Volume 184 THE Number 1 BIOLOGICAL BULLETIN LIBRARY FEB191993 WuOv'S 'I-' 1 ''!. Vi.TSS. FEBRUARY, 1993 Published by the Marine Biological Laboratory al Laooratoiy THE 1 9 1993 BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Associate Editors PETER A. V. ANDERSON, The Whitney Laboratory, University of Florida DAVID EPEL. Hopkins Marine Station, Stanford University J. MALCOLM SHICK, University of Maine, Orono Editorial Board WILLIAM D. COHEN, Hunter College DAPHNE GAIL FAUTIN, University of Kansas WILLIAM F. GILLY, Hopkins Marine Station, Stanford University ROGER T. HANLON, Marine Biomedical Institute, University of Texas Medical Branch CHARLES B. METZ, Llniversity of Miami K. RANGA RAO, University of West Florida RICHARD STRATHMANN, Friday Harbor Laboratories, University of Washington STEVEN VOGEL, Duke University SARAH ANN WOODIN, University of South Carolina Editor: MICHAEL J. GREENBERG, The Whitney Laboratory, University of Florida Managing Editor PAMELA L. CLAPP, Marine Biological Laboratory FEBRUARY, 1993 Printed and Issued by LANCASTER PRESS, Inc. 3575 HEMPLAND ROAD LANCASTER, PA THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory, MBL Street, Woods Hole, Massachusetts 02543. Subscriptions and similar matter should be addressed to Subscription Manager. THE BIOLOGICAL BUL- LETIN, Marine Biological Laboratory, Woods Hole. Massachusetts 02543. Single numbers, $35.00. Sub- scription per volume (three issues), $87.50 ($175.00 per year for six issues). Communications relative to manuscripts should be sent to Michael J. Greenberg. Editor-in-Chief, or Pamela L. Clapp, Managing Editor, at the Marine Biological Laboratory. Woods Hole, Massachusetts 02543. Telephone: (508) 548-3705, ext. 428. FAX: 508-540-6902. E-mail: pamclig'hoh. mbl.edu. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN. Marine Biological Laboratory. Woods Hole, MA 02543. Copyright 1993, by the Marine Biological Laboratory Second-class postage paid at Woods Hole, MA, and additional mailing offices. ISSN 0006-3 185 INSTRUCTIONS TO AUTHORS The Biological Bulletin accepts outstanding original research reports of general interest to biologists throughout the world. Papers are usually of intermediate length (10-40 manuscript pages). A limited number of solicited review papers may be ac- cepted after formal review. A paper will usually appear within four months after its acceptance. Very short, especially topical papers (less than 9 manuscript pages including tables, figures, and bibliography) will be pub- lished in a separate section entitled "Research Notes." A Re- search Note in The Biological Bulletin follows the format of similar notes in Nature. It should open with a summary para- graph of 150 to 200 words comprising the introduction and the conclusions. The rest of the text should continue on without subheadings, and there should be no more than 30 references. References should be referred to in the text by number, and listed in the Literature Cited section in the order that they appear in the text. Unlike references in Nature, references in the Re- search Notes section should conform in punctuation and ar- rangement to the style of recent issues of The Biological Bulletin Materials and Methods should be incorporated into appropriate figure legends. See the article by Lohmann et al. (October 1990, Vol. 179: 2 1 4-2 1 8) for sample style. A Research Note will usually appear within two months after its acceptance. The Editorial Board requests that regular manuscripts con- form to the requirements set below: those manuscripts that do not conform will be returned to authors for correction before 1. Manuscripts. Manuscripts, including figures, should be submitted in triplicate. (Xerox copies of photographs are not acceptable for review purposes.) The original manuscript must be typed in no smaller than 12 pitch, using double spacing (in- cluding figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper, 8'/2 by 1 1 inches. Please, no right justification. Manuscripts should be proofread carefully and er- rors corrected legibly in black ink. Pages should be numbered consecutively. Margins on all sides should be at least 1 inch (2.5 cm). Manuscripts should conform to the Council of Biology Ed- itors Style Manual. 5th Edition (Council of Biology Editors, 1983) and to American spelling. Unusual abbreviations should be kept to a minimum and should be spelled out on first reference as well as defined in a footnote on the title page. Manuscripts should be divided into the following components: Title page. Abstract (of no more than 200 words). Introduction, Materials and Methods, Results, Discussion. Acknowledgments. Literature Cited, Tables, and Figure Legends. In addition, authors should supply a list of words and phrases under which the article should be indexed. 2. Title page. The title page consists of: a condensed title or running head of no more than 35 letters and spaces, the manuscript title, authors' names and appropriate addresses, and footnotes listing present addresses, acknowledgments or contri- bution numbers, and explanation of unusual abbreviations. 3. Figures. The dimensions of the printed page. 7 by 9 inches, should be kept in mind in preparing figures for publi- cation. We recommend that figures be about 1 '/: times the linear dimensions of the final printing desired, and that the ratio of the largest to the smallest letter or number and of the thickest to the thinnest line not exceed 1:1.5. Explanatory matter gen- erally should be included in legends, although axes should always be identified on the illustration itself. Figures should be prepared for reproduction as either line cuts or halftones. Figures to be reproduced as line cuts should be unmounted glossy photo- graphic reproductions or drawn in black ink on white paper, good-quality tracing cloth or plastic, or blue-lined coordinate paper. Those to be reproduced as halftones should be mounted on board, with both designating numbers or letters and scale bars affixed directly to the figures. All figures should be numbered in consecutive order, with no distinction between text and plate figures. The author's name and an arrow indicating orientation should appear on the reverse side of all figures. 4. Tables, footnotes, figure legends, etc. Authors should follow the style in a recent issue of The Biological Bulletin in preparing table headings, figure legends, and the like. Because of the high cost of setting tabular material in type, authors are asked to limit such material as much as possible. Tables, with their headings and footnotes, should be typed on separate sheets, numbered with consecutive Roman numerals, and placed after CONTENTS No. I, FEBRTARY 1993 CELL BIOLOGY Costas, Eduardo, Angeles Aguilera, Sonsoles Gon- zalez-Gil, and Victoria Lopez-Rodas Contact inhibition: also a control for cell prolifer- ation in unicellular algae? DEVELOPMENT AND REPRODUCTION Fenteany. Gabriel, and Daniel E. Morse Specific inhibitors of protein synthesis do not block RNA synthesis or settlement in larvae of a marine gastropod mollusk (Haliati* n//o//v) Freeman, Gary Metamorphosis in the brachiopod Ti >< 'Innlulni: ev- idence for a role of calcium channel function and the dissociation of shell formation from settlement ECOLOGY AND EVOLUTION Curtis, Lawrence A., and Karen M. K. Hubbard Species relationships in a marine gastropod-tre- ntatode ecological system Douillet, Philippe, and Christopher J. Langdon Effects of marine bacteria on the culture of axenic oyster Crassostrea gigas (Thunberg) larvae 15 25 36 Okamura, Beth, and Lita Ann Doolan Patterns of suspension feeding in the freshwater bi vo/oan Plumatella irpens 52 Scheltema, Amelie H. Aplacophora as progenetic aculiferans and the coe- lomate origin of mollusks as the sister taxon of Si- puncula 57 IMMUNOLOGY Rinkevich, B., Y. Saito, and I. L. Weissman A colonial invertebrate species that displays a hi- erarchy of allorecognition responses 79 S.i u .id.*, Tomoo, Jeffrey Zhang, and Edwin L. Cooper Classification and characterization of hemocytes in Sl\t-/a f/ava . 87 PHYSIOLOGY Hidaka. Michio, and Kiwamu Afuso Effects of cations on the volume and elemental composition of nematocysts isolated from acontia of the sea anemone ('.ulliacti* f>nl\j)ii.\ 97 Mangum, Charlotte P. Hemocyanin subunit composition and oxygen binding in two species of the lobster genus Homiinn, and their hybrids 105 No. 2, APRIL 1993 DEVELOPMENT AND REPRODUCTION Abraham, Vivek C., Sunita Gupta, and Richard A. Fluck Ooplasmic segregation in the medaka (On~ias ta- lipes) egg 115 Hamel, Jean-Francois, John H. Himmelman, and Louise Dufresne Gametogenesis and spawning of the sea cucumber Psolus fabririi (Duben and Koren) 125 ECOLOGY, EVOLUTION AND BEHAVIOR Bridges, Todd S. Reproductive investment in four developmental morphs of Streblospw (Polychaeta: Spionidae) .... 144 Emschermann, Peter On Antarctic Entoprocta: nematocyst-like organs in a loxosomatid, adaptive developmental strategies, host specificity, and bipolar occurrence of species 153 Saigusa, Masayuki Control of hatching in an estuarine terrestrial crab. II. Exchange of a cluster of embryos between two females 186 Takeda, Satoshi, and Minoru Murai Asymmetry in male fiddler crabs is related to the basic pattern of claw-waving display 203 PHYSIOLOGY Ellington, W. Ross Studies of intracellular pH regulation in cardiac myocytes from the marine bivalve mollusk, Merce- naria campechiensu 209 CONTENTS Matsushima, O., T. Takahashi, F. Morishita, M. Fujimoto, T. Ikeda, I. Kubota, T. Nose, and W. Miki Two S-Iamidf peptides, AKSGEYRIamide and VSSEYRIamide, isolated from an annelid. McFarland, F. K., and G. Muller-Parker Photosynthesis and retention of zooxanthellae within the aeolid nudibranch Amluhti papillosa . . 216 223 Rees, Bernard B., and Steven C. Hand Biochemical correlates of estivation tolerance in the moiiniainsnailOwi//i-//\(Piilmonata:Oreohelicidae) 230 Wright, Jonathan C., and John Machin Atmospheric water absorption and the water budget of terrestrial isopods (Crustacea, Isopoda. Onisci- dea) 243 No. 3, JUNE 1993 REVIEW McEdward, Larry R., and Daniel A. Janies Life cycle evolution in asteroids: what is a larva.' PHYSIOLOGY 255 DEVELOPMENT AND REPRODUCTION Buckland-Nicks, John Hull cupules of chiton eggs: parachute si rue lures and sperm focusing devices? 269 Bollner, Tomas, and I. A. Meinertzhagen The patterns of bromodeoxyuridine incorporation in the nervous system of a larval ascidian, Ciinin ni- testinalis 277 Harvell, C. Drew, and Richard Helling Experimental induction of localized reproduction in a marine bryozoan 286 Montgomery, Mary K., and Margaret McFall-Ngai Embryonic development of the light organ of the sepiolid squid Euprymna sculopes Berry 296 BIOCHEMISTRY Weis, Virginia M., Mary K. Montgomery, and Mar- garet J. McFall-Ngai Enhanced production of ALDH-like protein in the bacterial light organ of the sepiolid squid Eiijji-yiiniii seolupes 309 Gaus, Gabriele, Karen E. Doble, David A. Price, Mi- chael J. Greenberg, Terry D. Lee, and Barbara-Anne Battelle The sequences of five neuropeptides isolated from Liinulii-, using antisera to FMRFamide 322 Tamura, Shouhei, Takahiko Shimizu, and Susumu Ikegami Endocytosis in adult eel intestine: immunological detection of phagocytic cells in the surface epithe- lium . 330 RESEARCH NOTE Smith, Andrew M., William M. Kier, and Sonke Johnsen The effect of depth on the attachment force of lim- pets 338 VIEWS AND DISCUSSION Rinkevich, Baruch Immunological resorption in Bo/n7/in schlosseri (Tunicata) chimeras is characterized by multilevel hierarchial organization of histocompatibility alleles. A speculative endeavor 342 Index to Volume 184 . 346 the Literature Cited. Figure legends should contain enough in- formation to make the figure intelligible separate from the text. Legends should be typed double spaced, with consecutive Arabic numbers, on a separate sheet at the end of the paper. Footnotes should be limited to authors' current addresses, acknowledg- ments or contribution numbers, and explanation of unusual abbreviations. All such footnotes should appear on the title page. Footnotes are not normally permitted in the body of the text. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more than two authors cited as Jones el al.. 1980. Personal communications and ma- terial in preparation or in press should be cited in the text only, with author's initials and institutions, unless the material has been formally accepted and a volume number can be supplied. The list of references following the text should be headed Lit- erature Cited, and must be typed double spaced on separate pages, conforming in punctuation and arrangement to the style of recent issues of The Biological Bulletin. Citations should in- clude complete titles and inclusive pagination. Journal abbre- viations should normally follow those of the U. S. A. Standards Institute (USASI). as adopted by BIOLOGICAL ABSTRACTS and CHEMICAL ABSTRACTS, with the minor differences set out below. The most generally useful list of biological journal titles is that published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of Serials; the most recent issue). Foreign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PE- RIODICALS, may find a booklet published by the Biological Council of the LI.R. (obtainable from the Institute of Biology, 41 Queen's Gate, London, S.W.7, England, U.K.) useful, since it sets out the WORLD LIST abbreviations for most biological journals with notes of the USASI abbreviations where these differ. CHEMICAL ABSTRACTS publishes quarterly supplements of ad- ditional abbreviations. The following points of reference style for THE BIOLOGICAL BULLETIN differ from USASI (or modified WORLD LIST) usage: A. Journal abbreviations, and book titles, all underlined (for italics) B. All components of abbreviations with initial capitals (not as European usage in WORLD LIST e.g.. J. Cell. Comp. Physiol. NOT/ cell. comp. Physiol.) C. All abbreviated components must be followed by a pe- riod, whole word components must not (i.e.. J. Cancer Res.) D. Space between all components (e.g., J Cell. Comp. Physiol., not J. Cell. Comp. Physiol.) E. Unusual words in journal titles should be spelled out in full, rather than employing new abbreviations invented by the author. For example, use Rii I 'fsindafjelags Islendinga with- out abbreviation. F. All single word journal titles in full (e.g.. I 'eliger. Ecol- ogy. Brain). G. The order of abbreviated components should be the same as the word order of the complete title (i.e.. Proc. and Trans, placed where they appear, not transposed as in some BIOLOGICAL ABSTRACTS listings). H. A few well-known international journals in their pre- ferred forms rather than WORLD LIST or USASI usage (e.g.. Nature. Science, Evolution NOT Nature, Land.. Science, N.Y.; Evolution. Lancaster. Pa.) 6. Reprints, page proofs, and charges. Authors receive their first 100 reprints (without covers) free of charge. Additional re- prints may be ordered at time of publication and normally will be delivered about two to three months after the issue date. Authors (or delegates for foreign authors) will receive page proofs of articles shortly before publication. They will be charged the current cost of printers' time for corrections to these (other than corrections of printers' or editors' errors). Other than these charges for authors' alterations. The Biological Bulletin does not have page charges. Reference: Biol Bull 184: 1-5. (February. Contact Inhibition: Also a Control for Cell Proliferation in Unicellular Algae? EDUARDO COSTAS, ANGELES AGUILERA, SONSOLES GONZALEZ-GIL, AND VICTORIA LOPEZ-RODAS Genetica Production Animal, Facnltad de 1'eterinaria. Universidad Complutense, Madrid, Spain Abstract. According to traditional views, the prolifera- tion of unicellular algae is controlled primarily by envi- ronmental conditions. But as in mammalian cells, other biological mechanisms, such as growth factors, cellular aging, and contact inhibition, might also control algal proliferation. Here we ask whether contact inhibition reg- ulates growth in several species of unicellular algae as it does in mammalian cells. Laboratory cultures of the di- noflagellate Prowcentrum lima (Ehrenberg) Dodge show contact inhibition at low cell density, so this would be an autocontrol mechanism of cell proliferation that could also act in natural populations of P. limn. But, Synecho- cystis spp., Phaeodactylum tricornntnm (Bohlin), Skele- tonema costatmn (Greville), and Tetraselmis spp. do not exhibit contact inhibition in laboratory cultures because they are able to grow at high cellular density. Apparently their growth is limited by nutrient depletion or catabolite accumulation instead of contact inhibition. Spirogyra in- signis (Hassall) Kutz, Prorocentrum triestinum Schiller, andAlexandrium tamarense (Hsfiaa) Balech show a com- plex response, as they are able to grow in both low and high cell density medium. These results suggest that con- tact inhibition is more adaptative in benthic unicellular algae. Introduction Environmental conditions (light, nutrients, tempera- ture, and turbulence) are thought to be the main controls of proliferation in unicellular algae. Thus, axenic cultures of algae progressively increase in cell number until division slows due to nutrient depletion, the shadowing of some cells by others, or metabolite accumulation. But other mechanisms could play an important role in autocontrol Received 26 March 1992; accepted 23 October 1992. of algal proliferation. In this respect, endogenous rhythms have been proposed as pacemakers of algal proliferation (reviewed by Edmunds, 1988). Also, mucilage production has been considered a mechanism of biological autocon- trol in unicellular algae (Margalef, 1989). Recently, Wyatt and Reguera (1989) proposed that the onset of phyto- plankton blooms and red tides are due to a mechanism of ecological autocontrol acting at the Gaian level. Several biological mechanisms that control the cell di- vision cycle in mammalian cells have recently been elu- cidated. They are based on growth factors, genes, and gene products that respond to growth factors (Baserga ft ai. 1986; Goustin ft ai, 1986; Cantley et ai. 1991; North, 1991 ). Although these mechanisms have been interpreted as adaptations for regulating cellular proliferation in mul- ticellular organisms, they are common to all eukaryotic cells, even regulating the cleavage of zygotes (Murray and Kirschner, 1989). Recently, we have proven that the cell division cycle in unicellular algae from different phyla (Cyanophyceae, Dinophyceae, Bacillarophyceae, and Chlorophyceae) are regulated by growth factors just as are mammalian cells (Costas and Lopez-Rodas, 199 la; L6- pez-Rodas et ai, 1991). In addition to regulation by growth factors, other mechanisms control the cell proliferation of mammalian cells. For example, some cells are genetically programmed to degenerate and die of old age after a determined number of generations. Also, the unicellular algae Spirogyra in- signis (Conjugatophyceae) undergoes cellular aging as do mammalian cells (Costas and Lopez-Rodas, 1991b). In mammals, another important regulator of cellular proliferation is contact inhibition. Mammalian cells grow in monolayers, colonizing the bottom of culture flasks, but they only increase until their growth is inhibited by contact with neighboring cells. Various mechanisms seem to be involved in this complex phenomenon, from growth E. COSTAS ET AL Table Characteristics ofllie species used Species Phyla Charactenstic Synechocystis spp. Cyanobacteria unicellular. planktonic Prorocenlrum lima Dmophyceae unicellular. (Ehrenberg) Dodge benthic Prorocenlrum Iriestinum Dinophyceae unicellular. Schiller swimming Alexandrium tamarense Dinophyceae unicellular. (Halim) Balech swimming Tetraselmis spp. Praxinophyceae unicellular. swimming Skek'lonema costatuin Bacillarophyceae cenobial (Greville) filamentous. planktonic Pliaeodaclylum tricormitum Bacillarophyceae unicellular. (Bohlin) benthic Spirogyra insignis* Conjugatophyceae cenobial (Hasak) Kutz filamentous. benthic * Spirogyra insignis grows in cenobial filaments anchored to the bottom of the flask by the distal cell. Every cell of the filament can divide (more detail in Costas and Lopez-Rodas, 1991b). factor competence to cell shape changes related to intercell contacts (review Alberts el a!., 1983). This paper attempts to determine whether contact in- hibition can limit the growth of unicellular algae, as is the case in mammalian cells. Several species of unicellular algae from different phyla are analyzed in a combined ecological and evolutionary approach. Materials and Methods Cultures Isolation and culture procedures for the species used were previously described in detail (Costas, 1990; Costas and Lopez-Rodas, 199 la, b, c), so only a brief description is provided here. The characteristics of the eight species employed are summarized in Table I. Freshwater and marine species were grown, respectively, in Petri dishes with 20 ml of WC medium or f/2 medium (Guillard, 1975), at 22.5 0.5C and 80 Mmol nT 2 s~', 12:12 h light-dark cycle. Cultures were treated with 150 mg 1 ' penicillin and 100 mg 1~' streptomycin and were, therefore, axenic. Be- fore the experiments were performed, the cultures were tested for the presence of bacteria using epifluorescence procedures as previously described (Costas, 1990). The possible effects of antibiotics on algal proliferation were obviated, because the antibiotic treatment was applied two months before the experiments took place, so the cultures were grown under axenic conditions. Cultures were maintained by serial transfers of a 500 30 cell inoculum to fresh medium once every day. The cells grew exponentially for 20-30 days, and then the cul- tures showed density-dependent inhibition of growth. We determined that a culture reached saturation when its growth rate approached zero and its cell density reached the maximum. Saturation was easily detected because, growth rates and cell densities were determined daily. The experiments took place three days after the cultures were saturated. Experimental Design Many factors act in the cell density-dependent inhibi- tion of growth. In this investigation, we attempted to an- alyze whether contact inhibition also takes part in this process. Clonal cultures of each species were grown until saturation density was reached, and then the following two experiments were performed. Experiment 1: Cells at saturation density growing in fresh medium. All the cells of each saturated culture were collected (by centrifugation at 1000 rpm for 20 min), and resuspended in the same quantity of fresh medium. In this way we obtained a culture in fresh medium with sat- urated density of cells. Growth rates and cellular densities were measured during the five following days. Five rep- licates were performed for each species. Experiment 2: Cells at low density growing in saturated medium. In the second experiment, the saturated medium, after centrifugation, was filtered through a 0.22 ^m pore filter to produce a completely axenic, saturated medium that was free of cells. In this saturated, cell-free medium, a centrifuged inoculum of the same species growing ex- ponentially, was cultured. Growth rates and cellular den- sities were measured during the five following days. Five replicates were performed for each species. If inhibition of growth by contact inhibition and other factors are mutually exclusive, then contact inhibition of growth can be detected by this system, according to the following logic. If a species exhibits contact inhibition, then it will probably be able to grow in Experiment 2. but it won't be able to proliferate in Experiment 1. On the contrary, if the growth inhibition is due to other factors (nutrient depletion or catabolite accumulation), then it will probably be able to grow in Experiment 1 but not in Experiment 2. But, if other factors (i.e., soluble factors), as well as contact inhibition affect growth, then the simple two possibility choice won't happen. To determine whether contact inhibition is a factor in growth inhibition of those algae that grow in monolayers, the following experiment was performed; i.e.. the same method used to detect contact inhibition in mammalian cells was applied to algae. The cells from half a Petri dish were removed mechanically from each saturated culture CONTACT INHIBITION IN MICROALGAE Table II Growth rates and percentage increase oj cell density in fresh medium at saturation density and in saturated medium at low density Cells at saturation density in fresh medium Cells at low density in saturated medium Exponential growth-rates Saturated growth-rates Growth rates % increase cell density Growth rates cell density Syiicchocystis spp. 0.79 0.01 -0.07 0.006 0.49 0.06" 64 5% -0.04 0.01** -4 i"; Prorocentrum lima 0.38 0.03 0.01 0.01 0.03 0.02** 2 1% 0.39 0.07** 47 2% Prorocentmm triestinitm 0.91 0.05 0.01 0.01 0.29 0.02** 33 1% 0.07 0.02** 7 2% Alexandnum tamarense 0.43 0.03 -0.03 0.02 0.12 + 0.05* 12 5% 0.07 0.02* 7 1% Telraselmis spp. 0.96 0.05 -0.04 0.01 0.87 0.07** 1 38 6% 0.01 0.01** 1 1% Ske/etonema costaturn 1.01 0.07 -0.02 0.03 0.58 0.03** 78 3% -0.07 0.01" -6 r; Phaeodactilum tricomiaum 0.88 0.04 0.02 0.01 0.52 0.01** 68 2% -0.03 0.02** -2 3% Spirnt>yra insignis 0.94 0.05 0.03 0.02 0.18 0.05** 19 4% 0.68 0.05** 97 6% * Statistically no significant differences were found (P > 0.05). ** Statistically significant differences (P < 0.01) were found between growth rates of Exp. 1 and Exp. 2. sample of monolayer species. If contact inhibition exists, the cells on the full side will continue growing into the cell-free half of the dish. Five replicates were performed in each case. A continuous recording by video microscopy helped us to evaluate this experiment. Control of hand/ ing effects Because some dinoflagellates are very sensitive to shear stress, we performed the following two preliminary ex- periments to determine whether manipulation would have detectable effects on the analyzed species. (a) Exponentially growing cells of each species were collected by centrifugation at 1000 rpm for 20 min and resuspended in the same quantity of fresh medium. Their growth rates (5 replicates of each species) were measured during the following five days and compared with the growth rates of uncentrifuged exponentially growing con- trols (5 replicates of each species). ANOVA analysis showed no significant differences (P > 0.05) between growth rates of centrifuged and uncentrifuged cells. Fur- thermore, the number of dead cells was estimated by the yellow cosine exclusion procedure (more details in Costas, 1986; Gonzalez-Chavarri, 1991), and ANOVA showed Table III Growth rates of Prorocentrum lima and Spirogyra insignis after cells were mechanically removed from half a Petri dish Border where Zone where cells cells had been had not been removed removed F no significant differences (P > 0.05) in the rate of cell death between centrifuged and uncentrifuged cells. (b) A similar procedure was employed with saturated cells, and the same results were obtained; (i.e.. there were no significant differences (P > 0.05) between centrifuged and uncentrifuged cells). More details about the proce- dures used to control the effects of handling are set out in Costas ( 1 986) and Gonzalez-Chavarri (1991). Experimental evaluation Once an experiment was initiated for each of the five replicates, both the mean growth rates (during the sub- sequent five days) and the percentage of cell density in- crease (during the subsequent 24 h) were determined. Cell density was estimated as the number of cells per square or cubic centimeter in monolayer or suspension cultures, respectively. The number of cells in each culture was de- termined by counting samples in a hemocylometer. The number of samples counted was determined according to the mean progressive technique (Williams, 1977) to obtain 95% accuracy. Growth rates were calculated as doublings per day: Prorocentrum lima Spirogyra insignis 0.31 0.07 0.47 0.03 0.01 0.01 0.13 0.02 dd' 1 = l/Ln2 Ln(Nt/No)/t, Where Nt = cells at time t; No = cells at time 0; and t = number of days between times t and (more detail in Costas, 1990). Results and Discussion Growth inhibition of saturated cultures of unicellular algae is a complex process, influenced by various factors, such as nutrient depletion, catabolite accumulation, shading effects, and possibly by contact inhibition. Be- cause these factors do not act independently, their inter- E. COSTAS ET AL Figure 1 . Growth of Prorocenlntm lima and Spirogyra insignia when cells were mechanically removed from half a Petn dish. The arrows represent the border produced in the experiment. Only the cells bordering the cell-free zone were able to grow, (a) Saturated P. lima culture at the time of removal, (b) P. lima culture 72 h after the removal. New cells have only proliferated into the open half of the plate, (c) Saturated 5. inaignis culture at the time of removal, (d) 5. insignia culture 72 h after the removal. New cells have only proliferated into Ihe free half of the plate. actions complicate a precise evaluation of the relative im- portance of each. Thus, our experimental design was aimed only at detecting whether contact inhibition takes part in cell dependent inhibition of growth. Table II summarizes the growth rates and the percent- age of cell density increases in both fresh and saturated culture media. Apparently, the dinoflagellate P. lima showed contact inhibition of growth. Both the growth rates and the cell densities of Experiments 1 and 2 were sig- nificantly different (P < 0.0 1 ). P. lima cells were not able to grow at saturation density in fresh medium (Experiment 1 ), but their growth started again in saturated medium when their cell density decreased (Experiment 2). In contrast, Synechocystis spp., Phaeodactilum iricor- nutitm, Skeletonema costatum and Tetraselmis spp. did not exhibit contact inhibition. In all the cases, statistically significant differences (P < 0.01) were detected between both the growth rates and the cell densities of Experiments 1 and 2. Apparently, their growth was limited by nutrient depletion or catabolite accumulation; thus they could proliferate at high cellular density in fresh medium (Ex- periment 1 ), but were not able to grow in saturated me- dium at low cell density (Experiment 2). Contact inhibition of growth may be an important mechanism in Spirogyra insignis. Although this species grew slowly at saturation density in fresh medium (Ex- CONTACT INHIBITION IN MICROALGAE periment 1), its growth was significantly increased (P > 0.01) at low density in saturated medium (Experiment 2). So, in S. insignis. the contact inhibition component seems to prevail because proliferation is faster in a satu- rated medium with low cell density than in fresh medium with high cell density. In Pmwcentrum tricstimun. however, a nutrient de- pendent inhibition or catabolite accumulation seemed to be more important than contact inhibition. P. tricstinum was able to grow in both experiments, although its growth in fresh medium at high cellular density was significantly (P > 0.01) faster than that in saturated medium at low cell density. In Ak'xandnwn tamarcnsc, all of the factors seemed to slow down proliferation. A. tamarcnsc cells were scarcely able to grow in either experiment. The cells of P. lima and 5. insignis were mechanically removed from half a Petri dish, and the resulting growth rates are summarized in Table III. In agreement with pre- vious experiments, the growth of P. lima and S. insignis seemed to be inhibited by a contact inhibition mechanism. In particular, only the cells bordering the cell-free zone were able to grow (Fig. 1 ). This experiment, which employs the traditional method of detecting contact inhibition in mammalian cells (Alberts et a/.. 1983), supports the hy- pothesis that contact inhibition takes place in the growth inhibition of P. lima and S. insignis saturated cultures. Only two of the three benthic species analyzed seemed to exhibit contact inhibition. These results suggest that contact inhibition is a more adaptative mechanism in benthic unicellular algae. Contact inhibition is usually thought of as a mechanism developed by animal cells to limit cell division. The results obtained in these experiments suggest an alternative in- terpretation. The dinoflagellates, which could be consid- ered the earliest group of protist, but which are also far removed from actual eukaryotes (Dodge. 1955; Herzog et ai, 1984; Costas and Goyanes, 1988), have developed contact inhibition, thereby suggesting that such a mech- anism had already been developed by unicellular organ- isms in an early era, probably as an autocontrol mecha- nism regulating natural populations. Nevertheless, contact inhibition has also evolved in the Conjugatophyceae (a recent group of higher algae that are phylogenetically far removed from dinoflagellates), suggesting that such mechanisms may have been developed independently in phylogenetically different groups of unicellular organisms. Acknowledgments Supported by DGICYT grants IN89-0163 and PS89- 0014. Literature Cited Alberts, B., D. Bray, J. Lewis, M. Raff, K. Roberts, and J. D. Watson. 1983. Molecular Biology <>/ I lie Cell. Garland Publishing, New York. Baserga, R., L. Kaczmarek, B. Calabrelta, R. Battini, and S. Ferrari. 1986. Cell cycle genes as potential oncogenes. Pp. 3-12 in Cell Crc/e ami Oncogenes. W. Tanner ami D. Gallwitz. eds. Springer- Verlag. New York. Cantley, L. C., K. R. Auger, C. Carpenter, B. Duckworth, A. Graciani, R. Kapeller, and S. Snlloff. 1 99 1 . Oncogenes and signal transduc- tion. CV//64: 281-302. Costas, E. 1986. Vltraesinictura cromosomica en dinoflagelados. Con- sidcntcioncs fvo/n/m/.v. Ph.D. Thesis. Univ. Santiago de Compostela. 240 pp. Costas, E. 1990. Genetic variability in growth rates of marine dinofla- gellates. Genenea 83: 99-102. Costas, E., and V. J. Goyanes. 1988. Comparative analysis of dinofla- gellate chromosomes and nuclei. Genet. (Lije Sci. Adv.) 7: 15-18. Costas, E., and V. Lopez-Rodas. 1 99 la. On growth factors, cell division cycle and the eukaryotic origin. Endocytobiosis & Cell Res 8: 89- 92. Costas, E., and V. Lopez-Rodas. 1991 b. Persistence of cell division synchrony in S/'/n^vra mvn,vi/s (Gamophyceae): membrane pro- teoglycans transmitting synchronizing information throughout gen- erations. Chronohiol Int 8(2): 85-92. Costas, E., and V. Lopez-Rodas. I99lc. Evidence for an annual rhythm in cell aging in Spin>f>yra m.v/#m.v (Chlorophyceae). Phmilngia 30(6): S97-S99. Dodge, J. D. 1955. Chromosome structure in the dinoflagellates and the problem of the mesokaryotic cell. 2ml. Internal. Con] on Pro- lo-ool. Exc Med Inter Cont-r Ser. No. 91: 39. Edmunds, L. N. 1988. Cellular and molecular hasis <>/ 'biological docks Spnnger-Verlag, New York. 497 pp. Gonzalez-Chavarri, E. 1991. I'roduccion de biomasa a base de 1111- eroalgas v sus a/ilicacioncs en la prutliiecion animal. Ph.D. Thesis. Universidad Complutense. 142 pp. Goustin, A. S., E. B. Leof, G. D. Shipley, and H. L. Moses. 1986. Growth factors and cancer. Cancer Res 46: 1015-1029. Guillard. R. 1975. Culture of phytoplankton for feeding marine in- vertebrates. Pp: 26-60 in Culture of Marine Invertebrate Animals, W. Smith and M. Chanley. eds. Plenum Publ. Co., New York. Herzog, M., S. Boletzky, and M. O. Soyer. 1984. Ultrastructural and biochemical nuclear aspects of eukaryote classification: independent evolution of the dinoflagellates as a sister group of the actual eu- karyotes. Origins of Lite 13: 205-215. Lopez-Rodas, V., M. Navarro, L. De La Campa, E. Gonzalez De Cha- varri, S. Gonzalez-Gil, A. Aguilera, R. Segura, and E. Costas. 1991. Tras las pistas de los primeros mecanismosde control de la division celular: Una aproximacion evolutiva. Pp. 94-108 in Crimo- canccrolngia. F. Chavarria, ed. Fundacion Cientifica A.E.C.C. Ma- drid. Margalef, R. 1989. Condiciones de aparicion de la purga de mar y presiones de seleccion sobre sus componentes. Cuademos da Area de Ciencias Marinas 4: 1 3-20. Murray, A. W., and M. W. Kirschner. 1989. Dominions and clocks: The union of two views of the cell cycle. Science 246: 614-621. North, G. 1991. Starting and stopping. Nature 351: 604-605. Williams, M. 1977. Stereological techniques. Pp. 226 in Practical methods in Electron Microscopy !'<>/ 17 M. Hayat. ed. Elsevier Sci. Publ. Co., New York. Wyatt, T., and B. Reguera. 1 989. ( , Ha alcanzado el cultivo de mejillon en Galicia su masa critica? Cuademos da Area de Ciencias Marinas 4:63-71. Reference: Bio/, Bull. 184: 6-14. (February, 1993) Specific Inhibitors of Protein Synthesis Do Not Block RNA Synthesis or Settlement in Larvae of a Marine Gastropod Mollusk (Haliotis mfescens) GABRIEL FENTEANY 1 AND DANIEL E. MORSE 2 Department of Biological Sciences and the Marine Biotechnology Center, Marine Science Institute. University of California, Santa Barbara, California 93106 Abstract. Antibiotic inhibitors of protein synthesis were tested for their effectiveness in larvae of the red abalone, Haliotis mfescens (gastropod mollusk). Emetine and an- isomycin proved highly effective in this system, while cy- cloheximide, fusidic acid, puromycin. and tetracycline were less effective. Emetine and anisomycin specifically inhibited protein synthesis but not RNA synthesis. The contribution to protein synthesis by chloramphenicol- sensitive prokaryotic contaminants was found to be un- detectable, except following the onset of symptoms of toxicity resulting from prolonged exposure to emetine or anisomycin. The induction of larval settlement and plan- tigrade attachment by 7-aminobutyric acid (GABA), a functional analog of the natural inducer of settlement, occurred even under conditions in which most protein synthesis was inhibited, as expected for a chemosensory system response, whereas subsequent developmental metamorphosis was completely blocked. Because emetine and anisomycin block protein synthesis including the synthesis of new transcription factors but do not block early transcription, treatment of marine invertebrate em- bryos and larvae with these inhibitors can be used to ob- tain a selective enrichment in the mRNA population of "early gene" transcripts induced directly by GABA and other morphogenetic signals, without dilution by new mRNAs, the appearance of which is dependent on the synthesis of new protein transcription factors. Received 5 August 1992; accepted 18 November 1992. Abbreviations: GABA, -j-aminobutync acid; TCA, tnchloroacetic acid; SSC, standard sodium chlonde-sodium citrate buffer. Present address: ' Program in Cell and Developmental Biology, Har- vard Medical School. Boston, Massachusetts 021 15. 2 Author to whom correspondence should be addressed. Introduction Developmentally competent larvae (0.2 mm in diam- eter) of the marine gastropod mollusk Haliotis mfescens (red abalone) are induced to settle from the plankton and begin metamorphosis by oligopeptides and proteins as- sociated with the surfaces of crustose red algae (Morse et al., 1979a, b, 1984; Morse and Morse, 1984), and by functional analogs of these natural inducers, such as 7-aminobutyric acid (GABA), muscimol, and baclofen (Morse et al.. 1979a.b, 1980a. Morse, 1992). These com- pounds apparently bind to chemosensory receptors, with subsequent transduction of the signal mediated by the second messengers cyclic AMP and Ca ++ (Trapido- Rosenthal and Morse, 1986a; Morse, 1992). This trans- duction pathway culminates in an excitatory depolariza- tion that is apparently triggered by the regulated opening of chloride ion channels (Morse et al.. 1980; Baloun and Morse, 1984; Morse, 1990, 1992). The morphogenetic response can be facilitated or amplified by the presence of lysine or lysine analogs (Trapido-Rosenthal and Morse, 1985, 1986b), acting through a separate lysine receptor that, in turn, stimulates a G protein-diacylglycerol signal transduction cascade (Baxter and Morse, 1987, 1992; Wodicka and Morse, 1991; Morse, 1990, 1992). Before we can understand the molecular mechanisms by which these convergent chemosensory pathways regulate larval settlement behavior and the subsequent induction of gene expression controlling cellular differentiation and prolif- eration (cf. Cariolou and Morse, 1988; Groppe and Morse, 1989; Spaulding and Morse, 1991; Degnan and Morse, 1993), we must first determine the requirements for de novo protein synthesis in these processes. Specific inhibitors of protein synthesis, such as cyclo- heximide and puromycin, have proved invaluable for such INHIBITORS OF PROTEIN SYNTHESIS studies in many other systems. Yet these well-known in- hibitors were ineffective with larvae of the red abalone (Haliotis mfesL'cns: gastropod mollusk) developing in sea- water. This finding prompted our search for inhibitors of mRNA translation that would specifically block protein synthesis in abalone larvae in seawater media, while not inhibiting RNA synthesis. The antibiotic protein synthesis inhibitors emetine and anisomycin exhibit the necessary effectiveness and spec- ificity. These compounds do not inhibit RNA synthesis or the induction of settlement and plantigrade attachment of the planktonic abalone larvae, as would be expected if these processes are mediated by a chemosensory system, but they completely block the subsequent metamorphosis which, as expected, is apparently dependent on de novo protein synthesis. These inhibitors should therefore help investigators identify the primary response genes (the transcription of which does not depend on de novo protein synthesis) and messenger RNAs responsible for the in- duction of metamorphosis. Materials and Methods Haliotis rufescens broodstock was collected off the coast of Santa Barbara, California, and production and culti- vation of larvae conducted as previously described (Morse et ai. 1977. 1978, 1979b). Spawning was induced by ex- posing gravid adults to 10 mA/ hydrogen peroxide. Male and female gametes were collected and washed separately, and then mixed to allow fertilization. Embryos and larvae were maintained in 5 ^m-filtered, U.V. -sterilized flowing seawater at 15 1C. Antibiotic inhibitors of protein synthesis were pur- chased from Sigma Chemical Company (St. Louis, Mis- souri), dissolved to make concentrated stock solutions and used fresh on the day of preparation. Tetracycline was purchased as the hydrochloride, fusidic acid as the sodium salt, and puromycin and emetine as the dihydrochlorides. Stock solutions were prepared in 0.22 /urn-filtered distilled water, either alone, or containing the minimum amount of ethanol required to completely solubilize the antibiotic. After addition of antibiotics to experimental samples, no more than 0.2% (v/v) ethanol was present in any seawater sample. Control experiments showed that the presence of 0.3% ethanol had no effect on larval behavior and settle- ment or on the level of [ 3 H]leucine incorporation into TCA-insoluble material in the larvae; higher concentra- tions of ethanol (above ca. 0.75%) and other organic sol- vents induced settlement of the larvae, with a rapidity corresponding to the concentration of solvent (data not shown). Similar results were reported earlier by Penning- ton and Hadfield (1989) for larvae of the nudibranch mollusk Phestilla sibogae. Synthesis of protein and RNA was measured by incor- poration of radioactive amino acid or nucleoside into acid- precipitable macromolecules. For each assay ca. 2000 lar- vae were placed in 10 ml of 5 /urn-filtered, U.V. -sterilized seawater in 40 ml Oakridge tubes. Rifampicin, a specific inhibitor of bacterial RNA synthesis, was added to a final concentration of 2.4 jtA/ in all samples to limit bacterial growth, except where otherwise noted. After incubations in the presence or absence of inhibitors, either L-[4,5- 3 H]leucine (150 Ci/mmol; Amersham Corporation, Ar- lington Heights, Illinois) or [5,6- 3 H]uridine (42 Ci/mmol; Amersham) was added to 0.1 or 0.2 jiCi/ml, as noted in the figure legends. For each treatment at each time point, three larval samples were used. Larvae were kept at 1 5 1 C, except as noted. To end the labeling, nonradioac- tive L-leucine or uridine was added to a final concentra- tion of 0.8 mA/ or 0.4 mA/, respectively. The Oakridge tubes then were centrifuged (16,000 rpm; 4C) for 5 min (Sorvall RC5 or RC5C Superspeed Centrifuges, Clare- mont, California); Sorvall SA-600 or SS-34 fixed angle rotors were used to pellet the larvae. The tubes were placed on ice, the water was drained off, 2 ml of cold 1 X SSC was added, and the larvae were re-suspended and ho- mogenized completely in an ice-cold Dounce homoge- nizer (7 ml Pyrex tissue grinder). One aliquot of 0.5 ml for each sample was removed, placed in a 1.5ml microfuge tube, and frozen for future quantitation of protein. For each sample, 1.5 ml of the homogenate was then placed into another microfuge tube on ice, and 100% (w/v) tri- chloroacetic acid (TCA) was added to yield a final con- centration of 10% (w/v). After acid precipitation for 30 min at 4C, each sample was poured through a 2.4 cm glass microfiber filter (GF/C; Whatman International Ltd., Maidstone, UK), washed three times with 5% TCA, then washed twice with 100% ethanol. The filters were com- pletely dried in an oven at 55C and then placed in liquid scintillation vials; 1 ml of scintillation cocktail (Bio-Safe II; Research Products International, Mount Prospect, IL) was added and radioactivity determined by liquid scin- tillation. Protein was quantitated by the method of Brad- ford ( 1976) according to the protocol of the reagent man- ufacturer (Bio-Rad Protein Assay; Bio-Rad Laboratories, Richmond, California); assays were conducted in triplicate and evaluated relative to a bovine serum albumin standard measured in parallel. The incorporation data presented are the means of triplicate determinations, with error bars representing one standard deviation. Assays of settlement and metamorphosis were con- ducted with larvae in glass scintillation vials (ca. 200 larvae in 10 ml of rifampicin-containing 5 ^m-filtered, U.V.- sterilized seawater) maintained under low illumination and observed with a dissecting microscope. Each treat- ment was conducted in triplicate. Larvae also were placed at a density comparable to that used in the incorporation G. FENTEANY AND D. E. MORSE 150 ^ 100 O O 2 o Q. o o _c 50 100 200 300 400 200 - g 150 - m CD CC 100 200 300 400 [Antibiotic] (\iM) 500 Figure I. Incorporation of ['H]leucine as a function of the concen- tration of antibiotic. After incubating 8-l()-day-old larvae (ca. 2.000 lar- vae/10 ml rifampicin-containing seawater) for 2 h in the presence or absence of blocker at the concentrations indicated, [ 3 H]leucine was added to 0.1 /iCi/ml, and the pulse allowed to proceed for 2 h. Mean control values (representing incorporation in the absence of antibiotic) are dis- placed on the abscissa for clarity. (A) Incorporation in the presence of emetine (diamonds) or anisomycin (rectangles). (B) Incorporation in the presence of cycloheximide (diamonds), puromycin (rectangles), tetra- cycline (triangles), or fusidic acid (circles). Details as described in Materials and Methods. assays (ca. 2000 larvae/ 10 ml seawater), and observed for mortality and other responses to the protein synthesis in- hibitors. Results Inhibition of protein synthesis Larvae of H. rufescens take up exogenous amino acids from seawater, as demonstrated by these and other in- vestigations (Jaeckle and Manahan, 1989), although these larvae are lecithotrophic. Several commonly used inhib- itors of protein synthesis, including cycloheximide, fusidic acid, puromycin, and tetracycline, had little or no inhib- itory effect on the overall incorporation of [ 3 H]leucine into TCA-insoluble material at concentrations that were not toxic to the larvae (Fig. 1). In marked contrast, both emetine and anisomycin proved strongly inhibitory in a concentration-dependent manner. Emetine (9 n.\f) efficiently blocked the incorporation of [ 3 H]leucine into Haliotis larvae under conditions in which 100 nM chloramphenicol (an inhibitor of protein synthesis only in prokaryotes) had no significant effect (Fig. 2A). Identical results were obtained for a range of chloramphenicol concentrations (50-600 n.\f), both in the presence or absence of 2.4 pM rifampicin (an inhibitor of bacterial RNA polymerase). Thus, in the absence of emetine, prokaryotic incorporation of [-'H]leucine was not detectable. Inhibition by emetine was quite rapid; incu- bation for 10 min with 9 fiM emetine prior to addition of radiolabel was sufficient to block incorporation to a level comparable to that produced by an incubation for 2 h (data not shown). The inhibitory effect of a single 200 150 c 100 ' 50 o "I 200 2 o 9- 8 15 c 100 50 Figure 2. Incorporation of [ 3 H]leucine in 7-day-old larvae in the presence or absence of emetine (9 nM) or anisomycin (200 //A/) and chloramphenicol (150 nKf). 7-day-old larvae were used. Pulse-labeling at 24 h following the time of initial addition of emetine was for 2 h (0.1 nCi/ml). (A) Treatments: ( 1) No emetine or chloramphenicol. (2) Chlor- amphenicol added at 2 1 h. (3) Emetine added at h. (4) Emetine added at h and chloramphenicol added at 21 h. (5) Emetine added to a con- centration of 9 ^M at h and the same amount added again at 12 h. (B) Treatments: (1) No anisomycin or chloramphenicol. (2) Chloram- phenicol added at 21 h. (3) Anisomycin added at h. (4) Anisomycin added at h and chloramphenicol added at 2 1 h. (5) Anisomycin added to a concentration of 200 /iA/ at h and the same amount added again at 12 h. INHIBITORS OF PROTEIN SYNTHESIS addition of emetine relaxed with time (Fig. 3A), and a second addition of the same amount of emetine at 12 h reduced incorporation slightly further (Fig. 2A). However, following prolonged incubation in the presence of eme- tine, the addition of chloramphenicol 3 h before labeling led to significantly lower levels of incorporation, partic- ularly at 24 h and 48 h following the addition of emetine (Figs. 2A, 3A). Therefore, some of the apparent relaxation of inhibition by emetine may be due to an increase in the proportion of protein synthesis attributable to contami- nating chloramphenicol-sensitive prokaryotes. This is likely to be the result of bacterial growth on the emetine- treated larvae themselves, as these larvae become weaker, although rifampicin (2.4 (iAf) was present throughout. The inhibitory effect of a single addition of anisomycin (200 pM) persisted longer than that caused by 9 /J.M eme- tine (Figs. 2B, 3B), and a second addition 12 h after the first did not reduce the incorporation of [ 3 H]leucine fur- ther (Fig. 2B). In the presence of anisomycin, the addition of chloramphenicol 3 h before pulse-labeling did not lead to significantly lower levels of incorporation up to 24 h (Figs. 2B, 3B). Much of the inhibition of protein synthesis by a single addition of anisomycin was reversed between 24 and 48 h (Fig. 3B). This late apparent relaxation was blocked by chloramphenicol (Fig. 3B), suggesting that it was due to an increase in prokaryotic incorporation. Effects of emetine and anisomycin on RNA synthesis To test whether emetine affects RNA synthesis, larvae were pulsed with [ 3 H]uridine both in the presence and absence of 10 6 M GABA (added 30 min following the addition of emetine). No significant inhibition of RNA synthesis was observed except at 6 h in the presence of both GABA and 9 nM emetine (Fig. 4A). The presence of emetine (9 n.M) may have a stimulatory effect on the incorporation of ['H]uridine in Haliolis larvae after 12.5 h. A similar experiment showed that the incorpo- ration of [ 3 H]uridine also was not inhibited by the addition of 200 pM anisomycin (added 60 min before addition of GABA; Fig. 4B). Effects of emetine and anisomycin on settlement, metamorphosis, and survival At concentrations sufficient to inhibit most protein synthesis, emetine and anisomycin did not block the initial induction of larval settlement and plantigrade attachment by GABA, although subsequent metamorphosis was completely blocked. Toxicity of these inhibitors was both time- and concentration-dependent. Initial settlement and plantigrade attachment of larvae induced by GABA ( 10~ 6 M and 10~ 3 M) occurred normally in the presence of 9 nM emetine (Fig. 5 A, B). Both in the presence and absence of emetine, larvae ceased their swimming behavior after c 'CD 2 Q. O) 300 200 100 . A 12 18 24 30 36 42 48 = 200 2 o Q. 150 100 50 B 12 18 24 30 Time (h) 36 42 48 Figure 3. Incorporation of ['HJleucme in the presence or absence of emetine (9 pM) or anisomycin (200 n\I) and chioramphenicol (150 nM) as a function of time following addition of emetine. Larvae were pulsed at the times after addition of emetine or anisomycin indicated for 2 h (O.I jiCi/ml). In the chloramphemcol-treated samples, chlor- amphenicol was added 3 h prior to pulse-labeling. Mean values are dis- placed slightly on the abscissa for clarity. (A) 4-day-old larvae were used (6 days old by the end of the experiment in the last 3 sets of samples). No antibiotic (diamonds); emetine (rectangles); emetine plus chioram- phenicol (triangles). (B) 5-day-old larvae were used (7 days old by the end of the experiment in the last 3 sets of samples). No antibiotic (dia- monds); anisomycin (rectangles); anisomycin plus chloramphenicol (tri- angles). addition of GABA, and plantigrade attachment followed. Attached larvae exhibited normal pedal locomotion in the presence of emetine. Abscission of the velum was also observed in the presence of emetine and occurred whether GABA (10~ 6 A/) was present or not, although at 9 /j.M emetine abscission occurred at lower levels when GABA was not present. Abscission was often premature or in- complete, particularly at higher concentrations of emetine. By 6 h after the addition of 10~ 6 M GABA, most of the larvae had settled both in the presence and absence of 9 juM emetine (Fig. 5 A). New shell growth was not observed in the presence of emetine when larvae were induced to settle with 10~ 6 M GABA, although it was observed nor- mally in settled larvae in the absence of the inhibitor by 48 h. Attachment proceeded more rapidly at the higher G. FENTEANY AND D. E. MORSE c '0) "o Q. O5 40 30 20 10 I -o 12 15 18 21 24 -.5 40 2 o Q. O o 30 20 10 B 9 12 15 18 Time (h) 21 24 Figure 4. Incorporation of [ 3 H]uridine in the presence or absence of emetine or anisomycin as a function of time after addition of GABA ( ICT 6 M). The larvae were pulsed for 20 min with radiolabeled nucleoside (0.2 ^Ci/ml) at the times indicated. Mean values slightly displaced on the abscissa for clarity. (A) I0-day-old larvae were used. Where indicated, emetine (9 p.\l) was added 30 min prior to the addition of GABA. No emetine or GABA (diamonds); emetine with no GABA (rectangles): no emetine, plus GABA (triangles); emetine plus GABA (circles). (B) 9-day- old larvae were used. Where indicated, anisomycin (200 nl\f) was added 60 min prior to the addition of GABA. No anisomycin or GABA (dia- monds); anisomycin with no GABA (rectangles): no anisomycin, plus GABA (triangles): anisomycin plus GABA, (circles). concentration of GABA; virtually all of the larvae were attached within 20 min, with no inhibition by emetine (Fig. 5B). Although emetine did not inhibit the initial rate of attachment of the larvae induced by GABA. the larvae failed to maintain their plantigrade attachment (Fig. 5A. B), and progressively more were found on their sides, ap- parently due to the toxic effect of prolonged exposure to emetine. There also was some attachment in the presence of emetine when GABA was absent (Fig. 5A. B). Prolonged exposure of larvae to emetine proved lethal. Even before any mortality was observed, larvae treated with 9 fiM emetine appeared to spend more time on the bottom of the test vial than larvae in control vials. By 36 h after the addition of emetine (9 nAf) both in the presence and absence of GABA (ca. 20 larvae/ml), few larvae were swimming and many appeared dead, while in the control vials lacking GABA, many of the larvae remained swimming and virtually all remained alive. All the larvae were dead by 54 h in the presence of 9 ^M emetine at ca. 20 larvae/ml, and by 72 h at ca. 200 larvae/ ml. Toxicity was progressively accelerated by higher con- centrations, although the initial rate of GABA-induced attachment remained unimpaired below 80 ^/emetine; in the presence of 18 nM and 40 nM emetine, virtually all of the larvae were attached within 20 min following the addition of 1(T 3 M GABA (data not shown). Exposure to 80 pM or 160 fiM emetine produced marked symptoms of toxicity; GABA-induced settlement was reduced, pre- mature abscission of the velum occurred in the presence and absence of GABA. and all of the larvae died within 6-12 h at the low density. Anisomycin appeared to exert a stimulatory effect on the activity of the larvae, particularly on the movement of the cilia. The level of swimming activity of the larvae was markedly greater in the presence of 200 nAI aniso- mycin than in control or emetine-treated vials, even after only 20 min following addition. This effect appeared to partially antagonize the initial GABA-induced attach- ment, with the attached larvae abnormally continuing sustained beating of their swimming cilia and displaying little pedal locomotion; plantigrade larvae often were dis- placed by collision with other swimming larvae, and sometimes began swimming again. The initial rates of settlement and attachment induced by 10 3 M GABA were relatively unaffected by anisomycin, although long- term attachment was reduced (Fig. 5C, D). [The weak settlement-inducing activity of high concentrations of an- isomycin itself (cf. Fig. 5C, D) may explain the biphasic settlement observed in the presence of GABA.] Aniso- mycin also produced concentration-dependent and time- dependent symptoms of toxicity, with complete mortality resulting from prolonged exposure (96 h) of larvae, at high or low density, to 200 ^M concentration. Discussion Inhibition of protein synthesis Emetine and anisomycin were found to be highly ef- fective inhibitors of protein synthesis in Haliotis larvae, whereas cycloheximide, fusidic acid, puromycin, and tet- racycline proved far less effective. Possible reasons for the limited effectiveness of these widely used inhibitors may include their instability or low solubility in seawater, or their inefficient diffusion into the deeper layers of larval tissue. There is some structural evidence supporting the suggestion that membrane permeability may be an im- portant determinant of effectiveness in the marine larval system. Emetine contains four methoxy groups, while an- isomycin contains one methoxy and one acetoxy group, all carbon-linked to cyclic nuclei in both of these anti- INHIBITORS OF PROTEIN SYNTHESIS 11 40 60 Time (h) 20 40 60 Time (h) 80 100 Figure 5. Attachment of larvae in the presence or absence of emetine or anisomycin as a function of time following addition of GABA. Cu 200 were placed in 10 ml rifampicin-treated seawater in triplicate, as described in Materials and Methods. Following a 30-min incubation with or without emetine (A and B) or a 2-h incubation with or without anisomycin (C and D), GABA was added to the samples indicated, and the mean percentage of the larvae showing attachment was scored at the times indicated. Standard deviations for all points, assayed in triplicate, were <4%/. 80 .. -. - \ \ss 1 1 -nr | i [ 20 40 60 80 Percent Metamorphosed Figure 7. (A) Histograms showing the effect of pretreatment at four days of larvae with high K + seawater that either induced partial meta- morphosis or no metamorphosis on the ability of these two categories of larvae to undergo metamorphosis after treatment with a metamor- phosis substrate or high K + seawater at five days. (B) Histograms showing effect of treatment with a metamorphosis substrate or high K* seawater at five days on metamorphosis of larvae that had not been pretreated with high K + seawater at four days. The hatched segment of each bar indicates the proportion of cases that underwent normal metamorphosis. The clear segment ol the bar indicates the percentage of cases that un- derwent partial metamorphosis. The number of cases is at the top of the bar. were not disturbed underwent normal metamorphosis by the next day (six cases). The larvae that were removed from their settlement site did not resettle but underwent partial metamorphosis (four out of six cases). This ex- periment suggests that settlement is necessary for normal metamorphosis. Discussion The role of voltage-dependent Ca 2+ channels in metamorphosis The following lines of evidence indicate that voltage- dependent calcium channels may play a role in meta- morphosis: ( 1 ) Treatment of larvae with high K + seawater which presumably depolarizes the cells of the larva induces metamorphosis and treatment of larvae with high K + in Na + -free seawater is just as effective in inducing meta- morphosis, (2) Treatment of larvae with high K 1 in Ca 2+ - free seawater inhibits metamorphosis, (3) Treatment of larvae with high K + in seawater with elevated Ca 2+ levels or Mg 2+ -free seawater increases the percentage of cases metamorphosing, (4) Treatment of larvae with high K + seawater in the presence of the calcium channel blockers Co 2+ and Nifedipine inhibits metamorphosis. In order to make this work more convincing one would have to dem- onstrate electrophysiologically that target cells are not only depolarized but give an action potential which is typical of voltage-dependent Ca 2+ channels and that Ca 2+ moves into the target cells from the external environment during depolarization. The identities of the target cells where voltage-depen- dent calcium channels function to mediate the meta- morphic stimulus is not known. One possible target cell IONIC CONTROL OF METAMORPHOSIS 23 candidate is a subset of cells in the larval nervous system. There is evidence that the nervous system receives and mediates the metamorphic stimulus in echinoid larvae (Burke, 1983b). Unfortunately, virtually nothing is known about the organization of the nervous system in articulate brachiopod larvae; however, nerve cell processes have been noted in ultrastructural studies done on these larvae for other purposes (Strieker and Reed, 1985a). Another possible set of target cells could be some of the cells that make up the surface epithelium of the larva (e.g., the cells of the distal part of the pedicle lobe). After these cells receive a metamorphic stimulus, it could be transferred to other epithelial cells of the larva by epithelial conduc- tion. There is evidence that epithelial conduction mediates the metamorphic stimulus in hydrozoans (Freeman and Ridgway, 1990). Both substrate and high K + seawater induced meta- morphosis appear to depend on calcium channel function. Substrate induced metamorphosis also depends on the pedicle lobe while high K + seawater induced metamor- phosis does not. The simplest model that accounts for these results is that there is a substrate-induced meta- morphosis receptor at the distal end of the pedicle lobe. When this is activated a metamorphic signal is sent from this site to cells outside of the distal region of the pedicle lobe that must have their putative voltage-dependent cal- cium channels activated in order to spread the metamor- phic stimulus (Fig. 8). When the cells outside the distal region of the pedicle lobe are activated, they also send an inhibitory signal to the cells in the distal region of the pedicle lobe preventing them from responding to substrate mediated metamorphic cues (Fig. 7). The signijii'iiiur <>/ '"partially metamorphosed" larvae The partially metamorphosed larva is most probably the result of an abnormal metamorphic response. This larva is characterized as a larva that forms a protegulum in the absence of mantle reversal and settlement. Because the formation of a protegulum under these conditions probably renders the mantle lobe incapable of reversal and because the mantle lobe does not spread out to occupy a larger area as it does after reversal, this metamorphic response is probably maladaptive. I have made only lim- ited attempts to look for later manifestations of normal metamorphosis in partially metamorphosed larvae. Two partially metamorphosed larvae were fixed and sectioned four days after the initiation of high K + seawater induced metamorphosis. Both of these larvae showed suggestions of cuticle deposition by the pedicle. In order to make this point with certainty, it would be necessary to do a study of these larvae at an electron microscope level of resolu- tion. I did not observe any indication of mouth formation in these partial larvae; however, they may not have been cultured long enough. N PI / !/ i >* i \ \ \ Figure 8. Diagrammatic view of a swimming larva with an apical lobe (AL). mantle lobe (ML), and pedicle lobe (PL). At the distal end of the pedicle lobe there is a postulated center composed of cells ( 1 ) which may use voltage-dependent calcium channels to transduce a substrate mediated metamorphic signal. This center sends a stimulatory meta- morphic signal to other cells in the larva including center (2) which functions via voltage-dependent calcium channels that acts as a secondary metamorphic center. Here this center is shown in the mantle lobe but it could be any place outside of the distal end of the pedicle lobe. The cells of this secondary metamorphic center send a stimulatory meta- morphic signal to other cells of the larva and an inhibitory signal to the cells that transduce the substrate mediated metamorphic signal turning off the metamorphic stimulus from these cells. This model accounts for the experiments described in this paper. A variety of factors probably play a role in generating the partial metamorphosis phenotype. In larvae that have been reared for a number of days in a sterile environment intrinsic maturational changes may occur so that various parts of the metamorphosis signaling pathway or cells that respond to the signaling pathway may be activated. If the postulated distal pedicle lobe substrate receptor cells were activated, an aged larvae may undergo normal metamor- phosis. This happened in a small percentage of cases (Ta- ble II). If cells that are part of the metamorphic pathway that reside outside of the distal region of the pedicle lobe are activated or if cells that will form the protegulum are activated, a larva that shows the partial metamorphosis phenotype would be generated. There is evidence that in some species with a bathy-pelagic life cycle that larvae which do not see an appropriate metamorphic cue in na- ture will metamorphose or partially metamorphose and still continue a pelagic existence (Thorson, 1946; Paine, 1963). The mechanics of mantle lobe reversal during meta- morphosis are not understood. There is a pair of muscles that insert in the mantle lobe and the pedicle lobe that are thought to contract during metamorphosis causing the mantle lobe to flip (Franzen, 1969; Long, 1964). Sub- strate adhesion by the pedicle lobe may be necessary for these muscles to contract or to cause the pedicle lobe to be compressed in an appropriate way as the muscles con- tract so that the mantle lobe is reversed. The production of larvae that show partial metamorphosis following sub- strate detachment could occur because protegulum for- mation is activated even though mantle reversal is inhib- ited. The small number of cases where partial metamor- 24 G. FREEMAN phosis occurs following the culture of larvae in the presence of substrates that induce metamorphosis can be explained in this way. Partially metamorphosed larvae and the conditions where they are formed provide an in- sight into the normal metamorphosis process. Acknowledgments I am grateful to Dr. A. O. D. Willows and the staff of the Friday Harbor Laboratories for their hospitality. I want to thank Sarah Cohen and her diving companions for collecting animals using SCUBA, Dr. Craig Staude for saving animals for me that were collected on dredging trips, and Drs. Alan Kohn and Patricia Morse for letting me use animals that were dredged for class use. I want to thank Judith Lundelius, Bob Goldstein, and Hyla Sweet for their comments on this manuscript. This work was supported by NSF grant DCB-8904333 and a URI re- search leave from The University of Texas. Literature Cited Burke, R. D. 1983a. The induction of metamorphosis of marine in- vertebrate larvae: stimulus and response. Can. J Zoo/. 61: 1701- 1719. Burke, R. D. I983b. Neural control of metamorphosis in Dendraster excentricus. Biol. Bull. 164: 176-188. Cameron, A., T. Tosteson, and V. llensley. 1989. The control of sea urchin metamorphosis: ionic effects. Develop. Growth Differ. 31: 589- 594. Franzen, A. 1969. On larval development and metamorphosis of Ter- cbralulina Brachiopoda. /<><>/, Bid. Uppsala 38: 155-174. Freeman, G. 1981. The role of polarity in the development of the hy- drozoan planula larva. Rou.\'s Arch. Dev Biol. 190: 168-184. Freeman, G.. and K. B. RidgHa\. 1990. Cellular and intracellular path- ways mediating the metamorphic stimulus in hydrozoan planulae. Rou\'sArch. Dcv. Biol. 199: 63-79. Hille, B. 198-1. Ionic Channels and Excitable Membranes Sinauer As- soc., Sunderland, MA. Long, J. A. 1964. The embryology of three species representing three superfamilies of articulate brachiopoda. Ph.D. Dissertation. University of Washington. Long, J. A., and S. A. Strieker. 1991. Brachiopoda. Pp. 47-84 In Re- production in Marine Invertebrates, I 'ol. 6 Echinoderms ami Lopho- phorales. A. C. Giese, J. S. Pearse. and V. B. Pearse, eds. Boxwood Press. Palo Alto. CA. Morse, D. K. 1990. Recent progress in larval settlement and meta- morphosis: closing the gaps between molecular biology and ecology. Bull. Mar Sci. 46: 465-483. Paine, R. 1963. Ecology of the brachiopod Glottidia pyramidata. Ecol. Monogr. 33: 187-213. Strathman, M. 1987. Reproduction ami Development of Marine I/t- vertehrates oj the Northern Pacific Coast. University of Washington Press, Seattle. WA. Strieker, S. A., and C. G. Reed. 1985a. The ontogeny of shell secretion in Terehratalia transversa (Brachiopoda, Articulata) I. Development of the mantle. ./ Morphol. 183: 233-250. Strieker, S. A., and C. G. Reed. 1985b. The protegulum and juvenile shell of a recent articulate brachiopod: patterns of growth and chemical composition. Lclhaia 18: 295-303. Strieker, S. A., and C. G. Reed. 1985c. Development of the pedicle in the articulate brachiopod Terebratalia transversa ( Brachiopoda, Ter- ebratulida) '/.oomorphology 105: 253-264. Thorson, G. 1946. Reproduction and larval development of Danish marine bottom invertebrates. Medd. Komm. Dan Fisk: Havundersog Ser. Plankton 4: 1-523. White, B. II., and C. S. Nicoll. 1981. Hormonal control of amphibian metamorphosis. Pp. 363-396 in Metamorphosis: A Problem in De- velopmental Biology. L. I. Gilbert and E. Frieden, eds. Plenum, New York. Yool, A. J., S. M. Green, M. Hadfield, R. Jensen, D. Markell, and I). Morse. 1986. Excess potassium induces larval metamorphosis in four marine invertebrate species. Biol. Bull. 170: 255-266. Reference: Biol. Bull 184: 25-35. (February, 1993) Species Relationships in a Marine Gastropod-Trematode Ecological System LAWRENCE A. CURTIS' AND KAREN M. K. HUBBARD* University Parallel, * School of Life Sciences, and College of Marine Studies. University of Delaware. Newark, Delaware 19716 Abstract. Individual snails (Ilyanaxsa obsoleta) on Cape Henlopen, Delaware, frequently are host to one or more trematode species. When different species occupy the same host, interactions might be expected. We investigated five species of parasites to determine whether their existence in different combinations would lead to altered within- host distributions or changed numbers of shed cercariae. Snails (32 samples, total = 379) were collected from June to August, in 1989, and microscopically examined. Par- asite species and stages present in five sections through each snail were recorded. Before examination, 206 of these snails were held in individual chambers in the field. After two high tides (ca. 24 h), the chambers were checked for species and the numbers of cercariae shed. Overall, 22 trematode combinations in single hosts were observed. Analysis revealed that co-occurrence with other species had no significant effects on any trema- tode. Further, analyses of species richness of infecting assemblages over two distinct intervals failed to show that competition is important in determining assem- blage richness. One pair of trematodes (Himasthla quissetensis and Lepocreadium setiferoides) has been reported not to co-occur. We observed co-occurrences, but so few that the apparent conflict between them could not be statistically demonstrated. We suggest that, in this system, parasites are adapted to the host only, they may interact, but they are not adapted to each other. Chances for a parasite to live free from other parasites seem too great for evolved (adapted) relationships to develop. The host, for similar reasons, is probably not adapted to the parasites. Received 14 February 1992: accepted 6 October 1992. 1 Mailing address: Cape Henlopen Laboratory, College of Marine Studies, University of Delaware, Lewes, DE 19958. Introduction For one species to be adapted to another, they must interact in such a manner that one consistently exerts a selective pressure on the other. Species interactions may be thought of as a continuum from local to global. A local interaction (as used here) results in genetic changes in restricted parts of a gene pool (and may result in local ecotypes). On the other hand, if an interaction is global, one species can be a source of biotic selective pressure over the whole operating gene pool of another. Reciprocal genetic changes between species amount to coevolution (Futuyma and Slatkin, 1983). This paper considers the species interactions in a marine gastropod-trematode system. Because the host gastropod has a planktonic larva and the trematodes are dispersed by highly mobile defin- itive hosts, both local and global phenomena must be considered. The interactions between hosts and parasites have been much discussed (see Moore, 1987 for an extensive review), and the levels at which such discussions may be focussed should be distinguished. In this work, two levels are nec- essary. The component community includes all parasite species using a particular host species (population). The infracommunity includes all the parasites in a single host (Esch el ai, 1990). An individual host, harboring a mul- tispecies parasite assemblage, is a biological unit where parasite-parasite as well as host-parasite interactions can occur. There are four basic patterns of evolutionary relation- ships that may be found in any host-parasite system (Fig. la-d). In scheme a, the parasites are adapted to the host (the minimal condition), whereas in scheme b, the host is also adapted to the parasites. Scheme c illustrates the case where the parasites are adapted to the host and to 25 26 L. A. CURTIS AND K. M. K. HUBBARD PARASITE A PARASITE B HOST PARASITE A HOST PARASITE B PARASITE A PARASITE B PARASITE A HOST PARASITE B Figure 1. Four models of possible adaptive relationships among spe- cies in a snail-trematode system. Parasites A and B may coexist in a single host. An arrow from one participant to another indicates that the participant at the origin of the arrow has evolved adaptations to selection pressures coming from the other (i.e.. "PARASITE A - PARASITE B" means A is adapted to B). One-way interactions between parasites (/.('., A adapts to B but not the reverse) are possible, but not figured. each other. Scheme d shows the case where parasites are coevolved with the host and with each other. There should be evidence of an adaptive relationship between species before it is assumed to exist (Williams, 1966). In this work, we have tested for species interactions among trematodes inhabiting the same gastropod host. The goal is to gather evidence to support the elimination of one or more of the above schemes and thereby improve our understanding of host-parasite systems. Studies of trematodes infecting gastropod populations have often revealed patterns of species co-occurrence that suggest interactions (see Rohde, 1981 for references). However, few workers have examined trematode assem- blages in individual gastropods taken from their natural habitat, to determine whether fitness of certain members is consistently affected by co-occurrence with other members (see DeCoursey and Vernberg, 1974). This is largely because multiply-infected hosts are difficult to ob- tain in numbers for study. The prevalence of trematodes in the population of Ilyanassa obsoleta (Prosobranchia, Neogastropoda) on Cape Henlopen, Delaware Bay is high. and a diversity of multiply-infected snails may be ob- tained (Curtis, 1985, 1987, 1990; Curtis and Hubbard, 1990). This allowed us to test for species interactions in a variety of trematode ensembles. Of the nine trematode species in Ilyanassa obsoleta ob- served in Delaware, five are commonly observed in the Cape Henlopen population and figure in this study: ///- masihla quissetensis, Lepocreadium setiferoides, Zoo- gonus rubellus. Aiistmbilharziu variglandis, and Gvnae- cotyla adunca. The snail is the first intermediate host. A variety of second intermediate hosts is used by these spe- cies. Various shorebirds serve as definitive hosts for H. quissetensis. A. variglandis and G. adunca, whereas fish species are used by L. setiferoides and Z. rubellus (see Stunkard, 1983 for life-cycles and taxonomic matters). Any direct species interactions among these parasites must occur in the snail, the only host they all have in common. There is no indication that Ilyanassa obsolete! lose in- fections (Curtis and Hurd, 1983). so the ensembles ob- served in snails probably represent relatively longstanding (period unknown) assemblages. Enduring species assem- blages, proximity in a natural habitat unit, and utilization of similar resources (Smyth and Halton. 1983). suggest that strong interspecific interactions might occur. If competitive interactions are frequent within individ- ual hosts whereby dominant species come to monopolize the host population through time, a pattern should emerge at the component community level. Early on, most snails should have single species infections; as time progresses species accumulate and there should be a preponderance of double and triple infections; and eventually there should be mostly single infections again, as the dominant species evict subordinates (Sousa, 1990). We searched for such a component community pattern among our snails at two time scales, through the summer and over several years. To examine within-snail parasite interactions, we tested individual species to see whether existence in different assemblages had consequences in terms of ( 1 ) alterations of within-snail spatial distributions, (2) complete suppres- sion of cercarial production, and (3) changes in numbers of cercariae released from hosts. Materials and Methods One sandbar (Fig. 2), located near the mouth of the Delaware Bay on Cape Henlopen (75 06'W, 38 471^), was chosen as the source for snails. Certain species of trematodes affect the behavior, distribution and temporal occurrence of Ilyanassa obsoleta on sandbars (Curtis, 1987, 1990). To avoid over-representing snails harboring particular parasite ensembles, we randomly chose collec- tion sites according to the angle and distance from a ref- erence point at the peak of the sandbar (Fig. 2). Samples GASTROPOD-TREMATODE INTERACTIONS 27 "0 10 20 30 40 METERS (ALONG BEACH) NE Figure 2. An elevational contour map of the 1989 sandbar on Cape Henlopen. Delaware where samples ofllyanassa ohsolctu were collected for this study. The 32 randomly selected sample sites are indicated by filled diamonds. The highest point on the map (the sandbar peak at center) is 56 cm above the lowest. were taken between 16 June and 17 August 1989 on both day and night low tides. We wanted many multiply-in- fected snails in the samples, and the 379 snails obtained (Table I) were purposely biased to include them. The snails came from an area where many multiples were likely to be found (e.g., Curtis, 1987), and large snails that were likely to be infected were chosen (Curtis and Hurd, 1983). Usually, two collections of 10 to 13 snails were collected and processed at a time. All the snails were dissected and 206 were also tested for cercarial release. We were interested in revealing gross within-snail dis- placements of individual parasite species by other species or combinations of species. Such displacements would be required if dominant species gradually evicted subordi- nates from the snail. During dissection each snail was removed from its shell and examined in sections to de- termine how individual parasite species, and stages thereof, were distributed within. Heavily parasitized snails are virtual bags of trematodes; they retain no consistent morphological landmarks that are useable as standard points of reference. Consequently, each snail was pinned to a board and cut crosswise into five equal lengths with a razor blade. Section 1 was the most dorsal portion of the snail (the spire), and section 5 was the most ventral (head and foot). The razor was cleaned between cuts and scrupulous care was taken to prevent contamination of one section with material from another. Sections were placed separately in small vials containing 5 ml filtered baywater. Each vial was vigorously shaken 50 times to release the contained trematode stages into the water. A small amount of the water was placed on a slide and examined with the aid of dissecting (32X) and compound (100X) microscopes. We took two samples from each vial. The species and stages of the trematodes were recorded for each section of each snail as follows: parental stages (rediae or sporocysts) and cercariae (PC); cercariae only (C); parental stage without mature cercariae (P); or absent (A). Observed cercariae may have been lib- erated from parental stages during the procedure, but this does not matter as we were only interested in whether formed (mature) cercariae were present. Trematodes were never found in section 5, and after the hundredth snail we stopped examining this section. Table I Trematode infections in Ilyanassa obsoleta collected for this study from a sandbar area (Fig. 2) on Cape Henlopen. Delaware Infecting species n Mean shell height (mm) Range shell height (mm) uninfected 18 21 17-25 singles Hq 74 24 20-27 Ls 25 24 20-27 Zr 29 22 20-26 Av 5 23 22-24 Ga 29 22 18-25 doubles HL 4 23 22-23 HZ 42 24 20-27 HG 10 23 21-25 LZ 8 24 20-28 LA 1 25 LG 32 23 20-26 ZA 1 24 ZG 24 23 20-27 AG 1C) 22 19-24 AD 1 23 GD 1 21 triples HZG 33 24 18-26 LZA 1 28 LZG 22 23 17-26 LAG 4 21 20-24 ZAG 5 23 23-25 Total = 379 Overall = 23 17-28 For each infection, number collected (n), and mean and range of shell heights are given. Snails infected by a single species (singles) are repre- sented by the genus and species initials of the trematode (Hq = Himasllila qitissetensis, Ls = Lepocreadium setiferoides, Zr = Zoogonus rubellits. Av = Austrobilhariia variglandix. Ga = Gynaecotyla adunca). Double and triple infections are represented with the generic initials of the species involved (e.g., a snail infected with H. quissetensis, Z. rubellus. and G adunca goes in the HZG category). Diplostomum nassa (D) occurred only in double infections. Shell height = siphonal canal to apex of shell (e.g., 21 = 20.5 to 21.4 mm). L. A. CURTIS AND R. M. K. HUBBARD The frequencies of parasite presence or absence in the snails were crosstabulated according to the following cri- teria: parasite assemblage (those species infecting the snail); snail section (1 - 4); and the stage of the parasite (sporocysts, rediae, cercariae). Contingency table analyses were employed to test for significant displacements of parasite stages within snails. For each parasite, we used log linear models (Sokal and Rohlf, 1981 ) to calculate the expected frequencies of occurrence of the stages (PC, P, C, or A) in various sections of hosts harboring various trematode ensembles. A saturated log linear model for this kind of analysis includes seven terms: Assemblage; Section; Stage; Assemblage x Section; Assemblage X Stage; Section X Stage; and Assemblage X Section X Stage. The purpose of this analysis is to learn which of these terms are necessary to calculate a set of expected frequencies that do not deviate significantly from the ob- served frequencies. After unnecessary terms are elimi- nated, we are left with the accepted model. The accepted model is expressed in hierarchical form. For example, an Assemblage X Section X Stage hierarchical model would nest all seven terms of the full model; and an Assemblage, Section X Stage model would nest all three one-way terms and the Section X Stage two-way term. If species interactions lead to spatial rearrangements within snails, a 3-way interaction term (i.e.. Assemblage X Section X Stage) would be necessary in the accepted model for any displaced species. For example, suppose species "a" were usually distributed throughout the snail from spire to mantle when it occurred alone, but in the presence of species "b" (i.e., assemblage "ab"), "a" were consistently absent from the spire section. The three-way interaction term would be necessary in the accepted model because absence (A) of species "a" from Section 1 would be a consequence of Assemblage composition. That is, the presence of species "b" in Section 1 would change Stage entries for "a" in Section I to absent (A) from one of the present categories (PC, P, or C). Therefore, a table of expected frequencies that matched observed frequencies could not be calculated without the Assemblage X Section X Stage term in the accepted model. We used cercarial release as a measure of fitness to learn whether parasites were affected by within-host in- teractions with other infecting species. We evaluated cer- carial output from assemblages during one short period, and tested similar assemblages throughout the summer. (An alternative, more manipulation laden, approach would be to follow cercarial output from individual as- semblages over a longer period of time.) Individual snails were confined in chambers in the natural environment for two high tides (ca. 24 h), and the water in which they had been immersed was examined for numbers and spe- cies of cercariae. We used 24-h periods to encompass any daily shedding patterns. The procedure is described in more detail in Curtis and Hubbard (1990). We used a Kruskal-Wallis test (Hollander and Wolfe, 1973) to de- termine, for each species, whether the number of cercariae shed was significantly different when in various co-oc- curring assemblages of parasites. All statistical calculations were done with the software package. Number Cruncher Statistical System, 5X Series. Results A competition model (Sousa, 1990) suggests that trem- atodes might invade a snail population, accumulate in snails, compete, and eventually complete the process by having dominant trematodes evict subordinates. If true, then over the relevant time we should see infecting species richness start low (mostly single infections), increase (mostly doubles and triples), and then decrease again. We looked for such a pattern within two distinct intervals, over the summer (Table II A), and over several years (Table IIB). We divided the sampling period into four two-week intervals; the fourth interval was extended to encompass the 25 snails collected on August 1 7. There were significant changes in richness from one period to the next, but the expected pattern was not seen. In particular, triple infec- tions were quite abundant early in summer and were most abundant in the last sampling period. This would not have been observed if dominant trematodes had defeated sub- ordinates in this period of time. Using size-classes of snails (Table IIB), the interval can be extended from months to years. At the beginning of its third summer, a snail on Cape Henlopen is about 14-15 mm; by the end of that summer, it has grown to about 17-18 mm (Curtis and Hurd, 1983). This means that the smallest snails we collected (17 mm. Table I) were probably in their fourth summer. If 3 mm/summer is used as an estimate of growth for parasitized snails, then the < =22 group in Table IIB is 4-5 yr old; the 23- 25 group is 5-6 yr old; and the 26-28 group is 6-7 yr old. The interval encompassed by Table IIB is about three years using this estimate. Parasitized snails may not grow this rapidly, and the interval is possibly longer. In this years-long interval (size-class range), there were signifi- cant changes in infecting species richness. Note (Table IIB) that single infections were more abundant than triple infections in the youngest snails, but that the proportion of triples increased among older snails. This is not the pattern predicted by the competition model. Occurrence of stages of five trematodes in sections of Ilyanassa obsoleta harboring different assemblages is shown in Table III. Recall that section 1 was dorsal (spire) and section 4 ventral (mantle). In Table IV. models for all five species (except Austrobilliarzia vanglandis) require the one-way Assemblage term because of the widely dif- ferent numbers of snails infected with each assemblage GASTROPOD-TREMATODE INTERACTIONS 29 Table II Tremalode infections in Ilyanassa obsoleta examined during tins work crosstabulated by number of inled IIIR Iremalode species (richness), time of collection in summer 19S9 (A), and si:e (age) of snail (B) Infecting Species Richness (n = 18) % Singles (n = 162) % Doubles (n = 134) % Triples (n = 65) A. Time Collected 16 Jun-29 Jun 30Jun-13Jul 14 Jul-27 Jul 28 Jul-17 Aug B. Size Class (mm) < = 22 23-25 26-28 2-way contingency analyses: Time X richness, Xf,, = 43.69, P < 0.001 Size X richness, X, 2 6) = 25.16, P < 0.001 0.9 15.0 1.1 4.1 10.5 1.0 0.0 39.8 38.7 50.0 42.9 43.8 43.6 32.3 44.3 40.0 30.7 25.5 33.3 36.9 35.5 15.0 6.3 18.2 27.5 12.4 18.5 32.3 113 80 X8 48 153 195 31 Size class ranges are in terms of shell height as in Table I. (see Table III). The frequencies for Lepocreadium setifer- oides can be modeled by taking into account, beyond the Assemblage term, only the one-way Stage term because most of the stage entries are in the PC category. The rest of the models require the Section X Stage term because there was some specificity as to what sections were likely to harbor which stages. This is clearest for Zoogonus ru- bcllus and Gynaecotyla adunca. Stages were often (clearly not always) absent (A) from sections 1 and 4. However, this was not significantly correlated with the assemblage of species infecting the snail. For none of the five species tabulated is an Assemblage X Section X Stage (three- way) interaction term necessary in its accepted hierarchical log linear model (Table IV). That is, co-occurring trem- atodes did not significantly affect the distribution of any of the five species tested. In most snails, parasite stages of all species present occurred throughout. Parasite species interactions could lead to cercarial suppression in a section rather than species eviction. For example, if the presence of species "a" suppressed cercarial production by species "b", the accepted log linear model for species "b" would have to include the Assemblage X Stage term. This would be necessary because, for species "b" in the presence of species "a", the frequency of the PC category would decrease, while the frequency of the P category would increase as compared to other assem- blages involving species "b". Expected values that matched this shift in observed frequencies could not be predicted (modeled) without incorporating the influence of Assemblage on Stage. No species' cercarial production was completely suppressed in this manner (Table I V, lack of Assemblage X Stage terms). The question now becomes: given that cercariae were being produced, was the number released from snails changed as a function of assemblage composition? To answer this, we used data from cercarial release chambers. Prepatent infections (those with no cercariae present) were eliminated from this analysis because their prepatency was not caused by assemblage composition (Table IV, no Assemblage X Stage terms in the accepted models). In- cluding prepatents would add meaningless variability. Absent cercariae are not germane to this analysis if they are not caused by the presence of other species. Table V describes statistically the cercarial output of each of the five species in various assemblages. The magnitude of variability should be noted. Table VI presents the results of Kruskal-Wallis tests that were used to determine whether the assemblage com- position significantly affected the numbers of cercariae released by particular (patent) assemblage members. The results show that although cercarial output (mean rank) did decrease for all species when additional species were present, there was not a significant depression of cercarial output for any one species. Finally, because Himasthla quissetensis and Lepo- creadium setiferoides have not previously been observed together, note that in Table V such a co-occurrence is listed, and that both species shed cercariae concurrently. Four snails contained both H. quissetensis and L. setifer- oides (Table I). Based on observations of a few mature (often moribund) H. quissetensis rediae and cercariae among many L. setiferoides rediae and cercariae, it ap- peared that L. setiferoides was evicting H. quissetensis from the snails. There were not enough of these snails to 30 L. A. CURTIS AND K. M. K.. HUBBARD Table III Spatial distributions of five trematode species (see Table I for parasite abbreviations) within sint>l\- and multiply-infected Ilyanassa obsoleta. Observed frequencies of parasite occurrence thy stage*), in snail sections 1-4 (see text), are given for each species Section Infecting trematodes Species tabulated 1 Stage 2 Stage 3 Stage 4 Stage PC C p A PC C p A PC C p A PC C p A Hq(n = 74) Hq 72 1 1 73 1 73 1 69 2 1 2 HZ(n = 42) Hq 38 2 2 40 2 40 1 1 33 6 2 1 HG(n = 10) Hq 6 3 1 10 9 1 8 1 1 HZG (n = 33) Hq 30 2 1 33 32 1 27 2 4 Ls(n = 25) Ls 25 25 24 1 22 3 LZ (n = 8) Ls 8 8 8 7 1 LG (n = 32) Ls 26 4 1 32 (I 30 2 23 2 3 4 LZG (n = 22) Ls 19 2 1 21 1 (] 21 1 17 5 Zr(n = 29) Zr 27 1 27 2 27 2 17 12 HZ Zr 25 17 37 5 38 4 26 2 14 LZ Zr 4 4 8 7 1 2 6 ZG (n = 24) Zr 22 2 24 23 1 10 14 HZG Zr 20 13 30 3 32 1 17 16 LZG Zr 12 10 21 1 21 1 8 1 13 ZAG (n = 5) Zr 3 2 4 1 5 5 Av (n = 5) Av 4 1 5 4 1 2 3 AG(n = 10) Av 7 3 10 10 2 8 ZAG (n = 5) Av 3 2 5 1 4 5 Ga(n = 24) Ga 25 4 28 1 26 3 16 1 12 HG Ga 6 4 9 1 10 4 1 5 LG Ga 23 1 1 7 29 1 2 29 1 2 16 2 14 ZG Ga 20 4 22 2 24 15 9 AG Ga 9 1 9 1 9 1 8 2 HZG Ga 21 12 28 5 32 1 10 23 ZG Ga 14 8 17 5 21 1 13 9 AG Ga 4 1 5 5 4 1 * Stage abbreviations: PC = parental stage (i.e., sporocysts or rediae) plus cercariae: P = parental stage only; C = cercariae only; and A = all stages absent. Individual species occurred in the context of several different combinations of infecting species (e.g.. Hg occurred alone, in HZ and HG doubles, and in HZG triples). For each species, frequencies are tabulated for each context. The number (n) of snails infected by particular trematode assemblages is indicated. Assemblages found in fewer than four snails are not tabulated. be included in the above log linear or Kruskal-Wallis analyses. Discussion Ilyanassa obsoleta is the only shared host in the life- cycles of these trematode species and is, therefore, the only place they might directly interact. They are tightly packed together in the snail, gather resources in similar ways, and are abundant on Cape Henlopen. Antagonistic interactions between trematode assemblage members have been noted by several investigators (e.g.. Lie el a!.. 1965: Basch el al, 1969; DeCoursey and Vernberg, 1974; Kuris, 1990; Sousa, 1990). On such grounds we anticipated that trematodes co-occurring in /. obsoleta would interact and most likely compete. A between-snail (component com- munity) analysis indicated that competition within snails was not an important determinant of the number of trem- atodes infecting individual snails. Further, regarding within-snail phenomena, no effect of assemblage com- position on any individual species could be discerned sta- tistically. However. Himasthla quissetensis and Lepo- creadiwn setiferoides were seen to co-occur in this study for the first time (Tables I, V), and this observation de- serves special comment. By virtue of their rare co-occur- rence, which eliminated the pair from our statistical anal- yses, these species apparently do interact negatively when they occur in the same snail. Our sample of trematode assemblages from the Cape Henlopen sandflat naturally included only those species combinations that can coexist long enough to be observed by the methods used. These included most of the possible GASTROPOD-TREMATODE INTERACTIONS 31 Table IV Results l log! i near analyses testing the influence of three factors ttrenuitode Assemblage, snail Section, and parasite Stage) on the frequencies of within snail occurrence reported in Table I' Species analyzed Hierarchical log-linear model accepted x : d.o.f. P =* Hq Assemblage, Section 54.21 45 0.163 x Stage Ls Assemblage, Stage 67.33 57 0.165 Zr Assemblage. Section 63.25 90 0.985 X Stage Av Section x Stage 25.79 32 0.773 Ga Assemblage, Section 60.99 105 0.999 x Stage * An insignificant X : (P > 0.05) without the three-way interaction term means that it is unnecessary; no significant displacement occurred. If a trematode's stages (/.<.. sporocysts. rediae. cercariae) were displaced from one snail section to another by the presence of a co-occurring species or combination of species, the accepted model for that trematode would require the three-way interaction term (i.e.. Assemblage X Section > Stage) to calculate expected frequencies without significant deviation from the observed. assemblages and virtually all that might have been ex- pected to occur. Twenty, of the 32 possible for five species analyzed, were actually observed (Table I). Missing as- semblages were the quintuple, the five quadruples, mul- tiples involving the scarce Amtrobilhania variglandis, and two triples involving Himasthla quissetensis and Lepo- creadiwn seiiferoides. A major concern is whether interparasite competitions occur that require considerable time for completion. In the early to middle phases of competition there may be no noticeable effect on any one species. We may have examined most assemblages at a time when coexistence is possible, and erroneously concluded that species do not interact. If such a time-course for competition is involved, how much time is necessary, and was our collection of parasite assemblages (in snails) biased by this? Two pos- sibilities present themselves: competitions could play themselves out over the summer; or over several summers. There was no indication that trematodes assemble in snails, compete, and ultimately evict subordinate species in either the short or the long interval (Table II). To the contrary, species appear to collect in snails as a function of time. Note that older snails, and not either younger group, have the largest proportion of triple infections (Ta- ble IIB). Sousa (1990) looked for a hyperbolic relationship between snail size and infecting species richness and sim- ilarly did not find one. Direct measurements of within-snail species dynamics also indicate no interactions among assemblage members. The occurrence of parasitic stages in different sections of variously infected snails is shown in Table III. No species was excluded from sections of snails because of co-oc- curring species (lack of Assemblage X Section x Stage terms in Table IV). If one species (or combinations of species) leads to gradual eviction of another species from snails, this phenomenon should have been quite common. Neither was cercarial production (from existing parental stages) of any species shut down by co-occurring species (lack of Assemblage X Stage terms in Table IV). Also, there was no indication that cercarial output from hosts (an estimate of fitness) was influenced by co-occurring species. There was no statistically significant reduction of cercarial output of any species as a function of assemblage Table V Descriptive statistics associated with numbers oftremalode cercariae released per hosl (Ilyanassa obsoleta) in 24 h in the field. Information is grouped bv species of cercariae being tabulated (.see Table I for species abbreviations) Infecting Cercariae trematodes tabulated Mean* S.D. Max. Med. Min. Hq(n = 42) Hq 527 709 2739 225 HL(n = 1) Hq 18 18 18 18 HZ(n = 23) Hq 211 344 1428 90 HG (n = 4) Hq 696 1135 2388 177 42 HZG(n = 22) Hq 155 274 1233 60 Ls(n = 18) Ls 319 590 2394 129 HL(n = 1) Ls 567 567 567 567 LZ(n = 2) Ls 130 185 261 131 LG(n = 10) Ls 42 66 165 9 LZG(n = 15) Ls 121 153 483 45 LAG(n = 2) Ls 18 25 36 18 Zr(n = 12) Zr 249 378 1095 15 HZ(n = 20) Zr 68 199 882 1 LZ(n = 2) Zr ZA(n = 1) Zr 1065 1065 1065 1065 ZG (n = 8) Zr 61 67 189 42 HZGln = 19) Zr 47 73 210 6 LZG(n = 13) Zr 79 138 474 21 ZAG (n = 2) Zr 12 8 18 12 6 Av (n = 3) Av ZA (n = 1) Av AC (n = 5) Av 9 16 36 LAG(n = 3) Av 21 16 33 27 3 ZAG (n = 1 ) Av 6 6 6 6 Gain = 14) Ga 222 454 1398 HG(n = 3) Ga 6 10 18 LG(n = 8) Ga 74 168 483 ZG (n = 6) Ga AG(n = 5) Ga 3 5 12 HZG(n = 17) Ga 2 9 36 LZG(n = 13) Ga 27 96 345 LAG (n = 4) Ga 6 12 24 ZAG(n = 1) Ga 18 18 18 18 Only infections that were patent for the species being tabulated are considered (n). For example, there were 33 HZG-infected snails (from Table I): 22 of these were patent for Hq; 19 for Zr; and 17 for Ga. A total of 206 snails were tested for cercarial release. 32 L. A. CURTIS AND K. M. K. HUBBARD Table VI Ri'Htlts ofKruskal-H 'ulli\ tests I'vuliuitmx the null hypothesis for each trematode species (see Table I for species abbreviations), thai the inimher of cercariat! shed was unaffected hv coeMSlhig species Effect of coexisting species on Infecting Mean rank fitness of species (# cercariae) Kruskal- d.o.f. Wallis H P = Hq Hq(n = 42) 52.583 HZ(n = 23) 41.957 HG (n = 4) 53.500 HZG (n = 22) 36.295 3 6.454 0.09 1 Ls Ls(n = 18) 23.972 LG (n = 10) 15.850 LZGIn = 15) 23.733 2 3.186 0.203 Zr Zr(n= 12) 42.417 HZ(n = 20) 29.575 ZG (n = 8) 42.563 HZG(n = 19) 33.763 LZG(n = 13) 41.962 4 5.254 0.262 Av No test Ga Ga(n = 14) 39.536 LG(n = 8) 39.688 ZG (n = 6) 26.500 AG (n = 5) 38.100 HZG(n = 17) 28.471 LZG(n = 13) 33.537 LAG(n = 4) 34.375 6 8.075 0.233 Prepatent infections and trematode assemblages observed fewer than four times were excluded. composition (Table VI). There was much variation, even in single infections (Table V), suggesting that sources of variability other than co-occurring species control cercarial output. DeCoursey and Vernberg (1974) studied assemblages of trematodes infecting Ilyanassa obsoleta in North and South Carolina. At the level of the component commu- nity, they noted that some species co-occur in multiple infections more or less often than would be expected based on the abundance of each in the system. They proposed that such patterns are produced by antagonisms or affin- ities among assemblage members. About 80 snails were dissected, with 30 of these being serially sectioned. The number of snails examined in each assemblage category is not reported. The authors noted "marked overlap in territory and habitat preferences," as we did in this study. Contrary to our conclusion (based on arbitrary snail sec- tions) that the parasites are not displaced, they concluded that some species are displaced from preferred sites (spe- cific snail organs) by other species. Even if small scale displacements (i.e., from organ to organ within our ar- bitrary sections) do occur, they would have to result in reductions in cercarial output (fitness) to have evolution- ary consequences. Cercarial output was not significantly reduced (Table VI). We also note that, if the interest is in adaptation of one parasite to others (Fig. 1 ), then section- ing snails along snail organ boundaries confounds adap- tation to other parasites with adaptation to the host. In the laboratory, DeCoursey and Vernberg ( 1974) also counted the cercariae released from 10 infected snails. Three were infected with Zoogonus lasius ( = rubellus) and five with Lepocreadium setiferoides. The remaining two were doubly infected with these same species. The num- bers of cercariae released in the laboratory by each species of trematode were averaged and compared. When Z. ru- be/Ins and L. setiferoides occurred alone, they each re- leased approximately 3500 cercariae in 24 h. When the species co-occurred, they released 901 and 1477, respec- tively. The authors concluded that L. setiferoides sup- pressed cercarial release by Z rubellus. Data show that cercarial production of both species was lower when they co-occurred. In any case, the number of observations pre- cludes meaningful statistical inference. We are interested in eliminating inoperative models from the four presented in Figure 1. Williams' (1966) dis- tinction between "functions" and "effects" seems useful here. Functions are biological characteristics that are direct products of natural selection (adaptations), whereas effects are characteristics that are a consequence of functions ("side" effects, not directly selected). Holmes (1986) points out that parasitic ". . . interactions should be important [in structuring helminth communities] only when species regularly co-occur at substantial population densities" (p. 203, brackets ours). We note, more specifically, that interactions based on adaptive responses (functions) of one parasite species to another cannot arise unless there is frequent co-occurrence over global gene pools. We cannot imagine how the parasites under study here could have adapted to one another. Definitive hosts (fish and birds) are highly mobile and scatter parasite eggs widely and unevenly. Consequently, spatial distribution of these trematodes within and among host snail popu- lations is patchy (Curtis and Hurd, 1 983; Curtis and Hub- bard, 1990), and there are abundant opportunities for trematode species to exist in isolation. The probability of co-occurrence generation after generation, particularly for specific parasites, is very low. Therefore, evolved parasite- parasite relationships are unlikely in this system. If inter- actions occur, they most likely result from effects, not functions. Our data indicating the lack of interactions among the majority of co-occurring trematodes. and the above considerations, justify eliminating models "c" and "d" (Fig. 1 ). Any evolved features of this system probably stem ultimately from the evolution of parasites to host, or possibly of host to parasites (models "a" and "b". Fig. 1 ). In deciding between models "a" and "b", many of the same arguments apply. Gooch et al. (1972) found that GASTROPOD-TREMATODE INTERACTIONS 33 Ilvanassa obsolete! were electrophoretically homogeneous all along the eastern seaboard, pointing to extensive dis- persal of larvae as the main cause. The planktonic larvae of/, obsoleta would then function analogously to parasite definitive hosts in the dispersal of progeny. Given the het- erogeneity of trematode prevalence in /. obsoleta popu- lations, many snail larvae would settle where parasites are not a frequent environmental challenge. If a snail were to obtain, by mutation, resistance to infection by one or more trematode species, its fitness probably would be en- hanced in parasite-ridden environments such as parts of Cape Henlopen. Yet its progeny would very possibly settle where parasites are infrequent. The mutation, there, would be at best neutral. These considerations suggest that model "a" (Fig. 1) is the operative one the only adaptive re- sponses between species in the /. obsoleta system are most likely those of the parasites to the host. The negative interaction between Hiniasthla quisse- lensis and Lepocreadium setiferoides in the Ilyanassa ob- soleta system deserves comment because it seems to counter the proposition that these parasites are not adapted to each other. A lack of co-occurrence of these abundant species in / obsoleta has been reported (Vern- bergetal., 1969; Curtis. 1985), but the detailed dissection methods used in this study revealed four co-occurrences (Table I). Obviously, miracidia of both species reach the same host, and there is a subsequent eviction (apparently of H. quissetensis). This eviction is important in terms of determining composition of the infra- and component assemblages observed, but is it based on adaptation of one parasite to another? In keeping with the above reasoning that parasite co-occurrence is not globally predictable enough to result in adaptations to other parasites we interpret this negative co-occurrence as based on an effect rather than a function [an exaptation (Gould and Vbra, 1982)] because it results from the way these species have evolved to the host, not to each other. In ecological terms, such a phenomenon is a competitive exclusion. However, in our hypothesis, the exclusion occurs between two spe- cies that are adaptively unaware of each other. If species interactions are an evolutionary force driving the struc- turing of interactive, co-adapted species assemblages, then we should distinguish between function- and effect-based relationships among species. A deeper appreciation of causal relationships in ecological systems will require un- derstanding these relationships. Factors structuring the assemblage of larval trematodes in populations of the California estuarine snail Cerithidea californica have been examined by Sousa (1990) and Kuris (1990). Two direct lines of evidence convinced these au- thors that competitive exclusions were occurring among C. californica trematodes. Sousa (1990) cites personal laboratory observations in which dominant species preyed upon stages of subordinates. Both authors reported trem- atode species replacements in individual snails periodically reexamined for infection by cercarial release. Kuris (1990) constructed a competitive hierarchy among trematode species in infracommunities, which he concluded would produce component community structure. In the Ily- anassa obsoleta system, such cercarial release data would have to be used judiciously because cercariae, even if present, often are not shed (Curtis and Hubbard, 1990). Data are not presented that assess this source of error for the C. californica system. In any event, there is consid- erable heterogeneity in prevalence of trematodes among C. californica populations (Kuris, 1990; Sousa, 1990). Parasite progeny are dispersed by definitive hosts similar to those in the I. obsoleta system, giving species the same opportunities to exist in isolation. This may mean that, whether they interact or not, parasites are not co-adapted in the C. californica system either. Because C. californica has direct development (Sousa, 1990) making popula- tions more insular the host may have the ability to evolve to its parasites. Several authors have examined snail-trematode sys- tems for interactions among parasites infecting the same host individual. Some have emphasized direct micro- scopical observations of antagonisms occurring in fresh- water snails (e.g.. Lie el al, 1965; Basch el al. 1969; Mouahid and Mone, 1990). Based upon such observa- tions, there can be no doubt that antagonisms between trematodes can and do occur, but their frequencies in natural snail populations are less certain. Other authors have emphasized observations of multiple infections in marine (e.g., Kuris, 1990; Sousa, 1990) and freshwater (e.g.. Fernandez and Esch, 199 la, b; Williams and Esch, 1991 (gastropods. In no case are multispecies assemblages reported to be particularly frequent. Such species-rich as- semblages are more frequent and various in the Ilyanassa obsoleta system on Cape Henlopen (Curtis, 1985, 1987, 1990; present study) than in any studied so far (see Cort el al. ( 1937)). The most frequent assemblage observed on Cape Henlopen is Lepocreadium setiferoides with Gy- naecotyla adiinca. and it occurred in only 4.4% of snails (n = 4870) examined by dissection (Curtis, unpub. data). Individual occurrence of each was 16.9 and 20.3%, re- spectively. Thus, even when species can and do co-occur, the probability of co-occurrence is slight. The opportunity to evolve adaptive responses to other particular trematodes seems minimal or nonexistent, which suggests that models "c" and "d" (Fig. 1 ) may be generally inoperative. The best opportunity for trematode-trematode adaptive re- sponses would be in a situation where all the necessary hosts are confined to one habitat, such as in a freshwater pond, as described by Williams and Esch ( 1 99 1 ) and Fer- nandez and Esch (199 la). However, Williams and Esch (1991) and Fernandez and Esch (1991b) conclude that within-snail trematode interactions in their system are in- 34 L. A. CURTIS AND R. M. K HUBBARD frequent and not the factor structuring the infra- and component communities. Can the host be adapted to its parasites? The evolution of a host to several parasites is a problem of "overwhelm- ing complexity" (McLennan and Brooks, 1991). and the issue is not resolvable with the data at hand. Dobson and Merenlender ( 199 1 ) suggest, as we content here, that the probability of such evolutionary responses would depend on host and parasite dispersal abilities, llyanassa obsoleta. because of its widespread dispersal, is unlikely to evolve to its parasites (model "b"), but it is a possibility with a snail in a more insular system, such as a pond. How can the coexistence, in a small habitat unit, of several species with similar resource requirements be ex- plained? This study has provided considerable compar- ative data on the fitness of parasites when they occur in different assemblages. The extensive variation in cercarial output (Table V) is not explainable by looking to presence or absence of other species. Perhaps resources for trem- atodes living in llyanassa obsoleta are somehow not lim- iting. We have suggested that the only adaptations (func- tions) in the system are those of the parasites enabling them to live in the snail (model "a". Fig. 1 ). We offer the following possible explanation. Each of these five trem- todes has evolved to castrate the host snail. Castration of the host stems from a parasite adaptation to channel en- ergy to the parasite that would otherwise go to the support of host gonadal tissue (Baudoin, 1975). llyanassa obsoleta is a long-lived host (7 years or more); the largest (oldest) snails are nearly all parasitized where trematodes are prevalent; and they appear not to lose infections (Curtis and Hurd, 1983). The host must survive the rigors of suc- ceeding winters. A trematode adapted to such a host may have been selected to exact intermediate to minimal damage (besides castration) because it could then "farm" the host for many years (see Minchella et al. (1985) and Gill and Mock (1985) for similar interpretations of host- parasite systems). We propose that, if the trematodes of /. obsoleta operate this way, then they should not singly, or in multiples, drain resources to the extent that they become limiting. In brief, they can coexist if they are all adapted to live well below the level at which the host is stressed. Acknowledgments We would like to thank the Undergraduate Research Office and the School of Life Sciences, University of Del- aware for a Science and Engineering Scholar grant, and two Peter White Fellowships awarded to support K. H.'s undergraduate thesis, from which this paper is adapted. We thank J. Moore for helpful comments on an earlier version. We are also grateful for the efforts and comments of two anonymous reviewers. This is contribution number 173 from the Program in Ecology, School of Life Sciences. Literature Cited Basch, P. K., K. J. Lie, and D. Heyneman. 1969. Antagonistic inter- action between strigeid and schistosome sporocysts within a snail host. J. Punisiiol. 55: 753-758. Baudoin, M. 1975. Host castration as a parasitic strategy. Evolution 29: 335-352. Curtis, L. A. 1985. The influence of sex and trematode parasites on carrion response of the estuarine snail llyanassa obsoleta. Biol. Bull. 169: 377-390. Curtis, L. A. 1987. Vertical distribution of an estuarine snail altered by a parasite. Science 235: 1 509- 1511. Curtis, L. A. 1990. Parasitism and the movements of intertidal gastro- pod individuals. Biol. Bull 179: 105-1 12. Curtis, L. A., and K. M. Hubbard. 1990. Trematode infections in a gastropod host misrepresented by observing shed cercanae. ./. IC.\[t Mar Biol Ecu/ 1-43: 131-137. Curtis, L. A., and L. E. Hurd. 1983. Age. sex and parasites: spatial heterogeneity in a sandflat population of llyanassa obsoleta. Ecology 64: 819-828. Cort, VV. W., D. B. McMullen, and S. Bracken. 1937. Ecological studies on the cercariae in Stugnicolu cmar^inala angulata (Sowerby) in the Douglas Lake region. Michigan. ./ Parusilol. 23: 504-532. DeCoursey, P. J., and \V. B. Vernberg. 1974. Double infections of larval trematodes: competitive interactions. Pp. 93-109 in Symbiosis ;/; l/ie Sea. W. B. Vernberg. ed. University of South Carolina Press. Columbia. SC. Dobson, A. P., and A. Merenlender. 1991. Coevolution of macropar- asites and their hosts. Pp. 83-101 in Parasite-Host Associations Co- existence or Conflict? C. A. Toft, A. Aeschlimann, and L. Bolis. eds. Oxford University Press. New York. Esch, G. W., A. W. Shostak, D. J. Marcogliese, and T. M. Goater. 1990. Patterns and processes in helminth parasite communities: an overview. Pp. 1-19 in Parasite Communities: Patterns and Processes, C. Esch. A. Bush, and J. Aho, eds. Chapman and Hall. London. Fernandez, J., and G. \V. Esch. 1991a. Guild structure of larval trem- atodes in the snail Helisoma anceps: patterns and processes at the individual host level. / Parasitol. 77: 528-539. Fernandez, J., and G. \V. Esch. 1991 b. The component community structure of larval trematodes in the pulmonate snail Helisoma anccps J Panisiiol 11: 540-550. Futuyma, D. J., and M. Slatkin. 1983. Introduction. Pp. 1-13 in Co- solution. D. J. Futuyma and M. Slatkin. eds. Sinauer Associates. Sundcrland, MA. Gill, D. E., and B. A. Mock. 1985. Ecological and evolutionary dy- namics of parasites: the case of Trypanosnma diemyctyli in the red- spotted newt Nolophthalmus viridescens. Pp. 157-183 in Ecology and Genetics of Host-Parasite Interactions, D. Rollinson and R. M. Anderson, eds. Academic Press, New York. Gooch, J. L., B. S. Smith, and D. knupp. 1972. Regional survey of gene frequencies in the mud snail Nassarius obsoletus. Biol. Bull. 142: 36-48. Gould, S. J., and E. Vbra. 1982. Exaptation a missing term in the science of form. PuleohiologyS: 4-15. Hollander, M., and D. A. Wolfe. 1973. Nonparametric Statistical Methods. John Wiiey and Sons. New York. Holmes, J. C. 1986. The structure of helminth communities. Pp. 203- 208 in PurasilologyQuo vadil^. M. J. Howell, ed. Proceedings 6th International Congress of Parasitology, Australian Academy of Sci- ences, Canberra. Kuris, A. 1990. Guild structure of larval trematodes in molluscan hosts: prevalence, dominance and significance in competition. Pp. 69-100 in Parasite Communities: Patterns and Processes, G. Esch, A. Bush, and J. Aho, eds. Chapman and Hall, London. GASTROPOD-TREMATODE INTERACTIONS 35 Lie, K. J., P. F. Basch, and T. Umathevy. 1965. Antagonism between two species of larval trematodes in the same snail. Nature 206: 422-423. McLennan, D. A., and D. R. Brooks. 1991. Parasites and sexual selection: a macroevolutionary perspective. Q Rev Biol 66: 255-286. Minchella, D. J., B. K. Leathers, K. M. Brown, J. M. McNair. 1985. Host and parasite counteradaptations: an example from a freshwater snail. Am Nat. 126: 843-854. Moore, J. 1987. Some roles of parasitic helminths in trophic interac- tions. A view from North America. Revixta Chilena de Hisloria Nat- ural 60: 159- 179. Mouahid, A., and H. Mone. 1990. Interference of Echinoparyphium elegans with the host-parasite system Bulmus truncatus-Schistosoma bovis in natural conditions. Ann. Trop. Med. Parasitol. 84: 341-348. Rohde, K. 1981. Population dynamics of two snail species, Planaxis sulcatus and Cerithium moniliferum, and their trematode species at Heron Island, Great Barrier Reef. Oecologia 49: 344-352. Smyth, J. D., and D. W. Halton. 1983. The Physiology of Trematodes. 2nd ed. Cambridge University Press, Cambridge. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, 2nd ed. Freeman and Company, New York. Sousa, \V. P. 1990. Spatial scale and the processes structuring a guild of larval trematode parasites. Pp. 4 1 -67 in Parasite Communities Patterns and Processes. G. Esch, A. Bush, and J. Aho. eds. Chapman and Hall, London. Stunkard, H. \V. 1983. The marine cercariae of the Woods Hole, Massachusetts region, a review and a revision. Biol. Bull. 164: 143-162. Vernberg, \V. B., F. J. Vernberg, and F. W. Beckerdite. 1969. Larval trematodes: double infections in the common mud-flat snail. Seienee 167: 1287. Williams, G. C. 1966. Adaptation and Natural Selection. Princeton University Press, Princeton, NJ. Williams, J. A., and G. \V. Esch. 1991. Infra- and component com- munity dynamics in the pulmonate snail Helisoma anceps. with spe- cial emphasis on the hemmnd trematode Ha/ipegus occidualis. J Parasitol. 77: 246-253. Reference: Biol. Bull 184: 36-51. (February, 1993) Effects of Marine Bacteria on the Culture of Axenic Oyster Crassostrea gigas (Thunberg) Larvae PHILIPPE DOUILLET 1 AND CHRISTOPHER J. LANGDON Oregon State University, Department of Fisheries, Hatfield Marine Science Center, Newport, Oregon 97365 Abstract. Bacteria-free oyster larvae ( Crassostrea gigas) were cultured under aseptic conditions; they were fed axenic algae (Isochrysis ga/bana), and the medium was inoculated with isolated strains of marine bacteria. Twenty-one bacterial strains were tested, and most were detrimental to larval survival and growth. However, ad- ditions of strain CA2 consistently enhanced larval survival (21-22%) and growth (16-21%) in comparison with con- trol cultures that were fed only algae. Size-frequency dis- tributions of populations of larvae cultured for 10 days on axenic algae were skewed due to the poor growth of many individuals; whereas size-frequencies from popu- lations of larvae fed axenic algae supplemented with CA2 bacteria were distributed normally. Strain CA2 may therefore make a nutritional contribution to the growth of oyster larvae. /. galbana did not grow under the light intensities used for larval culture; thus the improvement in larval growth cannot be attributed to bacterial en- hancement of algal growth and, consequently, food avail- ability. Naturally occurring microflora from Yaquina Bay, Oregon, depressed survival or growth of larvae-fed live algae. Introduction Bivalve larvae in culture vary substantially in survival and growth (Davis, 1953;Loosanoff, 1954;Walne, 1956a). Between 25% and 50% of the variability in the growth of a single population of mussel larvae (Innes and Haley, 1977), or different populations of larval Crassostrea vir- ginica (Newkirk el al, 1977), are due to genetic factors. A significant proportion of the variability in the survival of C. gigas larvae was similarly attributed to genetic factors Received 3 June 1992; accepted 10 November 1992. ' Present address: The University of Texas at Austin, Marine Science Institute, P.O. Box 1267, Port Aransas. Texas 78373. (Lannan, 1980). Exogenous factors, such as temperature (Loosanoff, 1959), salinity (Bayne, 1965), pH (Calabrese and Davis, 1970), food quantity (Walne, 1965), food quality (Davis, 1953), age of the algal food (Dupuy, 1975), larval concentration (Loosanoff et al., 1953), size of con- tainer (Dupuy, 1975), silt (Davis and Hidu, 1969), exu- dates of unfavorable algal species (Bayne, 1965), water quality (Millar and Scott, 1967) and toxicants (Walne, 1970) have been found to contribute significantly to vari- ability in larval growth. Nonetheless, even different cul- tures of larvae obtained from the same parents and grown under identical conditions of temperature, salinity and ration have been commonly reported to vary in their growth (Bayne, 1983). The role of bacteria as beneficial or harmful agents in the culture of bivalve larvae has been the subject of many investigations, but this role has not been fully evaluated. Thirteen different isolates of marine bacteria did not sup- port the growth of oyster larvae when provided as the sole source of paniculate food (Davis, 1950, 1953). High bac- terial densities in cultures of bivalve larvae are generally considered to be deleterious to the larvae (Walne, 1956a, 1956b, 1958), and even innocuous bacteria in large num- bers have been reported to depress the rate of algal inges- tion (Ukeles and Sweeney, 1969). Some bacterial strains are reportedly able to invade larvae, to produce toxins, or both (Guillard, 1959; Tubiash et al.. 1965; Tubiash et al., 1970; Brown, 1973; Di Salvo, 1978; Nottage and Birkbeck, 1986). In contrast, bacteria have also been im- plicated as a food source for bivalve larvae (Carriker, 1956; Hidu and Tubiash, 1963) or as improving the growth of larvae fed on algae (Martin and Mengus, 1977; Beese, in Prieur et al., 1990). The elimination of microbial contaminants is prereq- uisite to a study of the effects of a bacterial strain on an organism in culture. This approach has been used to study the effects of several bacterial strains on cultures of the 36 BACTERIAL EFFECTS ON OYSTER LARVAE 37 protozoan Amoeba nitrophila (Frosch, 1897 in Luck et al., 1931); the cladoceran Moina macrocopa (Stuart et ai, 1931); and larvae of the clam Mercenaria mercenaria (Guillard, 1959). In the present study, axenic larval Crassostrea gigas, obtained without the use of antibiotics, were used in a series of experiments meant to reveal whether selected strains of marine bacteria can consistently improve the survival and growth of algal-fed oyster larvae. Materials and Methods Maintenance of larvae, bacteria and algae Bacteria-free oyster larvae were obtained according to the method of Langdon ( 1983). Adult oysters Crassostrea gigas were held at 1 8C in a recirculating seawater system for a period of 4 to 6 weeks, depending on the initial reproductive condition of the broodstock. After this con- ditioning period, the oysters were opened and shucked. Using aseptic techniques in a laminar-flow hood, we dis- infected the external surface of the gonads of each oyster with a 1% solution of sodium hypochlorite. A small in- cision was made through the surface of the gonads with a heat-sterilized scalpel, and gametes from each oyster were removed with sterile Pasteur pipettes and transferred to separate sterile flasks containing 0.2 ^m-filtered, au- toclaved seawater (FSSW). Eggs were fertilized by the ad- dition of a few drops of sperm suspension and then were transferred to Erlenmeyer flasks containing FSSW at a density of 1 00 eggs ml ' . Eggs were incubated on an orbital shaker at 25 C for 48 h. When the trocophore larvae had developed into veligers (straight-hinged larvae), subsam- ples of larvae were aseptically withdrawn for axenicity tests, and the remaining larvae were then held at 5C for 5 days. Axenicity of larvae was determined by epifluo- rescence microscopy using 4'6-diamidino-2-phenylindole (DAPI) staining techniques (Porter and Feig, 1980). Sam- ples of larvae were also added to 1/10 recommended con- centration of Difco marine broth 2216 (3.74gr', salinity 30 ppt) and incubated at 25C under aerobic or anaerobic conditions (BBL GasPak Pouch). Larvae from cultures that showed no evidence of microbial contamination from either the epifluorescence test or the 5 day broth incu- bations were considered adequate for experimentation. To confirm that the larvae were axenic, broth incubations were continued for 30 days. Axenic straight-hinged larvae were transferred to 250 ml Erlenmeyer flasks, each con- taining 1 50 ml of FSSW, closed with cotton plugs and capped with aluminum foil. Final larval density was 5 ml" 1 . Growth experiments were then initiated by the addition to the culture flasks of axenic algae and the dif- ferent bacterial strains. Shell lengths of 100 randomly se- lected larvae were measured, either with an optical mi- crometer fitted to a compound microscope, or with an image analysis system (Zeiss Videoplan 2). Strains of marine bacteria were isolated from cultures of algae or oyster larvae at the Whiskey Creek Hatchery in Netarts Bay, Oregon. Other bacteria were isolated, ei- ther from the guts of adult oysters, or from incubations of protein capsules (Langdon, 1989) suspended in unfil- tered seawater. Pure bacterial strains were obtained by the dilution method of Rodina (1972). Strains were grown, at 25C, on marine agar 22 16 or brain heart infusion agar (Difco). Bacteria grown on such solid media for 3 to 5 days were resuspended for 24 h in FSSW; they were then washed by centrifugation at 20,000 X g for 10 min and resuspended in FSSW. Strains were added to larval cultures at concentrations of 10 5 -10 6 cells ml" 1 . Cell concentrations were derived from equations relating spectrophotometric absorbance (600 nm) and bacterial concentration; the latter value was determined by direct count after staining with DAPI (Porter and Feig, 1980). Such equations were developed and used for each strain tested. Axenic Isochrysis galbana Parke (clone ISO) was ob- tained from the Culture Collection of Marine Phyto- plankton (Maine). Algal cultures were grown at 20C in 200 ml f/2 medium (Guillard and Ryther, 1962) illumi- nated by 1000-1500 lux of cool white fluorescent light under a 12 h light/ 12 h dark photoperiod. Algal axenicity was determined as described above for larvae. All glassware was washed in 10% nitric acid, rinsed seven times with distilled water, and baked overnight at 450C. Disodium ethylenediamine-tetraacetate (EDTA) was added at a final concentration of 1 ppm to all seawater to reduce the load of dissolved organic matter (Utting and Helm, 1985). Salinity of seawater after sterilization varied between 28 and 3 1 ppt. Heat sterilization was carried out for 15 min at 121C and 1.06 kg cm 2 pressure. Larvae fed on live algae and bacteria Twenty-one marine bacterial isolates were tested in three culture experiments for their effects on the survival and growth of larvae fed axenic Isochrysis galbana. Ex- periment I included seven microbial isolates from the Whiskey Creek Hatchery (H1-H7) and five isolates from the guts of adult oysters (G1-G5). Control treatments were either larvae fed only algae or starved larvae. In Experiment II, two strains (H6, H7) that improved larval growth in Experiment I were tested along with five strains isolated from the Whiskey Creek Hatchery (H8- H12), one strain isolated from the gut of an adult oyster (G6), and three strains isolated from protein capsules in- cubated in seawater (CA1-CA3). Control treatments in- cluded starved larvae and larvae fed only algae. In third control (SW), cultures of larvae were inoculated at the beginning of the experiment with naturally occurring bacteria present in 5 ml samples of 1 ^m-filtered seawater collected from Yaquina Bay, Oregon. The larvae in the 38 P. DOUILLET AND C. J. LANGDON third control treatment were fed axenic algae every other day. Experiments I and II were carried out with four rep- licates per treatment. Experiment III was designed to retest strains that had enhanced larval survival and growth in Experiment II (H7, CA2). Control treatments similar to those described for Experiment II were included. Experiment III was carried out with eight replicates per treatment. Cultures of bacteria-free oyster larvae (75.5-82 ^m shell length) were inoculated once at the beginning of each ex- periment with bacterial strains. Bacteria-free algal cells, harvested from cultures in exponential growth phase, were added to the larval cultures every two days. The seawater of the larval cultures was not renewed during the culture period. The concentration of algal cells in each larval cul- ture flask was estimated, as follows, before each feeding. A 2-ml sample of the larval culture medium was asepti- cally removed from each flask with a pipet; to prevent removal of larvae, the end of the pipet was covered with a 64 j/m Nitex screen. Algal cells were preserved with formalin, concentrated by centrifugation, and re-sus- pended in 100 jul of 0.2 ^m-nltered seawater. Algal con- centrations in the samples were then determined with a hemocytometer. Fresh algae were then added to larval culture flasks to provide cell concentrations at pre-deter- mined levels. Algal cell concentrations were increased by 15,000 cells ml ', from 40,000 to 100,000 cells ml" 1 over a 10 day culture period. To provide uniform food quality during the experiments, algae from a single culture were added at each feeding period, to all larval cultures receiving an algal diet. Larval culture flasks were placed randomly on orbital shakers in a temperature-controlled room at 25C. The cultures were exposed to a light intensity of 50-70 lux for 12 h each day. No algal growth occurred at this low light intensity. After 10 days of culture, samples of water were aseptically withdrawn from flasks containing starved lar- vae or larvae fed only axenic algae; these samples were analyzed for microbial contamination as described above. The experimental data were analyzed only if these control treatments were bacteria-free at the end of the 10 day culture period. Effects ofCAl bacteria on the growth of algae in tan-al cultures Cells of axenic /. galhana were initially suspended at a concentration of 40,000 ml~' in f/2 medium and then subdivided in sixteen 250 ml Erlenmeyer flasks. CA2 cells were added at 10 5 cells ml ' (final concentration) to eight flasks, while FSSW was added to the other eight flasks to maintain similar initial algal concentrations in all flasks. The final volume of each algal culture was 200 ml. Four algal cultures inoculated with bacteria and four cultures that had received only FSSW were placed in conditions conducive to the growth of I. galbana (1000-1500 lux and 20C); the remaining algal cultures were exposed to the conditions used for larval culture (50-70 lux and 25 C). The algal cultures were incubated on orbital shak- ers for three weeks. Every second day, 10 ml samples were removed aseptically from each algal culture, and algal concentrations determined with a Coulter counter (Mo- del ZB1). Larvae fed on dead algae and bacteria Interactions between strain CA2 and living Isochrvsis galbana that could modify algal food quality were not addressed in the previous experiments. To determine whether bacteria could enhance cultures of larvae fed on non-living diets, live /. galbana were replaced with dead algae. In Experiment IV, known concentrations of axenic /. galbana were frozen at -5C. Freezing and thawing broke the cell walls and membranes of the algal cells. Larvae were fed dead freeze-killed algae (FA) every two days ac- cording to the same protocol used with live algae. One group of larval cultures fed FA was maintained bacteria- free, and two groups were inoculated at the beginning of the experiment with either strain H6 at 10 5 cells ml" 1 (final concentration), or with an inoculum of naturally occurring bacteria (SW). The wild strains were added in 5 ml samples of 1 /jm-nltered seawater collected from Yaquina Bay, Oregon, at a concentration of 10 5 -10 6 cells ml" 1 . Other larval cultures received on alternate days, ei- ther additions of strain H6 (at a final concentration of 10 ? cells ml ') alone, or naturally occurring bacteria (SW) (5 ml of 1 ^m-nltered seawater) alone. Control treatments included starved larvae and larvae fed every second day on live axenic /. galbana. Culture conditions and sample treatments were similar to those of experiments carried out with live algae. Four replicates were tested per treat- ment. Algal cells were also killed by w 'Co-irradiation (5 me- garads) at the Radiation Center at Oregon State Univer- sity. Non-viability of irradiated algae (IA) was evident by the lack of growth of cells in f/2 medium at 20C under 1000-1500 lux of fluorescent light emitted 12 h a day. The irradiation process also destroyed contaminants, as demonstrated by incubations, at 25 C, of irradiated algae in 1/10 diluted marine broth 2216 (3.74 g 1~', salinity of 30 ppt) under either aerobic or anaerobic conditions (BBL GasPak Pouch). The integrity of the irradiated algal cells was verified by microscopic examination. Cell volumes of irradiated and non-irradiated algae from seven different cultures were determined with a Coulter counter (Model ZB1) equipped with a calibrated Coulter channelyser (Model 256). To ensure that IA were acceptable to larvae as a food source, the ingestion rates of larvae fed on either IA or BACTERIAL EFFECTS ON OYSTER LARVAE 39 live /.SW///T.V/.V galbana were compared. Ingestion rates were calculated according to the methods described by Checkley (1980). Larval ingestion rates for live and ''"Co- irradiated algae were compared with a 2 sample t-test, after verifying homocedasticity by Cochran's test for ho- mogeneity of variances at the 0.05 level of probability (Douillet, 1991). In Experiment V, oyster larvae were fed IA every second day according to the methods employed with live algae in Experiments I to III. Three groups of larval cultures were fed IA. One group was maintained bacteria-free, while the two others were inoculated at the beginning of the experiment with strains H7 or CA2. Control treat- ments included starved larvae or larvae fed every two days on live axenic Isocfuysis galbana. Eight replicates were tested per treatment. Larval survival and growth were de- termined as described below. Data collection and analysis At the end of each experiment, the larvae were carefully transferred to scintillation vials containing buffered form- aldehyde (2% final concentration, pH = 8). The larval tissues were stained with rose of Bengal, so that the larvae that were alive could be distinguished from empty shells. The whole larval population in each flask was counted with a dissecting microscope, and the shell lengths of 100 randomly selected larvae were measured, either with an optical micrometer fitted to a compound microscope, or with an image analysis system (Zeiss Videoplan 2). Sur- vival and growth data were transformed to satisfy as- sumptions of ANOVA. Survival data were transformed as: arcsin (square root (percent survival 100 ')) Growth data were transformed as: arcsin (square root ((In L, - In L,,)t ')) where L, is the final mean shell length (^m); L u is the initial mean shell length (^m); and t is the culture period (10 days). These transformations were successful in reducing the heterocedasticity of the survival data but not of the growth data (Cochran's test for heterogeneity of variances, at the 0.05 level of probability). Treatment effects on larval sur- vival were tested with one-way ANOVA. Where significant differences were indicated, Tukey's honestly significant difference test (T-HSD) was applied to determine the sta- tistical significance of differences among individual treat- ments at the 0.05 level of probability. Treatment effects on larval growth were analyzed with the Kruskal-Wallis test (KW). Differences among individual treatments were determined by means of the Games and Howell test (G&H) of equality of means with heterogeneous variances (Sokal and Rohlf, 1981), at the 0.05 level of probability. All tests were performed with the computer program Sta- tistix (NH Analytical Software), except the Games and Howell test which was carried out with the program Biom (Rohlf, 1982). The size-frequency distributions of populations of algae- fed larvae that were bacteria-free were compared with those fed algae supplemented with CA2 bacteria in Ex- periments II and III. Skewness coefficients (gl : Sokal and Rohlf, 1981) of larval populations from each replicate flask were calculated and used to compare larval size fre- quency distributions. A normal size distribution would have a gl coefficient equal to 0. A skewness coefficient higher than indicates that the size distribution is posi- tively skewed (higher proportion of small-sized individ- uals), while a coefficient smaller than indicates negative skewness. After confirmation of homocedasticity of gl values by Cochran's test at the 0.05 probability level, data were analyzed by two-way ANOVA with treatment (algae, algae + CA2) and experiment as factors. As dictated by the results of ANOVA, appropriate multiple comparisons of means were conducted at the 0.05 level of probability using the Student-Newman-Keuls procedure (SNK), controlling for experiment-wide error (Underwood, 198 1 ). Cryopreservation ot bacteria Bacteria have been described as adaptable chimaeras, the metabolic plasticity of which results from widespread transfer of genetic information though plasmids or pro- phages (Sonea. 1988). This strategy for adaptation to changing environments may result, during evolution, in the loss of beneficial characteristics of selected bacterial strains. In order to reduce the possibility of changes in bacterial characteristics between successive experiments, selected strains were cryopreserved at -70C in 10% (V/V) glycerol in sterile 1/10 diluted marine broth 2216. Identification of strain CA2 The identification of bacterial strain CA2 was based on Bergey 's Manual of Systematic Bacteriology (Holt, 1 984). The methodology used for different procedures followed the Manual of Methods of General Bacteriology (Gerhardt et ai, 1981). Exponentially growing cells cultured on ma- rine agar 22 1 6 were used for the following tests performed at the Hatfield Marine Science Center, Newport, Oregon, (a) Cells were Gram stained, (b) Motility was determined by observations of wet mounts with light microscopy, (c) Oxidase activity was determined by spreading CA2 cells with sterile cotton swabs over Pathotec cytochrome oxi- dase test strips (General Diagnostics), which contained a derivative of dimethyl-p-phenylenediamine and -naph- thol. (d) Cultures of CA2 cells were flooded with 3% hy- drogen peroxide for catalase testing, (e) Oxidation and fermentation of glucose was assayed with the modified O-F medium of Leifson (1963). (0 Utilization of inorganic ORIGIN OF BACTERIA ADDED TO LARVAL CULTURES m H rcn c Q on I > cc Z) A 100- r\ 80- / / / 1 60- / f / / / 1 X 7 y 1 ' / / / X / . / . / / ' X 40- J, / ', ^ ; / / / ; A X X X X X 20- / / ' / ^ / / X X 1 X / y / / ^ / / a.. X X M X - r 1 , y / / / / / / / M v X H X V STARVED ALGAE H1 H2 " 2 H4 H5 H6 H7 G1 G2 G3 G4 G5 ALGAE + BACTERIA 155 in 145- E 135- O 125- 1 15- 105- T JL I / STARVED ALGAE HI H2 H3 H4 H5 H6 H7 Gl G2 G3 G4 G5 ALGAE + BACTERIA Figure I. Effects of different bacterial strains on oyster larvae cultured on a diet of axenic Isochrrsis galhana for 10 days. (A) Survival in Experiment I. (B) Survival in Experiment II. (C) Growth in Experiment I. (D) Growth in Experiment II. Bacteria were isolated from the Whiskey Creek Hatchery. Oregon (H). from the guts of adult oysters (G), from incubations of protein capsules in seawater (CA), or were naturally- occurring in 1 nm-nltered seawater (SW). Larval control treatments were starved or fed axenic /. galbana. Results of Tukey's HSD pairwise comparisons and Games and Howell's tests are displayed below the his- tograms of survival and growth, respectively. Squares that occur together on any one of the horizontal lines indicate mean values that are not different at the 0.05% level of significance. 40 BACTERIAL EFFECTS ON OYSTER LARVAE 41 ORIGIN OF BACTERIA ADDED TO LARVAL CULTURES KS Sw I~TI H GZ) C E\S CA 100- B 1 1 I 1 1 oo 80- / / 1 / f X A ;' 6 - > / 1 / 1 ^ 1 X X | / / / / . X v ^ > y / ' / ./ ' ~x V X s^ 40- co i ^ / / / ^ ', / x x x 1 20- X / / '. V, / / / ~x ^ ^ N v \ - i 1 ^ ^ / / / / / >(. X 1 1 1 cTiOv/irn !r.r SW H6 H7 H8 H9 HlO H1 1 HI 2 G6 CA1 CA2 CA3 ALGAE + BACTERIA o CO E X I o I CO 75 150- 125- iOO- 75 I I I STARVED ALGAE SW H6 H7 H8 H9 HlO H11 H12 G6 CA1 CA2 CA3 ALGAE BACTERIA Figure 1. (CiiHtinued) 42 P. DOUILLET AND C. J. LANGDON sources of nitrogen was evaluated by culturing CA2 cells on media prepared with NH 4 C1 or NaNO, (0.5 g 1 '), glucose (0.1 g 1 '), Na 2 HPO 4 (0.1 g 1 '), FePO 4 (0.004 g 1 '(and 1 ml 1 ' off/2 vitamin mix (Guillard and Ryther, 1962). The culture media used as controls were prepared by replacing NaNO, or NH 4 C1 with peptone or tryptone (Difco) at 0.5 g 1 '. (g) Anaerobic growth was determined by transferring CA2 cells either into solid media in Petri dishes, or into 25 ml 1/10 diluted marine broth 2216 (3.74 gT 1 ; salinity 30 ppt) contained in 50 ml Erlenmeyer flasks, placing these cultures in anaerobic GasPak pouches (BBL), and incubating the cells at 20C for up to one month. The following tests were carried out by Dr. Ronald Weiner (University of Maryland at College Park). Meth- odology followed the Manual of Methods for General Bacteriology (Gerhardt et ai, 1 98 1 ). (a) Salt requirements were evaluated by culturing CA2 cells in tryptic soy agar (TSA) prepared at different salt concentrations; NaCl was added at 1% increments up to 10% of the control level. (b) As evidence of anaerobic growth and motility, CA2 cells on a straight needle were used to inoculate a tube containing semisolid tryptic soy broth enriched with 0.8% agar and 1%. NaCl, and the pattern of growth observed. (c) Flagellar staining was carried out by the Leifson method (Gerhardt et ai, 1981). (d) Synthesis of exopoly- saccharides was evaluated by the phenol-sulfuric acid re- action (Gerhardt et ai, 198 1 ). (e) The mole percent gua- nine plus cytosine (mol% G + C) in extracted deoxyri- bonucleic acid (DNA) was determined by the thermal melting (denaturation) methods of Marmur and Doty ( 1962) with a Gilford UV programmable spectrophotom- eter. (f) Antibodies of 20 different bacteria strains be- longing to the Alteromonas/Shewanella group were tested for reaction with exopolysaccharides of CA2 cells, (g) Fatty acid analyses of strain CA2 were carried out for compar- ison with profiles of other marine bacteria by Dr. Fred Singleton (Center for Marine Biotechnology, University of Maryland) and by Dr. Warren L. Landry (Food and Drug Administration, Dallas, Texas). Results Larvae fed on live algae and bacteria Single additions of marine bacterial isolates to oyster larvae cultures significantly affected larval survival (AN- OVA, P< 0.01) and growth (KW, P < 0.01 ) after 10 days of culture in all experiments (Figs. 1, 2). The microbes tested can be divided into categories depending on their effects upon oyster larvae: adverse, neutral, or beneficial. Bacteria belonging to the last category were tested further, and their effects upon oyster larvae were designated as either variable or consistently beneficial. Adverse strains. Strains Gl. G2 and G4 adversely af- fected larval survival (T-HSD, P < 0.05), whereas strains Gl, G2, G4, G5, H8, and H10 adversely affected larval growth (G&H, P < 0.05). Bacteria present in 5 ml aliquots of 1 ^m-filtered seawater depressed larval survival (T- HSD. P < 0.05 ) in Experiment II and larval growth (G&H. P < 0.05) in Experiment III. Neutral strains. A large proportion of the strains (HI, H2. H3. H4. H5. H9. Hll, HI 2, G3, CA1, and CA3) added to cultures of oyster larvae had no significant effect on larval survival (T-HSD, P > 0.05) or growth (G&H, P > 0.05) compared with cultures fed algae alone. ( 'ariable strains. Addition of strains H6 and H7 to larval cultures caused inconsistent improvements of larval growth. For example, larval growth was enhanced (G&H, P < 0.05) in cultures inoculated with strains H6 and H7 in Experiment I. but the enhancement with strain H7 was statistically insignificant in Experiments II and III (G&H, P> 0.05). Moreover, larval growth was depressed (G&H, P < 0.05) when strain H6 was added to larval cultures in Experiment II. Beneficial strains. In both Experiments II and III. larvae grown in cultures inoculated with strain CA2 had a sig- nificantly greater shell length than control larvae fed only axenic algae (G&H, P < 0.05). Larval survival was en- hanced in cultures inoculated with strain H7 and CA2, but this enhancement was statistically significant only in Experiment III (T-HSD, P < 0.05). Size frequency distributions of populations of larvae fed axenic algae were skewed compared to those from cultures fed algae supplemented with CA2 bacteria (Fig. 3; Table 1 ). Analysis of variance indicates a significant interaction between treatment and experimental factors (Table 2). In both Experiments II and III, skewness coef- ficients for populations of larvae fed axenic algae alone were significantly larger (SNK, P < 0.05) that those for populations of larvae fed algae and inoculated with CA2 bacteria. The difference between the skewness coefficients of treatments in Experiment II is larger than that in Ex- periment III, explaining the significant interaction deter- mined by the two-way ANOVA test. Effects ofCA2 bacteria on the growth of algae in larval cultures Cells of Isochrysis galbana, with or without inoculations of CA2 bacteria, did not grow under the conditions used to culture larvae (Fig. 4). The occurrence of CA2 cells in the culture medium had no effect on algal growth under favorable light intensity (1000-1500 lux) and tempera- ture (20C). Larvae fed on dead algae and bacteria Significant differences among treatments in Experi- ments IV and V were determined for larval survival (AN- OVA, P < 0.0 1 ) and growth (KW. P < 0.0 1 ). The survival of larvae cultured on axenic FA or IA alone was signifi- cantly lower (T-HSD, P < 0.05) than that of larvae BACTERIAL EFFECTS ON OYSTER LARVAE 20 STARVED ALGAE ALGAE + BACTERIA O uo + 1 I \ O UJ I in STARVED ALGAE ALGAE + BACTERIA Figure 2. Survival and growth of oyster larvae after 10 days of culture on axenic Isochrysis galbana supplemented with different bacterial strains (Experiment III). Bacteria were isolated from the Whiskey Creek Hatchery, Oregon (H) or from incubations of protein capsules in seawater (CA). Naturally-occurring bacteria present in 1 Aim-filtered seawater (SW) were added in a control treatment. Other control treatments included larvae fed axenic / ga/hana or starved. Results of Tukey's HSD pairwise comparisons and Games and Howell's tests are displayed below the histograms of survival and growth, respectively. Squares that occur together on any one of the horizontal lines indicate mean values that are not different at the 0.05% level of significance. 44 P. DOUILLET AND C. J. LANGDON CO i rr z CO o UJ M LO IT DIETS O O AXENIC ALGAE AI GAF 4- CA2 DIETS O O AXENIC ALGAE ALGAE + CA2 105 140 175 210 245 280 LARVAL SHELL LENGTH Figure 3. Size frequency distributions of larvae cultured for 10 days on a diet oflsuchrysis galhana with or without addition of CA2 cells. Points represent percent larvae for each shell length interval of 30 ^m. Lines used for illustrative purposes only. Data from Experiments II (n = 400) and III (n = 800) for each treatment. cultured on live axenic algae alone (Figs. 5, 6). However, the survival of larvae fed FA or IA was higher (T-HSD, P < 0.05) than that of starved larvae. In contrast, no sig- nificant differences in larval survival were detected be- tween cultures fed live algae and cultures fed FA or IA inoculated with strains H6 and H7, respectively (T-HSD, P> 0.05). Survival of larvae fed every two days on bacteria H6 alone was not significantly different (T-HSD, P > 0.05) from that of larvae fed live algae, and was significantly higher (T-HSD, P < 0.05) than that of starved larvae (Fig. 5). Larvae from cultures inoculated every two days with 5 ml of 1 /jm-filtered seawater (SW) also showed higher survival (T-HSD, P < 0.05) than that of starved larvae. Larvae fed on FA or IA were significantly smaller than larvae fed on live axenic algae (G&H, P < 0.05), and were not different from the size of starved larvae (G&H, P > 0.05) at the end of the experiment (Figs. 5, 6). Additions of single bacterial strains to cultures of larvae fed FA or IA did not improve larval growth compared to larvae fed FA or I A alone (G&H, P > 0.05). In contrast, growth of larvae fed FA inoculated with 5 ml of 1 ^in-filtered sea- water was significantly enhanced (G&H, P < 0.05) com- pared to that of larvae fed FA alone or starved larvae (Fig. 5). Similarly, additions every two days of 5 ml of 1 //m- filtered seawater or strain H6 alone to larval cultures sig- nificantly enhanced the growth of larvae (G&H, P< 0.05) compared to that of starved larvae. The poor growth of larvae fed FA may have been due to the rupture of the freeze-killed algal cells. 60 Co-irradia- tion did not affect the integrity of the algal cells but re- BACTERIAL EFFECTS ON OYSTER LARVAE 45 Table I Skewness coefficients (gl) jrom si:efm/neney ilistrihulianx of populations oj larvae cultured in Experiments II and III Experiment Diet Average skewness of populations 1 S.D. 11 II II! Ill ISO ISO + CA2 ISO ISO + CA2 0.7906 0.21 34 (n = 4) -0.0605 0.2235 (n = 4) 0.3801 0.1720 (n = 8) -0.0466 0.29 10 (n = 8) Larvae were cultured with either axenic Isoclin'sis galbana (ISO) alone or /. galbana plus CA2 bacteria. duced their volume from 44.4 1 .92 ^m 3 to 26.3 0.59 /urn 3 (x 1 SD; n = 7). A high proportion of irradiated cells remained intact while in suspension in seawater, as demonstrated by the small decrease in cell concentration in control flasks, from 59,043 1,1 19 cells ml~ ', to 58,539 1,505 cells mr' (x 1 SD; n = 4) in 105 min. IA cells were ingested by oyster larvae at rates significantly (2 sample t-test, P < 0.01 ) greater than that for live cells. Identification of strain CA2 Strain CA2 was presumptively identified as Altero- nwnas sp. on the basis of the following characteristics: Gram negative rod; aerobic; oxidase positive; requires 250 nA/ salt; motile with polar flagella; exopolysaccharide synthesis; and guanine plus cytosine 43 mol% (T m ). The exopolysaccharides of CA2 bacteria did not react with antibodies to 20 species of A/temmonas. Further- more, both analyses of fatty acids revealed a very unusual fatty acid profile with a high proportion of C-14, C-15 fatty acids (Table 3); this is not characteristic of the genus Alteromonas. However, the fatty acid profile was not sim- ilar to any of the species profiles listed in Dr. Landry's marine library. Therefore, strain CA2 may be an Altero- monas species not typical of the genus. Further characteristics of strain CA2 include yellow pigment production, oxidation and fermentation of glu- cose, but no gas production, and inability to utilize in- organic sources of nitrogen, such as NH 4 C1 or NaNO 3 for growth. Catalase was weakly positive. Discussion Axenic larval Crassostrea gigas were used to determine the effects of additions of single bacterial strains on the survival and growth of larvae cultured with algae. Bacteria can be categorized as adverse, neutral or beneficial, de- pending on their effects upon oyster larvae. Furthermore, bacteria found beneficial in one experiment were reiested in subsequent experiments and could be further catego- rized as either variable or consistently beneficial strains. Additions of strain CA2 to larval cultures consistently enhanced larval survival (21-22%) and growth (16-21%) compared with that of larvae fed on algae alone. The specificity of bacterial strains as food for grazers has frequently been reported (Frosch, 1897 in Luck a ai. 1931; Stuart el al., 1931; Curds and Vandyke, 1966). Fur- thermore, Curds and VanDyke (1966) found that one bacterial strain was either slightly toxic, unfavorable, or favorable depending on the ciliate species tested. In con- trast, a single bacterial strain (PM-4) was found to promote the growth of both shrimp (Penaeus monodon) and crab (Port units tridentatus) larvae (Maeda, 1988; Maeda and Nogami, 1989). Consequently, no generalization about the beneficial effects of specific bacterial strains can be made; i.e., each strain must be tested again with each new target species. Bacteria may be used directly as a food item by oyster larvae (Douillet, 1991). Starved axenic oyster larvae showed poor survival and did not grow after 10 days of culture. In contrast, larvae in cultures inoculated with single bacterial strains or mixtures of naturally-occurring marine bacteria had higher survival rates than starved lar- vae, but lower growth rates than larvae fed on algal diets. Consequently, the bacterial strains tested did not provide all the nutritional requirements for larvae, but appeared, at least, to partially satisfy larval metabolic requirements, as demonstrated by the beneficial effects of bacteria on larval survival and growth. Straight-hinged oyster larvae, fed for 10 min on l4 C-labeled CA2 cells at 1.5 X 10 7 cells ml ' and purged of undigested l4 C-material, retained enough bacterial carbon to meet over 140% of their active carbon metabolic requirements during a 10 min period (Douillet, 1991). Beese(in Prieureta/.. 1990) determined that xenic, starved larval Crassostrea gigas grew 60% in size after seven days of culture, whereas starved axenic larvae did not grow. The ability of starved xenic bivalve larvae to grow has been determined to be greater for larvae Table II Two-way analysis of variance of skewness coefficients (gl)for size frequency distributions of populations of larvae cultured in Experiments II and III Source of variation d.f. Sum of squares Mean squares F-ratio Sig. level Experiment (A) 1 0.31468 0.31468 5.79 0.0259 CA2 addition (B) 1 3.2656 3.2656 60.12 0.0000 Interactions (A*B) 1 0.36039 0.36039 6.63 0.0180 Replicates (C) Residual (A*B*C) 20 1.0864 0.05432 Total 23 5.0271 Larvae were cultured with Isochrysis galbana alone or /. galbana plus CA2 bacteria. 46 P. DOUILLET AND C. J. LANGDON = 3 V -*-- CO DC c o 60 50 40 30 20- 0) 0) - 10-1 1 2 I Slow Fast Flow Figure 1. Ingestion rates (mean number of particles per fecal pellet per colony) of small (S) and large (L) colonies in slow and fast ambient flow (numbers above columns indicate number of colonies sampled). Bars represent two standard errors. SUSPENSION FEEDING IN PLUMATELLA 55 colonies may have relatively greater metabolic demands (possibly they invest more in statoblast or larval produc- tion per unit mass) than small colonies and therefore have a higher propensity to feed. However, the greater increase in the rate of feeding in small colonies with increased flow (by a factor of five) relative to large colonies (in which feeding increased by a factor of 1.8) (see Fig. 1 ) suggests that small colonies respond more strongly to increases in particle flux (or flow). Why small colonies should show such a marked response is not apparent. Perhaps small colonies create stronger ciliary currents in response to an abundance of food (particle flux serving as a cue) as has been observed in marine bryozoans (Best and Thorpe, 1983). Concerted pumping in large colonies may preclude the necessity to create individually stronger feeding cur- rents and may provide for a more constant food supply. Our results contrast with those obtained by Bishop and Bahr (1973) who found that clearance rates of the phy- lactolaemate Lophopodella carteri decreased with colony size. This discrepancy may relate, in part, to differences in colony morphology and growth in the two species, but it also is complicated by comparing feeding studies con- ducted under static and dynamic conditions and in dis- similar volumes of suspension. Lophopodella is a higher phylactolaemate, producing gelatinous, globular colonies with no branching (Wood. 1991 ). Colonies of Lophopodella do not grow indefinitely but undergo fission, the resulting colonies slowly creeping apart. Fission in Lophopodella may result in avoidance of lophophoral feeding interference that occurs as colonies get bigger, hence maximizing filtering efficiency (Bishop and Bahr. 1973; Hughes, 1989). To some extent, our re- sults for feeding in Plumalella support this contention. Plumatella does not undergo fission and its feeding does not decrease with increased colony size. The lack of in- terference in feeding in Plumatella may partly reflect its morphology. Plumatella colonies are tubular and branch- ing and their lophophores are spaced much further apart than those of Lophopodella. However, we also believe it is crucial to consider differing patterns of excurrent flow and food depletion in our experiments and in those of Bishop and Bahr (1973). In Bishop and Bahr's study (1973), Lophopodella col- onies were placed in small vials (diameter = 22 mm) that contained 10 ml of an algal suspension. Thus colonies will have had ample opportunity to resample previously filtered water because the total volume of water was small and because, under conditions of still water, previously filtered water was not carried away. Thus, it is not sur- prising that clearance rates were lower for large colonies. The volume of suspension in our study was large (25 1), and food depletion was not significant. Furthermore, in- corporation of ambient flow meant food-depleted water was carried away from colony surfaces. Conclusion This study indicates that feeding by freshwater bry- ozoans is less constrained by increased flow than it is in marine forms. As suggested above, the relatively large lophophores of phylactolaemates create powerful feeding currents that may be beneficial in both lotic and lentic environments. The complex hydrodynamics characteristic of marine habitats (see Denny, 1988) may ensure delivery of food to the level of small, circular lophophores of ma- rine bryozoans. Furthermore, small, circular lophophores maximize the collective surface area for feeding, and col- onies can benefit from the larger energy surplus associated with small size (Sebens, 1979, 1982; Ryland and Warner, 1986; Hughes, 1989). Thus lophophore size and shape in marine and freshwater bryozoans may reflect different so- lutions to different kinds of problems faced by small, co- lonial suspension feeders in the two sorts of environments. However, the role of phylogenetic constraint in deter- mining lophophore morphology cannot be ruled out (tra- ditional views hold U-shaped lophophores to be primi- tive). Although the majority of freshwater bryozoans possess large, U-shaped lophophores, small, circular lo- phophores are found in the phylactolaemate Fredericella and in the few gymnolaemates that have invaded fresh- water habitats. These exceptions to the rule indicate that the significance of lophophore size and shape in freshwater habitats merits further investigation. Acknowledgments We thank Pauline and David Whittington for their friendly interest and kind permission to collect Plumatella from their pond at Cassington Nurseries and Mark Brown for technical help. This work was submitted in partial fulfillment for the Zoology Honours Degree in the De- partment of Zoology, University of Oxford by L. Doolan. The manuscript has been improved by comments from two reviewers. Literature Cited Best, M. A., and J. P. Thorpe. 1983. Effects of particle concentration on clearance rate and feeding current velocity in the marine bryozoan Flustrellidra Impida. Mar. Biol. 77: 85-92. Best, M. A., and J. P. Thorpe. 1986. Effects of food particle concen- tration on feeding current velocity in six species of marine Bryozoa Mar. Bid/ 93: 255-262. Bishop, J. VV., and L. M. Bahr. 1973. Effects of colony size on feeding by Loplwpiiciella carteri. Pp. 433-437 in Animal Colonies: Devel- opment and Function through Time. R. S. Boardman. A. H. Chee- tham, and W. A. Oliver, eds. Dowden. Hutchinson and Ross, Stroudsberg, PA. Bushnell, J. H. 1966. Environmental relations of Michigan Ectoprocta. and dynamics of natural populations of Plumalella repens. Ecol. Monogr. 36: 95-123. Buss, L. W. 1980. Bryozoan overgrowth interactions-the interdepen- dence of competition for space and food. Nature 281: 475-477. 56 B. OKAMURA AND L. A. DOOLAN Chance, M. M., and D. A. Craig. 1986. Hydrodynamics and behaviour of Simuliidae larvae (Diptera). Can. J Zoo/. 64: 1295-1304. Ciborowski, J. J. H., and I). A. ( rain. 1989. Factors influencing dis- persion of larval black (lies (Diptera: Simuliidae): effects of current velocity and food concentration. Can. J. Fi\h. Aqnat. Sci. 46: 1329- 1341. ' DcMott, \V. R. 1986. The role of taste in food selection by freshwater zooplankton. (Ic^'logia 69: 334-340. Denny, M. W. 1988. Biology and the Mechanics of the Wave-Swept Environment. Pnnceton University Press, Princeton. N. J. 329 pp. Frechette, M., C. A. Butman, and W. R. Geyer. 1989. The importance of boundary -layer flows in supplying phytoplankton to the benthic suspension feeder, Mytilus edulis L. Limmil. Oceanogr. 34: 19-36. Frost, T. M. 1991. Porifera. Pp. 95-124 in Ecology and Classification nt \nrili American Freshwater Invertebrates J. H. Thorp and A. P. Covich, eds. Academic Press, Inc.. San Diego. CA. Hart, D. D. 1985. Causes and consequences of territonality in a grazing stream insect. Ecology 66: 404-414. Hildrew, A. G., and C. R. Townsend. 1980. Aggregation, interference, and foraging by larvae of Pleetroenemia conspersa (Trichoptera: Polycentropodidae). Amm. Behav. 28: 553-560. Hughes, R. N. 1989. .-I Functional Biology of Clonal Animals. Chapman and Hall, London. 331 pp. Jackson, J. B. C. 1979. Overgrowth competition between encrusting cheilostome ectoprocts in a Jamaican cryptic reef environment. / Anim. Kcul 48: 805-823. kaminski, M. 1984. Food composition of three bryozoan species (Bry- ozoa, Phylactolaemata) in a mesotrophic lake. Pol Arch. Hviirobtol. 31:45-53. karlson, R. H. 1991. Recruitment and local persistence of a freshwater bryozoan in stream riffles. Hydrobiologia 226: 91-101. Kay, A. M, and M. J. Keough. 1981. Occupation of patches in the epifaunal communities on pier pilings and the bivalve Pinna hicolor at Edithburgh, South Australia. Oecohgia 48: 123-130. Lacourt, A. W. 1968. A monograph of the freshwater Bryozoa-Phy- lactolaemata. /.oologische Verhandelingen No. 93. Lamberti, G. A., J. W. Feminella, and V. H. Resh. 1987. Herbivory and intraspeciftc competition in a stream caddisfly population. Oeeologia 73: 75-81. Lopez Gappa, J. J. 1989. Overgrowth competition in an assemblage of encrusting bryozoans settled on artificial substrata. Mar. Ecol. Prog. Ser. 51: 121-130. McAuliffe, J. R. 1984. Competition for space, disturbance, and the structure of a benthic stream community. Ecology 65: 894-908. McKinney, F. K., M. R. A. Listokin, and C. D. Phifer. 1986. Row and polypide distribution in the cheilostome bryozoan Bitgula and their inference in Archimedes. Letliaia 19: 81-93. Muschenheim, D. K. 1987. The dynamics of near-bed seston flux and suspension-feeding benthos. J Mar Res 45: 473-496. Okamura, B. 1984. The effects of ambient flow velocity, colony size, and upstream colonies on the feeding success of Bryozoa. I. Bitgula stolotujera Ryland, an arborescent species. J. E\p. Mar. Biol Ecol. 83: 179-193. Okamura, B. 1985. The effects of ambient flow velocity, colony size, and upstream colonies on the feeding success of Bryozoa. II. Cono- pcuin rcticiiluin (Linnaeus), an encrusting species. J. E.\p. Mar Biol. Ecol 89: 69-80. Okamura, B. 1988. The influence of neighbors on the feeding of an epifaunal bryozoan. / Exp. Mar. Biol Ecol. 120: 105-123. Okamura, B. 1990. Particle size, flow velocity, and suspension-feeding by the erect bryozoans Bugula neruina and B stolonifera. Mar. Biol. 105: 33-3X. Okamura, B. 1992. Microhabitat variation and patterns of colony growth and feeding in a marine bryozoan. Ecology 73: 1502-1513. Ricciardi. A., and D. J. Lewis. 1991. Occurrence and ecology of Lo- phupodella carleri (Hyatt) and other freshwater Bryozoa in the lower Ottawa River near Montreal, Quebec. Can J Zoo/. 69: 1401-1404. Rubin. J. S. 1982. The degree of intransitivity and its measurement in an assemblage of encrusting cheilostome Bryozoa. J. Exp. Mar. Biol Ecol. 60: 119-128. Ryland, J. S., and G. F. Warner. 1986. Growth and form in modular animals: ideas on the size and arrangement of zooids. Phil. Trans R Soc Loud. B313: 53-76. Sebens, K. P. 1979. The energetics of asexual reproduction and colony formation in benthic marine invertebrates. Am. Zool. 19: 683-697. Sebens. K. P. 1982. The limits to indeterminate growth: an optimal size model applied to passive suspension feeders. Ecology 63: 209- 222 Stebbing, A. R. D. 1973. Competition for space between the epiphytes of Fiicits semilu.s L. / Mar. Biol. Assoc. (/.A". 53: 247-261. Vogel, S. 1981. Life in Moving Fluids. Princeton University Press, Pnnceton. N. J. 352 pp. Winston, J. K. 1977. Feeding in marine bryozoans. Pp. 233-271 in Biology of Bryozoans. R. M. Woollacott and R. L. Zimmer, eds. Academic Press, New York. Winston, .J. E. 1978. Polypide morphology and feeding behavior in marine ectoprocts. Bull. Mar. Sci. 28: 1-31. Wood, T. S. 1973. Colony development in species of Plumalella and Fredericella (Ectoprocta: Phylactolaemata). Pp. 395-432 in Devel- opment and Function of Animal Colonies through Time. R. S. Board- man, A. Cheetham, and J. Oliver, eds. Dowden. Hutchinson & Ross. Stroudsburg. PA. Wood, T. S. 1989. Ectoproct bryozoans of Ohio. Ohio Biol. Surv Bull \ew Series 8(2): X + 70 pp. Wood, T. S. 1991. Bryozoans. Pp. 481-499 in Ecology and Classifi- cation />/ \orih American Freshwater Invertebrates. J. H. Thorp and A. P. Covich. eds. Academic Press. Inc.. San Diego. CA. Reference: Bin/. Bull 184: 57-78. (February, 1993) Aplacophora as Progenetic Aculiferans and the Coelomate Origin of Mollusks as the Sister Taxon of Sipuncula 1 AMELIE H. SCHELTEMA \Vootls Hole Oceanographic Institution, Woods Hole, Massachusetts 02543 Abstract. Evidence is presented in support of the fol- lowing phylogenetic hypotheses: ( 1 ) Sipuncula are the sis- ter taxon of Mollusca; (2) the two aplacophoran taxa, Neomeniomorpha (= neomenioids) and Chaetodermo- morpha (= chaetoderms), are monophyletic with a com- mon neomenioid-like ancestor, and of the two taxa, Chaetodermomorpha are more derived; (3) Aplacophora and Polyplacophora are sister taxa and form a clade, Acu- lifera; (4) Aculifera are the sister group of the remaining extant mollusks, Conchifera; and (5) Aplacophora are progenetic Aculifera. The evidence is based on homologies of early and late emhryological development, adult morphologies, and molecular analyses. Embryological development in si- punculans and mollusks shows a close relationship be- tween them, and embryological development of the shell separates Aculifera and Conchifera. Adult morphologies indicate: ( 1 ) monophyly of Aplacophora; (2) sister-group relationship between Aplacophora and Polyplacophora; (3) a molluscan plesiomorphy of nonsegmented serial replication of organs; and (4) progenesis in Aplacophora. Molecular evidence supports the embryological and mor- phological relationships between Sipuncula and Mollusca. Mollusca are thus hypothesized to be coelomate Eu- trochozoa, which share an ancestor that probably had se- rial replication of organs. Differences in size and structure of the coelom among Eutrochozoa are hypothesized to have been brought about by changes in the timing and the process of cavitation of the mesodermal bands that arise from cell 4d. Through the process of progenesis Aplacophora retained an ovoid embryological shape and Received 19 August 1992; accepted 25 November 1992. ' Contribution Nos. 8205 from the Woods Hole Oceanographic In- stitution, and 314 from the Smithsonian Marine Station at Link Port. several internal structures that, although they appear to be in a primitive state, are actually secondarily derived as is quadrant D specification during early cleavage. Introduction The uniqueness of Aplacophora among Mollusca lies in their derived vermiform body in combination with an internal organization that appears to reflect a primitive molluscan state, especially the simple ladderlike nervous system, serial musculature, distichous radula (two teeth per row) in its plesiomorphic aplacophoran state, simple digestive system, and epidermis that produces an aculif- erous cuticle. Their evolutionary significance to the phy- lum has long been a matter for conjecture. First came the question of whether Aplacophora were even mollusks, as they lack a number of "typical" characters such as a shell, mantle, and kidneys (e.g., Thiele, 1902; H. Hoffmann, 1929-30), but they have more usually been considered to belong within the phylum because of similarities to chitons in their nervous system ( Amphineura) and spicules (Acu- lifera) (e.g., Spengel. 1881; Heath, 1911). Further discus- sions were concerned with whether aplacophorans were "degraded" or truly "primitive" mollusks (see Hyman, 1967, pp. 68-70 for a historical account). There have been no current arguments which separate Aplacophora from Mollusca since evidence for a close relationship between Aplacophora and Polyplacophora was published by S. Hoffman (1949), but under present discussion is their origin and position within the phylum (Salvini-Plawen, 1972. 1981a, 1985; Scheltema, 1978, 1988), as well as the origin of the phylum Mollusca itself. Mollusca have been argued either to have a noncoelomate origin and to be the sister taxon of the eucoelomate An- nelia-Echiura-Sipuncula (Salvini-Plawen, 1972, 1985 fig. 42), or to be eucoelomates with an ancestor in common 57 58 A. H. SCHELTEMA with other coelomates (Wingstrand, 1985; Scheltema. 1988). In either argument, Aplacophora have been con- sidered stem forms and therefore preceded the Monopla- cophora with serial replication of organs. Hypotheses for a noncoelomate origin rest on the ar- gument that the worm-like Aplacophora with replicated lateroventral musculature evolved from a turbellario- morph ancestor, and that consequently the molluscan coelom is not homologous to that in the Eutrochozoa. A coelomate origin has been hypothesized from annelid- mollusk relationships, including the presence of a cell 4d that gives rise to mesoblasts and consequently a homol- ogous coelom, the presence of a trochophore larva, and serial repetition of body parts. Because the molluscan coelom is small and unsegmented, the idea that annelids and mollusks form a clade with a common segmented ancestor is poorly accepted. The dichotomous choice be- tween either a turbellariomorph or an annelid-like ances- tor for mollusks has dominated recent thinking about molluscan evolution (e.g., Hyman, 1967; Haszprunar, 1992), and the relationship of mollusks to other Eutro- chozoa has not been examined. However, recent molec- ular data discussed below urge reconsideration of mol- luscan relationships to other phyla. Evidence is presented here to support the hypotheses that ( 1 ) Mollusca are eucoelomates with their closest living relatives in Sipuncula, their sister group; (2) Aplacophora and Polyplacophora are sister groups in the subphylum Aculifera (contradicting Scheltema, 1978, 1988); (3) Acu- lifera are the sister group of the remaining living mollusks, Conchifera; (4) the aplacophoran taxa Chaetodermo- morpha (= Caudofoveata, here also called chaetoderms) and Neomeniomorpha (= Solenogastres sensu nomine Salvini-Plawen, here also called neomenioids) are mono- phyletic, sharing a neomenioid-like ancestor; and (5) aplacophorans are progenetic Aculifera. Considered in the discussion is the homology of the eutrochozoan coelom and the evolutionary difference between metamerism, or segmentation as it occurs in the annelids, and serial rep- lication of organs, as found in Neopilina and Vema (Wingstrand, 1985). The term "metamerism" is used here only to denote a segmented coelom; "serial replication" is used to denote the more general case of serial repetition of organs, whether or not by metameres. Evidence that Mollusca are Descended from Coelomates Mollusca have a coelom consisting of gonadal lumina, pericardium, and kidneys, as well as part of the gameto- ducts in Aplacophora. A noncoelomate ancestry calls for the widening of a pericardial space lined by mesoderm as protection for a heart (Salvini-Plawen, 1968a, 1972; not discussed 1985, 1990) and for gonads separate from the pericardium. This development of coelomic spaces would be a molluscan apomorphy. not homologous with annelid or sipunculan coelom. Alternatively, the molluscan peri- cardium can be considered as reduced from a large coe- lomic space homologous to that in other eutrochozoa. The involvement of the pericardial coelom in excretion is unique to mollusks. Ultranltration of blood occurs through podocytes that are present in most molluscan classes including Aplacophora (Andrews, 1988; Reynolds and Morse, 1991). Five independent lines of evidence indicate that re- duction of coelom is the case, and that Mollusca are eu- trochozoan coelomates: ( 1 ) presence of the molluscan cross in mollusks and sipunculans and (2) homology of certain characters in larvae of mollusks and sipunculans indicate that mollusks and sipunculans are sister taxa; (3) a large pericardium among "primitive" mollusks indicates that it is a molluscan plesiomorphy; (4) the embryological development of mesoderm in annelids, mollusks, sipun- culans, and nemertines is similar, and the coelom in the four groups is homologous; and (5) molecular data groups mollusks with other eutrochozoans. Sipunculans as sister taxon of the mollusks An evolutionary relationship between sipunculans and mollusks lies in their early embryological development and in morphological features of sipunculan pelagosphera and molluscan larvae. Molluscan cross. The molluscan cross is found in the embryological development of Gastropoda, Polyplaco- phora, Scaphopoda, and Aplacophora by the end of the 64-cell stage (Verdonk and van den Biggelaar, 1983; Heath, 1899; van Dongen and Geilenkirchen. 1974; Baba, 1951). It is formed by la' 2 -ld 12 cells and their descen- dents, with cells la" 2 -ld" 2 , called peripheral rosette cells, forming the angle between the arms of the cross (Fig. 1 A, B, D, peripheral cells solid black). In Annelida, however, it is cells la" 2 - Id" 2 that form the cross (Fig. IE, cross cells solid black) (Wilson, 1892). In the 64-cell stage of the neomenioid aplacophoran Epimenia vermcosa figured by Baba (1951), a molluscan cross seems apparent from Baba's shading (Fig. ID), although Salvini-Plawen (1985) found "no definite cross formation" in the same source. Manuscript drawings by G. Gustafson of developing Chaetoderma nitidulum eggs likewise show a molluscan cross. In contrast to most mollusks, early cleavage in Pe- lecypoda is asynchronous and bilateral, and no cross is formed; its absence would seem to be an apomorphy. Likewise, development in Cephalopoda seems an apo- morphy of that group, which has telolecithal eggs, early bilateral cleavage, and no molluscan cross. In Sipuncula, a molluscan not an annelid cross is formed, as Rice ( 1975, 1985) has emphasized and refig- ured from Gerould ( 1906), who first described its presence APLACOPHORA: PROGENETIC COELOMATES 59 Figure 1. (A-D) The molluscan cross. (A) Gastropoda (Lyniiiucu siagnalis. after Verdonk and van den Biggelaar. 1983, p. 1 1 1 fig. 3b); (B) Polyplacophora (Stenoplax heathiana. after Heath, 1899. pi. 32, fig. 23); (C) Sipuncula (Golfingia vulgaris. after Gerould, 1 906. p. 99, fig. D, as published in Rice, 1975, p. 99. fig. 17); (D) Aplacophora (Epimenia verrucosa. after Baba, 1951. p. 46, fig. 18). The apical rosette la'"-ld'" is shown in fine, close stippling; arms of the cross la' : - Id' 2 and daughter cells are shown in fine, open stippling; tip cells of cross 2a"-2d" are shown in coarse stippling; peripheral rosette cells la" 2 -ld" : are solid; and trochoblast cells I a 2 - Id 2 are clear. In Epimenia (D), the cleavage stage appears to be earlier than shown in A-C, as the tip cells have not yet separated from 2a' and 2c' (indicated by question marks), and the arms of the cross are not quite straight, similar to an earlier stage in Polyplacophora (Heath. 1899. pi. 32, fig. 17). In B, only one tip cell was discernible in Heath's illustration, and in C tip cells were not indicated in Gerould's original figure. (E) Annelid cross, Polychaeta (Nereis) (after Wilson, 1892, p. 396. diagram II). The apical rosette la 1 "-Id"' is shown in fine, close stippling; peripheral cells la' 2 -ld 12 are shown in fine, open stippling; and the arms of the cross from la" 2 -ld" 2 are solid. in sipunculan development (Fig. 1C). The presence of a molluscan cross during embryological development is understood here to be of phylogenetic importance, and sipunculans and mollusks share a character not found in either annelids or flatworms (Freeman and Lundelius, 1992). Its presence can be considered apomorphic to the embryonic morphology of turbellarians, which lack a cross. Similarities between sipunculan and molluscan lan'ae. Gerould (1906) noticed certain other resemblances to mollusks besides the molluscan cross in the development of sipunculans. In particular, he found similarity between sipunculan pelagosphera and molluscan larvae. The pela- gosphera is unique to sipunculans. It is a swimming larva that metamorphoses from a trochophore stage (Rice, 1975, 1985). Gerould noted the resemblance of the pela- gosphera lip glands to chiton larval pedal glands, and of the pelagosphera buccal organ to the radula sac in chiton larvae (Figs. 2, 3, 4). Pelagosphera larvae can either swim upright with the large metatroch or creep, head-down, along a solid surface. These activities are lost along with the larval head at metamorphosis. Jagersten (1963) first described creeping in living pelagosphera, and he related it to a creeping gastropod. He also noted that the buccal organ (= pharyngeal bulb, Schlundkopf) was used in feeding. Later Jagersten (1972) proposed a possible, but not certain, homology of the pelagosphera lip, which is the creeping surface posterior to the mouth, and the creeping lobe, or foot, between mouth and anus of mol- luscan larvae. Rice (1975, pp. 120-121) described the creeping lo- comotion of pelagosphera as follows: "The larva is able to ... glide along with . . . [the] head flattened against the bottom. Frequently the larvae . . . may crawl in the manner of an inchworm. presumably scraping material from the bottom. The continual eversion of the buccal organ during feeding probably aids in the removal of food from the substratum. This tough muscular organ [covered by cuticle. Rice, 1973] is believed to function in breaking up material into small panicles for feeding. . ." A mucus- like substance from the lip glands is secreted as the animal moves along a natural substratum (Rice, 1981 and pers. comm.). My own recent observations on living pela- gosphera corroborate many of Rice's. Precise descriptions of the protrusible buccal organ and lip gland have been given by Rice (1973). The buccal organ is a muscular sac, ventral and posterior to a cuticle- lined invagination called the buccal groove that lies below the esophagus. The epithelium of the buccal organ is overlain by the cuticle of the ventral side of the buccal groove and is the area first protruded (Fig. 2). Although the precise innervation of the buccal organ was not dem- onstrated, the circumesophageal connectives, which arise from the dorsal cerebral ganglion, are closely associated with the organ. Both the topography and function of the buccal organ and groove are remarkably similar to those of the radular apparatus in mollusks: ventral odontophore = buccal organ; ventral radula sac = buccal groove; and ventral cuticular radula = ventral portion of the cuticular buccal groove. Furthermore, the odontophore and prob- ably the buccal organ are innervated through connectives united with the cerebral ganglion. The homology would be more certain if it were known whether the buccal organ musculature is formed from mesoderm, as is the odon- tophore of mollusks (Raven, 1966), or whether it is myo- epithelial as in archiannelids (Jagersten, 1947; Rice, 1973). The lip gland takes several forms in various pelago- sphera, from a bilobed to a paired or four-lobed body A. H. SCHELTEMA Figure 2. Peiagosphera larvae of Sipuncula. (A) Frontal view of head. Siinmailu* sp. (from Rice, 1981, fig. 4). (B) Entire larva. Aspidosiphon sp. (from Rice, 1981, fig. 6). Numbers as in Figure 3: I huccal gland, 3 pore of lip gland, 4 mouth, 5 lip. which opens either directly, or by way of a ciliated duct or ducts, into the lip pore. In comparison, the anterior pedal gland in larval chitons and in Aplacophora is duct- less (cf.. Figs. 2A, 4B, 6C). Aplacophora, but not chitons, have a central ciliated pit. Similarities in form and function in these three struc- tures lip and foot, lip glands and pedal glands, and buc- cal organ and radula with its sac are striking. Their morphologies are particularly clear in sagittal sections of a pelagosphera and a chiton larva (Figs. 3, 4). There are also similarities in their development, as they all arise from posttrochal ectoderm, with these differences: in si- punculans, the origin of all three structures is stomodeal, whereas in mollusks, the ventral somatic plate, usually from cell 2d, gives rise to the foot and its glands, and only the radula sac is stomodeal (Raven, 1966). In Sipuncula as well, cell 2d gives rise to the somatic plate, which forms the ectoderm of the trunk (Rice, 1976). In mollusks, how- ever, the proximity and functional interdependence of the somatic and stomodeal structures are indicated by the pedal contribution to feeding in veliger larvae. An anterior, medial ciliary tract is formed on the foot by which particles unsuitable for ingestion are rejected (Moor, 1983). Only the head region of the pelagosphera, which is rad- ically altered during metamorphosis to a juvenile sipun- culan. can be compared to the Mollusca. The posterior part of the body with its large coelomic sac, nephridia, mid-dorsal anus, and ventral nerve cord, are already de- finitive adult structures. Evidence from the presence of the molluscan cross and from locomotary and feeding structures that are similar in mollusks and larval sipunculans is sufficiently strong that the two phyla can be considered as sister groups, and mollusks, therefore, as eucoelomates. Of course, if the primitive mode of sipunculan development should prove to be by way of a nonfeeding, lecithotrophic larva, then the similarities between planktotrophic pelagosphera and molluscan larvae would be convergent. However, Rice (1985) most recently considered evolutionary questions of sipunculan larval development and concluded that a yolky egg and short-lived planktotrophic pelagosphera was the primitive mode of development. Other considerations. Two further observations can be made to support arguments for a sipunculan-molluscan sister relationship, one embryological, the other paleon- tological. The first is the embryological development of Echiura (Newby, 1940) compared to that of the sipun- culans. Echiurans have traditionally been linked with si- punculans, both having worm- or sac-like, unsegmented coelomate bodies, but echiurans afford a contrast to si- punculans in their closer relationship to annelids. They have an annelid cross rather than a molluscan cross during early cleavage, and as in annelids, the major ciliary band API ACOPHORA: PROGENETIC COELOMATES 61 Figure 3. Midsagittal section of the pelagosphera larva, Phascolosoma agassi:ii (from Rice, 1973, pi. 5). 1 buccal organ, 2 lip gland, 3 pore of lip gland, 4 mouth, 5 lip, 6 stomach, 7 coelom, 8 esophagus. of older echiuran larvae is the prototroch anterior to the mouth. In sipunculan pelagosphera. the metatroch below the mouth, not the prototroch, is the major swimming organ. Indeed, the region in pelagosphera that forms the head with its locomotory lip, lip gland, and buccal organ, is represented in echiuran larvae by only a few rows of cells between the prototroch and metatroch, and no larval organs are present. If sipunculans are sister taxon of the Mollusca, they must have arisen, like mollusks, early in the evolution of metazoans. One piece of evidence for an early sipunculan history is the mid-Cambrian genus Ottoia from the Bur- gess Shale. Considered priapulids by Conway Morris ( Whittington, 1985) and close to priapulids by Banta and Rice (1976), the genus indicates great diversity of spe- cialized sacciform, coelomate or pseudocoelomate, worm- like animals already in the early Paleozoic. Sipunculans therefore could have a very long, but unobservable and unverified, geologic history. A second piece of evidence is that sipunculans contain hemerythrins, found also only in priapulids, lingulid brachiopods, and some annelids (Curry and Runnegar, 1 990). Because lingulids and prob- ably priapulids and annelids are known from the early Cambrian, the presence of hemerythrins indicates a very long history for all forms having these oxygen transport molecules. Si:e of the pericardium in "primitive" mollusks The pericardium is larger relative to the heart in Apla- cophora, Monoplacophora, and Polyplacophora than it is in Gastropoda, Pelecypoda, and Cephalopoda (Schel- tema, 1973, 1988; Scheltema and Kuzirian, 1991) (Figs. 5, 6A). Ontogenetically, the pericardium is already large before the heart develops from pericardia! epithelium in Aplacophora (Baba, 1938), and in Polyplacophora de- velopment of the pericardium precedes development of the gonad (Hammersten and Runnstrom, 1925). Thus the polarity of pericardia! size is from large to small in Mollusca, and the continued reduction within the phylum is considered to be a derived condition of the Mollusca. B Figure 4. Newly settled larvae of Polyplacophora. (A) Midsagittal section of Acanthochiton discrepans (after Hammarsten and Runnstrom, 1925, fig. E, figure reversed). (B) Ventral view of Stetwplax hcalhiana just after metamorphosis (after Heath. 1899, fig. 59). The opening of the pedal gland (3) lies posterior to the mouth (4); the gland opens through "a series of. . . intercellular channels" rather than a duct (Heath, 1899, p. 631); compare with Figure 6C. 1 radula sac, 2 anterior pedal gland, 3 opening of pedal gland, 4 mouth, 5 foot, 6 larval eye. Structures num- bered 1-5 are homologous to structures with the same numbers in Figs. 2 3. 62 A 9 9 9 Figure 5. Large pericardia! space and heart in primitive molluscs. (A) Aplacophoran, Chaetaderma nitidulum, sagittal section. Gametes pass from the gonads through the pericardium with its large, paired, lateral extensions ("horns") and thence into gametoducts leading to the mantle cavity (from Scheltema. 1973, fig. 2, and Scheltema, 1988, fig. 1 3). (B) Polyplacophoran, Chiton sine/am, cross section (after Wissel, 1904, pi. 24, fig. 49). (C) Monoplacophoran. Neopilina galathea. dorsal view, with paired pericardia! sacs, paired ventricles, and two pairs of auricles (after Lemche and Wingstrand, 1959, from Scheltema, 1988, fig. 1 3). (D) Polyplacophoran, Acanlhopleiira echinata, dorsal view, with two pairs of openings between auricles and ventricle (after Plate, 1898. from Scheltema, 1988. fig. 13). 1 pericardium, 2 ventricle, 3 auricle, 4 opening between auricle and ventricle, 5 auriculoventricular valve, 6 aortal bulb, 7 gonopericardial duct, 8 lateral extension of pericardium, 9 gametoduct. Development ofmesoderm The interpretation that the coelom is reduced in Mol- lusca assumes that the molluscan pericardium is homol- ogous to the coelom in other spiralian coelomates, namely Annelida and Sipuncula. In all three, the coelom is formed from mesoderm that originates from embryonic cell 4d. This cell gives rise to a pair of mesodermal teloblasts, which migrate inward to a ventrolateral position, one on each side of the midline ( Verdonk and van den Biggelaar, 1983; Anderson, 1973; Rice, 1975) and proliferate forward into two lateral mesodermal bands. Mesodermal bands APLACOPHORA: PROGENETIC COELOMATES 63 Figure 6. (A) Cross-section through the pericardium of a neomenioid aplacophoran, Helicoradomenia juani (from Scheltema and Kuzirian. 1991, fig. 5C). (B) Cross-section through the pedal gland and pedal pit of a neomenioid aplacophoran, Ocheyoherpia sp. The voluminous pedal gland occupies most of the head region; the lobes of the gland are in varying stages of secretion. (C) Ciliated pedal pit of Helicoradomenia juani. The pedal gland discharges into the pedal pit, not through distinct ducts, but through numerous channels as described for chitons (Fig. 4B). (D) Secretory epidermal papillae of the neomenioid aplacophoran, Helicoradomenia juani (from Scheltema and Kuzirian, 1991, fig. 2C). (E) Secretory epidermal papillae of the polyplacophoran. Acanthochiton fascicularis (from Fischer el ai, 1980, fig. 3). 1 pedal gland, 2 ciliated pedal pit, 3 dorsal blood sinus, 4 dorsal cecum of midgut, 5 cerebral ganglion, 6 oral cavity, 7 pericardium, 8 auricle, 9 ventricle, 10 ovum. 11 U-shaped gametoduct, 12 copulatory spicule pocket, 13 foot. Asterisks in D and E, cavities of dissolved spicules. 64 A. H. SCHELTEMA are present as well in Nemertini (Turbeville, 1986). In annelids, sipunculans, and nemertines, the coelom is formed by cavitation (schizocoely) of the bands. The coe- lom constitutes the major body cavity in annelids and sipunculans, but in nemertines it forms only vessels for blood circulation (Turbeville, 1986). In mollusks, the me- sodermal bands break up into masses of coelenchyme, from which is formed a solid anlage or pair of anlagen that cavitate to form the pericardium, heart and kidneys (Raven, 1966; Moor, 1983). In some mollusks with paired anlagen, the pericardium begins as paired cavities before becoming united (Raven, 1966). In Neopilina the peri- cardium is still paired (Fig. 5C), and the large pericardia! "horns" in some Aplacophora (Fig. 5A) may reflect an ancestral paired condition. The coelom among the spiralian protostomes described here is interpreted as being homologous because of sim- ilarities in early embryological development. Differences in coelom formation among the four phyla apparently arise from variations in the timing of cavitation after the mesodermal bands have formed; but the differences in process are not considered sufficient to deny homology of the coelom. A single pericardium formed by fusion in mollusks other than Neopilina is thus an apomorphy. Molecular evidence Recent sequencing of 18S ribosomal RNA among 22 classes (not including Aplacophora), in 10 animal phyla, split off acoelomate Platyhelminthes as sister group of the remaining bilaterian taxa, the eucoelomates, which fall into four closely rooted groups (Field ct ai, 1988). The group termed Eutrochozoa (Ghiselin, 1988) includes five analyzed phyla: Mollusca, Annelida. Brachiopoda, Po- gonophora, and Sipuncula. More recently Turbeville el al. (1992) have added Nemertini to the Eutrochozoa, bas- ing their results on 18S rRNA and analyzing two Platy- helminthes, in addition to the single flatworm analyzed by Field el al. (1988). A re-analysis by Lake (1990) of the 1988 data positioned Sipuncula closest to Mollusca and Brachiopoda, with Annelida and Pogonophora as sister groups. The presence of hemerythrins in Brachiopoda, Sipuncula, and some Annelida affords independent sup- port from molecular data for some of the results of Field t't al. (Curry and Runnegar, 1990). The relationships among Sipuncula, Mollusca, and Brachiopoda, however, remain unresolved, and possible synapomorphies of sipunculan and molluscan larval characters were not taken into account by Lake. Although the molecular evidence is still incomplete, it suggests that mollusks have descended from a coelomate ancestor, and that sipunculans are their closest sister group. In proposing that the last common ancestor of the Annelida-Mollusca lineage was hemocoelic and segmented. Lake did not dis- cuss the presence or absence of a coelom. Ghiselin ( 1988) considered the evolution of Mollusca in light of the mo- lecular evidence given in Field et al. (1988), amplifying the data with an analysis of specific nucleotides and a useful history of molluscan phylogenetic hypotheses. Ghiselin favored a segmented, coelomate eutrochozoan ancestor, with loss or reduction of segmentation in the Mollusca. Salvini-Plawen (1990), however, retained a preference for a turbellariomorph molluscan ancestry and refuted the validity of the sequencing by Field et al. (1988) and Ghiselin (1988), because "for some selected, tradi- tionally monophyletic groups [including mollusks] eu- phemistic premises are made" by eliminating some data as convergences. Willmer and Holland (1991) also con- sidered that mollusks had a flatworm origin and suggested that RNA analysis of several Platyhelminthes might show them to be poly- or paraphyletic, but the work of Turbe- ville et al. (1992) indicates that they are monophyletic. Monophyly of Aplacophora A proposed homology of the chaetoderm oral shield with the creeping sole of the archimollusk was the basis for separating the two aplacophoran taxa into two classes (Fig. 7B, C; Fig. 8A) (Salvini-Plawen, 1972, 1985, 1990). This homology was based on the innervation of the oral shield (Salvini-Plawen, 1972), the character of the epi- dermis, and the presumed homology of cuticular struc- tures (Fig. 8C, arrowhead) (S. Hoffman, 1949), but it is not upheld either by light or transmission electron mi- croscopy (Scheltema et al., in press, fig. 9; Tscherkassky. 1989). The oral shield cuticle is continuous with that of the pharynx and is a lip. and the innervation of the shield is cerebral, lying anterior to that part of the anterior ner- vous system considered "tentacular," and thus part of the head region, by Ivanov ( 1 99 1 ). Accordingly the two apla- cophoran taxa cannot be separated on the basis of the chaetoderm oral shield, although Salvini-Plawen (1990) recently argued that the homology holds because the fore- gut and oral-shield epithelia are different, and the presence of the cuticle is secondary. In a schematic drawing through an oral shield, Salvini-Plawen (1990, fig. 7) showed a sep- aration, the "mantle rim," between the oral shield cuticle and body cuticle, but this separation does not exist in my experience (Scheltema et al., in press, fig. 9B). The ar- gument would be clarified if it were known whether the oral shield is stomadeal in origin. Several synapomorphies suggest that the two aplacoph- oran taxa are monophyletic. The outgroup for comparison is Polyplacophora. The tetraneural nervous system, including the cerebral commissure, lateral and ventral nerve cords, and supra- rectal commissure, is more heavily ganglionated in both neomenioids and chaetoderms than in chitons. The radula APLACOPHORA: PROGENETIC COELOMATES 65 Figure 7. (A-C) Chaetodermomorpha. (A. B) (.'kern 'derma lurnerae. entire animal (antenor to left) and divided oral shield (I'rom Scheltema, 1985, fig. 3L, O. P). (C) Oral shield of SculOfiim megaradulalu.i (cf., Fig. 8A) (from Scheltema, 1988, fig. 6). (D. E) Neomeniomorpha. (D) Dorymenia sp. (E) A new neomenioid genus and species in the family Simrothicllidae. D and E are drawn to the same scale, anterior to left; the midgut and gonad lie between X-X and Y-Y. in its plesiomorphic state in Aplacophora is distichous, that is, only two teeth per row (Scheltema, 1988; Schel- tema el al, 1989), a reduction in number from the doco- glossate chiton radula. Both neomenioids and chaeto- derms have a dorsoterminal sense organ (= dorsocaudal sensory pit), or sometimes several, in the epidermis. It is of unknown function, although homology to the osphra- dium has been conjectured (Spengel, 1881; Haszprunar, 1987). Whether or not this homology is correct, the po- sition of the dorsoterminal sense organ is an autapomor- phy of the Aplacophora, for there is no compelling evi- dence that this position, postulated to be primitive for the molluscan osphradium (Salvini-Plawen, 1985), is other than an apomorphy shared only by neomenioids and chaetoderms. The two aplacophoran taxa share a similar reproductive system unique among mollusks. Paired gonads, some- times fused, open directly into the pericardium, and paired U-shaped gametoducts lead from the posterior end of the pericardium, first anteriorly and then posteriorly, to the mantle cavity (Figs. 5 A, 6 A, 9D). Separate gonaduct openings in species of Phyllomenia (Salvini-Plawen, 1978) are interpreted here as a derived condition of that genus. The mantle cavity in both neomenioids and chaeto- derms is small and posterior, acting as little more than a cloaca. In neomenioids, the groove on either side of the foot-fold can also be considered as reduced mantle grooves (Figs. 6A, 8C). The paired ctenidia in chaetoderms, which fill most of the mantle cavity, is probably a plesiomorphy. with loss in the neomenioids resulting from the space re- quirements of a secondarily more complicated reproduc- tive system, including sometimes very large copulatory spicules. Finally, the worm shape itself is here considered a syn- apomorphy of the Aplacophora. and not separate. Figure 8. (A) Cross-section through the oral shield of a chaetoderm, Scutopus megaradulatus, showing continuity between pharyngeal and oral-shield cuticle. Arrow indicates transition between homogeneous pharyngeal cuticle and more specialized fibrillar oral-shield cuticle with a thickened outer layer (from Schel- tema, 1988, fig. 5). (B) Sagittal section through a neomenioid, liymnomenia sp., showing serial lateroventral musculature. (C) Cross-section through the nonmuscular, heavily ciliated foot of a neomenioid, Helicora- domeniajuani. The arrowhead indicates the nonspiculose cuticle of the mantle cavity extending along each side of the foot groove, which was considered homologous to the chaetoderm oral shield by S. Hoffman (1949). (D) Cross-section through the radula. radula bolsters, and paired, hollow radula vesicles in Helicor- adomenia /lunu (from Scheltema and Kuzinan, 1991. fig. 4D). 1 oral-shield cuticle, 2 pharyngeal cuticle, 3 cuticle of body wall, 4 nerve fibers from precerebral ganglion, 5 ovarian region of hermaphroditic gonad, 6 digestive cells of undifferentiated midgut. 7 copulatory spicule pocket, 8 foot. 9 radula vesicle, 10 radula, 1 1 dorsal cecum of stomach/digestive gland. 66 APLACOPHORA: PROGENETIC COELOMATES 67 Figure 9. Nervous system and reproductive system in a neomenioid, Strophomenia scandens (A, D) and nervous system in a chaetoderm, Limifossor lalpoideus (B, C). (A) Lateral (= pleural, visceral) cord with its ongm in the cerebral ganglion separate from the origin of the ventral (= pedal) cord. Lateral and ventral cords remain separate posteriorly (after Heath, 1904, pi. 27, fig. 2). (B) Anterior end; the lateral and ventral cord have a single origin in the cerebral ganglion (after Heath, 191 1, pi. 10, fig. 8). (C) Posterior end: the ventral cord runs close to the lateral cord and fuses with it. The suprarectal commissure is ganglionated (after Heath, 1905. pi. 43, fig. 18). (D) Posterior end; the lateral and ventral cords are well separated, with the separation maintained throughout. The gonad empties into the pericardium, which is shown in fine stippling. The U-shaped gametoduct, with a many-lobed seminal receptacle, is shown in coarse stippling; it runs from the posterior end of the pericardium to the mantle cavity, indicated by dashed lines (after Heath, 1904, pi. 27, fig. 6). 1 cerebral ganglion, 2 lateral cord, 3 pedal cord, 4 suprarectal ganglion/commissure, 5 buccal ganglion, 6 gonad, 7 pericardium, 8 seminal receptacle, 9 mantle cavity. 68 A. H. SCHELTEMA convergent apomorphies in the two taxa. When this char- acter and those mentioned above are considered together, the Aplacophora clearly emerge as a monophyletic taxon. Chaetodermomorpha, derived Aplacophorans Neomeniomorpha are more similar than Chaetoder- momorpha to the outgroup, the Polyplacophora, in ner- vous system (Fig. 9A, D), form of epidermal papillae (Fig. 6D, E), presence of anterior pedal glands (present only in the larvae of chitons) (Figs. 4, 6B), presence of paired pharyngeal glands, serial lateroventral musculature (Fig. 8B), and inequality of height and width dimensions. Sev- eral autapomorphies indicate that the burrowing Chae- todermomorpha have been derived from a creeping neo- menioid-like ancestor. Criteria for considering a structure to be apomorphic are fusion or elaboration. Changes in the nervous system are pronounced. In chaetoderms the lateral and ventral cord on each side have a single origin from the cerebral ganglion, whereas in nearly all neomenioids lateral and ventral cords have sep- arate origins (Fig. 9A, B). In chaetoderms, the lateral and ventral cords on each side soon run close to each other, finally fusing into a single cord anterior to the suprarectal ganglion (Fig. 9C). In neomenioids, the cords remain apart and are well separated from each other (Fig. 9 A, D). There are few commissures between the ventral cords in chae- toderms, and many in neomenioids. In chaetoderms, the suprarectal commissure and precerebral ganglia are larger and more swollen than in neomenioids. Related to the loss of the ventral cord commissures, chaetoderms have entirely lost the foot and anterior pedal glands. The homology of mucous glands of the oral shield with pedal glands, proposed by S. Hoffman (1949), does not hold in TEM studies (Scheltema el ai. in press. Fig. 9A). In some species of Scut opus and Psilodens, ventral fusion of the mantle is marked by a longitudinal furrow between the spicules (Salvini-Plawen, 1968b; author's unpub. data). The gut of chaetoderms is modified from the simple combined stomach-digestive gland midgut of neomenioids to a separate stomach and blind digestive gland. In its most derived state in Chaetodermatidae, there is a gastric shield and style sac with a mucoid rod (Scheltema, 1978; Salvini-Plawen, 1981b). The serial lateroventral musculature of neomenioids (Fig. 8B) is lost in chaetoderms, although a few vestigial anterior bundles have been reported in a species of Scu- topus (Salvini-Plawen, 1985). Body form in chaetoderms is circular in cross-section; in neomenioids there is usually a small but measurable difference between height and width. The circulatory system is somewhat better denned in chaetoderms than in neomenioids, with anterior and posterior vertical septa defining hemocoelic sinuses, and with an often thick-walled aorta and aortal bulb (Schel- tema, 1973) (Fig. 5A). Finally, the chaetoderm oral shield represents a specialized cuticular structure. The autapomorphies of Chaetodermomorpha all seem to be related to their form of locomotion burrowing in muds and silts and feeding habits, either as carnivores on small benthic organisms, or as detritivores. Autapo- morphies also exist in the Neomeniomorpha, particularly the sensory vestibule and rather complicated reproductive system with accompanying loss of mantle cavity ctenidia, but specializations of the Chaetodermomorpha mark them as the more derived of the two taxa. Relationship of Aplacophora and Polyplacophora Aplacophora and Polyplacophora are here considered to be sister taxa, the Aculifera, on the basis of shared char- acters of nervous system, spicules, and epidermal papillae. An attempt is made to determine the polarities of these characters, and other anatomical similarities are noted. Nervous system Aplacophora, Polyplacophora, and the monoplacoph- oran Tryblidiacea Neopilina and Vema all have a fully developed tetraneury, with paired lateral and pedal nerve cords arising from a cerebral commissure or ganglia and a circumoral or circumesophageal nerve ring. In the monoplacophorans, both cords are joined posteriorly ventral to the rectum, whereas in Aplacophora (Fig. 9) and Polyplacophora, only the lateral cords are joined, and they unite above the rectum in a commissure (chitons) or ganglion (aplacophorans). There is only a single cross- pedal commissure in the monoplacophorans and numer- ous ones in chitons and neomenioid aplacophorans. What is the polarity of these two plans, both of which are plesiomorphic to more specialized nervous systems in other mollusks? Obvious outgroups for comparison, An- nelida, Echiura, Nemertini, and Sipuncula, appear to have a reduced nervous system and offer no clues. In Annelida there is only a paired ventral cord, except for a secondarily derived tetraneury in Amphinomidae (Gustafson. 1930); in Nemertini there is a pair of lateral cords joined either above or below the rectum; in Echiura, there is a single ventral cord; and in sipunculans there is also a single ven- tral cord which is paired in the larval pelagosphera of Phascolosoma agassizii (Rice, 1973). One might surmise that Neopilina, a deep-sea deposit or xenophyophore feeder (Tendal, 1985), is less mobile than either aplacoph- orans or polyplacophorans and has retained a simpler nervous system, and the aplacophoran-polyplacophoran system is more specialized (derived) owing to habitat (chi- tons) or to carnivory (aplacophorans). Of course, a sec- ondary loss and shifting of nerve elements in the mono- placophorans might also be considered, and Wingstrand APLACOPHORA: PROGENETIC COELOMATES 69 (1985) and Salvini-Plawen (1972) suggest that the sub- rectal commissure is an apomorphy. Whichever interpre- tation is correct, one can say that monoplacophoran and aplacophoran-polyplacophoran nervous systems are each apomorphic to some unknown ancestral state, and the suprarectal ganglion or commissure of the Aplacophora- Polyplacophora serves to relate them phylogenetically and set them apart from the Monoplacophora. Spicule formation Spicule formation in aplacophorans and polyplacoph- orans has most recently been investigated by Haas (1981) (Fig. 10). Spicules in both taxa are aragonite and formed extracellularly within an invagination of a single basal cell, which secretes calcium carbonate within a crystalli- zation chamber sealed by neighboring cells (Scheltema a al., in press, fig. 6D). In chitons, megaspines are formed from a proliferation of the basal cell and do not occur in aplacophorans. Spicules of the Aculifera are usually considered to be a plesiomorphic state of calcium carbonate formation within Mollusca, since both spines and shell occur in chi- tons and only spines occur in Aplacophora, both being "primitive" groups in the general sense. However, Mono- placophora, likewise considered primitive, have no spines. The dorsal, calcium-carbonate-secreting epidermis of Mollusca, in combination with a ventral locomotary sur- face, is probably an apomorphy. However, the shell-bear- ing Brachiopoda are rooted with the Mollusca-Annelida group by RNA sequencing (Field el al.. 1988). and some boring Sipuncula have calcium carbonate deposits at the dorsal anterior end of the trunk (Rice, 1969). Further comparative work needs to be done to compare calcium carbonate secretion among the Eutrochozoa before ho- mology can be assumed. It cannot be concluded from outgroup comparison that spicules and shell are homologous structures (and the ar- gument will be made further on that they are not), or that either is the plesiomorphic state. It can be concluded, however, that because of the way in which they are formed, spicules of Aplacophora and Polyplacophora are homol- ogous and can be construed as a synapomorphy. Epidermal papillae The epidermis of both chitons and aplacophorans are liberally supplied with secretory papillae (Fig. 6D, E). In chitons, papillae are homologous with aesthetes (Fischer c/ al.. 1980). Although homology with other conchiferan shell-penetrating structures has been suggested (Salvini- Plawen, 1985), the homology was considered spurious by Wingstrand, who reviewed the literature on the subject (1985, pp. 58-59). The presence of these papillae is COn- Kigure 10. Spicule formation in (A) Aplacophora and (B) Polypla- cophora (after Haas. 1981. figs. 6, 12). The spicule is formed within an invagination of a basal cell which secretes CaCOj. The crystallization chamber is sealed by a nng of neighboring cells, which in Polyplacophora produce a pellicle around the spicule. 1 spicule, 2 neighboring cell, 3 CaCO r secreting basal cell. sidered here to be an apomorphy of the Aplacophora- Polyplacophora. Rciluccd aerial replication Compared to Monoplacophora, there is less serial rep- lication in both Polyplacophora and Aplacophora, but both have greater serial replication than other mollusks. Serial replication appears as regular, lateroventral mus- culature in Neomeniomorpha (Fig. 8B) and as 8-fold rep- etition of muscles and shell plates in chitons. Other anatomical homologies Aplacophorans and polyplacophorans share certain other anatomical structures that are probably homolo- gous, but they may be plesiomorphies of the Mollusca. Dorsal paired gonads, becoming fused during ontogeny in chitons and most Chaetodermomorpha, lie like sacs more or less free above the gut and digestive gland in the dorsal hemocoel. In Neopilina, the gonad is ventral to the digestive system (Lemche and Wingstrand, 1959), and in many other Mollusca the gonad is intermingled closely with lobes of the digestive gland. The circulatory system in both groups is extremely open with posterior paired auricles and a ventricle, a dorsal aorta leading to the head (lacking in many Neomeniomorpha), and open sinuses, the latter more profuse in chitons. Taken together, the above reasons are sufficient for concluding that Aplacophora and Polyplacophora belong together in a single taxon, the Aculifera, which is therefore a clade, and not a grade. 70 A. H. SCHELTEMA Aculifera as the Sister Taxon of the Conchifera Chitons provide evidence that Aculifera are separate from their sister group, the Conchifera. The evidence is based on shell ontogeny, shell structure, and perhaps mo- lecular data. She/I ontogeny In Conchifera, the shell originates within an ectodermal invagination. the shell-field invagination. which is covered by an organic pellicle (Eyster and Morse, 1984) (Fig. 1 1 ). In Aeolidia papillosa, long cytoplasmic processes overlie the pellicle. In chitons, there is no shell field invagination, and shell plate anlagen are deposited within transverse depressions which are sealed, not by a pellicle, but by long, overlapping microvilli that lie beneath a gelatinous mucoid substance (Kniprath, 1980; Haas cl a/., 1980; Haas, 1981; see Scheltema, 1988, for a more complete discussion). Furthermore, in healthy larvae, shell is not deposited as separate granules, as illustrated by Kowalev- sky (1883), but as uninterrupted rods (Kniprath, 1980). This fact conflicts with the hypothesis that chiton shell arose from fused spicules (Salvini-Plawen, 1985. 1990). Shell structure The crystallography of chiton shell has been said to indicate an autapomorphy of chitons by Haas ( 1976), who found that "The . . . c-axis of [the] hypostracum lies in the bisectrix of the crystalline fibers. The whole complex acts crystallographically as a single crystal" (p. 392). If this crystallographic orientation is correct, then no ho- mology exists between polyplacophoran and conchiferan shell. Further differences are a lack of true periostracum in chitons (although Haas [ 1 98 1 ] has demonstrated a thin cuticle overlying the shell plates) and a lack of a nacreous layer (for further discussion see Wingstrand, 1985; and Scheltema, 1988). On the other hand, the shell of the try- blidiacean Monoplacophora does not differ from other primitive conchiferan shells (Lemche and Wingstrand, 1959). Molecular evidence The evolutionary tree of 1 8S rRN A has three branches for three classes of mollusks a nudibranch, two species of clam, and a chiton. This trifurcation of mollusks also appears in Lake's (1990) re-analysis of the data. Further molecular data for all molluscan classes should resolve the branching, but there is a hint of molecular distance between chitons and the two other classes analyzed. Evidence for Progenesis in Aplacophora A vermiform body is a character that could have been added rapidly by a small change in a regulatory gene or Figure 11. Shell deposition in larvae of Conchifera and Aculifera. (A, B) Gastropod, Aolidia papillosa- An organic pellicle (arrowheads) covers the lumen of the shell held invagination; a cytoplasmic extension shown in B seals the edge of the pellicle (after Eyster and Morse, 1984, figs. 1,2; from Scheltema, 1988, fig. 4). (C, D) Polyplacophoran, Ischno- chiton rissoi. The shell plate is first secreted beneath microvilli (stragulum) which are covered by a layer of mucus (C); later (D) the microvillar processes have pulled apart and a cuticle begins to form (after Kniprath. 1980, fig. 5, from Scheltema. 1988, fig. 4). Haas (1981) illustrated a similar process except for showing that cuticle covered the stragulum before CaCO, deposition. 1 shell field invagination, 2 cytoplasmic ex- tension, 3 microvillar process (stragulum), 4 calcium carbonate of shell plate, 5 mucous layer, 6 ?mucous cell, 7 cuticle. in timing of cell assembly early in the ontogeny of an aculiferan mollusk (for mechanisms and examples see Raff and Kaufman, 1983; McKinney and McNamara, 1991). In the embryological development of the chiton Lepido- pleums asellus, swimming larvae are first oval and then become secondarily flattened and sink to the bottom (Christiansen. 1954). Even with development of the foot, chiton larvae remain ovoid for a time (Heath, 1899, Fig. 52; Eernisse, 1988, Fig. 7). One can imagine that larvae of some aculiferan, not necessarily a chiton, might not APLACOPHORA: PROGENETIC COELOMATES 71 have become dorsoventrally flattened through a small change in gene regulation and the worm-like shape arose. The change to a vermiform shape could have occurred either early in the evolution of Mollusca or late. Recent phylogenies presume that a vermiform shape evolved as an early offshoot of the Mollusca, placing Aplacophora closest to the stem form, either as a monophyletic clade (Scheltema, 1988; Wingstrand, 1985), or as two separate clades, with the Chaetodermomorpha evolving first as the sister-group to all extant Mollusca (Salvini-Plawen, 1972, 1985). Serial replication thus was seen to he an apomor- phy. If Aplacophora are closest to the molluscan ancestor, then the imperatives following from that phylogenetic construct fit poorly with the arguments given above, that is: ( 1 ) Aplacophora and Polyplacophora are a clade; (2) shell is not formed by fusion of spicules; and (3) chiton shell is not homologous to conchiferan shell. The question of when serial replication evolved in mollusks becomes critical, for it is either a plesiomorphy of mollusks. or not. Polyplacophora, belonging to Aculifera, have some structures homologous with Monoplacophora, belonging to Conchifera, that are not shared with the Aplacophora (Wingstrand, 1985); radula dentition and radular appa- ratus including musculature; 8-serial pedal retractors; pre- oral unpaired fold, or velum; perhaps the heart with two pairs of atria; and coiled intestine. Wingstrand noted that some of these structures "could be plesiomorphic, i.e., could have been present already in some Aplacophoran ancestors" (1985, p. 74), but considered that the radula and radular apparatus, in particular the paired, hollow radula vesicles, are synapomorphies. It was not then known that paired radular vesicles are also present in some neomenioids (Fig. 8D). Here, structures argued to be apo- morphic by Wingstrand are considered plesiomorphic with exception of the coiled gut, a character widely con- vergent among mollusks. Thus, serial replication is here considered a plesiomorphy of Mollusca. The possibility that a worm shape was acquired by aplacophorans late in aculiferan evolution leads to a wholly different concept of molluscan phylogeny. It calls for progenesis in Aplacophora, wherein nonserial but ple- siomorphic-appearing anatomical characters are retained. The following evidence supports the hypothesis that Aplacophora are progenetic; i.e., that they have retained ancestral juvenile characters in adult form through ac- celeration of sexual maturation (Gould, 1977). ( 1 ) If narrowing of the body by acquisition of a worm shape arose early in aculiferan evolution without proge- nesis, then this process should be reflected somehow in the internal anatomy, and the more elongate (that is, nar- rower) the shape, the more pronounced should the internal changes become. Within the Neomeniomorpha, the least derived aplacophoran taxon, there is little organizational difference between short and elongate species in anterior and posterior ends or in musculature. Elongation of ex- ternal form is accompanied internally by a simple length- ening of the gonad and midgut (Fig. 7D, E). The situation in the more derived chaetoderms differs and does not serve the argument. A comparison can be made to Cryptoplax. a genus of chiton with a derived worm-like shape. In Cryptoplax there are at least four specializations of adult characters: (a) the mantle is very thick relative to internal body di- ameter; (b) there is loss of circulatory pathways; (c) there is loss of shell and shell musculature; and (d) the intestinal tract is remarkably long and complicated, turning back on itself in numerous spirals (Wettstein, 1904; H. Hoff- mann, 1929-30). Furthermore, an analysis of the allo- metric equation defining shape in 408 chiton species in 39 genera indicated great uniformity in allometry, except in the carnivorous Placiphorella and in genera of Cryp- toplacidae (Watters, 1991). Species ofCryptoplacidae, ex- cept those in the most primitive genus, are allometrically similar to each other but have shifted markedly from the allometry of other chitons. Although no allometric studies have been made of neomenioids, the extremes in ver- miformity (Fig. 7D, E) do not predict uniformity. Thus there may be an ontogenetic difference in the evolutionary pathway to a worm-like shape taken by the two aculiferan taxa. Progenesis, an intrinsic process, is hypothesized for Aplacophora, and selection working on structural genes, an extrinsic process, for Cryptoplax. (2) Progenesis results in early reproduction (Gould, 1977). One abundant northwestern Atlantic aplacophoran species living at 2000 m, Prochaetoderma yongei, is known to mature within one year, a remarkably rapid rate, given the ambient temperature (~3C) and in comparison with other cold-water mollusks. P. yongei is interpreted as being an opportunistic species (Scheltema, 1987), but since it is the only aplacophoran for which even part of the life history is known, one cannot be sure that early repro- duction is the usual case in Aplacophora. (3) Progenesis results in a reduced body size (Gould. 1977), but the size of the nearest ancestor to Aplacophora is, of course, unknown. Most neomenioids are usually less than 5 mm long, and one can only infer from the generally larger size of chitons that the first ancestral apla- cophoran was already small. Like some other deep-sea taxa, such as protobranch bivalves (Sanders and Allen. 1973) and isopods (Hessler et ai, 1979), aplacophorans have evolved primarily in the deep sea, where they reach their greatest diversity (Scheltema, 1990). Food is limiting there, and small body size of macrobenthic organisms is the norm (Monniot and Monniot, 1978; Allen, 1983; Soetaert and Heip, 1989). Large neomenioids do exist in the deep sea, but they are usually either specialized (giant Neomenia species: Baba, 1975; Kaiser, 1976) or live in environments where there is high productivity (e.g., high 72 A. H. SCHELTEMA latitudes: Proncomenia sluitcri, Derjugin. 1915, 1928). Large body size in Aplacophora is probably an apo- morphic character because it is found scattered amongst unrelated families, some of which have derived characters such as loss of radula or a thick dermis. (4) Certain structures in Aplacophora are less devel- oped than homologous structures in Polyplacophora or other mollusks. (a) The organic compos : *ion of the cuticle is simpler than in chitons (Beedham ano Trueman, 1968). (b) The radula in its plesiomorphic state in neomenioids has only two teeth per row (distichous), a condition found in the early ontogeny of several gastropods (Kerth, 1983; Scheltema, 1988; Scheltema ct a!., 1989). (c) The apla- cophoran mantle cavity, located ventroposterior to pos- terior, is small, serving as little more than a cloaca (Fig. 7A, D). (d) Both neomenioids and chaetoderms lack kid- neys, (e) The foot is developed only as a ciliated ridge without musculature in neomenioids (Fig. 6A). (f ) Gonads and pericardium are united in aplacophorans, reflecting the early ontogenic state in chitons, where the gonad orig- inates as an anlage of the pericardium (Hammarsten and Runnstrom, 1925) (Figs. 5 A, 6A, 9D). (g) The gut in neo- menioids is simple, with a united stomach and digestive gland; the digestive gland is separate from the stomach in other mollusks. (5) Aplacophora have retained a structure found in chitons only as larvae. The anterior pedal glands are large and specialized in neomenioids (Fig. 6B, C), but are lost soon after metamorphosis in chitons, where they serve only for early postmetamorphic attachment (Heath, 1899) (Fig. 4B). Although progenesis results in primitive-appearing structures, they are actually derived. Therefore, some process within the Aculifera should be primitive in the Polyplacophora but derived in the progenetic Aplacoph- ora. Such seems to be the case in early embryological development. Freeman and Lundelius (1992) have pro- posed that, among the spiralian coelomates Mollusca, Annelida, Sipuncula and Echiura, two mechanisms de- termine which blastomere is specified as the D quadrant. They hypothesized that the primitive mechanism for D quadrant specification is by induction after the fifth cleav- age, when one of the four macromeres has maximum contact with the micromeres. The derived mechanism is by segregation of the cytoplasm into one macromere, which is then specified as the D quadrant; it occurs by the second cleavage. In Polyplacophora, macromeres cleave equally and the D quadrant is specified by induc- tion, the primitive mechanism. But in the cleaving egg of the neomenioid Epimenia, a polar body is formed and therefore macromeres of unequal size; thus the D quadrant is specified by cytoplasmic determinants, the derived mechanism (Baba, 1951; Freeman and Lundelius, 1992). The evidence for progenesis presented here argues for heterochrony in the Aplacophora, but this idea cannot be tested either against fossils, which are unknown, or against a more complete phyletic lineage, as has been done for progenetic meiofaunal forms (Westheide, 1987) and deep- sea tunicates (Monniot and Monniot, 1978). When the early embryological development of aplacophorans is better known, and with further intrataxon comparative studies, the validity of the hypothesis may be clarified. Phytogeny of the Mollusca The phylogeny represented in Figure 12 proposes a coelomate molluscan ancestor with serial replication; two separate evolutionary molluscan lineages, the Conchifera and the Aculifera, based on synapomorphies of differences in CaCO, deposits; and morphologies arising from pro- genesis in the Aplacophora. The molluscan ancestor is considered to have had the following plesiomorphies: (1) extracellular CaCO 3 depo- sition by the dorsal epidermis (Mollusca generally); (2) serial replication, probably originally 8-fold (Monopla- cophora, Polyplacophora, Nautilus, neomenioids, some bivalves); (3) coelom from the 4d cell, paired pericardial cavities (in Monoplacophora, and fused but large in Apla- cophora and Polyplacophora); (4) radula, radular appa- ratus with hollow radula vesicles (Polyplacophora, Monoplacophora. Aplacophora, Fig. 8D); (5) nervous system poorly ganglionated, with cerebral ganglia and commissure, circumenteric ring, and paired lateral and pedal cords with cross-commissures and posterior con- nection (Monoplacophora in part, Aplacophora and Polyplacophora); (6) dorsoventrally flattened, small size (Cambrian Mollusca: Runnegar and Pojeta, 1985; Hasz- prunar, 1992; but note that the Cambrian fossil halkierids and Wiwaxia, perhaps near relatives of mollusks, are cen- timeters in length [Conway Morris, 1985; Conway Morris and Peel, 1990]); (7) dorsal cuticle (Aplacophora, Poly- placophora); (8) ventral ciliated locomotory sole (Mollusca generally); (9) head separate from the locomotory sole and with cerebral ganglia (Mollusca generally); (10) a groove between the dorsal and ventral surfaces, the future mantle cavity (Mollusca generally): (11) pre-oral fold; (12) the presence of podocytes in pericardial tissue (mollusks generally); (13) ductless anterior pedal mucous glands (as a glandular epithelium in Monoplacophora; Lemche and Wingstrand, 1959); (14) a one-way gut with mouth, anus, large digestive gland poorly differentiated from stomach (Neomeniomorpha, Monoplacophora); (15) paired pha- ryngeal diverticula; (16) poorly defined circulatory system; and (17) gonad and pericardium joined at least during ontogeny (Mollusca generally). The phylogeny presented in Figure 12 requires that the original calcium carbonate deposition in mollusks was APLACOPHORA: PROGENETIC COELOMATES 73 ACULIFERA APLACOPHORA CONCHIFERA 0. O tr oc 65 ili < o 39' 23 B -- 22' 39' 18 6-17 [c?| - 5 4 [d?| - 3 2 1 Figure 12. Proposed phylogeny of extant "primitive" Mollusca. (A) Apomorphies of Mollusca: I extracellular CaCO, deposition by dorsal epidermis; 1 eight-fold serial replication; 3 paired coelom, including pericardium; 4 radula; 5 poorly ganglionated tetraneury; 6 small size, dorsoventrally flattened; 7 dorsal cuticle; 8 ventral locomotory sole; 9 head separate from sole; 10 groove between dorsal and ventral surfaces; 1 1 pre-oral fold; 12 nonsegmented pericardium, pencardial tissue with podocytes; 13 ductless anterior pedal gland; 14 poorly differentiated stomach/digestive gland (model: Neopilina); 15 paired pharyngeal di- verticula; 16 poorly denned circulatory system; 17 joined gonad/pen- cardium during early ontogeny. (B) Separation of Conchifera and Acu- lifera: 18 calcareous shell; 19 spicules; 20 epidermal papillae; 21 supra- rectal ganglion/commissure; 22 reduced serial replication and fused pericardium. (C) Separation of Polyplacophora and Aplacophora (24-31 the result of progenesis): 23 eight shell plates; 24 worm shape; 25 reduced foot; 26 reduced mantle cavity; 27 joined gonad/pericardium; 28 kidneys absent; 29 chemically simple cuticle; 30 serial lateroventral musculature; 31 distichous radula; 32 U-shaped gametoducts; 33 gan- glionated nervous system; 34 dorsoterminal sense organ. (D) Separation of Chaetodermomorpha and Neomeniomorpha: 35 ventrally fused cu- ticle, foot lacking; 36 oral shield; 37 fused, reduced nervous system; 38 serial replication absent; 39 stomach separate from digestive gland; 40 large anterior pedal gland; 41 elaborated reproductive system; 42 ctenidia absent. * = convergent morphologies; c? = presence of ctenidia ques- tionable; d? = radula questionably docoglossate. neither as spicules nor as shell. CaCO, was first deposited, perhaps, as granules within a dorsal cuticle, which was thereby stiffened. Such a reinforced cuticle could act as the antagonist to the dorsoventral pedal musculature. During chiton ontogeny, the pedal musculature develops earlier than the shell plates (Hammarsten and Runnstrom, 1925). One can speculate from this fact that, perhaps, the various forms of shell and spicules among mollusks have resulted from selection for different modes of locomotion in various habitats, rather than selection just for protec- tion. In terms of CaCO 3 secretion among phyla, the impor- tant synapomorphy for mollusks, which sets them off from other spiralian coelomates, is the locomotary sole in com- bination with a cuticle- and CaCO r secreting dorsal epi- dermis. Certain rock-boring sipunculans also secrete CaCOj dorsally, forming a plug for their tubes (Rice, 1969), and Brachiopoda, which fall in with spiralian coe- lomates in molecular analysis, also have calcium carbon- ate shells. However, animals in neither of these phyla have the combination of dorsally produced CaCO, and a ven- tral locomotary surface unique to mollusks. It is hypothesized that after, or as, Conchifera diverged from the stem line, the mantle deepened and gills devel- oped. Serial replication was retained in Monoplacophora but lost in the rest of the Conchifera, except for serial pedal musculature in some taxa and the renal system in cephalopods. Aculifera may have evolved either at the same time as Conchifera or later. By the Upper Cambrian or Lower Ordovician, the serial shell plates of Polypla- cophora had evolved (Runnegar and Pojeta, 1985). This event was preceded by the loss of serial replication other than lateroventral muscles and perhaps by an increase in size. In a separate evolutionary event of progenesis, the Aplacophora evolved with probable reduction in size, further loss of serial replication, loss of nephridia, retention of gonad-pericardial connection, and acquisition of a worm shape with concomitant reduction of the foot. Chaetodermomorpha were derived from the neomenioid- like stem with complete loss of foot, reduction and fusion of the nervous system, and specializations of the gut. This hypothesized phylogeny does not call for an evo- lutionary process in which CaCO, deposits, or the cells that produce them, become fused. Furthermore, it should allow some of the Early Cambrian sclerite-bearing forms now coming to light, such as the shell-bearing, articulated halkieriid described recently from the Lower Cambrian of Greenland (Conway Morris and Peel, 1990), to find their place in relation to the extant Mollusca. In this phylogeny, the Monoplacophora with clear serial replication are not evolved after Aplacophora, and mol- luscan serial replication is considered to be a plesiomor- phy. As Wingstrand (1985) pointed out, it is difficult to imagine that serial replication evolved after the shell. The 74 A. H. SCHELTEMA careful and original anatomical analysis of Wingstrand showing close affinities of the monoplacophoran Trybli- diacea and Polyplacophora are upheld here as retained plesiomorphies of the common ancestor. Whether Pru- vot's neomenioid larva with its supposed seven rows of spicules actually exists does not change the argument (see Salvini-Plawen, 1972, 1981a, 1985; Scheltema, 1988 for discussions and figures of the larva). Manuscript drawings of Chaetoderma nitidithtm larvae made by G. Gustafson show eight rows of spicules for this taxon as well. If further observations on aplacophoran development prove that serial rows of spicules do exist, the larva still would not necessarily reflect progressive evolution from spicules to fused shell plate formation, but more likely would indicate a breakdown of plate formation similar to the breakdown of larval chiton shell plates caused experimentally by Kniprath (1980) (See also Scheltema, 1988). Age of the Aplacophora If known fossils reflect the actual time of evolutionary events, then the evolution of Polyplacophora late in the Cambrian (Runnegar and Pojeta, 1985) from a continuing line of aculiferous creatures was probable, with increased size and muscles being the determinants of shell plates rather than vice versa (see Hammarsten and Runnstrom, 1925, p. 276, for ontogenetic development of muscle be- fore shell). Aplacophora, with their highly derived shape and paedomorphic internal organization, give information about the primitive conditions of mollusks without being themselves primitive. A Late Cambrian-Early Ordovician origin from an aculiferan form with a developed mantle groove and posterior mantle cavity is postulated for Apla- cophora, with the 8-fold dorsoventral muscles rearranged in neomenioids into a series of indeterminate number. Cautionary Notes on Convergences Digestive system The molluscan gut appears to have evolved similar morphologies more than once (Fig. 12, no. 39). Evidence for convergence lies in presence of the style sac and gastric shield, found in a number of molluscan classes. In the aplacophoran family Chaetodermatidae, one of the most derived of the chaetoderm groups based on radula mor- phology (Scheltema, 1972, 1981), the gut is the most complicated among chaetoderms. with a gastric shield and a mucoid rod in a style sac (Scheltema, 1978; Salvini- Plawen, 1981b). The polarity of a less to a more compli- cated gut configuration within the chaetoderms is clear (Scheltema, 1981). Thus, the presence of a style sac and gastric shield is convergent among Mollusca. Metamerism Reduction of serial replication (Fig. 12, no. 22) is hy- pothesized for several molluscan classes Cephalopoda, Bivalvia, Polyplacophora. and Aplacophora. The evidence from morphology, ontogeny, and molecular analysis seems not to favor the hypothesis that replication origi- nated in annelids. If the altogether unsegmented Sipuncula are sister taxon of the mollusks, then arguments that the molluscan coelom is the result of a reduced annelid-like segmented coelom are not convincing. Evidence presented here could be interpreted in three ways (Fig. 13, s'-s 4 ). (1) A nonsegmented ancestor that had serial replication of organs and a coelom lies at the base of the lineage giving rise to Eutrochozoa (s 1 ). (2) The eutrochozoan ancestor had no serial replication, which O z O 0) Q _l UJ 9* 6-8 o 2. 10 11 9* S3 1-5 Figure 13. Phylogenetic relationship among Sipuncula, Mollusca. and Annelida. 1 spiral cleavage; 2 paired coelom originating from two teloblasts derived from 4d; 3 trochophore larva (?); 4 tetraneury; 5 ciliated creeping sole; 6 molluscan cross; 7 ventral, cuticular, pharyngeal (sto- madeal), protrusihle invagmation and attendant musculature; 8 anterior pedal gland; 9 fused nerve cords; 10 reduced coelom; 1 1 loss of creeping sole, s = serial replication: s 1 symplesiomorphic for all three taxa, but lost in Sipuncula; s 2 symplesiomorphic for Sipuncula and Mollusca, but lost in Sipuncula, and convergent with s 1 as metamerism in annelids; s 3 metamerism plesiomorphic for Annelida, convergent with either s 2 or s 4 ; s 4 plesiomorphic for Mollusca, convergent with s 1 . * = convergent morphologies. APLACOPHORA: PROGENET1C COELOMATES 75 later arose de novo twice: once in the stem form leading to mollusks and sipunculans (s 2 ). which was lost in the latter, and secondly in the ancestral annelid as metamer- ism (s 3 ). (3) Molluscan 8-fold serial replication (s 4 ) evolved after the stem form that gave rise to Sipuncula and Mol- lusca; and annelidan metamery (s 3 ) [as in (2)], arose as an unrelated evolutionary event. The first interpretation is perhaps closest to what may actually have occurred and seems the most parsimonious explanation. Differences in the coelom among eutrochozoan groups can be related to locomotion, a theme emphasized cor- rectly, I believe by Salvini-Plawen (e.g., 1972, 1985). Locomotion among Eutrochozoa is most rapid in annelids and mollusks. Serial pedal musculature is related to a creeping locomotion and is the most conservative serial structure in mollusks, present as a plesiomorphy in Monoplacophora. Polyplacophora, Aplacophora, and (much reduced) Pelecypoda and perhaps the neritid Gas- tropoda. In Annelida, coelom and muscle have combined in the perfection of a hydraulic locomotion (Clark, 1964). Perhaps, then, a re-examination of the relationship of muscles and coelom during ontogeny would be a useful exercise in providing insights into understanding the de- velopment of metamerism in Eutrochozoa. For instance, in at least some Annelida, ectodermal segmentation of the three anterior segments precedes segmentation of me- soderm (Anderson, 1973, pp. 36-37). Radula Wingstrand (1985) gave a detailed description of the radular apparatus in Polyplacophora and Monoplacoph- ora, demonstrating their great similarity, especially the docoglossate radula and radula vesicles. There are three possibilities: such a radula is a molluscan plesiomorphy; it is an apomorphy of Polyplacophora and Monopla- cophora; or it is convergent. Evidence from Aplacophora and ontogeny of some Gastropoda suggests that the plesiomorphic radula in mollusks was distichous (Keith, 1983; Scheltema el ai, 1989). An outgroup for comparison is the Cambrian sclerite-bearing Wiwaxia (Conway Morris, 1985) with two or three rows of teeth which appear much like the ple- siomorphic radula in Aplacophora. The phylogenetic po- sition of Wiwaxia. however, remains enigmatic, consid- ered either to be close to mollusks (Conway Morris, 1985) or to be an annelid (Butterfield, 1990). If the plesiomor- phic radula is distichous, then the docoglossate radula is convergent in Polyplacophora, Monoplacophora, and patellacean Gastropoda. If the docoglossate radula is a molluscan plesiomorphy, it is difficult to imagine how it functioned in a small Cambrian mollusk and what evo- lutionary steps would be necessary to account for all other molluscan radulae. The strongest evidence given by Wingstrand (1985) for monophyly of polyplacophorans and conchiferans is presence of a pair of hollow, presumably liquid-filled rad- ula vesicles found at that time only in Polyplacophora and Monoplacophora. None had been reported in Apla- cophora. However, a re-examination of the neomenioid Helicoradomenia juani and other species in the genus, which have a plesiomorphic aplacophoran radula, has led me to conjecture that paired, elongate, hollow vesicles present in this genus are a homolog to the radula vesicles in Polyplacophora and Monoplacophora (Fig. 8D). Therefore these vesicles are a molluscan plesiomorphy. However, further study of the aplacophoran radula and its apparatus is needed. Larval forms The phylogenetic significance of larval forms in Spiralia is not addressed here. There is still no agreement on whether a pelagic organism gave rise to benthic forms U'.t,'.. Nielsen and Norrevang, 1985), or vice versa, and whether the trochophore larva arose once or several times (Ivanova-Kazas, 1985a, b, for careful discussions). Within Mollusca. Salvini-Plawen ( 1972, 1985) regarded the peri- calymma larva, which lacks purely larval organs except the swimming test and is found only in aplacophorans and protobranch bivalves, as the ancestral type. The questions are left here as unresolved and not affecting the arguments for homology of early cell fate among Eutro- chozoa, although my preference is indicated by use of the latter term. Classification of Extant Molluscan Classes With shell and spicules considered as synapomorphies for Conchifera and Aculifera, respectively, the following classification of extant Mollusca emerges: Phylum Mollusca Subphylum Conchifera Class Monoplacophora Class Bivalvia Class Gastropoda Class Scaphopoda Class Cephalopoda Subphylum Aculifera Class Polyplacophora Class Aplacophora Subclass Neomeniomorpha Subclass Chaetodermomorpha This arrangement is similar to that already proposed in the last century with little knowledge of the soft anat- omy of Monoplacophora. Garstang (1896) considered the 76 A. H. SCHELTEMA Aplacophora as "degraded" from an ancestral chiton-like form, but although he later stressed the importance of paedomorphosis in evolution, he did not see it as per- taining to Aplacophora. It is curious that a classification based on what are here inferred to be synapomorphies and on progenesis should be much the same as classifi- cations of a hundred years ago. Conclusions The hypotheses, arguments, and pieces of evidence presented here lead to the conclusions that Mollusca ( 1 ) are eucolomates with an ancestry in common with spiral- ian trochozoans; (2) are related to Annelida, but not as closely as they are to Sipuncula: (3) have a reduced coelom which was never segmented; (4) are not directly descended from an aplacophoran-like or turbellariomorph prede- cessor; and (5) are descended from an ancestor with serial replication. Acknowledgments The idea that aplacophorans may have evolved through progenesis originally came from David R. Lindberg. I have benefitted from critical discussions with Dave and with Carole S. Hickman, Bruce Runnegar, Claus Nielsen, Tom Waller, Douglas Eernisse, and Bertil Akesson, all of whom also steered me towards relevant literature. Doug Eernisse and Bruce Runnegar read an earlier version of this paper as well. Thanks are also due to Gerhard Haszprunar, who read the manuscript in its present form. Two reviewers most helpfully suggested literature that I had overlooked. I have tried to keep the phylogeny presented here as straightforward as criticisms and discussions suggested, and hopefully there is not too much "story telling." My gratitude goes to each of my critics. Mary E. Rice opened the way for an understanding of the Sipuncula-Mollusca relationships presented here, and the visit with her at the Smithsonian Marine Station at Link Port, Ft. Pierce, Florida, afforded the opportunity to work with both living pelagosphera larvae and apla- cophorans. I thank her deeply, and for prints of the splen- did photographs of pelagosphera. I thank Franz P. Fischer for providing me with a copy of his photograph of polyplacophoran epidermis, and Claus Nielsen for a copy of G. Gustafson's original draw- ings of Chaetoderma nitidulum larvae. As always. I gratefully acknowledge helpful discussions with Rudolf Scheltema, who has provided me space and who has always taken an energetic interest in my work. The following credits for previously published illustra- tions are acknowledged: figures 5A, C, D, 7C, 8A, and 1 1 from American Malacological Bulletin 6 (1988): 57-68, figs. 4, 5, 6, 13; figure 2A, B from American Zoologist 21 (1981): 605-619, figs. 4, 6; figure 3 from Smithsonian Contributions to Zoology 132 ( 1973): pi. 5; figures 6A, D, and 8D from Ve liger 34 (1991): 195-203, figs. 2C, 4D, and 5C; figure 6E from Zoomorphologie 94 (1980): 1 2 1 - 1 3 1 , fig. 3; figure 7 A, B from Biological Bulletin 169 (1985): 484-529, fig. 3L, O, P. Note added in proof: Two papers have just been published that have direct bearing on the ideas presented here. (1) Bengston S. 1992. The cap-shaped Cambrian fossil Maikhanella and the relationship between coelosclentophorants and molluscs. Lethaia 25: 401-420. Maikhanella is a genus of Lower Cambrian halkienids with a cap-shaped shell formed of rows of embedded spicules, which were added by marginal accretion. Bengston discusses the possible homology with spicules of extant mollusks and with polyplacophoran shells. (2) Eernissee. D. J.. J. S. Albert, and F. E. Anderson. 1992. Annelida and Anthropoda are not sister taxa: a phylogenetic analysis of spiralian and metazoan morphology. Sysl. Biol. 41: 331-344. Analysis by maximum parsimony among 141 morpholog- ical and embryological characters supports the concept of Eutrochoza. including the Mollusca. Literature Cited Allen, J. A. 1983. The ecology of deep-sea molluscs. Pp. 29-75 in The Mollusca. Vol. 6- Ecology. W. D. Russell-Hunter, ed. Academic Press, Orlando, FL. Anderson, D. T. 1973. Embryology and Phytogeny in Annelids and Arthropods. Pergamon Press, Oxford and New York, xiv + 495 pp. Andrews, E. B. 1988. Excretory systems of molluscs. Pp. 381-448 in The Mollusca. Vol. 11. Form and Function, E. R. Trueman and M. R. Clarke, eds. Academic Press, San Diego, CA. Baba, K. 1938. The later development of a Solenogastre, Epimenia verrucosa (Nierstrasz). / Dept Agric. Kyitsyu Imp. Univ. 6: 21-40. Baba, K. 1951. General sketch of the development in a Solenogastre. Epimenia verrucosa (Nierstrasz). Misc. Repts. Res. Insl. Nat. Res. 19-21: 38-46. [In Japanese with English summary'-] Baba, K. 1975. Neomenia yamamoloi spec, nov., a gigantic Solenogaster (Mollusca: Class Solenogastres. Family Neomeniidae). occurring in the north-eastern part of Japan. Publ. Seto Mar. Biol. Lab. 22(5): 277-284. Banta, W. C., and M. E. Rice. 1976. A restudy of the Middle Cambrian Burgess Shale fossil worm, Oltoia prolifica. Pp. 79-90 in Proceedings of the International Symposium on the Biology oj the Sipuncula and Echiura, M. E. Rice and M. Todorovic, eds. Institute for Biological Research "SmisStankovic," Belgrade, and Smithsonian Institution, Washington, DC. Beedham, G. E., and E. R. Trueman. 1968. The cuticle of the Apla- cophora and its evolutionary significance in the Mollusca. ,/ Zoo/. 154: 443-451. Butterfield, N. J. 1990. A reassessment of the enigmatic Burgess Shale fossil Wiwaxia corrugata (Matthew) and its relationship to the poly- chaete Canadia spinosa Walcott. Paleobiology 16: 287-303. Christiansen, M. E. 1954. The life-history of Lepidopleurus asellus (Spengler) (Placophora). Nytt Mag. Zool. 2: 52-72. Clark, R. B. 1964. Dynamics in Metazoan Evolution: The Origin of the Coelom and Segments Clarendon Press, Oxford. 313 pp. Conway Morris, S. 1985. The Middle Cambrian metazoan \fiwaxia corrugata (Matthew) from the Burgess Shale and Ogygopsis Shale. British Columbia. Canada. Phil Trans. Roy. Soc. London. B 307: 507-586. Conway Morris, S., and J. S. Peel. 1990. Articulated halkienids from the Lower Cambrian of north Greenland. Nature 345: 802-805. APLACOPHORA: PROGENETIC COELOMATES 77 Curry, G. B., and B. Runnegar. 1990. Partial amino acid sequences of hemerythrins from Lmgula and a pnapulid worm, and the evolution of oxygen transport in early metazoans. No. 4314, Geol. Soc. Am Abstr Dallas, TX. Derjugin, K. M. 1915. [Fauna of the Kola gulf and its life conditions.]. .Mem Acad Imp Sci. Petrograd (ser. 8)34(1): 1-929, t. 1-14 [Pp. 499-501 Solenogastres]. [In Russian] Derjugin, K. M. 1928. Fauna des Weissen Meeres und ihre Existenz- bedingungen. Exp. Mcrs V.R.S.S. fasc. 7-8, 1-511 [Pp. 294-295, Solenogastres]. Eernisse. D. J. 1988. Reproductive patterns in six species of Lepi- dochitona (Mollusca: Polyplacophora) from the Pacific Coast of North America. Biol Bull 174: 287-302. Eyster, L. S., and M. P. Morse. 1984. Early shell formation during molluscan embryogenesis, with new studies on the surf clam, Spisula so/idissima. Am. Zoo/ 24: 871-882. Field, K. G., G. J. Olsen, D. J. Lane, S. J. Giovannoni, M. T. Ghiselin, E. C. Raff, N. R. Pace, and R. A. Raff. 1988. Molecular phylogeny of the animal kingdom. Science 239: 748-753. Fischer, F. P., W. Maile, and M. Renner. 1980. Die Mantelpapillen und Stacheln von Acanthochitonfascicularis L. (Mollusca, Polypla- cophora). Zoomorphologie 94: 121-131. Freeman, G., and J. \V. Lundelius. 1992. Evolutionary implications of the mode of D quadrant specification in coelomates with spiral cleavage. ./ Evol Biol 5: 205-247. Garstang, VV. 1896. On the aplacophorous Amphineura of the Bntish Seas. Proc. Malacol. Soc. London 2: 123-125. Gerould, J. H. 1906. The development of Phascolosoma. Studies on the embryology of the Sipunculidae I. Zoo/ Jalirh.. Anal. 23: 77-162. Ghiselin. M. T. 1988. The origin of molluscs in the light of molecular evidence. Pp. 66-95 in Oxford Surveys in Evolutionary Biology Vol. >. P. H. Harvey and L. Partridge, eds. Oxford University Press. Ox- ford. Gould. S. J. 1977. Ontogeny and Phylogeny. Harvard University Press, Cambridge. MA, and London. Gustafson, G. 1930. Anatomische Studien uber die Polychaten-familien Amphinomidae und Euphrosynidae. Zoo/. Bidr. Uppsala 12: 305-471. Haas, W. 1976. Observations of the shell and mantle of the Placophora. Pp. 389-402 in The Mechanisms of Mineralisation in Invertebrates and Plants, N. Watabe and K. M. Wilbur, eds. Belle W. Baruch Library in Marine Science 5. Haas, VV. 1 98 1 . Evolution of calcareous hardparts in primitive molluscs. Malacologia 21: 403-418. Haas, W ., K. Kriesten, and N. Watabe. 1980. Preliminary note on the calcification of the shell plates in chiton larvae (Placophora). Pp. 67-72 in The Mechanisms of Biomineralization in Animals and Plants. M. Omori and N. Watabe, eds. Tokai University Press, Tokyo. Hammarsten, O. D., and J. Runnstrom. 1925. Zur Embryologie von Acanthochiton discrepans Brown. Zoo/. Jahrb. Anal. 47: 261-318. Haszprunar, G. 1987. The fine morphology of the osphradial sense organs of the Mollusca. IV. Caudofoveata and Solenogastres. Phil Trans Roy Soc. London 5315: 63-73. Haszprunar, G. 1992. The first molluscs small animals. Boll. Zoo/. 59: 1-16. HeaCh, H. 1899. The development of Ischnochiton. Zoo/. Jahrb.. Anal. 12: 567-656. Heath, H. 1904. The nervous system and subradular organ in two genera of Solenogastres. Zoo/. Jahrb . Anal. 20: 399-408. Heath, H. 1905. The morphology of a Solenogastre. Zoo/. Jahrb. Anal. 21: 703-734. Heath, H. 1911. The Solenogastres. Reports on the scientific results of the expedition to the tropical Pacific . . . by the "Albatross" . . . Mem. Mus. Comp Zoo/. (Harvard University) 45(1): 1-179. Hessler, R. R., G. D. Wilson, and D. Thistle. 1979. The deep-sea isopods: a biogeographic and phylogenetic overview. Sarsia 64: 67- 75. Hoffman, S. 1949. Studien uber das Integument der Solenogastren nebst Bemerkungen uber die Verwandtschaft zwischen den Solenogastren und Placophoren. Zoo/ Bidr. Uppsala 27: 293-427. Hoffmann, H. 1929-30. Aplacophora. Pp. 1-134, 383-453 in Bronns Klasscn und Ordnungen des Tier-Reichs. Band III, Abt. I. Hyman, L. H. 1967. The Invertebrates. Vol. VI. Mollusca I McGraw- Hill, Inc., New York, vii + 792 pp. Ivanov, D. L. 1991. [Tentacular region of the central nervous system of the Mollusca and the homology of the organs of locomotion.] Rulhenica 1: 81-89. [In Russian] Ivanova-Kazas, O. M. 1985a. Origin and phylogenetic significance of trochophore larvae. Zoo/. Zluir 64(4): 485-497. [In Russian. English translation no. 1112119, National Museum of Canada, in manu- script.] Ivanova-Kazas, O. M. 1985b. Origin and phylogenetic significance of trochophore larvae. Zoo/. Zhur. 64(5): 650-660. [In Russian. English translation no. 1112120, National Museum of Canada, in manu- script.] Jagersten, G. 1947. On the structure of the pharynx of the Archian- nelida with special reference to there occurring muscle cells of aber- rant type. Zoo/ Bull. Uppsala 25: 551-570. Jagersten, G. 1963. On the morphology and behaviour of Pelagosphaera larvae (Sipunculoidea). Zoo/ Bidr. Uppsala 36: 27-35. Jagersten, G. 1972. Evolution of the Metazoan Life Cycle Academic Press. London and New York. Kaiser, P. 1976. Neomenia herwigi. sp. n.. ein bemerkenswerter Ver- treter der Solenogastren (Mollusca, Aculifera) aus Argentimschen Schelfgewassern. Mitt llamb Zoo/ Mus Insi 73: 57-63. Kerth, K. 1983. Radulaapparat und Radulabildung der Mollusken. II. Zahnbildung. Abbau und Radulawachstum. Zoo/. Jahrb.. Anal 110: 239-269. Kniprath. E. 1980. Ontogenetic plate and plate field development in two chitons, Middendorffia [= Lepidochitona] and Ischnochiton. IVilhelm Roux's Arch. Devel. Biol. 189: 97-106. Kowalevsky, M. A. 1883. Embryogeme du Chiton polii Phil., avec quelques remarques sur le developpement des autres Chitons. Ann. Mus. Hist. Nat. Marseille 1: 1-46. Lake, J. A. 1990. Origin of the Metazoa. Proc. Nat. Acad. Sci. USA 87: 763-766. Lemche, H., and K. G. Wingstrand. 1959. The anatomy of Neopilina galatheae Lemche, 1957 (Mollusca Tryblidiacea). Galathea Kept. 3:9-71. McKinney, M.L., and K. L. McNamara. 1991. Heterochrony: the Evo- lution of Ontogeny. Plenum Press, New York and London, xi.x + 437 pp. Monniot, C., and F. Monniot. 1978. Recent work on the deep-sea tu- nicates. Oceanogr Mar. Biol. Ann. Rev 16: 181-228. Moor, B. 1983. Organogenesis. Pp. 123-177 in The Mollusca. Vol. 3. Development. N. H. Verdonk, J. A. M. van den Biggelaar, and A. S. Tompa, eds. Academic Press, New York. Newby, W. W. 1940. The embryology of the echiuroid worm Vreehis caupo. Mem. Am. Philos. Soc. 16: 1-213. Nielsen, C., and A. Narrevang. 1985. The trochaea theory: an example of life cycle phylogeny. Pp. 28-41 in The Origins and Relationships of Lower Invertebrates. Syst. Assoc. Spec. Vol. No. 28. S. Conway Morris, J. D. George, R. Gibson, and H. M. Platt. eds. Clarendon Press, Oxford. 78 A. H. SCHELTEMA Plate, L. H. 1898. Die Anatomic und Phylogenie der Chitonen. Zool. Jahrb.. Suppl IV (Fauna Cln/ensis I): 1-243. Raff, R. A., and T. C. Kaufman. 1983. Embryos, Genes, and Evolution. Macmillan Publishing Co., Inc., New York & Collier Macmillan Publishers, London. Raven, C. P. 1966. Morphogenesis: The Analysis of Mottuscan Devel- opment. 2nd ed. Pergamon Press, Oxford etc. xiii + 365 pp. Reynolds, P. I)., and M. P. Morse. 1991. Morphological evidence for ultrafiltration of blood in the Aplacophora. Amer Zool 31: 137A. Rice, M. K. 1969. Possible bonng structures of Sipunculids. Amer /.ool 9: 803-812. Rice, M. E. 1973. Morphology, behavior, and histogenesis of the pe- lagosphera larva of Phascolosoma agassizii (Sipuncula). Smithsonian Contr. Zool. 132: 51 pp. Rice, M. E. 1975. Sipuncula. Pp. 67-127 in Reproduction of Marine Inver/ehrali's. \'ol. 2. Enloprocls and lesser coeloma/es, A. Giese, and J. Pearse, eds. Academic Press, New York. Rice, M. E. 1981. Larvae adrift: Patterns and problems in life histories of sipunculans. Am. Zool. 21: 605-619. Rice, M. E. 1985. Sipuncula: developmental evidence for phylogenetic inference. Pp. 274-296 in The Origins and Relationships of Lower Invertebrates, S. Conway Morns, J. D. George, R. Gibson, and H. M. Platt, eds. Oxford University Press, Oxford. Runnegar, B., and J. Pojeta, Jr. 1985. Origin and diversification of the Mollusca. Pp. 1-57 in The Mollusca. I '<>!. 10, Evolution. E. R. True- man, and M. R. Clarke, eds. Academic Press, Orlando, FL. Salvini-Plawen, L. v. 1968a. Die 'Funktions-Coelom Theorie' in der Evolution der Mollusken. Sysl. Zool. 17: 192-208. Salvini-Plawen, L. v. 1968b. Uber Lebendbeobachtungen an Caudo- foveata (Mollusca. Aculifera), nebst Bemerkungen zum System der Klasse. Sarsia 31: 105-126. Salvini-Plawen, L. v. 1972. Zur Morphologic und Phylogenie der Mol- lusken: Die Beziehungen der Caudofoveata und der Solcnogastres als Aculifera, als Mollusca und als Spiralia. Z. wiss. Zool. 184: 205-394. Salvini-Plavven, L. v. 1978. Antarktische und subantarktische Solen- ogastres (eine Monographic: 1898-1974). Zoohgica 44: 1-315. Salvini-Plawen, L. v. I981a. On the origin and evolution of the Mol- lusca. Alt Convegni Lincei 49: 235-293. Salvini-Plawen, L. v. 1981 b. The molluscan digestive system in evo- lution. Malacologia 21: 371-401. Salvini-Plawen, L. v. 1985. Early evolution and the primitive groups. Pp. 59-150 in The Mollusca. Vol. 10. Evolution. E. R. Trueman, and M. R. Clarke, eds. Academic Press, Orlando, FL. Salvini-Plawen, L. v. 1990. Origin, phylogeny and classification of the phylum Mollusca. Ibents9: 1-33. Sanders, H. L., and J. A. Allen. 1973. Studies on deep-sea Protobran- chia (Bivalvia); Prologue and the Pristiglomidae. Bull. Mus. Comp. Zool. 145: 237-262. Scheltema, A. H. 1972. The radula of the Chaetodermatidae (Mollusca. Aplacophora). Z Morph. Tiere72: 361-370. Scheltema, A. H. 1973. Heart, pericardium, coelomoduct openings, and juvenile gonad in Chaetoderma nitidulum and Falcidens cau- dalus (Mollusca, Aplacophora). Z. Morph. Ticreld: 97-107. Scheltema, A. H. 1978. Position of the class Aplacophora in the phylum Mollusca. Malacologia 17: 99-109. Scheltema, A. II. 1981. Comparative morphology of the radula and alimentary tracts in the Aplacophora. Malacologia 20: 361-383. Scheltema, A. H. 198?. The aplacophoran family Prochaetodermatidac in the North American Basin, including Chevroderma n.g. and Spa- ihoderma n.g. (Mollusca; Chaetodermomorpha). Biol. Bull 169: 484-529. Scheltema, A. H. 1987. Reproduction and rapid growth in a deep-sea aplacophoran mollusc, Prochaetoderma rongei. Mar. Eeol. Progr. Set: 37: 171-180. Scheltema, A. H. 1988. Ancestors and descendents: relationships of the Aplacophora and Polyplacophora. Amer. Malacol. Bull 6: 57-68. Scheltema, A. H. 1990. Aplacophora as a Tethyan slope taxon: evidence from the Pacific. Bull. Mai: Sci. 47: 50-61. Scheltema, A. H., K. Kerth, and A. M. Kn/irian. 1989. The primitive molluscan radula. P. 220 in L'nitas Malacologicu Iff 1 ' International Congress, Tubingen. [Abstract] Scheltema, A. H., and A. M. ku/irian. 1991. Helieoradomema juani n.g. n. sp.. a Pacific hydrothermal vent Aplacophora (Mollusca, Neomeniomorpha). I'eliger34: 195-203. Scheltema, A. H., M. Tscherkassky, and A. M. Kuzirian. (In press). Aplacophora. In Microscopic Anatomy 01 'Invertebrates. \'ol. 5. Mollusca: Monoplacophora, Aplacophora. Polyplacophora, and Gastropoda. F. W. Harrison and A. J. Kohn. eds. Wiley-Liss, New- York. Soetaert, K., and C. Heip. 1989. The size structure of nematode as- semblages along a Mediterranean deep-sea transect. Deep-Sea Res 36:93-102. Spengel, J. \V. 1881. Die Geruchsorgane und das Nervensystem der Mollusken. Z wiss. Zool. 35: 333-383. Tendal, O. S. 1985. Xenophyophores (Protozoa, Sarcodina) in the diet of Neopilma galatheae (Mollusca. Monoplacophora). Galalhea Repi 16: 95-98. Thiele, J. 1902. Die systematische Stellung der Solenogastren und die Phylogenie der Mollusken. Z wiss. Zool. 12: 249-466. Tscherkassky, M. 1989. Pedal shield or oral shield in Caudofoveata (Mollusca. Aculifera)? P. 254 in Unilas Malacologica, Abstracts of l(f h International Malacological Congress. Tubingen. [Abstract] Turbeville, J. M. 1986. An ultrastructural analysis of coelomogenesis in the hoplonemertine Prosorhochmus americanus and the poly- chaete Magelona sp. / Morphol. 187: 51-60. Turbeville, J. M., K. G. Field, and R. A. Raff. 1992. Phylogenetic po- sition of phylum Nemertini, inferred from 18S rRNA sequences: molecular data as a test of morphological character homology. Mol. Biol. Evol. 9: 235-249. van Dongen, C. A. M., and W. L. M. Geilenkirchen. 1974. The de- velopment of Denlalium with special reference to the significance of the polar lobe. I. II and III. Division chronology and development of the cell pattern in Dentalium 3)'3C> /o i _ ] 45(138)*- 3uu% JTTT 1 ' 62 (081)* 63% " ' i D47( 17 VS 10) " - 065(14 VS 8) 53(061)* 194% 041 (38VS 27) 073 16 VS 13) 68 (073)*233/ 093 ( 16 VS 13) L-l 2O r- CM fi 9 CM 3 f 8 =) d 100 _*_ j CO c f ~ _i ie K-) CO O ID -| d 51 g 2 f T i/i > S OJ QD 31 CO, _J r 1 ~ ^ d 16 l j5 CD i* 6 i P in o O Q c g J-T 3 49~1 o 5?^ o -i si 5 Co 9 CD ^ d 5] Figure 2. a. A hierarchy in the resorption between tour eolonies of generation II (refer to Fig. 1). The arrowheads point to the inferior partner. Numbers printed along the arrows refer to, respectively: days for complete resorption, zooid ratio (in parentheses, calculated as: the number of zooids in the inferior/superior partners on the day of fusion), percent increase or decrease of zooids in the superior partner from the day of fusion until a complete resorption of the inferior partner. The letter D refers to a case where a disconnection between the partners in a specific chimera occurs. In that case, the numbers along the arrow indicate: days from fusion to disconnection, number of zooids of the left colony on the day effusion vs. the number of zooids of the right colony (in parentheses). Disconnection between the partners within a Botryllus chimera is one of the variations in the outcome to chimera formation, resulting from unsuccessful fusion (Rinkevich and Weissman, 1989), reciprocal resorption (Rinkevich and Weissman, 1987a, 1989), or from a retreat growth phenomenon (Rinkevich and Weissman, 1988). These physiological-genetic-morphological parameters may lead to early separation between the partners before a complete resorption of the inferior partners in a chimera is obtained (Rinkevich and Weissman, 1988, 1989). The hierarchial tendency in the resorption phenomenon is, in most of the cases, already observed before separation between the candidates cancels this reaction. However, we did not count disconnection even when figuring hierarchy. In each such case, at least one additional chimera, where full resorption was accomplished, is assayed. It should be noted, however, that the incompleled results of disconnections are always in agreement with the results where resorption is completed. Subclone sizes may alter the direction of chimera resorption. However, this occurs only when ALLORECOGNITION RESPONSES IN COLONIAL INVERTEBRATE 85 the subordinate partner is much larger than the winner. All subclones used in the present study were matched to pairs with zooid ratios, below that may reverse the direction of resorption. b. Hierarchy in the resorption within the self-crossed offspring of generation III (the two left schemes) and within the detined-cross offspring of generation IV (refer to Fig. 1). The letter M refers to a case where the chimera dies. In that case, the numbers along the arrow indicate: days from fusion until the death of the chimera, number of zooids of the left and the right partners, respectively, on the day of fusion (in parentheses). A dashed arrow with an arrowhead points to a case where the direction of resorption is evident; however, the chimera either died or the partners disconnected before the resorption was completed. Additional subclones for doing new chimeras were absent; therefore, the hierarchy in resorption was not fully determined, c. Hierarchy in the resorption between generation II colonies and the self-crossed offspring of generation III. A dashed arrow without an arrowhead indicates a case where hierarchy is not evident before interactions of the partners in a specific chimera were interrupted by chimera mortality or disconnection. In those cases, no more chimeras were done because of the lack of additional subclones. d. Hierarchy in the resorption between generation II colonies and the defined-cross offspring of generation IV. Dashed arrow with arrowhead indicates two cases where hierarchy became evident before the interactions in the CAAs were interrupted by disconnection. 86 B. RINKEVICH ET AL in maintaining the Botryllus colonies, and M. Greenberg for valuable remarks on an earlier version. This study was supported by NCI Grant CA 42551, partly by a grant from the United States-Israel Binational Science Foun- dation, by a grant from the Basic Research Foundation administered by the Israel Academy of Sciences and Hu- manities, and by a Career Development Award from the Israel Cancer Research Fund, USA. Literature Cited Bailey. D. VV. 1978. Sources of subline divergence and their relative importance forsublines of six major inhred strains of mice. Pp. 197- 215 in Origins of Inbred Mice. H. C. Morse, ed. Academic Press. New York. Bevan, M. J. 1975. Interaction antigens detected by cytotoxic T cells with the major histocompatibility complex as modifier. Nature 256: 419-421. Bjorkman, P., M. A. Sapir, B Samraoui. W. S. Bennett, J. L. Strominger, and D. C. Wiley. 1987. The foreign antigen binding site and T cell recognition regions of class I histocompatibility cell recognition an- tigens. Nature 329: 512-518. Black, F. L., and F. M. Salzano. 1981. Evidence for heterosis in the HLA system. Am. J. Hum. Genet. 33: 894-899. Boyd, H. C, S. K. Brown, J. A. Harp, and I. L. Weissman. 1986. Growth and sexual maturation of laboratory-cultured Mon- terey Bolryllm schlosseri Biol Bull- 170: 91-109. Buss, L. W. 1982. Somatic cell parasitism and the evolution of somatic tissue compatibility. Proc Nail. Acad Sci. i'SA 79: 5337-5341. Eichwald, E. J., and I. L. Weissman. 1966. Weak histocompatibility loci. Annals N.Y. Acad. Sci. 129: 94-101. Eichwald, E. J., C. R. Silmser, and I. L. Weissman. 1958. Sex linked rejection of normal and neoplastic tissue. 1. Distribution and speci- ficity. J Nail Cancer 1ml 20: 563-575. Flaherty, L. 1988. Major histocompatibility complex polymorphism: A non immune theory for selection. Human Immiin. 21: 3-13. Galton, M. 1967. Factors involved in the rejection of skin transplanted across a weak histocompatibility barrier: gene dosage, sex of recipient, and nature of expression of histocompatibility genes. Transplantation 5: 154-168. Graff, R. J. 1978. Minor histocompatibility genes and their antigens. Transplant Proc 10: 701-705. Graff, R. J., W. K. Silvers, R. E. Billingham, W. H. Hildemann, and G. D. Snell. 1966. The cumulative effect ol histocompatibility an- tigens. Transplantation 4: 605-617. Grosberg, R. K. 1988. The evolution of allorecognition specificity in clonal invertebrates. Q Rev Biol. 63: 377-412. Grosberg, R. K., and J. F. Quinn. 1986. The genetic control and con- sequences of kin recognition by the larvae of a colonial marine in- vertebrate. Nature 322: 456-459. Harp, J. A., C. B. Tsuchida, I. L. Weissman, and V. L. Scofield. 1988. Autoreactive blood cells and programmed cell death in growth and development of protochordates. J. E.\p. Zooi 247: 257-262. Hedrick, P. W., and G. Thomson. 1983. Evidence for balancing selec- tion at HLA. Genetics 104: 449-456. Hughes, A. L., and M. Nei. 1988. Pattern of nucleotide substitution at major histocompatibility complex class I loci reveals overdominant selection. Nature 335: 167-170. Johnson, L. L. 1981. At how many histocompatibility loci do congenic mouse strains differ? Probability estimates and some implications. / llered 72: 27-31. Klein, J . 1 986. Natural History of the Major Histocompatihility (. '< im- plex. John Wiley & Sons, New York. Lappe, M. A., R. G. Graff, and G. D. Snell. 1969. The importance of target size in the destruction of skin grafts with non H-2 incompat- ibility. Transplantation 1: 372-377. Loveland, B., and E. Simpson. 1986. The non-MHC transplantation antigens: neither weak or minor. Immunol. Today 1: 223-224. Rinkevich, B., and I. L. Weissman. I987a. A long-term study on fused subclones in the ascidian Botryllus schlosseri: The resorption phe- nomenon (Protochordata: Tunicata). J. Zoo/. (Land.) 213: 717-733. Rinkevich, B., and I. L. Weissman. 1987b. The fate of Botryllus (As- cidiacea) larvae cosettled with parental colonies: beneficial or dele- terious consequences? Biol Bull. 173: 474-488. Rinkevich, B., and I. L. Weissman. 1987c. Chimeras in colonial in- vertebrates: A synergistic symbiosis or somatic and germ-cell para- sitism? Symbiosis 4: 117-1 34. Rinkevich, B., and I. L. Weissman. 1988. Retreat growth in the ascidian Botryllus schlosseri: A consequence of nonself recognition. Pp. 93- 109 in Invertebrate Historecognition, R. K. Grosberg, D. Hedgecock. and K. Nelson, eds. Plenum. New York. Rinkevich, B., and I. L. Weissman. 1989. Variation in the outcomes following chimera formation in the colonial tunicate Botryllus schlosseri. Bull Mar. Sci. 45: 213-222. Rinkevich, B., and I. L. Weissman. 1990. Bolryllus schlosseri (Tunicata) whole colony irradiation: do senescent zooid resorption and im- munological resorption involve similar recognition events? J Exp 7,ool 253: 189-201. Rinkevich, B., and I. L. Weissman. 1992. Allogeneic resorption in co- lonial protochordates: consequences of nonself recognition. Dc\. Camp. Immun 16: 275-286. Robertson, M. 1982. The evolutionary past of the major histocom- patibility complex and the future of cellular immunology. Nature 297: 629-632. Sabbadin, A., and G. Zaniolo. 1979. Sexual differentiation and germ cell transfer in the colonial ascidian Botryllus schlosseri. J. Exp. Zoo/. 207: 289-304. Scofield, V. L., J. M. Schlumpberger, L. A. West, and I. L. \\eissman. 1982. Protochordate allorecognition is controlled by an MHC-like gene system. Nature 295: 499-502. Taneda. Y., Y. Saito, and H. \\atanabe. 1985. Self or nonself discrim- ination in ascidians. Zoo/. Sci. 2: 433-442. Townsend, A. R. M., J. Bastin, K. Gould, and G. G. Brownlee. 1986. Cytotoxic T lymphocytes recognize influenza haemagglutinin that lacks a signal sequence. Nature 324: 575-577. Weissman, I. L. 1988. Was the MHC made for the immune system, or did immunity take advantage of an ancient polymorphic gene family encoding cell surface interaction molecules? A speculative es- say. Int. Rev Immiin. 3: 393-416. Weissman, I. L., Y. Saito, and B. Rinkevich. 1990. Allorecognition histocompatibility in a protochordate species: is the relationship to MHC semantic or structural? Immunol. Rev 113: 227-241. Zaleski, M. B., S. Dubiski, E. G. Niles, and R. K. Cunningham. 1983. Invnunogenelics. Pitman. Boston. Reference: Biol Bull 184: 87-96. (February, 1993) Classification and Characterization of Hemocytes in Styela clava TOMOO SAWADA 1 , JEFFREY ZHANG : , AND EDWIN L. COOPER Department of Anatomy and Cell Biology. School of Medicine, University of California, Los Angeles. California 90024 Abstract. Viable hemocytes of the common tunicate Styela clava are classified into four groups designated as eosinophilic granulocytes, basophilic granulocytes, hyaline cells and lymphocyte-like cells. Eosinophilic granulocytes, actively amoeboid, have large refractive granules that stain with neutral red. Basophilic granulocytes do not stain with neutral red and formed couplets or triplets. Hyaline cells, which often contain phagosomes. have electron-dense small vesicles recognizable only by electron microscopy. Hemoblasts have a characteristic large nucleolus which is visible by light microscopy. Eosinophilic granulocytes and hyaline cells actively ingest yeast particles in vitro. This classification simplifies former ones by correlating electron microscopy, with light microscopy, and viable with fixed hemocytes. Clearly viable tunicate hemocytes can be identified by simple methods. We have provided clear and more accurate descriptions which will lessen the con- troversy often associated with assigning hemocyte func- tions in immunodefense responses both in vivo and in vitro. Introduction The classification of tunicate hemocytes remains con- fused, not withstanding Wright's attempt ( 1 98 1 ) to devise useful categories. Recent progress in tunicate biology, however, requires a precise correlation between various cellular functions and particular types of hemocytes. Styela clava. especially, has been used in investigations of immunological responses including those associated with hemocytes: allogeneic reactions (Raftos and Cooper. 1991); cytotoxic reactions (Kelly et ai. 1992a); humoral Received 20 May 1992; accepted 9 November 1992. Present address: ' Department of Anatomy, Yamaguchi University School of Medicine, Ube-city, 755 Japan. 2 Undergraduate under the SRP program at UCLA. opsonin (Kelly et ai. 1992b, 1993a, b, in press); and the production of cytokines (Beck et ai. 1989; Raftos et ai. 1991). Humoral lectins (Yokozawa el ai. 1986; Harada- Azumi et ai. 1987). antibacterial substances (Azumi et ai. 1990), and a metallo-protease (Azumi el ai. 1991) were studied in another species, Halocynthia roretzi. Although the classification of Styela clava hemocytes began early (Ohue, 1936) and the site of hemopoiesis is described (Ermak. 1975. 1976), the literature includes de- scriptive morphologies with a plethora of terms, but rel- atively little experimental information uniting structure with function. Previous analyses of hemocytes failed to correlate age, season, and cell behavior in a systematic way, and these variables were not related to the various techniques used for examining them (e.g.. staining and fixation versus observation of live cells). Recent molecular and cytological studies focusing on the hemocytes and immune system of Stye/a will reveal a more precise picture of the functional contribution of individual effector cells. But this development depends on a thorough and con- sistent classification of the hemocytes. To establish an acceptable and predictable classification scheme, we examined hemocytes from Stye/a clava and correlated the morphological and behavioral character- istics of living hemocytes, and compared appearance of viable cells with those analyzed by light and electron mi- croscopy. Our work offers a strategy for classifying hem- ocytes in any invertebrate, especially tunicates which are becoming increasingly more important as we decipher the nature of effector cell activity during immune re- sponses. Materials and Methods Hcmocvtes Hemocytes were harvested by severing the stolons of Styela clava after rinsing the outside with 70% ethanol. 87 T. SAWADA ET AL Exuding hemolymph was collected into 0.5 M NaCl (NaCl-solution, pH 7.0 by 0.01 N NaOH) in polystyrene tubes; this prevented the nonspecific coagulation of he- mocytes and allowed individual hemocytes to be observed. Hemolymph was mixed with the NaCl solution one to one in final volume. Staining Hemolymph or hemocyte suspensions in NaCl-solution were loaded onto glass slides. After 10 min, adhering hemocytes were fixed for 15 min and stained with he- matoxylin and eosin (H&E). Cold ethanol, cold methanol or 4% paraformaldehyde (0. 1 M sodium cacodylate buffer, pH 7.0) were used as fixatives, and the morphological preservation was compared. For vital staining, neutral red (NR, 0.01% in final concentration) was added to hemocyte suspensions: 1 5-30 min later, the hemocytes were loaded onto glass slides and observed. Correlation of NR-staining with H&E-staining We photographed NR-stained hemocytes adhering on glass slides, then fixed them for regular light microscopy without moving the slides, and photographed them again under phase-contrast microscopy. After H&E-staining, we found exactly the same cells as in the former two pho- tographs (NR-staining and phase-contrast) to compare their appearance. Transmission electron microscopy (TEM) Hemocytes in the hemolymph and inside pharyngeal tissue were examined by TEM. Hemolymph collected into polystyrene tubes was centrifuged (400 X g for 5 min) and the pellet fixed. Pieces of pharynx (about 1.5 mm square) were dissected and fixed. Specimens were pre- fixed in a mixture of 2% glutaraldehyde and 2% parafor- maldehyde (0.75 M sucrose, 0.2 M sodium cacodylate buffer, pH 7.0), then post-fixed with 1% osmium tetroxide in the same buffer. The specimens were dehydrated in ethanol series and embedded in Medcast (Ted Pella, Red- ding, CA). Propylene oxide was used to infiltrate the resin. Autonomous fluorescence of viable hemocytes Hemocytes suspended in NaCl-solution were loaded on glass slides and observed with a Nikon EFD2 fluores- cence microscope with blue (420-490 nm)-and ultraviolet (330-380 nm)-illumination. Composition of hemocytes Different hemocyte types were counted by light mi- croscopy after H&E or NR-staining. and also by TEM. A sample of hemocytes was taken from 6 animals, and five to ten different viewing fields (1 10-130 cells in total) from each sample were examined in light microscopy with a 100X objective lens. Five pharyngeal pieces, one each from 5 animals (one TEM-section for each piece), and a hemocyte-pellet from one animal were examined by TEM. About one hundred cells were examined on each section. Phagocytic activity against yeast particles Saccharomyces cerevisiae (baker's yeast, type II; Sigma Chemicals, St. Louis, MO) was stained with Congo red and suspended in artificial seawater (approximately 1 X 10 8 particles/ml), according to Kelly et ai (1993a). Hemocyte suspensions in NaCl-solution (100 p\) were loaded on cover slips, and yeast particle suspension (100 ^1) was added 5 min later. The hemocytes were incubated for 30 min. After the cover slips were gently rinsed to remove excess yeast particles, 0.01% neutral-red solution was added. Hemocyte types were identified by NR-stain- ing. Types of hemocytes which phagocytized yeast par- ticles were identified. Results Light microscopy of hemocytes Most hemocytes adhered to glass slides, and some of them exhibited amoeboid movement within 5 min. How- ever, many small, transparent cells did not adhere well enough to resist water movement caused by pressure on the cover slip. By phase contrast microscopy, four different types were observed (Table !):(!) hyaline cells, which ex- hibited significant extensions ( 1 5-20 j/m in diameter); (2) round cells (basophilic granulocytes, 6-10 ^m in diam- eter), which contained highly refractive small granules and often formed couplets or triplets; (3) amoeboid cells (eo- sinophilic granulocytes; 8-15 /jm in diameter), which contained large granules and exhibited more active amoeboid movement than the other types; and (4) small spherical cells that did not spread (hemoblasts; 4-6 ^m in diameter), which contained a small amount of cyto- plasm and had a nucleolus clearly visible by light mi- croscopy. The nuclei of hemocytes other than hemoblasts were not visible unless they were spread flat on a glass slide. Phase contrast microscopy was not sufficient to distin- guish all eosinophilic and basophilic granulocytes with certainty. The eosinophilic granulocytes often contained granules as small as those of basophilic granulocytes, and when they were not moving they were just as round as basophilic granulocytes. We observed large cells that had a hyaline cytoplasm lacking visible granules, but they did contain pigmented or non-pigmented large vacuoles. These cells were also identifiable as hyaline cells because they spread wide and flat. The spreading of hyaline cells was rapid once it began, and these cells did not exhibit active amoeboid movement after they spread. HEMOCYTES OF TUNICATE Table I Classification and sonic characteristics o/'Styela clava licmocvies 89 Size 8 ! 5 ^m 610 f*m 15-20 JOT 4-6 j/m H&E-staining Orange Purple Very weak purple or Purple NR-staining Orange or red-violet pink Negative or orange at vacuoles Granules in LM Many retractile Many small G refractile Granules in TEM Not uniform heterogeneous Uniform spherical Small vesicles _ electron dense Adhesion to glass + + ii Phagocytosis + + + + + + Other characteristics Active amoeboid movement Forming couplets or Widely spread (cell Nucleolus visible in LM * Blue fluorecence in red-violet cells aggregates fusion?) Previous classifications Compartment signet-ring vesiculated Finely granular Hyaline signet-ring? Hyaline lymphocyte-like morula coarsely granular amoebocyte hemoblasts Under UV-illumination (330-380 nm). NR-staining and some characteristics of viable cells NR mainly stained the cytoplasmic granules of amoe- boid cells (Fig. 1 ). Hyaline cells were usually not stained except for some cases in which small or large cytoplasmic vacuoles were stained (Fig. IE, F). Two groups of amoe- boid cells were positively stained by NR (eosinophilic granulocytes), but the intensity differed. One was stained a dense red- violet, whereas the other was orange (Fig. 1 ). Both types of cells contained 5-20 large cytoplasmic granules, and they were morula-shaped before starting amoeboid movement. Other granular cells were usually round, unstained by NR, and contained highly refractive small granules (Fig. 1; basophilic granulocytes). Most of the hemoblasts were not stained (Fig. 1 ). but a few some- times stained faintly orange. Following staining with NR. hemocytes were easily distinguished and their characteristic behaviors examined. Both types of NR-positive granulocytes (eosinophilic granulocytes) were active in amoeboid movement, ex- tending many spine-like pseudopodia. NR-negative gran- ular cells (basophilic granulocytes) were less active in amoeboid movement, but extended long pseudopodia. These cells were often observed as couplets (Fig. ID), triplets or small aggregates composed only of this type of cell, and they did not separate once they came in contact. Occasionally, these cells spread flat after 30-60 min in- cubation. H&E-staining We used three different methods to observe hemocytes with H&E-staining after fixation. Ethanol and methanol significantly modified hemocyte morphology, so only paraformaldehyde fixation was utilized. We observed the same cells with NR-staining and H&E-staining. Two NR-positive amoeboid cells (red-violet and orange cells) were stained intensely red or pink. Both cells con- tained various sizes of cytoplasmic granules that stained with eosin, so both cells were classified as eosinophilic granulocytes. The appearance of the cytoplasmic granules was altered by fixation, especially by ethanol and meth- anol. Cells that were fixed with these agents appeared as vacuolated cells, granular cells, compartment cells or sig- net-ring cells. NR-negative amoeboid cells that contain small refrac- tive granules were stained purple with H&E and were des- ignated as basophilic granulocytes. Their cytoplasmic granules were no longer evident after fixation, and cyto- plasmic staining was relatively weak after they spread on slides. However, the nuclei of basophilic granulocytes were smaller and more dense than those of hyaline cells. The cytoplasm of hyaline cells was very thin after spreading so H&E stained them only weakly purple with some orange-stained cytoplasmic vesicles. Phagocytic vacuoles in some of them were stained red. Some hyaline cells contained large cytoplasmic vacuoles and thus ap- peared as signet-rings. Also their nuclei became larger as they spread. A few encapsulations of several small cells were ob- served (small encapsulation; Fig. 2). In addition, small numbers of multinuclear cells were observed in H&E- staining (Fig. 2). These cells spread flat and contained several large nuclei and small vesicles stained with eosin. Finally, hemoblasts stained purple, and their nucleoli be- came unclear after fixation. TEM oj hemocytes In centrifuged pellets of hemolymph, we observed five different hemocyte types (Figs. 3, 4): (1) large cells 90 T. SAWADA ET AL bg hy Figure 1. Living hemocytes on glass slides after NR-staining. (A) Eosinophilic granulocytes included two groups of granulocytes that stained in different colors (o = orange and r = red-violet). The sizes of the cytoplasmic granules are variable in each hemocyte. (B) Neither hemoblasts (hb) and basophilic granulocytes (bg) were stained. Nucleoli were evident in hemoblasts. (C) Basophilic granulocytes (bg) contained many refractive granules which were smaller than those of eosinophilic granulocytes (o = orange cells). (D) A couplet of basophilic granulocytes (bg): these were frequently observed. (E) Hyaline cells (hy) spread wide and flat on the glass slide to form a thin cytoplasmic sheet. (F)Some hyaline cells contained granules (arrow) that stained with NR. x 1250 (10-12 ^m in diameter: hyaline cells) containing signifi- cant amounts of endoplasmic reticulum (ER); (2) small spherical cells (5-6 /urn in diameter: hemoblasts) with little cytoplasm and large nuclei; (3-5) three different granu- locytes (6-12 nm in diameter: basophilic and eosinophilic granulocytes) containing abundant cytoplasmic granules. These five types also constituted the entire hemocyte pop- ulation within the pharyngeal tissue. The large cells contained numerous rough-surfaced and smooth-surfaced ER and also small vesicles (0.1 ^m in HEMOCYTES OF TUNICATE 91 A , Figure 2. Fixed hemocytes on glass slides with H&E-staining. (A) Multinuclear cells with seven nuclei spread wide and flat. Vesicular structures in the cytoplasm were slightly stained. (B) Small encapsulation (arrow) containing 3-7 small cells; these were sometimes observed in the hemolymph. xl 180 diameter) of high electron density (Fig. 3). A few of these cells contained large vacuoles or phagosomes. Their nuclei often had nucleoli and coarse and uniform euchromatin, although heterochromatin was sometimes observed. These cells corresponded to hyaline cells on the basis of size, phagosomes, and the absence of large cytoplasmic gran- ules. The small cells with little cytoplasm contained mito- chondria and small amounts of ER (Fig. 4C). Their nuclei, with characteristic large nucleoli, were usually larger than those of other hemocytes. Chromatin was uniformly dis- tributed and slightly more dense in comparison with hya- line cells. These cells corresponded to hemoblasts in cell size, i.e., little cytoplasm and characteristically large nu- cleoli. The three different granulocytes (temporarily designated as type 1, 2 and 3 granulocytes according to TEM) had the same nuclear pattern (usually with dense heterochro- matin at the periphery and sometimes small nucleoli) but differed in their cytoplasmic granules. Type 1 granulocytes (6-10 yum in diameter) contained electron-dense and spherical granules with a diameter range of 0.2-0.5 ^m (Fig. 4A). These cells correspond to basophilic granulo- cytes on the basis of size, the sizes of their cytoplasmic granules (they had smallest granules among granulocytes), and their frequency in the hemolymph. Type 2 gmnit/o- cytes (8-10 /urn in diameter) contained irregular-shaped granules that varied in size (0.1 -1.3 /urn in diameter). The granules contained homogeneous material of intermediate electron-density (Fig. 4B). Type 3 granulocytes ( 8- 1 2 ^m in diameter) had cytoplasmic granules that were also ir- regularly-shaped and remarkably varied in size (0.1-1.5 nm in diameter). These granules were composed of het- erogeneous materials central spheres with high electron density and surrounding material of intermediate electron density (Fig. 4B). Type 2 and 3 granulocytes corresponded to eosinophilic granulocytes on the basis of size, the ir- regular shape of their cytoplasmic granules, and their fre- quency of occurrence. Hemocyte composition We examined percentages of the various hemocytes in hemolymph by counting each type after NR- and H&E- staining and TEM (Table II). The order of dominance for each type was the same in all cases, but the exact values were somewhat different. The most abundant cells were eosinophilic granulocytes (46.3% in NR-staining); second were the basophilic granulocytes (21.0%); hyaline cells were third ( 18.5%-): and the smallest population was that of the hemoblasts (14.1%). The percentage of eosinophilic cells in H&E-staining (68.5%) was about the same as the sum of type 2 and 3 granulocytes in the hemocyte pellets observed by TEM (67.8%), but it was larger than the sum of orange- and red-violet cells in NR-staining (46.3%). Many fewer he- moblasts were found in both H&E-staining (2.3%.) and TEM (2.0%) than in NR-staining (14.1%). The proportion of hyaline cells ranged from 5.5 to 18.5%, even after the percentages of multinuclear cells and phagocytosis were added. Multinuclear cells (1.2% in H&E-staining) were not found in NR-staining or TEM of pharyngeal tissue. A utonomous fluorescence oj hemocytes Blue fluorescence was observed in certain granulocytes under ultraviolet-illumination. NR-staining of those flu- orescent hemocytes, in the same field of view, revealed that the autonomous fluorescence was from eosinophilic granulocytes which stained in red-violet (Table I). Under blue-illumination, no hemocytes exhibited autonomous fluorescence. Phagocytosis Four hemocyte types i.e., hyaline cells, eosinophilic granulocytes (including red-violet and orange cells in 92 T. SAWADA ET AL m t!+^& .. 'JQratfSH * ' : * *. '?& . Figure 3. Transmission electron microscopy of hyaline cells and a hinuclear cell. (A) Hyaline cell (h) in the centrifuged pellet, with heterochromatin at the nuclear periphery. (B) A binuclear cell in the centrifuged pellet. (C) Hyaline cell (h) in pharyngeal tissue; the nucleus has a large nucleolus and uniform euchromatin. All cells (A, B, C) contained electron-dense small vesicles, numerous vesicular structures, and endoplasmic reticulum. Bar = 1 jim. NR-staining). and basophilic cells ingested yeast parti- cles. Among them, hyaline cells and eosinophilic granu- locytes had significantly higher activity than basophilic granulocytes (Table III). In the cell population that had ingested yeast particles, hyaline cells (36-42%) were fewer than eosinophilic granulocytes (5 1-68%), as shown in Ta- ble IA. However, phagocytic activity was higher in hyaline cells, because the phagocytic ratios were higher in hyaline cells (32-78%) than in eosinophilic granulocytes (13- 35%), as shown in Table IB. Many hyaline cells engulfed 2-5 yeast particles, whereas most eosinophilic granulo- cytes incorporated only one particle. Discussion Classification of hemocytes Hemocytes from many species of tunicates have been classified by both light and electron microscopy (Ohue, 1936; George, 1939;Endean, 1960; Andrew, 1961, 1962; Overton, 1966; Smith, 1970; Botte and Scippa, 1977; Figure 4. Transmission electron microscopy of hemocytes in the centnt'uge pellet of hemolymph. (A) Type 1 granular cells ( 1 = basophilic granulocytes) containing relatively uniform and spherical granules. (B) Both type 2 (2) and 3 (3) granular cells (eosinophilic granulocytes) containing irregularly shaped granules. The granules of type 3 cells contain electron dense cores. (C) Hemoblast (hb) with little cytoplasm and without cytoplasmic granules, except for mitochondria and vesicles. The relatively large nucleus contains characteristic large nucleolus. Bar = 1 ^m. 93 94 lli'inncvtc composition examined under different conditions T. SAWADA ET AL Table II Hemocyte types Viable cells (NR-staining) Fixed cells (H&E-staining) EM (pellet) EM (pharynx) Eosinophilic Orange 68.5 5.6% 38.9% 15.8 7.5% granulocytes 16.5 5.0% Red-violet 28.9 32.8 11.0 29.8 8.5 Basophilic 21.0 5.1 21.1 4.4 2 1 .0 21.3 7.0 granulocytes Hyaline cells 18.5 7.9 5.5 2.9 8.0 16.8 3.8 Hemoblasts 14.1 3.2 2.3 2.2 2.0 8.5 6.0 Multinuclear cells 0.0 1.2 1.2 0.2 0.0 Cells of 0.1 0.2 1.4 1.2 1.1 2.3 1.6 phagocytosis No. of individuals 6 6 1 5 examined average S.D. Milanesi and Bunghel, 1978; Fuke. 1979, 1980: Rowley, 1981, 1982; Mukai et ai, 1990), and in morphological terms, such as vacuolated or granular cells, hyaline cells, hemoblasts or lymphocytes (Wright, 198 1 ). or by functions (Freeman, 1964; Fuke, 1980; Fujimoto and Watanabe, 1976; Burighel et ai, 1976; Rowley, 1983; Azumi et al, 1 990, 1 99 1 ; Raftos et al. . 1 990; Raftos and Cooper, 1 99 1 ). However, the inapplicability of these classifications from one species to the next, and the lack of correspondence between different methods (e.g., light versus electron mi- croscopy) have produced confusion. Hemocytes of S. clava have been classified into morula cells, compartment cells, signet-ring cells, granular amoe- bocytes, hyaline cells and lymphocyte-like cells (Ohue, 1936; Wright, 1981). But, among fresh and living hemo- cytes, we observed no signet-ring cells nor any cells with a stable, morula shape. Instead, there were granulocytes that frequently changed their appearance during amoe- boid movement. They appeared morula-like when they rounded up, and could be compartment cells or granular amoebocytes after they had become extended and flat- tened. Fixation, especially with ethanol or methanol, modified hemocyte morphology significantly, and some of the eosinophilic granulocytes and hyaline cells became signet-ring in shape. Therefore, we adopted two cautious guidelines. First, we avoided using such terms as morula, compartment, or signet-ring. Second, we employed no Wright- or Giemsa staining because they require methanol as the fixative. Instead, we preferred to use formaldehyde fixation and H&E-staining. We identified five different hemocyte types by vital NR- staining and TEM, and four types by H&E-staining of fixed cells. We estimated that the granules of the orange cells in NR-staining contain less dense material, and so correspond to type 2 granulocytes in TEM; similarly red- violet cells in NR-staining correspond to type 3 granu- locytes in TEM. The difference between type 2 and 3 cells, or between orange and red-violet cells, is not significant enough to separate them into two cell types. Moreover, both the orange and red-violet cells evidently correspond to eosinophilic granulocytes in H&E-staining. These two granulocytes appear to be similar in amoeboid movement and phagocytic activity. Therefore, we classified both of them into the same group as eosinophilic granulocytes. We suggest that type 2 granulocytes (orange cells) are an earlier stage in cell differentiation than type 3 granulocytes (red-violet cells). The correspondences between the light microscopical and TEM images of basophilic granulocytes (type 1 gran- ulocytes in TEM), hyaline cells, and hemoblasts were clear on the basis of their morphological characteristics and their frequencies of appearance. Multinuclear cells were classified as hyaline cells for the following reasons: (1) morphologically multinuclear cells are in all other respects similar to hyaline cells; (2) they sometimes contain large, eosinophilic vacuoles that we assume to be phagosomes; (3) the morphology and behavior of hyaline cells are quite similar to phagocytes type 1 (pl-ceQs)ofHalocynthiaroretzi(H. rorelii), which evidently fuse together and form multinuclear cell sheets (Sawada el al., 1991). But, we have no strong evidence for cell fusion between the hyaline cells of S. clava. More- over, the frequency of multinuclear cells in fresh hemo- lymph is not clear, because they could be identified only after spreading on glass. Both of these points require fur- ther investigation. In this study, therefore, we have identified four hemo- cyte types in S clava. ( 1 ) Eosinophilic granulocytes con- tain several refractive vacuoles that appeared red in neutral red vital stain, red by H&E. and exhibit active amoeboid HEMOCVTES OF TUNICATE 95 Table III Phagocytosis / yeast particles by hemocytes from three Jil/erent individuals (A) Composition of hemocytes which ingested yeast particles Eosinophilic granulocytes Total Basophilic Hyaline cells Animals (red-violet)* 1 (orange)* 1 granulocytes cells Hemoblasts examined a* 2 29.2% 22.1% 6.2% 42.5% 0.0% 113 b 41.0 11.0 10.0 38.0 0.0 100 c 41.6 16.8 5.0 36.6 0.0 nil (B) Phagocytosis against yeast particles within each kcinocyte type Hemocyte types Animals Ingesting cells Non-ingesting cells Total cells examined eosinophilic granulocytes a* 2 19.4% 80.6% 108 (red-violet) b 30.8 69.2 52 c 12.9 87.1 101 (orange) a 16.4 83.6 110 b ND* 3 ND ND c 1.6 98.4 61 basophilic granulocytes a 7.3 92.7 124 b 3.8 96.2 53 c 0.0 100 57 hyaline cells a 78.0 22.0 100 b 63.6 36.4 22 c 32.8 67.2 61 *' Two sub-populations of eosinophilic granulocytes different in colors of NR-staimng are indicated in parenthesis. * : Animals (a. b. c) in Table A correspond to the animals in Table B. * 3 No data. movement and phagocytosis. (2) Basophilic granulocytes contain numerous small granules that do not stain with neutral red, are purple in H&E, and form specific aggre- gations with the same cell type. (3) Hyaline cells contain fine electron-dense granules in TEM, occasionally contain phagosomes that stain red with neutral red and H&E, and extend into thin circular sheets on glass. (4) Hemoblasts possessed little cytoplasm, large nucleoli visible by light microscopy, but adhere only weakly to glass. Possible cor- respondence between former classifications are shown in Table I. Functions and characteristic behavior of each hetnocyte type Phagocytosis, as is well known, is a ubiquitous and im- portant immuno-defense response found throughout the animal kingdom. Hyaline cells exhibited the highest phagocytic activity, and some of them engulfed more than five yeast particles. Eosinophilic granulocytes were less active than hyaline cells, but they accounted for the largest population because of their abundance and active motility. Hyaline cells were the most likely candidates for ef- fecting encapsulation of larger particles by their ability to spread and form flat sheets and to fuse together into larger multinuclear sheets. Hemoblasts have been referred to as lymphocyte-like cells (Wright, 198 1 ) and as proliferative stem cells (Ermak, 1976). We also observed the charac- teristically large nucleolus also in viable cells and con- firmed their equivalents by light (Wright, 1 98 1 ) and elec- tron microscopy (Ermak, 1976). Motility was also an important and definitive, behav- ioral characteristic. Only eosinophilic granulocytes ex- hibited active movement. In contrast, the basophilic granulocytes did not separate after once contacting others, which resulted in the formation of couplets or triplets. This behavior continues when augmented, resulting in small aggregates. Similar behavior was also observed on gl -cells of//, roretii (Sawada et ai, 1991), and we suggest the presence of common granulocytes that can form spe- cific aggregates within the same cell type. Correspondence to the hemocytes in other tunicate species Hemocyte types found in many species have been cat- egorized into several groups by Wright (1981). However, the hemocytes of a single category often include several different types. In addition, certain hemocytes of one spe- cies are apparently absent in other species. It would not be instructive to compare only morphological aspects of hemo- cytes, and only under a single condition, such as in paraffin 96 T. SAW A DA ET AL sections. Observations of living hemocytes, under different conditions and stained with simple dye, coupled with func- tional analysis, e.g.. of phagocytosis, would be more useful. In such a manner, we compared the hemocytes of St vela clava and Halocvntlria rorctzi which have also been clas- sified in the living state (Sawada et ai, 1991 ), and found interesting correspondences between types. Hyaline cells and basophilic granulocytes were similar to the pi -cells and gl -cells of Halocynthia roretzi. respectively, in mor- phological and behavioral aspects. Hemoblasts, as the candidate for hematopoietic stem cells, may correspond to the ly-cells of Halocynthia roretzi, but their function as the stem cells has not been established in either species. Eosinophilic granulocytes seemed to be similar to the v3- and v4-cells of Halocynthia roretzi in that refractive vacuoles occupy most of the cell volume, and active amoeboid movement and addphilic staining occur. But eosinophilic granulocytes ofStyela clava were evidently more phagocytic. The correspondence between these species of at least two to three cell types may be consistent with their phylogeny. Acknowledgments We thank Sharon Sampogna and Monica Eiserling for technical help in light and electron microscopy. This study was supported by the National Science Foundation (Grant #DCB 90 05061). Literature Cited Andrew. VV. 1961. Phase microscope studies of living blood-cells of the tunicates under normal and experimental conditions, with a description of a new type of motile cell appendage. Q. .1. Micr. Sci. 102: 89-105. Andrew, VV. 1962. Cells of the blood and coelomic fluids of tunicates and echinoderms. Am. Zoo/. 2: 285-297. Azumi, K., H. Yokozawa, and S. Ishii. 1990. Halocyammes: novel anti- microbial tetrapeptide-like substances isolated from the hemocytes of the solitary ascidian. Halocynthia roretzi Biochemistry 29: 159-165. Azumi, K., H. Yokozawa, and S. Ishii. 1991. Lipopolysacchande in- duces release of a metallo-protease from hemocytes of the ascidian, Halocynthia roretzi. Dev. Comp. Immune/ 15: 1-7. Beck, G., G. R. Vasta, J. J. Marchalonis, and G. S. Habicht. 1989. Characterization of mterleukin-1 activity in tunicates. Comp. Bioclwm. Physiol. 92B: 93-98. Botte, I,., and S. Scippa. 1977. Ultrastructural study of vanadocytes in .-I iciilia malaca. Experienlia 33: 80-8 1 . Burighel, P., R. Brunelti, and G. Zaniolo. 1976. Hibernation of the colonial ascidian Botrylloides leachi (Savigny): histological obser- vations. Boll. Zoo/. 43: 293-301. F.ndean, R. 1960. The blood-cells of the ascidian. Phalhisia mammilala. Q J Micr. Sii. 101: 177-197. Ermak, T. H. 1975. An autoradiographic demonstration of blood cell renewal in Styela clava (Urochordata: Ascidiacea). Experientia 31: 837-839. Ermak, T. II. 1976. The hematogenic tissue of tunicates. Pp. 45-56 in Phylogeny of Thymus and Bone Marrow-bursa Cells, R. K. Wright and E. L. Cooper, eds. Elsevier/North-Holland, Amsterdam. Freeman, G. 1964. The role of blood cells in the process of asexual reproduction in the tunicate Perophora viridis. J Exp. Zool. 156: 157-183. Fujimoto, H., and II. VVatanabe. 1976. The characterization of granular amoebocytes and their possible roles in the asexual reproduction of the polystyelid ascidian, Poly:oa vesiciiliphora. J. Morphol 150: 623-638. Fuke, M. T. 1979. Studies on the coelorruc cells of some Japanese ascidians. Bull. Mar. Biol. Sin. Asamushi, Tohoku University 16: 143-159. Fuke, M. T. 1980. "Contact reactions" between xenogeneic or allo- geneic coelomic cells of solitary ascidians. Biol. Bull. 158: 304-315. George, VV. C. 1939. A comparative study of the blood of the tunicates. Quart. J Micr Sci. 81: 391-431. Harada-Azumi, K., II. Yokozawa, and S. Ishii. 1987. N-acetyl-galac- tosamine-specific lectin, a novel lectin in the hemolymph of the ascidian Halocynthia roretzi: Isolation, characterization and com- parison with galactose-specific lectin. Com. Bioclwm. Physiol 88B: 375-381. Kelly, K. L., E. L. Cooper, and D. A. Raftos. 1992a. In vitro allogeneic cytotoxicity in the solitary urochordates. J. Exp. Zool. 262: 202-208. Kelly, K. L., E. L. Cooper, and D. A. Raftos. 1993a. A humoral opsonin from the solitary urochordate Styela clava. Dev Comp. Imnumol. 17: (in press). Kelly, K. L., E. L. Cooper, and D. A. Raftos. 1992b. Purification and characterization of a humoral opsonin from the Styela clava. Comp. Biochem. Physiol 103B: 749-753. Kelly, K. L, E. L. Cooper, and D. A. Raftos. 1993b. Cytokine-like activities of a humoral opsonin from Styela clava. Zool. Sci. (in press). Milanesi, C., and P. Burighel. 1978. Blood cell ultrastructure of the ascidian Bolryllux schlosseri. I. Hemoblast, granulocytes macrophage. morula cell and nephrocyte. Acta Zool 59: 135-147. Mukai, H., K. Hashimoto, and H. VV atanabe. 1990. Tunic cords, glom- erulocytes, and eosinophilic bodies in a styelid ascidian. Polyandro- carpa misakiensis. J. Morphol. 206: 197-210. Ohue, T. 1936. On the coelomic corpuscles in the body fluid of some invertebrates. III. The histology of the blood of some Japanese as- cidians. Sci. Rep Tohoku L'niv. 11: 191-206. Overton, J. 1966. The fine structure of blood cells in the ascidian Per- ophora viriilix. J. Morph. 119: 305-326. Raftos, D. A., D. L. Stillman, and E. L. Cooper. 1990. //; vitro culture of tissue from the tunicate Styela clava. In Vitro Cell Dev Biol 26: 962-970. Raftos, D. A., and E. L. Cooper. 1991. Proliferation of lymphocyte- like cells from the solitary tunicate. Styela clava. in response to al- logeneic stimuli. J. Exp. Zool. 260: 391-400. Raftos, D. A, E. L. Cooper, G. S. Habicht, and G. Beck. 1991 . Invertebrate cytokines Tunicate cell proliferation stimulated by an interleukin-1- like molecule. Proc. Nat. Acad. Sci. USA 88: 9518-9522. Rowley, A. F. 1981. The blood cells of the sea squirt, dona intestinalis: morphology, differential counts and in vitro phagocytic activity. J. Invertebr. Palhol. 37:91-200. Rowley, A. F. 1982. Ultrastructural and cytochemical studies on the blood cells of the sea squirt, dona intestinalis. I. Stem cells and amoebocytes. Cell Tissue Res. 223: 403-414. Rowley, A. F. 1983. Preliminary investigations on the possible anti- microbial properties of tunicate blood cell vanadium. J. Exp. Zool. 27: 319-323. Sawada, T., Y. Fujikura, S. Tomonaga, and T. Fukumoto. 1991. Classification and characterization often types of hemocytes in tunicate Halocynthia roretzi. Zool. Sci. 8: 939-950. Smith, M. J. 1970. The blood cells and tunic of the ascidian Halo- cynthia aurantium (Pallas). I. Hematology. tunic morphology and partition of cells between blood and tunic. Biol. Bull 138: 354-378. Wright, R. K. 1981. Urochordates. Pp. 565-626 in Invertebrate Blood Cells 2. N. A. Ratcliffe and A. F. Rowley, eds. Academic Press, London. Yokozawa, H., K. Harada, K. Igarashi, Y. Abe, K. Takahashi, and S. Ishii. 1986. Galactose-specific lectin in the hemolymph of solitary ascidian. Halocynthia roretzi. Molecular, binding and functional properties. Biochim. Biophys. Acta, 870: 242-247. Reference: Bin! Bull 184: 97-104. (February, 1993) Effects of Cations on the Volume and Elemental Composition of Nematocysts Isolated from Acontia of the Sea Anemone Calliactis polypus MICHIO HIDAKA AND KIWAMU AFUSO Department of Biology, University of the Ryiikyus, Nishihara Okinawa, 903-01 Japan Abstract. The hypothesis that exchange of intracapsular divalent cations with Na + in seawater increases the inter- nal osmotic pressure during discharge of nematocysts of marine cnidarians was tested by examining effects of ex- ternally applied cations on the volume and elemental composition of nematocysts isolated from acontia of the sea anemone Calliactis polypus. The volume of isolated nematocysts increased with increasing concentrations of cations if the cation was monovalent but appeared to de- crease if the cation was divalent. Ca 2+ reduced the internal osmotic pressure of the nematocysts more efficiently than Mg 2+ . X-ray microanalysis of nematocysts incubated in 1 M solutions of various salts showed that Ca 2 + in isolated nematocysts was only partially replaced, if at all, by ex- ternally applied Na + and Mg 2+ while most Mg :+ was re- placed by Na + and Ca 2+ . The present results suggest that exchange of intracapsular divalent cations with external monovalent cations increases the internal osmotic pres- sure, and that selective binding of Ca 2+ to polyanions in the capsule decreases it. Whether the increase in the in- ternal osmotic pressure caused by the cation exchange is large enough to trigger discharge remains to be investi- gated. Introduction Lubbock and his colleagues proposed that loss of Ca 2+ from a nematocyst increases the osmotic pressure of the intracapsular fluid and thus causes discharge of the ne- matocyst (Lubbock and Amos, 1981; Lubbock el ai, 1981; Gupta and Hall, 1984). They proposed that poly- peptides in undischarged nematocysts are crosslinked by Ca 2+ to form polypeptide chains and that the release of Received 14 May 1992; accepted 20 October 1992. calcium from the nematocyst dissociates the polypeptide chains, thereby increasing the number of osmotically ac- tive molecules. Because of this report, Ca 2i has been con- sidered to play a major role in nematocyst discharge. Recently Weber (1989) demonstrated that naturally occurring cations of Hydra nematocysts can be replaced by externally applied cations. Nematocysts loaded with other cations generally retain discharge capabilities. Gerke d t/l. ( 199 1 ) found that in situ nematocysts of Hydra con- tained high concentrations of potassium (K) instead of calcium (Ca). These observations suggest that Ca 2+ is not indispensable for the discharge of certain kinds of ne- matocysts. Weber (1989) proposed that Hydra nematocysts can be considered as Donnan-equilibrium dominated osmotic systems and that cations associated with polyanions in the capsule, rather than polyanions themselves, contribute to high intracapsular osmotic pressure. Because Hydra nematocysts contain high concentrations of K (Gerke et a/., 1991 ) and are surrounded by a membrane that might serve as a diffusion barrier against ions of low molecular weight (Lubbock et a/., 1981), nematocysts of Hydra might be in equilibrium with high concentrations of K + . If such nematocysts are exposed to freshwater as a result of exocytosis, the osmotic pressure difference across the capsule wall would increase, leading to the discharge of the nematocysts. Indeed, isolated Hydra nematocysts im- mersed in concentrated NaCl or KC1 solutions swell up to 1 1 5% of the original volume and tend to discharge when the external concentration of the salts is lowered (Weber, 1989). The above process, however, may not account for the discharge of nematocysts of marine cnidarians, because nematocysts of marine cnidarians must discharge in sea- water, which contains high concentrations of salts. X-ray 97 98 M. HIDAKA AND K.. AFUSO microanalysis of frozen sections of various marine cni- darians show that the predominant cation of nematocysts /// situ is either Ca :+ , Mg : + , or K* (Tardent el a/.. 1990). If nematocysts of marine cnidarians also behave as Don- nan-equilibrium dominated systems, exchange of intra- capsular cations with cations in seawater will occur when nematocysts are exposed to seawater as a result of exo- cytosis. In Ca- or Mg-containing nematocysts, the ex- change of divalent cations in the capsule with monovalent cations such as Na + in seawater might increase the internal osmotic pressure, since one divalent cation is replaced by two monovalent cations to maintain electroneutrality. If the increase in the internal osmotic pressure is large enough, the nematocysts would discharge. The purpose of the present study is to examine the hypothesis of divalent-monovalent cation exchange. Un- discharged nematocysts isolated from various cnidarians contain high concentrations of Ca and Mg (Weber et a!., 1987; Mariscal, 1988; Hidaka, 1993). These isolated ne- matocysts provide a useful model for studying the re- sponses of Ca- and/or Mg-containing nematocysts to var- ious cations. We determined the effects of mono- and divalent cations on the volume of nematocysts isolated from acontia of the sea anemone Calliactis polypus. We also studied whether Ca 2+ and Mg 2+ found in the isolated nematocysts could be replaced by externally applied cat- ions as in Hydra nematocysts (Weber, 1989). Materials and Methods Specimens of Calliactis polypus on the shells of hermit crabs belonging to the genus Dardanus. were collected from the reef around the Okinawa island, and maintained in an aquarium supplied with a subgravel filter. The her- mit crabs were fed with chopped Tapes every 2-4 days. The anemones were used as a source of acontial nema- tocysts 1-3 days after feeding. Acontial filaments were obtained by prodding the sea anemone with blunt-tipped forceps. Undischarged basitrichous isorhiza nematocysts were isolated either in artificial seawater (ASW) or in distilled water (DW), since nematocysts isolated in ASW and those isolated in DW display different discharge capabilities (Hidaka and Mariscal, 1988). A piece of acontium was placed in a drop of ASW or DW on a glass slide. The glass slide was treated with a drop of a 0.1% solution of poly- 1-lysine (Sigma; approx. mol. wt. 90,000) in distilled water for 10 min in a wet chamber prior to use (Mazia et ai, 1975). Two strips of thin adhesive tape (Scotch 3M) were placed on both sides of the drop to make a narrow space between the glass slide and a cover slip, and to make it easy to replace the solution in this space. When nema- tocysts were isolated in ASW, the acontium was squashed under a cover slip. Only a small percentage (about 5%) of nematocysts discharged during this procedure, and most of them only partially discharged eversion of the tubule stopped halfway. Cellular debris and partially discharged nematocysts were removed by washing the squashed acontium with a few drops of ASW. Most of the nema- tocysts isolated in this manner from acontia ofCalliaclis tricolor discharged when immersed in 5 mAf EGTA (Hi- daka and Mariscal, 1988), suggesting that the isolated ne- matocysts are functional. When nematocysts were isolated in DW, the acontium was immersed in a drop of DW for 5 min and then the remaining acontium was removed. The extruded undischarged nematocysts were allowed to settle onto the glass slide for 10 min. Then, the isolated nematocysts were washed with more than five drops of ASW or DW to remove unattached nematocysts. Photomicrographs of nematocysts were taken in ASW or DW using a plan objective lens (X 100). Then, test so- lutions were applied by perfusing the nematocysts with at least eight drops of each test solution (Hidaka and Mar- iscal, 1988). Nematocysts isolated in ASW were treated with decreasing concentrations of salt solutions, that is, 1000, 100, 10, 1, and mAf solutions. Nematocysts iso- lated in DW were treated with increasing concentrations of salt solutions. Nematocysts were immersed in each test solution for 10 min, because changes in the volume of isolated nematocysts in solutions with or without Ca 2+ were completed within 10 min (Hidaka, 1992). After 10 min a pair of photomicrographs of the nematocysts was taken for each test solution. The length and diameter of nematocysts capsules were measured to the nearest 0.1 mm (corresponding to 0.05 /*m) on two photomicrograph prints using calipers. The average value was used for each capsule. The volume was calculated assuming that the capsule was an ellipsoid. The volume of nematocysts im- mersed in test solutions was normalized by the original volume of the nematocysts in ASW or DW, and expressed as relative volume. The original volume (mean SD) was 1 17.7 18.8 Mm 1 (n = 37) in ASW-isolated nematocysts and 100.7 10.1 M' (n = 37) in DW-isolated nema- tocysts. For each salt solution, three or four experiments were performed and at least seven nematocysts were mea- sured. The significance of regression of the relative volume of nematocysts on log (salt concentration) was tested in each salt solution. When the regression was not significant, the relative volume at 1 M salt concentration was com- pared with that at 1 mAf using Duncan's multiple range test. The difference in the relative volume of nematocysts was tested among pairs of cations at 1 M salt concentration using the multiple range test. For substitution experiments, nematocysts were isolated by immersing 20 acontia in 5 ml of DW for 5-10 min. Remaining acontial tissues were removed by filtering the nematocyst suspension through 60 ^m nylon mesh. Ali- quots (0.8 ml) of the filtrate were placed in each of six EFFECTS OF CATIONS ON NEMATOCYSTS 99 microtubes and centrifuged at 1940 X g for 5 min. One ml of each test solution was added to the pellet. Nema- tocysts were resuspended in the test solutions and allowed to stand for 10 min. Test solutions were ASW and 1 M solutions of NaCl, KG, CaCl : , MgCl 2 , and SrCl 2 . Next, the nematocysts were washed in DW by centrifuging at 1940 X g for 5 min and by resuspending the nematocysts in 1 ml of DW. The nematocysts were washed in DW and collected by centrifugation three times. Finally, the nematocysts were resuspended in 1 50 n\ of DW. Aliquots (20 n\) of the nematocyst suspension were placed on meshes with formval or collodion membranes that had been coated with carbon and treated with poly-1-lysine. The nematocysts were allowed to settle on the membrane for 1 h, then air-dried after the remaining solution was soaked up with a piece of filter paper. Specimens were observed under a scanning transmission electron micro- scope (JEOL JEM-2000EX) equipped with an energy dis- persive spectrometer (TN 42 U). X-ray spectra were ac- quired at an acceleration voltage of 100 kV. Semiquan- titative elemental analyses were performed using an application software (Noran Instruments Inc. SMTP) on 4-6 nematocysts for each test solution. The software, which was designed for standardless semiquantitative analysis of metallurgical thin films, removes background and integrates peak areas. The peak intensities were con- verted to ratios of element concentrations by multiplying by calculated K-factors. Correction for absorption was not made. The ASW contained (in mAI): NaCl, 480; KC1, 10; CaCl 2 , 10; MgCl 2 , 26; MgSO 4 , 29; and was adjusted to pH 8.0 with 10 mM HEPES. All the salt solutions and DW were buffered to pH 8.0 with 10 mM HEPES, and all the experiments were done at room temperature (23-26C). Results Nematocysts isolated from acontia of Calliaclis polypus in ASW swelled in concentrated solutions of monovalent cations (Fig. 1 ). There was a significant positive regression between the volume of nematocysts and the concentration of Na + and K + (regression analysis, P < 0.05). Though there was no significant regression between the volume of nematocysts and the concentration of divalent cations, the volume of nematocysts was significantly smaller in 1 M MgCl 2 and SrCl 2 than in 1 mM solutions (Duncan's multiple range test, P < 0.01). At 1 M concentration, nematocysts immersed in divalent cations were signifi- cantly smaller than those immersed in monovalent cations (Duncan's multiple range test, P < 0.01). When nema- tocysts that had been immersed in various salt solutions were immersed in a buffer solution without added salts, there was no significant difference in the mean volume 120 110 g 100 JS 90 0> oc 80 10 100 1000 Concentration of salts (mM) Figure 1. Effects of cations of various concentrations on the volume of nematocysts isolated from acontia of Calliaclis polypus in ASW. Ne- matocysts isolated in ASW were immersed successively in salt solutions of decreasing concentrations. The salt solutions examined were NaCl (). KCI (), CaCN (O), SrCl : (A), and MgCl 2 (O). The volume of ne- matocysts in each solution is expressed as a percentage of the original volume of nematocysts in ASW. Vertical bars represent standard devia- tions; some SD bars are omitted for clarity. (one-way ANOVA, P > 0.25). Thus the volume of the nematocysts increased with increasing concentration of cations if the cation was monovalent, but decreased if the cation was Mg :+ or Sr + . The volume of the nematocysts was smaller in 1 A/CaCl 2 than in 1 A/MgQ 2 (P < 0.01). The volumetric behavior of nematocysts isolated in DW was almost the same as that of nematocysts isolated in ASW (Fig. 2). When the concentration of external K + was increased, the volume of isolated nematocysts increased (regression analysis, P < 0.05). Though regression between the volume of nematocysts and concentration of Na + was not significant, nematocysts immersed in 1 M NaCl were larger than those immersed in 1 mAf NaCl (Duncan's multiple range test, P < 0.05). Nematocysts immersed in 1 M Cad: and SrCl 2 were smaller than those immersed in 1 mM solutions (P < 0.01). Nematocysts immersed in 1 M CaCl 2 or SrCl 2 were smaller than those immersed in 1 A/MgCl 2 (/ > <0.01). A scanning electron micrograph of a nematocyst sample prepared for X-ray microanalysis is shown in Figure 3. X-ray spectra of nematocysts were different depending on the incubation solutions (Fig. 4). The major elements of nematocysts incubated in ASW were Ca and Mg in ad- dition to sulfur (S), though a small Na-peak was present (Fig. 4A). When nematocysts were incubated in 1 M NaCl, the Na-peak increased, the Ca-peak remained high but the Mg-peak disappeared (Fig. 4B). When nematocysts were immersed in 1 M KCI, a small K-peak appeared, but the peaks of the other elements were not affected (Fig. 100 M. HIDAKA AND K. AFUSO 120 r - 110 0) 100 o > 0) 2 90 4) DC 80 10 100 1000 Concentration of salts (mM) Figure 2. Effects of cations of various concentrations on the volume of nematocysts isolated from acontia of Calliactis polypus in DW. Ne- matocysts isolated in DW were immersed successively in salt solutions of increasing concentrations. The symbols are the same as in Figure 1. The volume of nematocysts in each solution is expressed as a percentage of the original volume of the nematocysts in DW. The verlical bars represent standard deviations; some SD bars are omitted for clarity. 4C). Nematocysts incubated in 1 A/ CaCl 2 showed large Ca- and small Na-peaks in addition to the S-peak (Fig. 4D). When nematocysts were incubated in 1 M MgCl 2 , the Mg-peak increased (Fig. 4E). Nematocysts incubated in 1 A/ SrG 2 showed large Sr- and small Ca-peaks in ad- dition to the S-peak (Fig. 4F). When discharged nema- tocysts were analyzed, peaks of metals were absent re- gardless of the incubation solutions. Table I shows the relative abundance of metal cations in undischarged nematocysts that were isolated in DW and then incubated in various salt solutions. Ca accounted for about 50% of the metals in nematocysts immersed in ASW or 1 A/ MgCl 2 and more than 50% in nematocysts immersed in 1 A/ NaCl or KC1. Ca was replaced substan- tially only by strontium (Sr). Most of the Mg disappeared when nematocysts were immersed in 1 A/ NaCl, CaCl 2 , and SrCl : . Only a small amount of K was present in ne- matocysts incubated in 1 A/ KC1. Discussion Weber (1989) studied the volumetric behavior of iso- lated stenoteles of Hydra under different ionic conditions. He showed that nematocysts immersed in 1 A/ solutions of various salts swell when the concentration of salts is lowered, regardless of whether the cations are monovalent or divalent. The volumetric behavior of isolated nema- tocysts of the sea anemone Calliactis polypus was different from that of Hydra nematocysts. Calliactis nematocysts appeared to shrink in concentrated solutions of divalent cations as in Hydra nematocysts, but swelled in concen- trated solutions of monovalent cations. Thus the volu- metric behaviors of the marine anemone nematocysts and the freshwater Hydra nematocysts are different in solu- tions of monovalent cations. Weber (1989) showed that the volumetric behavior of Hydra nematocysts immersed in salt solutions of various concentrations can be accounted for by a Donnan-equi- librium model. The Donnan potential generates an asym- metrical distribution of ions across the capsule wall. Ac- cording to Weber's simulation studies, the difference in total ion concentration between the inside and outside of the capsule increases as the external salt concentration is lowered from 3 M to 0. 1-0.01 A/. When the external salt concentration is further lowered, the osmolarity difference drops due to protonation of polyanions, unless the exter- nal pH is high. The volume of Calliactis nematocysts, however, decreased as the external concentration of monovalent cations was lowered from 1 to A/. Thus the volumetric response of the sea anemone nematocysts to monovalent cations cannot be accounted for by the simple Donnan-equilibrium model. Weber ( 1989) showed that naturally occurring cations in Hydra nematocysts can be replaced by externally ap- plied cations. If this is true for nematocysts of marine cnidarians, cations contained in the isolated capsule might be replaced by external cations when nematocysts are im- mersed in various salt solutions. Isolated Calliactis ne- matocysts contained predominantly Ca 2+ and Mg :+ (Hi- daka, 1993). IfCa 2+ and Mg 2+ in the isolated nematocysts are replaced by monovalent cations, the internal osmotic pressure would increase as one divalent cation is replaced by the two monovalent cations required to maintain elec- troneutrality. The swelling of the sea anemone nemato- Figure 3. Scanning electron micrograph of isolated nematocysts used for X-ray microanalysis. The white spot represents the site irradiated with an electron beam during the acquisition of spectra. These nema- tocysts were isolated in DW and incubated in 1 A/ MgCl : for 10 mm. I I I I ( IS in (A [ IONS ON NEMATOCYSTS 101 flSM- 1 o.d.d C- 1 od-glutamic acid)s are the major constituents of nematocysts in Hydra (Hydrozoa, Cnidana). ./. Binl. Chcm 265: 9664-9669. Weber, J. 1991. A novel kind of polyanions as principal components of cnidarian nematocysts. Cnnip. Biochem. Physiol. 98A: 285-291. Weber, J.. M. Klug, and P. lardent. 1987. Detection of high concen- trations of Mg and Ca in the nematocysts of various cnidanans. E.\- pcricntia43: 1022-1025. Vanagita, T. M. 1959. Physiological mechanisms of nematocyst re- sponses in sea-anemone II. Effects of electrolyte ions upon the isolated cnidae. / Fac. Sci. Univ. Tokyo Sect. II' 8: 381-400. Reference: Biol. Bull 184: 105-113. (February. 1993) Hemocyanin Subunit Composition and Oxygen Binding in Two Species of the Lobster Genus Homarus and Their Hybrids CHARLOTTE P. MANGUM Bodega Marine Laboratory. University of California. P. O. Box 247. Bodega Bay, California 94923 [ Abstract. The monomeric subunit composition and O : binding properties of the hemocyanins (Hcs) of Homarus americamts. H. gammarus and their hybrids are very similar, though not identical. //. americanus He has six major electrophoretically separable polypeptide chains; H. gammarus He has four major and two minor chains; and the hybrid He has four major and one minor chain. Four chains co-migrate in all three groups, and the fifth chain in the hybrid co-migrates with a fifth chain in H. gammarus. Thus, qualitatively, the hybrid He is more like that of//, gammarus than H. americanus. a similarity reflected in respiratory properties. Although the O : affinity of the hybrid hemocyanin appears to lie intermediate be- tween that of the two parent hemocyanins at 25C, in fact it is significantly different from that of//, americanus but not H. gammarus. The cooperativity of the hybrid He also differs significantly from that of //. americanus but not H. gammarus He. The distinctive properties of H. americanus hemocyanin at 25C are believed to be due to either or both of two chains: a unique and also invariant chain in H. americanus. and one that is present in H. gammarus and the hybrids but not in H. ameri- canus. H. americanus He also appears to be slightly less sensitive to the allosteric modulator L-lactate. No differ- ence in CaCl 2 sensitivity was found. At lower temperatures respiratory properties are indistinguishable. In adult H. americanus that had been held under identical conditions for long periods, variation in subunit pattern was not en- tirely absent, but it was smaller than that found in natural populations of other species. No differences in O 2 binding at 25C were found in morphs differing qualitatively in Received 3 June 1992; accepted 8 October 1992. Permanent address: 'Department of Biology, College of William & Mary, Williamsburg, VA 23185-8795. one chain and quantitatively in two others. No effect of a combination of rearing temperature and diet was found on the He subunit composition of juveniles. Introduction The arthropod hemocyanins (Hcs) are multiples of hexamers built of 70-80 kDa polypeptide chains. Often the 2 X 6-mers predominate in the bloods of adult decapod crustaceans, including the lobster Homarus americanus (Olson et ul.. 1988). The number of different monomers is usually large, with a dozen or more found in several species of Uca (Sullivan et ai. 1983: Callicott and Man- gum, 1992; Mangum, 1992 and unpub. data). The mono- mers have been grouped into one of four categories on the basis of their electrophoretic mobilities, immunolog- ical reactions and roles in oligomer assembly (Markl, 1986). Within a species, the monomeric heterogeneity also plays a functional role in respiratory adaptation during the adult stage (Mason et ai, 1983; Mangum and Rainer, 1988; deFur et ai. 1990; Mangum et ai. 1991). By com- paring morphs, the functional differences have been at- tributed to particular electrophoretic bands (Mangum and Rainer, 1988; Mangum et ai, 1991; Mangum, 1992). The role of He subunit composition in bringing about functional differences between species is less clear. A sur- vey of forty-two species of various degrees of taxonomic relatedness suggests a high degree of specificity (Reese and Mangum, 1992). More intensive investigation of the Hcs of seven species of the genus Uca. which are extremely polymorphic as well as heterogeneous, supports the in- ference of species specificity (Callicott and Mangum, 1992; Mangum, 1992; C. P. Mangum, unpub. data). In every case, including sibling species such as U. panacea and pugilator, even low frequency He morphs of a species can be readily distinguished from those of another. Functional 105 106 C. P. MANGUM properties can also differ in sibling congeners with different latitudinal ranges. In comparisons of congeners that are less closely related, however, functional properties are more clearly related to environmental factors than to phylogenetic allinity or subunit composition (Reese and Mangum, 1992). In the only two species in which the effect of laboratory acclimation has been examined, the subunit phenotype of an adult individual is not fixed (Mason ct a/.. 1983; deFur el a!.. 1990; Callicott and Mangum, 1992). In Cal- lincctes sapidus. moreover, the variation both in the lab- oratory and in nature can be related to environmental factors such as salinity and hypoxia (Mangum, 1990; Pihl ct ai. 199 1;C. P. Mangum and S. P. Baden, unpub. obs.). Thus the members of the highly polymorphic samples of natural populations may have been acclimated to different environmental (or nutritional) conditions. Here I report data for the monomeric subunit com- position and oxygen binding of the Hcs of the adults of two species of lobsters in the genus Homarus and the hybrid progeny of their spontaneous matings. The two parent species had been brought from their native Atlantic habitats to the Bodega Marine Laboratory, where they were held under identical conditions for periods far longer than the species investigated previously. They are known to be highly homozygous at 41 loci and, at most loci, the allozymic phenotypes of the two are either extremely sim- ilar or identical. It is believed that the two speciated al- lopatrically when isolated for the first time during the Pleistocene (Hedgecockrt al.. 1977). In one species, I also examined the He subunit composition of juveniles which had been reared on either of two diets, and at different temperatures. Materials and Methods The sample All available adults, a total of 36. were examined; they were large (28-42 cm from rostrum to tail), intermolt individuals. Two (one of each sex) belong to Homarus gammarus (Linnaeus), formerly known as H. vulgaris: they were collected near lona, Scotland in 1975. They are the sole survivors of the larger sample characterized by Hedgecock et al. in 1977. Twenty-five adults ( 16 females, 9 males) are members of Homarus americanus H. Milne Edwards. All but one were caught on various dates in 1988-92 in waters surrounding Martha's Vineyard, Mas- sachusetts, and had been held in the mariculture facility at the Bodega Marine Laboratory for periods ranging from three months ( 1 individual) to more than three years (3 individuals). One individual of H. americanus (age > 6 years) was born in the Bodega Marine Laboratory. Nine hybrid adults (7 fertile females, 2 infertile males) were progeny of spontaneous matings between //. gammarus and americanus. Most were produced in 1983 by a female H. gammarus and a male H. americanus: one, of un- known parentage, was born in 1978. All adults had been fed the same diet of surf fish and shrimp, and had been held under identical photoperiods in the running seawater system of the Bodega Marine Laboratory. A seven year compilation of data ( 1 985-9 1 ) indicates that the water temperature ranges from 10 to 15C, and usually varies within about two degrees; over the four month period of sampling the salinity varied from 32.5-33.5%o, which is typical. The He subunit composition of 14 juvenile //. amer- icanus (8-10 cm length, both sexes), which had hatched in the Bodega Marine Laboratory 20-22 months earlier, was also examined. Half of these animals, which were their natural color, had been fed since stage IV a diet of brine shrimp, fish and crabs, and had been held at the seawater system temperature. The other half, phenotypic albinos, had been fed a diet based on casein; for the past year they had been held at room temperature (ca., 23C). The diets, rearing conditions and molt history of animals such as these were described in detail by Baum ( 1990). Preparation oj material and electrophoresis Blood was taken from the base of the last leg and serum expressed from the clot in a tissue grinder. After centrif- ugation an aliquot of the material was dissociated to its monomers by dilution with 0.01 mol I" 1 EDTA + 0.05 mol r 1 Tris (pH 8.9), to reduce light scattering; He con- centration was estimated from the absorbance of disso- ciated material at 338 nm (Bausch & Lomb Spectronic 2000 spectrophotometer), using the extinction coefficient reported by Nickerson and van Holde (1971). An addi- tional aliquot was diluted (1:10 or 1:30, depending on concentration) with the dissociating buffer for electro- phoresis, and the remainder frozen for future use. Ab- sorbance of the material from several individuals, detailed below, was compared at 280 and 338 nm. PAGE electrophoresis of native monomers was carried out at constant current according to Hames and Rickwood (1985). Following determination of the He phenotype in each individual, the variants among H. americanus were examined several times in adjacent lanes on the same gels. Representatives of each of the three groups were also compared many times on the same gels. Finally, gels were overloaded with six times the usual amount of material, the presence of Cu was determined according to Bruyn- inckx et al. (1978), and then the gels were stained as usual with Coomassie Blue. Oxygen binding On the basis of the PAGE, particular individuals were selected for a second bleeding, performed within a week HEMOCVANIN-O, BINDING AND SUBUNIT COMPOSITION IN LOBSTERS 107 of the final phenotype determination. Serum was dialyzed overnight, against seawater for most of the measurements or against a Tris maleate buffered salt for the experiments on inorganic ion sensitivity. Oxygen binding was deter- mined within a few days, using the cell respiration method (Mangum and Lykkeboe, 1979). Data analysis Bohr plots of the values for P 50 (oxygen affinity) were described by regression lines and their 95% confidence intervals compared. Mean values for n 50 (cooperativity) and He concentration were compared by Student's /-test. The data for CK binding as a function of [Cad:] were analyzed similarly. However, the nonlinearity of the re- sponse of Psn to [NaCl] and [Na^SOj] precluded statistical analysis. Results Hemocyanin concentration Adults of Homarus americanus had significantly (P = .02) higher levels of He [6. 1 1 (0.41 S.E.) g 100ml '] than the hybrids [4.18 (0.70 S.E.) g 100 ml" 1 ]. The values for the two members of//, gammarus (2.75 and 4.84 g 100 mr 1 ) also fall below the 95% confidence interval around the mean of the H. americanus sample. In the // americanus data there is no clear trend with length of time in the laboratory, suggesting that the nutritional state of the animals was good. The juveniles of this species had considerably lower He concentrations [1.02 (0.15) g 100 mr 1 ]. which were unrelated to diet. Monomeric subunit composition The two adult Homarus gammarus had identical He phenotypes, which were also the same as that of one of the two individuals examined two years earlier (C. P. Mangum, unpub. obs.). Four high density (or major) and two intermediate density (or minor) electrophoretic bands separated by charge (Fig. 1 ). All six were positive for Cu. The 25 adults of//, americanus exhibited very similar but not identical He phenotypes (Fig. 1 ). As many as eight bands separated on the lower third of the gels, four of which had co-migrants in H. gammarus (Fig. 1 ). The two most anodic bands ( 1 and 2) were always present in trace quantities, if at all. Material at their position appeared to quench the fluorescence of bathocuproine sulfonate, in- dicating the presence of Cu. However, it was not possible to ascertain the site of the quenching more precisely, only one of the two may contain Cu. These bands had no co- migrants in H. gammarus or the hybrids. In H. americanus bands 3-8 could reach high concen- trations. The gels on which the best separation was ob- tained exhibited less density in the middle of the material designated as bands 3 and 4, suggesting the presence of two chains that are similar in charge and extremely dif- ficult to resolve. Moreover, the leading edge of this ma- terial clearly co-migrated with bands 2 in the hybrid and 3 in H. gammarus. whereas the trailing edge clearly lagged behind. Thus I assigned two numbers (3 and 4) to this position of the H. americanus material, even though the separation was not great enough to photograph. In ad- dition, I was not able to decide whether the trailing ma- terial was present in all 25 individuals. The quantities of chains 6 and 7 in //. americanus are similar to the cor- responding ones in //. gammarus, but chains 4 and 8 always occurred in higher levels in H. americanus than H. gammarus (Fig. 1 ). In the early PAGE, band 5 in H. americanus did not appear to be sharply delineated at its leading and trailing edges. It was the only high density band that clearly varied qualitatively as well as quantitatively, ranging from absent (3 adults) to low concentration (5) to high concentration (17). Since I suspected that this band might not be a He chain, I examined the ratio of the absorbance at the protein peak (280 nm) and the active site (338 nm). According to this index, however, the total Cu content of H. amer- icanus samples containing maximal levels of band 5 did not clearly differ from those that lacked it; nor did it differ from the samples from //. gammarus and the hybrids. none of which contained co-migratory material. In ad- dition, on subsequent gels band 5 was as sharply delineated as the rest (Fig. 1). Bands 6 and 8 varied quantitatively, though the mag- nitude was not great. They decreased concomitantly to intermediate levels in two of the 25 animals; band 8 was intermediate in two additional individuals in which band 6 remained maximal. Band 7 appeared to be absent in a single individual in which 6 and 8 were maximal. This female, from which larvae had hatched two months ear- lier, had been in the laboratory for only three months. All other individuals had maximal levels of these three chains. Although I did not investigate material held at tem- peratures above freezing, there was no correlation between phenotype and age of frozen preparations; this has been true in my experience with Hcs from all species examined thus far. In the present case the same banding pattern was observed before and after four months. Finally, material prepared on a second occasion from four of the same individuals three months after the first bleeding showed no change in phenotype. The hybrids, all adults, exhibited a single phenotype which did not vary quantitatively or qualitatively. With the exception of the higher levels of band 2, it resembles the phenotype of H. gammarus more than that of H. americanus (Fig. 1 ). The hybrid He has four major chains and one minor chain, all of which correspond in mobility 108 C. P. MANGUM 1 2 H. americanus hybrid H. gammarus 1 i a 3 4 m 5 6 Figure 1. Banding patterns of the dissociated hemocyanins of the two parent species of Homarus and their hybrids, and a diagram illustrating the correspondence (arrows) of band positions. The anode is at the top. In the gel for // americanus, each pair of lanes shows a sample from a different individual in higher (left) and lower (right) concentrations. The lanes on the far right were overloaded to show the anodic material (numbered I and 2) that occurs in trace quantities. The cathodic triplet of bands in this species is more clearly shown in lanes that were not overloaded. The middle panel shows // gammarus ( 1 individual) and the hybrid He ( 1 individual) in alternating lanes. to one of the six chains of H. gammarus. The hybrids differ from both parents, but from H. gammarus only in the absence of the most anodic band (H. gammarus band 1 ) and the higher levels of hybrid band 2. They differ from H. americanus in the absence of its three most anodic bands (H. americanus bands 1-3), in the absence of H. americanus band 5, in the consistently lower quantity of their most cathodic chain (hybrid band 5) and in the pres- ence of a distinctive band 1. As in all other species I have examined (e.g., Mangum et a!., 1985; Mangum, 1992), the phenotypes of males and females in each of the three groups were indistinguishable no sex specific material was present. HEMOCYANIN-O, BINDING AND SUBUNIT COMPOSITION IN LOBSTERS 109 The juvenile H. americanus were indistinguishable from the adults. Six of the seven members of each dietary- thermal group had the maximum number of bands. One in each group lacked chain 5. Chains 6-8 were invariably present in maximal concentrations. Oxygen binding First, the intraspecific variation in //. americanus was examined. One adult lacked band 5 and also had minimal (= intermediate) levels of chains 6 and 8; at 25C, however, the oxygen binding properties of its He (stripped of organic co-factors) were indistinguishable from those of another individual containing maximal amounts of all eight bands. Therefore the data have been combined for presentation. The coefficient of determination (r) for the regression line describing P 50 in Figure 2 is 0.962, further affirming the absence of a perceptible effect of phenotype. The single individual with low quantities of band 7 had been sacri- ficed at the time the O 2 binding measurements were per- formed. Second, the Hcs of the two parent species were com- pared. At all but the lowest pH investigated. He O : affinity at 25 C is significantly lower in H. gammarus than H. americanus, though the difference is fairly small (Fig. 2). Third, the hybrid He was compared with each of the par- ent Hcs. Whereas the data for the hybrid He appear to be intermediate between those for the two parent species, the difference from //. gammarus is not significant even in the middle of the pH range investigated. In contrast, the difference between the hybrid and //. americanus is significant throughout most of the pH range exam- ined (>7. 2). The mean value for cooperativity is somewhat smaller (P = .001) in H. americanus (3.24 .11 S.E.) than //. gammarus and the hybrids (3.95 .13), which do not differ from one another (P = .15). Thus the respiratory properties of the hybrid He are also more like those of//. gammarus than H. americanus. At lower temperatures, the significant differences dis- appear completely. Ninety-five percent confidence inter- vals around regression lines fit to the O : affinity data in Figure 3 overlap fully throughout the pH range investi- gated. Often this is true because the numerical values di- minish and are thus more difficult to distinguish, but in this example there is not even an apparent trend. Mean values for cooperativity do not differ (P = .2-. 8). As a result, //. americanus He is less temperature sensitive than the other two Hcs, though only in the 15-25C range. For that range, the apparent heat of oxygenation (AH) is only 2.4 kcal mor' for H. americanus He (pH 7.6), whereas the value for the hybrid He is 5.6, and the value for H. gammarus He is -6.6. For the range 5-15C the value of AH for all three Hcs is -9.4 kcal mol~' (same pH). //. americanus He is slightly less sensitive to the allo- steric effector L-lactate (Fig. 4) than H. gammarus He; once again, the sensitivity of the hybrid He appears to be intermediate. At pH 7.6 the addition of 10 mmol 1~' lac- tate changes log P 50 of H. americanus He by 0.166, H. gammarus He by 0.232, and the hybrid He by 0.203. However, O ; affinity in H. gammarus and the hybrids is so 40 20 10 8 o . 5 - A A c ' o * 7.2 7.6 8.0 7.2 7.6 8.0 Figure 2. Oxygen binding at 25C of Homarus americanus (closed circles, solid lines), H. gammarus (open circles, dotted lines) and their hybrid (triangles, dashed lines) hemocyanins. The curves are fitted regression lines 95% confidence intervals. 0.05 mol T' Tris maleate buffered seawater. Material obtained from two individuals of H americanus was used (see text), whereas H gammarus and the hybrids are represented by a single individual. 110 C. P. MANGUM S 4 15 C 5 C ;-.** , i , i A 00 o in Q_ 70 SO 30 10 7 5 o, 7.0 7.5 8.0 7.0 7.5 PH 8.0 Figure 3. Oxygen binding al 1 5 and 5C of//, americanus (closed circles), H gammarus (open circles) and their hybrid (triangles) hemocyanins. The regression lines and confidence intervals were omitted for clarity. 0.05 mol 1~' Tns maleate buffered seawater. Origin of material as in Figure 2. not quite significantly different in the presence of lactate, even in the middle of the pH range. In the presence of lactate, O : affinity of //. gammarits He remains signifi- cantly lower than that of H. americanus throughout the pH range investigated. In contrast, the hybrid He has a significantly lower O 2 affinity than that of H. americanus He only at high pH. In all three groups cooperativity is significantly diminished in the presence of lactate. The mean value drops from 3.2 to 2.86 .05 S.E. (P = .05) for H americanus He, from almost 4 to 3.06 .24 (P = .05) for H. gammarus He, and from almost 4 to 2.95 .29 (P = .002) for the hybrid He. The sensitivity of the three Hcs to CaCh is indistin- guishable (Fig. 5). Regression lines and their 95% confi- dence intervals overlap fully throughout the concentration range investigated. Mean values for cooperativity do not differ (P = .50-.75). NaCl clearly raises He O ; affinity and lowers cooper- ativity of//, americanus He (Fig. 5). Once again, however, the different morphs were indistinguishable, and the data were combined for presentation. In contrast to the allo- steric effect of Ca 2+ . the relationship between P 50 and NaCl is nonlinear on logarithmic coordinates. I used high con- centrations of Na2SO 4 , prepared from the decahydrate. too 70 50 30 o n 10 L 7 5 7.2 7.6 8.0 7.2 7.6 8.0 7.2 7.6 B.O Figure 4. Lactate sensitivity of//, americanus (left panel: closed circles, solid regression lines 95% confidence intervals), H. gammarus (right panel: open circles, dotted lines) and their hybrid (middle panel: triangles, dashed lines) hemocyanins. Origin of material as in Figure 2. Control curves reproduced in each panel from Figure 2. 25C, 0.05 mol I" 1 Tris maleate buttered seawater. HEMOCYANIN-O, BINDING AND SUBUNIT COMPOSITION IN LOBSTERS 111 o in 5.0 4.5 4.0 O* * * 3.5 j 3.0 7 S f 8 o * A * i i i o 7 i | O g 5 - o. e 4 3 ~ o 2 i l l I l i 1.8 - 0.0 0.5 1 .0 1.5 log free [CaCI ] 2.0 1.5 1 .0 o o o 8 0.0 0.5 1.0 1.5 2.0 2.5 3.0 log free [NaCI] or [Na 2 SOj Figure 5. Inorganic ion sensitivities of lobster hemocyanins. The units of free ion concentrations (antilogs) are mmol l~'. Left panels: H amcricanus (closed circles, solid lines). //. gammarus (open circles, dashes) and their hybrids (triangles, dots). 0.05 mol 1~' Tns maleate buffer (pH 7.7) + O.I mol T 1 NaCI. Right panels: the response of H americanus He to NaCI (closed circles) and Na : SO 4 (open circles). 0.05 mol 1' Tris maleate + 0.01 mol 1~' CaCl : . 25C. Ongin of matenal as in Figure 2. to examine specificity. The response differed very little from that of NaCI (Fig. 5), and the apparent difference may lie within the error of preparing an accurate solution of a highly hydrated salt (especially at a marine labora- tory). Discussion The essentially non-specific sensitivity of Homarus americanus He to NaCI is further evidence that the in- organic ion responses of the crustacean Hcs are not all alike. The response of this He differs from that of portunid crab Hcs (Truchot. 1975; Mason el al, 1983), which are insensitive to NaCI, but resembles that of penaid shrimp Hcs (Brouwer et al.. 1978; Mangum and Burnett. 1985). From a physiological point of view, however, NaCI sen- sitivity is unlikely to be important in H. americanus, a stenohaline species. According to Hedgecock et ill. ( 1977 and pers. comm.), the genetic distance between the two parent species of Homarus. though significant, is so small that the numer- ical value is closer to expectation for subspecies than spe- cies. Thus it is of particular interest that the present find- ings support the inference of species specificity of He sub- unit composition (Reese and Mangum, 1992). Although H. gammarus is monomorphic for the common H. amer- icanus allele at 30 allozymic loci, neither of the two parent Hcs in the present sample could be confused with the other. As in the sibling species of Uca (Mangum, 1992 and unpub. obs.), this inference is true in spite of intra- specific variation. In H. americanus. band 3 is both di- agnostic of the species and, at least in the present sample, invariant. Material that co-migrates with chains 1 and 2 of H. gammarus is clearly absent from H. americanus. Furthermore, the hybrid He is structurally distinct from either parent. The present findings also support the inference of little interspecific genetic distance. Even though they are not identical, the two parent Hcs are more similar than any of the ca. 50 Hcs we have examined thus far, with the exception of Menippe adina and M. mercenaria Hcs (Reese 1989). Like the lobsters, these two sibling species of stone crabs are also believed to have speciated recently, and they also hybridize spontaneously (Bert, 1986). In both structural and functional properties the hybrid He resembles that of one parent more than the other. It has only one less chain than H. gammarus He but several fewer than H. americanus He. The electrophoretic be- 112 C. P. MANGUM havior of each of the five hybrid chains is identical to that of some one of the // gammanis chains, whereas hybrid chain 1 has no co-migrant in H. americanus. These re- lationships are reflected in the O : binding of the Hcs in a complete saline, though only at high temperature. In stage IV through adult H. americanus, SDS PAGE separates three He chains (Olson el ai. 1988; Olson and McDowell, 1989). As is often the case (e.g., Sullivan et al., 1983), additional bands are revealed when the sepa- ration is made by charge. In neither juveniles nor adults of this species can the He be categorized as strictly monomorphic at the level of quaternary structure, despite prolonged acclimation of the donors. Moreover, in juveniles the variation is distinctive of neither the stage nor the thermal-nutritional history. In both stages, however, the variation is much smaller than in samples of natural populations of several species of brachyuran crabs (Mangum, 1990, 1992; Callicott and Mangum, 1992). This generalization is true of respiratory properties of the adults as well. Although the sample size is much smaller in the present investigation, the inference remains unchanged when the comparison is made with, for example, the 14-20 individuals of Callinectes sapidus investigated by Mangum el al. in 1991. I emphasize that the small amount of He variation found here may not accurately represent natural popu- lations of H. americanus (much less H. gammanis). Nor is it clear that the lack of variation results from prolonged acclimation rather than limited genetic diversity, as found at other loci (Tracey et al., 1975; Hedgecock el al., 1977). However, I note that the inference of allozymic similarity in the species was made from samples of populations on either side of Cape Cod but not Cape Hatteras, the greater geographic barrier (e.g., Friedrich, 1973; National Geo- graphic Society, 1985). The present findings suggest that, given the common acclimation, the differences observed both within H. americanus, and between this species and the other two groups, represent a fixed condition in an adult individual. Although only intermolt animals were investigated here, the finding of no change with molt stage in Callinectes sapidus (Mangum et al.. 1985) has recently been con- firmed in H. americanus (N. B. Terwilliger, pers. comm.). More important in the present context, the virtual identity of most respiratory properties of the three Hcs appears to reflect the notable similarity of the electrophoretic phe- notypes. Conversely, it is reasonable to suggest that the slightly higher O 2 affinity and lower cooperativity of H. americanus He at high temperature are due to the chains that are unique to one of the three groups. Perhaps the most likely candidate is chain 2 in //. gammanis (= 1 in the hybrids), which is absent in H. americanus. However, the possibility that band 3 is invariant as well as unique to //. americanus cannot be excluded. Bands 1 and 2 are never present in H. americanus in more than trace quan- tities, and morphs containing or lacking chain 5 did not differ in O 2 binding. The latter inference would be un- warranted only if the effect of chain 5 was exactly com- pensated by an equal and opposite effect of chains 6 and 8, which were also variables in the comparison. Acknowledgments Supported by NSF DCB 88-16172 (Physiological Pro- cesses). I am extremely grateful to the University of Cal- ifornia for a Research Fellowship and to the Bodega Ma- rine Laboratory for its unfailing hospitality. Literature Cited |{:iiiin. N. A. 1990. Studies on the role of dietary protein and lecithin in molting and cholesterol transport in juvenile lobsters, Homanis sp. M. A. Thesis. Sonoma State University, Rohnert Park. Cal.. 65 pp. Bert, T. M. 1986. Speciation in western Atlantic stone crabs (genus MenippeY the role of geologic processes and climatic events in the formation and distribution of species. Mar. Biol. 93: 157-170. Brouwer, M., C. Bonaventura, and J. Bonaventura. 1978. Analysis of the effect of three different allosteric ligands on oxygen binding by hemocyanin of the shrimp Pcnaeus setiferus. Biochem. 17: 2148- 2154. Bruyninckx, W. J., S. Gutteridge, and H. S. Mason. 1978. Detection of copper on polyacrylamide gels. Analyt. Biochem. 89: 174-177. Callicott, K. A., and C. P. Mangum. 1992. Phenotypic variation and lability of the subunit composition of the hemocyanin of I'ca pugi- lulor. J Exp. Afar. Biol. Ecol. (in press). deFur, P. L., C. P. Mangum. and J. E. Reese. 1990. Respiratory re- sponses of the blue crab Callinectes sapidus to longterm hypoxia. Biol. Bull 178:46-54. Friedrich, H. 1973. Marine Biology. University of Washington Press. United Kingdom. 474 pp. Hames, B. D., and D. Rickwood. 1985. Gel Electrophoresis of Proteins. 1RL Press, Oxford. 290 pp. Hedgecock, D., K. Nelson, J. Simons, and R. Shleser. 1977. Genie similarity of American and European species of the lobster genus Homanis. Biol. Bull. 152:41-50. Mangum, C. P. 1990. Inducible O, carriers in the crustaceans. Pp. 92- 103 in Animal Nutrition and Transport Processes. 2. Transport, Res- piration and Excretion. Comparative and Environmental Aspects. J.-P. Truchot and B. Lalou, eds. Karger, Basel, Switzerland. Mangum, C. P. 1992. Structural and functional polymorphism of the hemocyanin O 2 transport system of the sand fiddler crab. L'ca pug- ilalor. J. Exp. Mar. Biol. Ecoi (in press). Mangum, C. P., and L. E. Burnett. 1986. The CO 2 sensitivity of the hemocyanins and its relationship to Cl sensitivity. Biol. Bull. 171: 248-263. Mangum, C. P., and G. Lykkeboe. 1979. The influence of inorganic ions and pH on the oxygenation properties of the blood in the gas- tropod mollusc Busycon canaliculalum. J Exp. Zoo/. 207: 417-430. Mangum, C. P., and J. S. Rainer. 1988. The relationship between sub- unit composition and oxygen binding of blue crab hemocyanin. Biol. Bull 174: 77-82. Mangum, C. P., J. Greaves, and J. S. Rainer. 1991. Oligomer com- position and oxygen binding of the hemocyanin of the blue crab Callinectes sapidus. Biol Bull 181:453-458. Mangum, C. P., B. A. McMahon, P. L. deFur, and M. I. VVheatly. 1985. Gas exchange, acid-base balance and the oxygen supply to HEMOCYANIN-O: BINDING AND SUBUNIT COMPOSITION IN LOBSTERS 1 13 tissues during a molt of the blue crab. Callincctcx sapidiis Rathbun. J. Crn.it. Bioi 5:207-215. Mason, R. P., C. P. Mangum, and G. Godette. 1983. The inllucnce of inorganic ions and acclimation salinity on hemocyanin-oxygen binding in the blue crab Callmeaes sapidiis. Biol. Bull. 164: 104-123. Markl. J. 1986. Evolution and function of structurally diverse subunits in the respiratory protein hemocyanin from arthropods. Biol. Bull. 171: 90-115. National Geographic Society. 1985. .-I/An <>/ An///; America Wash- ington, DC. Pp. 66-67. Nickerson. K. \\ ., and K. E. van ilolde. 1971. A comparison of mol- luscan and arthropod hemocyanin. I. Circular dichroism and ab- sorption spectra. Comp Biochcin P/IVMO! 39B: 855-872. Olson, K., N. B. I erwilliger, and J. McDowell Capuz/.o. 1988. Structure of hemocyanin in larval and adult lobsters. Am /.mil 28: 47A. Olson, K. S., and J. McDowell. 1989. Structure and function of he- mocvamn in American lobsters. Am /.mil. 29: 20A. Pihl, I.., S. P. Baden, and R. J. Dia/. 1991. Effects of periodic hypoxia on distribution of demersal fish and crustaceans. Mar Binl 108: 349-360. Reese, J. E. 1989. Structure and function of crustacean hemocy- anins. MA Thesis. College of William and Mary, Williamsburg. VA. 75 pp. Sullinin, B., L. Pennell, B. Hutchison, and R. Mulchings. 1983. Genetics and evolution of the hemocyanin multigene-I. Genetic variability in L'ca pugilator from Beaufort, NC. Comp. Biochem. Physiol 76B: 615-618. Tracey, M. L., K. Nelson, D. lledgecock, R. A. Shleser, and M. I,. Pressick. 1975. Biochemical genetics of lobsters: Genetic variation and the structure of American lobsters (Homanis americanus) pop- ulations. ./ Fish. Res Board Can 32: 2091-2101. Truchot. J.-P. 1975. Factors controlling the in vitro and in vivo oxygen affinity of the hemocyanin of the crab, Carcinus meanas. Resp. Phvswl 24: 173-189. CONTENTS CELL BIOLOGY Costas, Eduardo, Angeles Aguilera, Sonsoles Gon- zalez-Gil, and Victoria Lopez-Rodas Contact inhibition: also a control for cell prolifer- ation in unicellular algae? DEVELOPMENT AND REPRODUCTION Fenteany, Gabriel, and Daniel E. Morse Specific inhibitors of protein synthesis do not block RNA synthesis or settlement in larvae of a marine gastropod mollusk (Haliotis rujescens) 6 Freeman, Gary Metamorphosis in the brachiopod Terebratalia: ev- idence for a role of calcium channel function and the dissociation of shell formation from settlement 15 ECOLOGY AND EVOLUTION Curtis, Lawrence A., and Karen M. K. Hubbard Species relationships in a marine gastropod-tre- matode ecological system 25 Douillet, Philippe, and Christopher J. Langdon Effects of marine bacteria on the culture of axenic oyster Crassostrea gigas (Thunberg) larvae 36 Okamura, Beth, and Lita Ann Doolan Patterns of suspension feeding in the freshwater bryozoan Pltimatella repens 52 Scheltema, Amelie H. Aplacophora as progenetic aculiferans and the coe- lomate origin of mollusks as the sister taxon of Si- puncula 57 IMMUNOLOGY Rinkevich, H., Y. Saito, and I. L. Weissman A colonial invertebrate species that displays a hi- erarchy of allorecognition responses 79 Sawada, Tomoo, Jeffrey Zhang, and Edwin L. Cooper Classification and characterization of hemocytes in Stvela dava . 87 PHYSIOLOGY Hidaka, Michio, and Kiwamu Afuso Effects of cations on the volume and elemental composition of nematocysts isolated from acontia of the sea anemone Calliactis polypus 97 Mangum, Charlotte P. Hemocyanin subunit composition and oxygen binding in two species of the lobster genus Homarus and their hybrids 105 Volume 184 THE Number 2 BIOLOGICAL BULLETIN APRIL, 1993 Published by the Marine Biological Laboratory 1993 LATE SUMMER COURSES AT THE MBL History of Biology: Human Genetics in the Twentieth Century (AUGUST 1- AUGUST 11, 1993) APPLICATION DEADLINE: MAY 21, 1993 Open to students from a wide variety of backgrounds and ranks who share an interest in the history and philosophy of human genetics and eugenics. This course will focus on the history of human genetics in the United Slates, Great Britain, France, Germany and Russia in the twentieth century. Themes will include clinical and eugenic aspects of human genetic studies, the history of efforts to control human evolution, ethical questions arising from present as well as past attempts at such control, and the social construction of scientific knowledge. Directors: Garland Allen, Washington Univer- sity; John Beany, University of Minnesota; and Jane Maienschein, Arizona Stale University. Methods in Computational Neurosdence (AUGUST 3- AUGUST 31, 1993) APPLICATION DEADLINE: MAY 21, 1993 This intensive, four week computer laboratory and lecture course for 23 advanced graduate students, postdoctoral fellows and faculty, will examine how the biophysical and biochemical properties of neurons and synapses, together with the architecture of neural circuits, produce observed animal behavior. Lecture material will include the dynamics of individual neurons and synapses; the use of exact models of single cells versus reduced neural models in analysis of networks; the coding and processing of external stimuli within nervous systems; development of the nervous system; and applied mathematics. In the laboratory, specific aspects of nervous systems will be modeled using a UNIX graphic-color workstation and software designed forthe analysis of both single-cell dynamics and large network properties. Directors: David Kleinfeld andDavidW. Tank, Biological Computation Research Department, Bell Laboratories, Murray Hill, NJ. Molecular Evolution (AUGUST 8- AUGUST 20, 1993) APPLICATION DEADLINE: JUNE 1, 1993 A series of lectures and discussions exploring multiple approaches to molecular evolution, and a computer laboratory for phylogenetic and sequence analysis. Designed for a class of 60 established investigators, postdoctoral fellows, and advanced graduate students, this two week program will provide a forum for exchange of information among organismic and molecular biologists and ecologists. Director: Mitchell L. Sogin, Marine Biological Laboratory. Cellular and Molecular Neurobiology and Development of the Leech (AUGUST 8-Aucusr28. 1993) APPLICATION DEADLINE: JUNE 1, 1993 This course is for 12 graduate and postdoctoral students and independent investigators interested in applying diverse experimental approaches (electrophysiology, biophysics, cellular and molecular biology) to the study of the nervous system and development of a single organism. Students will leam techniques and concepts of modern experimental embryology, including lineage tracing, cell cycle analysis and the analysis of gene expression by in situ hybridization and immunohistochemical procedures. Students will also use patch clamp techniques to study whole cell and single channel transmembrane currents in situ and in isolated neurons in culture . Directors: Pierre Drapeau, McGill University; and David Weisblat, University of California, Berkeley. Pathogenesis of Neuroimmunologic Diseases (AUGUST 15- AUGUST 27, 1 993) APPLICATION DEADLINE: MAY 4, 1993 This course for 30 advanced graduate students, postdoctoral fellows, and junior faculty in neurosciences and immunology, and for residents in neurology, neurosurgery or psychiatry , will consist of lectures and discussions describing the application of genetic, molecular, and cell physiologic concepts and techniques in current use in immunology and neurophysiology to the analysis of pathogenesis in the belter known neurologic and psychiatric diseases thought to have an immunologic basis. Directors: J. Murdoch Ritchie, Yale University; and Byron H. Waksman, Harvard and New York Universities. Optical Microscopy and Imaging in the Biomedical Sciences (OCTOBER 20-OcroBER 28, 1993) APPLICATION DEADLINE: JULY 19, 1993 Designed for 22 research scientists, physicians, postdoctoral trainees, and advanced graduate students in animal, plant, medical, and material sciences, as well as non- biologists seeking a comprehensive introduction to microscopy and video- imaging. The course consists of lectures, laboratory exercises, demonstrations, and discussions that will enable the participant to obtain and interpret microscope images of high quality, to perform quantitative optical measurements, and to produce photographic and video records for documentation and analysis. Instruction on state-of-the-art equipment will be provided by experienced staff from universities and industry. Director: Colin S. Izzard, State University of New York at Albany. FOR FURTHER INFORMATION AND APPLICATION FORMS contact: Ms. Donanne Chrysler, Admissions Coordinator, Marine Biological Laboratory, Woods Hole, MA 02543, USA; (508) 548-3705, ext 401 . THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Associate Editors WAV 1 PETER A. V. ANDERSON. The Whitney Laboratory. University of Florida DAVID EPEL. Hopkins Marine Station. Stanford LIniversity J. MALCOLM SHICK. University of Maine. Orono I Editorial Board WILLIAM D. COHEN, Hunter College DAPHNE GAIL FAUTIN. University of Kansas WILLIAM F. GILLY. Hopkins Marine Station. Stanford LIniversity ROGER T. HANLON, Marine Biomedical Institute. LIniversity of Texas Medical Branch CHARLES B. METZ. LIniversity of Miami K. RANGA RAO. University of West Florida RICHARD STRATHMANN, Friday Harbor Laboratories, University of Washington STEVEN VOGEL, Duke University SARAH ANN WOODIN. University of South Carolina Editor MICHAEL J. GREENBERG, The Whitney Laboratory, University of Florida Managing Editor: PAMELA L. CLAPP. Marine Biological Laboratory APRIL, 1993 Printed and Issued by LANCASTER PRESS, Inc. 3575 HEMPLAND ROAD LANCASTER, PA THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory. MBL Street, Woods Hole. Massachusetts 02543. Subscriptions and similar matter should be addressed to Subscription Manager. THE BIOLOGICAL BUL- LETIN. Marine Biological Laboratory, Woods Hole, Massachusetts 02543. Single numbers. $35.00. Sub- scription per volume (three issues), $87.50 ($175.00 per year for six issues). Communications relative to manuscripts should be sent to Michael J. Greenberg. Editor-in-Chief, or Pamela L. Clapp. Managing Editor, at the Marine Biological Laboratory. Woods Hole. Massachusetts 02543. Telephone: (508) 548-3705, ext. 428. FAX: 508-540-6902. E-mail: pamclirt'hoh. mbl.edu. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Copyright ( Q> 1993. by the Marine Biological Laboratory Second-class postage paid at Woods Hole. MA. and additional mailing offices. ISSN 0006-3185 INSTRUCTIONS TO AUTHORS The Biological Bulletin accepts outstanding original research reports of general interest to biologists throughout the world. Papers are usually of intermediate length (10-40 manuscript pages). A limited number of solicited review papers may be ac- cepted after formal review. A paper will usually appear within four months after its acceptance. Very short, especially topical papers (less than 9 manuscript pages including tables, figures, and bibliography) will be pub- lished in a separate section entitled "Research Notes." A Re- search Note in The Biological Bulletin follows the format of similar notes in Nature. It should open with a summary para- graph of 150 to 200 words comprising the introduction and the conclusions. The rest of the text should continue on without subheadings, and there should be no more than 30 references. References should be referred to in the text by number, and listed in the Literature Cited section in the order that they appear in the text. Unlike references in Nature, references in the Re- search Notes section should conform in punctuation and ar- rangement to the style of recent issues of The Biological Bulletin. Materials and Methods should be incorporated into appropriate figure legends. See the article by Lohmann el at. (October 1990, Vol. 179: 2 1 4-2 1 8) for sample style. A Research Note will usually appear within two months after its acceptance. The Editorial Board requests that regular manuscripts con- form to the requirements set below, those manuscripts that do not conform will be returned to authors for correction before review. 1. Manuscripts. Manuscripts, including figures, should be submitted in triplicate. (Xerox copies of photographs are not acceptable for review purposes.) The original manuscript must be typed in no smaller than 12 pitch, using double spacing (in- cluding figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper. 8'/2 by 1 1 inches. Please, no right justification. Manuscripts should be proofread carefully and er- rors corrected legibly in black ink. Pages should be numbered consecutively. Margins on all sides should be at least 1 inch (2.5 cm). Manuscripts should conform to the Council of Biology Ed- itors Style Manual, 5th Edition (Council of Biology Editors. 1983) and to American spelling. Unusual abbreviations should be kept to a minimum and should be spelled out on first reference as well as defined in a footnote on the title page. Manuscripts should be divided into the following components: Title page. Abstract (of no more than 200 words). Introduction. Materials and Methods. Results. Discussion. Acknowledgments. Literature Cited. Tables, and Figure Legends. In addition, authors should supply a list of words and phrases under which the article should be indexed. 2. Title page. The title page consists of: a condensed title or running head of no more than 35 letters and spaces, the manuscript title, authors' names and appropriate addresses, and footnotes listing present addresses, acknowledgments or contri- bution numbers, and explanation of unusual abbreviations. 3. Figures. The dimensions of the printed page, 7 by 9 inches, should be kept in mind in preparing figures for publi- cation. We recommend that figures be about 1 '/: times the linear dimensions of the final printing desired, and that the ratio of the largest to the smallest letter or number and of the thickest to the thinnest line not exceed 1:1.5. Explanatory matter gen- erally should be included in legends, although axes should always be identified on the illustration itself. Figures should be prepared for reproduction as either line cuts or halftones. Figures to be reproduced as line cuts should be unmounted glossy photo- graphic reproductions or drawn in black ink on white paper, good-quality tracing cloth or plastic, or blue-lined coordinate paper. Those to be reproduced as halftones should be mounted on board, with both designating numbers or letters and scale bars affixed directly to the figures. All figures should be numbered in consecutive order, with no distinction between text and plate figures. The author's name and an arrow indicating orientation should appear on the reverse side of all figures. 4. Tables, footnotes, figure legends, etc. Authors should follow the style in a recent issue of The Biological Bulletin in preparing table headings, figure legends, and the like. Because of the high cost of setting tabular material in type, authors are asked to limit such material as much as possible. Tables, with their headings and footnotes, should be typed on separate sheets, numbered with consecutive Roman numerals, and placed after the Literature Cited. Figure legends should contain enough in- formation to make the figure intelligible separate from the text. Legends should be typed double spaced, with consecutive Arabic numbers, on a separate sheet at the end of the paper. Footnotes should be limited to authors' current addresses, acknowledg- ments or contribution numbers, and explanation of unusual abbreviations. All such footnotes should appear on the title page. Footnotes are not normally permitted in the body of the text. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more than two authors cited as Jones el ai, 1980. Personal communications and ma- terial in preparation or in press should be cited in the text only, with author's initials and institutions, unless the material has been formally accepted and a volume number can be supplied. The list of references following the text should be headed Lit- erature Cited, and must be typed double spaced on separate pages, conforming in punctuation and arrangement to the style of recent issues of The Biological Bulletin. Citations should in- clude complete titles and inclusive pagination. Journal abbre- viations should normally follow those of the U. S. A. Standards Institute (USASI), as adopted by BIOLOGICAL ABSTRACTS and CHEMICAL ABSTRACTS, with the minor differences set out below. The most generally useful list of biological journal titles is that published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of Serials; the most recent issue). Foreign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PE- RIODICALS, may find a booklet published by the Biological Council of the U.K. (obtainable from the Institute of Biology, 41 Queen's Gate, London, S.W.7, England, U.K..) useful, since it sets out the WORLD LIST abbreviations for most biological journals with notes of the USASI abbreviations where these differ. CHEMICAL ABSTRACTS publishes quarterly supplements of ad- ditional abbreviations. The following points of reference style for THE BIOLOGICAL BULLETIN differ from USASI (or modified WORLD LIST) usage: A. Journal abbreviations, and book titles, all underlined (for italics) B. All components