Other, less quantitative studies (Summarized in Cameron, 1986) also describe diverse fau- nas from coniferous forests, including species of sizes and shapes not found in British Co- lumbia. Given the botanical and climatic sim- ilarities involved, and the absence of clear dis- tinctions in the faunas of coniferous and broadleaved forests in both regions, the dif- ferences seen would not be predicted on a simple a priori hypothesis that available niches should be filled. Direct effects of conifer cover should not, however, be completely excluded. In particu- lar, there might be a connection between the relatively high proportion of very small spe- cies of snail and the small average particle size of litter. Such small species (especially Vertiginidae) predominate also in conifer for- ests in colder climates (e.g. Wareborn, 1969). If characters considered relate to niche occupied, and present physical and botanical environments do not appear to prescribe a radically different range of potential niches, it would seem that some of the niches are empty. Why might this be so? One possible explanation might lie in differ- ential risk from predators between regions, rendering certain niches untenable in one region. There is, at present, insufficient evi- dence to refute or confirm such an hypothe- sis. Cain's (1977) scatters of height and breadth, based on many faunal regions, show strong representation above and below the bisector in regions of diverse climates and habitats, in which the range of predators presumably differs. Only in rather harsh, cold continental climates does the upper scatter attenuate or disappear (Cain, 1981). Given the slow active dispersal of snails, and the morphological conservatism shown in many families (Cain 1977, 1981), a part of the SHELL CONVERGENCES IN FOREST SNAILS 153 answer might lie in the histories of the regions concerned and the nature of the snail faunas available to colonize them. The environmen- tal history of both regions is heavily influenced by the Pleistocene glaciations. Both were subject to glacial or peri-glacial conditions at the last advance of the Pleistocene ¡ce sheets as recently as 14-15,000 years ago (Wright, 1983, 1984, West, 1968), and their present flora and fauna are derived from subsequent colonization from the south. Solem (1984), in a global review of snail diversity, has suggested that snail faunas of areas subject to such drastic changes may be far from the maximum diversity which the habitat could sustain. Evidence that niches are not always filled, nor constrained by com- petition comes also from studies on other communities and guilds (Lawton, 1984 and others in Strong, Simberloff, Abele & Thistle, 1984). Differences between faunas in such recent environments may depend on the ac- cident of which species were available to colonize the newly available territory. In this context, differences in the structure of the regions could also be important. The coastal zone of British Columbia is part of a narrow strip running from Alaska to California, bordered to the east by high mountain ranges, on and behind which climatic regimes are drastically different. Belts of arid and alpine zones seal off the coastal strip from the interior. Pleistocene movements of fauna and flora have been largely north-south, with a similar pattern of succession and regression in each interglacial or interstadial (Heusser & Heusser, 1981). By contrast, oceanic climate influences pen- etrate far into the Eurasian continent (c.f. dis- cussion in Cain, 1981). The east-west axis of European mountains may be responsible for a general impoverishment of the north Euro- pean fauna and flora relative to that of the Appalachians, east of the American continen- tal divide (e.g. MacArthur, 1972). Neverthe- less, it is clear that the present snail fauna of north-west Europe derives from movements on an east-west as well as a north-south axis, and that the structure of the mountain ranges, especially of the Alps and Pyrenees has favoured speciation and local radiations (Kerney & Cameron, 1979). In north America, similar, indeed greater, local radiations have occurred in the more southerly Appalachians and Ozarks (Pilsbry, 1939-1948) but nottothe west of the northern Cascades, where the coastal forest is hemmed in by alpine environ- ments hostile to snails. The European snail fauna may have a greater diversity of origins than that of British Columbia: given the very short time (in evolutionary terms) that each area has been occupied by forest, such a dif- ference would be more important than evolu- tion in situ (Solem, 1984). Western European forests are also more heterogeneous than those of British Colum- bia. In particular, submontane and hilly re- gions may have intimate mixtures of conifer- ous, mixed and broadleaved forests. Opportunities for species originating in one to colonize the other are great. Other effects might delay the filling of niches. Many families of snails show consid- erable morphological conservatism (Cain, 1977, 1981). The markedly bimodal nature of the height/diameter scatter for pulmonate land snails indicates that, in most environ- ments, there is an adaptive trough between the modes that is rarely crossed. The ab- sence of large, tall-spired species in coastal British Columbia may be a consequence of a lack of appropriate ancestors in areas from which colonization took place. Even temper- ate, deciduous forests in North America lack large tall-spired species (Coney, Tarpley, Warden & Nagel, 1982; Cain, 1981). Although the European faunas discussed here are more diverse than that of British Columbia, both taxonomically (Cameron, 1986) and in terms of size and shape, the latter nevertheless contains some forms (e.g. Haplotrema) not present in Europe. Whether this is another accident of history, or a con- sequence of the prior appropriation of the relevant niche by the great diversity of smaller, but similarly shaped Zonitids in Eu- rope remains to be determined. Given the still scanty knowledge of the determinants of land-snail niches, this discus- sion of causes of the differences between faunas is bound to be speculative. The obvi- ous experiment, the introduction of species of sizes and shapes not represented in the indigenous fauna, is clearly objectionable on grounds of conservation. For the same reason, detailed comparisons with faunas in very different climates would be premature. North-east Asian pulmonate fau- nas share with British Columbia (and the Pacific north-west generally) a deficiency in larger tall-spired shells. Cain (1981) points to the rigorous climatic regime there, in contrast to the milder climate at the same latitudes in Europe as a possible causal agent. This 154 CAMERON explanation cannot, in itself, account for the British Columbian scatter. The most diverse land-snail faunas known come from scrub woodlands on the North Is- land of New Zealand (Solem, Climo & Roscoe, 1981, Solem & Climo, 1985), where there is a fauna of more than 80 species, and where 60 species may be found т a single site. Analysis ofthese data reveals, however, that more than 80% of the species concerned are flattened or globular and have volumes less than 100 mm?. Numbers of species in all other cat- egories, and especially of high spired forms, are fewer than in European forests. Solem (1984) gives convincing reasons for the high level of taxonomic diversity in these New Zealand faunas (including a long history of climatic stability, enabling niches to be filled by both evolution and colonization). We now need to explore the reasons why this great diversity 1$ expressed only in a very limited part of the spectrum of size and shape found else- where. The incomplete convergence of the faunas compared here, and the results of Solem's (1984) survey suggest that, in certain circum- stances, empty niches may occur in snail fau- nas. If further confirmed by other studies, this would in turn suggest that interspecific com- petition is by no means the most important determinant of snail fauna diversity, whether considered taxonomically or morphologically (Strong et al., 1984). It does not follow that shell size and shape are of no adaptive sig- nificance. Cain (1977, 1981) shows that there is good reason to think that the mechanical consequences of modes of life exert powerful selective forces which constrain the range of morphologies actually found. ACKNOWLEDGEMENTS This work was supported by the award of a British Ecological Society Travelling Fellow- ship and by the University of Birmingham. | should like to thank Professor G. G. E. Scudder for the use of facilities at the Univer- sity of British Columbia, and Professor A. J. Cain for criticism of the manuscript. LITERATURE CITED ANONYMOUS, 1982A, Canadian Climatic Normals Vol. 2: Temperature 1951-1980. Environment Canada: Ottawa. ANONYMOUS, 1982B, Canadian Climatic Normals Vol. 3, Precipitation 1951-1980. Environment Canada: Ottawa. ANONYMOUS, 1984, Geographia: London. BRANSON, B. A., 1977, Freshwater and terrestrial Mollusca of the Olympic Peninsula, Washington. Veliger, 19: 310-330. CAIN, A. J., 1964, The perfection of animals. View- points in Biology, 3: 36-63. 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, B, 277: 377-428. CAIN, A. J., 1981, Variation in shell shape and size of helicid snails in relation to other pulmonates in faunas of the Palaearctic region. Malacologia, 21: 149-176. CAIN, A. J. & COWIE, R. H., 1978, Activity of different species of land-snail on surfaces of different inclinations. Journal of Conchology, 29: 267-272. CAMERON, R. A. D., 1973, Some woodland mol- lusc faunas from southern England. Malacologia, 14: 355-370. CAMERON, В. А. D., 1978, Differences in the sites of activity of co-existing species of land molluscs. Journal of Conchology, 29: 273-278. CAMERON, R. A. D., 1981, Functional aspects of shell geometry in some British snails. Biological Journal of the Linnean Society, 16: 157-167. CAMERON, R. A. D., 1986, Environment and di- versities of forest snail faunas from coastal Brit- ish Columbia. Malacologia, 27: 341-355. CONEY, С. C., TARPLEY, W. A., WARDEN, J. С. 8 NAGEL, J. W., 1982, Ecological studies of land snails in the Hiwassee River Basin of Tennes- see, U.S.A. Malacological Review, 15: 69-106. COOK, L. М. & JAFFAR, W. М., 1984, Spire index and preferred surface orientation in some land snails. Biological Journal of the Linnean Society, 21: 307-313. COX, C. B. & MOORE, P. D., 1973, Biogeography. Blackwell, Oxford. HEUSSER, С. J. 4 HEUSSER, L. E., 1981, Palynology of the Whidbey Formation, Puget Lowland, Washington. Canadian Journal of Earth Sciences, 18: 136-149. KERNEY, M. P. & CAMERON, R. A. D., 1979, A field guide to the land snails of Britain and north-west Europe. Collins, London. KLINKA, K., GREEN, В. N., TROWBRIDGE, В. L. 8 LOWE, L. E. 1981. Taxonomic classification of humus forms in ecosystems of British Columbia. Land Management Report 8, Ministry of Forests, Province of British Columbia. KÖRNIG, G., 1966, Die Molluskengesellschaften des mittledeutschen Húgellandes. Malakolo- gische Abhandlungen staatliches Museum für Tierkunde in Dresden, 2: 1-112. LAWTON, J., 1984, Non-competitive populations, non-convergent communities and vacant niches: the herbivores of bracken. in: STRONG, D. R., Atlas of the World. SHELL СОМУЕВСЕМСЕ$ IN FOREST SNAILS 155 SIMBERLOFF, D., ABELE, L. G. & THISTLE, A. B. eds.: Ecological Communities: Conceptual Issues and the Evidence. Princeton University Press. LOCKE, G. M. L., 1970, Forestry Commission Census of Woodlands 1965-67. H.M.S.O. Lon- don. MACARTHUR, В. Н., 1972, Geographical Ecology. Princeton University Press. PILSBRY, H. A., 1939-1948, Land Mollusca of North America (north of Mexico). Academy of Natural Sciences of Philadelphia, Philadelphia. 2 vols., each in 2 parts. SCHMID, G., 1966, Die Mollusken des Spitzbergs. In: Der Spitzberg bei Tübingen, Natur und Landschaftsschutzgebiete Baden— Wurttemburg, Ludwigsburg, 3: 596-701. SOLEM, A., 1984, A world model of land snail diversity and abundance. In: SOLEM, A. & УАМ BRUGGEN, A. C., eds. World-wide snails. E. J. Brill, Leiden. SOLEM, A., CLIMO, Е. M. & ROSCOE, D. J., 1981, Sympatric species diversity of New Zealand land snails. New Zealand Journal of Zoology, 8: 453-485. SOLEM, A. 8 CLIMO, F. M., 1985, Structure and habitat correlations of sympatric New Zealand land snail species. Malacologia, 26: 1-30. STRONG, D. R., SIMBERLOFF, D., ABELE, L. С. 8 THISTLE, A. B., 1984, Ecological Communi- ties: Conceptual Issues and the Evidence. Princeton University Press. WAREBORN, I., 1969, Land molluscs and their environments in an oligotrophic area in southern Sweden. Oikos, 20: 461-479. WARING, R. H. & FRANKLIN, J. F., 1979, Ever- green coniferous forests of the Pacific North- west. Science, 204: 1380-1386. WEST, R. G., 1968, Pleistocene geology and biol- ogy. Longmans, London. WRIGHT, H. E., ed., 1983 and 1984. Late Quater- nary environments of the United States. Vol 1. The Late Pleistocene, Vol. 2. The Holocene. Longmans, London. WRIGHT, S., 1932, The role of mutation, inbreed- ing, crossbreeding and selection in evolution. Proceedings of the 6th International Congress of Genetics, 1: 356-366. Revised Ms. accepted 23 December, 1986 156 CAMERON APPENDIX Lists of species found in the studies used in this paper. Asterisks mark species in genera common to both regions, and found in these studies. (a) Species recorded in British Columbia, and their frequencies of occurrence. For details of sites etc. see text and Cameron (1986). Allogona townsendiana 0.11 Pristiloma lansingi 1.00 *Carychium occidentale 0.53 Pristiloma stearnsi 0.37 *Cionella lubrica 0.16 *Punctum conspectum 0.37 "Columella edentula 0.95 *Punctum randolphi 1.00 *Discus cronkhitei 0.16 Striatura pugetensis 1.00 *Euconulus fulvus 0.79 Triodopsis germana 0.63 Haplotrema sportella 0.89 “Vertigo andrusiana 0.11 Haplotrema vancouverense 0.95 “Vertigo columbiana 0.89 Microphysula cookei 0.16 “Vertigo rowelli 0.11 Monadenia fidelis 0.84 Vespericola columbiana 0.89 *Мезоуйгеа binneyana 1.00 “Vitrina alaskana OH Planogyra clappi 1.00 Zonitoides arboreus 0.26 Pristiloma johnsoni 0.11 (b) Species recorded in 3 European studies, with frequencies of occurrence for those of Cameron (1973) and Körnig (1966). For details see text. Note that Cionella is given here as Cochlicopa. South Downs German beechwoods German conifer woods Cameron (1973) (Körnig, 1966) (Schmid, 1966) Abida secale 0.08 0.10 — Acanthinula aculeata 0.29 0.30 4 Acicula fusca 0.83 — — Acicula polita — — + Aegopinella nitens = 0.65 aa Aegopinella nitidula 1.00 0.70 — Aegopinella рига 0.92 0.95 ar Arianta arbustorum 0.12 0.05 — Azeca goodalli — 0.50 — Bradybaena fruticum — 0.40 — Bulgarica cana — 0.10 — *Carychium tridentatum 1.00 0.65 + “Carychium minimum — 0.05 + Cecilioides acicula 0.04 ONS — Cepaea hortensis 0.62 0.60 + Cepaea nemoralis 0.54 0.75 + Clausilia bidentata 0.71 0.90 = Clausilia dubia — 0.25 — Clausilia parvula — 0.15 — *Cochlicopa lubrica 0.51 0.40 + *Cochlicopa lubricella 0.20 — — Cochlodina laminata 1.00 1.00 ar *Columella edentula — 0.05 + *Discus rotundatus 1.00 1.00 + Ena montana — 0.85 + Ena obscura 0.75 0.90 + *Euconulus fulvus 0.54 0.60 + Euomphalia strigella — 0.05 — Helicigona lapicida 0.18 0.75 — Helicodonta obvoluta 0.50 0.95 + Helix aspersa 0.88 — = Helix pomatia — 0.95 + SHELL CONVERGENCES IN FOREST SNAILS South Downs German beechwoods Cameron (1973) (Körnig, 1966) Iphigena pliculata — 0.20 Iphigena ventricosa — 0.55 Isognomostoma isognomostoma — 0.45 Laciniaria biplicata — 0/35 Macrogastra rolphii 0.75 — Monacha cantiana 0.04 — *Мезоуйгеа hammonis à 0.05 Orcula doliolum — 0.25 Oxychilus alliarius 0.83 0.25 Oxychilus cellarius 0.79 0.90 Oxychilus helveticus 0.13 — Perforatella incarnata — — Pomatias elegans 1.00 — *Punctum рудтаеит 0.54 0.25 Semilimax semilimax — 0.05 Trichia hispida 0.67 0.85 Trichia plebeia = 0.05 Trichia striolata 0.62 -- “Vertigo pusilla — — “Vertigo substriata = — Vitrea contracta 0.96 0.80 Vitrea crystallina 0.04 0.10 Vitrea diaphana — 0.40 “Vitrina pellucida 0.58 0.55 Zenobiella subrufescens — *Species not recorded in the sites chosen for calculation of frequencies, see text. 197 German conifer woods (Schmid, 1966) + a [+++ | el MALACOLOGIA, 1988, 28(1-2): 159-273 THE GENITALIC, ALLOZYMIC, AND CONCHOLOGICAL EVOLUTION OF THE EASTERN NORTH AMERICAN TRIODOPSINAE (GASTROPODA: PULMONATA: POLYGYRIDAE) Kenneth C. Emberton' Committee on Evolutionary Biology, University of Chicago, Chicago, IL 60637, U.S.A. ABSTRACT The 40 species of triodopsines in eastern North America are useful for evolutionary studies because of their diverse genitalic and conchological radiations. Previous monographs were based on shells, many features of which are subject to convergence. Dissection of the uneverted penial tubes revealed a morphological diversity that was classified into 10 characters comprising 60 character states. Cladistic analysis yielded a single most parsimonious tree with a consistency index of .970. Starch-gel electrophoresis of foot tissue detected 74 alleles among 16 loci. Cladistic analysis using the independent alleles model resulted in a consensus, maximum-parsimony tree with a consistency index of .950. Electrophoresed populations were divided into two equal subsets for rooted distance-Wagner analyses based on Prevosti distances. The resulting trees had cophenetic correlations of .897 and .883. The anatomical and allelic cladograms and the two genetic-distance trees were weighted according to the sizes and reliabilities of the data bases used in their construction. Branch-by- branch comparison of the four weighted trees produced a consensus phylogeny that was quite robust, and with only a few species remaining problematic due to incomplete or conflicting data. Supraspecific revision based on this consensus phylogeny divides eastern triodopsines into four genera: Neohelix von lhering, 1892; Triodopsis Rafinesque, 1819; Webbhelix Emberton, new genus; and Xolotrema (Rafinesque, 1819). The revision differs most strongly from previous classifications in its species groupings within the large genus Triodopsis. Revision of the Neohelix albolabris group (the “white-lipped land snail”), based on 46 populations, discovered two new taxa: N. solemi and N. albolabris bogani. A cladogram (consistency index 1.00) based on genitalic morphometrics formed the basis for revision, which split the group into the albolabris and alleni groups. Shell differences among taxa are subtle and occasionally unreliable for identification, according to a multivariate discriminant analysis. Genitalic and geographic comparisons between 25 pairs of sister taxa detected a pattern: sister taxa with virtually identical penial morphologies generally have peripatric geographic ranges, those slightly different are generally allopatric, those moderately different are sympatric, whereas those greatly different are parapatric. Population-level comparisons for 12 species failed to find any trace of reproductive character displacement. These results, as well as the pattern of genitalic convergences and the geographic stability of within-species genitalic morphology, led to the hypotheses that (1) peripheral isolates generally do not differentiate, (2) vicariant isolates gen- erally differentiate slowly, (3) differentiation due to reproductive character displacement is mod- erate at most, and (4) major differentiation is rare, rapid, and occurs in isolates. Shell evolution’s pattern and inferred process differs among taxonomic levels. Genera show general conchological stasis despite extensive, overlapping ecological radiations; the process is probably canalization. Species groups show mosaic distributions of minor shell characters; the process is presumably genetic indeterminism of canalized developmental programs. Species and populations show two patterns: patchy, non-clinal variation in size and some aspects of shape and sculpture, probably induced by local microclimates; and iterated environmental correlations—e.g., between apertural obstruction and ground moisture, spire flatness and crevice-dwelling, and periostracal glossiness and water—presumably due to natural selection. The nature and definition of a species in eastern triodopsines remains both a problem and a fruitful avenue of research. The many sympatric shell convergences between eastern triodopsines and the polygyrine genus Mesodon provide naturally replicated experiments in evolutionary morphology. Key words: snails; evolution; genitalia; allozymes; shells; cladistics; character displacement; natural selection; convergence. Present address: Department of Malacology, Academy of Natural Sciences, 19th & the Parkway, Philadelphia, 19103 (159) 160 TABLE OF CONTENTS INNOGNENOR ne ae ae Materials and Methods.................. TAXA SUIS ter ie ones Collections 54-32. DISSOGUONS » 2-2 нина SH ANANSIS An ee Élegtophoresis. scene Data ane sis en Secret ee Patterns of genitalic evolution......... Patterns of shell evolution ............ Taxonomie ЮГУ zes. Genitalic analysis....................... VAHAHON deu ana ae DESCUIDO о gez Suggested character-state iransiomations: {520.100 Cladisheanalvsiss 4e Allozymicanalysesaus. sets. Consensus phylogeny .................. Conchological variation ................. Revision of the Neohelix albolabris OUD ce en steel 950 Genitalic analysis... once 8... CladiSic Е sis scene Snellanalyssae tar ne Revised classification................. General supraspecific revision .......... Patterns of genitalic evolution........... Patterns of shell evolution .............. DISCUSSION tios oe Genitalic analysis..................... Allozymic Analysis... Robustness of the consensus A ea ck Genitalic evolution: pattern and ГОО. Shell evolution: pattern and PROCESS en Baer What is a species in the eastern American triodopsines? ............ Recommendations for future ОБА Ya ae Acknowledgements ..................... ÉHeralnre CEE 6454 ea: Appendix A. Electrophoretic DKOCCOIN 25.0. ra eine Appendix B. Systematic review of the Neohelix albolabris and alleni OUPS Sate) re ee Appendix C. Systematic review of the supraspecific taxa of the eastern American Triodopsinae............. Appendix D. On the phylogeny of the Triodopsis fallax group ............. EMBERTON 248 INTRODUCTION The Polygyridae are an autochthonous North American family of pulmonate land snails comprising approximately 260 species currently classified into 14 genera in 3 subfamilies (Pilsbry, 1940; Webb, 1954a; Hubricht, 1985; Richardson, 1986). This pa- per deals with eastern members of the subfamily Triodopsinae. Western triodop- sines comprise the single genus Vespericola Pilsbry, 1939, which has some 9 species and ranges along the Pacific coastal zone from southern Alaska to northern California (Pilsbry, 1940; Roth, 1984) Eastern triodop- sines, as revised in this paper, comprise the four genera Neohelix von lhering, 1892, (7 species); Triodopsis Rafinesque, 1819 (26 species, not including the Siberian “Trio- dopsis” supersonatum—see Emberton, 1986); Webbhelix Emberton, new genus (1 species); and Xolotrema Rafinesque, 1819 (5 species). They range throughout temperate North America east of the Great Plains. The eastern triodopsines are a common, large (8-40 mm), and sometimes dominant element of the leaf-litter invertebrate fauna. Eastern triodopsines are important for several reasons. (1) Because of their multiple sympatric conchological convergences on the polygyrine genus Mesodon (Pilsbry, 1940; Solem, 1976; Emberton, 1986), they contain superb naturally replicated experiments in evolutionary morphology (see Emberton, 1986). (2) Their diversity of complex penial morphologies (Webb, 1947-1980) makes them useful for testing the recent general hypotheses of Eberhardt (1985) concerning genitalic evolution. (3) Their substantial con- chological variation (e.g., Vagvolgyi, 1968) makes them useful for advancing our very limited knowledge of the adaptive vs. ecolog- ically induced components of shell shape in land snails (see review by Goodfriend, 1986). (4) The large size, high density, low vagility, and easy markability of many species make them useful subjects for generalizable studies in population biology (McCracken, 1976), population genetics (McCracken, 1980; Mc- Cracken & Brussard, 1980), life history and ecology (Vail, 1978; Emberton, 1981), and anatomy (Simpson, 1901; Emberton, 1985). (5) They are economically and ecologically important as the intermediate hosts of some- times lethal parasites of elk, deer, and other EASTERN NORTH AMERICAN TRIODOPSINAE 161 game and non-game mammals (e.g., Maze & Johnstone, 1986). (6) Their larger species are potentially of economic value as sources of anti-A agglutinin for typing human blood (Miles, 1983). Previous monographic treatments of the eastern American triodopsines (Pilsbry, 1940; Vagvolgyi, 1968) were conchological. The purposes of this paper are (1) to derive a robust phylogenetic hypothesis for the east- ern triodopsines using two independent data sets: male genitalia and allozymes; (2) to revise the eastern triodopsines above the spe- cies level, based on this phylogeny; (3) to analyze phylogenetic patterns of variation in both genitalia and shell morphology and to generate hypotheses about the evolutionary processes which produced these patterns; and (4) to further revise the Neohelix albolabris group to the subspecies level. The Neohelix albolabris group contains the largest, most conspicuous triodopsine snails. McCracken 8 Brussard’s (1980) electro- phoretic survey of “the white-lipped snail” (Neohelix albolabris [Say, 1816]) showed a confusing geographic diversity in this group which pointed out the need for taxonomic resolution using anatomical and conchologi- cal characters. Penial morphology, presumed to be impor- tant in species recognition and of great poten- tial value for the systematics of eastern triodopsines (Pilsbry, 1940; Webb, 1947- 1980; Solem, 1976), has previously been exploited only to a very limited extent. There are three ways of studying penial sculpture in land pulmonates (Fig. 1): by killing and fixing the snail relaxed and extended from its shell, then dissecting open the uneverted penial tube (the dissective method); by killing and fixing the snail in copulo so as to keep its penis fully everted (the evertive method); or by clearing, staining, and mounting the uneverted penial tube (the slide-mount nethod). Until the beginning of Webb's publi- cations in 1947, the only triodopsine species for which details of penial sculpture were known was Neohelix albolabris, illustrated by Binney (1851), Pilsbry (1894, 1940), and Simpson (1901); in all four of these papers it was studied by the dissective method. Webb (1947, 1948, 1952, 1954, 1959) studied 12 species and Grimm (1975) studied one spe- cies of eastern triodopsines by the evertive method and illustrated the general aspects of penial sculpture. Solem (1976) illustrated the dissected uneverted penial tubes of three species, thereby redemonstrating the efficacy of the dissective method and showing the wealth of sculptural detail omitted by Webb and Grimm. The dissective method is in many respects superior to both the evertive and slide-mount methods (Fig. 1). Waiting for penial eversion, then killing and fixing without distorting the soft tissues, is labor-intensive and yields little additional information (but see Character 10 below). Clearing and mounting the uneverted penial tube is more time-consuming than cut- ting it open, and is much less effective for interpreting complex sculpture because of three-dimensional overlap further distorted by viewing through other tissues. Thus the most obvious source of useful characters for phylogenetic analysis was penial sculpture as viewed by the dissective method. For this character set, 27 of the 40 species of eastern triodopsines had never been examined before, and, of those that had, only 3 had been illustrated in sufficient detail. The other character set chosen for phylo- genetic analysis was that of allozymes as viewed by horizontal starch-gel electro- phoresis. Regardless of whether allozymes are adaptively significant (e.g., Hochachka & Somero, 1984; Nevo & Bar, 1976; Nevo et al. 1981; Nevo et al., 1982) or adaptively neutral (e.g., Kimura, 1979, 1982), they offer a mor- phological data set virtually independent of penial morphology. Four species of eastern triodopsines had previously been electro- phoresed (McCracken & Brussard, 1980, as reevaluated by Emberton, McCracken & Wooden, in preparation). These were exam- ined at 8 variable loci that showed sufficient variation to bode success for applying electrophoresis to the systematics of the en- tire group. Certain alleles of some loci had also been shown to be genetically heritable (McCracken, 1976, 1980). The value of allozymes for systematic stud- ies is well established (e.g., Avise, 1975; Sarich, 1977; Throckmorton, 1978; Davis, 1978; Nei et al., 1983; Patton & Avise, 1983; Buth, 1984). Although land-snail allozymes have been used extensively for studies on population genetics and breeding systems (reviewed by Clarke, 1978; Selander & Och- man, 1983; Selander & Whittam, 1983), only twice before have they been used for exten- sive phylogenetic studies. In neither of these previous efforts—on West Indian Cerion by Gould et al. (1975), and on Moorean Partula 162 EMBERTON ratantnr m CLÉTILU genital pore lastral pustules) (and p pore à | п columns) Pupper penis | (pustulose) Den enis } (smooth, with random folds) a у pılaster pore FIG. 1. Penial morphology of east American triodopsines: its major features and the three alternative methods for studying the sculpture of its functional surface. a. The evertive method. b. The slide-mount method (clearing and staining). c. The dissective method. by Johnson et al. (1977)—was sufficient electrophoretic variation found to be of much value in reconstructing species-level phylo- genies. Both these groups, however, appear to be relatively recent radiations (Woodruff & Gould, 1978; Murray & Clarke, 1980), much younger than eastern triodopsines (see Emberton, 1986). Thus penial morphology and allozymes were chosen because of their independence from each other and because each promised to be rich in phylogenetically useful variation. To avoid circularity in evaluating conchologi- cal evolution, no shell characters were used for phylogenetic analysis. Time constraints prohibited the use of other morphological character sets that previous studies had indicated were less information- rich than penial morphology and allozymes. Concerning radulae, Solem's (1976) study of three triodopsines (two of them sympatric) had found an “essential similarity”, with, “in terms of basic structure and pattern of func- tioning, . no major differences between species, much less between genera”. Like- wise, Binney's (1878) sketches of the radulae of 9 other triodopsine species, although rather inadequate in detail, showed a similar lack of variation. The macro- and microstructure of the jaw promised little useful information, be- cause Solem (1976) found “no significant differences” among three species of eastern triodopsines. The size and shape of the hermaphroditic duct, talon, albumen gland, prostate, uterus, and spermatheca undergo such significant and extreme seasonal variation in one spe- cies of Triodopsis (Emberton, 1985) that the use of these characters for systematics would have had to have been cautious and labor- intensive. Likewise, considerable variations in the diameter and length of the ovotesticular lobes, the basal penis, the free oviduct, and the vagina correlate with changes in repro- ductive state (Emberton, 1985: figs. 5—7). Preliminary studies (Emberton, unpub- lished data) showed that the internal structure of the functional vagina (the spermathecal or gametolytic duct) was identical in several triodopsine species. This lack of variation extended to several pairs of microsympatric species with similar shell sizes and penial morphology. The structure of the triodopsine functional vagina has been illustrated by EASTERN NORTH AMERICAN TRIODOPSINAE 163 Binney (1851: pl. 7, fig. 4; pl. 8, fig. 3), Simpson (1901: pl. 8, fig. 11), and Grimm (1975: fig. 3A). Technology for viewing the chromosomal bands of land snails (e.g., Babrakzai & Miller, 1975, 1984) seemed in too early a stage of development for a project of this scope. Sim- ple chromosomal counts promised little in- sight, because an early study (Husted & Burch, 1947) of 17 species of polygyrids, including 6 triodopsines, found a diploid num- ber of 58 in all except what was identified as Triodopsis fraudulenta, populations of which were reported to vary in diploid number from 58 to 62. Furthermore, the phylogenetic inter- pretation of chromosomal numbers can be highly problematic (e.g., Solem, 1978). Thus this phylogenetic analysis of the east- ern American triodopsines was restricted to male-genitalic and allozymic characters. MATERIALS AND METHODS Taxa studied Neohelix von lhering, 1892 albolabris (Say, 1816) alleni (Sampson, 1883) dentifera (Binney, 1837) divesta (Gould, 1848) lioderma (Pilsbry, 1902) major (Binney, 1837) solemi Emberton, new species Triodopsis Rafinesque, 1819 alabamensis (Pilsbry, 1902) anteridon (Pilsbry, 1940) burchi Hubricht, 1950 claibornensis Lutz, 1950 complanata (Pilsbry, 1898) cragini Call, 1886 discoidea (Pilsbry, 1904) fallax (Say, 1825) fraudulenta (Pilsbry, 1894) fulciden Hubricht, 1952 henriettae (Mazyck, 1877) hopetonensis (Shuttleworth, 1852) Juxtidens (Pilsbry, 1894) messana Hubricht, 1952 neglecta (Pilsbry, 1899) obsoleta (Pilsbry, 1894) palustris Hubricht, 1958 pendula Hubricht, 1952 picea Hubricht, 1958 platysayoides (Brooks, 1933) rugosa Brooks 8 MacMillan, 1940 soelneri (Henderson, 1907) tennesseensis (Walker 8 Pilsbry, 1902) tridentata (Say, 1816) vannostrandi (Bland, 1875) vulgata (Pilsbry, 1940) vultuosa (Gould, 1848) Webbhelix Emberton, new genus multilineata (Say, 1821) Xolotrema Rafinesque, 1819 caroliniensis (Lea, 1834) denotata (Férussac, 1821) fosteri (F. C. Baker, 1932) obstricta (Say, 1821) occidentalis (Pilsbry & Ferriss, 1907) Collections Principal field work was conducted April- June 1982 in the eastern United States (“GS” series), and was supplemented by collections from southeastern Ohio in March-July 1979 (“Ohio” series), from the lower Ohio River Valley in April 1980 (“H” series), and from the southern Appalachian area in March-June 1983 (“SC” series). All collections were do- nated to the Field Museum of Natural History, Chicago. County-level localities, field num- bers, and catalog numbers of dissected and electrophoresed material are listed under each species in the systematic reviews in Appendices B and C. Detailed locality data are available from the author on request or from the Field Museum catalog. Snails in each lot were individually marked on their shells: 1, 2, 3, etc. for snails from which tissue samples were taken; and A, B, C, etc. for snails that were not tissue-sampled. Appen- dices B and C record which individual snails from each lot were dissected, electro- phoresed, and illustrated anatomically and/or conchologically. Additional anatomical material (total 18 lots) was borrowed from the Field Museum (FMNH), the Academy of Natural Sciences of Philadelphia (ANSP), and the private collec- tion of Mr. Leslie Hubricht. For the Neohelix albolabris and alleni groups, 41 populations were collected or bor- rowed, and 5 additional populations were studied from published anatomical illustra- tions. Dissections The uneverted penial tubes of 252 snails from 108 populations comprising all 40 of Hubricht's (1985) species were dissected. Most populations were collected in the early 164 EMBERTON spring, but to make a crude check for signifi- cant seasonal variation which might bias interspecific comparisons, three T. tridentata from Strouds Run State Park, Ohio, were dissected, each at a different stage in the life cycle of this species: mating-ready neoadult (FMNH 209209, specimen C); post-mating neoadult (FMNH 209536, specimen C); and overwintered, mating-ready, old adult (FMNH 209209, specimen D) (see Emberton, 1985). Whenever possible, at least three adults of each species were dissected. Because of the limitations of available material, however, 10 species were represented by only two dissec- tions each (X. obstricta, X. caroliniensis, T. picea, T. claibornensis, T. fraudulenta, T. rugosa, T. vultuosa, T. cragini, T. alaba- mensis, and T. neglecta), and 5 species were represented by only a single dissection each (X. occidentalis, T. henriettae, T. discoidea, T. fulciden, and T. pendula). The remaining 25 species were represented by three or more dissections each, usually with at least three from a single population. A representative dissection was illustrated for 39 of the 40 species, by means of a drawing tube attached to a Zeiss dissecting microscope. Relaxed specimens were used for 35 species, but because of limited material X. occidentalis, T. complanata, T. obsoleta, and T. fallax were represented by contracted specimens. T. rugosa became available too late to be illustrated. Comparative anatomies of eastern Ameri- can triodopsine outgroups were available in published illustrations. According to Ember- ton's (1986) phylogenetic analysis of the Polygyridae, eastern triodopsines are the most primitive group in the family, and their closest outgroups are the western American triodopsine Vespericola, and the ashmunel- lines Cryptomastix (western) and Allogona (western, with one eastern species: A. profunda). Penial-morphological data on these genera were available from Pilsbry (1940) and Webb (1948, 1968, 1970a, 1970b, 1970c). A more distant polygyrid outgroup of the triodopsines is the ashmunelline genus Ashmunella, the penial morphology of which was gotten from Pilsbry (1940) and Webb (1954). The closest non-polygyrid outgroups of triodopsines, according to the Emberton (1986) hypothesis, are the Corillidae, Am- monitellidae, and Oreohelicidae. In the Coril- lidae, only the external, uneverted penial mor- phology of one species is known (Solem, 1966); in the Ammonitellidae, limited details of the penial sculpture are known for Polygyrella, Polygyroidea, and Ammonitella (Pilsbry, 1939: figs. 369 #5a, 371 #5a, 373 #19); in the Oreohelicidae, penial sculpture 1$ known for a number of Oreohelix species (Pilsbry, 1939; Solem, 1978b). Another, more distant outgroup to the triodopsines which was considered were the Camaenidae, the penial anatomy of many species of which is known through the work of Wurtz (1955) and Solem (1979, 1981a, 1981b, 1984). See Til- lier (1986) for an alternative view on trio- dopsine outgroups. Additional methods were used for the study of the Neohelix albolabris and alleni groups. In order to quantify genitalic differences among taxa, 7 measurements were taken from one dissection per population for three populations each of N. alleni (pooling the two subspecies, which did not differ in the mea- surements taken—see Fig. 3), N. albolabris albolabris, N. albolabris bogani Emberton, new subspecies, N. major, and N. solemi. For these measurements, the most relaxed spec- imens were chosen from widely distributed populations. The measurements were: (1) the length of the penis, in mm, from its junction with the vagina to the internal apex of the dissection; (2) the number of pilastral lappets (this and other terminology is defined later) per 2.6 mm at the midpoint of the pilaster; (3) the number of columns of wall pustules per 1.3 mm, measuring transversely across the penial wall adjacent to the pilaster about two-thirds of its distance from the internal penial apex; (4) the length of the verge in mm; (5) the maximum width of the pilaster in mm; (6) the distance in mm from the external apex of the penis to the midpoint of the origin of the penial retractor muscle on the vas deferens; and (7) the length of the vas deferens, in mm, from where it bends at the external junction of the penis and vagina to its point of insertion at the external penial apex. Shell analysis Phylogenetic analysis to the species-group level was entirely free of consideration of shell morphology. In the systematic reviews, how- ever (Appendices B and C), comparative conchological descriptions are included to allow identification to species group from shells alone. To aid identification, a represen- tative shell for each of 39 species (all but 7. rugosa) was illustrated in two views: perpen- dicular to the plane of the aperture, and in the EASTERN NORTH AMERICAN TRIODOPSINAE 165 plane of the aperture while parallel to the axis of rotation. These views were chosen be- cause they simultaneousiy show as many important shell features as possible, including apertual dentition, apertural dishing, apertural lip thickness, pre-apertural deflection of the body whorl, umbilicus, height, surface striae, and, in a rough way, whorl count. The shell drawings were made using a drawing tube mounted on a Zeiss dissecting microscope. For most species, the illustrated shell was from the same population from which the penial morphology was illustrated. Shells of the Neohelix albolabris and alleni groups were studied in much greater detail. Despite a great similarity in the overall aspect of the shells, and an overlap in shell size among the 6 species and subspecies of this group, subtle conchological differences were apparent. In order to quantify these differ- ences and to objectively test their reliability for identifying the taxa, a multivariate discrimin- ant analysis was performed, beginning with a set of 11 measurements on 55 shells from 28 populations. These populations, their species or subspecies, and the identification numbers of the shells measured from each, are listed in the first three columns of Table 6. For each of the six taxa, a set of populations was chosen which appeared to include its full range of shell variation; from each population, all un- damaged adult shells were measured if there were no more than three—if there were more than that, the three shells showing extremes in the population's variation were chosen for measurement. There were 8 shell variables in which the 6 species and subspecies of the N. albolabris and N. alleni groups appeared to differ: rela- tive spire height (henceforth called REL- SPIRE), whorl expansion rate (WHRLEXPN), relative width of the apertural lip (RELLIP), relative size of the baso-columellar lip node (RELNODE), relative degree of pre-apertural deflection of the body whorl (RELDEFL), den- sity of surface striae (STRIAE), color (BROWN), and sheen (GLOSSY). These variables and the method for quantifying each are listed in Table 5. STRIAE was a direct count, BROWN and GLOSSY were rank measurements, and the remaining 5 (REL- SPIRE, WHRLEXPN, RELNODE, RELLIP, AND RELDIFL) were ratios of directly mea- sured or calculated distances. The 11 mea- surements from which the 8 variables were derived are listed as column headings in Table 6. Electrophoresis Posterior foot tissues (“snail tails”) were excised from field-activated snails and stored in eryogenic vials in liquid nitrogen. Horizontal starch-gel electrophoresis followed methods of Selander et al. (1971) and Shaw & Prasad (1970), as modified by Davis et al. (1981). Twelve enzyme systems yielding 16 loci were used: Sordh, Mdh-1 & 2, Me, Icd, Pgd, Gd-1 & 2, Sod-1 & 2, Got-1 & 2, Pgm, Lap, Mpi, and Gpi (see Appendix A). These loci were chosen because they were genetically inter- pretable, because they represent a diversity of metabolic pathways, and because several of them have a proven heritability (Mc- Cracken, 1976, 1980). Deliberately excluded were enzymes that have been shown to be environmentally induced in pulmonates: ester- ases (Oxford, 1973, 1978), lactate dehydro- genase (Gill, 1978a), acid phosphatase, and alpha-glycerophosphate dehydrogenase (Gill, 1978b). Complete electrophoretic proce- dures are given in Appendix A. The electrophoresed material comprised 249 snails from 64 populations representing 35 of the 40 Hubrichtian (1985) species of eastern triodopsines. The 5 species for which tissue samples were lacking were T. discoidea, T. fallax, T. obsoleta, T. rugosa, and T. soelneri. Three electrophoresed spe- cies had incomplete data: X. fosteri (missing Gd-1, Gd-2, and Sod-2), T. fulciden, and T. henriettae (both missing Gd-1 and Gd-2). All other species (32 total) were represented by at least one population with complete data for all 16 loci. Seventeen species were represented by a single electrophoresed population each (T. albamensis, T. burchi, X. caroliniensis, T. claibornensis, T. complanata, X. fosteri, T. fraudulenta, T. fulciden, T. henriettae, N. lioderma, T. messana, T. neglecta, T. pendula, T. picea, T. platysayoides, N. solemi and T. vannostrandi); 13 species were represented by two populations each (T. anteridon, T. cragini, X. denotata, N. dentifera, N. divesta, T. hopetonensis, T. juxtidens, W. multilineata, X. occidentalis, T. palustris, T. tennesseensis, and T. vulgata; four species were represented by three populations each (N. albolabris, N. major, X. obstricta, and T. vultuosa); one spe- cies was represented by 5 populations (N. al- leni); and one species was represented by 6 populations (T. tridentata). Catalogue num- bers of the voucher specimens for all electrophoresed populations are given in col- 166 EMBERTON umn 2 of Table 2, and in Appendices BandC. Of the total 64 populations, 50 had complete electrophoretic data and 14 had missing data for one to 7 loci. The number of snails electrophoresed per population (Table 2, column 3) ranged from one to 12, with a mean of 3.9 and a standard deviation of 2.5. The closest outgroup of eastern trio- dopsines from which comparative material was available was Allogona profunda (Say, 1821), of which two populations with sample sizes of 2 and 10 were electrophoresed. Data analysis Penial morphology was analyzed cladisti- cally (Hennig, 1966; Eldredge & Cracraft, 1980; Wiley, 1981). А character-state phylogeny was proposed for each character, using criteria reviewed by Emberton (1986), and its polarity was determined by outgroup comparison (e.g., Watrous & Wheeler, 1981). A taxon-by-character-state matrix was pre- pared using additive binary coding (Farris et al., 1970). Cladograms were generated from this matrix using the Wagner criterion of un- restricted parsimony (Kluge & Farris, 1969; Farris, 1970), using global branch swapping to approach heuristically the most parsimoni- ous set of trees. The PAUP program (Swof- ford, 1983) was used for computing the trees. These trees were visually compared branch- by-branch, and each discrepancy was re- solved based on which combination of convergences and reversals seemed biologi- cally most plausible. The final result of these comparisons was a single, most parsimoni- ous cladogram that was designated the “Anatomy Tree”. Electrophoretic data were analyzed both cladistically and phenetically. Cladistic analy- sis employed the independent alleles model (Mickevich & Johnson, 1976), by which al- leles not present in the outgroup are consid- ered apomorphous. Mesodon was used as the outgroup, because it was the only other polygyrid group for which a comparable electrophoretic data set was available (Emberton, 1986). For each eastern-Amer- ican triodopsine species, the presence or absence of each apomorphous allele was binary-coded. The resulting data matrix was analyzed by PAUP (Swofford, 1983), using global branch swapping to obtain the first 50 trees with equal, maximum parsimony. These trees were then compared branch-by-branch to determine the most frequently occurring configuration of each branch. In this manner, a single maximum-parsimony, consensus cladogram was arrived at, and was desig- nated the “Alleles Tree.” For phenetic treatment, the electrophoretic data set was divided into two subsets, the first consisting of 32 species plus the outgroup (Allogona profunda), each represented by a single population with complete data for all 16 loci. The second subset consisted of three species not included in the first subset, plus additional populations of 18 species in the first subset, plus two outgroups (A. profunda and Mesodon zaletus [Binney, 1837]), for a total of 33 populations. In this second subset, all loci with incomplete data were deleted, leaving 8 loci: Sordh, Mdh-1, Mdh-2, Pod, Sod-1, Got-1, Pgm, and Gpi. For each of the two subsets, Prevosti distances (Wright, 1978) among populations were calculated and subjected to the distance-Wagner proce- dure (Farris, 1970), with branch-length opti- mization, using NT-SYS computer programs (Rohlf et al., 1972). The resulting trees were designated the “Wagner-1 Tree” and the “Wagner-2 Tree.” The Anatomy, Alleles, Wagner-1, and Wagner-2 Trees were combined to produce a Consensus Tree in the following manner. Each of the four trees was weighted by a combination of two criteria: the number of data units and the relative reliability of the data units. The data units were consid- ered to be character-state transformations in the Anatomy and the Alleles Trees, and to be alleles in the Wagner-1 and Wagner-2 Trees. The reliability of anatomical data-units relative to allozymic data-units was estimated by di- viding the number of convergences and re- versals in the Anatomy Tree by the number of convergences and reversals in the Alleles Tree. Multiplying this reliability index times the number of anatomical character-state transformations gave a relative weight for the Anatomy Tree. The relative weight of the Alleles Tree was taken as the number of single-allelic transformations. Relative weights of the Wagner-1 and Wagner-2 Trees were considered to be to the number of alleles comprising the data subset from which each tree was calculated. Using these weight- ings to resolve conflicts, the four trees were visually compared branch-by-branch to arrive at a Consensus Tree. For a more detailed cladistic analysis of the Neohelix albolabris and alleni groups, addi- EASTERN NORTH AMERICAN TRIODOPSINAE 167 tional anatomical character-state transforma- tions were proposed based on the quantita- tive comparisons in penial morphology. Allthe available transformations were then used to construct a maximum-parsimony cladogram by hand. Multivariate discriminant analysis of shells of the Neohelix albolabris and alleni groups employed SAS software (SAS Institute, 1982). The 6 taxa were discriminated on the basis of 8 shell variables (Table 5). Eight of the 55 measured shells had an incompletely matured apertural lip (Table 6, last column), which affected the values of RELNODE and RELLIP, therefore these shells were deleted from the analysis. Patterns of genitalic evolution Patterns of evolution in penial morphology were analyzed by comparing sister taxa (spe- cies or species clusters appearing dichoto- mously in the Consensus Tree). For each of 25 sister taxa, the difference in penial mor- phology was ranked as great, moderate, slight, or none; and the geographical relation- ship of their ranges was classified as al- lopatric, sympatric, parapatric, or peripatric (in which one taxon is a small-ranged endemic peripheral to the much broader range of the other). Geographic ranges were gotten from Hubricht (1985). The importance of reproductive character displacement was assessed by comparing, for each of 12 species, populations sympatric vs. allopatric with another triodopsine species of similar shell size and shape. Table 9 lists the species, the sympatric species, the local- ities of compared populations, and the num- ber of dissections per population. Allopatric populations of T. tridentata were compared with populations sympatric with 7. vulgata, X. obstricta, T. picea, and T. juxtidens; likewise T. vulgata was tested for penial differences due to sympatry with X. denotata, T. tennes- seensis, and T. tridentata. Also tested were N. albolabris against N. alleni and N. denti- Гега; T. juxtidens against T. tridentata; and both N. alleni and N. dentifera were tested against N. albolabris. Patterns of shell evolution To analyze conchological evolution at the generic and species-group levels, a represen- tative shell was chosen for each species and was mounted in its proper position on the Consensus Tree. Patterns of change through time were interpreted under the assumptions that (1) the Consensus Tree was an accurate estimate of true phylogeny, and (2) the shell morphology of each (unknown) ancestor was between the morphologies of its extant de- scendents. Patterns at the species level were as- sessed using Vagvolgyi's (1968) conchologi- cal monograph, in which the ranges of basic shell measurements and ratios are listed within each species description. Vagvolgyi's total data base comprises 31,269 shells from 556 museum lots. For the present analysis, his data were compiled into tabular form, and a “diameter range” index (Solem, 1981a) was calculated for each species: the greatest mi- nus the least measured shell diameter, di- vided by the least, and expressed as a per- cent. TAXONOMIC HISTORY Triodopsis and Xolotrema were erected by Rafinesque in 1819 to separate tridentata, denotata, and other tridentate North Ameri- can species from Helix, then a massive genus comprising most of the world's land snails. The new generic names were largely ignored, however (all polygyrids going by the name Polygyra Say, 1818), until “Tryon (1867), Binney & Bland (1869), and later authors, following von Martens (1860) used Triodopsis for all of the depressed, two- or three-toothed [land shells] of the eastern United States, and Mesodon for the more capacious, subglobose species with a small parietal tooth, or tooth- less [and thus abandoned the name Xolotrema]” (Pilsbry, 1940). Many authors (e.g., Simpson, 1901; F. C. Baker, 1939) continued, however, to synonyomize Trio- dopsis and Mesodon under the blanket genus Polygyra. lt wasn’t until Pilsbry’s (1940) monograph on North American land snails that Triodopsis was clearly characterized an- atomically, was distinguished anatomically from Mesodon, and was recognized as cov- ering most of the wide range of shell shapes also covered by Mesodon but formerly erro- neously divided between the two genera. In this monograph, Pilsbry (1940) divided Triodopsis into the four subgenera Triodopsis s. str. Rafinesque, 1819; Cryptomastix Pils- bry, 1939; Xolotrema Rafinesque, 1819 (bringing this name back from obscurity); and 168 EMBERTON Neohelix von lhering, 1892, based on shell shape and reproductive anatomy, with Crypto- mastix disjunct in the Pacific Northwest. Pilsbry's taxonomy of eastern Triodopsis (i.e., the eastern triodopsines) was almost exclu- sively based on shell morphology, despite the fact that he illustrated the reproductive sys- tems of several species. He recommended that future revisions make use of penial mor- phology. Additional species and subspecies of Triodopsis were subsequently described by Lutz (1950) and Hubricht (1950, 1952, 1958). A summary of new and emended taxa from 1948 to 1984 was provided by Miller et al. (1984). Webb (1947a, 1947b, 1948, 1952, 1954, 1959) published a series of reports on the reproductive behavior and anatomy of se- lected species of triodopsines, and pointed out—as Pilsbry had predicted—important variation in penial sculpture, upon which he based several taxonomic changes. In his 1952 paper, Webb elevated Xolotrema to a full genus (defined as possessing a penial verge) and transferred the subgenus Neohelix to it. In 1954, Webb elevated the Pacific Northwestern subgenus Crypoto- mastix to generic level within the new subfam- ily Ashmunellinae, thereby restricting Trio- dopsis to eastern North America. Also based on penial morphology, Webb erected the subgenus Wilcoxorbis for Xolotrema fosteri (Webb, 1952), the subgenus Haroldorbis for Triodopsis cragini and the section Shelfordorbis for Triodopsis vulgata (Webb, 1959). Vagvolgyi’s (1968) monograph, “Systemat- ics and Evolution of the Genus Triodopsis (Mollusca: Pulmonata: Polygyridae)”, sum- marized a massive amount of new data on conchological variation. This revision was based solely on shells and ignored Webb's (1947-1961) anatomical work and validly pro- posed supraspecific taxa (see Grimm, 1975; Solem, 1976). Additional shortcomings of this work were that (1) the numerical formulae used for separating and defining taxa were arbitrary, based on neither multivariate nor any other objective criterion; (2) designation of “hybrids” was based on the untested crite- rion of high within-populational variation, the presence of which can have other explana- tions; and (3) the ecological descriptions were often arrived at by comparing species ranges with broad-scale vegetation maps, thereby sometimes missing important finer-grained ecological differences (L. Hubricht, personal communication; personal observations). Vagvolgyis Triodopsis copei (Wetherby) was subsequently split into the three species cragini, vultuosa, and henriettae by Cheatum & Fullington (1971) in their monograph of Texas polygyrids. Grimm (1975) gave brief comparative de- scriptions of Triodopsis and its species groups, and summarized his systematic con- clusions concerning the Т. fallax group based on a ten-year study of geographic shell vari- ation, laboratory hybridization, and, to lesser extent, penial morphology. Grimm's conclu- sions based on these (largely undocumented) studies were concordant with those earlier postulated from geographic shell variation by Hubricht (1953), but ran counter to those of Vagvolgyi (1968). Solem’s (1976) “Comments on Eastern North American Polygyridae” compared the sympatric, conchologically similar Neohelix divesta and N. albolabris with each other and with three sympatric, conchologically similar species of Mesodon in shell, radular struc- ture, jaw structure, external aspect of the reproductive system, and dissected penial morphology. Comparative data on the rare Triodopsis platysayoides were also included. Solem emphasized the need for sympatric- species comparisons to establish criteria for distinguishing allopatric species, and showed through adequate illustrations that penial morphology was an even richer source of systematically useful characters than Webb's illustrations had indicated. McCracken & Brussard’s (1980) study of electrophoretic variation among populations of the “white-lipped land snail’, although in- correct in many of its conclusions due to taxonomic errors (Emberton, McCracken, & Wooden, in preparation), showed the pres- ence of significant allozymic variation within Neohelix and demonstrated the heritability of several loci in a New York population of N. albolabris. Emberton (1985) dissected a temporal se- ries of Triodopsis tridentata and found that extreme seasonal variation ruled out repro- ductive-organ volume and, to some extent, organ length as useful systematic characters for triodopsines. Hubricht’s (1985) book of range maps and ecological sketches of eastern North Ameri- can land snails discarded most of Vagvolgyi’s (1968) species-level taxonomic changes and elevated several of Pilsbry’s (1940) subspe- EASTERN NORTH AMERICAN TRIODOPSINAE 169 cies and the one subspecies of Lutz (1950) to full species, resulting in 40 total species; taxonomy between the genus and species levels was not included. Richardson's (1986) bibliographic catalog of polygyrid species did not incoprorate Hubricht's (1985) changes. GENITALIC ANALYSIS Variation The eastern-triodopsine penis and its major structural features are presented diagram- matically in Fig. 1. The uneverted penis is held internally by a single retractor muscle attached to the vas deferens near the penial apex. The penial tube varies from short to extremely long, from cylindrical to clubbed. The position of entry of the vas deferens at the ejaculatory pore varies from terminal to subterminal. The collar-like muscular sheath, which circularly attaches to the basal penis and connects to the vas deferens via the retentor muscle, varies from short (covering only the basal fourth of the penis) to very long (covering the entire penis). Dissection of the uneverted penial tube reveals its ornately sculpted functional sur- face (Fig. 1). The dorsal pilaster is a longitu- dinal outgrowth of the penial wall; it varies among species in both length and surface sculpture. The penial wall (exclusive of the dorsal pilaster) is covered with rows of pus- tules. These pustules vary in size and shape among species, and are sometimes lacking. The pustular rows vary in pattern; when their pustules are absent they appear as low, smooth ridges. The area surrounding the ejaculatory pore may be flat or may be elon- gated as a flap-like, conical verge of variable size and shape. Other features which also may be present (but are not shown in Fig. 1) are a smooth ventral sperm groove; a fleshy, knob-like peduncle beneath the ejaculatory pore; and a low ventral pilaster. Proximal to the upper, sculpted region of the penis lies the basal penis. This region is smooth, lacking pustules. Its walls vary from thin with random folds, to thick and muscular with regular folds produced by both longitudi- nal and circular muscle bands. Proximal to the basal penis, between the vaginal opening and the genital pore, is the atrium. The wall of this region is always smooth and thin, bearing random folds. Variation within any given population was minor in sculptural details, but major in such elastic features as penis length, sheath length, retractor-muscle and retentor-muscle lengths, and the configuration of folds in the basal penis. Much of this variation seemed to correlate with the contractile condition of the specimen. Seasonal variation in penial morphology in the studied population of Triodopsis tridentata was slight. The post-mating snail had a thin- ner wall, and its pustules were somewhat thin and flap-like compared to the more promi- nent, robust pustules of both mating-ready snails. The distribution and relative sizes of the pustules, however, remained constant. In all 13 species for which more than one population was dissected, upper penial sculp- ture was remarkably uniform. An example of this geographic stability is illustrated in Fig. 3, which shows the penial morphologies of two populations of Neohelix alleni separated by the Mississippi River Valley. Judging both from the wide range disjunction of this spe- cies (Fig. 49) and from the fossil-palynological evidence concerning its deciduous-forest habitat (Delcourt & Delcourt, 1981), these two populations had been genetically isolated for at least 20,000 years, which is probably equivalent to at least half as many genera- tions (see McCracken, 1976). Nevertheless, these populations had accumulated only mi- nor differences in penial sculpture: the east- ern population (N. alleni fuscolabris) differed from the western (N. alleni alleni) only in its somewhat larger verge and in having its pustulose region descend approximately 20% lower. Because of this general morphological con- servatism, the penial morphology of each species could be adequately represented by a single illustration (Figs. 2-18). The only spe- cies not illustrated (Triodopsis rugosa) was very similar to T. fulciden (Fig. 18b). Descriptions Measurements in the following descriptions were taken solely from the illustrations (Figs. 2-18) and do not in any way reflect natural variation. Penis length was measured from the apex to the genital pore. The verge was measured from its dorsal side. The terms “large”, “small”, etc. are used relative to total penis length. Neohelix albolabris (Say, 1816)—Dissec- tions: 27 from 14 populations. Fig. 29-9. Length 17 mm. Shape cylindrical, the apical 170 EMBERTON ÓN 2 mm in FIG. 2. Opened uneverted penial tubes. a. Neohelix dentifera (Binney, 1837). FMNH 214810 #8 (also dissected #1, 4; FMNH 214809 [sympatric with Neohelix albolabris] #2, 7, 14). b. Diagrammatic detail of 3 lappets from center of pilaster of a, showing substructure of pustules. c. Diagrammatic detail of central wall pustules of a, showing lateral cusps. d. Neohelix albolabris (Say, 1816). FMNH 214920 (sympatric with Neohelix dentifera) #14 (also dissected #9, 11, 17; and 8 other populations ——see Appendix В). e. Detail of 3 lappets from center of pilaster of d, showing substructure of apparently fused pustules. f. Detail of other side of verge in d, showing opening of vas deferens. g. Detail of wall pustules of d. half enlarged. Ejaculatory pore terminal, on a verge. Verge large (length 1.5 mm), terminal, dorso-laterally compressed, back-pointing, with a ventrally subterminal pore and sculpted with surface cords continuing into about 6 terminal papillae (Fig. 2f). Dorsal pilaster long (7 mm) and broad (mid-width 1.3 mm), and superficially resembling a stack of tongue-like lappets with edges slightly convex and regu- larly marked with slight indentations (Fig. 2e). Basal half of the penis smooth with random folds; upper half uniformly sculpted with 25-35 adjacent, generally unmerging, equi- lateral columns of distinct, equal-sized pus- tules (Fig. 2g), radiating from the pore region. Sheath enclosing less than half of the upper, sculpted region of the penis. Neohelix alleni (Sampson, 1883)—Dis- sections: 15 from 8 populations. Fig. 3. Length 20 mm. Shape cylindrical, the apical half enlarged. Ejaculatory pore terminal, on a verge. Verge relatively large (1.0 mm), terminal, dorso-laterally compressed, back- pointing, with a ventrally subterminal pore and sculpted with surface cords continuing into about 8 terminal рарШае (Fig. 3b). Dorsal pilaster long (11 mm) and broad (mid-width 1.4 mm), and superficially resem- bling a stack of tightly appressed tongue-like lappets with smooth edges. Basal one-fourth to one-third of the penis smooth with random folds; upper half uniformly sculpted with 25-35 adjacent, generally unmerging, equi- lateral columns of distinct, equal-sized pus- TA EASTERN NORTH AMERICAN TRIODOPSINAE N, A = QI nn EST ee JN UNE RISAS Ze >< prio SSS AA о A Q OS OR м $2 L> ЧАРЫ pty, OO a HENS co — feo 252 о - d9 a oo = 09 м OG HUE Ze RON го ое 9.288 > © о ОЕ oe А (©) SF DE These two subspecies have probably been separated by the Mississippi pened uneverted penial tubes. a. Neohelix alleni alleni Valley for at least 20,000 years but show little difference in penia dissected #11, 13; and 7 other populations—see Appendix B opening of vas deferens. c. Neohelix alleni fuscolabris (Pilsbry, 1903 FMNH uncat. #7, 11, 15). FIG. 3. О EMBERTON 172 Y Pt LH Г] ( SIR N FIG. 4. Opened uneverted penial tube. a. Neohelix major (BInney, 1837). ЕММН 214930 #6 (also dissected see systematic section). b. Detail of the reverse side of verge of a, showing #7, 8; and other populations opening of vas deferens. EASTERN NORTH AMERICAN TRIODOPSINAE a LS 9, eS > =, > os Ve LA Sees SH ES à SEE $ RA, LS PRA BERT qa 2 173 re ee С e SA KEELE MITEL UE ER 2 4 CE 7 “as Pio a ER cr Sse BEX AS OS E Ca EX (2 T2 as Е FIG. 5. Opened uneverted репа! tubes. a. Neohelix lioderma (Pilsbry, 1902). FMNH 214844 #A (also dissected #9 and #B, С). b. Reverse of verge of a, showing opening of vas deferens. с. Detail of 3 lappets from center of pilaster of a, showing substructure suggesting laterally fused pustule. d. Neohelix divesta (Gould, 1848). FMNH 214813 #1 (also dissected #7, 8, 10; FMNH 176089). e. Reverse of verge of d, showing opening of vas deferens. f. Detail of 3 central lappets of pilaster of d, showing substructure suggesting laterally fused pustules. tules radiating from the pore region. Sheath enclosing less than half of the upper, sculpted region of the penis. Neohelix dentifera (Binney, 1837)—Dis- sections: 6 from 2 populations. Fig. 2a-c. Length 12 mm. Shape cylindrical, the apical half enlarged. Ejaculatory pore terminal, on a verge. Verge moderate in size (length 0.9 mm), terminal, dorso-laterally compressed, back-pointing, with a ventrally subterminal pore and sculpted with surface cords continu- ing into about 6 terminal papillae. Dorsal pilaster long (7 mm) and broad (mid-width mm), and superficially resembling a stack of thin, plate-like lappets with edges comprised 174 EMBERTON mn Ss = iR = ZN = Bai Re = es. FIG. 6. Opened uneverted penial tubes. a. Webbhelix multilineata (Say, 1821). FMNH 214848 #2 (also dissected FMNH 214849 #1, 5, A; Hubricht 48600 #A, B, C). b. Neohelix solemi Emberton, new species. FMNH 214936 #1 (also dissected 13 other populations—see Appendix B). of equal, bi-lobed units (Fig. 2b). Ваза! one- jacent, generally unmerging, equilateral col- fourth of the penis smooth with random folds; umns of distinct, approximately equal-sized, upper half uniformly sculpted with 25-35 ad- lobed pustules (Fig. 2c), radiating from the EASTERN NORTH AMERICAN TRIODOPSINAE 175 2 mm 4 у or LA FIG. 7. Opened uneverted penial tubes. a. Xolotrema denotata (Férussac, 1821). FMNH 214806 #6 (also dissected #1, 2). b. Reverse of verge of a. c. Xolotrema obstricta (Say, 1821). FMNH 214854 #9 (also dissected #1). d. Reverse of verge of c. e. Xolotrema caroliniensis (Lea, 1834). FMNH 171142 #A (also dissected #B). pore region. Sheath enclosing one half to two-thirds of the upper, sculpted region of the penis. Neohelix divesta (Gould, 1848)—Dissec- tions: 4 from 1 population. Fig. 5d-f. Length 10 mm. Shape cylindrical, the apical half en- larged. Ejaculatory pore terminal, on a verge. Verge moderate in size (length 0.6 mm), terminal, dorso-laterally compressed, back- pointing, with a ventrally subterminal pore and sculpted with surface cords continuing into about 6 terminal papillae (Fig. 5e). Dorsal pilaster long (7 mm) and broad (mid-width 0.8 mm), and superficially resembling a stack of thin, plate-like lappets with regularly indented edges (Fig. 5f). Basal one-fourth of the penis smooth with random folds; upper half uni- formly sculpted with 25-35 adjacent, gener- ally unmerging, equilateral columns of dis- tinct, approximately equal-sized, lobed pustules, radiating from the pore region. Sheath enclosing less than half of the upper, sculpted region of the penis. Neohelix lioderma (Pilsbry, 1902)—Dis- sections: 4 Нот 1 population. Fig. 5a—c. Length 7 mm. Shape cylindrical, the apical half enlarged. Ejaculatory pore terminal, ona verge. Verge (shown partially inverted in Fig. 5a-b) moderate in size (0.3 mm), terminal, dorso-laterally compressed, back-pointing, with a ventrally subterminal pore and sculpted with surface cords continuing into about 6 terminal papillae (not visible in Fig. 5b). Dor- sal pilaster long (6 mm) and broad (mid-width 0.5 mm), and superficially resembling a stack of thin, plate-like lappets with regularly in- dented edges (Fig. 5c). Basal one-fourth of the penis smooth with random folds; upper half uniformly sculpted with 25-35 adjacent, generally unmerging, equilateral columns of EMBERTON 176 == un. = nn A a ыы u (F. C. Baker, 1932). FMNH 214817 #A (also FIG. 8. Opened uneverted penial tubes. a. Xolotrema fosteri (Pilsbry & Ferriss, 1907). FMNH dissected #B, C, D, E; FMNH 214819 #19). b. Xolotrema occidentalis 214856 #5. c. Reverse side of verge of b, showing opening of vas deferens. Neohelix major (Binney, 1837)—Dissec- tions: 10 from 5 populations. Fig. 4. Length: 23 mm. Shape cylindrical, the apical half enlarged. Ejaculatory pore terminal, on a Dec. = ® ® 352 © a D 3 = Ф TD Е a 285 ar — Oo © He Cie ge EN Ф вс Eco оо Le Ф ME оО = Less x += Е REES avoe оное © eo .o®o 59 =Ф 82а = ON À = Ys O D OU ws EASTERN NORTH AMERICAN TRIODOPSINAE 2 > =o Me В AL iff = 177. X = ‘ère II SICH Y AE > ARA 5 AD) ap WA 2 ASIEN ur 3 a CAR р ое: Ra "OIGO Ka ES, d'a Où) AA 3 } S XA ASS FIG. 9. Opened uneverted penial tubes. a. Triodopsis vulgata Pilsbry, 1940. FMNH 214884 #1 (also dissected FMNH 214883 #2, 3; FMNH 214885 #1, 2, 3, 4). b. Triodopsis picea Hubricht, 1958. FMNH 214860 #14 (also dissected #4). с. Triodopsis claibornensis Lutz, 1950. FMNH 214800 #18 (also dissected #5: has broader pilaster). verge. Verge very large (length 3.0 mm), terminal, dorso-laterally compressed, back- pointing, with a ventrally subterminal pore and scuipted with surface cords continuing into about 12 terminal papillae (Fig. 2b). Dorsal pilaster long (12 mm) and broad (mid-width 2.2 mm), and superficially resembling a stack of tongue-like lappets with edges pro- nouncedly convex and irregularly wavy, with no regularly-spaced indentations. Basal half of the penis smooth with random folds; upper half uniformly sculpted with 25-35 ad- jacent, generally unmerging, equilateral col- umns of distinct, equal-sized pustules radiat- ing from the pore region. Sheath enclosing less than half of the upper, sculpted region of the penis. Neohelix solemi Emberton, new species— Dissections: 24 from 13 populations. Fig. 6b. Length: 17 mm. Shape cylindrical, the apical half enlarged. Ejaculatory pore dorsally subterminal, on a tiny verge which lies on the apex of a thick, fleshy protuberance. Verge minute (length 0.1 mm), dorso-laterally com- pressed, backpointing, with a ventrally sub- terminal pore and sculpted with surface cords continuing into 6-8 terminal papillae. Dorsal pilaster relatively short (3 mm) and narrow (mid-width 0.5 mm), merging terminally with 178 EMBERTON FIG. 10. Opened uneverted penial tube. Triodopsis fraudulenta (Pilsbry, 1894). FMNH 214822 #6 (also dissected #8: has more and smaller parts in thick- est part of pilaster, with wall pustules more pro- nounced). the fleshy protuberance, and superficially re- sembling an indistinct stack of indistinct tongue-like lappets with variousiy shaped edges. Ventral wall bearing one to three fold- like pilasters, sculptured no differently than the adjacent penial wall. Basal three-fifths of the penis smooth with random folds; upper two-fifths uniformly sculpted with 25-35 adja- cent, generally unmerging, equilateral col- umns of distinct, equal-sized pustules radiat- ing from the pore region. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis alabamensis (Pilsbry, 1902)— Dissections: 2 from 1 population. Fig. 16a. Length: 7 mm. Shape like amace. Ejaculatory роге ventrally subterminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (1 mm) and tapered proximally (mid-width 0.4 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis anteridon (Pilsbry, 1940)— Dissections: 3 from 2 populations. Fig. 14b. Length: 7 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (1.5 mm) and tapered proximally (mid-width 0.6 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis burchi Hubricht, 1950—Dis- sections: 3 from 1 population. Fig. 11a. Length: 8 mm. Shape like a baseball bat. Ejaculatory pore terminal. Verge absent. Dor- sal pilaster long (3 mm) and broad (mid-width 1.4 mm), consisting of abutting, unequaly- sized polygons, each covered with knob-like pustules about twice as large as the wall- pustules. Basal half of the penis smooth with random folds and slight circular corrugations; upper half sculpted with 15-20 columns of EASTERN NORTH AMERICAN TRIODOPSINAE 179 2 mm FIG. 11. Opened uneverted penial tubes. a. Triodopsis burchi Hubricht, 1950. FMNH 214797 #3 (also dissected #5, 12). b.-c. Triodopsis tennesseensis (Walker & Pilsbry, 1902). b. FMNH 214864 #15 (also dissected #13, 14). c. Area around opening of vas deferens in #14, showing lack of verge. d. Triodopsis complanata (Pilsbry, 1898). Hubricht 17932 #C (also dissected #A, B). equal-sized pustules radiating directly from the pore, the ventral-most columns with pus- tules indistinct, appearing almost smooth. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis claibornensis Lutz, 1950—Dis- sected 2 from 1 population. Fig. 9c. Length: 8 mm. Shape like a baseball bat. Ejaculatory pore ventrally subterminal, about one-fifth- way from the penial apex in the upper, sculpted region. Verge absent. Dorsal pilaster long (2.3 mm) and broad (mid-width 0.8 mm), covered with knob-like pustules all about twice as large as the wall-pustules. Basal half of the penis smooth with random folds and slight circular corrugations; upper half sculpted with 15-20 columns of distinct, equal-sized pustules radiating directly from the pore. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis complanata (Pilsbry, 1898)— Dissections: 3 from 1 population. Fig. 11d (a contracted specimen). Length: 5 mm. Shape like a baseball bat. Ejaculatory pore terminal. Verge absent. Dorsal pilaster short (1 mm) and broad (mid-width 0.8 mm), consisting of a solid mass bearing three tiers of long, sharp spurs. Basal third of the penis smooth with random folds and slight circular corrupations; upper two-thirds sculpted with 15-20 columns radiating directly from the роге, the dorsal columns bearing indistinct, equal-sized pus- tules, and the ventral columns completely smooth. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis cragini Call, 1886—Dissec- tions: 2 from 1 population. Fig. 13b. Length: 8 mm. Shape like a needle. Ejaculatory роге terminal. Verge absent. Dorsal pilaster two- thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid- width 0.3 mm), consisting of abutting irregu- larly sized and shaped polygons, each bear- ing one to three short, blunt spurs. Basal third of the penis smooth with random folds; middle third with slight circular corrugations; upper third sculpted with equilateral, widely sepa- 180 EMBERTON 69°, ar Re > LITT MI he 0 qu ee, %— ze FIG. 12. Opened uneverted penial tube. Triodopsis platysayoides (Brooks, 1933). FMNH 214861 #1 (also dissected #2; examined Hubricht 11860 [il- lustrated in Solem, 1976)). rated columns of equal-sized pustules, merg- ing ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal third of the penis. Triodopsis discoidea (Pilsbry, 1904)— Dissections: 1 from 1 population. Fig. 14d. Length: 6 mm. Shape like amace. Ejaculatory pore ventrally subterminal, two-fifths-way from the apex in the upper, sculpted region. Verge absent. A large, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.5 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculp- ted with equilateral, widely separated col- umns of equal-sized pustules, merging ven- trally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis fallax (Say, 1825)—Dissec- tions: 3 from 1 population. Fig. 14b (a con- tracted specimen). Length: 7 mm. Shape like a mace. Ejaculatory pore ventrally sub- terminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dorsal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.7 mm), consisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely sepa- rated columns of equal-sized pustules, merg- ing ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis fraudulenta (Pilsbry, 1894)— Dissected 2 from 1 population. Fig. 10. Length: 6 mm. Shape like a baseball bat. Ejaculatory pore ventrally subterminal, about one-fifth-way from the penial apex in the upper, sculpted region. Verge absent. Dorsal pilaster long (3 mm) and broad (mid-width 0.8 mm), consisting of nesting horeshoe-shaped units covered with knob-like pustules about twice as large as the wall-pustules. Basal third of the penis smooth with random folds and slight circular corrugations; upper two- thirds sculpted with 15-20 columns radiating directly from the pore, the dorsal columns bearing distinct, equal-sized pustules, and the ventral columns smooth, the ventral-most merging basally into a complexly ridged pro- tuberance. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis fulciden Hubricht, 1952—Dis- EASTERN NORTH AMERICAN TRIODOPSINAE 181 EU MAA E ZU [Ze oe aT PS É ys a SPL, | al e FIG. 13. Opened uneverted penial tubes. a. Triodopsis vultuosa (Gould, 1848). FMNH 214887 #A (also dissected #13: no trace of a verge; vas deferens opening terminal). b. Triodopsis cragini Call, 1886. FMNH 214803 #18 (also dissected #3: more pronounced pilaster and no sign of verge). с. Triodopsis henriettae (Mazyck, 1877). FMNH 214824 #1: pilaster seemed partly deteriorated, with structure vague. 182 EMBERTON Ory Da y $ TH 2 $ Ao AZ FIG 14. Opened uneverted penial tubes. a. Triodopsis tridentata (Say, 1816). FMNH 214876 (sympatric with Triodopsis juxtidens) #32 (also dissected 7 other populations——see Appendix C). b. Triodopsis anteridon (Pilsbry, 1940). FMNH 214796 #18 (also dissected FMNH 214793 #13, 14). c. Triodopsis juxtidens (Pilsbry, 1894). FMNH 214841 #5 (also dissected 410; FMNH 214838 #1, 2, 3; FMNH 214839 #4; FMNH 214842 #5, 6). d. Triodopsis discoidea (Pilsbry, 1904). FMNH 214811 #5. sections: 1. Fig. 18b. Length: 3 mm. Shape like a baseball bat. Ejaculatory pore terminal. Verge absent. Dorsal pilaster two-thirds the length of the sculpted region of the penis (1 mm) and tapered proximally (mid-width 0.3 mm), consisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fifth of the penis smooth with random folds: middle fifth with thick muscular walls bearing slight circu- lar corrugations; upper three-fifths sculpted with 15-20 columns of equal-sized pustules radiating directly from the pore, the ventral columns with pustules indistinct, and the ventralmost columns merging basally. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis henriettae (Mazyck, 1877)— Dissections: 2 from 1 population. Fig. 13c. Length: 9 mm. Shape like a needle. Ejacula- tory pore terminal. Verge absent. Dorsal pi- laster two-thirds the length of the sculpted region of the penis (2 mm) and tapered prox- imally (mid-width 0.2 mm), consisting of abut- ting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal third of the penis smooth with random folds; middle third with slight circular corruga- tions; upper third sculpted with equilateral, widely separated columns of equal-sized pus- tules, merging ventrally into 5-7 acute V- shapes. Sheath enclosing only the basal third of the penis. Triodopsis hopetonensis (Shuttleworth, 1852) —Dissections: 3 from 1 population. Fig. 15a. Length: 7 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, one- fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath EASTERN NORTH AMERICAN TRIODOPSINAE 183 Q A HI > SES TE mue Sublanssans detras er ARE aye 7] Fe AQ Cag O 2 LLC TT RES Et : р ae ба dé € ‘ Man, AN N N EN | al TY V ee FIG. 15. Opened uneverted penial tubes. a. Triodopsis hopetonensis (Shuttleworth, 1852). FMNH 214827 #A (also dissected #15, 25). b. Triodopsis palustris Hubricht, 1958. FMNH 214857 #15 (also dissected #4, 5). с. Triodopsis obsoleta (Pilsbry, 1894). Hubricht 10300 #C (also dissected #A, В). the pore. Dorsal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.4 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis juxtidens (Pilsbry, 1894)— Dissections: 6 from 3 populations. Fig. 14c. Length: 8 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, two-fifths-way from the apex in the upper, sculpted region. Verge absent. A large, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (3 mm) and tapered proximally (mid-width 0.5 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing about half of the upper, sculpted region of the penis. Triodopsis messana Hubricth, 1952— Dissections: 3 from 1 population. Fig. 16b. Length: 8 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (1 mm) and tapered proximally (mid-width 0.8 mm), con- 184 EMBERTON 9 RES Y, € rove AT LP eo © => emo... 7 a CES COCA U) ^^“. APDO $ — úl e TU Y ais er 2 RO y TU П er A y — == [IT] CA HH ПИ O e DIE Ton ИТОГ ть Tes Я a it, rt FIG. 16. Opened uneverted penial tubes. a. Triodopsis alabamensis (Pilsbry, 1902). FMNH 214791 #4 (also dissected #2: pilaster smaller and more lobular). b. Triodopsis messana Hubricht, 1952. FMNH 214846 #6 (also dissected #1 and #5, both with sculpture more effaced and with wall less tightly contracted). с. Triodopsis vannostrandi (Bland, 1875). FMNH 214880 #8 (also dissected #1 and #12, both with wall sculpture more effaced). sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis neglecta (Pilsbry, 1899)—Dis- sections: 2 from 1 population. Fig. 18a. Length: 5 mm. Shape like amace. Ejaculatory pore ventrally subterminal, two-fifths-way from the apex in the upper, sculpted region. Verge absent. A large, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (1 mm) and tapered proximally (mid-width 0.5 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing less than half ofthe upper, sculpted region of the penis. Triodopsis obsoleta (Pilsbry, 1894)—Dis- sections: 3 from 1 population. Fig. 15c (a contracted specimen). Length: 5 mm. Shape like a mace. Ejaculatory роге ventrally subterminal, one-fourth-way from the apex in EASTERN NORTH AMERICAN TRIODOPSINAE 185 FIG. 17. Opened uneverted penial tubes. a. Triodopsis fallax (Say, 1825). Hubricht 10209 #C (also dissected #A, В). b. Triodopsis soelneri (Henderson, 1907). ANSP A2318 (alcohol-preserved soft parts pulled from shells of ANSP 93545) #B (also dissected #A, С). the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dorsal pilaster two-thirds the length of the sculpted region of the penis (1 mm) and tapered proximally (mid-width 0.5 mm), consisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely sepa- rated columns of equal-sized pustules, merg- ing ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis palustris Hubricht, 1958—Dis- sections: 3 from 1 population. Fig. 15b. Length: 6 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.7 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis pendula Hubricht, 1952—Dis- sections: 1 from 1 population. Fig. 18c. Length: 5 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, two-fifths-way from the apex in the upper, sculpted region. Verge absent. A large, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.4 mm), con- sisting of abutting irregularly sized and 186 EMBERTON 8, KE 28% ELO ео, > № О HO N) RE Visit 1mm | I Г FIG. 18. Opened uneverted рета! tubes. a. Triodopsis neglecta (Pilsbry, 1899). ЕММН 214850 #2 (also dissected #5: no verge). b. Triodopsis fulciden Hubricht, 1952. FMNH 214823 #3. No verge; opening of vas deferens a simple hole. c. Triodopsis pendula Hubricht, 1952. FMNH 214859 #8. shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing about half of the upper, sculpted region of the penis. Triodopsis picea Hubricht, 1958—Dis- sections: 2 from 1 population. Fig. 9b. Length: 10 mm. Shape like a baseball bat. Ejaculatory pore ventrally subterminal, about one-fifth- way from the penial apex in the upper, sculpted region. Verge absent. Dorsal pilaster long (3 mm) and broad (mid-width 0.9 mm), consisting of nesting horseshoe-shaped untis covered with knob-like pustules about twice as large as the wall-pustules. Basal half ofthe penis smooth with random folds and slight circular corrugations; upper half sculpted with 15-20 columns of distinct, equal-sized pus- tules radiating directly from the pore. Sheath enclosing less than half ofthe upper, sculpted region of the penis. Triodopsis platysayoides (Brooks, 1933)— Dissections: 3 from 1 or 2 populations. Fig. 12. Length: 13 mm. Shape like a baseball bat. Ejaculatory pore terminal. Verge absent. Dor- sal pilaster long (7 mm) and very broad (mid-width 1.7 mm), consisting of two interdigitating columns of rectangular boxes, EASTERN NORTH AMERICAN TRIODOPSINAE 187 each covered with knob-like pustules about twice as large as the wall-pustules. Basal third of the penis smooth with random folds; upper two-thirds sculpted with equilateral, widely spaced columns of distinct, equal- sized pustules, merging ventrally into 10-12 obtuse V-shapes. Sheath enclosing less than half ofthe upper, sculpted region of the penis. Triodopsis rugosa Brooks & MacMillan, 1940—Dissections: 2 from 1 population. Not illustrated but similar to Fig. 18b. Length: not measured. Shape like a baseball bat. Ejacu- latory pore terminal. Verge absent. Dorsal pilaster two-thirds the length of the sculpted region of the penis and tapered proximally, consisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fifth of the penis smooth with random folds; middle fifth with thick muscular walls bearing slight circular corrugations; upper three-fifths sculpted with 15-20 columns of equal-sized pustules radi- ating directly from the pore, the ventral col- umns with pustules indistinct, and the ventral- most columns merging basally. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis soelneri (Henderson, 1907)— Dissections: 3 from 1 population. Fig. 17b. Length: 5 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. А small, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (1 mm) and tapered proximally (mid-width 0.3 mm), con- sisting of abutting irregularly sized and shaped polygons, each smooth and without spurs. Basal half of the penis smooth with random folds; upper half smooth with dorsal traces of equilateral, widely separated col- umns. Sheath enclosing only the basal half of the penis. Triodopsis tennesseensis (Walker & Pils- bry, 1902)—Dissections: 3 from 1 population. Fig. 11b—<. Length: 6 mm. Shape like a baseball bat. Ejaculatory pore terminal. Verge absent. Dorsal pilaster short (1 mm) and broad (mid-width 0.8 mm), consisting of a solid mass bearing three tiers of long, sharp spurs. Basal third of the penis smooth with random folds and slight circular corrugations; upper two-thirds sculpted with 15-20 columns radiating directly from the pore, the dorsal columns bearing indistinct, equal-sized pus- tules, and the ventral columns completely smooth. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis tridentata (Say, 1816)—Dis- sections: 10 from 8 populations. Fig. 14a. Length: 6 mm. Shape like a mace. Ejaculatory pore ventrally subterminal, one-fourth-way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dor- sal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.7 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis vannostrandi (Bland, 1875)— Dissections: 3 from 1 population. Fig. 16c. Length: 8 mm. Shape like a mace. Ejacula- tory pore ventrally subterminal, one-fourth- way from the apex in the upper, sculpted region. Verge absent. A small, smooth, fleshy peduncle present just beneath the pore. Dorsal pilaster two-thirds the length of the sculpted region of the penis (2 mm) and tapered proximally (mid-width 0.4 mm), con- sisting of abutting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal fourth of the penis smooth with random folds; middle fourth with slight circular corrugations; upper half sculpted with equilateral, widely separated columns of equal-sized pustules, merging ventrally into 5-7 acute V-shapes. Sheath enclosing only the basal half of the penis. Triodopsis vulgata (Pilsbry, 1940)—Dis- sects: 7 from 3 populations. Fig. 9a. Length: 9 mm. Shape like a baseball bat. Ejaculatory pore ventrally subterminal, about one-fifth- way from the penial apex in the upper, sculpted region. Verge absent. Dorsal pilas- ter long (3 mm) and broad (mid-width 0.9 mm), covered with knob-like pustules all about twice as large as the wall-pustules. Basal half of the penis smooth with random folds and slight circular corrugations; upper 188 EMBERTON half sculpted with 15-20 columns of distinct, equal-sized pustules radiating directly from the pore. Sheath enclosing less than half of the upper, sculpted region of the penis. Triodopsis vultuosa (Gould, 1848)—Dis- sections: 2 from 1 population. Fig. 13a. Length: 7 mm. Shape like a needle. Ejacula- tory pore terminal. Verge absent. Dorsal pi- laster two-thirds the length of the sculpted region of the penis (2 mm) and tapered prox- imally (mid-width 0.2 mm), consisting of abut- ting irregularly sized and shaped polygons, each bearing one to three short, blunt spurs. Basal third of the penis smooth with random folds; middle third with slight circular corruga- tions; upper third sculpted with equilateral, widely separated columns of equal-sized pus- tules, merging ventrally into 5-7 acute V- shapes. Sheath enclosing only the basal third of the penis. Webbhelix тиШтеаа (Say, 1821)— Dissections: 7 from 3 populations. Fig. 6a. Length: 14 mm. Shape cylindrical, the apical half enlarged. Ejaculatory pore terminal, on a verge. Verge large (1.2 mm), terminal, dorso- laterally compressed, backpointing, with a ventrally subterminal pore, smoothly sculp- ted, and bearing two broad, prominent termi- nal papillae. Dorsal pilaster short (5 mm) and broad (mid-width 1.0 mm), proximally trun- cated, and covered with small, uniform, pointed pustules. Basal two-thirds of the pe- nis smooth with random folds; upper one-third uniformly sculpted with 25-35 adjacent, gen- erally unmerging, equilateral columns of dis- tinct, equal-sized pustules radiating from the pore region; the pustules are indistinct on the basal two-thirds of these columns. Sheath enclosing less than half of the upper, sculpted region of the penis. Xolotrema caroliniensis (Lea, 1834)— Dissections: 2 from 1 population. Fig. 7e. Length: 8 mm. Shaped like a pear. Ejacula- tory pore ventrally subterminal, about one- third-way from the penial apex in the upper, scupted region, and on a verge. Verge small (0.2 mm), wider than long, ventrally sub- terminal on a slight prominence, dorso-later- ally compressed, forward-pointing, with a ven- trally subterminal pore and sculpted with surface cords continuing into about 4 narrow terminal papillae. Dorsal pilaster indistinct from the columns of wall pustules, and con- sisting of 5 broad, nested A-shapes. Basal one-third of the penis smooth with random folds; upper two-thirds sculpted with slightly separated columns of cuboidal, rough-sur- faced pustules, enlarging and merging and ventrally into 6-10 tapered U-shapes. Sheath enclosing less than half of the upper, sculpted region of the penis. Xolotrema denotata (Férussac, 1821)— Dissections: 3 from 1 population. Fig. 7a-b. Length: 9 mm. Shaped like a pear. Ejacula- tory pore ventrally subterminal, about one- third-way from the penial apex in the upper, scupted region, and on a verge. Verge small (0.2 mm), wider than long, ventrally sub- terminal on a slight prominence, dorso-later- ally compressed, forward-pointing, with a ven- trally subterminal pore and sculpted with surface cords continuing into about 4 narrow terminal papillae (Fig. 7b). Dorsal pilaster indistinct from the columns of wall pustules, and consisting of 5 broad, nested A-shapes. Basal one-third of the penis smooth with random folds; upper two-thirds sculpted with slightly separated columns of cuboidal, rough-surfaced pustules, enlarging and merg- ing and ventrally into 6-10 tapered U-shapes. Sheath enclosing less than half of the upper, sculpted region of the penis. Xolotrema fosteri (F. C. Baker, 1932)— Dissections: 6 from 2 populations. Fig. 8a. Length: 8 mm. Shape cylindrical. Ejaculatory pore terminal, on a verge. Verge small (0.2 mm), longer than wide, terminal, dorso-lat- erally compressed, backpointing, with a ven- trally subterminal pore and sculpted with sur- face cords continuing into about 6 terminal papillae. Dorsal pilaster short (2 mm), moder- ately wide (mid-width 0.3 mm), and superfi- cially resembling a single column of abutting cubes. Ventral surface bearing a long, smooth-surfaced, fleshy column with a cen- tral, longitudinal, shallow groove. Basal third of the penis smooth with random folds; middle third slightly bulbous and corrugated by bands of circular and longitudinal muscles; upper third sculpted with slightly separated columns of cuboidal, smooth-surfaced pus- tules, enlarging and merging and ventrally into 6-10 tapered U-shapes. Sheath enclos- ing the entire upper, sculpted region of the penis. Xolotrema obstricta (Say, 1821)—Dis- sections: 2 from 1 population. Fig. 7c-d. Length: 11 mm. Shaped like an inverted pear. EASTERN NORTH AMERICAN TRIODOPSINAE 189 Ejaculatory pore ventrally subterminal, about one-third-way from the penial apex in the upper, sculpted region, and on a verge. Verge small (0.2 mm), wider than long, ventrally subterminal on a slight prominence, dorso- laterally compressed, forward-pointing, with a ventrally subterminal pore and sculpted with surface cords continuing into about 4 narrow terminal papillae (Fig. 7d). Dorsal pilaster indistinct from the columns of wall pustules, and consisting of 5 broad, nested A-shapes. Basal one-third of the penis smooth with random folds; upper two-thirds sculpted with slightly separated columns of cuboidal, rough-surfaced pustules, enlarging and merg- ing and ventrally into 6-10 tapered U-shapes. Sheath enclosing less than half of the upper, sculpted region of the penis. Xolotrema occidentalis (Pilsbry 8 Ferriss, 1907) —Dissections: 1 Нот 1 population. Fig. 8b—c. Length: 7 mm. Shape cylindrical (in Fig. 8b, it appears clubbed because of con- traction within the sheath). Ejaculatory pore terminal, on a verge. Verge small (0.3 mm), longer than wide, terminal, dorsolaterally compressed, back-pointing, with a ventrally subterminal pore and sculpted with surface cords continuing into about 6 terminal papillae (Fig. 8c). Dorsal pilaster short (1 mm), mod- erately wide (mid-width 0.3 mm), and super- ficially resembling a single column of abutting cubes. Ventral surface bearing a long, smooth-surfaced, fleshy column with a cen- tral, longitudinal, shallow groove (this struc- ture is contracted and distorted in Fig. 8b). Basal third of the penis smooth with random folds; middle third slightly bulbous and corru- gated by bands of circular and longitudinal muscles; upper third sculpted with slightly separated columns of cuboidal, smooth-sur- faced pustules, enlarging and merging and ventrally into 6-10 tapered U-shapes. Sheath enclosing the entire upper, sculpted region of the penis. Suggested character-state transformations The total variation in penial morphology was classified into 10 characters comprising 60 character states. These are arranged into their suggested phylogenies in Figs. 19-23, in which the suggested character-state transfor- mations are numbered 1-50. The dorsal pilaster (Character 1) was the most variable penial-morphological character. Twenty-two states (including its absence in the outgroups) were detected, none of which appeared to be convergent. Their suggested phylogeny (Fig. 19) contains transformations 1-21. Pustules on the penial wall (Character 2) yielded 11 character-states, with two sets of convergences, each involving three charac- ter-states: types 1, 2, and 3 chevrons; and 3 types of smooth columns (explained below). The suggested phylogeny (Fig. 20) involves transformations 22-33. Verges (Character 3) occur in several ofthe outgroups of eastern triodopsines: Ves- pericola, Oreohelicidae, and some Camaen- idae. These verges, because of their struc- tural differences (discussed below), pre- sumably are convergent on, rather than plesiomorphous with, the eastern-triodopsine verge. Six character states were detected, three of which appeared convergent (types 1, 2, and 3 small verges). The suggested char- acter-state phylogeny (Fig. 21) embodies transformations 34-38. The position on the penis of the ejaculatory pore, or opening of the vas deferens (Char- acter 4), varied as 6 distinct character states, 4 of which appeared convergent (dorsally subterminal, and types 1, 2, and 3 ventrally subterminal). Evolution of a ventrally sub- terminal pore is probably easily achieved de- velopmentally by overgrowth of the dorsal penial wall. This presumably is functionally adaptive because it both plugs the mate's spermathecal duct during copulation (with the overgrown apical knob of the penis) and emits the sperm mass beneath the plug and away from the digestive spermathecal bursa (Emberton, 1986). The convergences were detected by differences in penial shape and in details of pore position and structure, as explained below. The suggested character- state phylogeny (Fig. 22) contains transfor- mations 39-43. Characters 5-10 (ventral pilaster, basal pe- nis length, ventral sperm channel, sheath length, upper penis length, and peduncle) each had two or three character states, sug- gested to be linked by one or two transforma- tions (Fig. 23, transformations 44-50). In presenting each of the 50 suggested character-state transformations below, the same format has been used throughout: (1) the transformation's identification number as used in Figs. 19-23; (2) the identification numbers of the transformation or series of transformations suggested to have preceded it evolutionarily; (3) the suggested plesio- Xolotrema Triodopsis Neohelix = Pil All polygons fused Pi arı d into a solid mass 1 bro 16 bearing 3 tiers of 3 11 long, sharp spurs EN \ 8 Each polygon dis- | 11 12 18 tinct and bearing | 1-3 short, blunt | af appets ghtly 'ilastral pustule So polygons spurs | li appres 1 ery large and irregularly fused ‚raduall lecreasin | ventrall та и 9 17 Pustules fused into | les m irregular polygons | +___ | > t 6 | а; appets Pustules fused | at-surfacec into 2 inter- | separated 10 digitating columns | of rectangular | . Е boxes 16 | 5 | Pustules fused | into nesting Pilaster less than E <. Partial lateral horseshoe-shape 1/2-length and 4 fusion of pustules, 15 proximally truncated ing distinct | > lappet 4 14 20 | Pustules knob-like, 1 abruptly larger Pilaster 2/3-length | than the wall pus- > and proximally / 3 tules, and unfused 21 tapered y Z 7 = Dorsal pilaster pr == — sent; covered wit! Pila A-length, ———— small, uniform, pointed | prox lly truncated 2 pustules; full-length [ Webbhelix 1 >», Various outgroups Dorsal pilaster absent \ FIG. 19. Suggested character-state transformations in eastern American triodopsine penial morphology. Character 1, pilaster and pilastral pustules. morphous state; (4) the outgroup taxa роз- sessing the suggested plesiomorphous state; (5) the suggested apomorphous state; (6) the taxa suggested to have formerly possessed the apomorphous state, although lacking it now; (7) the taxa which now possess the suggested apomorphous state; and (8) a dis- cussion of the suggested transformation, in- cluding its further explanation, if necessary, and its justification. Transformation 1—Preceding transforma- tions: none. Plesiomorphous state: dorsal pilaster ab- sent. Present in: Cryptomastix, Allogona, Ashmunella, Oreohelix, Polygyrella, Poly- gyracea, and the camaenids Amplirhagada and Torresitrachea. Apomorphous state: dorsal pilaster pres- ent, covered with unmodified pustules, and full-length. Formerly present in: all eastern triodopsines (Figs. 2-18). Now present in: Webbhelix multilineata (Fig. 6a). Discussion. Emberton (1986) discussed the evidence that this type of single dorsal pilaster, formed by a longitudinal outgrowth from the penial wall, is unique to triodopsines within the Polygyridae and their outgroups (a similar pilaster in some Australian camaenids is considered convergent). It apparently oc- curs in all eastern triodopsines (Figs. 2-18); although it is not at а! obvious in the dissec- tions of Xolotrema (Figs. 7, 8), it shows up clearly in cross-sections of the penes of X. denotata and X. fosteri (Pilsbry 1940, fig. 473 #6b, 7b: 793), so presumably occurs EASTERN NORTH AMERICAN TRIODOPSINAE 191 Triodopsis Type 2 chevron: Type 3 chevron: equilateral, widely equilateral, widel; separated columns separated columns merging ventrally merging ventrall into 10-12 obtuse into 5-7 acute V-shapes V-shapes x 32 All columns 8-10 columns, the Ventral columns merging mid- ventral-most smooth ventrally merging basally 1 \ Neohelix 77.2. aa Basal pustules | \ 28 30 33 24 enla zed р \ y Ventral columns 15-20 columns, the with pustules ventral-most р Xolotrema eer ı Apical 1/5ths See E indistinct merging basally | of columns smooth I ¡Type 1 chevron: \ tapered, slightly \ separated columns A 27 29 all merging ventrally Webbhelix o 23 into 6-10 U-shapes 15-20 unmerging УБЫТКА АЕ С RES columns radiating / Basal 2/3rds directly from the ; of columns 25 pore with pustules 26 indistinct NE | Sage NEE ETF у 22 N A IE ee is Еее RS ariousoutgroups +S & 25-35 adjacent, unmerging, equilateral columns of dis- \ Y tinct, equal-sized pustules, radiating from the pore region FIG. 20. Suggested character-state transformations in eastern American triodopsine penial morphology. Character 2, wall pustules. Neohelix ~~ Es -- © ~~~. Xolotrema р Type 1 small verge: Type 2 small verge: i subterminal and basally / terminal, longer than \ directed, wider than N wide, bearing 6 long, bearing 6-8 x narrow papillae \ narrow papillae Type 3 small verge: à \ subterminal and apically \ directed, wider than wer) \ 36 long, bearing 4 = = =e \ narrow papillae / Papillae 2 in \ ! number, broad, x \ and prominent 3 es Large, terminal, compressed, 5 - back-pointing verge with а М bbhe lip 3 \ subterminal pore and surface \ webbnellx \ cords continuing into 6 or more \ sen SS SS 1 narrow terminal papillae N a Small-to-large, terminal, + | conical, forward-pointing \ verge with a terminal pore and a smooth surface RR AE ck ; = ı ‘ Verge absent, Triodopsis — ____--- Various outgroups FIG. 21. Suggested character-state transformations in eastern American triodopsine penial morphology. Character 3, verge. throughout the genus. It is assumed, for want pearance appears to be attributable to differ- of evidence to the contrary and because a ences in the patterns of fusion and enlarge- similar structure appears nowhere else ment of the pilastral pustules. among the Polygyridae or their outgroups, The most plesiomorphous pilastral sculp- that this single dorsal pilaster arose only ture appears to be that seen in Webbhelix once, so is homologous throughout eastern multilineata (Fig. 6a). This pilastral sculpture triodopsines. lts great variation in gross ap- is identical with the plesiomorphous wall 192 Xolotrema 40 EMBERTON Triodopsis Webbhelix and various outgroups FIG. 22. Suggested character-state transformations in eastern American triodopsine penial morphology. Character 4, pore position. : se ites Neohelix 5 Ying Е 44 45 al pilaster 1 or less of the 3 nt total penis length lying between the vaginal opening and the base of th sheath = Triodopsis \ ee ae ae Webbhelix and Peduncle large various Outgroups : Xolotrema | | es 50 rin AJ Upper penis | extremely long Peduncle present | and thread-like and small | | 46 pe ' 48 49 1 / ntr perr ath coverin Upper penis Peduncle absent / y ro at nt 1 ; than 1/2 ;hort to long mi | t upper peni } FIG. 23. Suggested character-state transformations in eastern American triodopsine penial morphology. Characters 5-10, ventral pilaster, basal penis length, ventral sperm groove, sheath length, upper penis length, and peduncle. sculpture: adjacent, longitudinal columns of equal-sized pustules. Transformation 2—Preceding transforma- tions: 1. Plesiomorphous state: pilaster full-length. Present in (outgroups): Neohelix (Figs. 2-5, 6b); Xolotrema (Figs. 7, 8); and Triodopsis vulgata, picea, claibornensis (Fig. 9), burchi (Fig. 11a), and platysayoides (Fig. 12). Apomorphous state: pilaster 3/4-length and proximally truncated. Formerly and now present in: Webbhelix multilineata (Fig. 6a). Discussion: This character state is unique to W. multilineata, so presumably is apo- morphous. EASTERN NORTH AMERICAN TRIODOPSINAE 193 Transformation 3—Preceding transforma- tions: 1. Plesiomorphous state: pilaster covered with unmodified pustules. Present in (out- group): Webbhelix multilineata (Fig. 6a). Apomorphous state: pilastral pustules par- tially fused laterally to form lappets. Formerly present in: all Neohelix (Figs. 2-5, 6b). Now present in: Neohelix dentifera (Fig. 2a, b), lioderma (Fig. 5a, с), and divesta (Fig. 5d, f). Discussion. There appears to be a contin- uum from the totally unfused pilastral pustules of W. multilineata (Fig. 6a), to the laterally appressed slightly fused pilastral pustules of N. dentifera (Fig. 2b), to the partially laterally fused pilastral pustules of N. lioderma (Fig. 5c) and N. divesta (Fig. 5f), and the assump- tion is that this represents a true transforma- tion series. This lateral fusion of pilastral pustules results in a column of overlapping, plate-like elements called “lappets” (Figs. 1, 2a, 5a, d). Transformation 4—Preceding transforma- tions: 1, 3. Plesiomorphous state: lappets approxi- mately equal in number to the number of columns of wall pustules. Present in (outgroup): Neohelix albolabris (Fig. 2d), al- leni (Fig. 3a, c), major (Fig. 4a), and solemi (Fig. 6b). Apomorphous state: lappet number dou- bled. Formerly and now present in: Neohelix dentifera (Fig. 2a), lioderma (Fig. 5a), and divesta (Fig. 5d). Discussion. There are two distinct types of lappetted dorsal pilaster. In the first type, the number of pilastral lappets is approximately equal to (or somewhat greater than) the num- ber of columns of wall pustules, as seen in albolabris, alleni, major, and solemi. In the other type, the number of pilastral lappets is approximately equal to twice the number of columns of wall pustules, as seen in dentifera, lioderma, and divesta. There are no interme- diates between these two types. It appears likely that the double-lappet sculpture is de- rived from the single-lappet sculpture, possi- bly ма a simple, one-step modification in a developmental program. Transformation 5—Preceding transforma- HONS: 1, 3. Plesiomorphous state: lappet pustules par- tially fused. Present in (outgroups): Neohelix dentifera (Figs. 2a, b), lioderma (Fig. 5a, с), and divesta (Fig. 5d, f). Apomorphous state: lappet pustules com- pletely fused. Formerly and now present in: Neohelix albolabris (Fig. 2d, e), alleni (Fig. 3a, c), major (Fig. 4a), and solemi (Fig. 6b). Discussion. Although the lappets of alleni, and major, and solemi appear to have lost all trace of the pustules from which they presum- ably originated by lateral fusion, those of albolabris and solemi show a regular pattern of indentations (Figs. 2e and 6b) which seem to correspond to pustules (compare with Fig. EC: В. Transformation 6—Preceding transforma- tions Ar 35; Plesiomorphous state: lappets distinct. Present in (outgroups): Neohelix dentifera, albolabris, alleni, major, lioderma, and divesta (Figs. 2-5). Apomorphous state: lappets indistinct. For- merly and now present in: Neohelix solemi (Fig. 6b). Discussion: The dorsal pilaster of solemi, which appears on the right in Fig. 6b and is not to be confused with the ventral pilaster (unique to solemi) which appears in the center in Fig. 6b, is reduced in size and length due to the uniquely dorsally sub- terminal pore position (Transformation 38). Its lappets are indistinct and unequal in size, and may be vestigial now that a ventral pilaster is present. Transformation 7—Preceding transforma- tions: 1, 3,5: Plesiomorphous state: lappets flat-sur- faced. Present in (outgroups): Neohelix dentifera (Fig. 2a), alleni (Fig. 3c), lioderma (Fig. 5a), and divesta (Fig. 5d). Apomorphous state: lappets slightly con- vexly surfaced. Formerly present in: albo- labris (Fig. 2d, e) and major (Fig. 4a). Now present in: albolabris (Fig. 2d, e). Discussion. The convexity of albolabris's pilastral lappets (Fig. 2e) seems to result from a trend toward enlargement of the lappets which is continued in Transformation 8. Transformation 8—Preceding transforma- tions: 1. 3,5, 7. Plesiomorphous state: lappets slightly con- vexly surfaced. Present in (outgroup): Neo- helix albolabris (Fig. 2d, e). Apomorphous state: lappet surfaces very convex and irregularly wavy. Formerly and now present in: Neohelix major (Fig. 4a). Discussion. This character state is unique 194 EMBERTON to major, which has the largest pilastral lap- pets, so presumably is apomorphous. Transformation 9—Preceding transforma- tions: 1,3, 3. Plesiomorphous state: lappets slightly sep- arated. Present in (outgroups): Neohelix albolabris (Fig. 2a, e), major (Fig. 4a), and solemi (Fig. 6b). Apomorphous state: lappets tightly appres- sed. Formerly and now present in: Neohelix alleni (Fig. 3c). Discussion: The lappets are extremely smooth-surfaced and fit tightly together so that the general pilastral surface is relatively flat (Fig. 3c). Fig. За was accidently incor- rectly shaded and does not properly repre- sent this character state. Transformation 10—Preceding transforma- tions: 1. Plesiomorphous state: pilastral pustules small and uniform in size. Present in (out- group): Webbhelix multilineata (Fig. 6a). Apomorphous state: pilastral pustules very large and gradually decreasing in size ven- trally. Formerly and now present in: Xolo- trema (Figs. 7a, с, e; 8a, b). Discussion. Xolotrema's dorsal pilaster is unique and quite disjunct from any other found in eastern triodopsines. lts derivation from the W. multilineata-type is a best guess which is supported by the homologous verge (Character 3) between Webbhelix and Xolo- trema. Whether the large pilastral pustules originated by enlargement, or fusion, or both, is beyond conjecture at this point. Transformation 11—Preceding transforma- tions: 1, 10. Plesiomorphous state: pilastral pustules very large and gradually decreasing in size ventrally, but of unknown mid-dorsal configu- ration. Formerly present in: hypothesized common ancestor of Xolotrema denotata, obstricta, caroliniensis, fosteri, and ос- cidentalis (Figs. 7, 8). Now present in: none. Apomorphous state: pilastral pustules a single column of abutting cubes. Formerly and now present in: Xolotrema fosteri (Fig. 8a) and occidentalis (Fig. 8b). Discussion. The superficial resemblance of fosteri's and occidentalis's pilastral elements (Fig. 8a, b) to lappets (e.g., Figs. 2d, 3a) and to polygons (e.g. Figs. 14a, 15а) breaks down on close examination; these three types ap- pear to be nonhomologous. The dorsal pilas- ter of fosteri and occidentalis should not be confused with the ventral sperm groove (Figs. 8a, b) which occurs in these two species. Transformation 12—Preceding transforma- tions: 1, 10. Plesiomorphous state: pilastral pustules very large and gradually decreasing in size ventrally, but of unknown mid-dorsal configu- ration. Present in: hypothesized common an- cestor of Xolotrema denotata, obstricta, caroliniensis, fosteri, and occidentalis (Figs. 1,8). Apomorphous state: pilastral pustules ar- ranged into 5 broad, nested A-shapes. For- merly and now present in: Xolotrema deno- tata (Fig. 7a), obstricta (Fig. 7c), and caroliniensis (Fig. 7e). Discussion. The dorsal pilaster in denotata, obstricta, and caroliniensis is less obvious than in any other eastern triodopsines. The thickening of the dorsal penial wall which forms the dorsal pilaster is reduced in these species (Pilsbry 1940, fig. 473 #6b) to the extent that it is not at all evident in Fig. 7a, c, e. The swollen area beneath the subterminal verge in these species is easily mistaken for the dorsal pilaster, but the fact that the verge points up indicates that this is actually the ventral side of the penis, so the dorsal pilaster is on the opposite side (Fig. 7a, с, €), which is heavily sculpted with A-shaped arrangements of broad, rugose pustules. The substructural complexity of these pustules suggests that they resulted from fusion of smaller, plesiomorphous pustules. Despite the great difference between the dorsal pilastral sculptures of denotata, obstricta, and caroliniensis on one the hand (Fig. 7), and fosteri and occidentalis on the other hand (Fig. 8), their similarity in the broad, flat- surfaced, ventrally-diminishing pilastral pus- tules, as well as their apparent homologies in wall pustules (Character 2) and verge (Char- acter 3), lead to the suggestion that their dorsal pilasters arose from a common, unknown ancestral type. Transformation 13—Preceding transforma- tion: 1. Plesiomorphous state: pilastral pustules pointed and uniformly equal in size to the wall pustules. Present in (outgroup): Webbhelix multilineata (Fig. 6a). Apomorphous state: pilastral pustules knob-like and abruptly larger than the wall pustules. Formerly present in: all Triodopsis EASTERN NORTH AMERICAN TRIODOPSINAE 195 (Figs. 9-18). Now present in: T. vulgata (Fig. Эа) and T. claibornensis (Fig. 9c). Discussion. Knob-like pilastral pustules about twice as large as the wall pustules, with no ventral intergradation in size, occur either unfused or fused in various ways in vulgata, picea, claibornensis (Fig. 9), burchi (Fig. 11a), and platysayoides (Fig. 12). The most similar pilastral sculpture, and one from which this type could easily have evolved, is that of W. multilineata (Fig. 6a), in which the pilastral pustules are more pointed and scale-like, and equal in size to the wall pustules: simple enlargement would have sufficed. Transformation 14—Preceding transforma- tions: 1, 13. Plesiomorphous state: knob-like pilastral pustules unfused. Present in (outgroup): Triodopsis vulgata (Fig. 9a) and claibornensis (Fig. 9c). Apomorphous state: pilaster sculpted with nesting horseshoe-shaped elements with a knobby surface. Formerly and now present in: Triodopsis picea (Fig. 9b). Discussion. Despite their apparent fusion into this pattern, the apical knobs of the pilastral pustules are readily apparent in picea. Transformation 15—Preceding transforma- tions: 1, 13. Plesiomorphous state: knob-like pilastral pustules unfused. Present in (outgroup): Triodopsis vulgata (Fig. 9a) and claibornensis (Fig. 9c). Apomorphous state: pilaster sculpted with rectangular box-like elements with knobby surfaces and arranged in two interdigitating columns. Formerly and now present in Triodopsis platysayoides (Fig. 12). Discussion. The knobby surface of playtsay- oides’s pilaster suggests that its box-like ele- ments derived by fusion from the knob-like pilastral pustules seen in vulgata and claibornensis. Transformation 16—Preceding transforma- tions: 1, 13. Plesiomorphous state: knob-like pilastral pustules unfused. Present in (outgroup): Triodopsis vulgata (Fig. 9a) and claibornensis (Fig. 9c). Apomorphous state: pilaster sculpted with elements (polygons) 4—10 times the size of wall pustules and bearing 2-5 knobs. For- merly present in: hypothesized ancestor of all of Triodopsis except vulgata, picea, clai- bornensis, and platysayoides (Figs. 10, 11, 13-18). Now present in: none. Discussion. These polygonal pilastral ele- ments, which occur with modification through- out most of Triodopsis, have a surface sculp- ture or substructure reminiscent of the unfused pilastral pustules of vulgata and claibornensis (Fig. Эа, c)—this is especially evident in the illustration of messana (Fig. 16b)—so the assumption is that they were derived from these by fusion. Transformation 17—Preceding transforma- tions 18 167 Plesiomorphous state (?): knobby-surfaced pilastral polygons. Present in (outgroup): hy- pothesized ancestor approximated by the il- lustration of Triodopsis messana (Fig. 16b). Apomorphous state: knobby-surfaced pilas- tral elements 1-2 times as large as polygons. Formerly and now present in: Triodopsis burchi (Fig. 11a). Discussion. The pilastral sculpture of burchi is unique and problematic. Because of its knobby surface, it probably derived from ei- ther a vulgata-type ancestor (Fig. 9a) or a hypothesized ancestor in which partial fusion (into polygons) of the pilastral pustules had already taken place. The latter alternative was chosen because the irregular size and pattern of burchi’s pilastral elements suggest that intermediate fusion has taken taken place. Transformation 18—Preceding transforma- tions: 1, 13, 162 Plesiomorphous state (?): knobby-surfaced pilastral polygons. Present in (outgroup): hy- pothesized ancestor approximated by the il- lustration of Triodopsis messana (Fig. 16b). Apomorphous state: dorsal pilaster a solid, rounded mass bearing about 3 tiers of long, sharp spurs. Formerly and now present in: Triodopsis tennesseensis (Fig. 11b) and complanata (Fig. 11d). Discussion. This is another unique and problematic form of the dorsal pilaster. The spurs are so much longer, sharper, and more regularly arranged than are the blunt spurs of Transformation 19 that they are probably not homologous. The substructure of compla- nata's pilaster (Fig. 11d) somewhat гезет- bles a hypertrophied and regularized form of, for example, the pilaster of tridentata (Fig. 14a), so it may have derived from a pilaster with polygonal elements. 196 EMBERTON Transformation 19—Preceding transforma- tions: 1, 13, 16. Plesiomorphous state: pilastral polygons with simple knobby surface. Present in (out- group): hypothesized ancestor most closely approximated by the illustration of Triodopsis messana (Fig. 16b). Apomorphous state: pilastral polygons bearing blunt spurs. Formerly and now present in: Triodopsis fraudulenta (Fig. 10), vultuosa, cragini, henriettae, tridentata, an- teridon, juxtidens, discoidea, hopetonensis, palustris, obsoleta, alabamensis, messana, vannostrandi (Flgs. 13-16); fallax (Fig. 17a), neglecta, fulciden, and pendula (Fig. 18). Discussion. The blunt spurs, which are most obvious in the illustrations of vultuosa (Fig. 13a), tridentata (Fig. 14a), anteridon (Fig. 14b), and alabamensis (Fig. 16a), ap- pear to be derived by outgrowth of the indi- vidual pustules which originally fused (Trans- formation 16) to form the polygons. Transformation 20—Preceding transforma- tions: 1. Plesiomorphous state: dorsal pilaster full- length. Present in (outgroups): Webbhelix (Fig. 6a), Neohelix (Figs. 2-5, 6b);Xolotrema (Figs. 7, 8); and Triodopsis vulgata, picea, claibornensis (Fig. 9), burchi (Fig. 11a), and platysayoides (Fig. 12). Apomorphous state: dorsal pilaster less than 1/2 length and proximally truncated. For- merly and now present in: Triodopsis tennes- seensis (Fig. 11b) and complanata (Fig. 11d). Discussion. This short pilaster appears apomorphous relative to the taxonomically widespread and probably ontogenetically more easily achieveable full-length pilaster. Transformation 21—Preceding transforma- tions: 1. Plesiomorphous state: dorsal pilaster full- length. Present in (outgroups): Webbhelix (Fig. ба), Neohelix (Flgs. 2-5, 6b); Xolotrema (Figs. 7, 8); and Triodopsis vulgata, picea, claibornensis (Flg. 9), burchi (Fig. 11a), and platysayoides (Fig. 12). Apomorphous state: dorsal pilaster 2/3 length and proximally tapered. Formerly and now present in: Triodopsis fraudulenta (Fig. 10), vultuosa, cragini, henriettae, tridentata, anteridon, juxtidens, discoidea, hopetonen- sis, palustris, obsoleta, alabamensis, mes- sana, vannostrandi (Figs. 13-16), fallax (Fig. 17a), neglecta, fulciden, and pendula (Fig. 18). Discussion. For the same reasons cited for Transformation 20, it is assumed that this shortened pilaster is apomorphous. Transformation 22—Preceding transforma- tions: none. Plesiomorphous state: all wall columns bearing distinct pustules along their entire lengths. Present in (outgroups): some Camaenidae, Oreohelicidae, and Ashmunel- linae; all Neohelix except alleni (Figs. 2, 4, 5, 6b). Apomorphous state: all wall columns with pustules indistinct basally. Formerly and now present in: Webbhelix multilineata (Fig. 6a). Discussion. Despite their partial fusion in multilineata, the wall pustules are still evident and give the penial wall a rough surface sculpture. Thus this character state differs from the smooth pustular columns discussed under Transformations 23, 27, and 28. Transformation 23—Preceding transforma- tions: none. Plesiomorphous state: all wall columns bearing distinct pustules along their entire lengths. Present in (outgroups): some Camaenidae, Oreohelicidae, and Ashmunel- linae; all Neohelix except alleni (Figs. 2, 3, 5, 6b). Apomorphous state: all wall columns with their apical 1/5th to 1/4th smooth, with no trace of pustules. Formerly and now present in: N. alleni (Fig. 3). Discussion. Although smooth wall columns occur in some other eastern triodopsines (Transformations 27 and 28), the apical local- ization seen in alleni is unique and surely nonhomologous. Transformation 24—Preceding transforma- tions: none. Plesiomorphous state: all wall pustules ap- proximately equal in size or smaller basally. Present in (outgroups): some Camaenidae, Oreohelicidae, and Ashmunellinae; all Neo- helix except dentifera, lioderma, and divesta (Figs. 2d, 3, 4, 6). Apomorphous state: basal wall pustules more than twice as large as the apical wall pustules. Formerly and now present in: Neohelix dentifera (Fig. 2a), lioderma (Fig. 5a), and divesta (Fig. 5d). Discussion. There is variation within the apomorphous state: in dentifera the basal enlargement is more localized and abrupt than in lioderma and divesta. Because it is EASTERN NORTH AMERICAN TRIODOPSINAE 197 unclear which of these variations would be plesiomorphous to the other, they were left together as one apparently homologous, apomorphous state. Transformation 25—Preceding transforma- tions: none. Plesiomorphous state: wall columns un- merging, 25-35 in number, linear, equilateral, adjacent, and bearing equal sized-pustules. Present in (outgroups): some Camaenidae, Oreohelicidae, and Ashmunellinae; Webb- helix; all Neohelix except dentifera, lioderma, and divesta (Figs. 2d, 3, 4, 6). Apomorphous state: Type 1 chevron: wall columns all merging mid-ventrally into 6-10 U-shapes, tapered, slightly separated, and bearing unequally sized pustules. Formerly and now present in: Xolotrema (Flgs. 7, 8). Discussion. This and similar patterns are being called “chevrons” because the ventral wall resembles an inverted chevron. Despite a superficial similarity to Types 2 and 3 chev- rons (Transformations 31 and 32), the Type 1 chevron can be recognized as convergent by its ventral U- rather than V-shapes, its ta- pered rather than equilateral columns, its slightly rather than widely separated columns, its unequally rather than equally sized pus- tules, and its flat-surfaced rather than pointed pustules. The gap between the Type 1 chev- ron and its hypothesized plesiomorphous state is great and there are no other character states that appear transitional. The series represented by Transformations 29, 30, and 31 or 32 (discussed below) outlines a possi- ble path similar to one by which the Type 1 chevron may have arisen independently. Transformation 26—Preceding transforma- tions: none. Plesiomorphous state: 25-35 columns ra- diating from the pore region and adjacent and equally sized along their entire lengths. Present in (outgroups): some Camaenidae, Oreohelicidae, and Ashmunellinae; Webb- helix; and Neohelix (Figs. 2-6). Apomorphous state: 15-20 columns radiat- ing directly from and diverging and/or enlarg- ing from the pore. Formerly present in: Triodopsis (Figs. 9-18). Now present in Triodopsis vulgata, picea, and claibornensis (Fig. 9). Discussion. This wall-pustular pattern, in combination with a subterminal pore (Trans- formation 41), takes on the distinctive appear- ance of a spider's orb-web (especially Fig. Эа). It is closest to the presumably plesio- morphous pattern seen in the triodopsine outgroups and in Webbhelix and Neohelix, and could have arisen by fusion and/or simple reduction of wall columns. Transformation 27—Preceding transforma- tions: 26. Plesiomorphous state: 15-20 radiating wall columns, all with distinct pustules. Present in (outgroup): Triodopsis vulgata, picea, and claibornensis (Fig. 9). Apomorphous state: 15-20 radiating wall columns, the ventral-most with indistinct pus- tules. Formerly present in Triodopsis fraudu- lenta, burchi, tennesseensis, complanata, and fulciden (Figs. 10, 11, 18b). Now present in: Triosopsis fraudulenta, rugosa (Fig. 10), burchi (Fig. 11a), and fulciden (Fig. 18b). Discussion. The ventral wall columns are semi-smooth, apparently due to either the partial fusion or partial loss of their pustules. Similarity to the indistinct wall pustules of Webbhelix multilineata (Fig. ба) is due to convergence. Transformation 28—Preceding transforma- tions: 26, 27. Plesiomorphous state: 15-20 radiating wall columns, the ventral-most with indistinct pus- tules. Present in (outgroup): Triodopsis fraudulenta (Fig. 10), burchi (Fig. 11a), and fulciden (Fig. 18b). Apomorphous state: 15-20 radiating wall columns, the ventral-most smooth, with no trace of pustules. Formerly and now present in Triodopsis tennesseensis (Fig. 11b, c) and complanata (Fig. 11d). Discussion. The assumption is that the ventral pustules of tennesseensis and com- planata did not become smooth directly, but passed through a semi-smooth stage homol- ogous with that of fraudulenta, burchi, and fulciden. Transformation 29—Preceding transforma- tions: 26. Plesiomorphous state: 15-20 radiating wall columns, the ventral-most unmerging. Pres- ent in (outgroup): Troiodopsis vulgata, picea, claibornensis (Fig. 9), burchi, tennesseensis, and complanata (Fig. 11). Apomorphous state: 15-20 radiating wall columns, the ventral-most merging basally. Formerly present in: all Triodopsis except vulgata, picea, claibornensis, burchi, tennes- seensis, and complanata (Figs. 10, 12-18). 198 EMBERTON Now present in: Triodopsis fraudulenta, ru- gosa (Fig. 10), and fulciden (Fig. 18b). Discussion. The ventral wall columns form spindle shapes by diverging from the роге, then merging basally. The basal merging pre- sumably derived from non-merging columns, probably by a simple change in the develop- mental program by which the pustular col- umns form. Transformation 30—Preceding transforma- tions: 26, 29. Plesiomorphous state: 15-20 wall columns, the ventral-most merging basally. Present in (outgroup): Triodopsis fraudulenta, rugosa (Fig. 10), and fulciden (Fig. 18b). Apomorphous state: all 15-20 wall columns merging midventrally to form a plesiomorph- ous inverted chevron pattern. Formerly pres- ent in: Triodopsis platysayoides, vultuosa, cragini, henriettae, tridentata, anteridon, juxtidens, discoidea, hopetonensis, palustris, obsoleta, alabamensis, messana, vanno- strandi, fallax, soelneri, neglecta, and pendula (Figs. 12-18). Now present in: none. Discussion: The ventral wall patterns of platysayoides (Type 2 chevron) and the re- maining species (Type 3 chevron) differ sig- nificantly, but have enough features in com- mon that they probably had a common ancestor of unknown appearance, but proba- bly closer to platysayoides because of the number of pustular columns involved. It is assumed that the mid-ventral merging of wall columns evolved in a basal-to-apical direc- tion, with an intermediate stage in this pro- cess represented by the basal merging seen in fraudulenta, rugosa, and fulciden. Onto- genetic studies of penial sculpture may prove useful in testing this assumption. Transformation 31—Preceding transforma- tions: 26, 29, 30. Plesiomorphous state: all 15-20 wall col- umns merging midventrally to form a plesiomorphous inverted chevron pattern. Present in (outgroup): hypothesized ancestor probably closest to Triodopsis platysayoides (Big 12): Apomorphous state: Type 2 chevron: wall columns all merging midventrally into 10-12 obtuse V-shapes, equilateral, widely sepa- rated, and bearing equally sized pustules. Formerly and now present in Triodopsis platysayoides (Fig. 12). Discussion. This Type 2 chevron is conver- gent on Types 1 and 3 chevrons. For differ- ences from the Type 1 chevron, see the discussion under Transfomation 25. The Type 2 differs from the Type 3 chevron (Transformation 32) by the greater number and more obtuse angle of its ventral V- shapes. Transformation 32—Preceding transforma- tions: 26, 29, 30. Plesiomorphous state: all 15-20 wall col- umns merging midventrally to form an in- verted chevron pattern. Present in (outgroup): hypothesized ancestor probably closest to Triodopsis platysayoides (Fig. 12). Apomorphous state: Type 3 chevron: wall columns all merging mid-ventrally into 5-7 acute V-shapes, equilateral, widely sepa- rated, and bearing equally sized pustules. Formerly and now present in: Triodopsis vultuosa, cragini, henriettae, tridentata, an- teridon, juxtidens, discoidea, hopetonensis, palustris, obsoleta, alabamensis, messana, vannostrandi, fallax, soelneri, neglecta, and pendula (Figs. 12-18). Discussion. There is considerable variation within this character state, and a more thor- ough and extensive study may break it down into a number of systematially useful catego- ries. What is considered the “basic” Type 3 chevron is well represented in the illustrations of hopetonensis (Fig. 26a), palustris (Fig. 15b), and alabamensis (Fig. 16a); it differs from the Type 2 chevron Transformation 31) by the lesser number and the more acute angle of its ventral V-shapes; for its difference from the Type 1 chevron, see the discussion under Transformation 25. Variations from the “basic” Type 3 chevron include the more acutely angled V-shapes, presumably due to penial elongation, in vultuosa, cragini, and henriettae (Fig. 13); the apparent partial ef- facement of the wall sculpture of juxtidens (Fig. 14c), possibly due to sympatry with the similar tridentata (Fig. 14a); the possible ef- facement of the wall sculpture of fallax (Fig. 17a), which may be an artifact of the strong contraction of this specimen; the apparently total effacement of the wall sculpture of soelneri (Fig. 17b), possibly due to sympatry with the similar hopetonensis (Fig. 15a) and messana (Fig. 16b) (see Table 8); and the anastomoses among the V-shapes in pendula (Fig. 18c). Deriving both Types 2 and 3 chevrons from a common ancestor is a parsimonious suggestion, but need not be true: the apparent independent origin of the Type 1 chevron is evidence that Types 2 and EASTERN NORTH AMERICAN TRIODOPSINAE 199 3 could have evolved independently of each other as well. Transformation 33—Preceding transforma- tions: 26, 29. Plesiomorphous state: 15-20 radiating wall columns, the ventral-most merging basally. Present in (outgroup): Triodopsis fraudulenta and rugosa (Fig. 10). Apomorphous state: 8-10 radiating wall columns, the ventral-most merging basally. Formerly and now present in: Triodopsis fulciden (Fig. 18b). Discussion. The wall pattern of fulciden is unique and problematic. It most closely re- sembles rugosa (not illustrated, but similar to fraudulenta (Fig. 10)), except that the number of pustular columns is approximately halved. This reduction would parallel the reduction in column number suggested in Transformation 32. Transformation 34—Preceding transforma- tions: none. Plesiomorphous state: pore flush with the penial wall. Present in (outgroups): some Camaenidae; and Ammonitellidae, Allogona, Cryptomastix, and Triodopsis (Figs. 9-18). Apomorphous state: pore ventrally subterminal on a large, apical, dorso- ventrally compressed verge which points backward along the everted penis, and with a surface sculpture of cords which continue into 6 or more narrow terminal papillae. Formerly present in: Webbhelix (Fig. 6a), Neohelix (Figs. 2-6), Xolotrema (Figs. 7, 8). Now present in: Neohelix dentifera, albolabris, alleni, major, lioderma, and divesta (Figs. 2-5). Discussion. The hypothesized generalized form of the eastern-triodopsine verge is illus- trated in dorsal view in Figs. 2a, d, 3a, c, 4a, 5d; and in magnified ventral view, showing the subterminal pore, in Figs. 2f, 3b, 4b, and 4e. The verge of the illustrated specimen of lioderma (Fig. 5a, b) is abnormally partially inverted; in undistorted specimens, this speci- es's verge resembles that of divesta (Fig. 5d, e). Despite careful search (e.g., Fig. 11c), no trace of a verge was detected in any species of Triodopsis—the peduncle (Transforma- tions 45 and 46), despite a superficial resem- blance to a verge (e.g. Fig. 14a-d), differs in both position and structural detail. Although Webb's (1961, 1974) hypothesis that Trio- dopsis has secondarily lost the verge cannot be ruled out, this genus’s complete lack of any vestige suggests rather that it never had a verge. Transformation 35—Preceding transforma- tions: 34. Plesiomorphous state: vergic papillae 6 or more, narrow. Present in (outgroups): Neohelix (Figs. 2—5, 6b). Apomorphous state: vergic papillae 2, broad. Formerly and now present in: Webbhelix multilineata (Fig. 6a). Discussion. Narrow papillae appear to be simple extensions of the basic cord-like sub- structure of the verge (e.g. Fig. 2f). These cords and papillae are probably homologous with wall-pustular columns. It seems likely that the broad, paired papillae of multilineata are derived by fusion of the plesiomorphous narrow papillae. The surface of multilineata's verge also appears to be smooth, lacking the cord-like substructure (Fig. 6a). Transformation 36—Preceding transforma- tions: 34. Plesiomorphous state: verge large, termi- nal, longer than wide, bearing 6 or more narrow papillae. Present in (outgroup): Neohelix dentifera, albolabris, alleni, major, lioderma, and divesta (Figs. 2-5). Apomorphous state: Type 1 small verge: subterminal and basally-directed, wider than long, bearing 6-8 narrow papillae. Formerly and now present in: Neohelix solemi (Fig. 6b). Discussion. Small size in the eastern- triodopsine verge is correlated with a sub- terminal position (Types 1 and 3) and with a long penis (slight in Type 3, pronounced in Type 2). Since both a subterminal pore (Char- acter 4) and a long penis (Character 9) seem to be apomorphous conditions (see discus- sions under Transformations 38-41, 44), small size of the verge may also be apo- morphous. In the case of subterminal pore position, this view is supported both by struc- tural differences, indicating convergence, be- tween dorsally subterminal (Type 1) and ven- trally subterminal (Type 3) small verges; and by the following theory on functional morphol- ogy. Because the terminal verge of eastern triodopsines unfolds backward during copula- tion (Webb 1948, 1952, 1954a; see Fig. 1), it is hypothesized that its function is to direct the emitted sperm backward and away from the proteolytic enzymes of the spermathecal bulb. This function seems also to be served, however, by the alternative adaptive “strat- 200 EMBERTON еду” of moving the роге from a terminal to a subterminal position. Therefore, when the pore becomes subterminal, the verge is no longer functional, and therefore becomes vestigial. This theory, however, does not ex- plain the apparent correlation between a long penis and a short type-2 verge. Transformation 37—Preceding transforma- tions: 34. Plesiomorphous state: verge large, termi- nal, longer than wide, bearing 6 or more narrow papillae. Present in (outgroup): Neo- Рейх dentifera, albolabris, alleni, major, lioderma, and divesta (Figs. 2-5). Apomorphous state: Type 2 small verge: terminal, longer than wide, bearing 6 narrow papillae. Formerly and now present in: Xolotrema fosteri and occidentalis (Fig. 8). Discussion. Structurally this Type 2 small verge differs from the Type 3 (Transformation 38) only in being longer than wide and in having two more papillae (Flg. 8c vs. Fig. 7b, d), but further comparative study may prove this latter difference insignificant. There is the possibility, therefore, that Types 2 and 3 small verges are homologous, but because of their different positions (terminal vs. subterminal), it is suggested that they have independently become vestigial for functionally different rea- sons. By similar reasoning, Types 2 and 1 small verges are also convergent. Transformation 38—Preceding transforma- tions: 34. Plesiomorphous state: verge large, termi- nal, longer than wide, bearing 6 or more narrow papillae. Present in (outgroup): Neo- helix dentifera, albolabris, alleni, major, lioderma, and divesta (Figs. 2-5). Apomorphous state: Type 3 small verge: subterminal and apically directed, wider than long, bearing 4 narrow papillae. Formerly and now present in: Xolotrema denotata, ob- stricta, and caroliniensis (Fig. 7). Discussion. Structurally, this Type 3 small verge (Fig. 7b, d) differs from the Type 1 (Transformation 36) only in having two less papillae, but its vastly different position— ventrally subterminal as opposed to dorsally subterminal—leads to the suggestion that Types 1 and 3 are convergent. The sug- gested homoplasy Types 3 and 2 is dis- cussed under Transformation 37. Transformation 39—Preceding transforma- tions: none. Plesiomorphous state: pore terminal. Present in (outgroups): many Camaenidae; Corillidae, Ammonitellidae, Oreohelicidae; all non-east-American-triodopsine Polygyridae; Webbhelix; all Neohelix except solemi; Xolotrema fosteri and occidentalis; and Tri- odopsis fraudulenta, burchi, tennesseensis, complanata, platysayoides, vultuosa, cragini, henriettae, and fulciden (Figs. 2-5, 6a, 8, 10-13, 18b). Apomorphous state: pore dorsally sub- terminal and on a fleshy pedestal. Formerly and now present in: Nehoelix solemi (Fig. 6b). Discussion. The orientation of the verge and the position of the reduced dorsal pilaster clearly indicate the uniquely dorsal position of the pore in solemi. Transformation 40—Preceding transforma- tions: none. Plesiomorphous state: pore terminal. Pres- ent in (outgroups): many Camaenacea; Coril- lidae, Ammonitellinidae, Oreohelicidae; all non-eastern-triodopsine Polygyridae; Webb- helix; all Neohelix except solemi; Xolotrema fosteri and occidentalis; and Triodopsis fraudulenta, burchi, tennesseensis, com- planata, platysayoides, vultuosa, cragini, henriettae, and fulciden (Figs. 2-5, 6a, 8, 10-13, 18b). Apomorphous state: Type 1 ventrally subterminal pore: everted penis shaped like an inverted pear, with the pore ca 1/3-way from the apex and on a verge mounted on a fleshy pedestal. Formerly and now present in: Xolotrema denotata, obstricta, and carolini- ensis (Fig. 7). Discussion. Webb published two illustra- tions of the everted penis of denotata (Webb 1948, Fig. 1a; Webb 1954a, Plate 10, Fig. 11), showing its pyrifom shape and its pore (verge) position. Fig. 7 shows that the dis- sected, uneverted penis also has this shape, that obstricta and caroliniensis are essentially identical to denotata in this respect, and that the pore is on a fleshy pedestal which is not evident in the everted penis. Transformation 41—Preceding transforma- tions: none. Plesiomorphous state: pore terminal. Pres- ent in (outgroups): many Camaenidae; Coril- lidae, Ammonitellinidae, Oreohelicidae; all non-eastern-triodopsine Polygyridae; Webb- helix; all Neohelix except solemi; Xolotrema fosteri and occidentalis; and Triodopsis EASTERN NORTH AMERICAN TRIODOPSINAE 201 fraudulenta, burchi, tennesseensis, com- planata, platysayoides, vultuosa, cragini, henriettae, and fulciden (Figs. 2-5, 5a, 8, 10-13, 185). Apomorphous state: Type 2 ventrally subterminal pore: everted penis shaped like an angled baseball bat, with the pore ca 1/5-way from the apex and indented into the penial wall. Formerly and now present in: Triodopsis vulgata, picea, and claibornensis (Fig. 9). Discussion. Five published illustrations of the everted penis of vulgata (Webb 1959, Figs. 22, 27, 34, 38) show the general shape and position of the pore. Fig. 9 (this paper) shows that the dissected, uneverted penis also has this general shape, that picea and claibornensis are essentially identical to vulgata in this respect, and that there is no sign of a pedestal or verge. Webb's figures indicate a sub-pore protuberance (“tubercle”) which is covered with pustular columns (Webb 1959, Fig. 38); this is not evident in the uneverted penis (Fig. 9). Transformation 42—Preceding transforma- tions: none. Plesiomorphous state: pore terminal. Pres- ent in (outgroups): many Camaenidae; Coril- lidae, Ammonitellinidae, Oreohelicidae; all non-eastern-triodopsine Polygyridae; Webb- helix; all Neohelix except solemi; Xolotrema fosteri and occidentalis; and Triodopsis fraudulenta, burchi, tennesseensis, com- planata, platysayoides, vultuosa, cragini, henriettae, and fulciden (Figs. 2-5, 5a, 8, 10-13, 18b). Apomorphous state: Type 3a ventrally subterminal pore: everted penis shaped like a thick-handled mace, with the pore ca 1/4-way from the apex and above a smooth peduncle. Formerly present in: Triodopsis juxtidens, discoidea, neglecta, and pendula (Figs. 14c, d; 18a, c). Now present in: Triodopsis tridentata, anteridon, hopetonensis, palustris, obsoleta, alabamensis, messana, van- nostrandi, fallax, and soelneri (Figs. 14a-b, 15-17). Discussion. Illustrations of the everted pe- nis of four of the ten species with a Type 3a ventrally subterminal pore have been pub- lished: fallax (Grimm 1975, Fig. ЗВ), hopetonensis (Webb 1959, Figs. 9, 42), tridentata (Webb 1948, Fig. 4; Webb 1954a, Fig. 13; Webb 1959, Figs. 14, 28, 35, 40), and vannostrandi (Webb 1959, Figs. 17, 25a, 43). Figures 14a-b, 15, 16, and 17 (this paper) show that the dissected, uneverted penes of these four species also have this shape and that the remaining 6 species are essentially identical in this respect. Transformation 43—Preceding transforma- tions: 42. Plesiomorphous state: Type 3a ventrally subterminal pore: everted penis shaped like a thick-handled mace, with the pore ca 1/4-way from the apex and above a smooth peduncle. Present in (outgroup): Triodopsis tridentata, anteridon, hopetonensis, palustris, obsoleta, alabamensis, messana, vannostrandi, fallax, and soelneri (Figs. 14a-b, 15-17). Apomorphous state: Type 3b ventrally subterminal pore: mace head large, with the pore ca 2/5-way from the apex. Formerly and now present in: Triodopsis juxtidens (Fig. 14c), discoidea (Fig. 14d), neglecta (Fig. 18a), and репаша (Flg. 18c). Discussion. Illustrations of the everted pe- nis of three of these four species have been published: discoidea (Webb 1959, Fig. 13), juxtidens (Webb 1959, Figs. 15, 41), and neglecta (Webb 1959, Figs. 10, 11, 12). Fig- ures 14c, d and 18a, c (this paper) show that in the dissected, uneverted penes of these three species, the Type 3b ventrally sub- terminal pore is more easily distinguished from the Type 3a by the size of the peduncle (see discussion under Transformation 50) than by the position of the pore, which varies with the state of contraction. Thus, although the pore position of discoidea (Fig. 14d) is as it appears in the everted penis, the pores of juxtidens (Fig. 14c) and neglecta (Fig. 18a) are much closer to the apex than they appear in the everted penes. The pore of pendula (Fig. 18c) is intermediate in position and its peduncle is large, so pendula is interpreted as having a Type 3b ventrally subterminal pore. Since the Type 3b has a very similar penial shape and sub-pore peduncle to the Type 3a, it is assumed to be homologous and derived from the Type 3a by further overgrowth of the dorsal penial wall, making the terminal knob larger and the pore more subterminal. Transformation 44—Preceding transforma- tions: none. Plesiomorphous state: ventral wall free of any pilastral outgrowth. Present in (out- groups): some Camaenidae; Oreohelicidae, Ammonitellidae, and all Triodopsinae except Neohelix solemi (Fig. 49). Apomorphous state: ventral wall bearing a 202 EMBERTON single pilaster. Formerly and now present in: Neohelix solemi (Fig. 6b). Discussion. Because ofits surface sculpture of close, uniform pustules, the ventral pilaster of solemi is convergent on the dorsal pilaster of multilineata (side-by-side comparison in Fig. 6). It almost certainly is apomorphous. Transformation 45—Preceding transforma- tions: none. Plesiomorphous state: 1/3 or less of the total penis length lying between the vaginal opening and the base of the sheath. Present in (outgroups): all eastern Triodopsinae ex- cept Neohelix solemi (Figs. 2-5, 6a, 7-18). Apomorphous state: 1/2 or more of the total penis length lying between the vaginal open- ing and the base of the sheath. Formerly and now present in: Neohelix solemi (Fig. 6b). Discussion. A long basal penis occurs in the polygyrid ashmunellines Cryptomastix, Al- logona, and Ashmunella, but it appears that this character state in Neohelix solemi is not homologous because of its absence in solemis more immediate outgroup, the re- maining species of Neohelix. Transformation 46—Preceding transforma- tions: none. Plesiomorphous state: mid-ventral wall free of sperm groove. Present in (outgroup): all eastern triodopsines except Xolotrema fosteri and occidentalis (Figs. 2-7, 9-18). Apomorphous state: mid-ventral sperm groove present. Formerly and now present in: Xolotrema fosteri and occidentalis (Fig. 19). Discussion. The term ‘ventral sperm groove” refers to the smooth, mid-ventral, raised channel shown (somewhat exaggerat- edly) in Fig. 8a, b, and shown in cross section in Pilsbry (1940, Fig. 473 #7b), even though its function is unknown. It is not conspicuous in any of Webb's illustrations of the everted penis of fosteri (Webb 1952, Figs. 6, 8, 10, 11; Webb 1954a, Fig. 12), but is pronounced enough in the dissected, uneverted penis to be mistaken for the dorsal pilaster—trans- verse folds across the sperm groove in a contracted specimen of occidentalis (Fig. 8b) even produce a superficial resemblance to pilastral lappets (e.g., Fig. 2d). Transformation 47—Preceding transforma- tions: none. Plesiomorphous state: sheath covering 1/2 or less of the upper penis. Present in (out- group): all eastern triodopsines except Xolotrema fosteri and occidentalis (Figs. 2-7, 9-18). Apomorphous state: sheath covering the entire upper penis. Formerly and now present in: Xolotrema fosteri and occidentalis (Fig. 8). Discussion. Penial sheath length varies a great deal depending on the preservational state of the individual snail, so no attempt was made to analyze interspecific differences, with the single exception of this distinct and obviously apomorphous character state. The apparently long sheath of the illustrated spec- imen of Neohelix dentifera (Fig. 2a) resulted from prolapse of the upper penis into the basal penis, evidenced by the pattern of folds at the upper-basal junction—in other dis- sected specimens of dentifera the sheath appeared relatively shorter. The illustrated penis of X. fosteri (Fig. 8a) appears to be normal in length, but that of X. occidentalis (Fig. 8b) is obviously contracted within its uncontracted sheath. Transformation 48—Preceding transforma- tions: none. Plesiomorphous state: upper penis short to long. Present in (outgroup): all eastern trio- dopsines except Triodopsis vultuosa, cragini, and henriettae (Figs. 2-12, 14-18). Apomorphous state: upper penis extremely long and thread-like. Formerly and now present in Triodopsis vultuosa, cragini, and henriettae (Fig. 13). Discussion. The length of the upper, pustu- lated, penis is subject to some individual variation depending on preservational state. Because of this, and because of the high probability of convergence in such a develop- mentally plastic character as overall length, penial length was not considered for system- atic analysis except in this extreme and obvi- ously apomorphous case. The everted penis (illustrated in Webb 1959, Figs. 26b, 26c, 32; and Grimm 1975, Fig. 3c) is conspicuously long and narrow. Transformation 49—Preceding transforma- tions: none. Plesiomorphous state: sub-pore region flat or gradually raised. Present in (outgroup): Webbhelix; Neohelix; Xolotrema; and Trio- dopsis vulgata, picea, claibornensis, fraudulenta, burchi, tennesseensis, com- planata, platysayoides, vultuosa, cragini, henriettae, and fulciden (Flgs. 2-13, 18b). Apomorphous state: sub-pore region erec- tile as a small, fleshy peduncle. Formerly EASTERN NORTH AMERICAN TRIODOPSINAE 203 present in: Triodopsis juxtidens, discoidea, neglecta, and pendula (Figs. 14c, d; 18a, c). Now present in: Triodopsis tridentata, an- teridon, hopetonensis, palustris, obsoleta, alabamensis, messana, vannostrandi, fallax, and soelneri (Flgs. 14a-b, 15-17). Discussion. The peduncle was defined by Webb (1959) as the smooth, fleshy knob just beneath the subterminal pore in the everted penis of hopetonensis, tridentata, vanno- strandi, discoidea, juxtidens, and neglecta. It occurs in two disjunct sizes: small and large. In neither of these is there any substructural detail suggesting homology with the verge, so independent derivation is assumed, with the large peduncle derived from the small peduncle. The sub-pore “tubercle” of vulgata (Webb 1959) is not smooth but pustulose and is not a distinct knob, so probably is not homologous with the peduncle. The peduncle appears to consist of erectile tissue which variously appears in the dissected, uneverted penis as a lobe (Fig. 14a, b, c, d), thickened region (Figs. 15a, c; 16b; 17a, b), or wrinkled sac (Figs. 15b; 16a, c; 18a, c) beneath the pore. The erect small peduncle is best illus- trated in Webb 1948, Fig. 4 (as the “protuber- ance”), and in Webb 1959, Figs. 14, 25a, 40, and 43. Transformation 50—Preceding transforma- tions: 49. Plesiomorphous state: peduncle small. Present in (outgroup): Triodopsis tridentata, anteridon, hopetonensis, palustris, obsoleta, alabamensis, messana, vannostrandi, fallax, and soelneri (Flgs. 14a—b, 15-17). Apomorphous state: peduncle large. For- merly and now present in: Triodopsis jux- tidens, discoidea, neglecta, and pendula (Figs. 14c, d; 18a, c). Discussion. The erect large peduncle is best illustrated in Webb 1959, Figs. 12, 13, 15, and 41. The large peduncle can be distin- guished from the small peduncle in the dis- sected, uneverted penis by its large size, whether it is inflated (Fig. 14c, d) or deflated (Fig. 18a, c). Cladistic analysis The presence or absence of each of the 50 suggested anatomical transformations in each species of eastern triodopsines is pre- sented in Table 1. To simplify cladistic analysis, genitalically identical species were pooled, reducing the number of operational taxa from 40 species to 18 groups (Table 1). Nine of these groups consisted of a single species (W. multilineata, N. solemi, N. albolabris, N. major, N. alleni, T. picea, T. fulciden, T. burchi, and T. platy- sayoides). Each of the remaining multispe- cies groups was temporarily named for one of its better-known species without regard for previously named supraspecific taxa. By far the largest of these (Table 1) was the 7. tridentata group, comprising 10 species (tridentata, anteridon, fallax, obsoleta, palustris, messana, soelneri, alabamensis, vannostrandi, and hopetonensis). One of these, soelneri, was problematic in that it lacks wall pustules and its pilaster lacks sur- face sculpture (Fig. 17b). However, because of its Type 3a ventrally subterminal pore, its small peduncle, and the basic similarity of its pilaster to a tridentata-type without the conical processes, soelneri was considered a highly derived member of the tridentata group. The remaining 8 multi-species groups were non- problematic. The N. dentifera group had three species (dentifera, divesta, and lioderma); the X. fosteri group had two species (fosteri and occidentalis); the X. denotata group had three species (denotata, obstricta, and carolini- ensis); the T. vulgata group had two species (vulgata and claibournensis); the T. fraudu- lenta group had two species (faudulenta and rugosa); the T. tennesseensis group had two species (tennesseensis and complanata); the T. cragini group had three species (cragini, vultuosa, and henriettae); and the T. juxtidens group had four species (juxtidens, discoidea, neglecta, and pendula). Cladistic analysis resulted in a single most parsimonious tree (Fig. 24) with converg- ences in three transformations and with no reversals (consistency index = 48.5/50 = .970). The three convergences were in trans- formations 27, 29, and 30. The convergence in 27 was biologically probable, because two easily identifiable convergences were already known for this character (Transformations 22 and 23). Convergences in 29 and 30 were also not unreasonable, as had been noted in the discussion of Transformation 31. The most parsimonious way to avoid these two homoplasies involved an alternative place- ment of T. platysayoides. This alternative produced a tree with an overall consistency index only slightly lower (48/50 = .960), but it invoked three reversals in the T. platysay- oides lineage (Transformations 16, 19, and 21). This was clearly an inferior alternative to 204 TABLE 1. Presence (1) or absence (0) of 50 hypothesized character-state transformations in each of the 40 species of eastern American triodopsines. The 18 groups consist of species with identical presence/absence patterns. Transformation number Temporary species group 123456789 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Species 0001010000 OOO O ORO AO O ооо 100: NO O0 10) 20. 0) 0) ND 0) 010140) 20.0 10 10 10) 10) 107 07 100 0004 0-20 FOO ROO 1914.07040..020.101.0770:0720 2:0 507505:02:07702077029 OUTGROUPS multilineata solemi OSO 10" 90% 0M0 20 RO 00 M0 00 720200 1 1] 1 Ü 1 1 1 1 1 1 1 1 1 07202.0710792.072.070=02.07:9 1 multilineata solemi 0202072030 020 0000 OOO OO AN OOOO OO OOOO OOO 00550105000 0:30. 07:07 0500-0550! ооо ооо 0:20” :07°0:2:07202.0202.072050720=070770 0070 10 "0 "0 M0 MOMOMOMOS0 707707070 0/2.0,.02:02:07:02.0508.080710:50%07.0°/0 1 ТО OI OO EC 1 1 1 1 1 1 ARO ORO OM O0 MONO оо ооо оо ооо ооо 0% 201) 0) 0" 700 TOMO ORIO NOOO OO (0 0! (0080.0) 01070 1020’ 10} 0310; 100510500: 0 albolabris major albolabris major 0011051 AMOO MO 0 40) AD/HOFS0" 10:00, 501040) 10/50; 202 100. 00/0/0520" 50 191010001 0: 0000000 0,70, ON 0" 70/0700: оо 1 OO #02:.02:07=.05:.072.050=5077.0730=.0=70 alleni alleni 1 1 1 TOF 1100501007100) 10750.10; 40740! 0% 0) 05,0 (01.0 dentifera divesta dentifera 0000700005030: <0 TAO Re1020 707010520) 0 100) 00050000) о dentifera 0000000000 120 1717070705070 020207070 07707107 07.07 07070 lioderma fosteri dentifera fosteri ооо 000 020907070 100010: 1050, 0000000000 00100900000 1 1 1 1 020 (0% <0) 0) 010080 0) 10% 10) 02010" 00 1 1 0 0 оо 00000000 1 1 1 1 1 0': 050, 0 0:00: 0:70" 40:00:50 ONO OO O00) 040 HON 10010 1 1 0 0 0 1 1 1 1 1l 100000000 010707 0 0 0 0 0 0, 10: (0) 0) 0720=0730 00000000 00000000 100000000 occidentalis denotata obstricta fosteri 1 1 1 0 0 0 1 1 1 ооо ооо ооо 0.80.0’ 0000 00000 1 1 1 100000000 denotata denotata denotata vulgata 0; 10: (0: 010; 02:07 0: 10 050 0 0-10 10: 0 0 10/30) 0 0 0.10 0 100000000 100000000 caroliniensis vulgata 0 0° 0 0 0 10 .0: OF 10 ооо ооо 020207070770. 020=:0 1 OO" 0 0 0 0 0 40 0 0 0 0 0 0 020720: 7020707 (0! 0:0) (0) от OO (0: 0) (0) 00010 000-000 00/50 1 100000000 07,0 0 © m I 3 O Zz 1 1 ооо ооо ооо ооо ооо ооо 02:0: О D 10) 10.10) 10) ооо обо ооо OO OOO O70! 207 NOOO CORO O ооо 00" 0% 10°10) 00 00:10 0.0 01.0870) (0750) 900: 10.40; <0 00010} 0 ONO MO 80" 10.0" O PO 0 0) 0000000 02.07.0020. 010) 507.07:0°.0°70707:0702070%0%:0770 2020 1 1 1500.20 41 000000000000 0 0 0 0 1 1 1 1 1 1 1 1 il 1 1 1 oooooo oooooo ===> =) oooooo OOo ooooor эго ooooor oooooo OOrrrr 0000 oroooo rrrrrr oooooo oooooo oooooo © ©. 0.010:0 oooooo0 oooooo (SMS) oooooo oooooo oooooo oooooo rrrrrr © 2 5 a с LIES © DS cie Е 500% SaQ4ga 295958 328935 Sar 222 a 2 OE er Sch © 335% UUDODODO Sessse 53 в 8-6 аз? 1 1 1 1 1 11 1 000.0 0 07:02 02 020 "000 0000000000 0 1 оо оо 1 1 1 1 1 1 1 1 1 4 1 1 1 1 1 1 1 1 1 1 complanata 100000000 0 0 0 burchi tennesseensis burchi 1 100000000 0 0 0 OO MONO O HOMO ORO ROO O ROO 20 007100 1 0 0 0 0 0 1 | 1 1 1 1 1 1 1 1 1 1 1 оо 0 1 оо оо 0 0 оо оо оо оо 100000000 0 0 O platysayoides cragini platysayoides cragini cragini cragini 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 0210100010100 05008020 5050-50 00 010 0, 0101007100. 0.207050 0001000 000000000 0907070207070 70770 077050702070 705%07:0 0000000 00 0000 №00 07070220720707:07.050 1 1 1 1 1 1 1 il il il il 1 1 1 1 1 1 00.0 0 00.0 0 1 1 il 1 1 il 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 il 1 1 1 1 1 1 1 1 1 1 1 1 1 1 12080 12200 100 1210830 1.100 100 100 1 1 1 1 1 1 1 1 1 1 100000000 0 0 0 1503070020107 010). 0) 30510 vultuosa 1 1 1 100000000 00 0 henriettae tridentata anteridon fallax 0 0 0 0 0 0 0 0 0 0 редьки (от) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1101040050000: 70:50:10 tridentata tridentata tridentata tridentata 0500: 4001510 100000000 0 0 0 0220200700770 0 0 0 0 0 0 0 0 0 0 0 То оО) 000 0 020720550 00020 07070750 005070 007200 100000000 0 0 0 020202020720 110 01010 00100000 messana palustris 0207072077070 000790550770 оо оо 0 0 оо 080 оо оо оо 0 0 0 0 0 0 00 оо 1010100000050’ 100 tridentata tridentata tridentata tridentata tridentata tridentata juxtidens juxtidens juxtidens juxtidens 050170" 10/ 40) 0) 0) 0) 0. 0710: 20: 10’ <0) 10" ооо OO 00:0) 107020770 0 00000000 0:05 0400.0; 0050180 0040210720770 702.050 OOO 40". 00) 1070770 OO FONO 0) AO 050" 10 OD 0/000) 10510700 11020 0020701010 {0 "0:0 obsoleta soelneri 00.00.0150 02050700720 09070502070 00100.00 1010 010020100) 000 100000000 0 0 0 alabamensis 100000000 0 0 0 vannostrandi hopetonensis juxtidens 1 1 100000000 0 0 0 1 1 1 1 1 оо 15010000000: 10’ 00 00020 оо 050 050 0 0 оо 100000000 0 0 0 discoidea neglecta pendula 1 1 0208050 0) 30) 0) <0 100000000 0 0 0 0 0 0 100000000 0 0 0 EASTERN NORTH AMERICAN TRIODOPSINAE 205 EN a \ va e N = \ Inlodopsis 2 э | / = \ Ja a | / / = Dal / are | / x \ = 0 | oo | | 5 | | == : O 43 | Webbhelix Xolotrema | SE : | © D 50 | == | ae | Neohelix | а = | Я ГВ > e = | N 5 и \ a eel a © 5 2 | lie y SN | Ww Sn & | | la 4 0 ES 49 | e ig \ 18 DVS ee ee (0) pe o & 2.023 | | © 5 > u 5 Sea г | ‘= О | по & Y | a = > с = | 146 >= o = E o 33 ] | | a O | = © + 30 | EN 5 I: | 36 м о | 27 32 ! | 2 | а | о 18 N / © | q O 41 O ES 1 + \ 444 о — © = ae A A E an 7 à © 11455 = 12 | | & = j Ей = 8 = 46 | 929 = 1 128 311) 2 93) a see ae 421 (QE 47 2 / | ® zu N poe = peak | 401 | | № = / И 7 4 MINE MN / т ‚| ia % / ZUR = ; CES О 41 16. y \ 22, pa \ he oh | р x / Хх 57 \ y \ 35, ~ 3 4 24 \ Sal | a N 13 Fi \ / \ Jf 34 о omen FIG. 24. “Anatomy Tree”: a phylogenetic hypothesis for the eastern American triodopsines based оп penial morphology (50 character-state transformations shown in Figs. 19-23). This is the single most parsimonious tree generated by PAUP, with a consistency index of .970. 206 EMBERTON having convergences in Transformations 29 and 30, so Fig. 24 was decidedly the best cladogram to fit the data. Thus Fig. 24 is the “Anatomy Tree”. The branch lengths of this cladogram are scaled to the number of transformations they con- tain, so are a rough indicator of the degree of evolutionary change in penial morphology. ALLOZYMIC ANALYSES Complete electrophoretic results are pre- sented in Table 2. In this table, each allele (electromorph) is represented by its migration distance on the gel in mm relative to the control (Mesodon zaletus from Monte Sano, Alabama: FMNH 214772 and 214773), the migration distance of which was arbitrarily set at 100 mm. Seventy-four alleles were de- tected in the eastern triodopsines and 9 in the outgroup Allogona profunda. The most vari- able loci were Lap and Pgm, with 12 and 11 alleles. Sordh and Me each had 8 alleles; lcd had 7; Gpi had 5; Mdh-1, Mdh-2, Gd-1, Gd-2, and Sod-1 each had 4; Sod-2, Got-2, and Mpi each had three; Got-1 had two; and Pgd was the only monomorphic locus. Heterozygosity within populations was ex- tremely low. Most populations were mono- morphic for all but two or three loci, with a maximum of three alleles per locus (Table 2). Twenty triodopsine alleles were absent from the outgroup Mesodon and therefore were presumed apomorphous. The distribu- tions of these alleles among triodopsine spe- cies are listed in Table 3. Twelve alleles were restricted to a single species; the remaining 8 were present in two to 10 species. Phylogenetic analysis produced the “Al- leles Tree” presented in Fig. 25. This cladogram is the consensus of the first 50 trees with a maximum, identical consistency index generated by PAUP; its numbered transformations refer to the alleles listed in Table 3. The Alleles Tree contains one con- vergence (Lapya between picea and tennes- seensis) and one reversal (loss of Icdgg in hopetonensis). Comparison of 50 trees showed that both this homoplasy and this reversal are robust, occurring in 100% and 88% of the trees respectively. Phenetic analyses of the two independent subsets of the allozymic data are presented in Figs. 26 and 27. The first of these, the “Wagner-1 Tree”, comprising 32 species evaluated over 16 loci, has a cophenetic correlations of .897, indicating only mild dis- tortion of the original genetic distance matrix. The “Wagner-2 Tree” (21 species, 8 loci) has a similarly high cophenetic correlation of .883. CONSENSUS PHYLOGENY To aid in comparing the Anatomy Tree (Fig. 24), the Alleles Tree (Fig. 25), the Wagner-1 Tree (Fig. 26), and the Wagner-2 Tree (Fig. 27), each was labeled in a consistent manner: genera and outgroups were enclosed by dashed lines. The trees were weighted for comparison. In the Anatomy Tree, 3 out of the 50 transfor- mations (.06) showed reversal or conver- gence, whereas the Alleles Tree had 2 out of 20 (.10). Dividing these gave a “reliability” of anatomical over allozymic data units of 1.6. The number of data units for each tree was: Anatomy 50, Alleles 20, Wagner-1 73, and Wagner-2 28. Multiplying the morphological data units by 1.6 and dividing all by 75 and rounding gave the following weights: 1.0 for the Anatomy Tree, 0.3 for the Alleles Tree, 1.0 for the Wagner-1 Tree, and 0.4 for the Wagner-2 Tree. The four genera—Neohelix, Triodopsis, Webbhelix, and Xolotrema—are distinct and coherent throughout all four Trees (Figs. 24-27). The four minor exceptions to this general pattern are readily resolved. (1) In the Alleles Tree, N. albolabris appears in Trio- dopsis due only to the presence of Icdgg (Transformation 10) in one of its three popu- lations (albolabris-2), which could easily be a homoplasy. (2) Also in the Alleles Tree, T. messana is grouped in Neohelix due only to its possession of Icdgg (Transformation 11), which could be a homoplasy. (3) In the Wagner-1 Tree, T. burchi groups within Neohelix; in the equally weighted Anatomy Tree, however it pairs with the T. tennes- seensis group, well within Triodopsis, and the occurrence of this pairing in the Alleles Tree gives it greater weight than a Neohelix posi- tion for T. burchi; its genetic similarity to Neohelix could be due either to homoplasy in alleles or to retention of plesiomorphous al- leles. (4) The isolated position of X. fosteri within Neohelix in the Wagner-2 Tree (weight 0.4) is outweighed by its firm position within Xolotrema in both the Anatomy and Alleles Trees (combined weight 1.3); X. fosteri does not occur in the Wagner-1 Tree. With these EASTERN NORTH AMERICAN TRIODOPSINAE 207 exceptions resolved, there remains no doubt of the robustness of the four genera. Neohelix, Webbhelix, and Xolotrema to- gether constitute a monophyletic group, ac- cording to the Anatomy Tree and both Wagner Trees. According to the Wagner-1 Tree, Webbhelix is the most plesiomorphous genus; in both the Anatomy and Alleles Trees, it is concordant with Neohelix and Xolotrema, but retains more plesiomorphous charcter-states than either of these genera. This combined evidence that Webbhelix is the most plesiomorphous genus of eastern triodopsines outweighs the Wagner-2 Tree’s placement of it as sister group to Xolotrema. The grouping of Neohelix albolabris, N. major, N. alleni, and the highly derived N. solemi in the Anatomy Tree receives enough verification in the two Wagner Trees for ac- ceptance as it stands. In the Wagner-1 Tree, alleni, major, and solemi cluster together but are isolated from albolabris. However, in the Wagner-2 Tree, albolabris (two populations) does cluster with alleni (three populations) and major (two populations), but this cluster is isolated from a single population of alleni (alleni-4); solemi is missing from this tree. Thus, except for one slightly errant population in each of the two Wagner Trees, the evi- dence is consistent for an albolabris-major- alleni-solemi cluster, with solemi the most highly derived member both anatomically (Fig. 24) and electrophoretically (Fig. 26) as indicated by its long branch length in each of these trees. These four species comprise the Neohelix albolabris group, which is revised in the following section. The Neohelix dentifera group (dentifera, divesta, and lioderma) is clearly coherent and isolated from the other Neohelix anatomically (Fig. 24) and electrophoretically (Figs. 26, 27). There is no evidence in any of these trees as to the relationships of the three species within the group, but dentifera ap- pears to have a less apomorphous form of enlarged basal pustules, as discussed under anatomical Transformation 4, somay be plesio- morphous within the group. The dentifera group's position primitive to the albolabris group is evident in the Anatomy, Wagner-1, and Wagner-2 Trees, and is only contradicted by alleni and lioderma sharing Icdgg (Trans- formation 11) in the Alleles Tree, which could easily be a convergence and is strongly out- weighed by the evidence of the other trees. The anatomical division of Xolotrema (Fig. 24) into the fosteri group (fosteri and oc- cidentalis) and the denotata group (denotata, obstricta, and caroliniensis) is only partially supported by the electrophoretic data. The pair denotata and obstricta is linked by a unique derived allele (Me;os) in the Alleles Tree (Fig. 25) and is also tightly linked in both Wagner Trees (Figs. 26, 27), but caroliniensis groups no closer to this pair than does fosteri in the Alleles Tree or occidentalis in both Wagner Trees. However, since caroliniensis is represented electrophoretically by only a single specimen (Table 2), its relative position in these trees should not be considered very precise. Complete electrophoretic data were lacking for fosteri (Table 2), so it does not occur in the Wagner-1 Tree. In the Wagner-2 Tree, fosteri is strongly isolated not only from occidentalis but from the remainder of Xolotrema in general, but this placement based on genetic distance is shown cladisti- Cally to be aberrant in the Alleles Tree: fosteri shares one derived allele, (Sordhgg) with the remainder of Xolotrema and another derived allele (Me:55) with the denotata group. The Alleles Tree also shows that occidentalis is separated from fosteri by its lack of Me; and its unique possession of @рнот. Despite these partial discrepancies with the electrophoretic trees, the anatomical disjunc- tion between the fosteri and denotata groups is so extreme (seven transformations in the anatomy Tree) and the penial sculpture within each of these two groups is so cohesive (Figs. 7 and 8), that there can be no doubt of their separation. The fact that there has been little electrophoretic differentiation between the fosteri and denotata groups suggests that their anatomical distinctions have evolved rel- atively recently. There is no clear evidence from any of the trees as to which of these two groups is the more plesiomorphous. Within Triodopsis there are some discrep- ancies among the four trees of such magni- tude that the original dissections were reex- amined and several interpretive errors were detected. The Consensus Tree (Fig. 28), therefore, differs from the Anatomy Tree (Fig. 24) more for this genus than for any other, and contains some redefinitions of species groups. In the Wagner-1 Tree (Fig. 26), the T. vulgata group (vulgata and claibornensis), T. picea, the T. fraudulenta group (fraudulenta and rugosa, but represented only by fraudu- lenta), and T. platysayoides form a single, shallowly rooted group with no differentiation into subgroups except for a shallow pairing of EMBERTON 208 (oL )ı6 (o8)ooı (06)26 (02)86 (02)26 00, (oz)eoı (01)66 (oz)o01 (0€)soL 001 001 (ol )o0oL 86 201 001 86 001 96 201 Ss 058712 ejoajbau 1 66 96 001 86 16 001 00+ 001 00: LOL 001 001 66 16 201 € 6r8bLz eJeauI ¡nu М (54001 (62`)56 (21) (se )rot 201 16 (82')26 16 — (88`)001 001 06 66 501 96 001 001 96 201 + 14:1 284 Buessauw Y (ol )z6 (05)s6 00! 00: (06)66 (05)86 16 001 101 00: 66 501 001 001 001 16 504 Ss 056712 Jolew ‘N (58`)96 001 96 66 (s1)/6 16 001 101 00: 66 501 96 16 001 96 001 OL #rerız виларо! N (05`)001 (02 )26 (06)06 (05 )poL zo! (08')66 16 16 001 00: (or )ooı 001 201 86 86 001 96 ool 5 ererız suapııxnl ‘1 (sz')001 (SZ )pot 201 86 GL 16 001 001 06 001 20} 001 86 001 96 201 + — sisuauojadoy `1 (11:)56 (80)96 (SZ)Z6 (26)56 (z9)oı (05)06 (213001 00: (26)001 (80)66 (80)6 16 16 (€€)Z01 (0S)001 oo! 20+ 86 (e8')ro! 5'66 96 0: 9 22812 ejuajnpney 1 001 96 66 16 16 001 101 00: 66 501 001 (0/)00! 001 96 20: $5 18:1 204 в5амр N (Or )16 (09)s6 (09)s6 001 96 (0y)6 (Or )26 16 001 201 00: 66 501 201 001 001 96 201 $5 608715 влэщиар 'N (21) 16 (v6')001 (zz)oot (90)s6 (v6')501 (90 )pOL 00, (90)201 (P6)26 16 001 LOL 06 001 501 50: (90`)80+ 66 16 86 6 908512 ввоиар`Х (26`)56 (80`)001 001 16 66 16 16 001 06 66 201 201 001 001 96 201 9 608 1Z ben ‘1 (05 )ooL (S2')26 (05')vo1 201 16 (6/66 16 001 001 06 66 201 001 201 001 96 tot 2 208912 ejeuejdwoo | (05 )S6 (21:)06 00! 201 66 (05‘)/6 16 001 00: (88)001 66 201 86 001 001 96 Е + 008715 SISUBUJOGIE}O `1 (05`)001 001 001 201 16 16 001 101 06 66 501 501 (05’)501 66 16 86 1 = SISUS/UI/OJEO “Xx (sz)001 001 001 16 (s2)eoL 16 001 201 00; 66 zo! 00! 201 001 96 го 9 1649715 Iyoung “1 (58`)001 (19°) 46 (65`)06 (21 )voLl zo! 16 (££)66 16 001 001 (29)001 66 201 86 66 001 96 zOL € €6Lb12 uopuajue ‘1 (s2)68 (er )26 (ZL) LOL 001 00! 66 (85)66 16 001 201 00: 66 501 96 001 001 16 (8s)eol 9 бр! IU8IIE 'N (08`)/6 (02)16 001 00: (02)66 (08')S6 16 001 LOL 06 66 501 001 001 00! 16 50 Ss 0z6r12 sugejogje ‘М (21 )S6 (80 )26 (64`)001 201 66 (v8)66 (er) 2e (80')rot 201 66 — (80')601 16 — (89')001 001 06 86 201 001 001 001 96 201 9 1649715 SISU&UEQE]E | (60`)86 (01')98 (oe )ooL (01 )eoı (5/`)56 (06)66 (08`)56 (04°) 701 00: (se)soı (s1')86 16 001 oot (so)olL 66 — (01)601 06 vol (oz)00! 201 76 Ol — врипзола виобоу sajalıe ids idw de ubd4 2-100 1109 г-ро$ |-ро$ 2-p9 L-P9 po} SN Z-UPW L-UPW yps0S N Joquinu salsads 12101 anbojeyed $1907 HNW4 ‘00, Je sıydıowouow элэм suoiyeindod ¡je yoıym 10} 'pBy ‘Sn90] yI9L е $! ajqe} ay} и! umoys JON ‘ePUNjOJd виобо|/\у dno16yno ay, jo зиоцетдо4 ом} snıd ‘sauisdopou} иеэнэш\ ua]sea jo $э09э4$ gE Buisudwoo suoyendod $9 10} eyep awAzoly ‘2 JIGVL 209 EASTERN NORTH AMERICAN TRIODOPSINAE 001 00! (64`)001 (sz’)poL 001 001 (08')001 (0z')voL 00! 00! 00! 001 001 (s4)001 (se’)vol (05`)96 (0s')001 00! 56 001 001 001 001 (05`)001 (08 )201 (06')001 (oL')voL 96 96 00! co! 00! col col co! cot col ool co! col cot (05)001 (05,201 00+ 00+ (55`)96 (S2)66 (05')s6 (0S")26 16 26 (or )96 (09)66 66 66 16 86 (20)68 (£6)86 (pL )86 (98')001 66 (29)v6 (£€)46 16 16 (05')26 (0b)v6 (01 )26 (21 )66 (88°)001 (05`)/6 (0/`)66 96 16 96 16 96 16 16 = (05')56 (05)26 16 (SZ°)26 (Sz )66 16 (01)16 (0€ )56 (09`)/6 16 16 00! 96 16 96 16 96 16 (s2’)ı6 (61`)56 16 (cz )98 (G2)L6 = (21')66 (21001 (99')501 16 (1s)s6 (v9)26 (67)86 (9€')00L (64:)56 (12)26 16 (88)96 (21:)86 16 (21)56 (99)46 (21)001 16 (05')56 (05')86 16 (8e )s6 (29)26 16 (02 )56 (02 )26 (01 )ooL 16 SL 16 (Ov')S6 (0t')86 (oz')001 16 16 16 (02')S6 (04)46 (OL )66 16 00! 00+ 00! 00! 00+ 001 001 00+ 001 001 001 001 16 001 001 001 (21)46 (e8)001 00! 16 (06°)26 (01 )o0L 00+ 00+ 001 001 201 20} 201 00+ 00! 001 00! (£6 )001 (20')401L 00! 00! 901 001 (0/`)001 (05') 101 001 00+ 201 201 001 06 06 06 (S2')06 (64`)001 (9=`)06 (6/`)00+ 001 001 001 (€9°)06 (Z€°)001 06 66 06 (9€)06 (b9')001L (56`)06 (20')001 06 (€8°)06 (21 )o0L 00L 00+ 001 06 06 06 06 66 501 001 66 501 col 66 cot 00! 66 col 86 86 501 96 00+ 501 001 66 501 001 66 = 86 = = 06 66 co! 0+ 001 co! 86 66 co! 86 00+ col 00+ 66 col 00+ 66 501 001 86 co! 0+ 00! co! 86 00+ col 86 66 col 00+ 00+ 60+ 501 00+ 501 501 001 00+ sol 00+ 66 00+ 00+ 00+ 00+ 00+ 00! $701 86 (20')86 (98)001 (20')voL 86 66 col 001 001 (02)86 (oL )ooL (02 )voL 86 001 001 (06')SOL (or )8oL 00! 00! 66 001 00! 00! 00! 00! 00! 00! 00! (05')56 (0s')001 00! 00! 00! 00! 00! 00+ 00+ 00+ 00+ 00+ 66 66 96 96 16 96 96 (06)16 (01)96 96 96 96 96 (S2')L6 (Sz')96 201 96 96 96 96 96 16 96 16 (S2')L6 (sz')96 96 16 16 cot co! 86 col co! (01)68 (oz')LoL (OZ )£0t co! 00! 00! 501 501 76 col LOL co! LOL LOL S LOL LOL 0+ (ZOOL (88')201 co! 86 86 el8blz oL8pLz 508912 70871 < 961912 Ol6rlz 606r 12 806715 616715 188P12 v88v le 088715 9L8r12 r98rL2 Сбт! С L98r12 098rL2 658r12 [58715 958912 ps8ple z—eisanıp N г—виэдигр ‘м г—ввюоиер ‘Хх Z—wibe ‘1 г— иориаше ‘| 5 шае | у ша | eee | ee с ша | $— 592/041 N e—sugejog/e N г—грипуола euobo/|y BSONIINA 1 веб/пл ‘| IPUEHSOUUEA ‘1 ввиарщ | SISU88SS8UUS) “/ 1w8jos 'N sapıofesAlejd “1 Paid ‘1 eynpuad ‘| sujsnjed ‘1 SIJeJuapi990 ‘X EIOUJSIO “Xx A + ___ == _z_—______ IA 2 — K <— ЕЕ АЕ» sajal|e ido Id dey ud 2.109 1109 2-pos роб 2-P9 L-PO pol aw Z-UPWN L-UPW ypos N Jequinu $90905 a. _ ___ __ _ —— — — — _ _ —_ ""--"Î- = anbojeyed $1907 HNW4 TA eee ee ee (=>) pues . © (panuyuo)) “2 3719v1 EASTERN NORTH AMERICAN TRIODOPSINAE 211 TABLE 3. Alleles in eastern American triodopsines which are considered apomorphous (i.e., absent from their outgroup Mesodon), and the species in which they were detected. Allele Locus Species 1. Sordh 101.5 solemi 2.Sordh 98 caroliniensis, denotata, fosteri, obstricta, occidentalis 3. Sordh 89 alleni 4. Mdh-2 99.5 fraudulenta 5. Me 108 denotata, obstricta 6. Me 105 caroliniensis, denotata, fosteri, obstricta 7.Me 102 burchi, complanata, tennes- seensis 8. Me 97 lioderma 9. Icd 104 henriettae 10. [са 98 albolabris, anteridon, claibornensis, fulciden, juxtidens, pendula, picea, vannostrandi, vulgata 11. [са 96 alleni, lioderma, messana 12.Gd-1 101 multilineata 13. Gd-2 97 major 14. Sod-2 106 solemi 15. Got-2 105 neglecta 16. Pgm 96 tridentata 17. Pgm 75 hopetonensis, pendula 18. Lap 94 picea, tennesseensis 19. Gpi 107 occidentalis 20. Gpi 99 multilineata fraudulenta and picea. This pattern contrasts strongly with that seen in the Anatomy Tree (Fig. 24), in which the fraudulenta group plus fulciden is quite isolated from the vulgata group, picea, and platysayoides, with the tennesseensis group plus burchi intervening. Selected dissections of vulgata, picea, fraudulenta, rugosa, fulciden, and platysay- oides were repinned and compared. It was found that Fig. 10 misrepresents the pilastral structure of fraudulenta, which is actually much more like that of picea (Fig. 9b), but is variable within a single population (FMNH 214822), with some individuals (e.g. #6) showing secondary fusion of pustules which only superficially resembles the knob-less, spurred pilastral polygons of the tridentata, juxtidens, and cragini groups. Also, the pore of fraudulenta is Type-2 ventrally subterminal, which, in retrospect, is apparent in Fig. 10. Thus, fraudulenta is actually anatomically closest to picea, and the basal merging of its ventral-most wall columns is homoplasic with this condition in rugosa and fulciden, which have the tridentata-like, rather than the picea- like pilastral sculpture, and which still appear to have a terminal pore. This dissolves the previous fraudulenta group (fraudulenta and rugosa) of Table 1 and Figs. 24-27, leaving rugosa by itself, and redefines a new two- species fradulenta group as picea and fradulenta, with fraudulenta its most derived member, and closest anatomically to the vulgata group. On re-inspection, it appears that platysayoides's unique pilastral sculpture could be directly evolved from the more undif- ferentiated sculpture of the pilaster vulgata and claibornensis, as was anticipated in the discussion of Transformation 15. These re- vised anatomical decisions are reflected in the positions of the (revised) vulgata group and platysayoides in Fig. 28, which, unlike the Anatomy Tree (Fig. 24), is compatible with the Wagner-1 Tree (Fig. 26). Only one of these reconsidered species (vulgata) occurs in the Wagner-2 Tree (Fig. 27), and its pairing there with tridentata is at variance with the consensus reached in Fig. 28, but carries little relative weight. The Alleles Tree (Fig. 25) offers no strong contradiction to the Consen- sus Tree's arrangement of the vulgata group (it lacks platysayoides), for fraudulenta only lacks one allele (Icdgg) which is shared by vulgata, picea, and claibornensis, and the value of this allele is apparently lessened by its patchy distribution among members of the tridentata and juxtidens groups as well. The T. tennesseensis group (tennes- seensis and complanata) is consistently sup- ported by the three trees (Anatomy, Alleles, and Wagner-1) which contain both species. Its grouping with 7. burchi, evident in the Anatomy and Alleles Trees (combined weight = 1.3), is strongly contradicted by the Wagner-1 Tree (weight = 1.0), in which burchi appears between Webbhelix and the Neohelix-Xolotrema lineage. Dissections of T. burchi were reexamined, but no evidence was found for reinterpreting it anatomically; there- fore in the Consensus Tree (Fig. 28) burchi is retained next to the tennesseensis group with a question mark denoting its problematic sta- tus. The position of the tennesseensis group- ?burchi lineage as sister group to the vulgata group platysayoides lineage is clear-cut on both the Anatomy and Wagner-1 Trees (com- bined weight = 2.0) and is only mildly con- tradicted by the grouping of tennesseensis with tridentata and juxtidens in the Wagner-2 Tree (weight = 0.4). 212 EMBERTON Mesodon = OUTGROUP = “+ multilineata 5 ee ge Fe EG SE == solemi re = er: Yo = | we = + major | > х = | | x | + allenı J | 7 - lioderma a = © Se messana >" y occidentalis 74 WO / foster! In \ denotata pa \ J © a > obstricta о Xcaroliniensis _ ‚gcomplanta TS, 7 EN + — tennesseensis \ © \ | о \ | burchi | ' \ \ fraudulenta SS | A x albolabris / re vulgata = =. 1 > Diced a | + > claibornensis u / un | fulciden aR anteridon >) vannostrandı | —— hopetonensis E ES 2 / = pendula / / juxtidens ий у +— neglecta / a 7 | / ras henriettae a tridentata „7° = dent 22 FIG. 25. “Alleles Tree”: a phylogenetic hypothesis for the eastern American triodopsines based on allozymes, with Mesodon as outgroup. The 20 uniquely derived alleles are listed in Table 3. This tree is the consensus of 50 trees of equal and maximum parsimony generated by PAUP, with a consistency index of .950. The phylogeny of the remainder of Tree (Fig. 24). The tridentata group (tri- Triodopsis is fairly consistent among the four dentata, anteridon, fallax, obsoleta, hope- trees, and generally supports the Anatomy tonensis, palustris, messana, alabamensis, EASTERN NORTH AMERICAN TRIODOPSINAE 213 A-PROFUNDA à outgroup 1 3 515 А T-ANTER IDON RAG I T-VULTUOSA T-HOPETONENSIS ate T-JUXTIDENS || LL. T-TRIDENTATA T-PENDULA T-VANNOSTRANDI T-MESSANA T-CLAIBORNENSIS T-FRAUDULENTA j= 13 T-PICEA | | T-PLATYSAYOIDES T-VULGATA T-COMPLANATA T-PALUSTRIS _ _ T-DIVESTA er ho roo ooo ho ooo 444 Triodopsis ) T-TENNESSEENSIS > T-ALBOLABRIS ? Xolotrema ‘T-CAROLINIENSIS 1 Т-РЕМОТАТА T-OBSTRICTA j T-OCCIDENTALIS_” T-MAJOR T-ALLENI 1 ‚ Neohelix T-LIODERMA / / == NS T-MULTILINEATA > -- 4-42 - a + FIG. 26. “Wagner-1 Tree”: а distance-Wagner tree for 32 species of eastern American triodopsines, with Allogona as outgroup. Computed from the Prevosti distance matrix based on 16 allozymic loci (Table 2, upper half). The cophenetic correlation is .897. The branch lengths are optimized. vannostrandi, and soelneri) and the juxtidens group (juxtidens, discoidea, pendula, and neglecta) cannot be distinguished electro- phoretically (Figs. 25-27); nevertheless their separation is accepted on both anatomical and biogeographic grounds. The differences in pore position and peduncle size between the tridentata and juxtidens groups, although not extreme and not always easy to detect in dissection, are disjunct (see previous discus- sions under Transformations 42, 43, 49, and 50). Biogeographically, none of the four spe- cies of the juxtidens group (0%) show range overlap (Fig. 49), whereas there are approx- imately 10 range overlaps among the 10 species of the tridentata group (10/ (10 take 2) = 10/45 = 22%); this is consistent with the view that the juxtidens group is more recently evolved and less differentiated. The lack of electrophoretic differentiation between the tridentata and juxtidens groups is interpreted as evidence that their split was relatively recent. The T. cragini group (cragini, vultuosa, and henriettae) is tightly coherent in both the Anatomy and Wagner-1 Trees (Figs. 24, 26), although henriettae is missing from the latter. In the Wagner-2 Tree (Fig. 27), henriettae and two populations of vultuosa cluster closely, with cragini more distantly connected and with fulciden intervening, which is dis- cussed below. The consensus of these three trees (the Alleles Tree contains no informa- tion beyond henriettae's possession of the unique, derived allele Icdjo4) is that the cragini group is well defined and that cragini is 214 EMBERTON A-FROFUNDA-2 — ES o ——outgroups M-ZALETUS — T-ALBOLABRIS-2 5 T-ALBOLABRIS-3 N T-ALLENI-5 \ | T-ALLENI-2 Neohelıx T-ALLENI-3 T-MAJOR-2 T-MAJOR-3 __ 7 T-FOSTERI ao [pao] ы = -—— T=DIVESTA=2 — =~ == = T-DENTIFERA-2 Sr Xolotreme T-ALLENI-4 / | EN "3 raies T-DENOTATA-2 \ | T-OBSTRICTA-2 N = T-OBSTRICTA-3 | T-OCCIDENTALIS-2 _/ ыы» T-MULTILINEATA-2 > T-ANTERIDON-2 Sr ==- nn Webbhelix RAG 2 = T-FULC IDEN a T-HENRIETTAE dl zu T-VULTUOSA-2 E ] T-VULTUOSA-3 \ T-HOPETONENSIS-2 T-PALUSTRIS-2 | Triodopsis | T-JUXTIDENS-2 T-TRIDENTATA-2 | T-TRIDENTATA-4 T-TRIDENTATA-& T-TRIDENTATA-S T-TENNESSEENSIS -2 T-TRIDENTATA-3 T-VULGATA-2 =-+----+----+----+ +--- - + +----+-- ho + +----+--- = 05 D, 12 0.18 o 24 0.30 0.37 0.43 0 49 0.55 0.61 +----+----+ ----+----+----+----+----+ FIG. 27. “Wagner-2 Tree”: а distance-Wagner tree for 3 additional species and 29 additional populations of 18 of the species of eastern American triodopsines represented in the Wagner-1 Tree (Fig. 26), with Allogona profunda and Mesodon zaletus as outgroups. Computed from the Prevosti distance matrix based on the 8 allozymic loci for which all populations had complete data (Table 2, lower half). The cophenetic correlation is .883; the branch lengths are optimized. probably the most plesiomorphous species of the group. The position of the cragini group is outside the tridentata-juxtidens lineage in the Anatomy Tree (weight 1.0), shallowly within this lineage in the Wagner-1 Tree (weight 1.0), and outside this lineage in the Wagner-2 Tree (weight 0.4). The consensus, therefore, is the separation of these lineages as sister groups. Complete electrophoretic data were lacking for T. fulciden, and none were available for 7. rugosa, so the only test of their paired position in the Anatomy Tree (in which rugosa equals the “fraudulenta group”), is fulciden's position in the Wagner-2 Tree. Its close relationship to the cragini group in this tree in entirely sup- portive of its being a sister group to the cragini-tridentata-juxtidens lineage (Fig. 24), but could also denote its being a sister group of the cragini group alone. In the Consensus Tree (Fig. 28), therefore, the fulciden-rugosa pair (named the rugosa group) is positioned as in the Anatomy Tree (Fig. 24), but with a question mark. Since data for rugosa are EASTERN NORTH AMERICAN TRIODOPSINAE Xolotrema Neohelıx fosterı group denotata group solemi Comes = © Webbhelix albolabris ` major allenı a dentifera group > vulgata® PE claibornensis multilineata 22 OUTGROUPS picea Triodopsis \ \ trıdentata group 43 50 vultuosa henriettae cragını 42 49 tennesseensis complanata fulcıden 48 — rugosa fraudulenta 30 32 platysayoides 27с burcht Ce 26 FIG. 28. “Consensus Tree”: the robust phylogenetic hypothesis for the eastern triodopsines, representing the weighted consensus of the Anatomy, Alleles, Wagner-1, and Wagner-2 Trees (Figs. 24-27). scant, this pairing is also uncertain, so rugosa’s position in is also marked with a question mark. The completed Consensus Tree for the eastern American triodopsines is presented in Fig. 28, labeled with the suggested anatomi- cal charcter-state transformations as reas- sessed in the light of electrophoretic evi- dence. The Consensus Tree carries two convergences in transformation 27 and a single convergence in transformation 29. It represents a robust consensus between electrophoretic and anatomical data. CONCHOLOGICAL VARIATION Conchological illustrations of eastern Ameri- can triodopsine species are presented in Figs. 29-45. Also included for comparative purposes is an illustration of Allogona profunda, the only eastern ashmunelline (Fig. 46a-b). Shell variation of eastern trio- dopsines has been thoroughly discussed by Pilsbry (1940), Vagvolgyi (1968), and Grimm (1975). An illustrated key to most of the species is contained in Burch (1962). It is important to remember when identifying any eastern American triodopsine that many spe- cies of the polygyrine genus Mesodon have closely convergent shells. REVISION OF THE NEOHELIX ALBOLABRIS GROUP The following classification is proposed, based on analyses of penis and shell. The complete systematic review is presented in Appendix B. 216 EMBERTON _10 mm FIG. 29. Shells. a-b. Neohelix dentifera (Binney, 1837). FMNH 214810 #8. c-d. Neohelix albolabris albolabris (Say, 1816). FMNH 214920 #14. albolabris group albolabris albolabris albolabris (Say, 1816) albolabris bogani Emberton, new subspecies major (Binney, 1837) alleni group alleni alleni alleni (Sampson, 1883) alleni fuscolabris (Pilsbry, 1903) solemi Emberton, new species Genitalic analysis Species identification of each of the 46 populations (Fig. 47) was made by comparing its penial morphology with Figs. 2d, 3, 4, and 6b. Differences among albolabris, alleni, ma- jor, and solemi in upper penial sculpture were extremely stable over their geographical ranges, which made identifications easy and straightforward. For 39 of the populations, penial sculpture was examined by dissecting one to three specimens per population; for two populations (numbers 5, 32), specimens had partially everted their penes in the drown- ing jar, so could be identified without dissec- tion; the remaining 5 populations (numbers 16-19, 27) were identified from published anatomical illustrations (Simpson, 1901; Pilsbry, 1940; Webb, 1952, 1954a). The results of the penial-morphological measurements are presented in Table 4 as ranges over the three measured populations (one specimen per population). For each of the 7 variables, value ranges are underlined which do no overlap the value range of albolabris albolabris. Because of the small sample sizes and non-normal distributions, these differences were not tested for statisti- cal significance. Between the two subspecies of albolabris there was no difference detected in any ofthe 7 penial-morphological variables, therefore they were pooled for cladistic anal- ysis. From Table 4, seven new penial-morpholo- gical transformations (transformations A-G) are proposed for a cladistic analysis of albolabris, alleni, major, and solemi, using EASTERN NORTH AMERICAN TRIODOPSINAE 217 FIG. 30. Shells. a-b. Neohelix alleni alleni (Sampson, 1883). FMNH 214913 #B. c-d. Neohelix major (Binney, 1837). ЕММН 214933 #H. Webbhelix and the dentifera group as outgroups (Fig. 28). The format used is the same as used previously for transformations 1-50. Transformation A—Preceding transforma- tions: none. Plesiomorphous state: penis, pilastral lap- pets, and wall pustules all moderate in size: penis length variable with median ca 13-14 mm, pilastral lappets less than .3 mm high, wall pustules less than .15 mm wide. Present in (outgroups): Webbhelix multilineata (penis and pustules), N. dentifera group (penis, lap- pets, upper pustules), albolabris, alleni, solemi (penis, lappets, basolateral pustules). Apomorphous state: penis, pilastral lap- pets, and wall pustules all large: penis length invariable at 17 mm, pilastral lappets higher than .5 mm, wall pustules wider than .16 mm. Formerly and now present in: major. Discussion. The penis of major (Fig. 4a) looks much like a hypertrophied version of albolabris's (Fig. 2d), so it appears that its longer penis and larger pilastral lappets and wall pustules are intercorrelated features of a general enlargement. Of the outgroups, all have a moderate penis length. Only albolabris (Fig. 2d), alleni (Fig. 3a, c), and solemi (Fig. 6b) have the plesiomorphous lappet height, since Webbhelix multilineata (Fig. 6a) lacks lappets and the dentifera group (Figs. 2a, 5a, d) has them doubled (Transformation 4). The plesiomorphous wall pustule size is un- modified only in W. multilineata (Fig. 6a), albolabris (Fig. 2d), and alleni (Flg. 3a, c); the 218 EMBERTON FIG. 31. Shells. a-b. Neohelix lioderma (Pilsbry, 1902). ЕММН 214844 #A. c-d. Neohelix divesta (Gould, 1848). FMNH 214813 #A. wall pustules are enlarged basally (Transfor- mation 23) in the dentifera group (Figs. 2a, 5a, d) and enlarged everywhere but baso- laterally (Transformation B, below) in solemi (Fig. 6b). Transformation B—Preceding transforma- tions: none. Plesiomorphous state: all wall pustules moderate and approximately equal in size, less than 0.15 mm wide. Present in (out- groups): W. multilineata (Fig. 6a), albolabris (Fig. 2d), alleni (Fig. За, с). Apomorphous state: all but the baso-lateral wall pustules large, wider than .20 mm. For- merly and now present in solemi (Fig. 6b). Discussion. Large wall pustules, 5-6 per 1.3 mm, occur in both major and solemi (Table 4, column 3), but this appears to be due to con- vergence. The large wall pustules of solemi (Fig. 6b) are neither uniformly sized nor ac- companied by a large penis and large pilastral lappets, as they are in major (Fig. 4a). Transformation C—Preceding transforma- tions: none. Plesiomorphous state: pilastral lappets about as wide as the wall pustules. Present in (outgroup): alleni (Fig. 3a, c), solemi (un- modified baso-lateral wall pustules: Fig. 6b). Apomorphous state: pilastral lappets ap- proximately twice as wide as the wall pus- tules. Formerly and now present in: albolabris (Fig. 2d), major (Fig. 4a). Discussion. According to Table 4 (third and fourth columns), the pilastral lappets are slightly less dense than the columns of wall pustules in alleni (15-18 vs. 18-22 per 2.6 mm), and slightly more dense than the en- larged, central columns of wall pustules in solemi (14-15 vs. 8-12 per 2.6 mm). In contrast, the pilastral lappets in albolabris are EASTERN NORTH AMERICAN TRIODOPSINAE 219 FIG. 32. Shells. a-b. Webbhelix multilineata (Say, 1821). FMNH 214848 #2. c-d. Neohelix solemi Emberton, new species. FMNH 214936 #1. about twice as dense as the columns of wall pustules (8-11 vs. 16-24 for albolabris albolabris, and 9-14 vs. 20-22 for albolabris bogani), and the same is true of major (4-5 vs. 10-16 per 2.6 mm). Since pilastral lappets seem to be derived from wall pustules by lateral fusion (see discussion under Transfor- mation 5), it is assumed that the equal den- sity, or equal width, seen in alleni and solemi is plesiomorphous. Close examination of Figs. 2d and 11a reveals evidence that the double-sized lappets of albolabris and major resulted from vertical fusion: one of albo- labris's lappets has a lateral groove, and two of major's lappets have pieces of lappets angled beneath them laterally. This rather obvious character-state transformation was overlooked in the previous analysis. Transformation D—Preceding transforma- tions: none. Plesiomorphous state: verge large, greater than .12 the penial length. Present in (out- groups): W. multilineata (Fig. 6a), dentifera group (Figs. 2a, 5a, d), albolabris (Fig. 2d), major (Fig. 4a). Apomorphous state: verge moderate in size, less than .09 the penial length. Formerly and now present in: alleni (Fig. 3a, c). Discussion. The unique, small verge of solemi, which is only .01—.05 the penial length (Table 11, column 4) was already used as Transformation 36 (Type 1 small verge). As discussed under that Transformation, the moderately sized terminal verge of alleni is not homologous with that of solemi; instead, it is structurally (Fig. 3b) very similar to the 220 EMBERTON FIG. 33. Shells. a-b. Xolotrema denotata (Férussac, 1821). FMNH 214806 #1. c-d. Xolotrema caroliniensis (Lea, 1834). FMNH 171142 #B. e-f. Xolotrema obstricta (Say, 1821). FMNH 214852 #1. verges of albolabris (Fig. 2f), major (Fig. 4b), and divesta (Fig. 5e), which differ from it only in their larger size. Transformation E-—-Preceding transfor- mations: none. Plesiomorphous state: pilaster moderate in breadth, .06-.12 the penial length. Present in (outgroups): W. multilineata (Fig. 6a), denti- fera group (Figs. 2a, 5a, d), albolabris (Fig. 2d), major (Fig. 4a), alleni (Fig. За, с). Apomorphous state: pilaster narrow, .02-.04 the penial length. Formerly and now present in: solemi (Fig. 6b). Discussion. The uniquely narrow dorsal pi- laster of solemi (table 4, column 5) probably EASTERN NORTH AMERICAN TRIODOPSINAE 221 ly, pe ull UL, 5 mm FIG. 34. Shells. a-b. Xolotrema fosteri (F. С. Baker, 1932). ЕММН 214817 #15. c-d. Xolotrema occidentalis (Pilsbry 8 Ferriss, 1907). FMNH 214856 #5. indicates that it is vestigial. The indistinct lappets on this dorsal pilaster (Transformation 6) and the apparently compensatory ventral pilaster (Transformation 44) support this view. It seems likely that when solemi evolved a (dorsally) subterminal pore (Transformation 39), the adaptive significance of which is hypothesized in Appendix A, both its verge (Transformation 36) and its dorsal pilaster (Transformation E) were no longer functional and so became vestigial. Transformation F—Preceding transforma- tions: none. Plesiomorphous state: retractor muscle's origin distant from the penial apex, .4—.7 the penial length along the vas deferens. Present in (outgroups): W. multilineata (Binney, 1851, pl. 8, fig. 2), dentifera (Pilsbry, 1940, fig. 491), divesta (Pilsbry, 1940, fig. 492; Solem, 1976, fig. 4), albolabris, and major (Table 4, column 6). Apomorphous state: retractor muscle's ori- gin close to the penial apex, .1-.3 the penial length along the vas deferens. Formerly and now present in: alleni, solemi (Table 4, col- umn 6). Discussion. In the absence of detailed dif- ferences suggesting convergence, it is sug- gested that this apomorphous character state is homologous in alleni and solemi. Transformation G—Preceding transforma- tions: none. Plesiomorphous state: vas deferens long, over 4 times as long as the penis. Present in (outgroups): dentifera (Pilsbry, 1940, fig. 491), (Pilsbry, 1940, fig. 492), albolabris (e.g. Pilsbry, 1940, fig. 488) (Table 4, last column), major (Table 4, last column). Apomorphous state: vas deferens short, about 2 times as long as the penis. Formerly and now present in: alleni, solemi (Table 4, last column). Discussion. This hypothesized transforma- tion is somewhat problematic. A short vas deferens occurs in multilineata (Binney, 1851, pl. 8, fig. 2), in some divesta (Solem, 1976, fig. 4a), and in juvenile albolabris (Webb, 1954a, pl. 7, fig. 29); which suggests that the 222 EMBERTON 5 mm FIG. 35. Shells. a-b. Triodopsis vulgata Pilsbry, 1940. ЕММН 214884 #A. c-d. Triodopsis picea Hubricht, 1958. FMNH 214860 #15. ef. Triodopsis claibornensis Lutz, 1950. ЕММН 214800 +A. long vas deferens of dentifera, divesta, albolabris, and major may be apomorphous rather than plesiomorphous. However, in keeping with the hypothesis that the dentifera group is the immediate outgroup of the albolabris group (Fig. 28), outgroup compari- son dictates that the short vas deferens of alleni and solemiis apomorphous. For lack of evidence to the contrary, it is assumed homol- ogous in these two species. Cladistic analysis With the addition of Transformations A-G to those used in the Anatomy Tree (Transfor- mations 5-9, 24, 36, 39, 44, 45), there were a total 17 penial-morphological transformations with which to construct a cladogram. These yielded a single, parsimonious cladogram, free of convergence and reversal, which is illustrated in Fig. 48. The only change this EASTERN NORTH AMERICAN TRIODOPSINAE 223 FIG. 36. Shell. a-b. Triodopsis fraudulenta (Pilsbry, 1894). FMNH 214822 +A. represents from the Consensus Tree (Fig. 28) is in grouping alleni and solemi in a mono- phyletic lineage, designated the alleni group. Shell analysis The complete shell measurements are pre- sented in Table 6, and are referred to in the systematic review of the albolabris and alleni groups (Appendix B). The 8 conchological variables, when calculated from the raw mea- surements (Table 6) by the methods de- scribed in Table 5, and when standardized and subjected to discriminant analysis, yielded a discriminant function (Table 7) which correctly classified to subspecies or species 44, or 94%, ofthe 47 analyzed shells. The three misclassified shells are marked by asterisks in Table 6. One shell of alleni alleni (population 3, specimen #8) was misclassified as alleni fuscolabris, with a pos- terior probability of membership in that subspecies of .66; its probability of correct classification was .34. Two shells of albolabris albolabris (population 11, specimen #4; and population 12, specimen #17) were misclas- sified as solemi, with posterior probabilities of membership in that species of .60 and .53 respectively; their probabilities of correct clas- sification were .24 and .47, with the former specimen also having a .16 probability of misclassification in albolabris bogani. Thus, overall, the discriminant function (Ta- ble 7) was quite successful in differentiating the 6 taxa by the 8 shell variables (Table 5). The fact that 8 of the 9 total shells of alleni were correctly classified to subspecies, and that the ninth shell had a .34 probability of correct classification, is persuasive evidence of the conchological differentiation between the western alleni alleni and the disjunct east- ern alleni fuscolabris (Figs. 46, 50). Likewise, the fact that all of the 21 shells of albolabris which were correctly classified to species were also correctly classified to subspecies, testifies to the conchological differentiation between the eastern albolabris albolabris and the western albolabris bogani (Figs. 47, 49). The discriminant function's marginal failure to differentiate two shells of albolabris albolabris from solemi points out the necessity of dis- section for reliably identifying albolabris-al- leni-group snails along the northern Piedmont and Coastal Plain (Figs. 47). Revised classification The systematic review of the albolabris and alleni groups is presented in Appendix B. In it, extensive use was made of the tabulated discriminant function (Table 7). Because all 8 variables were standardized (to mean = 0, standard deviation = 1), they received equal weight in the analysis. Therefore, in the discriminant function (Table 7), the total range of a variable is an indication of its value in taxonomically discriminating among the shells. Thus, for example, GLOSSY's range from — 10.1 to 8.0 indicates that it is a more powerful discriminator than RELSPIRE, with its smaller range of —1.7 to 1.5. This dis- criminant function is biased to some degree by the included taxa and the included shells, such that, for example, a reanalysis compar- ing only albolabris albolabris and albolabris bogani would produce a different discriminant function emphasizing different variables. In short, Table 7 is not the final or the best word on how to tell these taxa apart by shell characters; it is better viewed as an interim guideline. 224 EMBERTON 5 mm 5mm FIG. 37. Shells. a-b. Triodopsis burchi Hubricht, 1950. FMNH 214797 #10. c-d. Triodopsis tennesseensis (Walker & Pilsbry, 1902). FMNH 214864 #7. e-f. Triodopsis complanata (Pilsbry, 1898). Hubricht 17932 #A. GENERAL SUPRASPECIFIC REVISION The supraspecific revision of the eastern triodopsines based on the consensus phy- logeny (Fig. 28) is listed below and is pre- sented in detail in Appendix C. This revision groups the 40 species into 4 genera, 14 spe- cies groups, and 8 species subgroups. Most of the species groups are the same as those temporarily introduced in Table 1 and used throughout the Anatomical and the electro- phoretic Trees (Figs. 24-27). Changes from Table 1 are establishment of the albolabris and alleni groups, expansion of the vulgata group to include fraudulenta and picea, deletion of the fraudulenta group, and creation of the rugosa group (rugosa and fulciden). The de- cision was made reluctantly to submerge EASTERN NORTH AMERICAN TRIODOPSINAE 225 FIG. 38. Shell. a-b. Triodopsis platysayoides (Brooks, 1933). FMNH 214861 #2. Webb's subgenera Wilcoxorbis Webb, 1952 (= the fosteri group) and Haroldorbis Webb, 1959 (= the cragini group), and section Shelfordorbis Webb, 1959 (= the vulgata group), because retaining them would have required coining 11 additional subgenera in order to keep the taxonomy hierarchically con- sistent. The genera and species groups are arranged alphabetically here; in Appendix C they are arranged phylogenetically. Neohelix von Ihering, 1892 albolabris group albolabris (Say, 1816) major (Binney, 1837) alleni group alleni (Sampson, 1883) solemi Emberton, new species dentifera group dentifera subgroup dentifera (Binney, 1837) divesta subgroup divesta (Gould, 1848) lioderma (Pilsbry, 1902) Triodopsis burchi group burchi Hubricht, 1950 cragini group cragini Call, 1886 henriettae (Mazyck, 1877) vultuosa (Gould, 1848) fallax group alabamensis subgroup alabamensis (Pilsbry, 1902) hopetonensis (Shuttleworth, 1852) vannostrandi (Bland, 1875) fallax subgroup fallax (Say, 1825) messana Hubricht, 1952 obsoleta (Pilsbry, 1894) palustris Hubricht, 1958 soelneri (Henderson, 1907) Juxtidens group Juxtidens subgroup discoidea (Pilsbry, 1904) juxtidens (Pilsbry, 1894) neglecta subgroup neglecta (Pilsbry, 1899) pendula Hubricht, 1952 platysayoides group platysayoides (Brooks, 1933) rugosa group fulciden? Hubricht, 1952 rugosa Brooks & MacMillan, 1940 tennesseensis group complanata (Pilsbry, 1898) tennesseensis (Walker 1902) tridentata group anteridon (Pilsbry, 1940) tridentata (Say, 1816) vulgata group fraudulenta subgroup fraudulenta (Pilsbry, 1894) picea Hubricht, 1958 vulgata subgroup claibornensis Lutz, 1950 vulgata Pilsbry, 1940 Webbhelix multilineata (Say, 1821) Xolotrema denotata group denotata (Ferussac, 1821) caroliniensis (Lea, 1834) obstricta (Say, 1821) fosteri group fosteri (F. C. Baker, 1932) occidentalis (Pilsbry & Ferris, 1907) & Pilsbry, Table 8 compares this classification with those of Pilsbry (1940) based on shell mor- phology; Webb (1952, 1954, 1959), based on 226 EMBERTON 5 mm FIG. 39. Shells. a-b. Triodopsis vultuosa (Gould, 1848). FMNH 214887 #7. c-d. Triodopsis cragini Call, 1886. ЕММН 214803 #2. e-f. Triodopsis henriettae (Mazyck, 1877). FMNH 214824 #2. reproductive anatomy and behavior; and Vagvolgyi (1968), based on shell morphology. Of the 40 species recognized here, Pilsbry classified 33, Webb 15, and Vagvolgyi 38. Ir- relevant ofthe number of species, this revision most closely resembles the classification of Pilsbry (1940) as modified by Hubricht (1985)—the major difference lies in the group- ing of species within Triodopsis (Table 8). The systematics of the Triodopsis fallax group presented in Appendix C and Table 9 is that of Grimm (1976), as discussed in Appen- EASTERN NORTH AMERICAN TRIODOPSINAE 227, N WWA 7S — И i! = ARA NN y DO 1 ) я = / IRA; NE Yfr ! S \ NÉE ) ИА С / yA / SE / / / / й / € FIG. 40. Shells. a-b. Triodopsis tridentata (Say, 1816). FMNH 214876 #4. c-d. Triodopsis anteridon (Pilsbry, 1940). FMNH 214796 #19. dix D. Division of the juxtidens group into Juxtidens and neglecta subgroups is based on shell morphology—see Appendix C. PATTERNS OF GENITALIC EVOLUTION The ranges of the 40 species of eastern triodopsines are presented in Fig. 49. These maps were compiled from Hubricht (1985), with corrections for the Neohelix albolabris and alleni groups. The maps were used to compare the de- gree of difference in penial morphology of sister taxa with their geographic range rela- tionship. The results based on 25 compari- sons (Table 9) are: sister taxa with virtually identical penes generally have peripatric ranges, those slightly different are generally allopatric, those moderately different are sympatric, but those greatly different are parapatric. The tests for population-level reproductive character displacement are summarized in Table 10. In none of these 12 tests was there any detectable difference in penial morphol- ogy between allopatric and sympatric popula- tions. PATTERNS OF SHELL EVOLUTION Fig. 50 shows the phylogenetic pattern of shell morphology among all known living spe- cies of eastern North American triodopsines. A general evolutionary pattern is of con- chological stasis within genera. In general, each genus is characterized by a distinct shell form: Neohelix and Webbhelix shells are large, globose, and toothless (Figs. 29-32); Xolotrema shells are medium-sized and de- pressed, with a blade-like parietal tooth and a long basal lamella (Figs. 33, 34); and Trio- dopsis shells are small, subglobose, and tridentate (Figs. 35-45). Shell convergences among these con- chologically distinct genera are rare. Webb- helix and Neohelix shells are similar appar- 228 EMBERTON 5 mm FIG. 41. Shells. a-b. Triodopsis juxtidens (Pilsbry, 1894). ЕММН 214841 #7. cd. Triodopsis discoidea (Pilsbry, 1904). FMNH 214811 #А. ently because they share the plesiomorphous shell morphology seen in some of their outgroups (Fig. 50). One Neohelix species— dentifera (Fig. 29a—b)—converged slightly on Xolotrema by its low spire, strong parietal tooth, and suggestion of a basal lamella. Two lineages of Triodopsis—platysayoides (Fig. 38) and tennesseensis group (Fig. 37c-f)— converged, apparently independently, on Xolotrema by evolving enlarged, depressed shells with reduced outer lip teeth. These convergences are not very close, hence the shells are easily assigned to the correct ge- nus. Within a genus, the distributional pattern of shell characters among species groups and among species is generally mosaic, with many cases of convergence or parallelism. Within Neohelix, a parietal tooth crops up in both the albolabris group (some albolabris populations—see Pilsbry, 1940) and the dentifera group (dentifera, Fig. 29a); a baso- columellar lip node appears in both the albolabris group (major, Fig. 30c) and the alleni group (alleni, Fig. 30a); and a glossy yellowish periostracum arises in all three spe- cies groups (albolabris hubrichti, alleni alleni, and lioderma). Within Xolotrema, a carinate shell occurs convergently in both the fosteri group (occidentalis, Fig. 34d) and the deno- {аа group (obstricta, Fig. 33f). Within Triodopsis, enlarged, flat shells with weak dentition appear in both platysayoides (Fig. 38) and the tennesseensis group (Fig. 37c-f); a squared-off parietal tooth occurs indepen- dently in the three species-groups vulgata (Figs. 35a, c, e, 36a), rugosa (Fig. 45c-), and juxtidens (Figs. 41c, 45a, e); a buttressed palatal tooth of identical appearance shows up in both rugosa of the rugosa group (not illustrated) and anteridon of the tridentata group (Fig. 40c); and toothless apertural lips occur convergently in three species in three disparate lineages: platysayoides (Fig. 38a), soelneri (Fig. 44c), and in rare populations of tridentata (see Pilsbry, 1940); a glossy periostracum appears in the tennesseensis group (complanata, Fig. 37f) and twice, ap- EASTERN NORTH AMERICAN TRIODOPSINAE 229 N \\ \ FIG. 42. Shells. a-b. Triodopsis hopetonensis (Shuttleworth, 1852). FMNH 214827 #22. c-d. Triodopsis palustris Hubricht, 1958. FMNH 214857 #1. e-f. Triodopsis obsoleta (Pilsbry, 1894). Hubricht 10300 #A. parently independently, in the fallax subgroup (palustris, Fig. 42d; and soelneri, Fig. 44d). Other examples of intrageneric shell converg- ences among species groups and species could be cited, but these are the most con- spicuous. Shell variation within species is summa- rized in Table 11, which compiles Vagvolgyi's (1968) data with taxonomic corrections. Shell size, spire height, umbilical relative width, and whorl count vary greatly. Diameter range covaried significantly with sample size, whether expressed as number of lots (r = 0.67, d.f. = 23) or as total number of shells (r = 0.64, 4.1. = 23). For wide-ranging, well- sampled species from all four genera (e.g., W. multilineata, N. albolabris, X. fosteri, and T. tridentata), diameter ranged approximately 70%. Maximum diameter range was found in Juxtidens: 95%. 230 EMBERTON FIG. 43. Shells. a-b. Triodopsis alabamensis (Pilsbry, 1902). FMNH 214791 #A. c-d. Triodopsis messana Hubricht, 1952. FMNH 214846 #A. e-f. Triodopsis vannostrandi (Bland, 1875). FMNH 214880 #11. DISCUSSION Genitalic analysis The penis proved to be an outstanding tool for the erection of a cladistic hypothesis for the eastern triodopsines. Its morphological diversity yielded an unprecedented number (for pulmonates) of character states, and its sculptural complexity permitted the detection of many convergences. The suggested character-state transforma- tions (Figs. 19-23) varied considerably in plausibility. The thoroughness of their docu- mentation, however, establishes an objective baseline for future, more enlightened revi- sions. The choice of PAUP (Swofford, 1983) for EASTERN NORTH AMERICAN TRIODOPSINAE 231 E 5 mm FIG. 44. Shells. a-b. Triodopsis fallax (Say, 1825). Hubricht 10209 XA. c-d. Triodopsis soelneri (Henderson, 1907). FMNH 159040 +A. constructing cladograms has recently re- ceived support by Fink's (1986) comparisons of available software: PAUP was clearly more reliable than PHYLIP for finding the shortest trees. The Anatomy Tree generated from the triodopsine data by PAUP (Fig. 24) is remark- able for its high consistency index and for its uniqueness as the single most parsimonious cladogram. See Fink (1986) for an introduc- tion to alternatives to the maximum-parsim- ony approach to cladogram construction used here. Allozymic analysis The number of snails electrophoresed per population averaged 3.9 (standard deviation 2.5). Although larger sample sizes would cer- tainly have been preferred, small samples are generally sufficient for detecting systematic affinity from allozymes (Sarich, 1977; Gorman 8 Renzi, 1979; Shaffer, 1984; also compare the “exemplar method” of Sokal & Sneath, 1963). Buth (1984) evaluated the available meth- ods for applying electrophoretic data to systematics studies. Of his concluding list of 8 recommendations—(1) sample intraspecific geographic variation, (2) list raw data, (3) code allozyme data with the locus as the character for cladistic analysis, but also consider distance methods, (4) state the procedure used for ordering the transforma- tions used for cladistic analysis, (5) construct minimum-length Wagner trees for cladistic analysis, because of their freedom from the assumption of constant evolutionary rates, (6) use outgroup comparison to determine the polarities of transformations, (7) check homoplasious steps in the constructed cladogram for possible introgressive origins, and (8) separately and identically analyze two independent data sets and examine them for congruency—all but numbers 3 and 7 were followed in this paper. Instead of Buth's number 3 recommendation to code loci as the cladistic characters, individual alleles were coded, thereby using the “independent allele model” introduced by Mickevich & Johnson (1976), “[treating] each allele as a 232 EMBERTON ES Е FIG. 45. Shells. a-b. Triodopsis neglecta (Pilsbry, 1899). FMNH 214850 #А. c-d. Triodopsis fulciden Hubricht, 1952. FMNH 214823 #A. e-f. Triodopsis pendula Hubricht, 1952. FMNH 214859 #14. binary character to be scored merely as present or absent” (Mickevich & Mitter, 1981). This scoring method has the disad- vantages—probably minor—of occasionally being biologically unrealistic by hypothesizing intermediates which lack alleles at a locus and by making the assumption that alleles are indeed always independent (Mickevich & Mitter, 1981). These disadvantages of the independent allele model are outweighed by its advantage of producing unquestionably ordered transformations (present/absent)—in this it differs importantly from coding the locus as the character, for which method “the problem of ordering is currently the most critical [unsolved] issue in this research area” (Buth, 1984). Of the several systems for coding independent alleles the present/ absent system used in this paper is the method of choice “[when] the cladistic infor- mativeness of frequency changes is suspect or demonstrably small” (Mickevich & Mitter, 1981), both of which conditions apply to the eastern-triodopsine data set (Table 2). Buth’s number 7 recommendation to check for introgressive origins of homoplasies was not feasible for the triodopsine data set because the degree of interspecific hybridization EASTERN NORTH AMERICAN TRIODOPSINAE 233 FIG. 46. Allogona profunda (Say, 1821). FMNH Uncat. #3. a-b. Shell. TABLE 4. Measurements of penial morphology for the 6 species and subspecies of the Neohelix albolabris and alleni groups, expressed as ranges over three measured dissections (one per population). Underlined ranges are those disjunctly different from albolabris. No. Retractor M. Vas pilastral No. Verge Pilaster distance deferens Penis lappets columns length: breadth: from penis length:? Species or length per of pustules penis penis apex: penis penis subspecies (тт)! 2.6 тт рег 1.3 тт? length length length length alleni plus fuscolabris 10-18 15-18 9-11 .08-.09 .09-.12 1-.3 2.1-2.4 albolabris 10-16 8-11 8-12 .15-.21 .06-.12 4-.7 5.1-5.6 bogani 10-13 9-14 10-1 .14—16 .07-.11 4-.7 4.6-5.7 major 17 4-5 5-8 .12-.15 .08-.11 4-.5 4.2—4.5 solemi 12-17 14-15 4-6 .01-.05 .02-.04 1-3 1.8-2.5 “From junction with atrium to internal apex. 2Adjacent to pilaster about two-thirds the distance from penial internal apex. 3From ‘У’ of the atrium to insertion at penial apex. within this group of snails is very poorly known. The second part of Buth’s (1984) number-3 recommendation to “[consider] the interpreta- tion of distance treatments Felsenstein (1984) advanced” was followed by transforming the electrophoretic data into genetic distances, then applying a clustering algorithm. The ad- vantages of the Prevosti genetic distance coefficient used in this analysis are its sim- plicity and its O-to-1 range; the “[single] theo- retical objection . . . [that it] gives equal weight to frequency differences throughout the range from O to 1” (Wright, 1978) does not seem critical for this data set, in which allelic fre- quencies can vary greatly within a species, in which heterozygosity is extremely low, and in which species usually differ by fixed, alterna- tive alleles rather than by frequency differ- ences among the same alleles (Table 2). The Prevosti coefficient was chosen over two which predominate in the literature—those of Nei (1972, 1978) and Rogers (1972). The failure of Nei’s distance coefficient to satisfy the triangle inequality (Farris, 1981), although probably not critical theoretically (Felsenstein, 1984), can produce the practical disadvan- tage of negative branch lengths, “an undesir- able and biologically uninterpretable result for a coefficient used in reconstructing phylo- genies” (Buth, 1984). Roger's distance coef- ficient satisfies the triangle inequality (Buth, 1984), but has the disadvantage of being “a mixed concept depending on the degree of [allelic] fixation as well as degree of difference in such a way that two populations with fixa- 234 EMBERTON albolabrıs from North Carolına ıntroduced са 1 FIG. 47. Geographical distribution of the 46 populations studied for the revision of the Neohelix albolabris group. tion of different alleles are considered farther apart than ones where both are heterallelic even though they have no common allele” (Wright, 1978). In retrospect, a better choice than the Prevosti coefficient would have been the Cavalli-Sforza & Edwards (1967) chord measure advocated by Wright (1978) and Felsenstein (1984). For clustering taxa from genetic distance data, three algorithms are most commonly used: UPGMA (unweighted pair-group with arithmetic averaging: Sokal 4 Sneath, 1963; EASTERN NORTH AMERICAN TRIODOPSINAE 235 allenı group albolabrıs group SEEN. allenı albolabrıs dentifera group FIG. 48. Revised phylogenetic hypothesis for the Neohelix albolabris group, based on the addition of penial-morphological transformations A-G, with the dentifera group as closest outgroup. This cladogram justifies splitting the albolabris group into the albolabris and alleni groups. Sneath & Sokal, 1973; Nei, 1978), Fitch- per, the distance-Wagner procedure was cho- Margoliash (Fitch & Margoliash, 1967), and sen because it provides the best fit to the distance-Wagner (Farris, 1972). For this pa- original distance data, it is free from the 236 EMBERTON TABLE 5. Shell variables used in discriminant analysis (Table 7) among the 6 species and subspecies of the Neohelix albolabris and alleni groups. Method of measurement or calculation (from Table 6) Variable Abbreviation Striae per 2.6 mm STRIAE Brownness BROWN Glossiness GLOSSY Relative height of spire RELSPIRE Whorl expansion rate WHRLEXPN Relative size of baso-columellar RELNODE node Relative width of apertural lip RELLIP Relative pre-apertural deflection RELDEFL of body whorl Striae count (number of striae per 2.6 mm on upper surface of the end of the fifth whorl). Color rank (1 to 7, ranging from light yellow to dark brown). Sheen rank (1 to 6, ranging from glossy to dull). Shell height divided by shell diameter. Shell diameter divided by whorl number. Width of apertural lip at midnode, divided by width of apertural lip at its narrowest basal point. Width of apertural lip at periphery, divided by shell diameter. Body-whorl depth behind lip, minus pre-deflection body-whorl depth, divided by shell diameter. assumption of constant evolutionary rates (unlike UPGMA and other agglomerative methods, and unlike the “Fitch-Margoliash” method of Prager & Wilson, 1978), and it is computationally feasible (unlike the true Fitch-Margoliash method) (Farris, 1981; Swof- ford, 1981; Tateno et al., 1982; review in Buth, 1984). Although Farris (1981) later re- pudiated his own distance-Wagner method (Farris, 1972) and all other methods of infer- ring phylogenies from distances, as inherently unable to reconstruct branch lengths consis- tent with evolutionary events, Felsenstein (1984) showed that “[Farris's] major criticisms of these methods lose their force” when an alternative, statistical (rather than absolute) interpretation of branch lengths is used. A drawback of the distance-Wagner algorithm is that the final tree topology is to some extent dependent on the order in which the distance data are read in for computation (e.g., Swof- ford & Selander, 1981). This drawback could have been (but was not for this paper) cor- rected by using rearrangement algorithms which shuffle and refeed the data, but even without such safeguards, “the distance- Wagner procedure is likely to be much more effective than [the Fitch-Margoliash proce- dure]” (Swofford, 1981). Robustness of the consensus phylogeny Anatomical and electrophoretic data sets were remarkably congruent. The final Con- sensus Tree (Figs. 28, 50) is virtually identical to the Anatomy Tree (Fig. 24), requiring little modification to comply with the electro- phoretic trees (Figs. 25-27). Thus the electrophoretic data validated the cladistic analysis of penial morphology. There is slight circularity in this statement, because Triodopsis fraudulenta and T. neglecta were anatomically reevaluated due to discrepan- cies between electrophoretic and anatomical data. This circularity is trivial in the context of the entire phylogeny, but it vouches strongly for the importance of comparing independent data sets to guard against misinterpretations. Genitalic evolution: pattern and process The results of 25 comparisons between sister taxa, presented in Table 9, showed a counter-intuitive trend. Sister taxa with virtu- ally identical penial morphologies generally have peripatric ranges, those slightly different are generally allopatric, those moderately dif- ferent are sympatric, but those greatly differ- ent are parapatric. Several shortcomings of the data need to be considered before interpreting this result. First, two of the taxon pairs are questionable (and are so marked in Table 9), because of inadequate data or a discrepancy between anatomical and electrophoretic data. Second, those ranges were called parapatric or peripatric which appear in Fig. 49 to be con- tiguous or only slightly overlapping, without an intervening geographical barrier; because these maps are imprecise, some of interpre- tations may be incorrect. Third, current range relationships may have little relevance to those under which the anatomical changes actually evolved, because of distortion by Plio-Pleistocene and perhaps earlier climatic and vegetational changes. EASTERN NORTH AMERICAN TRIODOPSINAE 237 If one assumes correct phylogeny and cor- rect interpretation of temporally stable ranges, then four hypotheses concerning ev- olutionary processes can be proposed. (1) Peripheral isolates generally do not differen- tiate. (2) Vicariant isolates generally differen- tiate slowly. (3) Differentiation due to repro- ductive character displacement is moderate at most. (4) Extreme differentiation is rare, rapid, and occurs in isolates. Each of these will be discussed in turn. Peripheral isolates generally do not appear to differentiate, because all 8 examples of peripatric sister taxa have identical genitalia (Table 9). This is consistent with the overall lack of intraspecific geographic variation found in the eastern triodopsines, as men- tioned previously. There may be some natu- ral-selected inertia to change in genital mor- phology due to founder effects or population bottlenecks, because these events are prob- ably common in triodopsine species. For ex- ample, populations of Neohelix albolabris are patchily distributed and ephemeral, with draught and predation periodically producing local die-backs or extinctions, followed during favorable years by rapid build-ups from survi- vors or founders (McCracken, 1976). There would be a selective advantage to groups of N. albolabris in which flush-crash populations conserved ancestral genitalic morphology and were thus able to restore their genetic diversity by remating with other populations during flushes. Indeed the penial morphology of this species is remarkably uniform over its very wide geographic range (Fig. 47, Table 4). Vicariants generally appear to differentiate slowly, because all four examples of allopatric sister taxa differ only slightly in their genitalia (Table 9). This hypothesis is further sup- ported by two species of Neohelix (albolabris and alleni), both of which have two subspe- cies that have been genetically isolated by the Mississippi River (Figs. 47, 49) for at least 20,000 years (see Delcourt & Delcourt, 1981)—equivalent to at least 10,000 genera- tions (McCracken, 1976)—and that have evolved significant shell differences (Table 7), yet are virtually identical in penial sculpture (Fig. 3, Table 4). Differentiation due to character displace- ment appears to be moderate at most, be- cause all 5 examples of sympatric sister taxa had only moderate genitalic differences, and because none of the 6 examples of sister taxa showing greater than moderate differences were sympatric. All 5 sympatric pairs are probably microsympatric. In three of them (T. soelneri vs. both T. messana and T. hopetonensis; T. tridentata vs. T. juxtidens; and N. albolabris vs. both N. dentifera and N. divesta) the taxa have been found within crawling distance of each other with no evi- dence of hybridization (personal observation); it is likely that the other two examples (N. albolabris bogani vs. N. alleni alleni, and W. multilineata vs. N. albolabris) also come into contact, with no hybrids known. In all of these cases, the penial differences were primarily in the dorsal pilaster, with occasional differ- ences in the wall pustulation as well. These differences may be sufficient in themselves for mate recognition, but there other possible isolation mechanisms that prevent sister-spe- cies hybridization and hence that diminish the role of reproductive character displacement in causing morphological divergence. Despite Webb's (1948, 1952, 1959, 1961) conclusion that penial sculpture is the basis of mate- recognition in triodopsines, interspecific matings do occur under laboratory conditions (Grimm, 1975), even between such genitali- cally different species T. tridentata and T. vulgata (Webb, 1948). Thus pheromones, courtship behavior, and post-mating isolating mechanisms may also play some role in mate recognition. In addition, in some of these cases of sympatric sister species, there are varying degrees of habitat difference, sug- gesting that ecological character displace- ment limits reproductive contact. Both W. multilineata and T. soelneri inhabit marshier habitats than their sister taxa (e.g., Vagvolgyi, 1968; Hubricht 1985); N. alleni fuscolabris inhabits a more alkaline, limestone habitat than N. major (Hubricht, personal communi- cation; personal observations); and T. dis- coidea is found on river bluffs, whereas T. tridentata is found in woods above the bluffs (Vagvolgyi, 1968; personal observations). No conspicuous habitat differences which would restrict contact are known for the four pairs N. alleni alleni vs. N. albolabris hubrichti, N. divesta vs. N. albolabris hubrichti, N. dentifera vs. N. albolabris albolabris, or T. tridentata vs. T. juxtidens. The first two of these pairs need investigation (see Solem, 1976); the third is currently under study in Virginia by T. Asami; and the fourth shows a mozaic distributional pattern (Pilsbry, 1940) suggestive of compe- tition, although they are occasionally found microsympatric (Vagvolgyi, 1968; personal observation). Op vy 85 55 ve 6€ se Ov er Se ve 55 сс 55 ce ce 9€ 05 LE Le er ce LE Ly Op 0€ Le ynpeqns EMBERTON ynpeqns ynpeqns ynpeqns ююючююююч SH LD OO << O9 EC EC CO + CD CD O M M оомомяотя MMM NS ST SU + OS ET ST SA S9JON yues Че) di uaayS 10109 pulyag 9¢ Le Ус Oc Le ve Oc ve Le OC Oc Oc vl Oc сс St Ge €c Sc Oc 9 ve £c 0€ ec ed 81 Se ve Se ve 05 ec €c Le co ce 91 Oc 6l SL [ra 81 Le Oc 6! el Oc 8! Zt 81 Le vi cl Ge Ge Le 8S 602 Gee €¢ ez Zt 85 G02 Gas ed ve ve 89 202 9 Le v7 ve OC LS Keil Eos ed Le OC 85 661 e Le 86 Ge cc LS gel 887 ce ca La 89 € 072 LC ed 61 85 86: L 62 de Sd 8! 9s 561 c'8c Ve OC ez LG €8b c 62 St Zar Le 8's GAZ 0'8c 02 ce 02 8S Г 69 9! St 61 rs cal 6Gc Sill all 02 85 08+ 922 Pa ce 02 LG SIS Li G92 8 67 02 GS 902 Ge OC АС 02 GS 96: 7'55 6 Ge ¿A 9'S Eke 67€ 8 Le 91 rs 98+ 20€ all 61 co es 991 Ove 8 92 91 rs 961 cle pl EC Le rg dst L'6c 9! le 6L rs 98| 00€ Zu vc 91 pS 961 € 62 61 Ge LE rs 881 € 6c all al 8 es 991 Ze et OL Le es 991 952 uonoayep АлэцаиэЧ9 эроч-рии ejljswnjoo дипоэ ‘OÙ (WW) (шш) -81d IV IV IV dep HOUM Apog .UIPIM dif jeseg ses HOUM зчбэн 1ээшеа eus “dod ‘ou SL sugejog/e s1/qe/0qje Sl si/qe/0q/e SUgEIOgIe vl S!/Qe/0q/e suqejoqg/e vl S!/qe/0q/e suqejoqg/e vl S/QEJOQIE sugejoqg/e el S//qe/0Q/8 s11qe/0q/8 el s//qe/0q/8 s11qe/0q/8 cl S//Qe/0Q/8 S1/qe/0Qq/8 ср S!Qqe/0q/8 $ИЧЕОЧЕ ср S//qe/0Qq/8 $ИЧЕОЧЕ LL sugejogje suUgejogIe LL sugejogje sUgeJOogIe OL S//qe/0q/e s11qe/0q/8 OL S//qe/0q/8 S1/qe/0q/8 Ol S//qe/0q/8 s/1qe/0q/8 6 sugejoosny IU8//E 8 sugejoosny Iuajje 8 suqgejoosny jual/e L lUS8I/E IU8/JE 9 1ua/¡e IU8//E G шаре IU8/JE G uajje IU8/JE © шаре Iuajje 5 uajje IU8/JE € uajje IU8/JE с ‚uajje 1иэ]е | uajje IU8//E "OU saloadsqns JO sainads ‘(Z эает) sısAjeue JUBUIWHOSIP 10} pasn ($ э!ае |) SAa|QeUeA ¡Jays 8 ay) payejnojeo элэм YOIYM шо ‘Sdnoib juayje pue 5иде/од/е XxıJayoaN эц} jo ззиэшалтлзеаш ||9YS ‘9 371991 238 239 EASTERN NORTH AMERICAN TRIODOPSINAE ynpeqgns упредп$ ynpeqns ynpeqns OWTODONNODO MMNNN-NNNNNON NS + O + OO + SST o Ae Mp Mee MS Mp ES > A Sh SSP > AO A o o LD CO SLO Sts 0€ ce Ge vv GE ce LS ge ve SE er 55 Lv ch Ov Lv 95 cv Sy ve Le 05 Le Le 6c ve Le ve Le Oc Be €c €c 91 ce co Ge 61 8° 81 ve ve 9 9 9 ec Le 91 ÿL 81 vl ZE cl Le 9! Oc 91! ZE vt 81 vi vl £c GIE 81 JAN сс 61 61 €c Oc [ra Ec Ge 61 Ol el el cl 9! vl LE 9, GE gl: Oc Sl 8 el vl €c 9! 9! 81 Le 61 81 LG 61 61 €c ec LE Ol GE el Ol el el GE SE vl 8 03 81 сс el Sl 8c Oc сс Le Le ve Ge 05 Le 6c ve 81 91 Ol ZE Sl cl SL 9! 81 LE QL Lo Le SI ZE Le Le сс [ra €c Le Oc сс сс 61 Oc ec сс сс LC 61 ra ce 81 ra 61 Ge сс 95 gS gig ES F9 GS os ZAG Sis LEG SS 89 55 8S 8S 6S 0'9 0'9 0'9 19 vs vs vs SiS SS os SE vs Эа ‘UONOUN) JUBUILULOSIP AY} Aq payıssejosuu пецс,.. un Jad ww вг50` = JO0JI{ UOISJ@AUOD,, ‘UOYM Yyy ay, JO риа ay, JO ээвип$ Jaddn ay} UO ww 9`г Jad эещз JO JEQWNN, 9°81 $91 GAGIL ers GZ sg ges 481 cle g6l 8'cc 9'81 Ole 917 ¡Ta 6' Lc 0'Sc Sve 0'Ec vs! Op! ДЗ LS! 691 Zeil: ¿ELL eb! OZ! 8 9c L'EC 0'Sc OVE 0:92 ove Le 6/7 8'0€ Cc 6c ace L'6C L Le 93€ BLE LE 1242 6'55 9'€€ eve ges eee ove 8 9c L'9c 9'9c € Ec 79 95 95 95 GE ve 55 ce Le Le Le 0€ 0€ 0€ 6c 6c 6c 8c 8c 8c Ge ve €c €c ce co Oc Oc Oc 1W8JOS ¡ajos 1W8JOS ¡ajos 1W8JOS ¡ajos ¡ajos Jolew Jofew Jofew Jofew Jofew Jofew Jofew Jolew Jofew Jolew Jofew Jofew Nyaugny Sugejogje Nyaugny sugejogje nyougny sugejogje nyougny sugejogje nyougny sugejogje nyougny sugejogje nyougny sugejogje nyougny sugejogje nyougny sugejogje 240 EMBERTON TABLE 7. Linearized discriminant function for the Neohelix albolabris and alleni groups, based on 8 shell variables (Table 5) standardized to mean = O and standard deviation = 1. Neohelix species and subspecies alleni Shell variable alleni fuscolabris STRIAE — 1.2 — 4 BROWN 35 —1.5 GLOSSY 5.1 1.9 RELSPIRE D —1.7 WHRLEXPN 15 4.2 RELNODE 3.1 57 RELLIP 1.1 — 0 RELDEFL —.6 —1.9 Constant —9;1 — 14.4 A prediction of this hypothesis is that repro- ductive character displacement at the level of populations within a species should be no more than moderate, and more likely should be slight to negligible. A test of this prediction is afforded by the results of Table 10. For 12 species, penial morphology was compared between populations sympatric vs. allopatric with a species of similar shell size and shape. In not one of these comparisons was there any detectable difference. Thus the prediction is strongly confirmed, and the hypothesis is supported that differentiation due to reproduc- tive character displacement is moderate at most. Major differentiation appears to occur un- commonly and rapidly in isolates, because the 6 pairs of greatly different sister taxa are all parapatric, and because the other 14 pairs of isolated (non-sympatric) sister taxa are either identical or only slightly different. Thus the evolutionary pattern suggests a punctu- ated process: when differentiation does occur in isolates, it is extreme and rapid, leaving no intermediates. If this hypothesis concerning major differ- entiation in the genitalia of eastern trio- dopsines is correct, then what evolutionary mechanisms produce this punctuated pro- cess? Since it occurs in parapatry, genetic drift in rare peripheral populations may have sidetracked the selectively canalized devel- opmental program which ordinarily blocks change. Once canalization was overcome, evolutionary change could proceed by any of a number of possible mechanisms, including selection for functional optimization, sexual selection, reproductive character displace- albolabris albolabris hubrichti major solemi —.2 .8 T4 => 1.1 — 9 .8 1.9 =.5 8.0 OA =.1 — 8 6 =x(0 eS —.4 — 1.9 8 nS — 2.8 —1:8 3.0 19 2.3 = 19 ANS — 175 =.3 152 2 Я —4.1 —9.0 —10.4 SA ment, direct environmental selection, contin- ued genetic drift, genetic linkage, and pleiotro- pism. Of these, selection for functional opti- mization and sexual selection seem the most likely causes of major genitalic differentiation. Selection for functional optimization could have acted to prevent the loss of sperm during transfer due either to (1) the penis slipping out or (2) the sperm being captured and digested by the mate's gametolytic gland (the spermatheca—see Tompa, 1984). These two selective pressures would have favored both sculptural modifications which improved the penis's frictional hold within the mate's gametolytic duct (the functional vagina), and structural modifications which improved the ejected sperm's chances of escaping back down this duct to reach the fertilization pouch (the talon—see Tompa, 1984). It seems reasonable that these selective pressures were responsible for such conspicuous fea- tures as (1) the grappling-hook-like pilaster of the Triodopsis tennesseensis group (Fig. 11b-d); (2) the chevron-like patterns of wall pustules convergent among the Xolotrema fosteri-denotata (Figs. 7, 8), the Triodopsis platysayoides (Fig. 12), and the Triodopsis cragini-tridentata-juxtidens (Figs. 13-18) lin- eages; (3) the backward-directed verge in Webbhelix (Fig. 6a) and Neohelix (Figs. 2-5); and (4) the clubbed apex with a subterminal ejaculatory pore convergent among the Neo- helix solemi (Fig 6b), the Xolotrema denotata (Fig. 7), the Triodopsis vulgata (Fig. 9), and the Triodopsis tridentata-fallax-juxtidens (Figs. 14-18) lineages (see Fig. 50). Con- vergences in these structures probably indi- cate that they are adaptive. EASTERN NORTH AMERICAN TRIODOPSINAE 241 FIG. 49. Range maps of eastern American triodopsines. Adapted from Hubricht (1985). If these suggestions are correct, why, then, is there so much diversity of form? That is, why are so many different responses to the same two selective pressures? In the first place, the selective pressures may not be equal in each clade. For example, one clade may have a thick mucus which would clog delicate sculpture and therefore would select for coarse sculpture. Or, for example, clades may differ in the strength of the digestive enzymes secreted by the gametolytic gland or in the presence or strength of a muscular pump in the wall of the gametolytic duct (the functional vagina), therefore selecting differ- ently for morphological “strategies” to chan- nel the ejaculate back down the duct. According to the hypothesis of sexual se- lection by female choice, “male genitalia func- tion as ‘internal courtship' devices used by females to discriminate among “males” (Eber- hard, 1986). Runaway sexual selection pro- duces rapid and arbitrary divergence in penial morphology, according to Eberhard's model. Much of the divergence in eastern-triodopsine genitalia, however, is not arbitrary but is con- vergent, suggesting natural selection for func- tion rather than sexual selection. Both forms of selection have probably played a role, however. Certainly the apparent rapidity of divergence is consistent with Eberhard's model. Reproductive character displacement has already been ruled out as a likely cause of major differentiation. ejoa¡bau Bapıoasıp suapııxnl (рыаАч) (риаАч) sısuswegeje EJ3J0SgO Pjajosqo (puugAy) хЕ|Е) (puqAu) (puqAu) IUI6E19 saepio{esAe;d ejebinA eJebinA (puqAy) EIOUISGO eJejouap SIEJUSPIIIO 118]50} EMBERTON TETE Jolew sugejogje eynpued ejoa¡bau suapyxní suapııxn! хв//Е) XEIIEJ XEIIE} 4UaUJ30S XB/]J хе|Е) XB/]8J хе esodny eJeJuapı 12d09 12d09 12d09 U8pI9/N} esobni ejeuejduwoo ejeuejduoo 142109 eJeue/duoo eJuaınpney Bjuajnpney ejoa¡beu ejoa¡bau в]24540 2124540 2]2145490 snuejuabies 118]50} EJS8AID Р/ащиар sugejogje sugejogje sugejogje eJeaunnnw sısdopon] sısdopou] sisdopo!/ sisdopo!/ sisdopol/ sisdopo!/ sisdopo!/ sisdopouy sisdopouy sisdopouy sisdopouy sisdopou sisdopou 5/5А0рои | sısdopou] sisdopo!/ 5/5А0рои | sisdopo!s/ sisdopo!/ 5/5аорои | sisdopou | 515Ч0рои | 5150 0!рои 1 5/5А0рои | 5/5аорои 1 sisdopou sısdopon] EWENOIOX вшало|ох вшало|ох (e1914) ‘И EWENOJOX xılay0aN X!|9409N x//94y09N XI2Y09N XI[2Y08N XI/9409N SIq10p10/9yS 5/5А0рои sisdopouy sisdopouy 5/5Ч0рои | 5/5Ч0рои | sisdopo!/ SIQIOPJOJEH sisdopo!/ EWELJOJOX SIALOXOJNM xI/9/409N x//9409N XII2UO8N XI|2YO8N xıjay03N sisdopo!/ sisdopou sisdopou sısdopom | sısdopou sısdopou] sisdopo!/ sisdopou вшало/ох EWENOJOX EWELNOJOX ешело/ох EWENOJOX EWELNOJOX EWENOJOX BapIoasıp suapıyxnl sısuawegeje 29/0540 uopuajue seyanuay ejeue/duwos sisuaassauua} eJebima S!¡ejuapi990 вшлэро! 515иа5лэлед sugejoasn, y иае Jolfew sugejogje ejoa¡bau вешарщ ввшиарщ sisuauojadoy /риедзоииел /риедзоииел иаиао$ sisuguojadoy XE//E} esobns eJeJuapı) ESONJINA ESONJNA бе esobni eJeJuapin ввиэрщ sepioAesAjejd виапрпел Pjuajnpnedy SISU8IUI[O129 2]214540 eJejou EIOLNSIO 118150} шплоивриш BJSOAIP eJaJuep sugejogje sugejogje sugejogje suqejoqe eJeau! nus sisdopou sisdopou $15аорои | sisdopo!/ sisdopo!/ sisdopo!/ sisdopo!/ 5/5А0ро | sısdopoul sisdopou sisdopou sısdopoul sısdopou] sısdopoi] sısdopou] sısdopoi] sısdopoi] sısdopoi] sısdopou] sisdopou | EU91J0/OX EWSNOJOX вшэлоох MENTE EWENOIOX (uoposew) W XIJ8Y08N XIJ8Y08N XII8Y08N XI/94y09N XIJ2Y08N x//9409N XI/9/09N eınpuad в)оэ/баи BapIoasıp suepyxn/ sisuauojadoy 'риедзоииел sısuswegeje 1auj9os Pjajosqo susnjed РИР55ЭШ xeJje} uopuajue ввшэрщ эецаииац PSONJINA 1uıbes0 uapıaıny esobns Pyeuejdwoo SISU88SS8UU8) 142109 sapıofesAjejd eaoid eJuaınpney sisuausoqgiejo вебпл SISUBIUNOJEO EJ91}SqO EJEJOU8p $/виар!эо 118}50} вшларо! EJS8AID eJa/nuep 119/05 шае Jolew sugejogje eJeauınw воэ/баи ejoa¡bau suapyxní suapyxní sisuawegeje sisuawegeyje sısuawegeje xejB} XE/JE} XE/JE} xejje} ХЕ|/Е} ejuajnpney ejuaınpney eJe6/nA eJe6¡na EJS8AIP EJS8AID BJ8/}U8p suapxní suapınl suapııxn! suapııxn! XB//8) XB//8) XB//8, XB//8, xejje} хв//Е) ХЕ//Е} ХЕ// Е} в)виэерщ ввиэрщ бе бе 16219 (¿Jeso6n1 esobn sisuaassauua} sisuaassauua} 1younq sepioAesAjejd вебпл EJeb ina EJeb¡nA eJebynA в)воиэр в)воиэр eJejouap 119/50) 119450} в/эдиер г/эщиэр в/эдиэр шае jaye suqejoqe sugejoqe sısdopoi] sisdopol/ sısdopoi] sisdopouy sisdopouy sisdopou sisdopouy sisdopouy sısdopoi] sısdopou] sisdopo! sısdopoul sisdopou sisdopou sisdopou 1 sisdopol sisdopoi sisdopo! sisdopol sısdopou] sısdopou] sisdopo!/ sisdopo! | sisdopol/ sisdopo! | sisdopoi sısdopoil EWENOJIOX EWENOJOX EWENOJIOX EWENOJOX EWENOIOX XI|9/09N x1/9409N XI|9/09N х/эцоэм xılayo0aN xılayoaN xılayoaN х!94999М saioadsqns salsads snuebqns uoneoissejo (8961) $1Абюлбел U01198S snuabqns snuan) uoneayissej9 (6961 'ErS6l 2561) $.999М Salsadsqns salsads snuabqns иоцеэ!5$е1о (OPEL) S.A1qS|I4 saisads дполбап$ saisads dno1B saisads LONEONISSEIO SIUL snuan) 242 "SUONEONISSEIO SNOIAGId ээлц} YIM Sauisdopoii} иеэнашу\ иаэ}зеэ JO UONEINISSEIO pasinar AU] jo UOSUedWOD 'g JIGVL EASTERN NORTH AMERICAN TRIODOPSINAE 243 TABLE 9. Comparison of the difference in penial morphology with the relationship between geographic ranges for 25 pairs of sister taxa of eastern triodopsines according to the phylogeny in Fig. 50. The taxa are designated by five-letter abbreviations. Question marks denote pairs of phylogenetically uncertain status. “—" is a minus sign. Phylogenetically adjacent taxa solem vs. rest of Neohelix fostr group vs. denot group platy vs. vulgt group tenns group vs. burch group (?) cragn group vs. rugos group cragn group vs. tridt group Webbhelix vs. Neohelix-solem soeln vs. rest of fallx subgroup tridt group vs. juxtd group dentf group vs. albol group + allen albol group vs. allen albol vs. major dentf group vs. divst group vulgt subgroup vs. fraud subgroup rugos vs. fulcd (?) liodm vs. divst occdt vs. fostr denot group (3 spp.) claib vs. vulgt picea vs. fraud compl vs. tenns cragn group (3 spp.) fallx subgroup-soeln (7 spp.) anter vs. tridt juxtd group (4 spp.) Penial Geographical shift relationship great parapatric great parapatric great parapatric great parapatric great parapatric great parapatric moderate sympatric moderate sympatric moderate sympatric moderate sympatric moderate sympatric slight allo or parapatric slight allopatric slight allo or parapatric slight allopatric none peripatric none peripatric none paraipatric none peripatric none peripatric none peripatric none peripatric none parapatric none peripatric none peri or allopatric Did the external environment select for penial-morphological differences in the east- ern triodopsines? It seems unlikely. The spe- cies groups and genera—that is, the major morphological types—do not segregate eco- logically (Emberton, 1986), nor is there any evidence of environmental correlation at any level, including within species groups. There seems to be no correlation between the size of the penis and its structural complexity. For example, Neohelix lioderma is as small in both body and penis as many species of Triodopsis, yet has the Neohelix penial sculp- ture in its full complexity (Fig. 5a). A correla- tion recurrent in stylommatophorans between arid habitat and short penial length (Solem, personal communication) does not apply to the eastern triodopsines, in which the great- est penis-length-to-shell-diameter ratio ос- curs in the Triodopsis cragini group (Fig. 13), which also occupies the most arid habitat of all known triodopsines (Emberton, 1986). Genetic drift, although possibly the instiga- tor of genitalic divergence by straying from canalized fitness peaks, is not likely to be the proximate cause of the divergence. Evidence for this view lies in the multiple convergences and in the apparent speed and morphological precision of evolution. Drift, however, can be held responsible for vestigialization: random variation in structures that are no longer func- tional. The dorsal pilaster of Neohelix solemi (Fig. 6b), as well as the verges of N. solemi, the Xolotrema fosteri group (Fig. 8), and the Xolotrema denotata group (Fig. 7), are pre- sumably vestigial. Although the rough concordance between conchology and penial morphology (Fig. 50) could indicate genetic linkage, with selective changes in the shell randomly inducing unselected changes in the penis, it is more likely that since both shell and genitalia have undergone (independent) evolutionary diver- gence, they both are correlated with time, and hence, secondarily, with each other. Eastern triodopsines have a relatively high chromo- some number (29 to 32 pairs, according to Husted & Burch, 1947), obviating the ne- 244 EMBERTON TABLE 10. The localities (state:county) of populations dissected in searches for reproductive character displacement between pairs of conchologically similar species of eastern American triodopsines. The number of specimens dissected from each population is in parentheses. Species A Allopatry Sympatry Allopatry Species B albolabris AR:Logan(1) AR:Crawford(2,3) AR:Izard(8) alleni AR:Washington(1) IA:Lynn(1) OK:Sequoyah(3) lA:Jackson(1) TX:Houston(3) IA:Clayton(1) LA:Washington(1) albolabris WV:Greenbrier(2) WV:Preston(3,3) WV:Pendleton(3) dentifera WV:Boone(3) NC:Watauga(3) OH:Athens(2) PA:Chester(2) vulgata KY:Harlan(4) KY:Fayette(1,3&3) — denotata 8 TN:Morgan(2) tennesseensis vulgata KY:Fayette(1) KY:Harlan(1,4) KY:Edmonson(2) tridentata TN:Morgan(2) WV:Pendleton(1) WV:Pocahontas(1) WV:Preston(1) OH:Athens(3) tridentata KY :Harlan(1) KY:Edmonson(2,2) — obstricta WV:Pocahontas(1) WV:Pendleton(1) WV:Preston(1) OH:Athens(3) tridentata WV:Pocahontas(1) WV:Pendleton(1,2) — picea WV:Preston(1) KY:Edmonson(2) KY:Harlan(1) OH:Athens(3) Juxtidens WV:Pendleton(2) WV:Pocahontas(2,1) WV:Pendleton(1) tridentata NC:Catawba(3) NC:Burke(1) cessity for, or the probability of, tight link- ages. Pleiotropy is also an unlikely explanation for the major morphological diversity of the eastern-triodopsine penis because the penis develops from mesoderm, whereas the shell- forming mantle develops from ectoderm (Raven, 1975). To summarize, neither reproductive char- acter displacement, environmental selection, genetic drift, genetic linkage, nor pleiotropy is a probable cause of major evolutionary change in eastern-triodopsine genitalia. This supports the suggestion that selection for functional optimization and sexual selection are the most likely causes. WV:Preston(1) KY :Edmonson(2) KY :Harlan(1) OH:Athens(3) Shell evolution: pattern and process Since the consensus phylogeny (Figs. 28, 50) was constructed strictly from soft-part anatomy and biochemistry, there is no circu- larity in using it to detect patterns of shell evolution. Shell variation was analyzed at three taxonomic levels: among genera, among species groups, and within species groups. Patterns—and inferred processes— of variation differ among these levels. In general, each genus has a characteristic shell morphology (Fig. 50). Neohelix and Webbhelix share the plesiomorphous shell: large, globose, and toothless. Xolotrema shells are medium-sized and depressed, with EASTERN NORTH AMERICAN TRIODOPSINAE 245 Xolotrema Neohelix & NS è р = й М x = A > S 4 \ Various = Ne SE outgroups в ) > N 4 x = = ой > == DES Triodopsis ZA ‘al i | Bi FIG. 50. Evolution of shell morphology and of upper penial sculpture in eastern American triodopsines. a blade-like parietal tooth and a long basal lamella. Triodopsis shells are small, sub- globose, and tridentate. The most common exceptions to this generality are in size. Intraspecific variation—discussed below— produces broad overlap in shell size, both within and between genera. Nevertheless, the largest occur in Neohelix, and by far the smallest occur in Triodopsis. The rare convergences (Neohelix dentifera, Triodopsis platysayoides, and the T. tennesseensis group) offer little contradiction to the general- ity that within major clades (genera) of east- ern triodopsines, shell morphology is distinct and virtually static. The evolutionary process behind this pat- tern is problematic. A preliminary study (Em- berton, 1986) concluded that the genera broadly overlap ecologically, with virtually no conchological changes accompanying eco- logical convergences. Ecological relation- ships are clearly in need of further investiga- tion (see Goodfriend, 1986). Within a genus, the distributional pattern of shell characters among species groups is generally mosaic, with many cases of conver- gence or parallelism. This rank mosaicism has confused past conchologically based sys- tematics (e.g. Pilsbry, 1940, and Vagvolgyi, 1968; see Table 8). lts pattern suggests a process encompassing both drift and selec- tion. Possible selective explanations for a few of the recurrent shell features are discussed in the next section. For many of these fea- tures, however, it seems more likely—but would be impossible to demonstrate unequiv- ocally—that their genetic program is ubiqui- tous in the subgenus, is selectively neutral with respect to its alternative states, and is expressed randomly among species of the clade due to genetic drift. This process is called genetic indeterminism by Throck- morton (1965)—also see Gould's (e.g. Gould 8 Woodruff, 1986) discussions of morpholog- ical canalization. Within species, there is great variation in shell size, spire height, umbilical relative width, and whorl count (Table 11). Diameter ranges up to 95% within a species; the extent of this range in shell size depends on the number of populations and specimens mea- sured. Such variation in adult size is common not only in land snails, whose time for shell growth (before it is interrupted by reproduc- tive maturity) depends heavily on the local humidity regime (e.g., Solem 8 Christensen, EMBERTON 246 6r Ir 89567 9661 LS’ Ov" 09-87 0805 = = Ec pl a = 6r ch 89—67 027195 ve-cl 59—65’ MA 9S'—8€" 9908 сес EZ veo pl 05-17 09 ES BETZ 09-27 GS-ey 6 Er 7: €c—9l = = 9577, = a LoS 57-75’ 9-09 ve-Sl 57—05 ÿ9-0'G 059 | | ооо ceba 6e-l'e 9'ÿ—c'e Gel’ O'e-c'l Loco СЕТ £ÿ0c 01-611 02-11 EZ 0'9-6'¢ Ziel 6'5—6'1 USE 201010001000 WRIP/SIUM SHOUM weıpyun snoilqun 69049 esp vs 0p Loo 99-87 SZ 9905; ANA 09'—8v' 19-87 vs ct" vs 0v 99-97 E97 59—17’ SS abu ¿S'-cv' 89—57 ДЭ 199° 99-99 71—69’ E29 188" WEIPAH ARE) 99-9 0'01-8'S 66 ES Sil бр ВО L6 63 OS SER) LICE GL VS 9'01-0'8 сэ =67 065229 9'01-8'S 9231-96 65198 0°L1-6°9 81128 951-901 811-2751 S'O£-8 81 6 Ec-G' tl c 8l-0 01 JybiaH %0 6€ JE EE %0 C9 %6 +6 SESS %e 6€ %S'ES %€ 89 %L'9E %5`07 %C OV %S 6€ %€ BV Sl CE %5`69 %G LE %b "ES %v'6S %7`91 %S Lp %c 6€ %6 ES %C OL %S'El obueı 1э}эшеа 8'€l-Z'01L c €l-6'6 602-6 21 L'61-8°6 L’EL-9'8 S'cl-ZL6 чер ¿'0c-€ CI ва 8Ol-L 2 S'GI—9 01 556—191 cel-6'8 691-251 861 сс: 65-261 657-071 O'cc-8 El S 61-691 9212-561 L'OE—E EC ЕЕ Le 998-205 L'82-2'91 Jayaweiq ve 18 v6 c8c УР! ver 281 965 9! 58 144 ce 6c ve LES Lp 26! SE ve ve ZS Ly 897 082 sılays [8101 ОЕ 05-1 ste ce-l Lec LES 08-2 SE ИС сс 8-1 DRE 91-3 OSE Get ИО vec сс E11 Set 216 8-1 Lo! Gc—cC Jo] 19d sIIaus 6 вприаа 6 ejoa¡bau el Baploosip 85 suapıyxn! Git sısuawegeje фе (xey/ej) + ejajosqo + siujsnjed 02 XB//8J 08 ввиэрщ G в5опупл 91 IUI6E19 el uopuajue + esobns гг BeJeueJduoo + sisugassauua) € ıyaıng 6 eaoid + ejuajnpney 16 sISUSuJOqieJo + вебпл LL EISLNSGO er BJEJOUSP ez 119,50] 1 EJSSAIP 51 PJ9/}U8p OL 1ua/¡ 61 (ilwajos) + sofew 98 (lwajos) + sugejoqie ce eJeau!)1 ¡nu $107 saloads ‘8 ajqe 1 о} Buipio99e sjuawsn{pe эниоцохе} цим ‘(8961) 'АбюлбБел Jo 1x8] eu} шоц pajıdwon ‘seuisdopou] иаэ}5еэ jo $эю0э4$ ui ззиэшалпзеаш |jeys Ul UOHeUeA jo зэбиен ‘LL JIGVL EASTERN NORTH AMERICAN TRIODOPSINAE 247 1984; Gould, 1985), but also in aquatic gas- tropods (e.g., Vermeij, 1980). Vagvolgyi (1968) documented that intra- specific shell variation is geographically patchy and non-clinal (the small number of clines he reported is no more than one would expect by chance). The same was true of apertural features, keel, fulcrum, and sculp- ture. Vagvolgyi attributed this patchy variation to ecophenotypic responses to patchily dis- tributed microclimates, “occur[ing] in spite of gene flow, not because of lack of it.” This interpretation is probably correct—see the documentation of this phenomenon in Cerion (Gould, 1985) and in Neohelix major and Mesodon normalis (Emberton, 1986))—but genetic drift and local selection could also play significant roles. In addition to this patchy, non-clinal pattern in size and shape, there are several correla- tions between niche and shell morphology which recur within species and species groups. These convergences, discussed in turn below, are probably due to environmental selection. Apertural obstruction correlates with ground moisture. Parallel altitudinal clines in the size of apertural teeth occur in Triodopsis tridentata, T. fallax, and T. fraudulenta (Vag- volgyi, 1968). The aperture becomes more obstructed with increasing elevation and, con- comitantly, increasing ground moisture. An altitudinally opposite cline exists in the cragini group (Vagvolgyi, 1968), with the most highly obstructed species (henriettae) inhabiting lowland, riverine forests; the least obstructed species (cragini) occupying dry uplands; and vultuosa intermediate in both apertural ob- struction and habitat. Thus the consistent correlation in all four clines is with ground moisture. This pattern supports the view of apertual teeth as barriers to insect predators, presuming that the density and/or diversity of insect predators increases with increasing ground moisture, but fails to support the view of apertural denticles as barriers to water loss (Goodfriend, 1986). An alternative view is that snails living in humid habitats have a longer season of activity, hence more time for the deposition of shell material, including the apertural teeth. Flatness correlates with crevice dwelling. Five separate species groups show the par- allel evolution of a flat-spired species associ- ated with rock crevices (see Emberton, 1986). In the Neohelix alleni group, N. alleni fusco- labris is flat for the group, and is restricted to limestone-cliff areas of northern Alabama and adjacent Tennessee (Hubricht, 1985, and personal communication; personal observa- tions). The Xolotrema fosteri group has X. occidentalis, a flat, subcarinate cliff-dweller; the Xolotrema denotata group has X. obstricta, which, with its pronounced keel and depressed spire, is the most rock-associated member of its group. The aberrant Triodopsis platysayoides inhabits crevices between sand- stone blocks in a restricted region of the New River Gorge, West Virginia, and has the flat- test spire of the entire genus. The Triodopsis Juxtidens group’s only exclusive cliff-dweller (along the Ohio River Valley) is the conspic- uously flat 7. discoidea. This parallel concor- dance with habitat suggests that a flat shell is adaptive for rock-crevice dwelling. Similar en- vironmental correlations occur in several lin- eages of Mediterranean helicids, suggesting the same selective pressures (Goodfriend, 1986). Glossiness correlates with water. Another iterated shell-habitat correlation—pointed out by Vagvolgyi (1968), and comparable, as he stated, to Rensch’s (1932) trends—is be- tween a glossy periostracum and nearness to a large body of water. The glossiest member of the Neohelix dentifera group (N. lioderma) appears to be restricted to the Arkansas River Valley. In the Triodopsis tennesseensis group, glossy T. complanata lives along the river, whereas dull T. tennesseensis occurs on the upper banks and farther inland. In the Triodopsis fallax group, two species have independently evolved a shiny periostracum: T. palustris along the Santee River floodplain, and Т. soelneri in the marshes of the Lake Waccamaw area. The riverine cliff snail Trio- dopsis discoidea is the glossiest member of the T. juxtidens group. Glossiness may be an exclusively ecophenotypic effect, but is prob- ably at least a partially selected trait. Its heritability has never been assessed, al- though Grimm's (1975) lab-reared T. soelneri and its hybrids should yield important data (see Appendix D). Juvenile apertural size correlates with arid- ity. Vagvolgyi (1968) also noted that arid- adapted species of eastern triodopsines have relatively smaller juvenile apertures, hence the tightly coiled shells of the Triodopsis alabamensis group and, to a lesser extent, the Triodopsis cragini group. The same pat- tern has been found in various other groups of land snails; it implies natural selection for water loss, although not all experimental re- 248 EMBERTON sults have been consistent with this inter- pretation (Goodfriend, 1986). To summarize, there are two major compo- nents to the pattern of shell variation within species and species groups of eastern triodopsines, and they appear to differ in the processes which produced them. First, the patchy, non-clinal variation in the size, and many features of form, of the shells is proba- bly due to ecophenotypic effects. And sec- ond, the several correlations between envi- ronment and shell morphology iterated among separate lineages are probably due primarily to natural selection, with perhaps some ecophenotypic contribution. What is a species in the eastern American triodopsines? If the biological species concept were used for the eastern triodopsines, species groups would probably be reduced to species. Spe- cies groups have, with two exceptions (Neohelix solemi and Triodopsis soelneri), virtually identical genitalia, hence are proba- bly capable of interbreeding. Indeed, hybrid- ization has been reported (based on analysis of geographical shell variation) within the Xolotrema denotata group (Vagvolgyi, 1968) and the Triodopsis fallax group (Hubricht, 1953; Vagvolgyi, 1968; Grimm, 1975). Vagvolgyi (1968) even concluded that certain Hubrichtian (1985) species are not species at all, but hybrid swarms: X. caroliniensis, T. vultuosa, T. henriettae, T. messana, T. van- nostrandi, and T. hopetonensis. The only reported cases of reproductive isolation within species groups occurs among some mem- bers of the Triodopsis fallax group, which nevertheless still hybridize in the laboratory (Grimm, 1975; see Appendix D). Laboratory hybridization has also been reported within the Xolotrema denotata group (Webb, 1980). In the revision of the Neohelix albolabris and alleni groups (Appendix B), the biological species concept was applied, using the “yard- stick method” of comparing sympatric spe- cies to determine the degree of penial differ- ence capable of reproductively isolating species (under the still unproven assumption that penial morphology is the predominant mate-recognition system). Thus, subspecific status was assigned to genetically isolated, conchologically differentiated taxa which had the same or only minutely different penial sculpture. Specific status was provisionally assigned to N. major because its penis may be different enough, by yardstick criteria, from the similar N. albolabris to prevent interbreed- ing, and this difference is disjunct, with no sign of clinal intergradation. If these same criteria were applied to a species-level revi- sion of all eastern triodopsines, then species groups would be reduced to species. This was not done for anumber of reasons. First, there are important precedents in land- snail taxonomy for species which hybridize. Gould & Woodruff (1986), for example, opted to assign specific status to two hybridizing “semispecies” of snails (Cerion glans and gubernatorium) of New Providence Island be- cause of a multitude of evolutionarily signifi- cant differences. Murray 8 Clarke (e.g., 1980) followed a similar taxonomic path with the “incipient species” of Partula on Moorea. Second, there is some evidence that spe- cies may be reproductively isolated despite close genitalic and conchological similarities. Recent discoveries in the confamilial genus Ashmunella indicate that morphological differ- ences among valid species can be slight. Ashmunella has all appearances of being oversplit, with specific status bestowed on a mosaic collection of often subtle shell differ- ence. Karyotypic and breeding studies (Babraksai & Miller, 1984) have shown, how- ever, that at least one such subtle shell dis- tinction marks true biological species: hybrids of A. proxima and A. lenticula suffer gametic disgenesis producing effective sterility. Thus, in the eastern triodopsines, Grimm's (1975) and Hubricht's (1953, 1985) assertions that messana and hopetonensis, as well as obsoleta and hopetonensis, live in sympatry without conchological evidence of hybridizing cannot necessarily be rejected (as Vagvolgyi, 1968, did) simply because of the subtlety of their shell differences, because apparent integrades exist elsewhere, or because—as reported in this paper—their genitalia appear identical. In view of these considerations, there sim- ply is not enough evidence on which to base a robust species-level revision of eastern triodopsines at the present time. Therefore Hubricht's (1985) species designations have been retained in the supraspecific revision (Appendix C). Recommendations for future research The nature and definition of a species need to be researched for eastern triodopsines. Because this is now one of the phylogeneti- EASTERN NORTH AMERICAN TRIODOPSINAE 249 cally best understood groups of pulmonates, such investigations will yield important gener- alizations concerning pulmonate systematics. In addition, despite the general congruence between the two data sets (genitalic and allozymic) used for phylogenetic reconstruc- tion, there are several problematic groups for which data were incomplete or in conflict. (1) The taxonomic status of Webbhelix multi- lineata chadwicki needs to be assessed (see Webb, 1952). (2) The zone of potential con- tact or integradation between Neohelix albolabris albolabris and N. major needs col- lection and assessment. (3) The precise ranges of N. albolabris hubrichti and N. alleni alleni, and the degree of range overlap and sympatry, need to be determined. (4) Topo- typic “Neohelix albolabris traversensis” needs collection and dissection to test the prediction that it is N. albolabris albolabris which conchologically converges on solemi; if it is anatomically what has been called solemi, then the name traversensis has pre- cedence for this species. (5) The status of Neohelix lioderma is in question: is it теге a small-sized population of N. divesta? (6) The systematic and ecological relationships of Xolotrema fosteri and X. occidentalis need evaluation; for example, do other “oc- cidentalis's” (flat-spired cliff dwellers) occur as ecophenotypic variants within the range of fosteri? (7) One of the most promising areas of investigation is in the Xolotrema denotata group. Despite a basic sameness of the ap- erture and of the penial morphology, and despite evidence of hybridization, shell varia- tion is extreme. lt ranges from subglobose, with a rounded periphery, bearing periostra- cal hairs, and ribless (denotata); to de- pressed, with a keeled periphery, hairless, and strongly ribbed (obstricta). These are the only hairy shells and the only keeled shells in the eastern American triodopsines. Vag- volgyis (1968: Fig. 21) claim that carolini- ensis is a hybrid zone around the circular range of obstricta where it is nearly sur- rounded by the range of denotata is quite plausible, but needs to be tested electro- phoretically and by more rigorous shell anal- ysis. The ecological significance, if any, of the disjunct shell forms has yet to be investigated. (8) The question of whether Triodopsis claibornensis is a local ecophenotypic dwarf of T. vulgata needs to be settled. (9) Likewise, what is the status of Triodopsis picea in relation to T. fraudulenta? Does its ecological separation (high-montane) and its shell differ- entiation (dwarf, pustulose) denote incipient or full speciation, or ecophenotypic variation? (10) The phylogenetic position of Triodopsis platysayoides as sister to the T. vulgata group needs corroboration from an independent data set to be considered truly robust, be- cause of its aberrant, unique penial morphol- оду. (11) The electrophoretic similarity of Triodopsis burchito Neohelix, in addition to its unique dorsal-pilastral sculpture of uncertain homology, make it a problematic species. It clearly needs further comparisons. (12) The phylogenetic position of the Triodopsis ten- nesseensis group is in question, and needs testing by other data sets. The possibility needs to be investigated that T. complanata is an ecophenotypic variant of T. tennes- seensis, its glossiness due to living near water. It the two are true species, do they hybridize? (13) Electrophoresis of fulciden should clarify its now dubious placement in the rugosa group. (14) The Triodopsis cragini group, despite its disjunctly different penial morphology, parallels the variation of the T. fallax subgroup. Vagvolgyi’s (1968) claim of hybridization, rejected by Cheatum & Fullington (1971) and Hubricht (1985), de- serves electrophoretic testing. Any shell stud- ies should explore ecological correlations. (15) Since Triodopsis anteridon’s range lies within that of 7. tridentata, is it an ecological variant (confusingly convergent, by the way, on 7. rugosa)? If not, do the two species interact? (16) The Grimm-Hubricht hypothe- sis on the evolution of the fallax subgroup (Appendix D) needs rigorous testing. Grimm’s unpublished lab-hybridization data and spec- imens should be evaluated. Multivariate shell morphometrics, coupled with targeted mito- chondrial DNA studies, should resolve the problem of this intriguing evolutionary micro- cosm. (17) The Triodopsis juxtidens-T. discoidea pair seems to be a case of incipient or recent speciation involving a major shift in habitat accompanied by an apparently adap- tive shell change. Vagvolgyi (1968) claimed conchological intermediates between juxti- dens and discoidea in the Kanawha River Valley of West Virginia, suggesting that speciation is not complete. Careful investiga- tion of this system, including ecological anal- yses and tests for directional selection for a flattened spire may be the best approach to generalities about the speciation process in triodopsines. The eastern triodopsines, because of their species diversity, their robust phylogenetic 250 EMBERTON hypothesis, their mapped species’ ranges, and their broad conchological, genitalic, and allozymic variation, are a superlative system for further evolutionary studies. For example, the three major clades (Neohelix, Triodopsis, and Xolotrema) could be compared as to (1) their modes of speciation; (2) their covaria- tions among the respective evolutionary rates of anatomy, shell, and allozymes; (3) their phylogenetic changes in shell ontogeny, as measured from sections or x-rays of adult shells (Raup, 1966; Schindel, in review, 1986); (4) their rates of spread from Pleistocene refugia as determined by al- lozymic geographic variation; (5) their strengths of selection—measured by compar- ing dead shells of juveniles with the juvenile whorls of living adults—in parallel adaptive trends (e.g., flattening of the spire as an adaptation for cliff dwelling); and (6) their ecophenotypic plasticity of shell shape. Perhaps the most promising aspect of the eastern triodopsines for the study of evolution is that their conchological radiation has been reiterated by the distantly related, confamilial genus Mesodon (Pilsbry, 1940; Emberton, 1986). These two radiations overlap each other almost perfectly in geography, ecology, conchology, and species richness (Emberton, 1986). Thus Mesodon represents a natural, in situ replication of the evolution of the eastern triodopsines. Such synchronous, sympatric, parallel radiations appear to be quite rare in nature, and present untapped opportunities for formulating and testing general hypothe- ses concerning evolutionary convergence. Since convergence can only be evaluated in the context of phylogeny (e.g., Bookstein et al., 1986), this monograph and a parallel monograph in progress on Mesodon lay groundwork for utilizing this system. ACKNOWLEDGEMENTS | take pleasure in thanking the people and organizations who have made this project possible. Alan Solem provided space, equip- ment, field funding, instruction, and speci- mens at the Division of Invertebrates, Field Museum of Natural History, Chicago. Linnea Lahlum did some or all of the stippling on several of the anatomical drawings. This paper is a contribution of the Molecular Genetics Laboratory of the Department of Malacology Academy of Natural Sciences of Philadelphia. George Davis generously gave of his time and facilities there, and was a continual source of instruction, discussion, and encouragement. Davis also was most helpful with advice on organizing the manu- script. Caryl Hesterman taught me to do starch-gel electrophoresis; | am grateful for her skill and patience. John Hendrickson, also of this Academy, generously ran all of the data analyses employing PAUP and BIOSYS, and provided invaluable advice and patient trouble-shooting. | am also grateful to members of my thesis proposal and defense committees: David Raup, Michael Wade, H. Bradley Shaffer, Russell Lande, Lynn Throckmorton, James Teeri, and Harold Voris. For assistance in the field-collection of spec- imens, | am grateful to Ellen Emberton, Lucia Emberton, Ned Walker, Gene Bryant, Tony Bryant, Eugene Keferl, Leslie Hubricht, John Ahrens, John Petranka, Betsy Kirkpatrick, Glenn Webb, Wayne Van Devender, Amy Van Devender, Martha Van Devender, Wayne Evans, Arthur Bogan, Bob Lawton, John Pinkerton, Mark Southerland, Dennis Herman, Greg Mueller, Kisa Nishikawa, Phil Service, Joe Bernardo, Ken Baker, Alan Lo, and David Kasmer. | also thank the many park rangers and private-property owners who gave per- mission to collect on their land, and the many people who provided camping sites or other hospitality. For their assistance in getting me collecting permits, | am grateful to Steven Chambers of the Office of Endangered Spe- cies, and to Ken Knight of the West Virginia Department of Natural Resources. Margaret Baker, Patricia Johnson, and Lucy Lyon gra- ciously and efficiently labeled and catalogued my collections at the Field Museum. Wayne and Amy Van Devender continually sent me live snails from all over the United States, some of them critical material for this study. Andi Garback was prompt and courteous in lending me specimens from the collection of the Academy of Natural Sciences of Philadel- phia. Glenn Webb kindly permitted me to study his slide-mounted voucher specimens, and generously shared his vast knowledge of the Polygyridae. Leslie Hubricht unstintingly provided col- lecting localities, identifications of question- able material, then unpublished range maps (Hubricht, 1985), critical specimens from his personal collection, and advice. Without Mr. Hubricht's help, the realized scope of this study would have been unthinkable. Frank Climo gave helpful comments on an EASTERN NORTH AMERICAN TRIODOPSINAE 251 early draft of this paper. | am also grateful to two anonymous reviewers for their valuable critiques. This work was funded by the following grants to the author: Public Health Service Genetics Training Grant GM07197-07; the Hinds Fund of the University of Chicago; the Jessup Fellowship Fund of the Academy of Natural Sciences of Philadelphia; the Louer Fund of the Field Museum of Natural History, Chicago; and the Student Computation Fund of the University of Chicago. Some additional funding was provided by a National Science Foundation Grant to George M. Davis, by a United States Department of Agriculture grant to Michael J. Wade, and by field funding from the Field Museum of Natural History to the author. LITERATURE CITED AVISE, J. C., 1975, Systematic value of electro- phoretic data. Systematic Zoology, 23: 465-481. BABRAKZAI, N., WARD, О. С. & MILLER, W. B., 1975, The introduction of Giemsa and centro- meric banding techniques of chromosomes to molluscan cytotaxonomy. Bulletin of the Ameri- can Malacological Union, 1975: 67. BABRAKZAI, N. & MILLER, W. B., 1984, Cyto- genetic study of interspecific hybrids of Ashmunella (Mollusca: Pulmonata: Polygyridae). |. А. proxima X A. lenticula F, hybrids. Malacologia, 25: 413-426. BAKER, F. C., 1939, Fieldbook of Illinois land snails. Illinois Natural History Survey, Manual 2, Urbana, 166 pp. BINNEY, A., 1851, The terrestrial air-breathing mollusks of the United States and the adjacent territories of North America, Vol. |. Boston, Mas- sachusetts, 366 pp., 16 pls. BINNEY, W. G., 1878, The terrestrial air-breathing mollusks of the United States and the adjacent territories of North America, Vol. V. University Press, Cambridge, Massachusetts (Bulletin of the Museum of Comparative Zoology, 4:, 449 pp., 16 pls. BURCH, J. B., 1962, How to know the eastern land snails. Brown, Dubuque, lowa, 214 pp. BUTH, D. G., 1984, The application of electro- phoretic data in systematic studies. Annual Re- view of Ecology and Systematics, 15: 501-522. CARSON, H. L., 1982, Speciation as a major reorganization of polygenic balance. In: WHITE, M. J. D., ed., Mechanisms of speciation, Liss, New York. CAVALLI-SFORZA, L. L. & EDWARDS, А. W. F., 1967, Phylogenetic analysis: models and estima- tion procedures. Evolution, 21: 550-570. CHEATUM, E. P. & FULLINGTON, R. W., 1971, The aquatic and land Mollusca of Texas. Part one. The Recent and Pleistocene members of the gastropod family Polygyridae in Texas. Dal- las Museum of Natural History, Bulletin 1: 1-74. CLARKE, B., 1978, Some contributions of snails to the development of ecological genetics. Pp. 159-170 in: BRUSSARD, P. F. ed., Ecological genetics: the interface. Springer-Verlag, New York. DAVIS, G. M., 1978, Experimental methods in molluscan systematics. Pp. 99-169 in: FRET- TER, V. 8 PEAKE, J., eds., Pulmonates. Vol. 2A. Systematics, evolution and ecology. Academic Press, London. DAVIS, С. M., HEARD, W. H., FULLER, S.L.H. 4 HESTERMAN, C., 1981, Molecular genetics and speciation in Elliptio and its relationships to other taxa of North American Unionidae (Bivalvia). Biological Journal of the Linean Society, 15: 131-150. DELCOURT, P. A. & DELCOURT, H. R., 1981, Vegetation maps for eastern North America: 40,000 yr. B. P. to the present. Pp. 123-165 in: ROMANS, R. C., ed., Geobotany Il. Plenum, New York. DIXON, W. J. & BROWN, M. B., 1979, BMDP-79, Biomedical Computer Programs, P-Series. Uni- versity of California Press, Berkeley, 880 pp. ELDREDGE, N. & CRACRAFT, J., 1980, Phylo- genetic patterns and the evolutionary process: method and theory in comparative biology. Co- lumbia University Press, New York, 349 pp. EMBERTON, K. C., 1981, Ecological notes on two sympatric, conchologically convergent polygyrid snails in Ohio. Bulletin of the American Mala- cological Union, 1980: 27-30. EMBERTON, K. C., 1985, Seasonal changes in the reproductive gross anatomy of the land snail Triodopsis tridentata tridentata (Pulmonata: Polygyridae). Malacologia, 26: 225-239. EMBERTON, K. C., 1986, The evolution of multiple sympatric homeomorphy among three genera of land snails. Doctoral Dissertation, University of Chicago, 780 pp. FARRIS, J. S., 1970, Methods for computing Wagner trees. Systematic Zoology, 19: 83-92. FARRIS, J. S., 1972, Estimating phylogenetic trees from distance matrices. American Naturalist, 106: 645-668. FARRIS, J. S., 1981, Distance data in phylogenetic analysis. Pp. 3-23, in: FUNK, V. A. & BROOKS, О. R., eds., Advances in Cladistics: Proceedings of the First Meeting of the Willi Hennig Society. New York Botanical Garden, Bronx, New York. FARRIS, J. S., KLUGE, A. G. & ECKARDT, M. J., 1970, A numerical approach to phylogenetic systematics. Systematic Zoology, 19: 172-191. FELSENSTEIN, J., 1984, Distance methods for inferring phylogenies: a justification. Evolution, 38: 16-24. FINK, W. L., 1986, Microcomputers and phylo- genetic analysis. Science, 234: 1135-1139. FITCH, W. M. & MARGOLIASH, E., 1967, Con- 252 EMBERTON struction of phylogenetic trees. Science, 155: 279-284. GILL, P. D., 1978a, Non-genetic variation in isoenzymes of lactate dehydrogenase of Cepaea nemoralis. Comparative Biochemical and Physi- ology, 59B: 271-276. GILL, P. D., 1978b, Non-genetic variation in isoenzymes of acid phosphatase and alpha- glycerophosphate dehydrogenase of Cepaea nemoralis. Comparative Biochemistry and Phys- iology, 60B: 365-368. GOODFRIEND, G. A., 1986, Variation in land-snail shell form and size and its causes: a review. Systematic Zoology, 35: 204-223. GORMAN, G. C. & RENZI, J., Jr., 1979, Genetic distance and heterozygosity estimates in electro- phoretic studies: effect of sample size. Copeia, 1979: 242-249. GOULD, S. J. & WOODRUFF, D. S., 1986, Evolu- tion and systematics of Cerion (Mollusca: Pulmonata) on New Providence Island: A radical revision. Bulletin of the American Museum of Natural History, 182: 389-490. GOULD, $. J., WOODRUFF, D. $. 8 MARTIN, J. P., 1975, Genetics and morphometrics of Cerion at Pongo Carpet: a new systematic approach to this enigmatic land snail. Systematic Zoology, 23: 518-535. GRIMM, F. W., 1975, Speciation within the Trio- dopsis fallax group (Pulmonata: Polygyridae)— a preliminary report. Bulletin of the American Malacological Union, 1974: 23-29. HENNIG, W., 1966, Phylogenetic systematics. Uni- versity of Illinois Press, Urbana, 263 pp. HOCHACHKA, P. W. & SOMERO, G. N., 1984, Biochemical Adaptation. Princeton University Press, Princeton, New Jersey, 537 рр. HUBRICHT, L., 1950, The distribution of Triodopsis soelneri (J. B. Henderson) in North Carolina. Nautilus, 64: 67. HUBRICHT, L., 1952, Three new species of Triodopsis from North Carolina. Nautilus, 65: 80-82. HUBRICHT, L., 1953, Land snails of the Southern Atlantic coastal plain. Nautilus, 67: 22-24. HUBRICHT, L., 1958, New species of land snails from the eastern United States. Transactions of the Kentucky Academy of Sciences, 19: 70-76. HUBRICHT, L., 1985, The distributions of the native land mollusks of the eastern United States. Fieldiana, Zoology, New Ser., no. 24: 1-191. HUSTED, L. & BURCH, P. R., 1946, The chromo- somes of polygyrid snails. American Naturalist, 80: 410-429. JOHNSON, M. S., CLARKE, B. & MURRAY, J. J., Jr., 1977, Genetic variation and reproductive isolation in Partula. Evolution, 31: 116-126. KIMURA, M., 1979, The neutral theory of molecular evolution. Scientific American, 241: 94-104. KIMURA, M., 1983, The neutral theory of molecular evolution, Cambridge University Press, Cam- bridge, Massachusetts, 367 pp. KLUGE, А. ©. & FARRIS, С. S., 1969, Quantitative phyletics and the evolution of anurans. System- atic Zoology, 18: 1-32. LUTZ, L., 1950, A list of the land Mollusca of Claiborne County, Tennessee, with a description of a new subspecies of Triodopsis. Nautilus, 63: 99-105, 121-123. McCRACKEN, С. F., 1976, The population biology of the white-lipped land snail, Triodopsis albo- labris. Doctoral Dissertation, Cornell University, 136 pp. McCRACKEN, С. F., 1980, Self-fertilization in the white-lipped land snail Triodopsis albolabris. Bi- ological Journal of the Linnean Society, 14: 429-434. McCRACKEN, С. Е. & BRUSSARD, P. F., 1980, The population biology of the white-lipped land snail, Triodopsis albolabris: genetic variability. Evolution, 34: 92-104. MICKEVICH, M. F. & JOHNSON, M. S., 1976, Congruence between morphological and al- lozyme data in evolutionary inference and char- acter evolution. Systematic Zoology, 25: 260-270. MICKEVICH, M. F. & MITTER, C., 1981, Treating polymorphic characters in systematics: a phylo- genetic treatment of electrophoretic data. Pp. 45-58 In: FUNK, V. A. & BROOKS, О. R., eds., Advances in Cladistics. Proceedings of the First Meeting of the Willi Hennig Society. New York Botanical Garden, Bronx, New York. MILLER, W. B., REEDER, В. L., BABRAKZAI, N. & FAIRBANKS, H. L., 1984, List of new and re- vised recent taxa in the North American terres- trial Mollusca (north of Mexico) published since 19 March 1948. Part 1. Tryonia, 11:1-14. MURRAY, J. J., Jr. & CLARKE, B., 1980, The genus Partula on Moorea: speciation in prog- ress. Proceedings of the Royal Society of Lon- don, B, 211: 82-117. 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. NEVO, Е. С. & BAR, Z., 1976, Natural selection of genetic polymorphisms along climatic gradients. Pp. 159-184, in: KARLIN, $. & NEVO, E., eds., Population Genetics and Evolution. Academic Press, New York. NEVO, Е. C., BAR-EL, C., BAR, Z. & BEILES, A., 1981, Genetic structure and climatic correlates of desert land snails. Oecologia, 48: 199-208. NEVO, E. C., BAR-EL, C., BEILES, A. & YOM- TOV, Y., 1982, Adaptive microgeographic differ- entiation of allozyme polymorphism in land snails. Genetica, 59: 61-67. OXFORD, G. S., 1973, The genetics of Cepaea esterases. |. Cepaea nemoralis. Heredity, 30: 127-139. OXFORD, G. S., 1978, The nature and distribution of food-induced esterase in helicid snails. Mala- cologia, 17: 331-339. EASTERN NORTH AMERICAN TRIODOPSINAE 253 PATTERSON, С. М. & BURCH, J. B., 1978, Chro- mosomes of pulmonate molluscs. Pp. 171-217, In: FRETTER, V. & РЕАКЕ, J., eds., Pulmon- ates. Vol. 2A. Systematics, Evolution and Ecol- ogy. Academic Press, London. PATTON, J. С. & AVISE, J. C., 1983, An empirical evaluation of qualitative Hennigian analyses of protein electrophoretic data. Journal of Molecular Evolution, 19: 244—254. PILSBRY, H. A., 1894, Manual of conchology. Ser. 2: Pulmonata. Vol. 9 (Helicidae, Vol, 7). Guide to the study of helices. Academy of Natural Sci- ences of Philadelphia, 366 pp. PILSBRY, H. A., 1895, Manual of conchology. Ser. 2: Pulmonata. Index to the helices. Academy of Natural Sciences of Philadelphia, 126 pp. PILSBRY, H. A., 1905, On Dorcasia, Trigonephrus, Corilla, Thersites, and Chloritis. Proceedings of the Malacological Society of London, 6: 286-291. PILSBRY, H. A., 1939, Land Mollusca of North America (north of Mexico), Vol. 1, Part 1. Acad- emy of Natural Sciences, Philadelphia, Mono- graphs, no. 3: 1-573. PILSBRY, H. A., 1940, Land Mollusca of North America (north of Mexico), Vol. 1, Part 2. Acad- emy of Natural Sciences, Philadelphia, Mono- graphs, no. 3: 575-994. PILSBRY, H. A., 1946, Land Mollusca of North America (north of Mexico), Vol. 2, Part 1. Acad- emy of Natural Sciences, Philadelphia, Mono- graphs, no. 3: 1-520. PILSBRY, H. A., 1948, Land Mollusca of North America (north of Mexico), Vol. 2, Part 2. Acad- emy of Natural Sciences, Philadelphia, Mono- graphs, no. 3: 521-1113. POULIK, M. D., 1957, Starch gel electrophoresis in a discontinuous system of buffers. Nature, 180: 1477-1479. PRAGER, Е. М. & WILSON, А. C., 1978, Construc- tion of phylogenetic trees for proteins and nucleic acids: empirical evaluation of alternative matrix methods. Journal of Molecular Evolution, 11: 129-142. RANDLES, W. B., 1900, On the anatomy of the genus Acavus, Montfort. Proceedings of the Malacological Society of London, 4: 103-113. RAUP, D. M., 1966, Geometric analysis of shell coiling: general problems. Journal of Paleontol- ogy, 40: 1178-1190. RAVEN, C. P., 1975, Development. Pp. 367-400, In: FRETTER, V. & PEAKE, J., eds., Pulmon- ates. Vol. 1. Functional anatomy and physiology. Academic Press, London. REEDER, R. L. & ROGERS, S. H., 1979, The histochemistry of the spermatheca in four spe- cies of Sonorella (Gastropoda: Pulmonata). Transactions of the American Microscopical So- ciety, 98: 267. RENSCH, B., 1932, Ueber de Abhaengigkeit der Groesse, des relativen Gewichtes und der Oberflaechenstruktur der Landschnecken- schalen von den Umweltsfaktoren. Zeitschrift für Morphologie und Okologie der Tiere, 25: 757-807. RICHARDSON, L., 1986, Polygyracea: catalog of species. Part 1, Polygyridae. Tryonia, 13: 1-139. ROHLF, R. J., KISPAUGH, J. & KIRK, D., 1972, NT-SYS: Numberical taxonomy system of multivariate statistical programs. Technical Re- port, State University of New York, Stony Brook, New York (June 1974 version). ROGERS, J. S., 1972, Measures of genetic simi- larity and genetic distance. University of Texas Publication 7213: 145-153. ROGERS, $. H., REEDER, В. L. & SHANNON, W. A., 1980, Ultrastructural analysis of the morphol- ogy and function of the spermatheca of the pulmonate snail, Sonorella santaritana. Journal of Morphology, 163: 319-329. ROTH, B., 1984, A new species of Vespericola (Gastropoda: Pulmonata: Polygyridae) from the Klamath Mountains, California. Wasman Journal of Biology, 42: 84-91. SARICH, V. M., 1977, Rates, sample sizes, and the neutrality hypothesis for electrophoresis in evo- lutionary studies. Nature, 265: 24-28. SAS INSTITUTE, Inc., 1982, SAS user's guide: statistics, 1982 Ed. Cary, North Carolina, 584 pp. SCHINDEL, D. E. (submitted June 1986), Morpho- metrics of gastropod coiling: new procedures, morphological integration and architectural con- straints. SELANDER, R. K. & OCHMAN, H., 1983, The genetic structure of populations as illustrated by molluscs. Pp. 93-123 in: Isozymes: current top- ics in biological and medical research. Vol. 10: Genetics and evolution. Liss, New York. SELANDER, R. K., SMITH, M. H., YONG, S. Y., JOHNSON, W. E. & GENTRY, J. B., 1971, Biochemical polymorphism and systematics in the genus Peromyscus. |. Variation in the old- field mouse (Peromyscus polionotus). University of Texas Studies in Genetics, 4: 49-90. SELANDER, В. К. 8 WHITTAM, T. S., 1983, Pro- tein polymorphism and the genetic structure of populations. Pp. 89-114 in: NEI, M. 8 KOEHN, R. K. eds., Evolution of genes and proteins. Sinauer, Sunderland, Massachusetts. SHAFFER, H. B., 1984, Evolution in a highly paedomorphic lineage: a case study of the Mex- ican ambystomatid salamanders. Doctoral Dis- sertation, University of Chicago. SHAW, С. В. & PRASAD, R., 1970, Starch gel electrophoresis of enzymes——a compilation of recipes. Biochemical Genetics, 4: 297-320. SIMPSON, G. B., 1901, Anatomy and physiology of Polygyra albolabris and Limax maximus and embryology of Limax maximus. Bulletin of the New York State Museum, 8: 237-314, pls. 1-28. SIMPSON, G. G., 1944, Tempo and mode in evo- lution. Columbia University Press, New York, 237 pp. SNEATH, P. H. A. & SOKAL, R. R., 1973, Numer- ical taxonomy. Freeman San Francisco, 573 pp. SOKAL, R. R. & SNEATH, P. H. A., 1963, Princi- 254 EMBERTON ples of numerical taxonomy. Freeman, San Francisco, 359 pp. SOLEM, A., 1966, Some non-marine mollusks from Thailand,withnotesonclassificationoftheHelicari- onidae. Spolia Zoologica Musei Hauniensis, Copenhagen, 14: 1-110. SOLEM, A., 1972, Microarmature and barriers in the aperture of land snails. Veliger, 15: 81-87. SOLEM, A., 1975, Notes on Salmon River Valley oreohelicid land snails, with description of Oreohelix waltoni. Veliger, 18: 16-30. SOLEM, A., 1976, Comments on eastern North American Polygyridae. Nautilus, 90: 25-36. SOLEM, A., 1978, Classification of the land Mollusca. Pp. 49-98 in: FRETTER, V. 8 PEAKE, J., eds., Pulmonates. Vol. 2A. Systematics, evo- lution and ecology. Academic Press, London. SOLEM, A., 1979, Camaenid land snails from west- ern and central Australia (Mollusca: Pulmonata: Camaenidae). |. Taxa with trans-Australian dis- tributions. Records of the Western Australian Museum, suppl. 10: 1-142. SOLEM, A., 1981a, Camaenid land snails from western and central Australia (Mollusca: Pul- monata: Camaenidae). Il. Taxa from the Kimberley, Amplirhagada lredale, 1933. Records of the Western Australian Museum, suppl. 11: 147-320. SOLEM, A., 1981b, Camaenid land snails from western and central Australia (Mollusca: Pulmonata: Camaenidae). Ill. Taxa from the Ningbing Ranges and nearby areas. Records of the Western Australian Museum, supple. 11: 321-425. SOLEM, A., 1984, Camaenid land snails from west- ern and central Australia (Mollusca: Pulmonata: Camaenidae). IV. Taxa from the Kimberley, Westraltrachia lredale, 1933 and related genera. Records of the Western Australian Museum, suppl. 17: 427-705. SOLEM, A. & CHRISTENSEN, C. C., 1984, Camaenid land snail reproductive cycle and growth patterns in semiarid areas of north-wes- tern Australia. Australian Journal of Zoology, 32: 471-491. SWOFFORD, D. L., 1981, On the utility of the distance wagner procedure. Pp. 25-43, in: FUNK, V. A. & BROOKS, D. R., eds., Advances in cladistics: proceedings of the first meeting of the Willi Kennig Society. New York Botanical garden, Bronx, New York. SWOFFORD, D. L., 1983, PAUP. Phylogenetic analysis using parsimony, Vers. 2.1. Illinois Nat- ural History Survey, Champaign. SWOFFORD, D. L. & SELANDER, R. B., 1981, BIOSYS-1: A FORTRAN program for compre- hensive analysis of electrophoretic data in pop- ulation genetics and systematics. Journal of Her- edity, July/August: 281-283. ТАТЕМО Y., NEI, М. & PAJIMA, S., 1982, Accuracy of estimated phylogenetic trees from molecular data. |. Distantly related species. Journal of Molecular Evolution, 18: 387—404. THROCKMORTON, L. H., 1965, Similarity versus relationship in Drosophila. Systematic Zoology, 14: 221-236. THROCKMORTON, L. H., 1978, Molecular phylo- genetics. Pp. 221-239, In: ROMBERGER, J. A., ed., Biosystematics in agriculture. Wiley, New York. TILLIER, S., 1985, Morphologie comparée, phylo- génie et classification des gasteropodes pulmones stylommatophores (Mollusca). Doc- toral Dissertation, Museum National d'Histoire Naturelle and l'Université Pierre et Marie Curie (Paris). TOMPA, A. S., 1984, Land snails (Stylom- matophora). Pp. 47-140 in: WILBUR, K. M., editor-in-chief, The Mollusca. Vol. 7. Reproduc- tion. Academic Press, New York. VAGVOLGYI, J., 1968, Systematics and evolution of the genus Triodopsis (Mollusca: Pulmonata). Bulletin of the Museum of Comparative Zoology, 136: 145-254. VAIL, V. A., 1978, Laboratory observations on the eggs and young of Triodopsis albolabris major (Pulmonata: Polygyridae). Malacological Re- view, 11: 39-46. VERMEIL, С. J., 1980, Gastropod shell, growth rate, allometry and adult size: environmental implications. Pp. 379-394 in: RHOADS, D. С. & LUTZ, R. A., eds., Skeletal growth of aquatic organisms: biological records of environmental change. Plenum Press, New York. WATROUS, L. E. & WHEELER, Q. D., 1981, On the out-group method of character analysis. Sys- tematic Zoology, 30: 1-11. WEBB, G. R., 1947a, Studies of the sex-organs of mating polygyrid landsnails. /llinois Academy of Science, Transactions, 40: 218-227. WEBB, G. R., 1947b, The mating-anatomy tech- nique as applied to polygyrid landsnails. Ameri- can Naturalis, 81: 134-147. WEBB, G. R., 1948, Comparative observations on the mating of certain Triodopsinae. Nautilus, 61: 97-103. WEBB, G. R., 1951, Sexological notes on the landsnail Oreohelix. Natural History Miscellanea, 78: 1-5. WEBB, G. R., 1952, A sexological revision of some triodopsin land-snails, Xolotrema, Neohelix, & Wilcoxorbis. Gastropodia, 1: 7-8. WEBB, G. R., 1954, The life-history and sexual anatomy data on Ashmunella with a revision of the triodopsin snails. Gastropodia, 1: 13-18. WEBB, G. R., 1959, Pulmonata, Polygyridae: notes on the sexology of Triodopsis, a new subgenus, Haroldorbis, and a new section, Shelfordorbis. Gastropodia, 1: 23-25. WEBB, G. R., 1961, The phylogeny of American land snails with emphasis on the Polygyridae, Arionidae, and Ammonitellidae. Gastropodia, 1: 31—44. WEBB, С. R., 1968, The Ashmunellinae: sexologi- cal notes on Allogona. Gastropodia, 1: 70-72. WEBB, G. R., 1970a, Sexological notes on EASTERN NORTH AMERICAN TRIODOPSINAE 255 Cryptomastix mullani (Bland and Cooper). Gastropodia, 1: 73-75. WEBB, С. R., 1970b, Observations on the sexology of Vespericola columbiana (Lea) from Olympic Peninsula, Washington. Gastropodia, 1: 75-77. WEBB, С. R., 1970c, Fragmentary observations on sexology of Cryptomastix hendersoni Pilsbry and C. magnidentata Pilsbry and a new subgenus (Pulmonata, Polygyridae, Ashmunellinae). Gas- tropodia, 1: 77-78. WEBB, G. R., 1974, The sexual evolution of the polygyrid snails. Gastropodia, 1: 85-90. WEBB, G. R., 1980, The hybridization of the snails Xolotrema denotata and X. caroliniensis in the laboratory (Pulmonata, Polygyridae, Trio- dopsinae). Gastropodia, 2: 1-2. WILEY, Е. O., 1981, Phylogenetics. the theory and practice of phylogenetic systematics. Wiley, New York, 439 pp. WRIGHT, S., 1978, Evolution and the genetics of populations, Vol. 4. Variability within and among natural populations. University of Chicago, Chi- cago, 580 pp. WOODRUFF, D. S., 1978, Evolution and adaptive radiation of Cerion: a remarkably diverse group of West Indian land snails. Malacologia, 17: 223-239. WURTZ, C. B., 1955, The American Camaenidae (Mollusca: Pulmonata)) Proceedings of the Academy of Natural Sciences of Philadelphia, 107: 99-143. APPENDIX A. ELECTROPHORETIC PROCEDURES Loci. The 16 loci are listed in Table 12. Gels. 33 grams starch (Electrostarch Com- pany, lot #392) to 250 ml buffer. Dimensions 18.4 mm x 14.4 mm x 0.6 mm. Paper wicks. Cut from filter paper, dimen- sions 7-8 mm x 1.2-1.3 mm; 25-30, rarely up to 40, per gel. Buffer systems and running times. All were run at 35 milliamperes or 350 volts, whichever was reached first. TC-6, Tris-Citrate pH6 (Shaw & Prasad, 1970): 2.5 hr. Poulik (dis- continuous tris-citrate): 3.5 hr. ТЕВ 9, tris- EDTA-borate pH 9.1 (Ayala et al., 1973): 4.5 hr. TEB 9/8, TEB 9 gel run in TEB 8 (Shaw & Prasad, 1970) tray buffer. Power supply. Heath Schlumberger Regu- lated High Voltage Power Supply; each gel run in a separate tray under a separate power supply. Grinding buffer. Modified from Selander et al. (1971): 0.01 molar Tris buffer, 0.001 molar EDTA, 5 x 10 ° molar NADP, 0.2 parts per thousand beta-mercaptoethanol, pH adjusted to 6.8. For making 500 ml: 0.6055 g Tris, 0.1681 д EDTA, 19.1 ml NADP, 0.01 ml beta-mercaptoethanol. Also used de-ionized distilled water for some tissues, with no de- tectable difference in results. Chemicals. All from Sigma Chemical Com- pany. Staining. Recipes from Shaw and Prasad (1970) unless otherwise indicated in Table 12. Stained in a tray for Got and Lap; for all others, stained using agar overlay: 10 ml of 2% agar solution (4 grams agar to 200 ml water) at 60°C per 10 ml of stain, freshly mixed at room temperature. Agar overlays conserve staining chemicals and allow the gel to be readon a light table as staining pro- ceeds, allowing greater scoring accuracy. Controls. Mesodon zaletus from the popu- lation at Monte Sano, Alabama (GS 20 = GS 101) was used as control on all but two runs which used the same species from White Oak Sink, Tennessee (GS 9). The runs made in 1982 had 5 controls in the center of each gel, and the runs made in 1983 had 2 controls on each end and 3 controls in the center of each gel. Scoring. Banding patterns on gels were measured on a light table and immediately copied onto graph paper to the nearest 0.5 mm, with compensations for apparent edge effects and local distortions. Questionable bands were labeled as such on the graph paper record to aid later interpretation. APPENDIX B. SYSTEMATIC REVIEW OF THE NEOHELIX ALBOLABRIS AND ALLENI GROUPS The studied populations are numbered from 1 to 46 as they appear in Fig. 47. Species group Neohelix albolabris (Figs. 2d-g, 4, 29c-d, 30c-d; Tables 2, 4, 6, 7, Fig. 49) Key characters. Penis: pilastral lappets ap- proximately half the number of columns of wall pustules; pilaster moderately wide; wall pustules all distinct and approximately equal in size; verge large; retractor-muscle's origin distant (ca 1/2 the penial length) from the penial apex; vas deferens more than 4 times as long as the penis. Neohelix albolabris (Say, 1816) (Figs. 2d-g, 29c-d; Tables 2, 4, 6, 7; Fig. 49) 256 EMBERTON TABLE 12. Enzyme systems used for electrophoretic analysis. Sorbitol dehydrogenase' Abbrevia- Name tion Sordh Malate dehydrogenase* Mdh Malic enzyme? Me Isocitrate dehydrogenase* led Phosphogluconate dehydrogenase? Pgd Glucose-6-phosphate dehydrogenase® Gd Superoxide dismutase’ Sod Glutamate-oxaloacetate transaminase® Got Phosphoglucomutase? Pgm Leucine aminopeptidase'° Lap Mannose phosphate isomerase'' Mpi Glucose phosphate isomerase '* Gpi Total Enzyme Buffer Number of commission system Molecular readable number used structure(s) loci 1.1.1.14 ТЕВ 9 Tetramer 1 1.121897 TC 6 Dimers 2 1.1.1.40 ТЕВ 9 Tetramer 1 leat 142 TC 6 Dimer 1 1.1.1.44 ТЕВ 9/8 Dimer 1 1.1.1.49 TEB 9/8 Dimers 2 AAA TEB 9/8 S-1 Dimer 2 S-2 Tetramer 265 TEB 9 Dimers 2 Я, Poulik Monomer 1 3.4.1.1 TC6 Monomer 1 5.3.1.8 ТЕВ 9 Мопотег 1 DIES Poulik Dimer 1 16 ‘Stains slowly, streaks a bit. Clear, stains in a few minutes, keeps well. 3Stain: 5 ml НС! developer, 5 ml MDH substrate solution, MgClo, MTT, NADP (0.15 ml), PMS, 10 ml agar. Stains slowly, keeps well. A second locus comes up with TC 6, but is unreliable. *Stains very slowly, streaks a bit. Second locus visible but too streaked to read. Must be read quickly, blurs badly if left too long. ®Second locus does not appear unless 5 mg NADP is added to gel before degassing, as per Brewer (1970). First locus blurs and streaks quickly, second is slow and keeps well. "Comes up slowly. Better if left under fluorescent light. Disappears with time. “Sometimes called asparate amino transferase. Stained in tray, recipe from Selander et al. (1971). Soluble (anodal) locus stains faster than mitochondria (cathodal) locus. Both streak, but in one direction, so clearly readable. °Strong satellite bands which had to be learned and discounted. A second locus is clear, but with too much overlap with the first locus to be scored. ‘°Stained in tray. Stains very slowly and keeps well. A second, slow locus is too streaked to read reliably. "Stain recipe from Nichols, Chapman & Ruddle (1973). Stains at a moderate rate, keeps well. '2Stains quickly and soon blurs with formation of satellite bands. Comparisons Penis. On its pilaster a/bolabris differs from major by the density of lappets, having more per unit length (Table 4); and by the shape of the pilastral lappets, having slightly as op- posed to greatly convex surfaces (Fig. 2d, e vs. Fig. 4). The wall pustules of a/bolabris are smaller than in major (Table 4). Shell. N. albolabris has fewer striae per unit distance than major (Table 7). It also differs from major in having a lower whorl expansion rate and a much smaller baso-columellar lip node (Table 7). Key characters Penis: internal length 10-16 mm; pilaster 1/20th to 1/10th as broad as the penis is long, and bearing 8-14 lappets per 2.6 mm; lappet surfaces slightly convex; wall-pustular col- umns 16-24 per 2.6 mm; verge 1/7th to 1/5th as long as the penis. Shell: diameter 23-39 mm, depressed-glo- bose, whorls 5 1/2-6; striae moderately raised, 17-26 per 2.6 mm on the 5th whorl; yellow-brown to brown; glossy to dull; whorls slowly expanding for the group; apertural lip narrow to wide for the group; basocolumellar node absent to inconspicuous; pre-apertural deflection of the body whorl moderate to weak. Neohelix albolabris albolabris (Say, 1816) (Figs. 2d-g, 29c-d; Tables 2, 4, 6, 7; Figs. 47, 49) Studied material (10) OH: Athens County (Ohio 35; FMNH 214917): 12 live adults—dissected #A, B; measured shells #A, B, C. (11) PA: Chester County (GS 129; FMNH 214919): 2 live EASTERN NORTH AMERICAN TRIODOPSINAE 257 adults, 4 tissue samples—dissected #3, 4 (measured #3); electrophoresed #1, 2, 3, 4; measured shells #3, 4. (12) WV: Preston County (GS 130; FMNH 214920); 20 live adults, 20 tissue samples—dissected #9, 11, 14 (illustrated #14); electrophoresed 8, 12, 16, 17, 20; measured shells #12, 14, 17 (illustrated #14). (13) WV: Greenbrier County (GS 139; FMNH 214921): 2 live adults— dissected #A, B; measured shells #A, B. (14) WV: Boone County (GS 142; FMNH 214922): 9 live adults—dissected #A, B, C (measured #A); measured shells #A, В, С. (15) МС: Watauga (GS 151, 152; FMNH 214924): 9 live adults—dissected #A, B, C (measured FA); measured shells +A, В, С. Published dissections (0) PA? (Binney, 1851, Plate VI, Fig. IV). (16) NY: Albany County (Simpson, 1901, Plate 8, Figs. 2, 3, 4, 6). (17) PA: Bucks County (Pilsbry, 1940, Fig. 488:7). (18) IN: Monroe County (Webb, 1952), Plate 4, Fig. 12:7). (19) IN: Monroe County (Webb, 1954, Plate 10, Fig. 9:16). Key characters Shell: striae 18-23 per 2.6 mm on the 5th whorl; color light to dark brown; surface dull; height to diameter ratio .59—.72; whorls 5.2 to 6.0; outer lip width 15-30 mm; pre-apertural body whorl deflection moderate. Neohelix albolabris bogani Emberton, new subspecies (Tables 2, 4, 6, 7; Figs. 47, 49) Synonymy Xolotrema albolabris alleni (“Wetherby” Sampson) of Webb, 1952, Gastropodia, 1 (1): 7-8, Figs. 2, 13. Triodopsis albolabris alleni (Wetherby) of Solem, 1976, Nautilus, 90: 25-36, Figs. 1a, b, 2a; с, 8-12. Studied material (23) OK: Sequoyah County (FMNH 176127): 2 live adults—dissected #A, B; measured shells #A, B. (20) TX: Houston County (GS 76; FMNH 214925): 10 live adults, 10 tissue samples—dissected #1, 5, 8. (21) AR: Crawford County (GS 90; FMNH 214926): 2 live adults, 2 tissue samples— dissected #3, 4; electrophoresed #1, 2. (22) AR: Logan County (FMNH 176087): 2 live adults—dissected #A (measured #A); mea- sured shells #A, B. (23) OK: Sequoyah County (FMNH 176144): 1 live adult—dis- sected; measured shell. (24) AR: Washington County (FMNH 176160): 1 live adult—dis- sected; measured shell. (25) AR: Washington County (FMNH 176160): 1 live adult—dis- sected; measured shell. (26) LA: Washington County (FMNH 195989): 1 live adult—dis- sected; measured shell. Published dissections (27) AR: Logan County (Webb, 1952, Plate 4, Fig. 13). (23) OK: Sequoyah County (Solem, 1976, Fig. 5a)—also included in stud- ied material above. Comparisons Neohelix albolabris bogani has previously been confused with Neohelix alleni alleni, from which it differs by its penial morphology (Figs. 2d vs. Fig. 3; Table 4) and subtle aspects of shell morphology (Table 7). The two occur sympatrically at Devils Den State Park, Crawford County, AR (Fig. 46, popula- tions 3 and 21). This is the western subspecies of Neohelix albolabris, occurring west of the Mississippi from at least Texas to Arkansas, but also getting east of the River in the Delta area (Fig. 46, population 26). It differs from the eastern N. albolabris by several shell characters which are convergent on western alleni: yellower color, glossier surface, moderately higher spire, and narrower lip (Table 7). It also differs from N. albolabris by its denser striae, its slower whorl expansion rate, and its stronger pre-apertural deflection (Table 7). In penial morphology (Table 4) and electromorphs (Ta- ble 2, Fig. 27) it shows no significant differen- tiation from albolabris albolabris. By shell characters, albolabris bogani can usually be distinguished from the sometimes sympatric alleni alleni by its denser striae, slower whorl expansion rate, smallness of the baso-columellar node, narrower lip, and more pronounced preapertural deflection (Table 7). At Devils Den State Park, albolabris bogani was smaller in diameter than alleni; it is not known whether they were microsympatric, as the collection covered a wide area of hard- wood forest. 258 EMBERTON Key characters Shell: striae 18-26 per 2.6 mm on the 5th whorl; color yellow-brown to light brown; sur- face glossy; height-to-diameter ratio .58-.68; whorls 5.2-5.7. Remarks Despite the virtually identical penial mor- phology and electromorphs, the disjunct shell morphologies and geographic ranges clearly indicate subspecific status for bogani. The pre- cise range relationships still need to be worked out (to fill in the gaps in Fig. 47) before sound hypotheses can be formulated about the rel- ative time of separation of the two subspecies of albolabris, but it appears likely that the Mis- sissippi River has kept them isolated for the past 20,000-40,000 years (Delcourt 8 Del- court 1981). Pilsbry (1940: 842) reported an introduction of Neohelix albolabris from North Carolina to Tyler, Texas, presumably in the 1890's (Fig. 47). Although it is tempting to speculate that this was the founder of albolabris bogani, both the degree of con- chological differentiation and the widespread occurrence of this subspecies in an arid terrain argue strongly against such a theory. The shell convergence on alleni alleni by which albolabris bogani has until now es- caped detection, is intriguing. Working out the degree of range overlap and sympatry of these two trans-Mississippian species would be a worthy contribution to malacology by providing valuable data on the sympatric- convergent evolution of shell morphology and color. This subspecies is named for Dr. Arthur Bogan of the Department of Malacology, Academy of Natural Sciences of Philadelphia. Neohelix major (Binney, 1837) (Figs. 4, 30c, d; Tables 2, 4, 6, 7; Figs. 47, 49) Studied material (28) TN: Blount County (GS-3; FMNH 214927): 7 live adults, 7 tissue samples— dissected #35; electrophoresed #1, 6, 7, 10, 22, 24, 28; measured shells #1, 3, 35. (29) TN: Meigs County (GS-105; FMNH 214928): 3 live adults, 3 tissue samples—dissected #2, 3 (measured #3); electrophoresed #1, 2, 3; measured shells #1, 2, 3. (30) SC: Mc- Cormick County (GS-176; FMNH 214930): 13 live adults, 13 tissue sample—dissected #6, 7, 8 (measured #7; illustrated #6); electro- phoresed #1, 2, 3, 6, 10; measured shells #4, 6, 11 (illustrated #H). (31) SC: Aiken County (GS-179; FMNH 214933): 11 live adults—dissected #A, B, C (measured #A); examined 3 partially everted penes; mea- sured shells #A, B, C. (32) AL: Cleburne County (GS-180; FMNH 214935): 1 live adult—examined partially everted penis; measured shell. Comparisons Penis. N. major has the largest pilastral lappets—twice as many per unit length as albolabris (Table 4)—with the most convex, wavy surfaces. This species also has the largest wall pustules of the albolabris group (Table 4). In all other aspects it is similar to N. albolabris, and in fact much resembles an overgrown version of this species (compare Figs. 4 and 29). Shell. N. major has the least glossy shell with the relatively narrowest lip of both the albolabris and alleni groups (Table 7). Its striae are less dense than any of these taxa except albolabris bogani, in which the striae are much lower and less distinct (Table 7). The shells of N. major and Mesodon normalis are often sympatric and sometimes indistin- guishable (Emberton, 1986). Key characters Penis: internal length ca 17 mm and rela- tively invariable; pilaster ca 1/10th as broad as the penis is long, and bearing 4—5 lappets per 2.6 mm; lappet surfaces wavy and very convex; verge 1/8th to 1/7th as long as the penis. Shell: diameter 27-40 mm, depressed-glo- bose, whorls 5 1/2-6; striae moderately raised, 16-34 per 2.6 mm on the 5th whorl; brown to dark brown; dull; whorls moderately expanding for the group; apertural lip rela- tively narrow for the group; basocolumellar node generally conspicuous; pre-apertural deflection moderate. Remarks Although the differences in penial morphol- ogy between major and albolabris (Figs. 4 and 2d) are arguably slight enough to denote only subspecific distinction, the available ev- idence supports Hubricht’s recognition of ma- EASTERN NORTH AMERICAN TRIODOPSINAE 259 jor as a full (sister) species. The penial differ- ence is extremely consistent and uniform geographically, and although albolabris and major have never been found sympatric, their differences in penial size and sculpture are comparable to those between sympatric albolabris and dentifera (Fig. 2a, d). The shell differences between major and albolabris are distinct (Table 7) and disjunct, with no sign of clinal or hybridal intergradation. The electro- phoretic difference is small (Table 2) but on the order of that found among other species pairs of the albolabris and alleni groups (Fig. 26, 27). There is a need for more fieldwork in Virginia to test for range overlap or intergradation. Species Group Neohelix alleni (Figs. 7, 6b, 30a-b, 32c-d; Tables 2, 4, 6, 7, Figs. 47, 49) Key Characters Penis: pilastral lappets approximately equal to the number of columns of wall pustules; pilaster moderately wide to narrow; wall col- umns with distinct pustules or locally smooth; wall pustules equal in size, or large except baso-laterally; verge moderate to minute, api- cal or dorsally subterminal; retractor-muscle origin close (less than 1/3rd the penial length) to the penial apex; vas deferens less than 2 1/2 times as long as the penis. Neohelix alleni (Sampson, 1883) (Figs. 3, 30a-b; Tables 2, 4, 6, 7; Figs. 47, 49) Comparisons Penis. N. alleni differs markedly from the albolabris group (albolabris and major) by its much shorter vas deferens, its retractor mus- cle attachment very close to the penial apex, its relatively short verge, and its flat- and smooth-surfaced, tightly appressed, densely packed pilastral lappets (Table 4, Fig. 3). Its differences from N. solemi are discussed un- der that species. Shell. The only single character which dis- tinguishes the shell of alleni from other spe- cies of the albolabris and alleni groups is its relatively faster whorl expansion rate (Table 7). It can also be separated from albolabris and solemi by its pronounced baso- columellar node (Fig. 30a), and from major by its glossier surface, yellower color, and denser striae (Table 7). Key Characters Penis: internal length 10-18 mm; pilaster ca 1/10th as broad as the penis is long, and bearing 15-18 lappets per 2.6 mm; lappet surfaces flat and smooth, lappets closely ap- pressed; verge apical, ca 1/10th as long as the penis. Shell: diameter 23-38 mm, depressed to depressed-globose, whorls 5-6; striae rela- tively low, 16-26 per 2.6 mm on the 5th whorl; yellow to yellow-brown; glossy; whorls rapidly expanding for the group; baso-columellar node large and conspicuous; preapertural de- flection slight to very slight. Neohelix alleni alleni (Sampson, 1883) (Fig. 3a; Tables 2, 7; Figs. 47, 49) Studied material (1) IA: Lynn County (GS-17; FMNH 214908): 1 live adult, 1 tissue sample—dis- sected; electrophoresed. (2) lA: Jackson County (GS-18; FMNH 214909): 1 live adult, 1 tissue sample—dissected; electrophoresed. (3) AR: Crawford County (GS-90; FMNH 214910): 8 live adults, 8 tissue samples— dissected #4, 5, 6; electrophoresed #1, 3, 4, 5, 8; measured shells +1, 2, 8. (4) AR: Izard County (GS-97; FMNH 214911): 15 live adults, 15 tissue samples—dissected #11, 12, 13 (measured #13; illustrated #12); electrophoresed #1, 2, 4, 5, 6, 8. (5) AR: Izard County (GS-98; FMNH 214913): 4 live adults—illustrated shells #A, B (illustrated #B). (6) IA: Clayton County (FMNH 171135): 1 live adult—dissected (measured); illustrated shell. (7) AR: Izard County (FMNH 176221): 1 live adult—dissected; illustrated shell. Comparison This is the western (trans-Mississippian), typical, widespread subspecies of alleni. It differs from ¡ts eastern counterpart in its high- er-spired, yellower, glossier, more densely striate, and generally smaller shell, with a slower whorl expansion rate, a less pro- nounced baso-columellar node, and a more pronounced pre-apertural deflection (Table ZA): 260 EMBERTON Key characters Shell: diameter 24-30 mm, depressed-glo- bose, whorls 5-6; striae low, 16-22 per 2.6 mm on the 5th whorl; yellow to yellow-brown: glossy; whorl expansion rate low for the spe- cies; apertural lip wide for the species; baso- columellar node small for the species; pre- apertural deflection pronounced for the species. Neohelix alleni fuscolabris (Pilsbry, 1903) (Fig. 3c; Tables 2, 7; Figs. 47, 49) Studied material (8) AL: Madison County (GS-20; FMNH 214 ): 14 live adults, 17 tissue samples— dissected #7, 11, 15; electrophoresed #1, 2, 4,5, 7, 8, 9, 10, 11, 12, 13; illustrated shells #7, 11. (9) AL: Madison County (GS-101; FMNH 214 ): 4 live adults, 4 tissue samples— dissected #1, 2 (measured #2; illustrated #1); illustrated shell #1. Comparison This is the disjunct eastern subspecies of alleni. For shell differences, see comparative remarks under subspecies alleni alleni above. Key characters Shell: diameter 33-39 mm, depressed, whorls 5 1/2-6; striae moderately raised, 17-20 per 2.6 mm on the 5th whorl; brownish yellow; dull for the species; whorls rapidly expanding for the species; apertual lip rela- tively narrow for the species; basocolumellar large and pronounced for the species; pre- apertural deflection very slight for the species. Neohelix solemi Emberton, new species (Figs. 6b, 32c-d; Tables 2, 4, 6, 7; Figs. 47, 49) Synonymy Helix albolabris var. maritima Pilsbry, 1890, Proc. Acad. Nat. Sci. Phila., p. 283, 3 figs. (shell, genitalia, radular teeth). Pilsbry, 1892, Nautilus 5: 142. Walker, 1906, Ill. Cat. Moll. Michigan, part 1, р. 465, fig. 13. Cockerell, 1918, Nautilus 31: 108 (Ram Island, MA). Not Helix maritima Draparnaud, 1805. Triodopsis albolabris form traversensis (Leach) Pilsbry, 1940, Land Moll. North Amer., pp. 836-839, Fig. 489 #9 (shell). Hackney, 1944, Nautilus, 58: 56 (Beaufort, NC). Jacobson, 1945, Nautilus, 59: 68 (Westchester County, NY). Alexander, 1947, Nautilus, 60: 97 (Cape May Point, NJ). Triodopsis albolabris (Say) of Rehder, 1949, Nautilus, 62: 121 (Lake Waccamaw, Columbus County, NC); and, in part, of Mc- Cracken & Brussard, 1980, Evolution, 34: 92 (“Moriello Orchard”, NY, and “Appledore Is- land”, ME, populations). Triodopsis albolabris albolabris (Say) plus T. a. major (Binney), in part, of Vagvolgyi, 1968, Bull. Mus. Comp. Zool., 136: 145 (northeastern Coastal Plain). Triodopsis albolabris (Say) plus T. major (Binney), in part, of Hubricht (1985) (north- eastern Coastal Plain). Studied material (holotype and paratypes) (33) NC: Catawba County (GS-32; FMNH 214936): 1 live adult—dissected (illustrated); illustrated shell (illustrated). (34) NC: Colum- bus County (GS-39; FMNH 214937): 1 live adult—dissected; illustrated shell. (35) SC: Williamsburg County (GS-41; FMNH 214939): 1 live adult—dissected (measured); illustrated shell. (36) NC: Columbus County (GS-164; FMNH 214941): 3 live adults—dis- sected #A, B, C (measured #B); illustrated shells #A, B, C. (37) NC: Columbus County (GS-165; FMNH 214942): 1 live adult—dis- sected. (38) NJ: Cape May County (GS-208; FMNH 214943): 1 live adult, 3 tissue sam- ples—dissected #1, electrophoresed #1, 2, 3. (39) NC: Scotland County (SC-101; FMNH 214945) (HOLOTYPES): 3 live adults, 7 tis- sue samples—dissected #5, 6, 7. (40) NC: Scotland County (SC-103; FMNH 214946): 3 live adults, 3 tissue samples—dissected #1, 2, 3. (41) NC: Wake County (SC-138; FMNH 214947): 1 live adult, 1 tissue sample—dis- sected. (42) NC: New Hanover County (SC-277; FMNH 214948): 5 live adults—dis- sected #A, B, C, D, E. (43) NJ: Cape May County (ANSP 63869-A2432): 6 live adults— dissected #A. (44) NJ: Cape May County (ANSP 72764-A2431): 13 live adults—dis- sected #A, B. (46) NY: Westchester County (ANSP 181296-A2410): 1 live adult—dis- sected. Comparisons Penis. This species is unique within Neo- helix in having a (dorsally) subterminal pore, a EASTERN NORTH AMERICAN TRIODOPSINAE 261 ventral pilaster, a greatly reduced dorsal pi- laster, and a greatly elongated basal penis (Fig. 6b). Shell. The only shell character distinguish- ing N. solemi from other members of the albolabris and alleni groups is its relatively dark brown color (Table 7), but there is over- lap in color (Table 6), so this is not reliable for identification. Practically speaking, solemi need only be distinguished from N. albolabris and major, both of whose ranges appear to be parapatric to it (Fig. 47). It usually differs from N. albolabris by its narrower apertural lip, slightly higher spire, and slightly slower whorl expansion rate; it differs from major by the weakness of its baso-columellar node and by its slightly higher spire, slower whorl expan- sion rate, glossier surface, and slightly denser striae (Table 7). These differences are based on statistical comparisons of small samples however, and should be used only as guide- lines, not as absolutes, for identification. The only reliable way to distinguish solemi from albolabris or major is by dissection, as shown by two misclassified shells in the discriminant analysis, discussed above. Key characters Penis: internal length 12-17 mm; pilaster 1/50th to 1/25th as broad as the penis is long, and bearing 14-15 lappets per 2.6 mm; pilas- ter abbreviated in length by the subterminal pore position; pore dorsally subterminal, mounted on a thick, fleshy pedestal; verge 1/100th to 1/20th as long as the penis; basal penis long, with 1/2 or more of the total penis length lying between the vaginal opening and the base of the sheath. Shell: diameter 24-35 mm, depressed-glo- bose, whorls 5—6; striae moderately raised, 16-21 per 2.6 mm on the 5th whorl; dark brown; moderately dull; whorl expansion rate relatively slow; relative apertural lip width relatively low; pre-apertural deflection moder- ate. Remarks Pilsbry (1940: 839) noticed one of the an- atomical distinctions of this species—its short vas deferens—but was led by shell similari- ties to synonomyze it with N. albolabris traversensis (Leach) of Traverse City, Michi- gan, and nearby localities. Except for this Michigan disjunct, he reported its range (based on shell material) as Coastal Plain Maine to North Carolina; this conforms well with distributional findings based on anatom- ical studies (Fig. 47), which further extend the range into Coastal Plain South Carolina. No material north of Westchester County, New York has yet been anatomically verified, to my knowledge, but electrophoretic and con- chological comparisons have convinced me (Emberton, McCracken, & Wooden, in prep- aration) that solemi occurs in Ulster County, New York (FMNH 214952) and York County, Maine (FMNH 214950). Although | have not yet dissected topotypic N. albolabris traversensis (Leach), | have little doubt about its being a different species because of its extreme western disjunction from the known range of solemi (Fig. 47). The discriminant analysis, as discussed above, has shown that the albolabris albolabris shell can be mis- taken for that of solemi. This species is named for Dr. Alan Solem, Curator and Head, Division of Invertebrates, Field Museum of Natural History, Chicago, eminent terrestrial malacologist and mentor. APPENDIX C. SYSTEMATIC REVIEW OF THE SUPRASPECIFIC TAXA OF THE EASTERN AMERICAN TRIODOPSINAE The eastern American triodopsines differ from all other polygyrids in having a single dorsal pilaster in the upper penis. As in the polygyrid genera Vespericola, Cryptomastix, and Allogona, they have a penial sheath, a retentor muscle, an upper penis, and the penial retractor muscle attaches to the vas deferens (Fig. 11), but they differ from these other genera in the following ways. First, eastern triodopsines lack both epiphallus and flagellum, both of which are present, although not always conspicuous, in Vespericola, Cryptomastix, and Allogona. Second, the basal penis of eastern triodopsines is never wider than, nor longer than, the upper penis and never contains any lobes, flaps, or non- random folds; this differentiates them from both Cryptomastix and Allogona. Third, when eastern triodopsines have a verge, it is al- ways flat with a subterminal pore and terminal papillae; the verge of Vespericola differs in being roundly conical with a simple terminal pore. Fourth, when the shells of eastern triodopsines are large and toothless, they are also always imperforate, and therefore are readily distinguishable from the widely umbilicate large shells of Allogona (Fig. 46). 262 EMBERTON Therefore, any snail which (a) is east of the 100th meridian, (b) has a polygyrid shell which is not Allogona profunda (see Fig. 46), and (c) has a penial sheath (and retentor muscle)——‘s a triodopsine. The penial char- acters are essential for identification because on shell characters alone many eastern triodopsines are easily confused or even in- distinguishable from the geographically over- lapping polygyrine genus Mesodon (see Pilsbry, 1940; Solem, 1976; Emberton, 1986). Anatomically these two lineages are readily distinguishable by the external aspect of the uneverted penis: Mesodon lacks the penial sheath, the retentor muscle, and the thick- ened spermathecal duct of triodopsines, and its penial retractor muscle inserts on the apex of the penis rather than on the vas deferens. In addition, the thick spermathecal duct of triodopsines distinguishes them from Meso- don. (An easy, though destructive way to identify an adult polygyrid as a Mesodon or an eastern-triodopsine in the field is to lightly step on it: the penis can then be diagnosed.) Genus Webbhelix Emberton, new genus (Figs. 6a, 32a—b; Table 2; Fig. 49) Comparisons Webbhelix is unique among triodopsines in having the dorsal pilaster covered with uni- form pustules equal in size to the wall pus- tules (Fig. 6a). It is also the only triodopsine known to have spiral color bands on the shell (Fig. 32b), although these are not always present. Key characters Penis: pilaster approximately 3/4-length, abruptly truncated basally, and covered with uniform sharply-pointed pustules equal in size to wall pustules; wall pustules arranged in approximately 25 contiguous longitudinal col- umns and partially fused along their columns basally; verge large, with two broad and prominent terminal papillae, and smooth-sur- faced. Shell: diameter 14.5-32 mm, depressed- globose, whorls 5 1/2-6; imperforate; thin, thin-lipped; usually marked with reddish- brown color bands. Discussion This genus, which occupies a basal phylogenetic position, is named for Dr. Glenn R. Webb, recently retired from Kutztown Uni- versity, Pennsylvania, whose forty years of dedicated research and publishing are so basic to our understanding not only of eastern triodopsines but of many other North Ameri- can land pulmonates. Webbhelix multilineata (Say, 1821) (Figs. 6a; 32a—b; 32a—b; Table 2; Fig. 49) Studied material (1) IL: Marshall County (GS 127; FMNH 214848): 2 live adults, 2 tissue samples— dissected #2 (illustrated #2); electro- phoresed #1; illustrated shell #2. (2) IL: Kane-Cook Counties: (GS 207; FMNH 214849): ca 11 live adults, 15 tissue sam- ples—dissected #1, 5, A; electrophoresed #1, 3, 5. (3) IL: Calhoun County: (Hubricht 48600) ca 15 live adults—dissected #A, В, С. Published anatomies (1) Binney 1851, Plate VIII. (2) Webb 1948, Figs. 2, 2a. (3) Webb 1952, Plate 5, Figs. 1-8. (4) Webb 1954, Plate 10, Fig. 10. Discussion Webb (1952: 8) elevated Pilsbry's (1940: 850) form chadwicki (Ferriss, 1907) to a full species, but both Vagvolgyi (1968) and Hubricht (1985) synonymized it with multi- lineata. Genus Neohelix von Ihering, 1892 (Figs. 2-5, 6b, 29-31, 32c-d; Table 2; Fig. 49) Comparisons Neohelix is the only genus of eastern triodosines which has pilastral lappets (Figs. 2b, e; 5c, f). It is the only genus besides Webbhelix which has its wall pustules ar- ranged in 25-35 contiguous, longitudinal col- umns; which has a large verge, although its verge size varies; and which has a large, toothless, imperforate shell. Its shell and apertural lip seem to be always thicker than in Webbhelix and its shell is never banded as in Webbhelix. Neohelix also differs from EASTERN NORTH AMERICAN TRIODOPSINAE 263 Xolotrema and Triodopsis in never having a ventrally subterminal pore and in never hav- ing either a palatal or a basal apertural bar- rier. It differs from Triodopsis in having a closed umbilicus. Generally, any eastern triodopsine with a shell which is imperforate, smooth-lipped, and unbanded is a Neohelix. Key characters Penis: dorsal pilaster full-length, smoothly terminating basally, and armed with lappets, lappet number either approximately the same or approximately twice the number of col- umns of wall pustules, pilaster rarely vestigial; wall pustules arranged in 25-35 contiguous, longitudinal columns, and either uniform in size or larger basally; verge large to vestigial, always with a corded surface and thin termi- nal papillae; pore terminal or, rarely, dorsally subterminal. Species Group Neohelix albolabris See Appendix B. Species group Neohelix alleni See Appendix B. Species group Neohelix dentifera (Figs. 2a—c, 5, 29a—b, 31; Table 2; Fig. 49) Comparisons Penis. The dentifera group differs from other Neohelix in the doubled number of its pilastral lappets (Figs. 2a, 5a, d) and the incomplete lateral fusion of the lappets' com- ponent pustules (Figs. 2b, 5c, f), as well as in the enlargement of its basal-most wall pus- tules (Figs. 2a, 5a, d). Shell. The dentifera group's shell is much more depressed than in other Neohelix (Figs. 29b, 31b, d). The only species of this group which occurs east of the Mississippi, dentifera, is readily distinguished from all other eastern Neohelix by its well developed parietal tooth and wide apertural lip (Fig. 31a); the parietal tooth which occurs rarely in albolabris (e.g. Pilsbry, 1940, fig. 489 #8) is always much weaker than dentifera’s. Key characters Penis: pilastral lappets equal in number to approximately twice the number of columns of body wall pustules; pilastral pustules compris- ing the lappets only partially fused laterally; basal-most wall pustules large; verge large, terminal. Shell: diameter 16-30 mm, depressed, whorls 4 1/2-5 1/2; parietal tooth absent or strong; lip smooth or with a small bump suggesting a basal tooth or lamella (Fig. 29a); lip narrow to broad; striae weak to moderately strong. Species subgroup Neohelix dentifera (Figs. 2a—c, 29a—b; Table 2; Fig. 49) Key characters Penis: basal-most 2-3 layers of wall pus- tules enlarged. Shell: diameter 20-30 mm, whorls 5-5 1/2; parietal tooth strongly developed; apertural lip very thick and wide; basal lip sometimes with a bump suggesting a tooth or lamella; striae moderately strong. Neohelix dentifera (Binney, 1837) (Figs. 2a—c, 29a—b; Table 2; Fig. 49) Studied material (1) WV: Preston County (GS-130; FMNH 214809): 20 live adults, 20 tissue samples— dissected #2, 7, 14; electrophoresed #1, 2, 5, 15, 16. (2) WV: Pendleton County (GS-134; FMNH 214810): 10 live adults, 10 tissue samples—dissected #1, 4, 8 (illustrated #8); electrophoresed #5; illustrated shell #8. Species subgroup Neohelix divesta (Figs. 5d, 21; Table 2; Fig. 49) Key characters Penis: Basal-most 8-15 layers of wall pus- tules enlarged. Shell: Diameter 14-18 mm; whorls 4 1/2-5; parietal tooth always absent; apertural lip evenly narrow and always perfectly smooth internally; striae weak. Neohelix divesta (Gould, 1848) (Figs. 5d-f, 31c-d; Table 2; Fig. 49) Studied material (1) AR: Crawford County (GS-90; FMNH 214813): 1 live adult, 2 tissue samples— dissected #1, 7, 8, 10 (illustrated #1); 264 EMBERTON electrophoresed #1, 2; illustrated shell #A (FMNH 214815). (2) AR: Logan county (GS-95; FMNH 214814): ca 4 live adults, 19 tissue samples—electrophoresed #3, 9, 13, 16, 18. Neohelix lioderma (Pilsbry, 1902) (Figs. 5a-b, 31a-b; Table 2; Fig. 49) Studied material (1) OK: Tulsa County (GS-82; FMNH 214844): 9 live adults, 15 tissue sample— dissected #9, A, В, С (illustrated +A); electrophoresed #1, 5, 6, 7, 9, 10, 12, 13, 14, 15; illustrated shell #A. Remarks N. lioderma was originally described as subspecies of the polygyrine Mesodon indianorum (see Pilsbry, 1940). It is obviously a very recently derived diminutive of divesta, with a restricted, relict range peripheral to that of divesta (Fig. 49). Genus Xolotrema (Rafinesque, 1819) (Figs. 7, 8, 33, 34; Table 2; Fig. 49) Comparisons Penis. Xolotrema differs from the other three genera of eastern triodopsines by the gradual dorsal enlargement of its pustules (wall-to-pilaster); its Type 3 chevron; and its very small, apical or ventrally subterminal— never dorsally subterminal—verge. Shell. Conchologically, Xolotrema is unique among eastern triodopsines in its long, smoothly curved parietal tooth which never abruptly changes height; its long, blade-like basal lamella; and its basally-pointing palatal tooth (Figs. 33a, c, e; 34a, c). It includes the only triodopsines with an angular (Figs. 33d, 34b, d) or keeled (Fig. 33f) periphery, or with hair-like periostracal processes (Fig. 33a, b). Xolotrema can always be distinguished from Webbhelix and Neohelix by its possession of a palatal tooth and a basal lamella, and from Triodopsis by the complete coverage of its umbilicus by an extension of the reflected apertural lip in the adult. Key characters Penis: pustules gradually enlarging dor- sally, largest on the pilaster; pilastral pustules arranged either in a single column of abutting cubes or in 5 broad, nested A-shapes; wall pustules arranged in tapered, slightly sepa- rated columns all merging ventrally into 6-10 U-shapes; verge small, bearing 4-6 narrow terminal papillae; verge either terminal or ventrally subterminal and apically directed; everted penis either tubular or shaped like an everted pear; ventral sperm groove present or absent; sheath either covering entire upper (uneverted) penis or covering less than half. Shell: diameter 8-27 mm, depressed, whorls 4 1/2-5 1/2; periphery keeled, angular, or rounded; parietal tooth long, high-standing, gently arched, smoothly decreasing in height toward the umbilicus; basal barrier in the form of a long, blade-like lamella; palatal tooth very strong to weak, pointing downward toward the basal lamella; striae either very to moderately strong, or weak and masked by dense hair- like periostracal processes. Species group Xolotrema fosteri (Figs. 8, 34; Table 2; Fig. 49) Key characters Penis: pilastral pustules a single column of abutting cubes; verge terminal, bearing 6 terminal papillae; everted penis tubular; ven- tral sperm groove present; sheath entirely covering uneverted upper penis. Shell: diameter 14-20, whorls 4 1/2-5 1/2; periphery slightly angled or with an angled shoulder; palatal tooth moderate to weak; striae moderately strong to strong; surface free of pustules or hair-like processes. Xolotrema fosteri (F. C. Baker, 1932) (Figs. 8a, 34a-b; Table 2; Fig. 49) Studied material (1) KY: Hancock County (H-22; FMNH 214817): 1 live adult—dissected #A, В, С, D, E (illustrated #A); illustrated shell #15. (2) KY: Hancock County (GS-15; FMNH 214819): 24 live adults, 24 tissue samples— dissected #19; electrophoresed #12, 13, 14, 15.102 17. 18: 19421327 Xolotrema occidentalis (Pilsbry & Ferriss, 1907) (Figs. 8b-c, 34c-d; Table 2; Fig. 49) EASTERN NORTH AMERICAN TRIODOPSINAE 265 Studied material (1) AR: Independence County (GS-99; FMNH 214855): 1 live adult, 10 tissue sam- ples—electrophoresed #2, 3. (2) AR: Inde- pendence County (GS-100; FMNH 214856): 5 live adults, 10 tissue samples—dissected #5 (illustrated #5); electrophoresed #1, 2, 3, 4, 10; illustrated shell #5. Species group Xolotrema denotata (Figs. 7, 33; Table 2; Fig. 49) Key characters Penis: pilastral pustules in 5 broad, nested A-shapes; verge subterminal, apically di- rected, bearing 4 terminal papillae; everted penis shaped like an inverted pear; ventral sperm groove absent; sheath covering less than half the uneverted upper penis. Shell: diameter 17-26 mm, whorls 5-6; periphery keeled, to angled, to rounded; pal- atal tooth very strong; striae either very to moderately strong, or weak and masked by sense hair-like periostracal processes. Xolotrema denotata (Férussac, 1821) (Figs. 7a-b, 33a—b; Table 2; Fig. 49) Studied material (1) IN: Jefferson County (GS-14; FMNH 214805): O live adults, 2 tissue samples— electrophoresed #1, 2. (2) KY: Fayette County (GS-112; FMNH 214806): 7 live adults, 13 tissue samples—dissected #1, 2, 6 (illustrated #6); electrophoresed #1, 2, 5, 6, 7, 8, 10, 11, 13; illustrated shell #1. Xolotrema obstricta (Say, 1821) (Figs. 7c-d, 33e-f; Table 2; Fig. 49) Studied material (1) KY: Henderson County (GS-16; FMNH 214852): 1 live adult, 1 tissue sample— electrophoresed #1; illustrated shell #1. (2) AL: Madison County (GS-20; FMNH 214853): 1 live adult, 1 tissue sample—electro- phoresed #1. (3) KY: Edmonson County (GS-125; FMNH 214854): 15 live adults, 16 tissue samples—dissected #1, 9 (illustrated #9); electrophoresed #1, 3, 4, 6, 10. Xolotrema caroliniensis (Lea, 1834) (Figs. 7e, 33c-d; Table 2; Fig. 49) Studied material (1) AL: DeKalb-Marshall Counties (GS-184; FMNH 214 ): 1 subadult, 1 tissue sample— electrophoresed #1. (2) TN: Franklin County (FMNH 171142): 5 live adults—dissected #A, В (illustrated +A); illustrated shell #B. Genus Triodopsis (Rafinesque, 1819) (Figs. 9-18, 35-45; Table 2; Fig. 49) Comparisons Penis. Triodopsis differs from the other three genera of eastern triodopsines by the abruptly larger pustules on its pilaster. Shell. Triodopsis is unique among eastern American triodopsines in having an open um- bilicus and a distinct, non-lamellar basal tooth. Key characters Penis: pilastral pustules abruptly larger than wall pustules; pilastral pustules either unfused, fused into nesting horeshoe shapes, fused into two columns of interdigitating rect- angular box shapes, fused into grossly irreg- ular elements, fused into a solid apical mass bearing three to four tiers of long and sharp spurs, or fused into irregular polygons bearing short and blunt spurs; wall-pustular columns separated and either radiating from the pore, 15-20 (or rarely 8-10) in number, and unmerging or incompletely merging basally; or completely merging ventrally to form either 10-12 obtuse V-shapes or 5-7 acute V- shapes; wall-pustular columns with pustules distinct, with pustules partially fused, or smooth with no sign of pustules; verge ab- sent; pore terminal or ventrally subterminal; penis short, to long, to extremely long and thread-like; erectile, fleshy peduncle below the pore large, small, or absent. Shell: diameter 8-27 mm, depressed-glo- bose to depressed, whorls 4 1/2-6 1/2; um- bilicus wide and open to minute and creviced; parietal tooth prominent, varying from straight to abruptly angled up to about 120 degrees, from uniformly high-standing to abruptly changing in height; basal tooth weak (rarely absent) to pronounced, varying from peg-like to tapered, from simple to buttressed to bidentate, and from marginal to deeply re- 266 EMBERTON cessed; palatal tooth pointing toward the um- bilicus, weak (rarely absent) to pronounced, varying from broad to narrow, from squared to tapered, from simple to buttressed, and from marginal to deeply recessed; striae very weak to very strong. Species group Triodopsis vulgata (Figs. 9, 10, 35, 36; Table 2; Fig. 49) Key characters Penis: pilastral pustules unfused or fused into nesting horseshoe shapes; wall pustular columns 15-20, radiating from the pore, unmerging or partially merging basally, and either with distinct pustules or nearly smooth; pore ventrally subterminal, about 1/5-way from the apex, everted penis shaped like an angled baseball bat. Shell: diameter 10-19.5 mm, depressed, whorls 4 1/2-6; aperture deeply dished; apertural periphery with a squared-off ap- pearance; parietal tooth straight, broadly wedge-like, and symmetrical or slightly an- gled and tapered toward the umbilicus; basal tooth peg-like, marginal; palatal tooth broad, squared, recessed; striae moderate to very strong. Species subgroup Triodopsis vulgata (Figs. 9a, c; 35a, b, e, f; Table 2; Fig. 49) Key characters Penis: pilastral pustules unfused; wall pustular columns never merging. Shell: parietal tooth slightly angled and tapered toward the umbilicus. Triodopsis vulgata Pilsbry, 1940 (Figs. 9a, 35a—b; Table 2; Fig. 49) Studied material (1) TN: Morgan County (GS-109; FMNH 214883): 20 live adults, 18 tissue samples— dissected #2, 3. (2) KY: Fayette County (GS-112; FMNH 214884): 7 live adults, 8 tissue samples—dissected #1 (illustrated #1); electrophoresed #1, 2, 3, 4, 6, 7, 8; illustrated shell #A. (3) KY: Harlan County (GS-119; FMNH 214885): 9 live adults, 11 tissue samples—dissected #1, 2, 3, 4; electrophoresed #2, 3, 6. Triodopsis claibornensis Lutz, 1950 (Figs. 9c, 35e-f; Table 2; Fig. 49) Studied material (1) TN: Claiborne County (GS-117; FMNH 214800): 22 live adults, 22 tissue samples— dissected #5, 18 (illustrated #18); electro- phoresed #1, 5, 16, 20; illustrated shell #A. Species subgroup Triodopsis fraudulenta (Figs. 9b, 10, 35c-d, 36; Table 2; Fig. 49) Key characters Penis: pilastral pustules fused into nesting horseshoe shapes; wall-pustular columns unmerging or partially merging basally, with distinct pustules or nearly smooth. Shell: parietal tooth straight, wedge-like, and symmetric. broadly Triodopsis fraudulenta (Pilsbry, 1894) (Figs. 10, 36; Table 2; Fig. 49) Studied material (1) WV: Greenbrier County (GS-139; FMNH 214822): ca 5 live adults, 11 tissue samples—dissected #6, 8 (illustrated #6); electrophoresed #2, 4, 5, 6, 7, 11; illustrated shell #A. Triodopsis picea Hubricht, 1958 (Flgs. 9b, 35c-d; Table 2; Fig. 49) Studied material (1) WV: Pendleton County (GS-134; FMNH 214860): 20 live adults, 20 tissue samples— dissected #4, 14 (illustrated #14); electro- phoresed #1, 5, 9, 11, 17; illustrated shell #15. Species group Triodopsis platysayoides (Figs. 12, 37; Table 2; Fig. 49) Key characters Penis: pilastral pustules fused into two col- umns of interdigitating rectangular box shapes; wall-pustular columns completely merging ventrally to form 10-12 obtuse V- shapes; pore terminal. Shell: diameter 27 mm; spire nearly flat; umbilicus very broad and open; parietal tooth short, nearly straight, high-standing, symmet- EASTERN NORTH AMERICAN TRIODOPSINAE 267 rical, and scooped internally; basal tooth very low, with broadly tapered sides; palatal tooth absent. Triodopsis platysayoides (Brooks, 1933) (Figs. 12, 37; Table 2; Fig. 49) Studied material (1) WV: Preston County (SC-273; FMNH 214861): 2 live adults, 5 tissue samples (col- lected under U.S. Dept. Interior Fish 8 Wildlife Permit 4 PRT-670226 and W. Va. Dept. Nat. Res. Scientific Collecting Permit No. 17, 1984, both to the author)—dissected #1, 2 (illustrated #1); electrophoresed #1, 3, 4, 5; illustrated shell 42. (2) WV: Preston County (Hubricht 11860): 1 live adult—examined dis- section done by Solem (1976). Species group Triodopsis burchi (Figs. 11a, 37a—b, Table 2; Fig. 49) Key characters Penis: pilastral pustules fused into grossly irregular elements irregular in size and shape; wall-pustular columns ca 15, radiating from the pore, unmerging, and semi-smooth; pore terminal. Shell: diameter 8-17 mm; spire extremely low; parietal tooth as in the platysayoides group; palatal tooth high, tiny, triangularly pointed, and marginal. Triodopsis burchi Hubricht, 1950 (Figs. 11a, 37a-b; Table 2; Fig. 49) Studied material (1) VA: Patrick County (GS-143; FMNH 214797): ca 10 live adults, 14 tissue sam- ples—dissected #3, 5, 12 (illustrated #3); electrophoresed #3, 4, 7, 9, 11, 14; illustrated shell #10. Species group Triodopsis tennesseensis (Figs. 11b-d, 37c-f; Table 2; Fig. 49) Key characters Penis: pilastral pustules fused into a solid apical mass bearing three to four tiers of long, sharp spurs; wall-pustular columns as in the burchi group, except completely smooth. Shell: diameter 9-25 mm; spire low; apertural teeth as in the burchi group; striae either very strong or very weak. Triodopsis tennesseensis (Walker 8 Pilsbry, 1902) (Figs. 11b—c, 37c-d; Table 2; Fig. 49) Studied material (1) KY: Fayette County (GS-112; FMNH 214864): 18 live adults, 18 tissue samples— dissected #13, 14, 15 (illustrated #15); electrophoresed #2, 5, 18; illustrated shell #7. (2) KY: Pulaski County (GS-124; FMNH 214865): 7 live adults, 12 tissue samples— electrophoresed #1, 6. Triodopsis complanata (Pilsbry, 1898) (Figs. 11d, 37e-f; Table 2; Fig. 49) Studied material (1) KY: Pulaski County (GS-13; FMNH 214802): О live adults, 1 tissue samples— electrophoresed #1, 2. (2) KY: Pulaski County (Hubricht 17932): ca 9 live adults (live into isopropynol)—dissected #A, В, С (illus- trated #C); illustrated shell #A. Species group Triodopsis rugosa (Figs. 186, 45c-d; Table 2; Fig. 49) Key characters Penis: pilaster ca 2/3-length and proximally tapered; pilastral pustules fused to form irreg- ular polygons each bearing 1-3 short, blunt spurs; wall-pustular columns either ca 15 or ca 9, partially fused basally, semi-smooth; pore terminal. Shell: diameter 8-11 mm; depressed; um- bilicus moderate; parietal tooth as in the vulgata group; basal and palatal teeth peg- like, strongly buttressed, slightly recessed; striae very strong, modertaely to widely spaced. Triodopsis rugosa Brooks & Macmillan, 1940 (Fig. 49) Studied material (1) WV: Logan County (SC-278; FMNH 214888): 6 live adults, 11 tissue samples— dissected #1, 2. 268 EMBERTON Triodopsis fulciden Hubricht, 1952 (Figs. 18b, 45c-d; Table 2; Fig. 49) Studied material (1) NC: Burke County (GS-35; FMNH 214823): 5 live adults, 5 tissue samples— dissected #3 (illustrated #3); electrophoresed #2, 3; illustrated shell #A. Species group Triodopsis cragini (Figs. 13, 39; Table 2; Fig. 49) Key characters Penis: penis extremely long and thread- like; pilaster as in the rugosa group; wall- pustular pilaster columns completely fused ventrally into 5-7 acute V-shapes. Shell: diameter 8.5-14.5 mm; depressed- globose; umbilicus small; parietal tooth slightly to pronouncedly scooped externally, umbilicad extension moderate to absent; basal tooth with an umbilicad extension vary- ing from weak to equal in size to the basal tooth itself, and slightly to deeply recessed; basal lip bearing a weak to strong convex ridge; palatal tooth broad, rounded, and vary- ing from moderately sized and recessed to very large and deeply recessed; striae weak to strong. Remarks The shells of the three species seem to form a continuum from least to most derived in the order cragini, vultuosa, henriettae, showing an increasing overgrowth of the apertural lip and dentition. This hypothesis is supported by the electrophoretically more primitive position of cragini in the Wagner-2 Tree (Fig. 27) and as depicted in the Consen- sus Tree (Fig. 28). Triodopsis cragini Call, 1886 (Figs. 13b, 39c-d; Table 2; Fig. 49) Studied material (1) TX: Polk County (GS-73; FMNH 214803): 20 live dults, 20 tissue samples— dissected #3, 18 (illustrated #18); electro- phoresed #1, 2, 4, 8, 10, 14; measured shell #2. (2) TX: Henderson County (GS-79; FMNH 214804): 7 live adults, 7 tissue sam- ples—electrophoresed #2, 3, 6. Triodopsis vultuosa (Gould, 1848) (Figs. 13a, 39a-b; Table 2; Fig. 49) Studied material (1) TX: Walker County (GS-71; FMNH 214887): 18 live adults, 15 tissue samples— dissected #A, B (illustrated #A); electro- phoresed #1, 9, 11; illustrated shell #7. (2) TX: Cherokee County (GS-78; FMNH uncat.): ? live adults, 11 tissue samples— electrophoresed #1, 6. (3) TX: Jefferson County (GS-208?; FMNH uncat.): ? live adults, ? tissue samples—electrophoresed #1, 2. Triodopsis henriettae (Mazyck, 1877) (Figs. 13c, 39e-f; Table 2; Fig. 49) Studied material (1) TX: Houston County (GS-76; FMNH 214824): 2 live adults, 2 tissue samples— dissected #1, 2 (illustrated #2); electro- phoresed #1, 2; illustrated shell #2. Species group Triodopsis tridentata (Figs. 14a-b, 14, 16, 17, 40, 42, 43, 44; Table 2; Fig. 49) Key characters Penis: penis length moderate; pilaster as in the rugosa and cragini groups; pore ventrally subterminal, ca 1/4-way from the apex; everted penis mace-shaped; moderate-sized peduncle beneath pore. Shell: diameter 8-15 mm; depressed-glo- bose to depressed; whorls 4 1/2-6 1/2; um- bilicus moderate to minute; parietal tooth vari- able, ranging from that of the vulgata and fraudulenta groups, to that of the burchi and tennesseensis groups, to that of the cragini group with a more pronounced umiblicad ex- tension, to a form superficially resembling that of the denotata group of Xolotrema; basal tooth marginal and variable, covering much of the range of shapes found in the vulgata, rugosa, and cragini groups, and rarely absent entirely; palatal tooth marginal and supra- peripheral, to moderately recessed and sub- peripheral, and either peg-like (and but- tressed or unbuttressed), or as in the cragini group (and buttressed or unbuttressed), or rarely absent; striae very weak to very strong. EASTERN NORTH AMERICAN TRIODOPSINAE 269 Species subgroup Triodopsis tridentata (Figs. 14a—b, 40; Table 2; Fig. 49) Key characters Shell: diameter 12-15 mm; depressed; whorls 4 1/2-5 1/2; umbilicus moderate; parietal tooth either as in the rugosa group or as inthe burchi or tennesseensis group; basal and parietal teeth as in the rugosa group, except either buttressed or unbuttressed (or rarely absent altogether), and with the palatal tooth marginal; striae very strong. Triodopsis tridentata (Say, 1816) (Figs. 14a, 40a—b; Table 2; Fig. 49) Studied material (1) TN: Blount County (GS-8; FMNH 214866): 1 live adult, 1 tissue sample— electrophoresed #1. (2) TN: Blount County (GS-9; FMNH uncat.): ? live adults, 9 tissue samples—electrophoresed #1, 2, 3, 4, 5, 6, 7, 8, 9. (3) NC: Haywood County (GS-10; FMNH 214867): ca 10 live adults, 10 tissue samples—electrophoresed #6. (4) WV: Pendleton County (GS-134; FMNH 214875): 10 live adults, 10 tissue samples—dissected #3. (5) WV: Pocahontas County (GS-135; FMNH 214876): 4 live adults, 5 tissue sam- ples—dissected #2 (illustrated #2); electro- phoresed #1 2, 3, 4; illustrated shell #4. (6) KY: Harlan County (GS-119; FMNH 214872): 1 live adult, 1 tissue sample—dissected #1. (7) KY: Edmonson County (GS-125; FMNH 214873): 2 live adults, 2 tissue samples— dissected #1, 2. (8) WV: Preston County (GS-126; FMNH 214874): 15 live adults, 15 tissue samples—dissected #15. (9) NC: Avery County (GS-153; FMNH 214878): 10 live adults, 10 tissue samples— electrophoresed #2, 5, 7. (10) OH: Athens County: Site IV-1 (FMNH 209209): 10 live adults—dissected #C, D. (11) OH: Athens County: Site IIl-3 (ЕММН 209536); 5 live adults—dissected #C. (12) Locality unknown (FMNH 171254): ? live adults—dissected #A. Triodopsis anteridon (Pilsbry, 1940) (Figs. 14b, 40c-d; Table 2; Fig. 49) Studied material (1) KY: Harlan County (GS-121; FMNH 214793): 21 live adults, 21 tissue samples— dissected #13, 14; electrophoresed #6, 7, 16. (2) WV: Boone County (GS-142; FMNH 214796): 20 live adults, 20 tissue samples— dissected #18 (illustrated #18); electro- phoresed #3, 10; illustrated shell #19. Species group Triodopsis fallax (Figs. 15, 16, 17, 42, 43, 44; Table 2; Fig. 49) Key characters Shell: diameter 8-14 mm; depressed-gl- obose; whorls 4 1/2-6 1/2; umbilicus moder- ate to minute; parietal tooth as in the cragini group, but with a more pronounced, angled umbilicad extension; apertural lip teeth as in the cragini group, but with the basal tooth marginal more strongly buttressed, the palatal tooth only slightly recessed; striae moderate to strong. Remarks The phylogeny of the fallax group is dis- cussed in Appendix D. For species diagnoses see Grimm (1975), except for palustris, for which see Hubricht (1958). Species subgroup Triodopsis fallax (Figs. 15b—, 16b, 17, 42c-f, 43c-d, 44a-d; Table 2; Fig. 49) Key characters Shell: whorls 4.5-5.0, lip edge generally sharp, apertural teeth relatively indistinct; lus- ter dull to very shiny. Triodopsis fallax (Say, 1825) (Figs. 17a, 44a—b; Fig. 49) Studied material (1) NC: Richmond County (Hubricht 10209): 6 live adults (dropped live into isopropynol)—dissected #A, В, C (illustrated #C); measured shell #A. Triodopsis messana Hubricht, 1952 (Figs. 16b, 43c-d; Table 2; Fig. 49) Studied material (1) NC: Columbus County (GS-163; FMNH 214846): са 5 live adults, 10 tissue samples— dissected #1, 5, 6 (illustrated #6); electro- phoresed #1, 7, 8, 9; illustrated shell #A. 270 EMBERTON Triodopsis palustris Hubricht, 1958 (Figs. 15b, 42c-d; Table 2; Fig. 49) Studied material (1) SC: Williamsburg County (GS-41; FMNH 214857): ca 5 live adults, 15 tissue samples—dissected #4, 5, 15 (illustrated #15); electrophoresed #5, 8, 10, 11, 15; illustrated shell #1. (2) GA: Wayne County (GS-49; FMNH 214858): ca 6 live adults, 15 tissue samples—electrophoresed #4, 9. Triodopsis obsoleta (Pilsbry, 1894) (Figs. 15c, 42e-f; Fig. 49) Studied material (1) NC: Chowan County (Hubricht 10300): 7 live adults (dropped live into isopropanol)— dissected #A, B, C (illustrated #C). Triodopsis soelneri (Henderson, 1907) (Figs. 17b, 44c-d; Fig. 49) Studied material (1) NC: New Hanover County (ANSP A2318): 3 live adults—dissected #A, B, C (illustrated #B). (2) NC: Columbus County (FMNH 159040): shells only-illustrated shell #A. Species subgroup Triodopsis alabamensis (Figs. 15a, 16a, c, 42a-b, 43a-b, e-f; Table 2; Fig. 49) Key characters Shell: whorls 4 1/2-6 1/2; lip edge swollen; apertural teeth relatively distinct; luster al- ways dull. Triodopsis alabamensis (Pilsbry, 1902) (Figs. 27a, 54a-b; Table 8; Fig. 49) Studied material (1) TN: Meigs County (GS-105; FMNH 214791): 3 live adults, 7 tissue samples— dissected #2, 4 (illustrated #4); electro- phoresed #1, 2, 3, 4, 6, 7; illustrated shell #A. Triodopsis vannostrandi (Bland, 1875) (Figs. 16c, 43e-f; Table 2; Fig. 49) Studied material (1) SC: Aiken County (GS-179; FMNH 214880): 12 live adults, 12 tissue samples— dissected #1, 8, 12; electrophoresed #1, 2, 3, 4, 5, 6, 10; illustrated shell #11. Triodopsis hopetonensis (Shuttleworth, 1852) (Figs. 15a, 42a-b; Table 2; Fig. 49) Studied material (1) NC: Catawba County (GS-33; FMNH uncat.): ? live adults, 12 tissue samples— electrophoresed #2, 4, 6, 9. (2) NC: Colum- bus County (GS-38; FMNH 214827): ca 25 live adults, 25 tissue samples—dissected #15, 25, A (illustrated #A); illustrated shell #22. (3) AL: Perry County (GS-57; FMNH 214832): ca 10 live adults, 13 tissue sam- ples—electrophoresed #8. Species group Triodopsis juxtidens (Figs. 14c-d, 18a, c, 41, 45a—b, e-f; Table 2; Fig. 49) Key characters Penis: penis length moderate; pilaster as in the rugosa, cragini, tridentata, and fallax groups; wall-pustular columns as in the cragini and tridentata groups; pore ventrally subterminal, ca 2/5-way from the apex; everted penis shaped as in the tridentata group, but with a broader apical knob; large- sized peduncle beneath pore. Shell: diameter 10-18 mm, moderately to very depressed, whorls 4 1/2-6; umbilicus moderately to very wide; aperture dished but not as deeply as in the vulgata group; parietal tooth as in the vulgata and rugosa groups and the tridentata group; palatal tooth marginal to moderately recessed, narrow to moderately broad, squared to pointed; basal tooth mar- ginal, as in the tridentata group, but rarely buttressed on the columellar side. Species subgroup Triodopsis juxtidens (Figs. 14c-d, 18a, c, 41a-d; Table 2; Fig. 49) EASTERN NORTH AMERICAN TRIODOPSINAE 2 Key characters Shell: palatal tooth rounded, umbilicus moderately depressed to very depressed. Triodopsis juxtidens (Pilsbry, 1894) (Figs. 14c, 41a-b; Table 2; Fig. 49) Studied material (1) NC: Catawba County (GS-33; FMNH 214838): са 9 live adults, 12 tissue samples— dissected #1, 2, 3; electrophoresed #2, 4, 6, 9. (2) NC: Burke County (GS-34; ЕММН 214839): 1 live adult, 2 tissue samples— dissected #4. (3) NC: Columbus County (GS-37; FMNH 214840): ca 30 live adults, 30 tissue samples—electrophoresed #4, 18. (4) WV: Pendleton County (GS-132; FMNH 214841): 10 live adults, 10 tissue samples— dissected #5, 10 (illustrated #5); illustrated shell #7. (5) WV: Pocahontas County (GS-135; FMNH 214842): 10 live adults, 11 tissue samples—dissected #5, 6; electro- phoresed #1, 2, 3, 8, 9. Triodopsis discoidea (Pilsbry, 1904) (Figs. 14d, 41c-d; Fig. 49) Studied material (1) IL: Hardin County (SC-217; FMNH 214811): 1 live adult, 8 tissue samples— dissected #5 (illustrated #5); illustrated shell HA. Species subgroup Triodopsis neglecta (Figs. 18a, c, 45a—b, e-f; Table 2; Fig. 49) Key Characters Shell: palatal tooth squared; unmbilicus moderately to very wide; depressed. Triodopsis neglecta (Pilsbry, 1899) (Figs. 18a, 45a-b; Table 2; Fig. 49) Studied material (1) MO: Barry County (GS-96; FMNH 214850): ca 7 live adults, 10 tissue samples— dissected +2, 5 (illustrated +2); electro- phoresed #1, 2, 4, 5, 8; illustrated shell 4 A. Triodopsis pendula Hubricht, 1952 (Figs. 18c, 45e-f; Table 2; Fig. 49) Studied material (1) NC: Wilkes County (GS-149; FMNH 214859): ca 5 live adults, 19 tissue samples— dissected #18 (illustrated #8); electro- phoresed #1, 4, 7, 18; measured shell #14. APPENDIX D. ON THE PHYLOGENY OF THE TRIODOPSIS FALLAX GROUP The fallax subgroup is believed to comprise 8 species (Hubricht, 1985). Hubricht (1953, 1971) discussed field evidence for hybridiza- tion or lack of it among 6 of these species. In 1975, Grimm cursorily summarized his 10 years of field and laboratory studies on hy- bridization or lack of it among 7 of these species, and proposed an evolutionary hy- pothesis based on shell lip and dentition, presence or absence of field hybridization, and current geographical distributions. In Fig. 51, Grimm’s (1975) verbal hypothesis is sum- marized in the form of a cladogram. Table 13 summarizes Grimm's hybridizational evi- dence in support of this cladogram, and adds the available genetic-distance data. One spe- cies is included (palustris) which Grimm omit- ted from his hypothesis. It is beyond the scope of this paper to evaluate Grimm’s (1975) conclusions con- cerning field hybridization. Grimm's speci- mens and notebooks are at the National Museum of Natural Sciences, Ottawa, Can- ada, and deserve morphometric study. The consistency of Grimm’s conclusions concern- ing hybridization with this cladogram is appar- ent in Table 13: of the 10 species pairs found sympatric, those which commonly hybridize in nature have an average patristic distance (number of transformations separating them) of 2.1 (n = 6), those which rarely hybridize in nature have a patristic distance of 3 (n = 1), and those which never hybridize in nature have an average patristic distance of 4.3 (n = 3). It would be tautological to consider this correlation as validating Grimm’s cladistic hy- pothesis (Fig. 51), however, because he based his hypothesis on these same hybrid- ization data. An independent test of the cladogram is afforded, however, by the electrophoretic data available for four of the species pairs (Table 13). The four Prevosti genetic dis- 272 EMBERTON North Res Pıedmont Coastal Plaın = = x т E с E Y > FT = а UN [eu oO Fe | о = E о n South Piedmont Coastal Plain alabamensis vannostrandı hopetonensis 1 2 FIG. 51. Cladogram summarizing Grimm's (1975) phylogenetic hypothesis for the Triodopsis fallax subgroup. Hypothesized character transformations: 1. Differentiation of apertural tooth prominence into less distinct (Fig. 44a)—to the left—vs. more distinct (Fig. 43a)—to the right—with the ancestral condition unknown. 2. Differentiation of the thickness of the internal edge of the apertural lip and teeth into relatively thin—to the left—vs. relatively thick—to the right—, with the ancestral condition unknown. 3. Type A, extreme reduction of the lip teeth (from Fig. 44a to Fig. 44c). 4. Reduction of overall penial sculpture (from Fig. 17a to 17b). 5. Туре В, pronounced reduction of lip teeth (from Fig. 44a to Fig. 42e). 6. Type С, moderate reduction of lip teeth (from Fig. 44a to Fig. 43c). 7. Type D, moderate reduction of lip teeth (from Fig. 43a to Fig. 42a). 8. Type E, slight reduction of lip teeth (from Fig. 43a to Fig. 43e). tances (taken from Emberton, 1986, Appen- dix B-1) do not support the cladogram, as is shown below: Cladistic distance Genetic distance(s) Z 3 TA 4 where “cladistic distance” equals the number of transformations separating the species in the cladogram (Fig. 51). These data are scant, and the differences not very great, so they are hardly conclusive. Because of the small and close genetic distances involved, this group has probably radiated so recently that electrophoresis will be of little value in deducing its phylogeny; of the currently avail- able biochemical methods, mitrochondrial DNA studies are more likely to provide an- EASTERN NORTH AMERICAN TRIODOPSINAE 273 TABLE 13. Supporting evidence for Grimm's (1975) implied cladogram of the Triodopsis fallax group. Species pair fallax & obsoleta fallax & alabamensis fallax & vannostrandi fallax & hopetonensis messana & obsoleta messana & soelneri messana & alabamensis messana & vannostrandi messana & hopetonensis hopetonensis & vannostrandi hopetonensis & obsoleta hopetonensis & soelneri palustris & messana palustris & alabamensis palustris & vannostrandi palustris & hopetonensis DPD PB BR CO CO M CO CO D — ¿Number of transformations on Grimm's cladogram (Fig. 51). >“fallax x vannostrandi x hopetonensis.” “Although a single hybrid population was found. swers because of the faster evolutionary rate of this molecule. According to Grimm's (1975) hypothesis, the fallax subgroup consists of a Piedmont stock which has successively invaded and speciated in the Coastal Plain during Plio- Pleistocene regressions, as first suggested by Hubricht (1953). The inland stock differen- tiated early between fallax in the north and alabamensis in the south, both with strongly developed apertural dentition. According to the hypothesis, fallax spun off three succes- sive Coastal Plain species—soelneri, obso- leta, and messana—each with a different type of reduced dentition, and ranging from ex- tremely reduced (Fig. 44c), to very reduced (Fig. 42e), to moderately reduced (Fig. 43c); and alabamensis spun off two successive Coastal Plain species—hopetonensis and vannostrandi—each with a different type of reduced dentition, and ranging from moder- ately reduced (Fig. 42a) to slightly reduced (Fig. 43e). Grimm suggested—for no clearly stated reason—that the longer a species of the fallax group remains on the Coastal Plain, the more reduced its apertural dentition be- comes. Purportedly, all 7 of these species Patristic distance? Genetic distance Grimm's field (Prevosti) observations — hybrids — hybrids — hybrids? — hybrids — hybrids — sympaters“ 27. (no overlap) 21 (no overlap) .35 sympaters 132 hybrids — sympaters — sympaters 21 = 22 = .30 = “33 = hybridize in the laboratory, but not all hybrid- ize when they come in contact in the field (Grimm, 1975). Based on the cladogram of Fig. 51, the fallax group into the two subgroups (called “herds” by Grimm, 1975) fallax and ala- bamensis. T. palustris, which was not analyzed by Grimm, is tentatively placed in the (northern) fallax subgroup, despite its somewhat south- ern range (Fig. 49)—and despite Hubricht's (1950-1953) placing it with alabamensis for that reason—because of its less prominent teeth and relatively thin inner lip (see trans- formations 1 and 2, Fig. 51), and because it is electrophoretically closer to messana (Pre- vosti distance .21) than to alabamensis, van- nostrandi, or hopetonensis (Prevosti dis- tances .22, .30, and .33). Following Grimm's concept of evolutionary trends, palustris's po- sition in the cladogram (Fig. 51) lies between obsoleta and messana, because the reduc- tion of its lip teeth (Fig. 42c) is intermediate between those two species (Fig. 42e and 430). Revised Ms. accepted 18 February, 1987 MALACOLOGIA, 1988, 28(1-2): 275-287 GENETIC HETEROGENEITY ON DIFFERENT GEOGRAPHIC SCALES IN NUCELLA LAMELLOSA (PROSOBRANCHIA, THAIDIDAE) W. Stewart Grant! School of Fisheries, University of Washington, Seattle, Washington 98195 USA & Fred М. Utter Northwest and Alaska Fisheries Center, NOAA, NMFS 2725 Montlake Blvd. E. Seattle, Washington 98112, USA ABSTRACT The magnitude of gene flow, and hence the potential for population differentiation, in marine mollusks is determined largely by the extent of larval and adult dispersal. In this study we measured population differentiation in an intertidal whelk, Nucella lamellosa, which has low levels of dispersal between populations because it lacks planktonic larvae and because adults show little tendency to migrate along shore. Using two polymorphic allozymes, Pep-2 and Pgm, as population markers, we found significant allele frequency differences among 12 breeding colonies sampled on a single low tide along a continuous 100 m boulder beach. These frequency differences, up to 0.12 for Pep-2 and 0.11 for Pgm, arise by random drift in aggregations maintained by homing of adults to previous breeding areas. On a scale of between 100 and 1,000 km, we found considerable differentiation among populations located along the Pacific Ocean coast and in Juan de Fuca Strait, Hood Canal, and Puget Sound. Variation among populations along some shorelines was haphazard as expected from genetic drift in small populations and limited gene flow. In other areas, allele frequencies varied clinally over distances of 300 to 600 km. These clines may have arisen by chance from genetic drift and limited gene flow, or may reflect natural selection on Pep-2 and Pgm or on closely linked loci. A gene diversity analysis indicated that 33% of the total gene diversity was due to population subdivision at various geographic scales and that 67% was contained, on average, within populations. This is the greatest amount of population subdivision yet reported for a marine gastropod and supports the postulate that gastropods with limited larval and adult dispersal should have genetically fragmented populations. Key words: Nucella, protein electrophoresis, allozyme variation, population genetics, microgeographic variation, Pacific Northwest. INTRODUCTION Marine gastropods exhibit different modes of larval development which potentially influ- ence the extent of gene flow between popu- lations. Electrophoretic studies of species having long-lived planktonic larvae with a large potential for passive dispersal by ocean currents show that gene flow can have an homogenizing effect on populations (Gooch et al., 1972; Berger, 1973) even in the face of strong regional selection (Johnson 8 Black, 1984). On the other hand, gastropods with gene flow reduced by larval brooding tend to show much greater levels of differentiation among populations (Berger, 1973; Snyder 8 Gooch, 1973; Janson 8 Ward, 1984; Janson, 1986). However, there have been few studies of gastropods with other modes of non- planktonic larval development to demonstrate that reduced gene flow produces a greater amount of genetic differentiation among pop- ulations. In this study, we measured genetic differ- entiation among populations of Nucella [Thais] lamellosa (Gmelin, 1791), an intertidal “Present address: Department of Microbiology, University of Cape Town, Rondebosch 7700, South Africa. After 31 December 1987: Department of Genetics, University of the Witwatersrand, Johannesburg 2050, South Africa. (275) 276 whelk with much-reduced dispersal during both its larval and adult stages (Spight, 1974). Gene flow is presumably limited by direct larval development in benthic egg capsules and by homing of adults to previous breeding colonies. Although there is some adult migra- tion along shore, populations as near as25 m can expand and contract independently of one another in response to food availability and reproductive success (Spight, 1974). This reduction in gene flow enhances the formation of local races of shell color, banding and shell sculpturing, which at one time formed the basis of subspecific nomenclature (Dall, 1915; Kincaid, 1957). A study of al- lozyme variation by Campbell (1978), how- ever, showed that these forms belong to a single polymorphic species. The goal of this study was to measure the amount of genetic differentiation among pop- ulations of this whelk on two different geo- graphic scales. We examined allozyme vari- ation among breeding colonies along 100 m of beach and along 1,000 km of shoreline in Juan de Fuca Strait, Hood Canal and Puget Sound, Washington. If gene flow is restricted to the extent suggested by previous studies of its life history and migratory patterns (Spight, 1974), then significant genetic heterogeneity should be apparent among populations or even among subpopulations. MATERIALS AND METHODS A total of 2286 whelks were collected from 27 intertidal locations in British Columbia, Canada, and Washington and Oregon, USA (Fig. 1), transported in damp cloth, and kept live in recirculating sea water tanks at 10°C until electrophoresis within one week. Soluble proteins were extracted from foot muscle and digestive gland by maceration ín a test tube with distilled water and by centrifugation at 1000 x g for 10 min. Horizontal starch-gel electrophoresis followed May et al. (1979) and histochemical stain protocols followed Harris & Hopkinson (1976). Gels consisted of 13% hydrolyzed potato starch (Electrostarch, Madison, WI). We initially resolved the products of 19 enzyme-coding loci in one sample using three different electrophoretic buffers. A discontin- uous system using 11$, citric acid, and lithium hydroxide (Ridgway et al., 1970; pH = 8.1) was used for aspartate aminotransferase (Aat-1, Aat-2; EC 2.6.1.1), esterase (Est-1; GRANT & UTTER + Vancouver) /slond ‘\f Se BER N Juon de Fuca . Strait 16 : : Puget Sound ~ Commencement с DOY, o . $ S © “= ay S Q S ; N Washington < N 50 km Oregon FIG. 1. Locations of samples of Nucella lamellosa in the Strait of Juan de Fuca, Hood Canal, and Puget Sound. Location numbers correspond to those in Table 1. ЕС 3.1.1.1), diaphorase (Dia; EC 1.6.2.2), lac- tate dehydrogenase (Ldh, EC 1.1.1.27), peptidase (Pep-2, Pep-3; substrate = glycyl- leucine; EC 3.4.11), phosphoglucomutase (Pgm; ЕС 2.7.5.1), sorbitol dehydrogenase (бар; ЕС 1.1.1.14), superoxide dismutase (Sod; ЕС 1.15.1.1). А continuous buffer sys- tem using tris, citric acid and N(3- aminopropyl)-morpholine (Clayton & Tretiak, 1972; pH = 6.5) was used for adenylate kinase (Ak; ЕС 2.7.4.3), isocitrate de- hydrogenase (/dh-1, Idh-2; ЕС 1.1.1.42), glycerol-3-phosphate dehydrogenase (Gpa; EC 1.1.1.8), leucine aminopeptidase (Lap-1; EC 3.4.1.1), malate dehydrogenase (Май; EC 1.1.1.37), and 6-phosphogluconate de- hydrogenase (Pgd; EC 1.1.1.44). A buffer containing tris, boric acid, and NaEDTA (Markert & Faulhaber, 1965; pH = 8.7) was used for mannosephosphate isomerase (Mpi; NUCELLA GENETIC HETEROGENEITY 277 TABLE 1. Allele frequencies, Wright's inbreeding coefficient (F) in breeding colonies of N. lamellosa at site 21 near Bellingham, Washington. Number of whelks In colony Electrophoresed Pep-2'°° |= Pgm'°° [= 83 50 0.100 0.111 0.890 —0.124 294 60 0.067 —0.066 0.950 —0.053 33 38 0.091 0.267 0.985 —0.026 275 47 0.032 —0.030 0.989 0.022 184 59 0.051 0.300 0.975 0225 124 53 0.009 —0.058 0.962 — 0.032 17 117 0.118 0.435 0.912 0.634 52 52 0.038 — 0:35 0.923 0.188 21 21 0.095 0.446 0.952 — 0.042 63 63 04127. 0.284 1.000 0.0 89 64 0.078 0.348 0.906 0.086 197 72 0.083 0.453 0.931 —0.081 Mean 119.3 49.3 0.071 0.230' 0.948 0.040 1P < 0.01 ЕС 5.3.1.8), and xanthine oxidase (Хо; ЕС 1.2.3.2). The gel banding patterns for Рер-2 and Pgm, which we scored in all of the samples, were generally well resolved in our samples. Individuals with questionable geno- types were reanalyzed. Representative gen- otypes of polymorphic loci in each sample were electrophoresed on the same gel to determine allelic identities. RESULTS Genetic variation We examined a total of 19 loci in a sample of 50 whelks and found that Pep-2 and Pgm were polymorphic. We subsequently scored these loci in all samples. There were three zones of banding for gels stained with peptidase using glycyl-leucine as a substrate. The second anodal zone, encoded by Pep-2, showed three-banded and one-banded phe- notypes reflecting heterozygotes and homo- zygotes of a dimeric enzyme. We observed a single zone of banding for Pgm, which had two-banded heterozygotes and single-band- ed homozygotes typical of a polymorphic monomer. Microgeographic variation We collected samples of whelks from 12 breeding colonies on a 100 m stretch of cobble beach near Bellingham, Washington (site 21) to measure microgeographic varia- tion. We censused all of the breeding colonies that could be found during a single nocturnal spring low tide (8 February 1977) and col- lected subsamples, or the entire aggregation if it were small (Table 1). Very few solitary whelks were observed away from the breed- ing colonies. Numbers of whelks in the colo- nies varied from 17 to 294 and averaged 119.3 (SD = 96.5). This is similar to the average of 145.8 whelks per breeding colony measured by Spight (1974) on San Juan Island. The average nearest-colony distance was 10.1 m which is also similar to the results of Spight (1974) who found intercolony dis- tances of 10 to 15 m. We had expected intermediate allozyme frequencies for Pgm and Pep-2 at this site based on our results from other localities. However, frequencies for Pgm'® in 12 sam- ples ranged from 0.890 to 1.00 with a weighted mean of 0.948, and frequencies for Pep-2'°° ranged from 0.009 to 0.127 with a mean of 0.071 (Table 1). Nonetheless, these data may provide some insight into breeding colony structure. We examined genotypic distributions in the colonies with Wright's (1943) fixation index, F;, which measures the effects of inbreeding, selection or other processes affecting geno- typic frequency. The G-test for goodness of fit (Sokal & Rohlf, 1981) was used to test for the significance of F; (i.e., departures of genotypic 278 GRANT & UTTER TABLE 2. Wright's unweighted Е statistics for N. lamellosa on two geographic scales. Fsr = gene differentiation among subpopulations relative to the total population. Fis = probability of identity of two homologous genes in an individual relative to its subpopulation. Fır = probability of identity of two genes in an individual relative to the total population. Locus Allele Ест Fis Fit Site 21 (12 colonies over 100 m) Pep-2 100 0.017 0.252 0.265 Pgm 100 0.024 0.093 0.115 Average 0.021 0.135 0.190 Pacific Northwest (30 samples over 1000 km) Pep-2 100 0.400 0.116 0.470 Pgm 100 (0117 0.056 0.218 Ауегаде 0.286 0.086 0.344 | . О «Ô co at proportions from Hardy-Weinberg expecta- 00 O cor” co" со tions in a sample). Fis is the unweighted e : ool A e oho average of F; over samples and is the corre- 0 4 és ? ue lation of two homologous genes in an individ- a } stat ual relative to the colony (Table 2). Fsy is the + correlation between randomly-chosen pairs of homologous genes in a colony relative to the total population. A significant departure from Hardy-Weinberg expectations (P < 0.01) ap- peared for the genotypes of Pep-2 pooled over colonies but not for Pgm. There were no T == —~o— —e— —+— == = [ | significant departures from Hardy-Weinberg Ч Pep-2!00 . | expectations in the colonies themselves, but 3 | $ 1} s Fis was 0.252 for Pep-2 and 0.093 for Pgm. $ re ! > The amount of differentiation among colonies , 007377 6 8 1012 4 16 18 20222426 measured by Fst was 0.017 for Pep-2 and à ГОГ + MAS" а 0.024 for Pgm. G-tests for independence < р Рот (0? (Зока! & Rohlf, 1981) of absolute allele fre- quencies among colonies were significant for both Pep-2 (Gıı = 23.1, 0.05 > P > 0.01) and Pgm (G:1 = 23.8, 0.05 > P > 0.01). Geographic Variation Relative allozyme frequencies and approx- imate 95% confidence intervals for the most common alleles of Pep-2 and Pgm are ar- ranged along the shoreline in Fig. 2. Frequen- cies of Pep-21% varied clinally from 0.96 at Newport, Oregon (site 1) on the outer coast to about 0.40 at Port Townsend (site 7) (Table 3). The direction of this cline in Hood Canal (sites 8-13), however, was reversed so that frequencies exceeded 0.90 in the southern part of the inlet. In Puget Sound (sites 14-19) the frequencies of Pep-2'°° were less than 0.22 and averaged 0.10. A cline was apparent along the San Juan Islands and the north shore of Juan de Fuca Strait; Pep-2'°° varied from 0.07 at Bellingham, Washington (site 21) f 05 ht 1 24 ое р E EE ERNEST LS] 6 8 10 12 1416 18 20222426 Location number FIG. 2. Allele frequencies of Pep-2'°° and Рдт"99 in populations of Nucella lamellosa. Vertical bars represent four binomial standard errors, [р(1-р)/2п]"?, where p is the frequency of the most common allele and n is sample size, and approxi- mate a 95% confidence interval. Broken vertical lines separate geographically isolated groups of samples. to 0.96 at Port Renfrew Harbor, Canada (site 26). Along the outer coast and Juan de Fuca Strait south, frequencies of Pgm'°° varied haphazardly from 0.48 at site 3 to 0.738 at site 5. In Hood Canal frequencies appeared to NUCELLA GENETIC HETEROGENEITY 279 TABLE 3. Locations, dates, number of whelks sampled, allelic frequencies for two polymorphic loci and fixation indices (F) for samples of Nucella lamellosa in the Pacific Northwest. Date Pep-2 Pgm Location (mo-yr) N 100 Fs 100 106 90 109 E; Oregon 1. Newport 8-75 88 0.920 0.38** 0.670 0.330 — — 0335 2. Tillamook 8-75 75 0.760 0.09 0.520 0.480 — — 0.07 Washington 3. Westport 8-75 90 0.710 0.38** 0.480 0.520 — -— 0.11 4. Mukkaw Bay 7-75 90 0.620 0.13 0.560 0.440 — — 0.10 5. Salt Creek 7-75 65 0.474 0.07 0.738 0.261 — — 0.40* 6. Middle Point 8-75 80 0.394 0.24* 0.630 0.370 — —- 0.12 7. Port Townsend 7-75 57 0.425 0.28* 0.465 0.535 — — 033: 3-78 80 0.440 —0.01 0.513 0.487 oa — 0.10 8. Hood Canal Bridge 8-75 50 0.640 0.05 0.510 0.490 — — —0.16 3-78 80 0.453 0.02 0.516 0.484 — — 0.03 9. Big Beef Bay 9-75 50 0.940 -0.06 0.680 0.320 — — —0.01 10. Dabob Вау 9-75 50 0.910 0.15 0.750 0.250 == — —0.01 11. Hoodsport 11-75 50 0.930 0.53** 0.810 0.100 0.090 — 0.02 12. Scenic State Park 8-75 50 0.960 -0.04 0.960 0.040 = —- —0.04 13. Kitsap Мет. Park 8=75 50 0.640 —0.04 0.530 0.470 = — —0.00 14. Olympia 4-75 40 a 0.00 1.000 — —- — 0.00 15. Ruston 4-75 40 — 0.00 1.000 — — -— 0.00 16. Alki 4-75 50 0.120 0.24 0.960 0.040 = — —0.04 8—78 90 0.111 0.32** 0.994 0.006 — — 0.07 17. Golden Gardens 4-75 50 0.220 0.10 0.970 0.030 — — —0.03 18. Edmonds 4-75 507 0.140’ 0:16 0.970 0.030 — — — 0103 19. Mukilteo 4-75 50 0.090 0.15 0.940 0.060 — — ON 20. Rosario Beach 7-78 80 0.112 0.25 0.875 0.119 = 0.006 -0.09 21. Bellingham 2-77 591 0.071 0.24** 0.948 0.052 — — 0.07 22. Snug Harbor 4-75 40 0.450 —0.01 0.700 0.300 -— — — 0.07. 23. Friday Harbor 4-75 40 0.275 0.25 0.725 0.275 — = 0.25 British Columbia 24. Victoria 5-75 40 0.750 -0.07 0.538 0.462 — — —0.06 25. River Jordan 5-75 40 0.763 -0.18 0.688 0.312 -- — 0.13 26. Port Renfrew Harbour 5-75 40 0.900 -0.11 0.612 0.388 — — ON 27. Port Renfrew Outer coast 5-75 40 0.650 0.34* 0.675 0.325 — — — 0.25 !Wright's (1943) inbreeding coefficient. *Significant departure from Hardy-Weinberg proportions 0.05 > P > 0.01. “P< 0.01. vary Clinally from 0.51 at the north-shore entrance (site 8) to 0.96 at site 12 on the south shore. In Puget Sound frequencies varied over a small ranged between 0.94 and 1.0 (sites 14-19). Along the north shore of Juan de Fuca Strait and the San Juan Is- lands, Pgm'® varied irregularly from 0.54 to 0.73 (sites 20-27). Samples from three sites (7, 8 and 16) were analyzed in 1975 and again in 1978 to measure the temporal stability of allele frequencies (Table 3). Allele frequencies for Pep-2 varied significantly (Р < 0.01) between years at site 8 and for Pgm at site 25 (0.05 > P > 0.01). The remaining 4 comparisons between years were not signifi- cant. F-statistics were originally developed for the analysis of subpopulations within a single population and not for the analysis of popula- tions within a species (Wright, 1943). None- theless, we applied this analysis to our allele frequency data for populations on a larger geographic scale knowing that our samples potentially included individuals from more than one subpopulation and making each population a ‘subpopulation’. Fis and Fsr are as before and Fır is the correlation between homologous genes in an individual relative to the total population that we sampled. For 280 GRANT & UTTER TABLE 4. Hierarchical analysis of gene diversity for Nucella lamellosa. Samples were subdivided into 5 groups for regional comparisons: group 1 locations 1-4; group 2 = locations 5-7; group 3 = locations 8-13; group 4 = locations 14-19; group 5 = locations 21-27. Hr = total heterozygosity. Hs = mean population heterozygosity averaged over all populations. Gcs = relative gene differentiation among colonies within populations. Суяз = relative gene differentiation between years. Gsr = relative differentiation among populations within regions. Gar = relataive differentiation among regions. Locus Hr Hs Ges Gyrs Gsr Gar Pep-2 0.472 0.256 0.001 0.002 0.190 0.265 Pgm 0.334 0.264 0.002 0.000 0.092 0.113 Average 0.042' 0.027' 0.002 0.001 0.141 0.189 ‘Average includes 17 additional monomorphic loci. Pep-2, 19 of 30 samples had heterozygote deficits, 8 of which represented significant departures from Hardy-Weinberg expecta- tions (Table 3). There was an average deficit of heterozygotes in the samples (Fis = 0.116) (Table 2). For Pgm, 14 samples had heterozygote deficits, of which 3 represented significant departures from Hardy-Weinberg expectations. Fis for this locus was 0.056. Fır was 0.470 for Pep-2, 0.218 for Pgm, and averaged 0.344. Fsr values for the 27 loca- tions were 0.400 for Pep-2 and 0.172 for Pgm and averaged 0.286. We further examined differences among samples with Nei’s (1973) gene diversity statistics using the hierarchical algorithm of Chakraborty et al. (1982) (Table 4). In this analysis, total gene diversity, Hr, (heterozygosity of pooled allele frequencies) was partitioned into its components which were due to differences (1) among breeding colonies within a location at the lowest level, (2) between years (roughly one generation), (3) among samples within regions, and (4) among regions at the highest level. For re- gional comparisons the samples were divided into 5 groups corresponding to the outer coast (sites 1-4), Juan de Fuca Strait-south (sites 5-7), Hood Canal (sites 8-13), Puget Sound (sites 14-19), and Juan de Fuca Strait-north including the San Juan Islands and Bellingham (sites 20-27). Assuming that the remaining 17 loci were monomorphic in all samples, Hr was 0.042 of which 0.02% was due to gene differences among breeding col- onies at a single location, 0.01% was due to differences between years (measured at three sites), 14.1% was due to differences between localities within regions and 18.9% was due to differences between the five re- gions. Average heterozygosity per sample varied between 0.0 and 0.053, averaged 0.027, and represented 66.7% of the total gene diversity. DISCUSSION Differentiation among breeding colonies Mature N. lamellosa typically aggregate in late winter into small groups in low-intertidal areas in which each female deposits 40 to 60 egg capsules in a common egg mass at- tached to rocks. Over a five year period, Spight (1974) measured all of the egg cap- sule masses along a 600 m rocky shore at Shady Cove on San Juan Island and found an average of 50.8 masses each year having an average area of 316.5 cm”. These egg masses were attended on average by 145.8 whelks. The populations at this site, however, were not necessarily typical of whelk popula- tions at other sites. At a more wave exposed site on San Juan Island, for instance, egg capsule masses were much larger averaging 1586 cm? in size and were produced by a correspondingly larger number of breeders. In our study, site 21 differed from Shady Cove in that it consisted of large cobbles and boulders instead of bedrock. The densities of whelks and of egg capsule masses, however, were similar to those observed at Shady Cove; we found 12 breeding colonies along a 100 m beach that were attended by an aver- age of 119.3 whelks. We therefore feel justi- fied in comparing the Shady Cove data with our own in the following analyses of mi- crogeographic variation. Our examination of allele frequencies of Pep-2 and Pgm at site 21 revealed a signifi- cant amount of allele frequency heterogeneity NUCELLA GENETIC HETEROGENEITY 281 among the 12 breeding aggregations. The relative measure of differentiation combined over loci, Fsr = 0.021, was large considering the physical proximity of the breeding groups. This estimate, however, may be inflated somewhat by sampling errors since our sam- ple sizes were not large. Nonetheless, these results confirm Spights (1974) conclusion that breeding colonies are not random aggre- gations of snails along the beach, but are structured to some degree by juvenile site fidelity and by homing of whelks to previous breeding areas. Island model of migration Using extensive tagging data and direct observation, Spight (1974) summarized the complex demography of N. lamellosa as fol- lows: about 94% of surviving hatchlings stay in their population for their first year and 71% of these remain until they reach maturity three years later, so that 67% of the surviving juveniles spawn for the first time in their own breeding group. Approximately 40% of these spawn a second time of which 71% remain in their original breeding group. Therefore, the probability of a hatchling remaining with the same spawning group is 0.59. This suggests that the migration rate among local colonies may be as high as 0.41. If we assume that breeding colonies are at equilibrium with re- spect to migration, we can compare this esti- mate of migration with that predicted by our estimate of Fs; and Wright's (1951) island model of migration. Ignoring mutation, relative differentiation among colonies for small mi- gration rates is approximately Fst — 1/(4 Мт ats ih): The model predicts 12 migrants (Nm) be- tween colonies per generation over this stretch of beach. Taking Spight's (1974) esti- mate of average colony size of 146 and our own of 119 whelks as estimates of N, the effective migration rate (m) is on the order of 0.08 and 0.10, respectively, rather than 0.41. Both genetic and empirical estimates of migration, however, are subject to several sources of error. First, our small sample sizes would tend to inflate Ест so that the real value of m may be larger than our estimate. Sec- ond, the island model of migration does not entirely reflect the real biology of Nucella in that equal exchange between all colonies is not likely. Spight (1974) has shown that mi- gration among breeding areas varies greatly by area and over time. This again would have the effect of underestimating m from the model. Third, the value of N used in the model is the effective population size which is un- doubtedly overestimated by the census num- ber of whelks in a colony. Longterm sperm storage by females, mating between only a few whelks in a colony, and the presence of immature, parasitized or senescent whelks would inflate estimates of effective population size by direct count. The large positive value of the average fixation index within colonies (Ест = 0.135) suggests that individuals in a colony originate from only a few matings. Together with the genetic data, an overesti- mate of N would produce migration rates that were too large. Fourth, the empirical estimate of emigration into other colonies is probably too large because of the mortality of tagged whelks or because of the emigration of whelks out of the study area. Genetic drift Microgeographic differentiation has also been reported for the high rocky intertidal, ovoviviparous periwinkle, Littorina saxatilis which also has limited adult dispersal (Jansen 8 Ward, 1984). An Fst = 0.095, averaged over 11 polymorphic loci, was found among 11 populations situated along a 1 km beach. Populations as close as 4 m from one another exhibited significantly different allele frequen- cies. A moderate amount of heterozygote deficit was also observed in these populations (Fis = 0.070) and was interpreted to result from partial isolation among populations. Al- though selection for shell shape between wave-exposed and sheltered sites was strong, allozyme differentiation did not appear to be related to environmental gradients. Similar small scale differentiation has been reported for other intertidal organisms with differing amounts of gene flow between areas and has been variously interpreted to reflect habitat selection by genotype (Giesel, 1970; Jansen, 1982), genotype dependent spawn- ing times and synchronous larval recruitment (Gosling and Wilkens, 1985), environmental selection among pre-recruit (Johnson 4 Black, 1984) or post-recruit larvae (Boyer, 1974; Koehn et al., 1976; Gartner-Kepkay et al., 1983; Kartavtsev 8 Zaslavskaya, 1983), and mixing of recruits from genetically differ- ent source populations (Tracey, Bellet 8 Gravem, 1975; Koehn et al., 1976; Milkman & Koehn, 1977; Lassen 8 Turano, 1978). None of these models, however, appears to 282 GRANT & UTTER explain the pattern of allozyme differentiation we observed among the breeding colonies of Nucella on a scale of 100 m. The colonies were located along a gently-sloping cobble beach with an even topography without any obvious longshore gradients in wave- exposure, food availability, or desiccation that could act as selective agents to produce the genetic differences. The formation of breed- ing colonies in Nucella is behavioral, and does not reflect habitat selection during peri- ods of non-breeding when Nucella are dis- persed along the beach (Spight, 1974). We conclude, therefore, that the allozyme heter- ogeneity is due to random genetic drift among the small breeding colonies which are par- tially isolated from one another by homing to previous breeding areas. Additional experi- ments are required to determine whether whelks converge on microhabitats that en- hance larval survival in the benthic egg cap- sules or whether whelks are attracted to one another by genotype (assortative mating). If assortative mating is important, the analyses of juveniles from a single egg capsule mass may show evidence of inbreeding. The significant degree of microgeographic variation that we found among breeding col- onies of Nucella calls attention to our sam- pling design for studying genetic variation on a larger geographic scale. Most of our sam- ples were taken during nonbreeding times of the year when whelks were dispersed from breeding aggregations. Most of these sam- ples probably included individuals from more than one colony. Thus, the larger number of heterozygote deficits in our samples may reflect the Wahlund effect in which genetically differentiated populations are included in a single sample. This method of sampling, how- ever, tends to average out microgeographic differences in allele frequency and may yield a more representative genetic profile for a region of shore than sampling individual col- onies. Allele frequencies did not vary much over time at the three sites which were resampled after three years. In the later samples smaller, younger whelks were collected to avoid sam- pling the same adult population twice. Given the degree of site fidelity in this whelk and overlapping generations, temporal changes in allele frequencies would not be expected over such a short period of time. Thus, the two significant differences between years proba- bly resulted from sampling different colonies having different allele frequencies. Changes over longer periods of time, however, might be expected to appear through genetic drift. Regional differentiation We observed a marked difference in the levels genetic diversity within and among samples from Hood Canal and those from Puget Sound. In Hood Canal, allele frequen- cies varied widely over a range of about 0.50 for both Pep-2 and Pgm, whereas in Puget Sound allele frequencies varied over a much narrower range of about 0.20. In the lower most reaches of Puget Sound to the south of Commencement Bay, populations of N. lamellosa are scarce and the total lack of heterozygosity at sites 14 and 15 is most likely due to the loss of alleles through drift in small populations. This whelk is much more abundant at other locations in Puget Sound, however, and drift in small population is an unlikely explanation for low levels of genetic diversity. Another explanation may be that Puget Sound is polluted by industrial waste to a greater extent than Hood Canal and pollut- ants may be acting as selective agents on Pep-2 and Pgm or on linked loci. For example, a smelter is located in Commencement Bay that historically released heavy metals into marine waters (Bromenshenk et al., 1985). Complexed metals such as mercurial oxides have been implicated as selective agents on allozymes in some marine organisms (Nevo et al., 1984). In addition, wood processing plants release sulfide compounds and acci- dental spills of oil and other toxic substances are not uncommon. In some areas, notably among sites 1-7 and 20-27 for Pgm, allele frequencies varied haphazardly among populations as expected if random genetic drift and restricted gene flow were the most important influences on allele frequencies. In other areas, principally along the Strait of Juan de Fuca for Pep-2 and among locations in Hood Canal for Pgm, allele frequencies varied clinally. One expla- nation is that these clines reflect contact between differentiated populations that in- vaded uninhabited shores after Pleistocene glaciers receded from the Pacific Northwest after a glacial maximum about 18,000 years ago (Esterbrook, 1969). Kincaid (1957) sug- gested that repeated post glacial invasions may explain the geographic distributions of the various shell morphs of N. lamellosa. It is, however, not possible rigorously to test these NUCELLA GENETIC HETEROGENEITY 283 historical hypotheses with the present set of genetic data. Another explanation is that the allele fre- quency clines appeared by chance (Endler, 1977). Clines resulting from drift and gene flow alone should appear independently of one another for different loci. This appears to be the case for some of the clines that we observed. A sharp cline exists for Pep-2 among outer coast and south shore Juan de Fuca Strait populations but not for Pgm over this same coastline. There are, on the other hand, parallel clines for both loci along the north shore of the Juan de Fuca Strait and into Puget Sound. These clines may simply be coincidental or may be the result of a common selective agent. Clines may also arise by adaptive differen- tiation in response to selection along smooth or abrupt environmental gradients (Endler, 1977). Such clines have been reported for a Lap locus in the intertidal mussel, Mytilus edulis (Hilbish & Koehn, 1985). It is difficult in most cases, however, to determine whether selection is acting on allelic products of a particular locus ог on those of linked loci. In addition to numerous physical, chemical and biological oceanographic gradients along the shores of Juan de Fuca Strait, Hood Canal, and Puget Sound, several sources of marine pollution exist in Puget Sound that may act as selective agents. Additional studies are re- quired, however, to assess the importance of selection on allele frequencies. Gene flow Although the lack of planktonic larvae sug- gest that gene flow is limited, other kinds of passive dispersal may still be important for gene flow in N. lamellosa. Passive transport can be achieved by the attachment of egg capsules to floating logs or algae, or capsules may be dislodged by storms and carried to other locations by currents (Kincaid, 1957). Palmer (1984b), however, discounted such mechanisms for N. emarginata, which depos- its intertidal egg capsules similar to N. lamellosa, because drifting capsules would most likely be captured by anenomes, drift into the strand line and die or settle into subtidal areas with little chance of survival. Our dem- onstration of differences among colonies at a single site and among populations separate by a few kilometers (e.g. sites 8, 9, 10 and 13), suggests that any form of passive migra- tion is not significant. Nonetheless, gene flow and migration along a shore must occur at some rate over long periods of time, else how would postglacial shores become inhabited? Even species such as the house mouse (Mus musculus) (Baker, 1981) and the land snail (Partula taeniata) (Murray 8 Clark, 1984), which consist of strongly isolated populations, have been shown, through the introduction of genetic markers, to have significant rates of gene flow over time. The existence of popu- lations of N. lamellosa that are fixed or nearly fixed for different alleles would facilitate trans- plant experiments to measure long term rates of gene flow. Such experiments would, of course depend on the successful introgres- sion of the introduced genes into a popula- tion. Geographic range and speciation Nucella lamellosa appears to have the greatest amount of genetic fragmentation among populations of any gastropods studied so far with electrophoresis (Table 5). Our results for N. lamellosa showed that 33% of the total gene diversity was due to subdivision on different geographic scales. Another intertidal gastropod, Littorina saxatilis, which is ovoviviparious giving birth to crawlaway juveniles, also shows considerable genetic differentiation among populations. An analy- sis of gene frequencies combined over three studies (Ward 8 Warwick, 1980; Janson & Ward, 1984; Janson, 1987) indicated that 11% of the total variation was due to all sources of microgeographic and geographic subdivision. Other gastropods, which have planktonic larvae and as a consequence pre- sumably greater gene flow between popula- tions, exhibit much less genetic fragmentation among populations. Studies of Nassarius obsoletus (Gooch et al., 1972), Crepidula fornicata (Hoagland, 1984) and Siphonaria jeanae (Johnson 4 Black, 1984a, b) show that populations separated by as much as 2,000 km have not diverged genetically from one another. In these species the amount of variation due to all sources of population subdivision is less than 5% and in some cases less than 1%. The salt marsh pulmon- ate, Melampus bidentatus, while having planktonic larvae, shows a large degree of population fragmentation (Schaeffer et al., 1985). In this case oceanographic barriers limit larval dispersal and gene flow. These studies substantiate the hypothesis that GRANT & UTTER 284 15 цим peinseayN; *10000'0 ueu]} SSa] Inq 0'0 Ley} 1а}еэ26 зэщел, _ рии (786) зая 3 чозицог 966 0 7000 000`0 000'0 000'0 000'0 1d LS „geueal eueuoydis (2161) ‘1e 19 49009 166 0 100'0 200'0 — — — г th snjajosgo SNIJESSEN (5861) ‘Je 19 Jayaeyos 27/`0 — ЕО ELL'O — — OL 9 snjejuapig sndwejaw (7861) pue¡beoH 176`0 Er0 0 0,00 — — 200°0 61 д eJeo1uJo, в1пр!ае19 aeAIE| DIUOPUB|d (7861) uosuer (+861) рем 3 uosuef 0861) A9IMIEM Y рем 068`0 1Z0'0 1Z0'0 es0'0 ez00 = S ov SIINEXES BUNONIT Jaded $141 899'0 681'0 РКО == 200'0 ,000'0 С Lp ESOJISUE] EJI99NN эелие| DIUO}YUL|GUON 99U919/0H aıdwes Wy 0001 Wy 001 Wy OL uy | jesodwa | 190] saıdwes salsads UIA 0} 001 0} OL 0} | O} uu | JO ‘ON JO ‘ON 159 a_i ‘зрододзеб 9 ul иоцеиел |елодша} pue эцаелбоэб jo (E261 ‘I8N) sishjeue AJISI8AIP эиэб |е1цолеле!н ‘6 97891 NUCELLA GENETIC HETEROGENEITY 285 planktonic larval dispersal—if it occurs—acts as a strong homogenizing force among pop- ulations. There is considerable interest in the rela- tionship between modes of larval develop- ment, its influence on gene flow and popula- tion structure on one hand, and geographic range, speciation and extinction on the other. Shuto (1974), Crisp (1978), Sheltema (1978), and Jablonski (1986) postulate, in part, that species with widely dispersing planktonic lar- vae resist speciation through the cohesive effect of gene flow. Such species also tend to have large geographic ranges because they can easily invade favorable habitats. On the other hand, species with reduced gene flow have much greater genetic fragmentation among populations which is thought to pro- duce a greater rate of speciation. Such pop- ulations are also thought to have shorter geographic ranges because colonization is retarded by reduced larval dispersal. Although the lack of larval and adult dis- persal in N. lamellosa produces a genetically fragmented population structure, this whelk does not appear to fulfill the predictions of the foregoing hypothesis. Contrary to prediction, N. lamellosa has one of the largest geo- graphic ranges of North Pacific Ocean gas- tropods extending over 30° of latitude from Monterey, California to the Bering Sea and along the Aleutian Archipelago. Nucella emarginata, which also has the same mode of larval development, also occupies an equally large geographic range along the Pacific Coast of North America (Palmer, 1984b). It may be that the threshold of gene flow re- quired to bring about species cohesiveness is generally much lower than is assumed in these arguments. From theory, substantial differentiation may be prevented by only a single migrant per generation (Spieth, 1974). Alternatively, N. lamellosa as it is presently defined may include more than one taxo- nomic unit. This appears to be the case for N. emarginata where Palmer (1984a) recently discovered a genetically distinct sibling spe- cies in California. Clearly, additional genetic studies of N. lamellosa are needed to test the predictions of the gene flow-speciation hy- pothesis more rigorously. ACKNOWLEDGMENTS We thank Risteen Stafford and Ken Dunton for help in the field and the laboratory, and Robert Dillon, Alan Kohn, Rich Palmer, and Nils Ryman for helpful comments on various drafts of the manuscript. This study was sup- ported in part by a grant-in-aid to WSG from Sigma XI, the Scientific Research Society and by the National Marine Fisheries Service, Seattle, WA. The paper was written while WSG was supported on a postdoctoral fellow- ship from the Council for Scientific and Indus- trial Research, Pretoria. LITERATURE CITED BAKER, A. E. M., 1981, Gene flow in house mice: introduction of a new allele into free-living popu- lations. Evolution, 243-258. BERGER, E., 1973, Gene-enzyme variation in three sympatric species of Littorina. Biological Bulletin, 145: 83-90. BOYER, J. F., 1974, Clinal and size-dependent variation at the LAP locus in Mytilus edulis. Biological Bulletin, 147: 535-549. BROMENSHENK, J. J., CARLSON, S. R., SIMPSON, J. C. & THOMAS, J. M., Pollution monitoring of Puget Sound with honey bees. Science, 227: 632-634. CAMPBELL, C. A., 1978, Genetic divergence be- tween populations of Thais lamellosa (Gmelin). In BATTAGLIA, B. & BEARDMORE, J. A., eds., Marine Organisms, Genetics, Ecology, and Evo- lution. Plenum Press, New York, p. 157-170. CHAKRABORTY, R., HAAG, M., RYMAN, N. & STAHL, G., 1982. Hierarchical gene diversity analysis and its application to brown trout popu- lation data. Hereditas, 97: 17-21. CLAYTON, J. W. & TRETIAK, D. N., 1972, Amino citrate buffer for pH control in starch gel electrophoresis. Journal of the Fisheries Re- search Board of Canada, 29: 1169-1172. CRISP, D. J., 1978, Genetic consequences of different reproductive strategies in marine inver- tebrates. in BATTAGLIA, B. & BEARDMORE, J. A., eds., Marine Organisms, Genetics, Ecology, and Evolution. Plenum Press, New York, p. 257-273. DALL, W. H., 1915, Notes on the molluscan subgenus Nucella inhabiting the northwest coast of America and adjacent regions. Proceedings of the United States National Museum, 49: 557-572. ENDLER, J. A., 1977, Geographic Variation, Speciation and Clines. Princeton University, Princeton, 246 p. ESTERBROOK, D. J., 1969, Pleistocene chronol- ogy ofthe Puget Lowland and San Juan Islands, Washington. Geological Society of America, Bul- letin, 80: 2273-2286. САВТМЕВ-КЕРКАУ, К. E., ZOUROS, E., DICKIE, L. М. & FREEMAN, К. R., 1983, Genetic differ- entiation in the face of gene flow: A study of 286 GRANT & UTTER mussel populations from a single Nova Scotian embayment. Canadian Journal of Fisheries and Aquatic Sciences, 40: 443—451. GIESEL, J. T., 1970, On the maintenance of a shell pattern and behavior polymorpism in Acmaea digitalis. Evolution, 24: 98-119. GOOCH, J. L., SMITH, В. $. & KNUPP, D., 1972, Regional survey of gene frequencies in the mud snail Nassarius obsoletus. Biological Bulletin, 142: 36-48. GOSLING, E. & WILKINS, N. P., 1985, Genetics of settling cohorts of Mytilus edulis (L); preliminary observations. Aquaculture, 44: 115-123. HARRIS, H. & D. A. HOPKINSON. 1976. Hand- book of Enzyme Electrophoresis in Human Ge- netics. American Elsevier Publishing Co., New York. HILBISH, T. J. & KOEHN, R. K., 1985, Exclusion of the role of secondary contact in an allele fre- quency cline in the mussel, Mytilus edulis. Evo- lution, 39: 432—443. HOAGLAND, K. E., 1984, Use of molecular genet- ics to distinguish species of the gastropod genus Crepidula (Prosobrancia: Calyptraeidae). Mala- cologia, 25: 607-628. JABLONSKI, D., 1986, Larval ecology and macroevolution in marine invertebrates. Bulletin of Marine Science, 39: 565-587. JANSEN, K., 1982, Genetic and environmental effects on the growth rate of Littorina saxatilis. Marine Biology, 69: 73-78. JANSON, K., 1987, Genetic drift in small and recently founded populations of the marine snail Littorina saxatilis. Heredity, 58: 31-37. JANSON, К. 8 WARD, В. 0. 1984. Micro- geographic variation in allozyme and shell char- acters in Littorina saxatilis Olivi (Prosobranchia: Littorinidae). Biological Journal of the Linnean Society, 22: 289-307. JOHNSON, М. $. & BLACK, R., 1984a, The Wahlund effect and the geographical scale of variation in the intertidal limpet Siphonaria sp. Marine Biology, 79: 295-302. JOHNSON, M. S. & BLACK, R., 1984b, Pattern beneath the chaos: the effect of recruitment on genetic patchiness in an intertidal limpet. Evolu- tion, 38: 1371-1383. KARTAVTSEV, Yu. Ph. 8 ZASLAVSKAYA, N. I. 1983, Allozyme polymorphism in a population of the common mussel Myilus edulis L. (Mytilidae) from the Sea of Japan. Marine Biology Letters, 4: 163-172. KINCAID, T., 1957, Local Races and Clines in the Marine Gastropod Thais lamellosa Gmelin: A Population Study. Calliostoma Co., Seattle, WA. 75 p., 64 pl. KOEHN, R. K., MILKMAN, R. & MITTON, J. B., 1976, Population genetics of marine pelecypods. IV. Selection, migration and genetic differentia- tion in the blue mussel, Mytilus edulis. Evolution, 30: 2-32. LASSEN, H. H. & TURANO, F. J., 1978, Clinal variation and heterozygote deficit at the Lap- locus in Mytilus edulis. Marine Biology, 49: 245-254. MARKERT, С. |. 8 FAULHABER, I., 1965, Lactate dehydrogenase isozyme patterns of fish. Journal of Experimental Zoology, 159: 319-332. MAY, B., WRIGHT, J. E. 8 STONEKING, M., 1979, Joint segregation of biochemical loci in Salmonidae: results from experiments with Salvelinus and review of the literature on other species. Journal of the Fisheries Research Board of Canada, 36: 1114-1128. MURRAY, J. & CLARK, B., 1984, Movement and gene flow in Partula taeniata. Malacologia, 25: 343-348. NEI, M., 1973, Analysis of gene diversity in subdivided populations. Proceedings of the National Academy of Sciences, U.S.A., 70: 3321-3323. NEVO, E., BEN-SHLOMO, R. & LAVIE, B., 1984. Mercury selection of allozymes in marine organ- isms: prediction and verification in nature. Pro- ceedings National of the Academy of Sciences, U.S.A., 80: 1258-1259. MILKMAN, В. D. & KOEHN, В. K., 1977. Temporal variation in the relationship between size, num- bers and allele frequency in a population of Mytilus edulis. Evolution, 31: 103-115. PALMER, A. R., 1984a, Genetic divergence and speciation in Thais (or Nucella) emarginata. An- nual Meeting, Western Society of Naturalists, 1983. Abstract. PALMER, А. R., 19846, Species cohesiveness and genetic control of shell color and form in Thais emarginata (Prosobranchia, Muricacea): prelim- inary results. Malacologia, 25: 477—491. RIDGWAY, G. J., SHERBURNE, S. W. & LEWIS, R. D., 1970, Polymorphism in the esterases of Atlantic herring. Transactions of the American Fisheries Society, 99: 147—151. SCHAEFFER, S. W., KELLER, E. C. 8 BUROKER, N. E., 1985, Population genetics of Melampus bidentatus (Gastropoda: Pulmonata): the effect of planktonic development on gene flow. Genetica, 66: 223-229. SCHELTEMA, R. S., 1978, On the relationship between dispersal of pelagic veliger larvae and the evolution of marine prosobrach gastropods. In BATTAGLIA, В. & BEARDMORE, J. A, eds., Marine Organisms, Genetics, Ecology, and Evolution. Plenum Press, New York, p. 303-322. SHUTO, T., 1974, Larval ecology of prosobranch gastropods and its bearing on biogeography and paleontology. Lethaia, 7: 239-256. SNYDER, T. P. & GOOCH, J. L., 1973, Genetic differentiation in Littorina saxatilis (Gastropoda). Marine Biology, 22: 177-182. SOKAL, В. В. & ROHLF, В. J., 1981, Biometry, Ed. 2. Freeman and Co., San Francisco, 859 p. SPIETH, P. T., 1974, Gene flow and genetic differ- entiation. Genetics, 78: 961-965. SPIGHT, T. M., 1974, Sizes of populations of a marine snail. Ecology, 55: 712-729. NUCELLA GENETIC HETEROGENEITY 287 TRACEY, M. L., BELLET, N. F. & GRAVEM, C. D., 1975, Excess allozyme homozygosity and breeding population structure in the mussel Mytilus californianus. Marine Biology, 32: 303-311. WARD, R. D. & WARWICK, T., 1980, Genetic differentiation in the molluscan species Littorina rudis and Littorina arcana (Prosobranchia: Littorinidae). Biological Journal of the Linnean Society, 14: 417-428. WRIGHT, S., 1943, Isolation by distance. Genetics, 28: 114-138. WRIGHT, S., 1951, The genetical structure of pop- ulations. Ann. Eugenics, 15: 323-354. Revised Ms. accepted 9 June 1987 MALACOLOGIA, 1988, 28(1-2): 289-318 THE COLOURS OF MARINE BIVALVE SHELLS WITH SPECIAL REFERENCE TO MACOMA BALTHICA A. J. Cain Department of Zoology, University of Liverpool, Brownlow Street, Р. О. Box 147, Liverpool, L69 3BX, England ABSTRACT Colours of the shells of marine bivalves have been dismissed as mere excretory products. Some internal colours may be present for reasons other than their visible properties. The British species, when scored for the brilliance or weakness of their external colours and for their degree of probable exposure to visual predators, suggest strongly that there is a connection between these variables, probably because of the action of visual predators. Very small shells show no patterning, and are left out of this comparison. While many species may be cryptically colored, others clearly are not. The shell colour variation in Macoma balthica is shown to be a definite polymorphism, not continuous variation, and descriptions of the morphs are given. As all are conspicuous against the normal background of mud or muddy sand, it is probable that the polymorphism is maintained by apostatic selection, except that a sample from the Baltic Sea may be cryptic. Examination of the nature of the variation in this species against predictions from a colour atlas of what would be expected if only apostatic selection were acting, suggests that some other factors must be acting as well as apostatic selection. Key words: coloration; marine bivalves; Macoma balthica; shell; polymorphism. I. INTRODUCTION In the general account of bivalves in the Treatise on Invertebrate Paleontology, Cox (1969, but probably written long before) re- marks “The pigments are thought to be waste products of metabolism, derived from the diet or other sources, and secreted in the shell as a means of disposal. As in the gastropods, it is improbable that the colour ornament can have any protective function in the great majority of bivalves which lie buried in sedi- ment. Some bottom-living forms, notably the pectinids, appear remarkably well camou- flaged.” Sufficient work has been done on terrestrial gastropods (Cain, 1983; Heller, 1975) and some on marine ones (Reimchen, 1979) to suggest that this viewpoint is untenable for them, and an excellent paper on the bivalve Donax faba (Gmelin) by D. A. S. Smith (1975) hardly encourages its application to bivalves. Moreover, a disposition to regard bivalve shell colours as mere waste products would no more conduce to their proper investigation than does the insistence by Gould & Lewontin (1979) that discordant patterns in bivalves are in origin non-adaptive, and merely the result of constraints. A survey of the colours (and pat- terns) of bivalve shells in relation to what is known of the ecology and habits of the species (this paper) suggests strongly that when cer- tain irrelevant colourings are put aside, colora- tion does show regularities like those found, for other animal groups, in Cott’s classic work (1940). While much coloration in bivalves may well be cryptic, there are various species, both richly coloured and variable, that seem to be conspicuous. When the intellectual stumbling- block of a priori non-adaptation is removed, it becomes possible to consider whether such species do show adaptation of any sort. In most bivalves that vary in colour, the nature of the variation has not been described accu- rately. The purpose of the present paper, after surveying the colours and patterns of British bivalves in general, is to examine the variation in the very common species Macoma balthica. The variation is found to be a true polymorph- ism with some peculiar features. Its signifi- cance in the life of the animal is discussed. |. COLOURS OF MARINE BIVALVES (1) Classification of bivalves For the purposes of this paper, the classi- fication given in the Treatise on Invertebrate (289) 290 CAIN Paleontology is used, except that the superfamily Mesodesmatacea is added (Yonge & Allen, 1985). The difficulties in determining the proper classification of early Palaeozoic bivalves (and indeed whether some of them are bivalves or Crustacea) can be appreciated from the papers of Newell & Boyd, Pojeta, and Scarlato and Starobogatov in the symposium edited by Yonge & Thompson (1978). While the classification of the Treatise may not be final, it is most likely to be altered either by reallocation of some of the earliest groups (which do not concern the present paper) or by recognition of conver- gence, which will merely strengthen the points made in the present discussion. (ii) Types of colours From the point of view taken in the present paper, the colours of bivalve shells can be divided into three types. (a) Iridescences The famous mother-of-pearl, and real pearls, which together with Tyrian purple caused Pliny to regard molluscs as the root of all evii—see Rackham (1940)—show irides- cence. This is the result of an extremely regular submicroscopic packing of the struc- tural elements of the shell, producing (in effect) diffraction gratings. Such iridescence is normally internal, on the insides of the valves, or (in gastropods) hidden under a thick periostracum. There is no need, there- fore, to consider them as subject to visual selection by predators (although this possibil- ity cannot be excluded for some trochid gas- tropods which normally wear to show some iridescence externally). (b) Structural melanins К was pointed out to me many years ago by Professor N. Tinbergen that the flight-feathers of gulls show less wear in their black than in their white areas, which is easily verified on moulted remiges picked up on the beach. Dr. Carol Jones tells me (personal communica- tion) that horses’ hooves, when of a light golden colour, split and wear much more than the normally-coloured ones. It seems, there- fore, either that melanins, and perhaps other dark pigments, confer additional hardness where they are secreted into skeletal struc- tures, or those structures are hardened when specially modified to receive such pigments. Many bivalve shells have dark blotches or flushes of pigment internally, sometimes in the muscle scars, often on the upper edges of the valves on either side of the umbones. Even in shells which are normally plain white inside, such marks may occur as individual variants, aS can be seen in the common cockle, Cerastoderma edule (L.). There ap- pears to be a range of intensities in melanins varying from dull yellow to brownish orange, brown, black-brown, and black. Some pur- ples, blues, violets and blue-blacks are found in the same positions, and, whether melanins or other genera of pigments, probably have the same effect. Such colours as these in the insides of shells may possibly be related to a strengthening of the shell and, like irides- cences, need not be considered either as subject to selection by visual predators or in any sense a disproof that externally visible colours are in any way adaptive. Other inter- nal colours may have a different function. (c) Externally visible colours These are the colours that are investigated in this paper. They range from brilliant reds, oranges, yellows, browns and blacks to occa- sional greens, blues and violets. In many bivalves they are more apparent in young shells than in old ones, young shells often being more translucent, so that pigment dis- tributed in the thickness of the shell contrib- utes more to the external appearance. Nor- mally or occasionally exposed parts of the animal itself often show cryptic, disruptive, or flash colour, for example the ends of the siphons of Hiatella arctica, the mantle tenta- cles and perhaps the whole body of Lima when forced to escape by rapid swimming, the brilliant red foot of species of cockle— again when jumping in an escape reaction. | have not been able to collect sufficient data to illustrate this point in detail. (iii) Types of variation To my knowledge, no exact description of the variation in colour and pattern of any British bivalve, has yet been published (and very few for any foreign one). Even in the standard monographs, such as Forbes & Hanley's (1853) or Jeffreys’s (1863-69), often no more than an indication of the range of MARINE BIVALVE SHELL COLOURS 291 variability is given. In particular, no careful separation of discontinuous from continuous variation based on random samples has been made. Museum samples cannot be taken to be random, and the bestowing of a varietal name is no indication whatever of whether the form concerned segregates or is part of a continuum, as has been shown extensively in the gastropod Cepaea (compare Cain, Shep- pard & King (1968) and references therein with Taylor (1914)). | have been forced, therefore, to use a very rough classification of bivalve variation from the data in the standard monographs and Tebble (1966), and from my own experience, which is not extensive. If a species is stated to show several markedly different shell colours, and/or different patterns, or the frequent pres- ence and absence of a pattern, it is classed as |, highly coloured and variable. This class certainly contains some true polymorphisms (Macoma balthica is an example) but other species in it may show continuous although extensive variation. If a species is described as very far from white, although with no great range of colora- tion, e.g. with a black or black-brown periostracum, or a bright brown shell, or, as in Glycimeris, with a constant mottling all over of yellow-brown, it is classed as Il, well coloured. № a species is given as pale- coloured, or occasionally tinged with colour, or with scattered or inconspicuous markings, it is classed as Ш, poorly coloured. If it is described as ofi-white, dirty white or white, it is classed as IV, white. Probably some rather translucent shells included here ought to be transferred elsewhere if well-marked colours ofthe animal show through; the genus Lima is an example. On this grouping, the percent- ages of British species are | 12.80, II 20.12, Il 27.44, and IV 39.63. The above classification is very rough, and exactly where some species should be is a matter of opinion. To make the allocations clear, appendix A gives a complete list of the British species, following Tebble (1966), with their classification in this paper. ı shall be glad to receive corrections accompanied by better data, and if possible random samples. The British archipelago is fortunately situ- ated for the present purpose, since it is the meeting-place of three faunas, Lusita- nian-Mediterranean, north-west European (sometimes called Celtic) and sub-boreal. For its area, it has a large number of species of marine bivalves. (iv) Sources of data for probable exposure to predators Tebble (1966) seems concerned to give the full range of occurrence in respect to both depth and types of substrate for each spe- cies, a treatment which tends to obscure their stations of greatest frequency and abun- dance. Much on this subject has been gleaned from Yonge & Thompson (1976) and Yonge's New Naturalist volume (1949). In addition, Barrett & Yonge's guide to the sea shore (1958) and the illustrated account of marine molluscs of the English Channel and French Atlantic coasts by Bouchet et al. (1979) have been used, as well as a variety of papers that are noted under each species in Appendix A. (v) Variation in relation to size If, as seems probable (Cott, 1940), pattern- ing of the shell serves to break it up visually, or to disguise it as part of a patterned back- ground, it might be expected that very small shells, being at or below the commonest blotch size in the environment, would not show patterns. A pattern means here any marked variation in colour over the outside of the shell. The 20 British marine bivalves given by Tebble as 0.5 cm or less in their greatest measurement when full-grown are given in Table 1, with relevant data. This size-limit was chosen somewhat arbitrarily but bearing in mind the breadths of the finer rays on mus- sels or Mactra corallina, and of the mottlings on Glycymeris glycymeris and various scal- lops. None of these small species have any patterning at all reminiscent of the forms just mentioned; indeed, the only trace of variation in colour over the shell is in Turtonia minuta, to which Lasaea rubra might be added, and the only marked variation between individuals is in Astarte triangularis which may have a colour polymorphism. Of these three species, the first two are common intertidally in rock crevices, empty barnacle shells and the like, and are probably often exposed to visual predators. The Astarte lives off-shore in sandy mud, sandy gravel and shell gravel (Tebble, 1966) and its exposure to visual predation needs investigation. Of the remain- ing species, of 0.51 cm and more, 23 have a definite pattern, 37 are recorded as with or without a pattern (perhaps polymorphic?) and 104 as without a pattern (Table 2). When those with a pattern are combined with those 292 CAIN TABLE 1. British marine bivalves less than 0.5 cm maximum dimension; external colours of shell. Nomenclature and arrangement as in Tebble, 1966. Species External colour Size (cm) Yoldiella lucida greyish green 0.32 Yoldiella tomlini greenish or brownish yellow 0.48 Phaseolus pusillus whitish 0.16 Arca pectunculoides straw-coloured 0.48 Crenella decussata yellow-brown 0.32 Crenella prideauxi pale yellow 0.32 Lima sarsi cream, translucent 0.32 Astarte triangularis light yellow, orange or dark brown 0.32 Thyasira croulinensis white 0.32 Thyasira ferruginea rusty encrustation 0.32 Thyasira subtrigona translucent 0.16 Lasaea rubra light yellow, tinted reddish 0.32 Lepton nitidum light yellow 0.32 Neolepton sulcatulum white 0.16 Neolepton sykesi translucent less than 0.16 Epilepton clarkiae pale yellow or white 0.16 Devonia perrieri white, occasionally tinted brown 0.48 Montacuta substriata whitish or translucent 0.32 Mysella bidentata light brown 0.32 Turtonia minuta brownish with purplish or rose at umbones 0.32 with one sometimes (to obviate low expec- tancies), (Table 3) a yx? with 5 degrees of freedom of 21.39 (P just less than 0.001) results, which is highly significant. A more extended analysis of pattern occurrence against size is given in Table 2, which sug- gests that 1.0 cm could be taken as the upper limit of small size in future analyses. It also suggests that the proportion of patterned forms is markedly higher for shells above 4 cm than for those from 1 to 4, and this is confirmed on a separate x? for the two classes of size 1.01-4 and 4.01-32; (x?, = 7.65, Р < 0.01). In the rest of this paper, therefore, shells below 0.5 cm are omitted. (vi) Variation in relation to exposure By means of the information gleaned from the sources mentioned above, plus some experience of my own of intertidal and other bivalves, a rough classification has been made into (i) those wholly or frequently ex- posed to view, (ii) those only partially exposed to view, and (iii) those probably entirely con- cealed. Under (i) are all epifaunal bivalves, such as edible mussels and oysters, and such forms as many scallops which, although they may nestle into sand, remain very superficial, and may be often fully visible, e.g. when swimming. This category merges into the next, and again, the allocation of particular species is a matter of opinion. Under (ii) are forms which are shallow burrowers, and those that lurk in shallow crevices such as occur in the holdfasts of large seaweeds. Some, such as Nucula, are so slightly buried that they could be very easily exposed by predators. Lima hians, which is a nest-builder, has been placed in group (i) because its well-known escape reaction by rapid swimming, the bril- liant colour of its tentacles, and its apparently distastefulness (Gilmour, 1963, 1967) sug- gest that it is not infrequently disturbed by predators; the other limids have been put in group (ii). In group (iii), again not separated by any clear division, are nestlers in deep crevices, burrowers in thick black muds such as Thyasira, permanent deep burrowers un- able to burrow again if exposed, such as Mya, actual borers, and deep-sea forms. The per- centages of British species in these catego- ries are (i) 20.12, (ii) 45.12, and (iii) 34.76. Stanley (1970) rightly refers to Donax as burrowing rapidly and deeply. However, as Ansell (1968, 1983) shows, several species of it feed on the surface of the sand, migrating upwards with the waves breaking on the beach. The rapid burrowing is an escape 293 MARINE BIVALVE SHELL COLOURS ee vel 1 | с ри 9 9, 119 02 6 Bra ci 0c $12101 ma A | a сор Pl № Ol 0 ШЕЯ ou UJIM Ze | с ew SG р пе еб ac ulsped INOYYM 10 UM ez | L рее EB лс uJayed WIM Ol Le 08 6c de La 92 Gc he Ge ee le Oc 61 el 2.9. Si rls) cla 1 О 68 OS O ROSS 10 no К _— ‘WO ul ZI ‘uoneueA jo adA} pue azis Aq зэмела эциеш ysınug JO UONNQUISIO ‘е 37891 294 CAIN TABLE 3. Distribution of British marine bivalves by size and type of variation. Lumped data. Size class 0-1.0 cm 1.01-2 2.01—4 4.01--6 6.01-8 8.01-32 Totals With any pattern 2 14 11 9 9 15 60 With по pattern 30 34 28 8 12 12 124 Totals 32 48 39 7 21 27, 184 % with pattern within size class 6.25 29.17 28.21 52.94 42.86 55.55 x?, 5 degrees of freedom = 21.39, P < 0.001. TABLE 4. Distribution of species of British marine bivalves according to probable degree of visibility to predators (categories (i)-(ili)) and degree of coloration and patterning (types I-IV). Shells with usual adult maximum measurement less than 0.5 cm excluded. Category (i) well exposed to predators Type | 10 highly coloured (7.89)e and variable Il 12 well coloured (4.31)e Ш 7 poorly coloured (0.47)d IV 4 white (6.30)d Totals 33 Category (ii) partially Category (iii) hidden hidden Totals 11 0 21 (0.25)e (7.30)d 20 1 33 (1.75)e (9.56)d 26 12 45 (1.60)e (0.85)d 17 44 65 (5.18)d (20.29)e 74 57 164 In each cell the upper number is the number of species; the number in parentheses is the contribution to the x?. e = excess, d = deficit of observed against expected numbers. x?, 6 degrees of freedom = 65.75, P << 0.001. reaction, but the important point for the present paper is the animal's exposure to predators, both birds and crabs, on the shore. Probably even the British ones, which are less mobile, should be in category (i) but | have preferred to leave the genus in (ii), erring on the cautious side. Cockles are well known to have a jumping reaction, using the muscular foot. Ropes & Merrill (1973) note leaping and gliding in Spisula solidissima, (Dillwyn) especially young ones. Ansell (1967) describes leaping in Gari tellinella and G. fervensis, and provides a review of the whole subject (1969). Many shallow burrow- ers (and even some razor shells, in spite of their ability to disappear down their burrows at amazing speed; see McMahon & McMahon, 1983) show some leaping movement. Even when it is an escape mechanism from star- fish, it may still expose them to other preda- tors. (vii) Tentative conclusions The distribution of species by their degree of coloration and probable degree of visibility to predators is given in Table 4. The resulting x° with 6 degrees of freedom is highly signif- icant, and inspection of the contributions from MARINE BIVALVE SHELL COLOURS 295 each cell shows that there are marked ex- cesses over expectation of white shells in category (ili) (Concealed), and of highly coloured and variable shells (type |) in cate- gory (i). Correspondingly, whites are deficient in (i), and type | and type Il shells in (iii). A closer consideration of the species in each cell suggests that a detailed examination of apparent exceptions would be well worth while. Thus in the shallow burrowers, poorly coloured or white shells, e.g. in the cockles and venerids, tend to have strong sculpture, or to resist breakage by their thickness. Montacuta ferruginosa (Il, iii) is not normally exposed and may owe its rusty coloration to staining in the rusty-coloured anal track of its host, an irregular urchin (Marshall, 1891; Morton J. E., 1962). Galeomma turtoni (IV, (1)?) is recorded as crawling about like a gas- tropod, and one species of the genus gives what appears to be a dymantic (frightening) colour display (Morton, B., 1975). Apparent exceptions, therefore, may be no exceptions in reality, when their habits are fully known. It appears, then, that overall, there is rea- son to believe that the superficial colours and patterns of British marine bivalves are highly influenced by selective agents related to de- gree of exposure; if these include visual pred- ators, type ll forms may be expected to be cryptic (e.g. some Nucula spp., Arctica islandica, Glossus humanus), being seen by predators on dark muds or muddy sands. Type | forms may be cryptic on diversified backgrounds, as Cox (1969) allowed for scal- lops, or apostatically coloured as Smith (1975) suggested for Donax. Type Ill forms may perhaps be cryptic, especially when young, on paler backgrounds, or more often be normally invisible, and so need no colora- tion. It seems unlikely that direct selective action by the physical factors of the environ- ment (e.g. insolation, winter temperature etc.) should produce the diversity of colours seen in Type | forms. A first hypothesis, therefore, is that colour variation is principally selected for by visual predators. It will be seen below for Macoma balthica that the situation cannot be as simple as this. Stanley (1970) has paid careful attention to the modes of life of a large number of marine bivalves in New World waters in relation to the shape and other characteristics of the shell. He did not consider colour, since he was primarily concerned with the interpretation of fossil bivalves, in which it is very seldom preserved, and indeed whitened artificially his modern shells before photographing them (in order to bring out the surface sculpture). A comparison of his data with the colours and patterns noted briefly by Abbott (1974) and Warmke & Abbott (1962) for each of his species, suggests the same conclusions as those reached for British bivalves. A fuller treatment is in preparation. It will be seen from the classification given in Appendix A that type | shells are found in the subclasses Pteriomorpha (Pectinidae) and Heterodonta (order Veneroida, super- family Tellinacea, families Tellinidae, Dona- cidae, Psammobiidae; superfamily Venera- cea, family Veneridae). A glance at Abbott (1974) confirms these families and possibly adds to the Pteriomorpha the superfamily Limopsacea (family Limopsidae or Philo- bryidae, Limopsis antillensis Dall, with pink, orange or yellow shells). Certainly the family Spondylidae is to be added to the Pectinacea. In the Heterodonta (order Veneroidea) the superfamily Chamacea (e.g. Chama ma- cerophylla (Gmelin), “lemon-yellow, reddish brown, deep- to dull-purple, orange, white or a combination of these colours” Abbott (1974) achieves type | status. In the same order but in the superfamily Carditacea, Pleuromeris tridentata (Say) appears to be of type | (“дгау- ish brown to bright-rose, sometimes with red- brown mottlings”). The subclasses Palaeo- taxodonta, Cryptodonta and Anomalodes- mata seem to achieve at best only type III or IV coloration, although a few palaeotaxodonts in the British fauna (Nucula spp.) can be ranked as type Il. It is obvious even from this brief survey that type | coloration shows much convergent evolution. It seems most probable that its distribution is determined by mode of life, not by taxonomic affinity. Ш. COLOUR VARIATION IN MACOMA BALTHICA (i) Materials and methods (a) Collections Samples of freshly dead shells were col- lected on the strandline, and nearby, at Red Rocks and Hoylake (Wirral Peninsula) after a considerable shell-wreck by A. J. C. and J. T. Cain in late December 1983 and early Janu- ary 1984. A first sampling was by picking up all shells seen within small areas, to avoid 296 CAIN visual bias. Later samples were scooped up with a sieve from the vast numbers of shells huddled along the strandline and partly buried in the churned sand. Much of the sand was washed away from the sieve in tide-pools, and the rest removed by more leisurely wash- ing at home. No visual bias could have been exerted in these samples; since they agreed well in proportion with earlier samples, these also can be accepted as without collectors bias. Only complete shells, with the valves attached to each other by the ligament were used in our samples, except that a few with one valve partly damaged were not rejected. A further collection of shells made at Red Rocks, Wirral, between the areas collected by us was kindly given us by Dr. lan Wallace (Liverpool City Museums) in 1985. This also was scooped up from a strandline, and is free from visual bias. A collection, almost entirely of separate valves, from near Camber Sands (south coast of England) by Dr. J. Mallet was picked up by eye, and Dr. Mallet queries (in litt.) its randomness on the grounds that yel- low shells are very noticeable. Interestingly, this sample does show an enhanced propor- tion of yellows as compared with the rest of those from the Wirral Peninsula but otherwise has much the same colour range as our own collections. Through the kindness of Dr. G. Russell a random collection of specimens (preserved) has been received from the Baltic Sea (Finnish coast; Tvarminne Zoological Station). This differs entirely from the others, as can be shown by a rough score. It was found impractical to remove the animals with- out considerable risk of breaking the shell. Linnaeus (1758) rightly characterised the Bal- tic population as fragilissima. Voucher speci- mens are being deposited at the Academy of Natural Sciences of Philadelphia. (b) The Villalobos atlas and colour scores To give repeatable colour scores, and to explore the actual in relation to the possible colour variation, the colour atlas by Villalobos and Villalobos (1947) was used. The atlas is an analytical one, with each of the 38 pages generated from a different hue, from red through the spectrum to blue and purple, the series being a continuum. On each page (Fig. 1), the top line of colour squares is the series of neutral tints from darkest (1) to palest (20) and successive lines down to the bottom of the page increase in the intensity of hue and decrease in the content of neutral tint. The first row (0°), the same on each page, is totally unsaturated, showing only neutral colour; the last (12°) is completely saturated with the hue. From left to right, the columns decrease in the intensity of pigment (both of hue and of neu- tral tint) from 1 to 20, several of the very palest in columns 18 and 19 not being printed, the very pale shades being extremely difficult to print. A few that were printed have been cancelled subsequently. Column 20 com- prises only the white square in row 0°. On each page, therefore, the most intense colour is seen in row 12°, columns 12-16, the numbers lower than these being more in- tensely pigmented and therefore more som- bre, the higher ones being paler. The general muddiness of colour increases up the page and to the left, until in row 0° column 1 we have virtual black, the intensest neutral tint. The pages can therefore be thought of as half sections of a cylinder, radiating from a central axis, the neutral-tinted row 0° which is the same on every page, with the brightest spectral colours as a band around the periph- ery of the cylinder and well above its middle. Pure white shells are scored as row 20. The principal difficulties in the use of the Villalobos atlas for scoring variation in Macoma balthica are twofold. Some hues fall between two pages, and while they can be scored to a good approximation, an atlas with twice the number of pages would be desir- able. In the following scores, (see Table 5) 2/3 (for example) indicates a hue falling between pages 2 and 3. If it is clearly closer to 3 than 2 it is given as —3, if closer to 2 than 3 it is 2 +. Further, and much more important, there are a number of very pale colours involved, and these are off the page. Where interpola- tion seems possible, the resulting score is placed in square brackets, e.g. 3 8° [19] means that the page and row are determined, but there is no actual colour square printed in column 19 for row 8°, and the value is arrived at by interpolation between the printed squares. Since in some of the shells there is a change of colour, only shells of more than 8 mm maximum diameter were used for scor- ing, inspection showing that any change had occurred by this size. Only intact bivalve shells were used from the samples collected by A. J. C. and J. T. C. since they were abundant (many in fact still with flesh en- closed) although it would be possible to use only left-hand or only right-hand valves in areas where the species is uncommon. Hav- MARINE BIVALVE SHELL COLOURS 297 2° 3° 4° 5° 6° 7° ge 9° 10° 115 122 <— Increasing saturation Decreasing intensity ——> FIG. 1. Colour atlas page summarizing intensity and saturation scores for Macoma balthica shells (black circles) and for muday sand and mud (white circles), irrespective of hue. Colour squares actually printed on every page are outined in black. The top row is of pure neutral colours, repeated on every page. Scores in unoutlined squares are estimated for saturation or intensity or both. Macoma balthica colours, irrespective of hue, cluster well away from their sandy mud background colours. ing both valves means that one can easily recognise small post-mortem stainings which are almost invariably asymmetrical. (c) Sorting for morphs Sorting must be done with the shells under water, in a good light. Both the outside and the inside of the valves can become very chalky and appear whitish when dry; wetting ensures that the full colour is seen, and of course gives an appearance corresponding to what is normal in the wild. A few shells were discarded since they were so badly stained black or rust-coloured by burial in the sand, or greened by algae. The true colour can be seen, when there is occasional doubt, by inspecting the muscle attachment scars on the inside of the valve. The shell here is densely glassy, not eroded into chalky pat- terns by remobilization of its surface calcium during prolonged closures, and gives a useful view into the interior of the shell. The paired valves were therefore laid out like butterflies in large flat-bottomed trays on a background of white absorbent paper and in about 3cm of clean water. They were grouped in columns according to hue and intensity so that each column (or series of columns) agreed in hue and progressed from the most to the least intense. Most of the resulting arrangements were checked by J. T. C. Only where there appeared to be a definite discontinuity between adjacent columns was a morph boundary recognised; and all the shells were classified by morphs and intensi- ties before any were colour-scored. This avoided any temptation to break up a contin- uum of colour variation by assigning it to successive pages of the atlas and produce apparent morphs corresponding to succes- sive pages. (d) Live material Some difficulty was found in getting live samples to compare with the empty shells, but one was obtained from the Dee estuary a mile or so up from the Red Rocks collecting points, and some small ones from the 298 CAIN Lancashire coast at Blundellsands. As the shell is somewhat translucent, these were necessary to check on the appearance of the live animal. In fact, the body is mostly trans- lucent white, there being only a slight blackish infusion near the hinge line, presumably cor- responding to food or pigment in the digestive gland and perhaps the kidneys. While this dulls the colour of the shell slightly near the umbones it made no difficulty in the scoring, and the score of the empty shell can be accepted as virtually identical with that of the living animal. (e) Rescoring Examination of these rather large samples shows at once that there is a true polymor- phism, with some remarkable features. Sev- eral major colour classes can be readily sorted, as can some intensities of colour within them, and shells with a clear change of colour occurring at about 5 тт maximum diameter. The exact scoring of some of the minor differences in shade of colour is not so easy. Consequently, the major samples were scored on 14 Jan. '84, re-sorted and scored on 15 Feb. '84, and finally scored in Nov. '85, having been put away and not looked at т the intervals. Anyone beginning an investigation of this species should score at first as large a sample as possible, certainly over 300 shells, since some of the colours and colour combi- nations seem to be rare. (ii) Notation for morphs and variants Since M. balthica has not been bred, all symbols must be in roman, not italics (Cain, 1987). The most expressive English words for the three shades of red (sensu lato) appear, unfortunately, to be purple, pink and peach; they are therefore designated by P;, Po, and Pz. Orange O, orange-yellow OY, yellow Y and ivory | are easy to symbolize; the initials of the qualifying words are added for warm yellow WY, chrome-yellow CY, lemon-yellow LY, and warm ivory WI. As W is used for the qualifier ‘warm’, white is symbolized by Wh. Nearly all the hues (except white, which is not susceptible of such qualification) vary considerably in intensity. The prefixes VD (very deep), D (deep), M (medium), P (pale) and F (faint) are therefore used as appropri- ate. A single class, pinkish white (discussed below) is so pale that it is symbolized sepa- rately as WhP;. For shells that change hue, the same sys- tem is adopted, with the initial colour, as seen at the umbones, placed first and separated by a hyphen from the second colour, e.g. PP;- FO is a pale purple changing to faint orange, MO-DCY a medium orange changing to deep chrome yellow. (iii) Morphs and variants The brief descriptions that follow are in- tended only to indicate the diversity of colour and to assist anyone who is scoring a sample. It would be impossible to print a colour plate with sufficient accuracy to show the exact shades of yellow and ivory, and words plus colour scores can be used to indicate the gaudier forms. The following list gives the hues in the order of the colour atlas. Each hue is a morph with respect to all the others. Different intensities are probably not morphs but parts of a con- tinuum; possible exceptions are noted. In most shells there is a reduction, with growth, of the intensity of the hue, which is greatest at the umbones (i.e. in the young shells) and decreases somewhat towards the adult shell margin. As the adductor muscle scars are composed of a glassier material than the rest of the shell, they show an intenser hue than the more opaque areas, and indeed in some fresh juveniles (with translucent or nearly transparent shells) produce the effect of two spots of intenser colour even on the outside of each valve. In late juveniles to adults, the valves are nearly opaque, and colour on the outside is confined to the umbones and nearby, and to occasional growth rings which vary in number, placing, and depth of colour from individual to individual. The rest of the outside is a rather dirty white, with pale brown or blackish brown near the margin where the periostracum still remains. The ligament, which is external, appears as a short thick dark-brown line just behind the umbones. Most shells have only one hue (mono- chrome). In a small percentage it changes markedly. The change does not correspond, in at least some individuals, and perhaps in all, to a growth line or radius (Fig. 2). So far, only two hues have been seen (dichrome shells). It is possible that the change in inten- sity usual in monochrome shells may cor- respond to that from the first to the second hue in dichromes, and that changes from an intense hue to a paler version of the same hue may require special care in scoring. The MARINE BIVALVE SHELL COLOURS 299 FIG. 2. Dichrome shell of Масота balthica 14.2 mm long. The initial colour (outlined on right- hand valve) is about 2 11° 15 (P1) turning rather abruptly to white. The edges of the coloured patch correspond neither to growth lines nor to radii. slight change in hue occasionally seen be- tween the umbonal region and the muscle scars, which are sometimes slightly less blu- ish, may be due to the obvious difference in texture of the shell, as noted above, perhaps producing a slight Tyndall blue in the opaquer areas; it might be an actual change of pig- ment. In the following scores, only the definite changes of hue are noted. These can usually be seen rather faintly on the outsides of the valves. The colour scores for each hue and inten- sity are given in Table 5. Taken on the inside of the valve, they give the clearest idea of the actual pigment hue and intensity, and, since morph variation within a single population is far more likely to be genetic than not, they are the best basis for trying to understand the nature of the variation, i.e. polymorphic or continuous. The outside of the valves, how- ever, is presumably what a predator would TABLE 5. Colour atlas scores for colours of Macoma balthica shells. Hue and intensity Symbol Very deep purple VDP, Deep purple DP, Medium purple MP, Pale purple PP; Faint purple ER: Pinkish white WhP, Very deep pink VDP> Deep pink DP> Medium pink MP; Pale pink PP; Very deep peach VDP3 Medium peach MP3 Deep orange DO Medium orange MO Pale orange PO Faint orange FO Deep orange-yellow DOY Medium orange-yellow MOY Pale orange-yellow POY Deep warm yellow DWY Medium warm yellow MWY Pale warm yellow PWY Faint warm yellow FWY Medium chrome yellow MCY Yellow У Lemon yellow LY Faint yellow РУ Warm ivory WI Ivory | White Wh Villalobos score Ridgway equivalent Раз ХИ Amaranth Purple — 318215 ХИ Chatenay Pink or хи Flesh Pink 3510217; XXVIII Shrimp Pink =o 10-18 XXVIII Shrimp Pink =o 12° 18 XXVIII Shrimp Pink —3 12° — XXVIII Shrimp Pink 3191 34.9213 ХИ Jasper Pink А XIII Coral Red —4 10° 16 near ХШ Coral Pink —4 10° 18 near XIV Pale Salmon 57012 near XIV Cornelian Red or XIV Apricot Orange Sk XXVIII Japan Rose 5+ 9 14 XIV Carrot Red or XIV Flesh Ocher —6 8° 14 XIV Apricot Buff Bar WO alte | Orange Pink 5+ 29° 219 — 6 9° 15 XV Ochraceous Salmon 6 102 17 Ш Саристе Вий 6 11° [18] пеаг Саристе Вий 6/7 10° 15 XV Zinc Orange 6/7 9° 15 near Ш Capucine Buff /7 9° 19 near Ш Capucine Buff METE — ИВО Ш Orange Buff 8/9 10° 17 Ш Light Orange Yellow 9 ?9° [18] near XXX Colonial Buff 9/10 ?9° [19+] near XVI Straw Yellow 27/8 ?7° [19+] near XV Light Buff 28 8° [19+] near XV Antimony Yellow — 0° 20 — 300 CAIN TABLE 6. Occurrences of hues and intensities in random samples of Macoma balthica, Wirral Penninsula. V V ?\ D D M P F Wh D D M B D M D M P F Sample Pi Р1 Р1 Р1 Р1 Р1 Р2 Р2 Р2 re ES P3 O O O O (1) Red Rocks, 29 Dec. 83 A.J. G 4 61 88 12 2 14 6 11 1 7 1 Té % 0.97 14.73 21.26 2.90 0.48 3.38 1.45 2.66 0.24 1.69 0.24 1.69 (2) Red Rocks, 1985 |. Wallace 2 79 48 15 23 12 89 20 13 11 4 у 15 7 6 % 0.34 13.25 8.05 2.52 3.86 2.01 14.93 3.36 2.18 1.85 0.67 1.17 2.52 1.17 1.01 (3) Hoylake 26 Dec. 83 А. J. C., J. T. C 8 77 48 8 1 3 6 6 10 % 2.70 26.01 16.22 2.70 0.34 1.01 2.03 2.03 3.38 (4) Hoylake 26 Dec. 83 А. J. C. 8 79 49 12 7 2 1 8 10 4 2 8 10 % 231 2283 14.16 3.47 2.02 0.58 0.29 2.31 2.89 1.16 0.58 231 2:89 (5) Ноуаке 26 Оес. 83 ТС: 3 21 93 15 6 11 6 6 5 7 % 0.89 6.25 27.68 4.46 1.79 3.27 1.79 1.79 1.49 2.08 Totals 25 317 326 62 2 51 14 110 34 13 11 37 12 35 16 40 % 1.26 15.95 16.40 312 0.10 257 0.70 5.53 1.71 0.65 0.55 1.86 0.60 1.76 0.80 2.01 usually see. In general, the umbonal region samples (Table 6) so far suggests a more or outside is the same hue as the inside of the less normal distribution of intensities. shell but a little paler; the rest of the adult shell 2) Pink, Po is only faintly tinged or plain white except for The hue is noticeably a more yellowish red the remains of the periostracum. than in P,, more obviously so on the inside than on the outside of the shell. In all but sample 2 (Table 6) only a few have been seen; in that sample, the distribution again A. Monochromes suggests a continuum rather than discrete morphs. 1) Purple, P 3) Peach, Ps Outside, the umbonal region shows vari- _ Both the umbonal region outside and the ous intensities of a slightly bluish red, a true inside are a definite peach-colour (colour of Tyrian purple (not violet). Inside, the shell is peach fruit, not flower), yellower again than more or less uniformly purple. The range of P2. Only small numbers have been seen in intensity is considerable (Table 5), from 13 any sample. to more than 18, giving very deep, deep, 4) Orange, O de medium, pale and faint purple shells. The A distinctly orange hue, both inside and shells called pinkish white are placed here out. Although only a few shells have been because although they are only recognis- seen, there is a marked variation in intensity, able by careful comparison with ivories and though not nearly the full range. whites and the colour (internal only) is so 5) Orange-yellow, OY faint that it is not easily scorable, they do not Although intermediate between (4) and seem to belong with other hues. (6), these seem to segregate clearly, again A careful scoring of shells may perhaps with a good range of intensity. show a discontinuity between MP,, PP, and 6) Warm yellow, WY — , FP,, but the distribution of numbers in the These occur only in Dr. Wallace's sample, MARINE BIVALVE SHELL COLOURS 301 D M Р D M Р Е D M F OY OY OY WY WY WY МУ CY CY У LY Y WI Wh Dichromes Total 8 27 3 A 5 20 48 56 — 28 414 1.93 6.52 0.72 0.97 0.24 1.21 4.83 11.59 13.53 6.76 5 IS №3 6 7 5 1 3 3 24 9 64 53 40 596 0.84 1.51 2.18 0.50 1.01 1.17 0.84 0.18 0.50 0.50 4.03 1.51 10.74 8.89 6.71 18 20 3 2 5 12 28 18 23 296 6.08 6.77 1.01 0.68 1.69 4.05 9.46 6.08 7.77 20228 4 3 18 35 24 32 346 0.58 8.09 1.16 0.87 5.20 10.11 6.94 9.25 ¡SO 4 10 1 12 63 Di 336 4.46 2.98 1.19 2.98 0.30 3.57 18.75 8.03 6.25 5 ES 6 7 5 1 1012314 AA E A 1988 0.25 2.62 493 015 0.30 0.35 0.25 0.05 050 1.16 0.20 1.91 357 11.97 8.95 7.24 no. 2 in Table 6, but segregate in it from O, morphs. The slight variation in intensity in OY and the various forms of Y, again with a (7), (8), (9) and (10) is probably due only to considerable range of intensity although in a the low numbers seen. Usually (Table 6) the total of only 21 shells. medium intensity is the commonest in each 7) Chrome yellow, CY hue. Very few have been seen, 11 in mono- While (10) has been labelled FY, the chromes, and the assignment of one of higher luminosity of yellow as a hue, com- these to DCY is tentative but supported by pared with purple, for example, makes direct variation in the dichromes, in which the comparison with the others difficult. Perhaps second colour appears to be MCY in 7 and a separate colour-designation and symbol DCY in 1. should be used. 8) Yellow, Y 11) Ivory, | Pure yellow. Very little variation in inten- This group, so pale that its exact hue and sity has been seen (see below under (10)). page in the atlas are dubious, segregates 9) Lemon yellow, LY quite clearly from the hues above and from Markedly more towards yellow green than white. Warm ivory also seems to segregate, (8). Only 4 shells seen, little variation in but it may be an extreme of ivory. Their intensity. relative numbers in Table 6 do not exclude 10) Faint yellow this possibility. Difficult to score, falling between two 12) White, Wh pages of the atlas. It is natural to wonder whether this and the last three hues are not merely due to differences in intensity of the same pigment. From the colour atlas this is not the case, as they cannot be matched to different intensities on the same page, and are therefore best regarded as separate This is a good plain white, segregating well from the ivories and other colours, ex- cept that careful comparison is needed to separate pinkish white WhP, (see under (1) above). It appears to be produced by the absence of pigments, and therefore has no variation in intensity. 302 CAIN TABLE 7. Macoma balthica dichrome shells in Wirral samples. Sample nos. as in Table 6. (1) (2) = ae = O < NO = oo — ay >» № — — D — NO Total 28 40 (3) (4) (5) Total 2 2 1 1 6 1 3 1 10 4 1 4 2 1 1 3 6 12 7 5 17 2 2 8 9 2 3 3 3 2 2 5 1 2 3 3 1 8 7 1 1 14 6 1 3 3 3 2 1 4 8 16 4 4 1 1 № [9%] [0%] № № = —^ > > 7.24 B. Dichromes The full list of dichromes seen is given in Table 7 with their occurrence in the samples. So many are represented by one, two or three shells (19 out of 30) that doubling the size of the samples might well double the number of sorts observed. If it is considered that differ- ent intensities of the same hue are parts of a continuum, we can lump the sorts in Table 7 by ignoring the variation of intensity of both the first and second colours; there still remain 11 distinct dichromic types. The total fre- quency of dichromes in each sample 1$ al- ways low (Table 7; mean 7.24%), but might be raised by very careful scoring of apparent monochromes. The full scores according to the scheme given above of all the Wirral samples are given in Table 6, with the details of dichromes in Table 7. Dr. Mallet's collection from near Camber Sands is of drifted single valves, somewhat worn. To avoid scoring the same individual twice, the valves were sorted by chirality as well as into colours, and the largest number, whether of left-hand or right- hand valves, was taken as the score for each colour. The Baltic specimens could only be scored from the outside. The scores for both are shown in Table 8. The lack of dichromes in the Baltic sample may or may not be genuine, but the lack of yellows probably is. (iv) Features of the polymorphism in Wirral samples The scores given in Tables 6 and 7, and the colour scores for the different hues in Table 5, MARINE BIVALVE SHELL COLOURS 303 TABLE 8. Scores of samples of Macoma balthica from non-Wirral localities. Purple, Pale pink Peach Orange orange A nr. Camber Sands J. Mallet 64 15 28 12 % 29.22 6.85 12.79 5:50 B Baltic coast G. Russell 52 ———— 8 % SU) STAN bring out a number of points with regard to the polymorphism. (a) Although the range of hues is wide, from purple through peach, orange and various yellows to ivory, it is continuous (atlas pp. 2 to 10) and restricted when compared with the full range of possible colours, only 9 pages out of 38. It cannot be said that this is due to an incapacity of molluscs, or even bivalves, for producing greens, blues and violets. Dark blues and occasional purples are well known in some mussels (e.g. Mytilus edulis, Modiolus modiolus) and greens in others (Musculus). Brown, as rays or mottles, is conspicuous in Glycymeris glycymeris, Mactra corallina and Venus striatula. A rich purple is seen inside the valves of Gari fervensis, violet in Donax vittatus. Numerous other examples could be given from foreign bivalves. The general area of muddy sand scores on the colour charts (Fig. 1) is around 7 4” 10, extending from thence towards 1° 1 as it gets blacker with contained mud. (At such dark colours as these, the page number can vary widely with little effect). The general ensem- ble of the shell colour scores, projected on to the same page is about as far from the muddy-sand and mud scores as it can be (Fig. 1). It is obvious, and equally so from a glance at the live animals, that there is no trace of cryptic coloration here. It is interest- ing to contrast with M. balthica the common cockle, Cerastoderma edule, often found with it or nearby. Large juveniles and adults have a white shell, very strongly constructed, and easily seen like the ivory and white morphs of M. balthica, but very small Orange- Ivory, yellow Yellow white Dichromes Total 18 62 20 219 8.22 28.31 9.13 407 467 87.15 juveniles usually have a scattering of dark markings (rows of short dashes) which effectively mottle the shell and, make it less conspicuous. If the coloration is not cryptic, since it is markedly diverse it is possible that we have here a form of apostatic polymorphism (Clarke, 1964). But if so, the coloration is not diverse enough for a fully apostatic poly- morphism. It is easy to predict from the atlas what such a polymorphism should be. If nis the number of morphs observed, and any one taken at random is found to fall on a particular page of the atlas, then the others should be spaced as far from it and each other as possible. They will occupy the vertex positions of a polygon, inscribed within the cylinder of pages (Fig. 3), with number of sides and of vertices equal to n, one of the vertices being located at the page with the observed score for a morph. Thus, if we take it that there are three principal morphs, one of which is purple, scoring on p. 2, the other two should be at the vertices of an inscribed equilateral triangle within the circle (Fig. 3), i.e. near to p. 15 (lime-green) and 27 (cobalt- ultramarine). Purple-pink is in fact one princi- pal class of hue, but the others actually ob- served are orange, р. 5-6, and yellow, р. 8 + to 10-. They are approximately equally spaced but far too close together for pure apostatic selection. A further separation could be made by alternately displacing the colour score up and down the pages. If orange is lowered in its intensity score, a rich brown results. Lowering the score for yellow would produce a dull green. Neither of these is found. Raising either would produce pale 304 FIG. 3. Predicted and observed hues of morphs of Macoma balthica on the hypothesis of apostatic selection acting alone. The 38 radii are the edges of the pages of the Villalobos colour atlas grouped as half-sections of a cylinder with the neutral colour row (common to all pages) as the central axis. Observed hues labelled with morph symbols (P1, OY etc., see Table 5). They fall into 3 major groups, purple-pink, orange, and yellow. If purple-pink is taken as the datum, then under plain apostatic selection the other two hues should be as far apart from it and each other as possible. White is also a morph but is not allocatable to any hue; if shown, it will be at the centre of the circle. The vertices of the inscribed equilateral triangle imposed on the circle give the predicted positions of non-white hues; these are close to lime green (1, p. 15) and cobalt-ultramarine (си, р. 27). P. 3 is scarlet, s, р. 7 orange, o, in the Villalobos terminology, which is marked for the principal hue-names by lower-case letters. colours, but these would then approximate to whites. (b) For apostatic polymorphism, not com- plicated by various degrees of crypsis (as in Cepaea, Cain, 1977, 1983) not only should the hues be as diverse as possible, they should also differ visibly as much as possible. There should be no admixture of neutral colour, which would render them more simi- lar, and all should score the maximum (12°) for saturation. While most are duller than this (Table 5) almost all approximate to that edge of the page. The dullness probably means only that the colour is not displayed on a pure white background. Equally, the most brilliant intensity should be used in all, since great intensities produce very similar browns, olives etc., and lesser ones produce pale colours, MARINE BIVALVE SHELL COLOURS 305 approximating to white. Here we find a marked discrepancy (Fig. 1). Tothe eye, hues of saturation 12° and intensities between 10 and 14 make the most noticeable difference in colour from page to page. Yet (Table 5) in М. balthica the same hue, e.g., P;, ranges from above 10 to beyond the limits of the page; this is true for P,, O, WY and one of the paler yellows, and nearly so for Po. Probably larger samples would extend the range for other hues less abundantly represented so far. This variation towards white would in itself argue against simple apostasy, but it is the more remarkable in that white and ivory (themselves strikingly alike) are abundant morphs in all the samples, and these pallid hues converge on them. (c) This tendency to similarity is taken much further when we look at the outsides of the shells. Small juveniles have glossy translu- cent or nearly transparent shells which show their hue and intensity almost as well exter- nally as internally. On larger shells the umbonal region continues to show the colour of the juvenile, but the rest of the shell is an Opaque greyish white, relieved only by the pigmentation of occasional growth lines, sel- dom as intense as the umbonal colour, and the remains of brown periostracum near the growing edge. Only a faint general tinge of the internal hue is visible outside. Juveniles are very diverse, other age-groups less so, al- though it is true that the eye is caught by the bright umbonal colours forming two patches sharply divided posteriorly by the conspicu- ous brown ligament. (d) Some of the colour-classes may be produced by combinations of the pigments mediating the principal colours. This is cer- tainly the case in populations of the winkle Littorina rudis (Maton) that | have examined, in which at the mouth of some shells two pigments can be scored because their distri- butions do not quite overlap. In the present polymorphism, no such convenient separa- tion has been observed, and it cannot be asserted that the orange-yellow class, for example, is in fact made by the suffusion of the shell by both the yellow and the orange pigments, but it may be; and if so, presumably it is heterozygous, with no dominance, for the two colours. (e) In some forms two pigments are cer- tainly present but they do not overlap. As the shell grows there is a fairly rapid switch from one to another. In almost every case, the change is from a redder to a yellower or whiter hue. In Cepaea, colour heterozygotes not infrequently begin with the recessive colour and then rapidly change to the domi- nant, so that the recessive is only seen close to the apex of the spire (only very rarely have | seen shells that changed much later). In Macoma balthica the change is regularly at about Y4 of the final area of the shell. A number of these dichrome shells, espe- cially those changing to pale yellow or white, hardly differ from monochromes of the same first (umbonal) hue since these are usually white externally except for the umbonal areas. (v) Frequencies of morphs in Wirral samples Many of the cells in Table 6 would give expectations of less than 5 for a x? test. The least further grouping that would give a prob- ably valid test is VDP1 + DP1, PP1 + FP1, VDR2=+:DP2 MP2 + РР2; ЭРЗ:- MP3; DO + MO, PO + FO, DOY + MOY, WY + DCY + МСУ, Y + LY + FY; all expectations are more than 5 except one, sample 3, WY + DCY + MCY, which is 4.76. This grouping shows a highly significant x? (d.f. 64) = 476, P << 0.001. Separation by locality into (1) + (2), Red Rocks, and (3) + (4) + (5), Hoylake, necessitates the further grouping of the Hoylake samples, namely PP1 + FP1 + Wh1, all P2, all O, and all Y after POY. Both sets of samples still show highly significant differences, as follows. Red Rocks x? (d.f. 16) = 153, P << 0.001. Hoylake x? (d.f. 24) = 121, P << 0.001. In the Red Rocks samples, the major dif- ferences are in the proportions of MP1 and VDP2 + DP2, POY, and WI. As the propor- tions of the P1 groups other than MP1 are in good agreement, it seems unlikely that a mistake has been made in scoring the inten- sity of pigmentation; the same is probably true of VDP2 + DP2, since the other P2 group (MP2 + PP2) differs only by 5%, and similarly of the OY groups and WI against |. In the Hoylake samples, there is a marked discrepancy in both VDP1 + DP1 and MP1 in sample (5), deficiencies in DOY + MOY in sample (4) and in POY in sample (5), and an excess of | in (5). Since the shells are from a shell-wreck, not collected alive in situ, it is difficult to suggest biological reasons for these discrepancies. 306 CAIN No shells were perforated by Natica or similar predators, nor were any single valves col- lected, so that the reasons for differential drifting demonstrated experimentally by Lever (1958; Lever et al., 1961, 1964) do not apply. It is conceivable that different samples have different proportions of intensities of hues because there is some fading with time, or indeed that the exact intensity produced in a given shell is partly dependent on its condi- tions of growth and therefore on weather or other environmental factors. What are unlikely to be affected by either of these factors are the exact hues, most of which seem to be true morphs. Regrouping under each locality by hue still gives highly significant differences for the two Red Rocks samples; the three Hoylake ones are now just not sig- nificant. A further grouping into all pinks, all oranges (including OY), all yellows, and all whites (| + Wh) produces the same results. Regrouping by effective hue requires the transfer of FP1, WhP1, FO, FWY and FY to a separate group comprising also WI, | and Wh, and to obtain expecteds greater than 5 re- quires grouping by hue. Again, samples 1 and 2 are highly significantly different, 3, 4 and 5 are on the borderline of significance (P = 0.01). Since the Wirral samples except no. 2 (Dr. Wallace’s) are taken from the same shell- wreck, it is probably better biologically to regard them as different subsamples from the same huge sample. As such, they agree well with one another (Tables 6, 7). Dr. Wallace’s sample was taken later from the same area of beach as no. 1. It differs strikingly in the nearly equal proportions of P; and P; instead of a vast preponderance of P, over Ps, and in the presence of WY (absent in the other samples). This last, and an apparently high proportion of FY (4% instead of about 1%) are compensated for to a large extent by the relatively low proportions of OY (4 as against 9%). The significance of these differences is considered further in relation to the south coast sample. If we group the hues into P, O, OY, WY + CY + Y, LY + FY, 1, Wh, and all dichromes, the Wirral samples are so similar that they can be combined and give the following percentages: Р 50.40 LY etc. 2.11 O 5.18 | 15.54 OY 7.80 Wh 8.95 WY etc. РУ Dichromes 7.24 By far the commonest are the pinks (broad sense), with ivories and whites well behind (together, 24.49) and the various oranges and yellows making up 17.86%. In actual appear- ance, however, the various faint intensities and pinkish white should be allocated with ivory and white, giving a percentage of effec- tive white overall as 31.33 (range in samples, 25.00 to 34.52), as against 24.49. In practice, looking at actual samples on the shore, there is no doubt of the visual effect—the majority are pinks (broad sense) and whites, with a sparse scatter of dichromes, yellows and or- anges. The finer distinctions that can be made on a sample of cleaned shells are largely obliterated. Since many shells are partly obscured by sand or mud, and often show growth-line variations in intensity of colour, the dichromes are not very conspicu- ous, and the whole, to a casual eye, becomes a scatter of pinks and whites, with a very few yellows тагке у conspicuous. To a careful observer, including no doubt predators, dis- tinctions are much more obvious. (vi) Non-Wirral samples Dr. Mallets sample from near Camber Sands, A, being somewhat waveworn, and that from the Baltic, B, not being removable from the bodies, less fine divisions into hues have been used; the scores are given in Table 8. As compared with the Wirral sam- ples, A is somewhat low in pinks (30% as against 45%), high in orange (13% vs. 3-6%), deficient in orange-yellows (absent), high in yellows taken together (8% vs. 3-4%), and very similar in ivory + white and dichromes. In its high values of orange and yellow it approaches sample 2 (Dr. Wallace’s) but this last was collected by scooping, not by hand- picking. Otherwise, the general agreement of A with the Wirral samples is good. The Baltic sample, in complete contrast, is 87% white, 1.7% orange and 11% pink (all hues taken in the broad sense). The pinks give the impression of being very uniform, much more so than in the English samples. Dr. Russell tells me that, at the place of collection, the bottom was largely covered with a whitish calcareous deposit. This is the only sample that could be protectively coloured. Smith (1975) noted that in Donax faba (Gmelin) the very small shells are about the size of the sand grains of the substrate and, scattered among them, may be hard to see because of their diversity of colours. The MARINE BIVALVE SHELL COLOURS 307 20 o o GO O OOOO A Е O ODE + it O O @) = intensity orange uby scarlet yellow © = = E | = emerald violet magenta FIG. 4. Variation of intensity of generating square for each hue in the Villalobos colour atlas; in effect, variation of luminosity of hues (crosses). Orange, yellow, and yellow-green are the most luminous. Colour scores for hue and intensity of Macoma balthica morphs and variants (open circles) show a close correspondence to the generating intensities but also a considerable scatter of paler colours. White shells are not allocable to any hue. muddy sand on which M. balthica is found appears to be finer in texture, and it is unlikely that small juveniles would be cryptic, although new-fallen spat possibly might. (vii) Frequencies and apostatic selection As pointed out above, except for the Baltic sample none of the hues and intensities come near to those of the background against which M. balthica would be seen. If all the morphs and forms were equally conspicuous on a given background and subject to apostatic selection, one would expect the frequencies of а! to be equal at equilibrium. If not, less conspicuous ones will be at a fre- quency higher in proportion to their degree of crypsis, at least to a first approximation. Equivalent molar concentrations of different pigments give different intensities of colour. It is noticeable in the Villalobos atlas that the colour-square on each page givina the colour from which the rest of the page is generated by concentration, or dilution with transparent medium, or dilution with neutral tint, varies rather regularly with hue (Fig. 4) such that in the sequence of colour used by M. balthica, at equivalent concentrations reds cut out more light than oranges, these than yellows, and these, of course, than ivories and whites. The intensest hues observed so far in M. balthica give intensity values of approximately Pink Peach Orange 14 15 17 Chrome Lemon Ivory 18 19 unscorable (all of which, of course can appear as fainter dilutions), while wet muddy sand and mud are from about 10 at the very highest to 0 (prob- ably usually below 6 if the substrate carries live М. balthica). Their colours are deep browns, black-browns and near-black, often of very indeterminate hue. Wet pure sand, of the sort usually described as yellow, falls about p. 7, which is between chromes and lemon yellows, so that its generating hue is within the range of colours seen in M. balthica, but is much darker than they are; and in fact yellows stand out, to the human eye, very conspicuously on dirty sand, let alone mud. 308 CAIN Comparing pages in the atlas, the distinc- tion between hues is most immediately recognisable in the purest colours (12°) be- tween about 12 and 17, those below being rather dark, those above too dilute, рае and washed-out. To a predator with colour-vision, therefore, different colour morphs should look most distinct at these intensities. In fact, colour scores (Table 5) in this range are shown by the deep and medium intensi- ties. The darkest are purple, pink and peach, and these, therefore, are closest to sandy mud in intensity. White and the palest colours are obviously furthest away. If we arrange the hues in increasing intrinsic paleness, this should be the ranking in decreasing fre- auency if all hues are equally conspicuous but intensity is acted on by apostatic selection. Table 6, which is in this order of increasing luminosity of hue (not intensity) immediately shows that the hypothesis is untenable, since the most abundant classes are at both ends of the distribution. Regrouping by effective hue, i.e. removing FP;, WhP,, FO, FWY, and FY to the white group (WI, | and Wh) de- creases the frequencies of purples and pinks and increases that of white which should be the least. Fig. 5 gives frequency distributions of the frequencies, for internal or effective scores, lumped by hue or completely sepa- rate, dichromes being omitted as heteroge- neous. None approximates to a scatter around the expected value. If colour is supposed to have any cryptic effect, then OY and WY are the likeliest to approximate to that of sand. They are not commoner than the less cryptic colours in any sample. (vil) Biology of Macoma balthica As an easily identified and often exceed- ingly abundant species, M. balthica has been used often for research, both ecological and physiological; the large literature can be ap- proached through the references in Beukema & Meehan (1985), Brafield (1963), Brafield & Newell (1961), McLusky & Allan (1976), Meehan (1985), and Meehan & Diaz (1984). As so often with well-known, abundant and variable species, the taxonomy is only now beginning to be worked out properly; there are strong indications (Meehan, 1985; Beukema & Meehan, 1985) that eastern North American populations (and presumably western North American also) are not conspecific with the European M. balthica. % occurrence Р.Р РО O0 WIE У ЕР Welewh 1 2.3 MAN МУ morph 60 50 40 Ф о = © 3.0 - о о o 20 x 10 р Es] E O Y Wh principal hues FIG. 5. Frequencies of hues plus white in the combined Wirral samples of Macoma balthica, in order of increasing luminosity, the few dichrome shells being omitted. 5a, percentage occurrences of all morphs, data from Table 6. If all morphs are equally conspicuous and apostatic selection is acting, the expected percentages are given by the horizontal line at 7.69%. Since hue and intensity vary together (fig. 4) if intensity is being selected (e.g. by colour-blind predators), the percentages should decrease steadily from left to right, the pink morphs being darker and more like the background. If any hue is cryptic it should be the yellow class, and percent- ages should decrease from it in both directions. 5b, faint colours and pinkish white lumped with white, and correspondingly all pinks, all oranges, all yellows and all effective whites lumped. The only result is to emphasize the bimodality of the percent- ages, unexpected on any of the three hypotheses just given. MARINE BIVALVE SHELL COLOURS 309 Of particular importance for the present work is the occurrence of M. balthica largely between the tidemarks, also extending just below but usually most abundant at the mid- tide level (Brafield & Newell, 1961) and so very much exposed to bird predators when the tide is out. It occurs in mingled sand and mud or actual mud (Yonge, 1949; Brady, 1943; Brafield & Newell, 1961) being largely replaced in clean sand by Tellina tenuis (Yonge, 1949; Barrett & Yonge, 1958); it is especially abundant in estuaries (e.g., Brady, 1943; Stopford, 1951). When the tide is out it is found at depths from the surface to 12 cm but can be seen burrowing through wet sub- strate (Stopford, 1951). Brafield (1963) shows that anoxic conditions in the substrate cause it to rise to the surface where it can breathe oxygenated water lying in ripple hollows, and Brafield 8 Newell (1961) document its moving about feeding on the beach and leaving fur- rows of various shapes in the substrate, straight, U-shaped or nearly circular. The siphons are highly extensile, and when the tide is in, M. balthica may feed with the body at several cm depth in the substrate, sucking up detritus on the substrate surface (deposit feeding) or if suspended food is available, taking that. Brady (1943) suggests that its ability to do both allows it to compete with obligatorily deposit-feeding polychaetes. Newell (1965) has investigated its feeding on detritus. Such data as these make it clear that M. balthica is very much at risk from visual predators (birds) when the tide is out; fish may also be important when the tide is in. Although many fish predators take only the siphons, or, as with flatfish, use touch not sight in their foraging, some at least may snatch out individuals in the top layer of the substrate. In an important paper, Beukema & Meehan (1985) survey the distribution of shell colour in M. balthica on both sides of the Atlantic, using the broad categories red, orange, yellow and white, and noting the occurrence of a few percent of bicoloured shells “mostly either white or yellow with a red spot near the umbo”. There are marked differences be- tween the eastern American and the Euro- pean populations. In Europe, as a result of extensive collections, they note that red tends to be most abundant in the north, yellow is largely confined to France, and the proportion of white is highly variable. They make the intriguing observation that in Europe “The share of these uncoloured shells appeared to be particularly high in brackish waters (as in the Baltic), but was also more than about 0.6 at some sampling places with a salinity near 30%. In very muddy sediments the propor- tions of indistinctly coloured shells was often high but not necessarily so.” “Shell colours were most vivid in samples from sandy (i.e. exposed) places; samples from muddy sedi- ments contained high proportions of whitish, greyish or bluish shells; at times, to such an extent that the colour sorting of the sample had to be abandoned.” My visits to numerous coastal localities in the British Isles and Brittany, and some in the Bay of Biscay (French and Spanish) and the French Mediterranean coast do not suggest that there is any special yellowness in the sand substrates in France as against those in Britain, whereas there is in France a marked development of chrysophycean algae that colour a band of rocks on the shore and affect the colour of the coincident Littorina species. The Baltic sample reported here agrees en- tirely with Beukema 8 Meehan's remarks, as do the others which are from sandy mud rather than only mud (probably including the Camber sands sample when alive). Beukema 8 Meehan comment on the diversity of shell colour “nor can we easily suggest a possible adaptive significance, as the growing individ- uals disappear below the sediment surface as soon as the shell colours develop”. DISCUSSION К was concluded tentatively (р. 295 above) from the general survey of British bivalve coloration in relation to probable exposure that it supports the expectation that the exter- nal colours and patterns of marine bivalves are strongly influenced by visual predators. Even infrequent exposure, or exposure mainly as the juvenile, may in time produce a very considerable selective effect. The biology of Macoma balthica (p. 308) certainly suggests that this species is open to considerable pressure by visual predators, perhaps even more to terrestrial than to aquatic ones. The vision of its principal pred- ators is not likely to be influenced by depth of water and consequent filtering of colours. Its display of a true colour polymorphism with no morphs resembling its background immedi- ately suggests apostatic selection. However, various features of the actual morphs, namely 310 CAIN the narrow range of hues, production of very pale forms of most hues resembling the com- mon white morphs, and the relative abun- dances of different morphs all raise difficulties on the supposition of pure apostatic selection. The very different distributions of major colour classes in Europe recorded by Beukema & Meehan (1985) might suggest some sort of climatic selection with darker morphs in the northern parts ofthe range, and perhaps even with white forms more common on the blacker backgrounds which would heat up more dur- ing neap tides in the summer. If this is true, selection should be stronger on the juveniles which, with their shorter siphons, would need to stay closer to the surface of the substrate. A further suggestion is that the animals are mimicking their own dead shells which often lie about in great abundance, and are useless for food; but on this supposition, all the dead shells should be alike. It is conceivable that the small juveniles, much more vulnerable to predators, are showing an apostatic poly- morphism and later become white for this very reason; this would not involve group selection. In this case the retention of colour inside the shell seems to be a functionless continuation of the juvenile colour. All such suggestions as these can only be tested by a full investigation into the predator- prey relationships of the species, especially in the young stages, with special attention to low frequencies of predation, the habits and modes of search of different predators, and especially the habits of the prey. More bivalves than yet suspected may be quite athletic, at least when juvenile. Speculation is useful only as it suggests ideas to work on. While no doubt paragraphs could be devoted to possible influences of developmental con- straints, linkage disequilibria, and the like, the total absence of any data on such matters suggests that they be postponed. ACKNOWLEDGEMENTS | am deeply grateful to J. Gittins, J. Mallet, S. С. Meredith and 1. Wallace for collections, to my wife for assistance in collecting and scoring, and to A. D. Ansell, B. C. Clarke, Е. М. Cook, G. M. Davis, К. E. Hoagland, Р. В. Mordan, and В. $. Morton for extremely useful comments. Fig. 2 was taken by B. Lewis. The other figures were drafted by N. Karnow. REFERENCES ABBOTT, R. T., 1974, American seashells. The marine Mollusca of the Atlantic and Pacific coasts of North America. New York, etc.: Van Nostrand Reinhold Co. Ed 2. ALLEN, J. A., 1954a, A comparative study of the British species of Nucula and Nuculana. Journal ofthe Marine Biological Association of the United Kingdom, 33: 457-472. ALLEN, J. A., 1954b, On the structure and adapta- tions of Pandora inaequivalvis and P. pinna. Quarterly Journal of Microscopical Science, 95: 473-482. ALLEN, J. A., 1958a, On the basic form and adaptations to habitat in the Lucinacea (Eula- mellibranchia). Philosophical Transactions of the Royal Society of London, ser. B, 241: 421-481. ALLEN, J. A., 1958b, Observations on Coch- lodesma praetenue (Pulteney) (Eulamelli- branchiata). Journal of the Marine Biological Association of the United Kingdom, 37: 97-112. ALLEN, J. A., 1960, Manganese deposition on the shells of living molluscs. Nature, 185: 336-337. ALLEN, M. F. 8 ALLEN, J. A., 1955, On the habits of Pandora inaequivalvis. Proceedings of the Malacological Society of London, 31: 175-185. ANSELL, A. D., 1961, The functional morphology of the British species of Veneracea (Eulamel- libranchia). Journal of the Marine Biological As- sociation of the United Kingdom, 41: 489-515. ANSELL, A. D., 1967a, Leaping and other move- ments in some cardiid bivalves. Animal Be- haviour, 15: 421-426. ANSELL, A. D., 1967b, Burrowing in Lyonsia norvegica (Gmelin) (Bivalvia: Lyonsiidae). Pro- ceedings of the Malacological Society of London, 37: 387-393. ANSELL, A. D., 1967c, Leaping movements in two species of Asaphidae (Bivalvia). Proceedings of the Malacological Society of London, 37: 395-398. ANSELL, A. D., 1968, Defensive adaptations to predation in the Mollusca. Proceedings of Sym- posium on Mollusca, Marine Biological Associa- tion of India, part Il: 487-512. ANSELL, A. D., 1969, Leaping movements in the Bivalvia. Proceedings of the Malacological Soci- ety of London, 38: 387-399. ANSELL, A. D., 1983, The biology of the genus Donax. In: MCLACHLAN, A. 8 ERASMUS, T., eds., Sandy beaches as ecosystems, рр. 607-635. The Hague, Junk. BAIRD, В. H., 1958, On the swimming behaviour of escallops. Proceedings of the Malacological So- ciety of London, 33: 67—71. MARINE BIVALVE SHELL COLOURS 311 BARRETT, J. & YONGE, C. M., 1958, Collins pocket guide to the seashore. London: Collins. BEUKEMA, J. J. & MEEHAN, В. W., 1985, Latitu- dinal variation in linear growth and other shell characteristics of Macoma balthica. Marine Biol- ogy, 90: 27-33. BOUCHET, P., DANRIGAL, R. & HUYGHENS, C., translated by PICTON, B. E., 1979, Living sea- shells. Molluscs of the English Channel and Atlantic Coasts [of France]. Poole, England: Blandford Press. BRADY, F., 1943, The distribution of the fauna of some intertidal sands and muds on the Northumberland coast. Journal of Animal Ecol- ogy, 12: 27-41. BRAFIELD, A. E., 1963, The effects of oxygen deficiency on the behaviour of Macoma balthica (L.). Animal Behaviour, 11: 345-346. BRAFIELD, A. E. & NEWELL, G. E., 1961, The behaviour of Macoma balthica (L.). Journal of the Marine Biological Association of the United King- dom, 41: 81-87. CAIN, A. J., 1977, The efficacy of natural selection in wild populations. In: GOULDEN, C., ed., The changing scenes in natural sciences, pp. 111-133. Academy of Natural Sciences, Phila- delphia, Special Publications no. 12. CAIN, A. J., 1983, Ecology and ecogenetics of terrestrial molluscan populations. In: WILBUR, K. M., ed.-in-chief, The Mollusca, 6 Ecology (RUS- SELL-HUNTER, W. D., ed.), pp. 597-647. Aca- demic Press, New York, London, etc. CAIN, A. J., 1987, The scoring of polymorphic colour and pattern variation and its genetic basis in molluscan shells. Malacologia, 28: 1-15. CAIN, A. J., SHEPPARD, Р. М. & KING, J. М. B., 1968, Studies on Cepaea |. The genetics of some morphs and varieties of Cepaea nemoralis (L.). Philosophical Transactions of the Royal Society of London, ser. B, 253: 383-386. CLARKE, B. C., 1964, Frequency-dependent se- lection for the dominance of rare polymorphic genes. Evolution, 18: 364-369. COTT, H. B., 1940, Adaptive colouration in ani- mals. Methuen, London, xxxii + 508 pp., 48 pl. COX, L. R., revised by NUTTALL, C. P., 1969, Coloration. In: MOORE, R. C., ed., Treatise on Invertebrate Paleontology, part N, Mollusca 6 (Bivalvia), pp. 70-72. University of Kansas and Geological Society of America, Lawrence, Kan- sas. FORBES, E. & HANLEY, S., 1853, A history of British Mollusca, and their shells. 4 vols. Van Voorst, London. GILMOUR, Т. H. J., 1963, A note on the tentacles of Lima hians (Gmelin) (Bivalvia). Proceedings of the Malacological Society of London, 35: 81-85. GILMOUR, Т. H. J., 1967, The defensive adapta- tions of Lima hians (Mollusca, Bivalvia). Journal of the Marine Biological Association of the United Kingdom, 47: 209-221. GOSLING, E. M., 1984, The systematic status of Mytilus galloprovincialis in Western Europe: a review. Malacologia, 25: 551-568. GOULD, S. J. & LEWONTIN, R. C., 1979, The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist pro- gramme. In: SMITH, J. MAYNARD & HOLLI- DAY, R., eds, The evolution of adaptation by natural selection. London: Royal Society, pp. 147-164. HELLER, J., 1975, Visual selection of shell colour in two littoral prosobranchs. Zoological Journal of the Linnean Society of London, 56: 153-170. JACKSON, J. B. C., 1973, The ecology of molluscs of Thalassia communities, Jamaica, West Indies, |. Distribution, environmental physiology, and ecology of common shallow-water species. Bul- letin of marine Science, 23: 313-350. JEFFREYS, J. G., 1863-1869, British Conchology, or an account of the Mollusca which now inhabit the British Isles and the surrounding seas. 6 vols. Van Voorst, London. KAUFFMAN, E. G., 1969, Form, function and evo- lution. In: COX, L. R., etc., eds. Treatise on Invertebrate Paleontology, Part N vol. 1. Mol- lusca, 6: Bivalvia 129-205. Geological Society of America and the University of Kansas: Lawr- ence, Kansas. LEVER, J., 1958, Quantitative beach research. 1. The “left-right-phenomenon”; sorting of lam- ellibranch valves on sandy beaches. Basteria, 22: 21-51. LEVER, J., KESSLER, A., VAN OVERBEEKE, A. P. & THIJSSEN, R., 1961, Quantitative beach research Il. The ‘hole-effect’. A second mode of sorting of lamellibranch valves on sandy beaches. Netherlands Journal of Sea Research, 1: 339-358. LEVER, J., VAN DEN BOSCH, M., COOK, H., VAN DIJK, T., THIADENS, A. J. H., S. J. [sic] & THIJSSEN, R., 1964, Quantitative beach re- search Ill. An experiment with artificial valves of Donax vittatus. Netherlands Journal of Sea Re- search, 2: 458-492. LINNAEUS, C., 1758, Systema Naturae ed. 10, 1. Holmiae: Laurentii Salvii. Facsimile 1956, Lon- don: British Museum (Natural History). [As Tellina balthica]. MARSHALL, J. T., 1891, The habitat of Montacuta ferruginosa. Journal of Conchology, 6: 399—404. McLUSKY, D. $. 8 ALLAN, D. G., 1976, Aspects of the biology of Macoma balthica (L.) from the estuarine Firth of Forth. Journal of Molluscan Studies, 42: 31—45. McMAHON, В. F. & MCMAHON, С. O'B., 1983, Leaping and swimming as predator escape re- sponses in the jackknife clam, Ensis minor Dall (Bivalvia: Pharellidae). Nautilus, 97: 55-58. MEEHAN, B. W., 1985, Genetic comparison of Macoma balthica (Bivalvia, Tellinidae) from the eastern and western North Atlantic Ocean. Ma- rine Ecology—Progress Series, 22: 69-76. МЕЕНАМ, В. W. & DIAZ, В. J., 1984, Comparison of Macoma balthica labial palps from the Eastern 312 CAIN and Western North Atlantic: another look. Jour- nal of Molluscan Studies, 50: 231-232. MITTON, J. B., 1977, Shell color and pattern vari- ation in Mytilus edulis and its adaptive signifi- cance. Chesapeake Science, 18: 387-390. MORTON, B., 1975, Dymantic display in Gale- omma polita Deshayes (Bivalvia: Leptonacea). Journal of Conchology, 28: 365-369. MORTON, J. E., 1954, The crevice faunas of the upper intertidal zone at Wembury. Journal of the Marine Biological Association of the United King- dom 33: 187-224. MORTON, J. E., 1962, Habit and orientation in the small commensal bivalve mollusc Montacuta fer- ruginosa. Animal Behaviour, 10: 126-133. NEWELL, R., 1965, The role of detritus in the nutrition of two marine deposit feeders, the prosobranch Hydrobia ulvae and the bivalve Macoma balthica. Proceedings of the Zoological Society of London, 144: 24-45. NEWKIRK, С. F., 1980, Genetics of shell color in Mytilus edulis and the association of growth rate with shell color. Journal of Experimental Marine Biology and Ecology, 47: 89-94. OCKELMAN, K. W., Turtonia minuta (Fabricius) a neotenous veneracean bivalve. Ophelia, 1: 121-146. RACKHAM, H., 1940, Pliny Natural History vol. 3, libri VIII-Xl. London: Heinemann. Cambridge, Mass.: Harvard University Press [see Book IX, section LIII.] REIMCHEN, T., 1979, Substratum heterogeneity, crypsis, and colour polymorphism in an intertidal snail (Littorina mariae). Canadian Journal of Zo- ology, 57: 1070-1085. ROPES, J. W. & MERRILL, A. S., 1973, To what extent do surf clams move? Nautilus, 87: 19-21. RUSSELL, P. C. J., 1971, A reappraisal of the geographical distributions of the cockles, Cardium edule L. and Cardium glaucum Bruguiere. Journal of Conchology, 27: 225-234. RUSSELL, P. C. J., 1972, A significance in the number of ribs on the shells of two closely related Cardium species. Journal of Conchology, 27: 401-409. RUSSELL, Р. С. J. 4 HOPNER PETERSEN, G., 1973, The use of ecological data in the elucida- tion of some shallow water European Cardium species. Malacologia, 14: 223-232. SEED, R. 4 ROBERTS, D., 1976, A study of three populations of saddle-oysters (family Anomiidae) from Strangford Lough, Northern Ireland. Irish Naturalists’ Journal, 18: 317-321. SMITH, D. A. S., 1975, Polymorphism and selective predation in Donax faba Gmelin (Bivalvia: Tellinacea). Journal of Experimental Marine Bi- ology and Ecology, 17: 205-219. STANLEY, 5. M., 1970, Relation of shell form to life habits of the Bivalvia (Mollusca). Memoirs of the Geological Society of America, 125, xiii, 1-296. STOPFORD, S. C., 1951, An ecological survey of the Cheshire foreshore of the Dee estuary. Jour- nal of Animal Ecology, 20: 103-122. TAYLOR, J. W., 1914, Monograph of the land and freshwater Mollusca of the British Isles, vol. 3. Taylor: Leeds. TEBBLE, N., 1966, British bivalve seashells. A handbook for identification. London, British Mu- seum (Natural History). VILLALOBOS-DOMINGUEZ, C. & VILLALOBOS, J., 1947, Atlas de los colores. Colour atlas. Libreria El Ateneo Editorial, Buenos Aires. WARMKE, С. |. & ABBOTT, В. T., 1962, Carib- bean seashells. A guide to the marine mollusks of Puerto Rico and other West Indian islands, Bermuda and the Lower Florida Keys. Wynne- wood, Pennsylvania: Livingston Publishing Co. YONGE, C. M., 1937, The formation of siphonal openings by Thracia pubescens. Proceedings of the Malacological Society of London, 22: 337-338. YONGE, C. M., 1946, On the habits and adapta- tions of Aloidis (Corbula) gibba. Journal of the Marine Biological Association of the United King- dom 26: 358-376. YONGE, C. M., 1949, The sea shore. London: Collins. (New Naturalist). YONGE, C. M., 1951, Observations on Sphenia binghami Turton. Journal of the Marine Biological Association of the United Kingdom 30, 387-392. YONGE, C. M. 8 ALLEN, J. A., 1985, On significant criteria in establishment of superfamilies in the Bivalvia: the creation of the superfamily Mesodesmatacea. Journal of Molluscan Studies, 51: 345-349. YONGE, C. M. & THOMPSON, T. E., 1976, Living marine Molluscs. London, Collins. YONGE, C. M. & THOMPSON, T. E., 1978, Evolu- tionary systematics of bivalve molluscs. Philo- sophical Transactions of the Royal Society of London, ser. B, 284: 199-436. MARINE BIVALVE SHELL COLOURS 313 APPENDIX A. British marine bivalves: types of coloration and probable degree of exposure. The types of colour and pattern variation (I-IV) and the probable degrees of exposure ((i)—(iii)) are defined in the text, section II (iii) and И (vi). Type of Maximum colour and Probable measurement pattern degree of Species and classification (cm'?) variation exposure BIVALVIA' PALAEOTAXODONTA Nuculoida Nuculacea Nuculidae? Nucula sulcata Bronn 1.90 Il (ii) Nucula nucleus (L.) 1827 Ш (ii) Nucula hanleyi Winckworth 2 Il (ii) Nucula turgida Leckenby & Marshall 1122774 Il (ii) Nucula tenuis (Montagu) 127% Ш (ii) Nuculanacea Nuculanidae Nuculana minuta (Miller)? 1.90 Ш (ii) Yoldiella lucida (Loven) 0.3273 Ш (ii) Yoldiella tomlini Winckworth 0.48" Ш (ii) Phaseolus pusillus (Jeffreys) 0.167 Ш (ii) (CRYPTODONTA) (Solemyoida) (Praecardioidat) PTERIOMORPHA Arcoida Arcacea Arcidae Arca tetragona Poli 5.08 Il (i) Arca lactea L. 1.90 IV (ii) Arca pectunculoides Scacchi 0.48" IV (ii) Limopsacea Glycimeridae Glycimeris glycimeris (L.)* 6.35 Il (ii) Limopsidae Limopsis aurita (Brocchi) 1.27 IV (ii) Mytiloida Mytilacea Mytilidae Mytilus edulis L.? 15.24 Il (i) Modiolus modiolus (L.) 22.86 Il (1)? Modiolus barbatus (L.) 6.35 Il (i)? Modiolus adriaticus Lam. 3.81 Il (i)? Modiolus phaseolinus (Philippi) 1.90 Il (1)? Adula simpsoni (Marshall) 1.90 Ш (iii) Musculus discors (L.) 127 Il (1)? Musculus marmoratus (Forbes) 1.90 Il (1)? Musculus costulatus (Risso) 1274 Il (i)? Musculus niger (Gray) 5.08 Il (i)? Crenella decussata (Montagu) 0.32" Ш (1)? Crenella prideauxi (Leach) 0:32" Ш (1)? Pinnacea Pinnidae Pinna fragilis Pennant 30.48 Il (1)? Pterioida Pteriina Pteriacea 314 CAIN APPENDIX A (Continued) Type of Maximum colour and Probable measurement pattern degree of Species and classification (cm??) variation exposure Pteriidae Pteria hirundo (L.) 7.26 Il (i) Pectinacea Pectinidae Pecten maximus (L.)*® 15.24 | (i) Chlamys sulcata (Muller) 2.54 | (1) Chlamys varia (L.) 6:35 | (1) Chlamys nivea (Macgillivray) 6.35 IV (i) Chlamys distorta (da Costa)? 3.81 | (i) Chlamys opercularis (L.) 8.89 | (1) Chlamys septemradiata (Müller)* 5.08 | (1) Chlamys tigerina (Müller) 2.54 | (i) Chlamys furtiva (Loven) 1.90 | (i) Chlamys striata (Muller) 1.90 | (i) Chlamys vitrea (Gmelin) 1.90 IV (i) Similipecten similis (Laskey) 0.95 | (i) Anomiacea Anomiidae Anomia ephippium L. 6.35 Il (1) Monia patelliformis (L.)? 3.81 Ш (i) Monia squama (Gmelin)® 3.81 Ш (i) Heteranomia squamula (L.) 1.27 Ш (1) Limacea Limidae Lima hians (Gmelin)? 2.54 IV? (i)?? Lima loscombi Sowerby 1.90 IV (ii) Lima sulcata Brown 1.27 IV (ii) Lima subauriculata Montagu 0.63 IV (ii) Lima sarsi (Loven) 0.320 IV (ii) Ostreina Ostreacea Ostreidae Ostrea edulis L. 10.16 Ш (i) Crassostrea virginica (Gmelin) 17.78 Ш (1) Crassostrea angulata (Lamarck) 17.78 Ш (i) (PALAEOHETERODONTA) (Modiomorphoidat) (Unionoida freshwater) (Trigonioida) HETERODONTA Veneroida Lucinacea'® Lucinidae Loripes lucinalis (Lam.) 1.90 IV (111) Myrtea spinifera (Montagu) 2.54 IV (11) Lucinoma borealis (L.) 3.81 IV (iil) Divaricella divaricata (L.) 1227. IV (iii) Ungulinidae** Diplodonta rotundata (Montagu) 2.54 IV (111) Thyasiridae Thyasira flexuosa (Montagu) 0.95 IV (ili) Thyasira croulinensis (Jeffreys) 0:32™ IV (ill) Thyasira ferruginea Winckworth 0.32™ Il (iii) Thyasira subtrigona (Jeffreys) 0.16" IV ii)? (Chamacea) Leptonacea MARINE BIVALVE SHELL COLOURS 315 APPENDIX A (Continued) _оооее”з3з3с3С3ццццццццемммммммммШШ"ШШ"м"|"Щ"|"|"|"|"|"|"|—|—|—"—"—"_ь Бр Д_-_дЗнрнжмлммььъьЪьух лнмс"чмс",ьмнньн— Туре of Maximum colour and Probable measurement pattern degree of Species and classification (ст"2) variation exposure Erycinidae Lasaea rubra (Montagu)'? 0327 Il (iii) Kelliidae Kellia suborbicularis (Montagu) 0.95 IV (111) Leptonidae Lepton squamosum (Montagu) 1:27 IV (iii) Lepton nitidum Turton 0:327 IV (11) Neolepton sulcatulum (Jeffreys) 0.16” IV (ii) Neolepton sykesi (Chaster) 0.167 IV (ii) Epilepton clarkiae (Clark) 0.16" IV (ii) Montacutidae Montacuta substriata (Montagu) 0327 IV (iii) Montacuta ferruginosa'? (Montagu) 0.79 Il (iii) Mysella bidentata* (Montagu) 0327 IV (ii) Galeommatidae Galeomma turtoni Sowerby 127 IV (i)? Devonia perrieri (Malard) 0.48" IV (iii) (Chlamydoconchacea) Cyamiacea Turtoniidae Turtonia minuta (Fabricius)'* 0.327 Il (iii) (Carditacea) Crassatellacea Astartidae Astarte sulcata (da Costa) 2.54 Ш (ii) Astarte elliptica (Brown) 3.17 Ш (ii) Astarte montagui (Dillwyn) 1227 Ш (ii) Astarte triangularis (Montagu) 0327 Ш (ii) Astarte borealis (Schumacher) 4.44 I (ii) Cardiacea Cardiidae Acanthocardia aculeata (L.) 10.16 Ш (ii) Acanthocardia echinata (L.)'° 7.62 Ш (ii) Acanthocardia tuberculata (L.) 8.90 Il (ii) Parvicardium minimum (Philippi) 1227. IV (ii) Parvicardium papillosum (Poli) 1.27 Ш (ii) Parvicardium ovale (Sowerby) 1827. IV (ii) Parvicardium scabrum (Philippi) 1.27 IV (ii) Parvicardium exiguum (Gmelin)'® 1.27 Ш (ii) Cerastoderma edule (L.)'° 5.08 Ш (п) Cerastoderma glaucum (Lam.)'’ 6? Ш (ii) Laevicardium crassum (Gmelin)'° 7162. Ш (ii) (Tridacnacea) Mactracea Mactridae Mactra corallina (L.)'® 5.08 Il (ii) Mactra glauca Born 11.43 Il (ii) Spisula elliptica (Brown) Salz IV (ii) Spisula solida (L.) 4.44 IV (ii) Spisula subtruncata (da Costa) 2.54 IV (ii) Lutraria lutraria (L.) 12.70 Ш (111) Lutraria magna (da Costa) 12.70 Ш (iii) Lutraria angustior Philippi 10.16 Ш (111) Mesodesmatacea'? Mesodesmatidae Ervilia castanea (Montagu) 1.27 Il (ii) Solenacea 316 CAIN APPENDIX A (Continued) Type of Maximum colour and Probable measurement pattern degree of Species and classification (cm'?) variation exposure Solenidae?° Ensis ensis (L.) 12.70 I (iii) Ensis arcuatus (Jeffreys) 15.24 Ш (iii) Ensis siliqua (L.) 20.32 Ш (iil) Solen marginatus Montagu 1270 Ш (ili) Cultellidae Cultellus pellucidus (Pennant) 3.81 Ш (iii) Tellinacea Tellinidae Tellina squalida (Montagu) 4.44 | (ii) Tellina tenuis (da Costa) 1.90 | (ii) Tellina fabula (Gmelin) 1.90 Ш (ii) Tellina donacina L. 2.54 Il (11) Tellina pygmaea Lovén 0.95 | (ii) Tellina crassa Pennant 6.35 Il (ii) Tellina balaustina L. 1.90 Il (ii) Gastrana fragilis (L.) 4.44 IV (ii) Macoma balthica (L.) 2.54 | (ii) Donacidae Donax*' vittatus (da Costa) 3.81 | (ii) Donax variegatus (Gmelin) 3.81 Il (ii) Psammobiidae**? Gari fervensis (Gmelin)?8 5.08 12 (ii)? Gari depressa (Pennant) 6.35 | (1)? Gari tellinella (Lam)? 2.54 | (ii) Gari costulata (Turton) 2.54 | (ii) Scrobiculariidae Scrobicularia plana (de Costa) 6.35 IV (111) Semelidae Abra tenuis (Montagu) 127 IV (iii) Abra alba (Wood) 2.54 IV (Ш)? Abra nitida (Müller) 1127 IV (iil)? Abra longicallus (Scacchi) 1.90 IV (iil)? Abra prismatica (Montagu) 2.54 IV (iii)? Solecurtidae Solecurtus scopula (Turton) 6:35 Ш (111) Solecurtus chamasolen (da Costa) 6.35 Ш (iil) Pharus legumen (L.) 12.70 Ш (iii) (Dreissenacea, freshwater in Britain) (Gaimardiacea) Arcticacea Arcticidae Arctica islandica (L.) 12.70 Il (ii) Glossacea Glossidae Glossus humanus (L.) 10.16 Il (ii) (Corbiculacea, freshwater in Britain) Veneracea?* Veneridae Dosinia exoleta (L.) Sal Il (ii) Dosinia lupinus (L.) 3.81 Il (ii) Gafrarium minimum (Montagu) 1.59 | (ii) Callista chione (L.) 8.89 Il (ii) Venus verrucosa (L.) 6.35 Il (ii) Venus casina (L.)* 5.08 Ш (ii) Venus ovata Pennant 1.90 I (ii) Venus fasciata (da Costa) 2.54 | (ii) MARINE BIVALVE SHELL COLOURS 317 APPENDIX A (Continued) Type of Maximum colour and Probable measurement pattern degree of Species and classification (cm'?) variation exposure Venus striatula (da Costa) 4.44 Il (ii) Venus mercenaria L. 12.70 Ш (ii) Venerupis aurea (Gmelin) 4.44 Ш (ii) Venerupis rhomboides (Pennant) 6.35 Ш (ii) Venerupis pullastra (Montagu) 5.08 Il (ii) Venerupis saxatilis (Fleuriau) 3.81 IV (iii) Venerupis decussata (L.) 7.62 Ш (ii) Notirus irus (L.) 2.54 IV (iii) Petricolidae Petricola pholadiformis (Lam.) 6.35 IV (iii) Mysia undata (Pennant) 3.81 IV (ii) Myoida Myina Myacea Myidae Mya truncata L. 7.62 IV (iii) Mya агепапа L. 15.24 IV (iii) Sphenia binghami Turton?° 1527 IV (iii) Corbulidae Corbula gibba (Olivi)*?® 1.27 Ш (ii) Gastrochaenacea Gastrochaenidae Gastrochaena dubia (Pennant) 2.54 IV (iii) Hiatellacea Hiatellidae Hiatella arctica (L.) 3.81 IV (iii) Panomya arctica (Lam.) 7.62 IV (iii) Saxicava jeffreysi Winckworth 0.95 IV (iii) Pholadina Pholadacea Pholadidae Pholas dactylus L. 15.24 IV (111) Barnea candida (L.) 6:35 IV (iii) Barnea parva (Pennant) 3.81 IV (ili) Zirfaea crispata (L.) 8.89 IV (iii) Pholadidea loscombiana Turton 3.81 IV (iii) Martesia striata (L.) 5.08 IV (iii) Xylophaga praestans Smith 1.90 IV (iii) Xylophaga dorsalis Turton WAU IV (iii) Teredinidae Teredo navalis L. 0.95 IV (iii) Lyrodus pedicellatus (Quatrefages) 0.63 IV (iii) Nototeredo norvagicus (Spengler) 1.90 IV (111) Psiloteredo megotara (Forbes 8 Hanley) 127 IV (iii) Teredora malleolus (Turton) 1527 IV (111) Bankia fimbriatula Moll 8 Roch 0.63 IV (111) (Hippuritoida‘) ANOMALODESMATA Pholadomyoida (Pholadomyacea) Pandoracea Pandoridae Pandora?” albida (Roding) 3.81 Ш (ii) Pandora pinna (Montagu) 1.90 Ш (ii) Lyonsiidae Lyonsia norwegica (Gmelin)*® 3.81 IV (ii) Periplomatidae 318 CAIN APPENDIX A (Continued) Type of Maximum colour and Probable measurement pattern degree of Species and classification (cm'?) variation exposure Cochlodesma praetenue (Рийепеу)?9 3.81 IV (111) Thraciidae Thracia®° phaseolina (Lamarck) 3.81 IV (iii) Thracia villosiuscula (Macgillivray) 2.54 IV (iil) Thracia pubescens (Montagu) 8.89 IV (111) Thracia convexa (Wood) 6:35 IV (iii) Thracia distorta (Montagu) 2.54 IV (iil) Poromyacea Poromyidae Poromya granulata (Westerdorp) 1727, Ш (ii) Cuspidariidae Cuspidaria cuspidata (Olivi) 1.90 Ш (ii) Cuspidaria rostrata (Spengler) 2.54 IV (ii) Cuspidaria costellata (Deshayes) 0.95 IV (ii) Cuspidaria abbreviata (Forbes) 0.95 IV (ii) (Clavagellacea) ‘Extinct or unrepresented subclasses and orders, but not superfamilies, are shown for completeness. Only families represented in the British fauna are included. The order down to families is that of the Treatise on Invertebrate Paleontology, that of genera and species, and the binominal nomenclature (except as noted otherwise) is that of Tebble (1966). Introduced species, which are very few, have been included. Orders etc. in parentheses are extant but not represented; extinct ones are marked with a +; for others that are omitted, the reason is given. Allen (1954a, 1960). m indicates the minute species disregarded in estimating the effect of size on colour and pattern. “Allen (1960), deposition of manganese compounds on shell. “The genetics of shell colour in Mytilus edulis, and the adaptive significance of its clinal variation with climate (east coast of North America) have been investigated by Newkirk (1980) and Mitton (1977) respectively. From the numerous studies on M. edulis L. and M. galloprovincialis Lamarck cited and added to by Gosling (1984) it appears best to treat the latter as a subspecies of edulis. ®Swimming, Baird (1958), Yonge (1949). “Usually, since it becomes fixed like an oyster, considered as a separate genus, Hinnites, e.g. Abbott (1974). ®Seed & Roberts (1976) cast some doubt on the distinctness of M. patelliformis and M. squama. Material from Strangford Lough (seen by courtesy of Dr. Roberts) and material of M. squama identified by Winckworth (BM(NH)), seen by courtesy of Dr. P. Mordan) suggests, but does not prove, their distinctness. °Defensive adaptations, Gilmour (1963, 1967). ‘For discussions of preferred habitats (and habits) of Lucinacea, British and foreign, see Allen (1958a), Jackson (1973) and Kauffmann (1969). "'Diplodontidae in Tebble (1966). '2Morton, J. E. (1954). '3Morton, J. E. (1962). '4Said to be a neotenous уепегасеап by Ockelman (1964). 'SLeaping, Ansell (1967a). ‘Habits of P. exiguum, Russell & Petersen (1973). "As С. lamarcki (Reeve) in Tebble (1966). Russell (1971) corrects the distribution in Tebble of edule and glaucum and (1972) reports character displacement in rib number between these two species. ‘8Leaping, Ansell (1969). #Yonge 4 Allen (1985). 20| eaping, Ansell (1968). "Biology of the genus, Ansell (1983). *2Gariidae in Tebble (1966). 23| еартд, Ansell (1967c). 24 Habits, Ansell (1961). 25Habits, Yonge (1951). 26Habits, Yonge (1946). Habits, Allen (1954b), Allen & Allen (1955). 28Habits, Ansell (1967b). 2#Habits, Allen (1958b). In Laternulidae, Tebble (1966). Habits, Yonge (1937). Revised Ms. accepted 15 July 1987 INDEX TO TAXA IN VOLUME 28 An asterisk (*) denotes a new taxon. abbreviata, Cuspidaria, 318 Abies, 148 Abra, 316 abyssorum, Boreotrophon, 77-79 Acanthinula, 156 Acanthocardia, 315 Acer, 135 Achatina, 119, 128 Acicula, 156 acicula, Ceciliodes, 156 Aclididae, 77, 78 Aclis, 67, 77, 78 aculeata, Acanthinula, 156 acuta, Cochlicella, 11 acuta, Physa, 17, 25, 26 Adalaria, 100 Adeorbis, 79 Admete, 77 adriaticus, Modiolus, 313 Adula, 313 Aegopinella, 156 Aeolidia, 62 Agriolimax, 119, 126, 128, 129 alabamensis, Triodopsis, 163-165, 178, 184, 196, 198, 201, 203, 204, 208, 212, 213, 225, 230, 241, 242, 246, 247, 270, 272, 273 alaskana, Vitrina, 156 alba, Abra, 316 Albida, 156 albida, Pandora, 317 albolabris, Helix, 260 albolabris, Neohelix, 159-161, 163-170, 193, 194, 199, 200, 203-209, 211-225, 227-229, 233- 238, 240-244, 246, 248, 249, 255-259, 261, 263 albolabris, Triodopsis, 260 albolabris, Xolotrema, 257 Alia, 72 alleni, Neohelix, 159, 163-167, 169, 170, 171, 193, 194, 196, 199, 200, 203-205, 207-209, 211-225, 227, 228, 233-235, 238, 240-244, 246-249, 255, 257-261, 263 alleni, Triodopsis, 257 alleni, Xolotrema, 257 alliarius, Oxychilus, 157 Allogona, 150, 151, 164, 166, 199, 202, 206, 208-210, 213-215, 233, 261, 262 alonensis, Iberus, 105 Ammonitella, 164 Ammonitellidae, 164, 199-201 Amplirhagada, 190 Anachis, 77 Ancylus, 43 andrusiana, Vertigo, 156 anglicus, Cucumis, 30, 42 angulata, Crassostrea, 314 angustior, Lutraria, 315 Anomalodesmata, 317 Anomia, 314 Anomiacea, 314 Anomiidae, 314 anteridon, Triodopsis, 163, 165, 178, 182, 196, 198, 201, 203, 204, 208, 209, 211-214, 225, 227, 228, 241-243, 246, 249, 269 antillensis, Limopsis, 295 antonia, Benthomangelia, 78, 79 Aplacophora, 95, 101, 102 Aplysia, 36, 37 Aporrhais, 77 appressa, Lymnaea, 119, 129 arboreus, Zonitoides, 148, 156 arbustorum, Arianta, 151, 156 Arca, 292, 313 Arcacea, 313 Archaeogastropoda, 73 Arcidae, 313 Arcoida, 313 Arctica, 295, 316 artica, Hiatella, 290, 317 Arcticacea, 316 arctica, Panomya, 317 Arcticidae, 316 Arctium, 30 arcuatus, Ensis, 316 arenaria, Mya, 317 Arianta, 151, 156 Ariolimax, 43 Arion, 43, 48, 119 Ashmunella, 164, 190, 202, 248 Ashmunellinae, 164, 168, 196, 197, 215 asiatica, Hanleyella, 101 aspersa, Helix, 10, 12, 43, 151, 156 Astarte, 291, 292, 315 Asteroidae, 73 ater, Arion, 43, 119 Athoracophoridae, 119 aurea, Venerupis, 317 aurita, Limopsis, 313 Azeca, 156 balaustina, Tellina, 316 balthica, Macoma, 289-318 bairdii, Gymnobela, 78 Bankia, 317 barbatus, Modiolus, 313 Barnea, 317 Bathysciadium, 77 Belomitra, 79 Benthobia, 79 Benthomangelia, 78, 79 Benthonella, 77-79 bergensis, Oenopota, 67 bidentata, Clausilia, 156 bidentalis, Hemicycla, 105-117 bidentatus, Melampus, 283, 284 bidentata, Mysella, 292, 315 (319) 320 INDEX binghami, Sphenia, 317 binneyana, Nesovitrea, 156 Biomphalaria, 37, 53, 129 biplicata, Laciniaria, 157 Bithynia, 81, 87, 91 Bivalvia, 313 *bogani, Neohelix, 159, 164, 216, 219, 223, 233, 234, 237, "257,258 borealis, Astarte, 315 borealis, Lucinoma, 314 Boreotrophon, 77, 79 Brachyura, 72 Bradybaena, 151, 156 Bradybaenidae, 150 brandaris, Murex, 128 brandtii, Schizoplax, 101 brevis, Gymnobela, 78 Brookula, 67, 79 brychia, Frigidoalvania, 67, 73, 77 Buccinidae, 77, 78 budapestensis, Milax, 43 Bulgarica, 156 burchi, Triodopsis, 163, 165, 178, 179, 195, 197, 200-206, 208, 211-213, 215, 224, 225, 241-243, 246, 249, 267-269 Calappa, 73 Calliotropis, 77 Callista, 316 Camaenacea, 200 Camaenidae, 164, 189, 196, 197, 199-201 cana, Bulgarica, 156 Cancellariidae, 78 candida, Barnea, 317 cantiana, Monarcha, 2, 8, 157 capensis, Brookula, 67 Carcinus, 71, 72 Cardiacea, 315 Cardiidae, 315 carinata, Alia, 72 caroliniensis, Xolotrema, 163-165, 175, 188, 194, 200, 203, 204, 207, 208, 211-213, 220, 225, 241, 242, 248, 249, 265 carota, Daucus, 30, 42 Carychium, 156 casina, Venus, 316 castanea, Ervilia, 315 Caudofoveata, 101, 102 caverna, Lepidochitona, 95, 96, 99, 100 Ceciliodes, 156 cellarius, Oxychilus, 119, 126, 128, 157 Cepaea, 1-12, 42, 43, 48, 151, 156, 291, 304, 305 Cephalopoda, 101 Cerastoderma, 290, 303, 315 Cerion, 105, 161, 247, 248 Cerithiella, 77, 78 chadwicki, Webbhelix, 249, 262 Chaetodermatida, 102 Chaetodermomorpha, 102 Chama, 295 Chamacea, 295, 314 chamasolen, Solecurtus, 316 chariessa, Theta, 78 Chimaeriformidae, 71 chione, Callista, 316 Chlamydoconchacea, 315 Chlamys, 314 cinerea, Lepidochitona, 96, 97 cinereoniger, Limax, 131 Cionella, 156 Cipangopaludina, 81, 87, 91 claibornensis, Triodopsis, 163-165, 177, 179, 192, 195, 196, 201-204, 207, 208, 211-213, 215, 222, 225, 241-243, 246, 249, 266 clappi, Planogyra, 151, 156 clarkiae, Epilepton, 292, 315 Clausiliidae, 152 Clausilia, 156 Clavagellacea, 318 Clithon, 12 Clupeoidae, 71 Cocculinidae, 77, 79 Cochlicella, 11 Cochlicopa, 156 Cochlodesma, 318 Cochlodina, 156 collarifera, Hemicycla, 106, 108, 112, 117 columbiana, Vertigo, 156 columbiana, Vespericola, 151, 156 columbianus, Ariolimax, 43 Columella, 156 Colus, 77, 81, 91, 93 complanata, Triodopsis, 163-165, 179, 195-197, 200-204, 208, 211-213, 215, 224, 225, 228, 241-243, 246, 247, 249, 267 Conchifera, 101 Conomurex, 81 conspectum, Punctum, 156 contabulata, Admete, 77 contectoides, Viviparus, 91 contortus, Planorbis, 43 contracta, Vitrea, 157 convexa, Thracia, 317 cookei, Microphysula, 156 copei, Triodopsis, 168, 242 corallina, Mactra, 291, 303, 315 Corbicula, 2 Corbiculacea, 316 Corbula, 317 Corbulidae, 317 Corillidae, 164, 200, 201 cornuarietis, Marisa, 129 costellata, Cuspidaria, 318 costulata, Gari, 316 costellatum, Bathysciadium, 77 costulatus, Musculus, 313 cragini, Triodopsis, 163-165, 168, 179, 181, 196, 198, 200-205, 208-210, 213-215, 225, 226, 240-243, 246, 247, 249, 268-270 Crangonidae, 72 crassa, Tellina, 316 Crassostrea, 314 crassum, Laevicardium, 315 Crenella, 292, 313 Crepidula, 283, 284 crispata, Zirfaea, 317 INDEX 321 cronkhitei, Discus, 156 croulinensis, Thyasıra, 292, 314 Crustacea, 65 Cryptodonta, 313 Cryptomastix, 164, 167, 168, 190, 199, 202, 261 crystallina, Vitrea, 157 Cucumis, 30, 42 Cultellidae, 316 Cultellus, 316 curta, Gymnobela, 79 Cuspidaria, 318 Cuspidariidae, 318 cuspidata, Cuspidaria, 318 Cyamiacea, 315 Cyclostrema, 78 Cyclostrematidae, 77 Cylichna, 66 dactylus, Pholas, 317 Daucus, 30, 42 Decapoda, 65, 72, 73 decussata, Crenella, 292, 313 decussata, Venerupis, 317 denotata, Xolotrema, 163, 165, 167, 175, 188, 190, 194, 200, 203-205, 207-209, 211-215, 220, 225, 228, 240-244, 246-249, 265, 268 dentifera, Neohelix, 163, 165, 167, 170, 173, 193, 196, 197, 199, 200, 202-205, 207-209, 213- 222, 225, 228, 235, 237, 241-247, 259, 263 depressa, Gari, 316 Deroceras, 17-19, 23, 24, 29-37, 42, 43, 48, 50, 131 Devonia, 292, 315 diaphana, Vitrea, 157 Diodon, 73 Diplodonta, 314 discoidea, Triodopsis, 163-165, 180, 182, 196, 198, 201, 203, 204, 213, 225, 228, 237, 241, 242, 246, 247, 249, 271 discors, Musculus, 313 Discus, 156 distorta, Chlamys, 314 distorta, Thracia, 317 divaricata, Divaricella, 314 Divaricella, 314 divesta, Neohelix, 163, 165, 168, 173, 175, 193, 196, 197, 199, 200, 203, 204, 207-209, 213, 214, 218, 220-222, 225, 237, 241-243, 246, 249, 263, 264 doliolum, Orcula, 157 dominula, Panaxia, 5 Donacidae, 295, 316 donacina, Tellina, 316 Donax, 303, 316 dorsalis, Xylophaga, 317 Dosinia, 316 Dreissenacea, 316 Drilliola, 79 dubia, Clausilia, 156 dubia, Gastrochaena, 317 echinata, Acanthocardia, 315 Echinodermata, 65, 73 edentula, Columella, 156 edule, Cerastoderma, 290, 303, 315 edulis, Mytilus, 303, 313 edulis, Ostrea, 314 elegans, Pomatias, 157 elliptica, Astarte, 315 elliptica, Spisula, 315 elodes, Stagnicola, 37, 61-63 emarginata, Nucella, 72, 283, 285 emarginata, Thais, 8, 9 Ena, 156 Enidae, 152 Ensis, 316 ensis, Ensis, 316 ephippium, Anomia, 314 Epilepton, 292, 315 Epitonium, 78, 79, 81 Eremina, 105 erinacea, Ocenebra, 81, 87, 91 Ervilia, 315 Erycinidae, 315 Euconulus, 148, 156 Eulimella, 77 Euomphalia, 156 Euphorbia, 106 exiguum, Parvicardium, 315 exoleta, Dosinia, 316 fabula, Tellina, 316 fallax, Triodopsis, 160, 163-165, 168, 180, 185, 196, 198, 201, 203, 204, 212, 225, 226, 229, 231, 240-243, 246-249, 269-273 fasciata, Venus, 316 fasciatus, Liguus, 13 fernaldi, Lepidochitona, 95, 96, 98-100 ferruginea, Thyasira, 292, 314 ferruginosa, Montacuta, 295, 315 fervensis, Gari, 294, 303, 316 fidelis, Monadenia, 7, 150, 151, 156 fimbriatula, Bankia, 317 flavus, Limax, 36, 131-146 flexuosa, Thyasira, 314 fluviatilis, Ancylus, 43 fornicata, Crepidula, 283, 284 fosteri, Xolotrema, 163, 165, 168, 176, 188, 190, 194, 200-207, 210-212, 214, 215, 221, 225, 228, 229, 240-243, 246, 247, 249, 264 fragilis, Gastrana, 316 fragilis, Pinna, 313 fragilissima, Macoma, 296 fraudulenta, Triodopsis, 163-165, 178, 180, 196— 205, 207, 208, 211-215, 223-225, 236, 241- 243, 246, 247, 249, 266, 268 frielei, Gymnobela, 78 Frigidoalvania, 67, 73, 77 fruticum, Bradybaena, 151, 156 fulciden, Triodopsis, 163-165, 169, 180, 186, 196-205, 210-215, 224, 225, 232, 241-243, 249, 268 fulica, Achatina, 119, 128 fulvus, Euconulus, 148, 156 funebralis, Tegula, 72 furtiva, Chlamys, 314 322 INDEX fusca, Acicula, 156 fuscolabris, Neohelix, 169, 171, 216, 223, 233, 234, 237, 238, 240, 242, 247, 260 Gadidae, 71 Gaimardiacea, 316 Galeomma, 295, 315 Galeommatidae, 315 Gari, 294, 303, 316 Gastrana, 316 Gastrochaena, 317 Gastrochaenacea, 317 Gastrochaenidae, 317 germana, Triodopsis, 151, 156 Geryon, 72, 73 gibba, Corbula, 317 glabrata, Biomphalaria, 37, 53, 93, 129 glans, Cerion, 248 glauca, Mactra, 315 glaucum, Cerastoderma, 315 globosus, Biomphalaria, 62, 63 Glossacea, 316 Glossidae, 316 Glossus, 295, 316 Glycymeridae, 313 Glycymeris, 291, 303, 313 glycymeris, Glycymeris, 291, 303, 313 goodalli, Azeca, 156 granulata, Poromya, 318 graphica, Oenopota, 77 gualtierianus, Iberus, 105, 117 gubernatorium, Cerion, 248 Gymnobela, 77, 79 gyrina, Physella, 37 haliaeeti, Anachis, 77 Halosauridae, 71 hamonis, Nesovitrea, 157 Hanleyella, 101 hanleyi, Nucula, 313 Haplotrema, 151, 153 Haroldorbis, 168, 225, 242 harpa, Pusillina, 77 harpularia, Lora, 78 Helicarionidae, 119, 128 Helicidae, 13, 47, 62, 105-117, 150, 152 Helicigona, 151, 156 Helicodonta, 156 Helix, 5, 10, 43, 62, 63, 119, 128, 151, 156, 167, 260 helveticus, Oxychilus, 157 Hemicycla, 105-117 henriettae, Triodopsis, 163-165, 168, 181, 182, 196, 198, 200-204, 210-215, 225, 226, 241, 242, 247, 248, 268 Heteranomia, 313 Heterodonta, 295, 314 hians, Lima, 292, 314 Hiatella, 290, 317 Hiatellacea, 317 Hiatellidae, 317 Hippuritoida, 317 hirundo, Pteria, 314 hispida, Trichia, 151, 157 hopetonensis, Triodopsis, 163, 165, 182, 183, 196, 198, 201, 203, 204, 206, 208, 210-214, 225, 229, 237, 241, 242, 248, 270, 272, 273 horridum, Oplopanax, 43 hortensis, Arion, 48 hortensis, Cepaea, 4, 5, 12, 151, 156 hubrichti, Neohelix, 228, 237, 239, 240, 249 humanus, Glossus, 295, 316 hupensis, Oncomelania, 81-94 Hymenoptera, 1 Iberus, 105, 117 incarnata, Perforatella, 151, 157 indianorum, Mesodon, 242, 264 ioxanus, lberus, 105 Iphigena, 157 irus, Notirus, 317 islandica, Arctica, 295, 316 Isognomostoma, 151 isognomostoma, Isognomostoma, 151, 157 japonicum, Schistosoma, 81 jeanae, Siphonaria, 283, 284 jeffreysi, Saxicava, 317 johnsoni, Pristiloma, 156 Juvenichiton, 97 juxtidens, Triodopsis, 163, 165, 167, 182, 183, 196, 198, 201, 203-205, 208, 210-215, 225, 227-229, 237, 240-244, 246, 247, 249, 270, 271 Kellia, 315 Kelliidae, 315 laeve, Deroceras, 43, 48 Laevicardium, 315 Lambis, 81, 91 lambis, Lambis, 81, 91 lamellosa, Nucella, 275-287 laminata, Cochlodina, 156 lansingi, Pristiloma, 156 lapicida, Helicigona, 151, 156 lapillus, Nucella, 8, 81, 87, 91 lapillus, Thais, 6, 8, 9 Lasaea, 291, 292, 315 legumen, Pharus, 316 Lehmannia, 17-27, 43 lenticula, Ashmunella, 248 Lepetella, 77 Lepidochitona, 95-100 Lepton, 292, 315 Leptonacea, 314 Leptonidae, 314, 315 Leucosyrinx, 78, 79 Levantina, 105 lignosa, Mopalia, 95-100 Liguus, 2, 8-10, 13 Lima, 290, 292, 314 Limacea, 314 Limacidae, 131-146 Limax, 17-19, 23, 24, 29, 32, 35-37, 42, 43, 48, 131-146 Limidae, 314 INDEX 323 Limopsacea, 295, 313 Limopsidae, 295, 313 Limopsis, 295, 313 lioderma, Mesodon, 242, 264 lioderma, Neohelix, 163, 165, 173, 175, 193, 196, 197, 199, 200, 203, 204, 207, 208, 211-213, 217, 218, 225, 228, 241-243, 247, 249, 264 Lissospira, 77, 79 littorea, Littorina, 71, 72 Littorina, 6, 7, 10, 67, 71, 72, 81, 87, 91, 283, 284, 305, 309 Littorinidae, 71 longicallus, Abra, 316 Lora, 78 Loripes, 314 loscombi, Lima, 314 loscombiana, Pholadidea, 317 lottae, Pleurotomella, 79 lubrica, Cionella, 156 lubrica, Cochlicopa, 156 lubricella, Cochlicopa, 156 lucida, Yoldiella, 292, 313 Lucinacea, 314 lucinalis, Loripes, 314 Lucinidae, 314 Lucinoma, 314 luhuanus, Сопотигех, 81 lupinus, Dosinia, 316 Lutraria, 315 lutraria, Lutraria, 315 Lymnaea, 25, 53-64, 119, 127, 129 Lymnaeidae, 53, 62 Lyonsia, 317 Lyonsiidae, 317 Lyrodus, 317 lyronuclea, Theta, 78, 79 macerophylla, Chama, 295 Macoma, 289-318 Macrogastra, 157 Macrouridae, 71, 73 Mactra, 291, 303, 315 Mactracea, 315 Mactridae, 315 Macularia, 105 maenas, Carcinus, 71, 72 magna, Lutraria, 315 major, Neohelix, 163-165, 172, 176, 193, 194, 199, 200, 203-205, 207, 208, 210-222, 225, 228, 233-235, 237, 239-243, 246-249, 256, 258, 259, 261 major, Triodopsis, 260 malleata, Cipangopaludina, 87 malleolus, Teredora, 317 marginata, Lehmannia, 43 marginatus, Limax, 131-146 marginatus, Solen, 316 mariae, Littorina, 6, 72 Marisa, 129 maritima, Helix, 260 marmoratus, Musculus, 313 Martesia, 317 maximus, Limax, 17-19, 23, 24, 29, 32, 35-37, 42, 43, 48, 131-146 maximus, Pecten, 314 megotara, Psiloteredo, 317 Melampus, 283, 284 mercenaria, Venus, 317 Mesodesmatacea, 290, 315 Mesodesmatidae, 315 Mesodon, 159, 160, 166-168, 206, 211, 212, 214, 215, 242, 247, 250, 255, 258, 262 messana, Triodopsis, 163, 165, 183, 184, 195, 196, 198, 201, 203, 204, 206, 208, 211-213, 225, 230, 237, 241, 242, 248, 269, 272, 273 Microphysula, 151, 156 Milax, 43 minimum, Carychium, 156 minimum, Parvicardium, 315 minus, Arctium, 30 minuta, Nuculana, 313 minuta, Turtonia, 291, 292, 315 Mitrella, 67, 77 Modiolus, 313 modiolus, Modiolus, 313 Modiomorphoida, 314 Monacha, 2, 8, 157 Monadenia, 7, 150, 151, 156 Monia, 314 Monoplacophora, 101 Montacuta, 292, 315 Montacutidae, 315 montagui, Astarte, 315 montana, Ena, 156 Mopalia, 95-100 morchi, Taranus, 77 multilineata, Neohelix, 242, 246, 262 multilineata, Webbhelix, 163, 165, 174, 188, 190— 196, 199, 202-205, 208, 210-215, 217-221, 225, 229, 237, 241, 242, 262 Murella, 105 Murex, 128 Mus, 283 muscosa, Mopalia, 95-99 Musculus, 303 musculus, Mus, 283 Mya, 292, 317 Myacea, 317 Myidae, 317 Myina, 317 Myoida, 317 Myrtea, 314 Mysella, 292, 315 Mysia, 317 Mytilacea, 313 Mytilidae, 313 Mytiloida, 313 Mytilus, 303, 313 Nassarius, 128, 129, 283 Natica, 77, 306 Nautilus, 101 navalis, Teredo, 317 neglecta, Triodopsis, 163-165, 184, 186, 196, 198, 324 INDEX 201, 203, 204, 208, 211-213, 225, 227, 232, 236, 241, 242, 246, 271 nemoralis, Cepaea, 4, 5, 12, 36, 42, 43, 48, 151 nemoralis, Helix, 5 Neohelix, 159-161, 163-173, 175-177, 190-194, 196, 197, 199-203, 205-219, 225, 227-229, 233-238, 240, 242-245, 247-250, 255-264 Neolepton, 292, 315 Neomeniomorpha, 95, 102 Nerita, 91 Nesovitrea, 156, 157 niger, Musculus, 313 nigrolineata, Littorina, 72 nitens, Aegopinella, 156 nitida, Abra, 316 nitidula, Aegopinella, 156 nitidum, Epitonium, 78 nitidum, Lepton, 292, 315 nivea, Chlamys, 314 normalis, Mesodon, 247, 258 norvagicus, Nototeredo, 317 norwegica, Lyonsia, 317 Nostoc, 82 notata, Xolotrema, 242 Notirus, 317 Nototeredo, 317 Nucella, 8, 72, 81, 87, 91, 275-287 nucleus, Nucula, 313 Nucula, 292, 295, 313 Nuculacea, 313 Nuculanacea, 313 Nuculanidae, 313 Nuculidae, 313 Nuculoida, 313 obscura, Ena, 156 obscura, Solariella, 77 obsoleta, Triodopsis, 163-165, 183, 184, 196, 198, 201, 203, 204, 212, 225, 229, 241, 242, 246, 248, 270, 272, 273 obsoletus, Nassarius, 128, 129, 283 obstricta, Xolotrema, 163-165, 167, 175, 188, 194, 200, 203, 204, 207, 209-214, 220, 225, 228, 241, 242, 244, 246, 247, 249, 265 obtusata, Littorina, 72 obvoluta, Helicodonta, 156 occidentale, Carychium, 156 occidentalis, Aporrhais, 77 occidentalis, Mesodon, 242 occidentalis, Patera, 242 occidentalis, Xolotrema, 163-165, 176, 189, 194, 200-204, 207, 209-214, 221, 225, 228, 241- 243, 247, 249, 264 Ocenebra, 81, 87, 91 Oenopota, 67, 77 Omalogyra, 78 Oncomelania, 81-94 Onoba, 77 opercularis, Chlamys, 314 Ophiuroidea, 73 Opisthobranchia, 62, 66, 100 Oplopanax, 43 Orcula, 157 Oreohelicidae, 164, 189, 196, 197, 200, 201 Oreohelix, 164, 190 Ostrea, 314 Ostreacea, 314 Ostreidae, 314 Ostreina, 314 oualaniensis, Clithon, 12 ovale, Parvicardium, 315 ovalis, Oenopota, 77 ovata, Venus, 316 Oxychilus, 119, 126, 128, 157 packardi, Pleurotomella, 77, 78 Pagurus, 73 Palaeoheterodonta, 314 Palaeotaxodonta, 313 palustris, Triodopsis, 163, 165, 183, 185, 196, 198, 201, 203, 204, 209, 210, 212-214, 225, 229, 241, 242, 246, 247, 269-271, 273 Рапама, 5 Pandalidae, 72 Pandora, 317 Pandoracea, 317 Pandoridae, 317 Panomya, 317 papillosa, Aeolidia, 62 papillosum, Parvicardium, 315 Partula, 2, 9, 13, 161, 248, 283 Partulida, 87 parva, Barnea, 317 Parvicardium, 315 parvula, Clausilia, 156 patelliformis, Monia, 314 Patera, 242 Pecten, 314 Pectinacea, 314 Pectinidae, 295, 313 pectunculoides, Arca, 292, 313 pedicellatus, Lyrodus, 317 pelagica, Onoba, 77 pellucida, Vitrina, 157 pellucidus, Cultellus, 316 Penaeidae, 72 pendula, Triodopsis, 163-165, 185, 186, 196, 198, 201, 203, 204, 209, 211-213, 225, 232, 241, 242, 246, 271 peregra, Lymnaea, 62 Perforatella, 151, 157 Periplomatidae, 317 perrieri, Devonia, 292, 315 Petricola, 317 Petricolidae, 317 pfeifferi, Biomphalaria, 37 Pharus, 316 phaseolina, Thracia, 317 phaseolinus, Modiolus, 313 Phaseolus, 292, 313 Philobryidae, 295 Pholadacea, 317 Pholadidae, 317 Pholadidea, 317 pholadiformis, Petricola, 317 Pholadina, 317 INDEX 325 Pholadomyacea, 317 Pholadomyoida, 317 Pholas, 317 Physa, 17, 25, 26 Physella, 37 Picea, 148 picea, Triodopsis, 163-165, 167, 177, 186, 192, 195-197, 201-207, 209, 211-213, 215, 222, 224, 225, 241-243, 244, 246, 249, 266 Pinna, 313 pinna, Pandora, 317 Pinnacea, 313 Pinnidae, 313 pisana, Theba, 2, 4, 6, 10, 119-130 plana, Scrobicularia, 316 Planogyra, 151, 156 Planorbis, 41, 43 platysayoides, Triodopsis, 163, 165, 168, 180, 186, 192, 195, 196, 198, 200-205, 207, 209, 211, 213, 215, 225, 228, 240-243, 245, 247, 249, 266, 267 plebeia, Trichia, 151, 157 Pleurococcus, 133 Pleuromeris, 295 Pleurotomella, 77, 79 pliculata, Iphigena, 157 polita, Acicula, 156 Polycera, 100 Polygyra, 167 Polygyracea, 190 Polygyrella, 164, 190 Polygyridae, 159-273 Polygyroidea, 164 Polyplacophora, 95-103 pomatia, Helix, 62, 63, 119, 128, 151, 156 Pomatias, 157 Poromya, 318 Poromyacea, 318 Poromyidae, 318 praestans, Xylophaga, 317 praetenue, Cochlodesma, 318 prideauxi, Crenella, 292, 313 prismatica, Abra, 316 Pristiloma, 151, 156 profunda, Allogona, 164, 166, 206, 208, 209, 213-215, 233, 262 Prosobranchia, 65-79, 81, 86, 87, 91, 93, 119, 275-287 proxima, Ashmunella, 248 Psammobiidae, 295, 316 pseudoareolata, Pusillina, 77 pseudoflavus, Limax, 131-146 pseudoplatanus, Acer, 135 Psiloteredo, 317 Pteria, 314 Pteriacea, 313 Pteriidae, 314 Pteriina, 313 Pterioida, 313 Pteriomorpha, 295 pubescens, Thracia, 317 pugetensis, Striatura, 151, 156 pulchella, Vallonia, 148 pullastra, Venerupis, 317 Pulmonata, 17-27, 62, 95, 100, 105, 119-130, 168 Punctum, 151, 156, 157 pura, Aegopinella, 156 pura, Mitrella, 67, 77 pusilla, Vertigo, 157 Pusillina, 77 pusillus, Phaseolus, 292, 313 putris, Succinea, 119, 126 pygmaea, Tellina, 316 pygmaeum, Punctum, 157 pygmaeus, Colus, 77 Pyramidellidae, 87 quadrasi, Oncomelania, 81-94 quinquidens, Geryon, 72 randolphi, Punctum, 156 Ranunculus, 43 repens, Ranunculus, 43 reticulatum, Deroceras, 17-19, 23, 24, 29-37, 42, 43, 48, 50, 131 reticulatus, Agriolimax, 119, 126, 128, 129 Retusa, 66 rhomboides, Venerupis, 317 Rissoidae, 78 rolphii, Macrogastra, 157 rositai, Iberus, 105 Rossmaessleria, 105 rostrata, Cuspidaria, 318 rotundata, Diplodonta, 314 rotundatus, Discus, 156 rowelli, Vertigo, 156 rubra, Lasaea, 291, 292, 315 rudis, Littorina, 72 rugosa, Triodopsis, 163-165, 169, 187, 198, 199, 203, 204, 207, 211, 214, 215, 224, 225, 228, 241-243, 246, 249, 267-270 sandersoni, Pleurotomella, 78, 79 sargentianus, Mesodon, 242 sargentianus, Patera, 242 sarsi, Lima, 292, 314 sativus, Cucumis, 30, 42 saxatilis, Littorina, 6, 283, 284 saxatilis, Venerupis, 317 Saxicava, 317 scabrum, Parvicardium, 315 Scaphander, 66 Schistosoma, 81 Schizoplax, 99, 101 scopula, Solecurtus, 316 Scrobicularia, 316 Scrobiculariidae, 316 secale, Abida, 156 Semelidae, 316 Semilimax, 157 semilimax, Semilimax, 157 senegalensis, Nerita, 91 septemradiata, Chlamys, 314 Serpuloides, 17 Shelfordorbis, 168, 225, 242 sigmoidae, Xanthodaphne, 79 siliqua, Ensis, 316 326 INDEX Similipecten, 314 similis, Similipecten, 314 simpsoni, Adula, 313 Simrothiella, 95 Siphonaria, 283, 284 sitkana, Littorina, 81, 87, 91 smithi, Cyclostrema, 78 soelneri, Triodopsis, 163, 165, 185, 187, 198, 201, 203, 204, 213, 225, 228, 229, 231, 237, 241-243, 247, 248, 270, 272, 273 Solanum, 31, 42 Solariella, 77, 78 Solecurtidae, 316 Solecurtus, 316 *solemi, Neohelix, 159, 163-165, 177, 193, 194, 199-205, 207, 209, 211-213, 215-223, 225, 233, 235, 239, 240, 241, 243, 246, 248, 249, 259, *260, 261 Solen, 316 Solenacea, 315 Solenidae, 315 Solenogastres, 95, 101, 102 solida, Spisula, 315 solidissima, Spisula, 294 Sphenia, 317 spinifera, Myrtea, 314 spiralis, Partulida, 87 Spisula, 294, 315 sportella, Haplotrema, 151, 156 squalida, Tellina, 316 squama, Monia, 314 squamosum, Lepton, 315 squamula, Heteranomia, 314 stagnalis, Lymnaea, 25, 53-64, 119, 127, 129 Stagnicola, 37, 61-63 stearnsi, Pristiloma, 156 stimpsoni, Colus, 81, 87, 91, 93 striata, Chlamys, 314 striata, Martesia, 317 striatula, Venus, 303, 317 Striatura, 151, 156 strigella, Euomphalia, 156 striolata, Trichia, 151, 157 Stylommatophora, 119 subauriculata, Lima, 314 subcylindrica, Truncatella, 87 suborbicularis, Kellia, 315 subrufescens, Zenobiella, 157 substriata, Montacuta, 292 substriata, Vertigo, 157 subtrigona, Thyasira, 292, 314 subtruncata, Spisula, 315 Succinea, 119, 126 Succineidae, 119 sulcata, Astarte, 315 sulcata, Chlamys, 314 sulcata, Lima, 314 sulcata, Nucula, 313 sulcatulum, Neolepton, 292, 315 supersonatum, “Triodopsis”, 160 sykesi, Neolepton, 292, 315 Tacita, 78, 79 taeniata, Partula, 9, 283 Taranus, 77 Tegula, 72 Tellina, 309, 316 Tellinacea, 295, 316 tellinella, Gari, 294, 316 Tellinidae, 295, 316 tenella, Benthonella, 78, 79 tennesseensis, Triodopsis, 163, 165, 167, 179, 187, 195-197, 200-206, 209-215, 224, 225, 228, 240, 241-247, 249, 267-269 tentaculata, Bithynia, 81, 87, 91 tenuis, Abra, 316 tenuis, Aclis, 77 tenuis, Nucula, 313 tenuis, Tellina, 309, 316 Terebridae, 71, 72 Teredinidae, 317 Teredo, 317 Teredora, 317 Testaria, 101 tetragona, Arca, 313 Thaididae, 71, 275-287 Thais, 6, 8, 9, 10, 13 Tharsiella, 79 Theba, 2, 4, 6, 10 Theta, 78, 79 Thracia, 317 Thraciidae, 318 Thyasira, 292, 314 Thyasiridae, 314 tigerina, Chlamys, 314 tincta, Gymnobela, 78 tinctum, Epitonium, 81 tomentosa, Lymnaea, 62 tomlini, Yoldiella, 292, 313 Torresitrachea, 190 townsendiana, Allogona, 150, 151, 156 traversensis, Neohelix, 242, 249, 261 traversensis, Triodopsis, 260 triangularis, Astarte, 291, 292, 315 Trichia, 151, 157 Tridacnacea, 315 tridentata, Pleuromeris, 295 tridentata, Triodopsis, 163-165, 167-169, 182, 187, 195, 196, 198, 201, 203-205, 209-215, 225, 227-229, 237, 240-244, 246, 247, 249, 268-270 tridentatum, Carychium, 156 Trigonioida, 314 Triodopsinae, 159-273 Triodopsis, 151, 156, 159, 160, 162-165, 167-169, 177-186, 187, 190-192, 194-203, 205-210, 212-215, 217, 222-232, 236, 237, 240, 242, 243, 245, 247-250, 257, 260, 263-273 truncata, Mya, 317 Truncatella, 87 truncatula, Lymnaea, 62 tryoni, Benthobia, 79 tuberculata, Acanthocardia, 315 tuberosum, Solanum, 31, 42 tubicola, Lepetella, 77 turgida, Nucula, 313 INDEX 327 Turridae, 77, 79 Turtonia, 291, 292, 315 turtoni, Galeomma, 295, 315 Turtoniidae, 315 Typhlomangelia, 79 Tyrrheniberus, 105 umbilicatus, Adeorbis, 79 undata, Mysia, 317 Ungulinidae, 314 unifasciata, Eulimella, 77 Unionoida, 314 valentiana, Lehmannia, 17-27 Vallonia, 148 vancouverense, Haplotrema, 151, 156 vannostrandi, Triodopsis, 163, 165, 184, 187, 196, 198, 201, 203, 204, 209, 211-213, 225, 230, 241, 242, 248, 270, 272, 273 varia, Chlamys, 314 variegatus, Donax, 316 Veneracea, 295, 316 Veneridae, 295, 316 Veneroida, 295, 314 Venerupis, 317 ventricosa, Iphigena, 157 Venus, 303, 316 vermicularis, Serpuloides, 17 verrucosa, Venus, 316 Vertigo, 156, 157 Vertiginidae, 152 Vespericola, 151, 156, 160, 164, 189, 261 villosiuscula, Thracia, 317 virginica, Crassostrea, 314 Vitrea, 151, 157 vitrea, Chlamys, 314 Vitrina, 156 vittatus, Donax, 303, 316 Viviparus, 81, 87 vulgata, Triodopsis, 163, 165, 167, 168, 177, 187, 192, 195-197, 201, 203-205, 207, 209-215, 222, 224, 225, 237, 240-244, 246, 249, 266, 268, 270 vultuosa, Triodopsis, 163-165, 168, 181, 188, 196, 198, 200-204, 209, 210, 213-215, 225, 226, 241, 242, 246, 248, 268 walleri, Aclis, 67, 77, 78 *Webbhelix, 159, 160, 163, 165, 174, 188, 190, 197, 199-203, 205-208, 210-215, 217-221, 225, 227, 229, 237, 240, 242-245, "262, 264 whiteavesii, Cerithiella, 77, 78 Wilcoxorbis, 168, 225, 242 Xanthodaphne, 79 Xolotrema, 159, 160, 163-165, 167, 168, 175, 176, 188-192, 194, 196, 197, 199-203, 205-215, 220, 221, 225, 227-229, 240, 242-245, 247-250, 257, 263-265, 268 Xylophaga, 317 Yoldiella, 292, 313 zaletus, Mesodon, 166, 206, 214, 255 Zenobiella, 157 Zirfaea, 317 Zoarchidae, 71 Zonitoides, 148, 156 VOL. 28 1988 MALACOLOGIA 25th Anniversary International Journal of Malacology Revista Internacional de Malacologia Journal International de Malacologie Международный Журнал Малакологии Internationale Malakologische Zeitschrift Publication date Vol. 27, No. 2-17 December, 1986 MALACOLOGIA, VOL. 28 CONTENTS A. J. CAIN The scoring of polymorphic colour and pattern variation and its genetic basis o eo o eye er ee ee 1 А. J. САМ The colours of marine bivalve shells with special reference to Масота PANIC EEE N ee ee 289 R. A. D. CAMERON Incomplete convergence of shell sizes and shapes in forest snail faunas from two continents: a relic of environmental history? ..................... 147 Е. G. CLAVERIA 4 Е. J. ETGES Spermatogenesis in Oncomelania hupensis quadrasi, a molluscan host of SEMSIOSOMA ADONIGUINM ee at e a en DES oooO tees 81 A. COOK 8 D. J. RADFORD The comparative ecology of four sympatric limacid slug species in Northern EE A ое A a 131 Y. A. VAN DUIVENBODEN 4 A. TER MAAT Mating behaviour of Lymnaea stagnalis ................................... 53 D. J. EERNISSE & K. KERTH The initial stages of radular development in chitons (Mollusca: Вора сорго 95 К. С. ЕМВЕАТОМ The genitalic, allozymic, and conchological evolution of the eastern North American Triodopsinae (Gastropoda: Pulmonata: Polygyridae) ............ 159 W. $. GRANT & Е. М. UTTER Genetic heterogeneity оп different geographic scales in Nucella lamellosa (Brosobranchia, IIihaldidae) rer res a ane 275 M. IBANEZ, J. BARQUIN, E. CAVERO & M. R. ALONSO | La variabilidad de Hemicycla bidentalis (Gastropoda, Helicidae) .......... 105 J. MASON & J. COPELAND The incidence and variety of Lehmannia valentiana conjoined twins: related breeding experiments (Gastropoda, Pulmonata) .......................... и С. ROLDAN & Р. GARCIA-CORRALES Anatomy and histology of the alimentary tract of the snail Theba pisana (GastropodaxPulmonata). ee ren ea en ee Ai 119 ©. Р. ROLLO The feeding of terrestrial slugs in relation to food characteristics, starvation, maturationsandıliferhistory: ee. es cele tec ee een 29 ©. НОО A quantitative analysis of food consumption for the terrestrial Mollusca: allometry, food hydration and temperature ................................ 41 F.K. VALE & M. A. REX Repaired shell damage in deep-sea prosobranch gastropods from the МЕ аа catas ee ое 65 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. 28(1-2) 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. While publishing symposia from time to time, MALACOLOGIA encour- ages submission of single manuscripts on diverse topics. Papers of local geographical or systematic interest should be submitted elsewhere, as should papers whose primary thrust is physiology 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 inthe Council of Biology Editors (CBE) Style Manual (ed. 5, 1983) available for U.S. $24.00 Нот CBE, 9650 Rockville Pike, Bethesda, MD 20814, 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 letters and spaces. MALACOLOGIA 1988 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 journal, so plan accordingly. The maximum size of a printed figure is 13.5 x 20.0cm (preferably 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 ог Ys. Letters and numbers should be at least 3mm high after reduction. Several drawings or photographs may be grouped together to fit apage. Photographs are to be high contrast. High contrast is especially im- portant for histological photographs. 13. All illustrations are to be numbered sequentially as figures (not grouped as plates or as lettered subseries), and are to be ar- ranged 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 nondi- agrammatic figures, and should be conve- nient lengths (e.g., “200 um,” not “163 um”). Magnifications in captions are not accept- able. 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 information given in the text. Each let- tered abbreviation labeling an individual fea- ture in a figure must either be explained 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 se- rials are cited unabbreviated. Supply pagina- tion for books. Supply information on plates, etc., only if they are not included in the pagination. 19. In 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 100 or more reprints are or- dered, an author receives 25 additional cop- ies free. Reprints must be ordered at the time proof is returned to the Editorial Office. Later orders cannot be considered. For each au- thors’ 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. 29, personal subscriptions are U.S. $17.00 and institutional subscriptions are U.S. $27.00. Address inquiries to the Subscription Office. VOL. 28, NO. 1-2 MALACOLOGIA 1988 CONTENTS a A. J. CAIN y = The scoring of polymorphic colour and pattern variation and its sone basis A ¡Nmolliscan shell: cuco is RE ON A J. MASON & J. COPELAND The incidence and variety of Lehmannia valentiana conjoined twins: oi, breeding experiments (Gastropoda, Pulmonata) ........................ oe C. D. ROLLO | The feeding of terrestrial slugs in relation to food characteristics, starvation, _ MAO and ею у ло ae ee TN в С. D. ROLLO AN A quantitative analysis of food consumption for the terrestrial Mollusca: allometry, food hydration and temperature ................................. Y. A. van DUIVENBODEN & A. TER MAAT a 14 Mating behaviour of Lymnaea stagnalis ........................... В: E, E : 11 ¿AQ F. K. VALE 8 M. A. REX Repaired shell damage in deep-sea prosobranch gastropods from ‘thes WESIER-INGEIRSAUARNG TL. AU. RE ANR lc ON . Е. G. CLAVERIA & Е. J. ETGES № Spermatogenesis in Oncomelania hupensis quadrasi, a molluscan ho í Schistosoma japonicum Be res De A A В. PEL D. J. EERNISSE 4 K. KERTH The initial stages of radular leal in chitons Е Poly a О И ec М. IBANEZ, J. BARQUIN, Е. CAVERO & М. В. ALONSO у La variabilidad de Hemicycla bidentalis (Gastropoda, Helicidae) В С. ROLDAN & P. GARCIA-CORRALES Br AN and ea of the alimentary tract of the snail Theba pisana г А. COOK & D. J. RADFORD Я The comparative ecology of four sympatric limacid slug species п Nort ern 7 Ur AS a ee RRE PO EA NG ne т R. A. D. CAMERON | Incomplete convergence of shell sizes and shapes in forest snail faunas from two continents: a relic of environmental history? ......... e К. С. ЕМВЕВТОМ ne The genitalic, allozymic, and И evolution of the eastern American Triodopsinae (Gastropoda: Pulmonata: Polygyridae) .. W.S. dde M. UTTER | A. J. CAIN The colours of marine bivalve shells with epoca reference to Macoma balthica .......... а ee ОА a о ones МЕХ ТОО. 28 INO! ЕР er А nr. 519 U О 4 A U ‘ Г h u 6 1 | PN WU oe e 3 eh CAT Г y 2 Kr 1 ‘ 1 , so У UN | y de Anl a AO nis Lin Fr ar Re LA г $ HE Sar Зина + EIER ET ER ее А ет" 4 u CCC RED É > Ben al Tres ER dons FE gone ARRE Mate A LEHE tas COLLE PCT EEE ESS ets nn ur, etree deere DELETE a IE FE Zu LA anlar ys terete AS APRA entró 2 2 ne ge . nee rain au . " deb Ir PA PA A A et m FAA «¿er ма u: pee a e ь и ele eres THERE ИЛЬ eee 7 ehem rr tal . у! MACON PPT os ed rev e ur ” Vip dur 0% VE wie № DEMO АК, * vera e+ ter hoja HP A LE + “ ee aa N ALA RE IIS owt ue. ee es Ua LIE ALLE PTE rit rats tether) were da 4 i > LICHT [LTÉE EEE LTÉE IE DE ВОС Y 1 ея NEE PAUSE ETES #4 PETE | venden 4 # LITA ANI st NACRE 4 22 ee HER ADS AE ri Pale f WU AA at