"""""""•"•••^SM Volume 172 Number 1 fl 1387 BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board RUSSELL F. DOOLITTLE, University of California at WILLIAM R. JEFFERY, University of Texas at Austin San Diego GEORGE M. LANGFORD, University of North WILLIAM R. ECKBERG, Howard University Carolina at Chapel Hill ROBERT D. GOLDMAN, Northwestern University GEORGE D. PAPPAS, University of Illinois at Chicago C. K. GOVIND, Scarborough Campus, University SIDNEY K. PIERCE, University of Maryland of Toronto HERBERT SCHUEL, State University of New York at JUDITH P. GRASSLE, Marine Biological Laboratory Buffalo MICHAEL J. GREENBERG, C. V. Whitney Marine VIRGINIA L. SCOFIELD, University of California at Laboratory, University of Florida Los Angeles School of Medicine MAUREEN R. HANSON, Cornell University LAWRENCE B. SLOBODKIN, State University of New York at Stony Brook JOHN E. HOBBIE, Marine Biological Laboratory JOHN D. STRANDBERG, Johns Hopkins University LIONEL JAFFE, Marine Biological Laboratory DONALD P. WOLF, Oregon Regional Primate Center HOLGER W. JANNASCH, Woods Hole Oceanographic Institution SEYMOUR ZIGMAN, University of Rochester Editor: CHARLES B. METZ, University of Miami FEBRUARY, 1987 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. 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 THE BIOLOGICAL BULLETIN, Marine Bio- logical Laboratory, Woods Hole, Massachusetts. Single numbers, $20.00. Subscription per volume (three issues), $50.00 ($ 1 00.00 per year for six issues). Communications relative to manuscripts should be sent to Dr. Charles B. Metz, Editor, or Pamela Clapp, Assistant Editor, at the Marine Biological Laboratory', Woods Hole, Massachusetts 02543. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Copyright © 1987, 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). Very short papers (less than 10 manuscript pages including tables, figures, and bibliography) will be published in a separate section entitled "Short Reports." A limited number of solicited review papers may be accepted after formal review. A paper will usually appear within four months after its acceptance. The Editorial Board requests that manuscripts conform to the requirements set below; those manu- scripts which 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 double spacing (including figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper, 8'/z by 1 1 inches. Manuscripts should be proofread carefully and errors 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 Editors Style Manual, 4th Edition (Council of Biology Editors, 1 978) 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 addi- tion, authors should supply a list of words and phrases under which the article should be indexed. 2. Figures. Figures should be no larger than 8'/2 by 1 1 inches. The dimensions of the printed page, 5 by 73/4 inches, should be kept in mind in preparing figures for publication. We recommend that figures be about l'/2 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 generally 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 photographic 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. 3. 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 information 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, acknowledgments 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. 4. A condensed title or running head of no more than 35 letters and spaces should appear at the top of the title page. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more than two authors cited as Jones et al, 1980. Personal communications and material 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 LITERATURE 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 include complete titles and inclusive pagination. Journal abbreviations should normally follow those of the U. S. A. Standards Institute (USASI), as adopted by BIOLOGICAL ABSTRACTS and CHEMICAL AB- STRACTS, 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). For- eign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PERIODICALS, may find a booklet published by the Biological Council of the U.K. (obtainable from the Institute of Biology, 4 1 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 additional 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 period, 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 abbrevi- ations invented by the author. For example, use Rit Visindajjelags Islendinga without abbreviation. F. All single word journal titles in full (e.g. Veliger, Ecology, 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 preferred forms rather than WORLD LIST or USASI usage (e.g. Nature, Science, Evolution NOT Nature, Lond., Science, N.Y.; Evolution, Lancaster, Pa.) 6. Reprints, charges. Authors will be charged the excess over $100 of the total of (a) $30 for each printed page beyond 1 5, (b) $30 for each table, (c) $ 1 5 for each formula more complex than a single line with simple subscripts or superscripts, and (d) $ 1 5 for each figure, with figures on a single plate all consid- ered one figure and parts of a single figure on separate sheets considered separate figures. Reprints may be ordered at time of publication and normally will be delivered about two to three months after the issue date. Authors (or delegates or 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). 11 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY i Editorial Board RUSSELL F. DOOLITTLE, University of California at WILLIAM R. JEFFERY, University of Texas at Austin San Diego GEORGE M. LANGFORD, University of North WILLIAM R. ECKBERG, Howard University Carolina at Chapel Hill ROBERT D. GOLDMAN, Northwestern University GEORGE D. PAPPAS, University of Illinois at Chicago C. K. GOVIND, Scarborough Campus, University SIDNEY K. PIERCE, University of Maryland of Toronto HERBERT SCHUEL, State University of New York at JUDITH P. GRASSLE, Marine Biological Laboratory Buffalo MICHAEL J. GREENBERG, C. V. Whitney Marine VIRGINIA L. SCOFIELD, University of California at Laboratory, University of Florida Los Angeles School of Medicine MAUREEN R. HANSON, Cornell University LAWRENCE B. SLOBODKJN, State University of New York at Stony Brook JOHN E. HOBBIE, Marine Biological Laboratory JOHN D. STRANDBERG, Johns Hopkins University LIONEL JAFFE, Marine Biological Laboratory DONALD P. WOLF, Oregon Regional Primate Center HOLGER W. JANNASCH, Woods Hole Oceanographic Institution SEYMOUR ZIGMAN, University of Rochester Editor: CHARLES B. METZ, University of Miami FEBRUARY, 1987 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. in The BIOLOGICAL BULLETIN is issued six times a year at the Lancaster Press, Inc., Prince and Lemon Streets, Lancaster, Penn- sylvania. Subscriptions and similar matter should be addressed to The Biological Bulletin, Marine Biological Laboratory, Woods Hole, Massachusetts. Single numbers, $20.00. Subscription per volume (three issues), $50.00 ($100.00 per year for six issues). Communications relative to manuscripts should be sent to Dr. Charles B. Metz, Editor, or Pamela Clapp, Assistant Editor, Marine Biological Laboratory, Woods Hole, Massachusetts 02543. THE BIOLOGICAL BULLETIN (ISSN 0006-3185) POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Second-class postage paid at Woods Hole, MA, and additional mailing offices. LANCASTER PRESS, INC., LANCASTER, PA IV Reference: Biol. Bull. 172: 1-9. (February, 1987) HORMONAL MODULATION OF PHEROMONE-MEDIATED BEHAVIOR IN A CRUSTACEAN RICHARD A. GLEESON, MICHAEL A. ADAMS', AND AMOS B. SMITH, III1 2 C. V. Whitney Laboratory, University of Florida, Rt. 1, Box 121, St. Augustine, Florida 32086; {Monell Chemical Senses Center and 2 Department of Chemistry, University of Pennsylvania, Philadelphia, Pennsylvania 19104 ABSTRACT A stereotyped courtship display is normally triggered in the male blue crab, Calli- nectes sapidus, by a pheromone released from pubertal females. Following bilateral eyestalk ligation, ablation, or optic tract transection, males do not respond to the pheromone, suggesting that neural pathways in the eyestalk ganglia are important for processing or transmitting pheromone stimulus information. Interestingly, males begin to exhibit spontaneous display behavior within a few days following eyestalk ligation or ablation, but not if only the optic tracts are transected. We propose that the loss of a circulating eyestalk factor, which moderates the activity of CNS pathways controlling courtship display, is responsible for the induction of the spontaneous be- havior. This factor may normally control pheromone receptivity in males by modu- lating the excitability of these CNS pathways either directly or by acting via an inter- mediate(s); possibly by regulating the activity of the androgenic glands which exhibit massive hypertrophy following eyestalk ligation. INTRODUCTION A fundamental goal of neuroethology is to understand the mechanisms which regulate the "behavioral state" of an animal (Truman and Weeks, 1 985). Such regula- tion assures that the organism's behavior represents an appropriate interface between its internal physiological condition and the external environment. Hormonal con- trols are frequently of major importance in this regulation, particularly in reproduc- tive behavior; although depending on the animal's reproductive strategy, social and environmental cues also can be of paramount importance (Crews and Moore, 1986). Pheromone communication plays a critical role in the reproductive activities of ani- mals in nearly all major phyla (Shorey, 1976). Frequently these chemical messages trigger very specific stereotyped" behaviors, which ultimately insure that mating is successful. In this report we consider a mechanism by which internal factors may regulate sensory activation of such a behavioral program. The reproductive behavior of C. sapidus is coordinated in part by a pheromone that is present in the urine of pubertal females and detected by males via chemorecep- tor sensilla (aesthetascs) on the outer flagellum of the first antennae (Gleeson, 1980, 1982). In receptive males this pheromone triggers a stereotyped courtship display which is characterized by lateral spreading of the chelae, extension of the walking legs, and lateral waving of the swimming appendages above the carapace. Immediately following the display, the male grasps the female and carries her beneath him for up to several days until she undergoes her maturity molt. Copulation is initiated soon Received 28 May 1986; accepted 30 October 1986. 2 R. A. GLEESON ET AL. thereafter, and this is normally the only time during her entire life that the female will mate. In experiments which have focused on determining the chemical nature of this pheromone, males are routinely screened for pheromone receptivity before use in bioassays (Gleeson et al, 1984). Interestingly, this screening procedure has revealed apparent cycles of receptivity within the male population; i.e., at times very few males will display when presented with a pheromone stimulus, whereas at other periods nearly all males are responsive. No obvious correlations between these peaks of recep- tivity and seasonal mating activity, molt stage, or cycles of environmental parameters have yet been identified. This apparent periodicity in receptivity suggested that hormonal controls may be operating. If so, the androgenic glands might be potential sources of such a regulatory factor since they are known to directly control the development and maintenance of both primary and secondary sexual characteristics in male crustaceans (Charniaux- Cotton and Payen, 1985). Because an eyestalk hormone (possibly gonad inhibiting hormone) has been shown to moderate the activity of the androgenic glands in several crustacean species (Adiyodi and Adiyodi, 1970; Adiyodi, 1985), we reasoned that ligating the eyestalks in male blue crabs may be an approach to increasing androgenic hormone production and consequently enhancing the males' pheromone receptivity; i.e., if, in fact, an androgenic hormone affects receptivity. This hypothesis is only testable, however, if the olfactory pathways in the eyestalk ganglia [e.g., the medulla terminalis (Ache and Fuzessery, 1979)] are not critical for processing pheromone stimulus information. In this study we present evidence suggesting the existence of hormonal modula- tion of pheromone-triggered courtship display behavior in C. sapidus and begin to define the neural substrate affected by this modulation. MATERIALS AND METHODS All animals were obtained from commercial sources, maintained in holding tanks with flow-through seawater systems, and sustained on a mixed diet offish, squid, and shrimp which was offered 2-3 times weekly. Molt stages were assessed according to the criteria of Van Engel (1958). Eyestalk ligations were performed by tightly tying off each eyestalk at its base in the region of the arthrodial membrane using a suitable length of 000 suturing silk. In animals in which the eyestalks were ablated, amputations were made just distal to a ligature tied at the base of each eyestalk; this procedure considerably reduced the loss of hemolymph from the stump. The ligation and ablation procedures are equivalent in their effects in that with ligation there is an immediate loss of visual input as judged by the animal's lack of response to visual cues that would normally trigger a defense posture. Furthermore, necrosis of eyestalk tissue is quite apparent within 24 h post- ligation with the eyestalk frequently becoming detached after a few days. Transec- tions of the optic tracts were achieved using small scissors (Mini-Vannas) which were inserted through a transverse slit, approximately 2 mm in length, made in the mesial surface of the arthrodial membrane at the base of each eyestalk. For behavioral observations, all animals were held individually in 40-liter aquaria, each having a flow-through seawater supply. Behavioral activity was monitored in one of two ways. In the first, observations were made daily for three 15-min periods (morning, midday, and afternoon) throughout the duration of the experiment. In a second approach, crab activity was automatically recorded on video cassette for 40 s intervals, every half hour, 24 h a day. Using this latter method, up to eight animals could be monitored simultaneously. O k_ O HORMONAL MODULATION OF BEHAVIOR Days Post-ligation I 2 3 4 5 6 7 8 9 10 I I 12 13 14 I 2 3 4 5 6 7 W FIGURE 1 . Occurrence of "spontaneous display behavior" (stippled squares) following bilateral eye- stalk ligation of adult Callinectes sapidus males. All animals were observed for three 15 min periods daily. No data were collected on day 1 1 . Pheromone receptivity of males was tested by introducing a 300 n\ sample of pubertal-female urine into the inlet flow of each aquarium and observing the male's behavior for two minutes. In all tests the urine samples were derived from pooled batches which, in our bioassay system, reliably trigger display behavior in intact, re- ceptive males within two minutes. For histological processing of the androgenic glands, the subterminal regions of the deferent ducts were dissected out, fixed with 2.5% glutaraldehyde in isosmotic Callinectes saline (Perkins and Wright, 1969) and refrigerated overnight. This was followed by treatment in Carnoy's fluid for 90 min at 4°C then dehydration and imbedding in paraffin. Tissue cross-sections (10 nm thick) were subsequently sub- jected to standard hematoxylin and eosin staining. RESULTS AND DISCUSSION A pilot study to examine the effects of eyestalk ligation on pheromone receptivity was performed using four non-receptive adult males. In these trials a pheromone stimulus was presented to each animal, daily, for several days following bilateral eye- stalk ligation. The results were surprising. Exposure to the pheromone during the first four days post-ligation did not trigger courtship behavior; however, on days five and six, three of the males began to exhibit courtship display activity in the absence of a pheromone stimulus. This behavior was, in all respects, identical to the stereotyped display activity that is normally triggered by the pheromone of pubertal females. Car- rying behavior, which normally follows courtship display, was also initiated by these spontaneously active animals when other males were placed in their tanks. Bouts of the "spontaneous display behavior" (SDB), lasting up to several hours, occurred for approximately three days; thereafter the activity appeared to decline. Because of this spontaneous behavior, the results of tests to evaluate pheromone receptivity in these animals were inconclusive. To explore this phenomenon further, an experiment was conducted using 30 in- termolt males of mature size (carapace short-width = 1 13.3 mm, SEM ± 0.8), 15 of which were randomly selected for bilateral eyestalk ligations; the remaining 1 5 served as controls. All animals were observed three times daily over a 14 day period to moni- tor the occurrence of SDB. Of the 10 males surviving ligation, 7 exhibited SDB, whereas no such activity was observed in the 1 5 control animals (Fig. 1 ). The sponta- R. A. GLEESON ET AL. LJ IX "CL Q >*— o O) .O E 3- 2- I- 0 ii 1 J I i 3 4 7 8 9 10 Days Post-ablation FIGURE 2. Frequency profile of "spontaneous display behavior" for 1 2 adult Callinectes sapidus males following bilateral eyestalk ablation. Behavior was sampled for 40 s intervals every half hour over 10 days. Shown are the mean number of intervals in which display activity occurred during each six hours post-ablation. neous behavior was not seen until the fifth day after ligation, reaching a peak on day seven before subsiding. The profile of this behavior was more precisely documented in a group of 1 2 adult, intermolt males (carapace short-width = 120.4 mm, SEM ± 1.2) in which the eyestalks were ablated. In this experiment, activity was automatically recorded on video cassette. The mean number of 40 s intervals in which SDB was observed during each 6 hours commencing on the day following ablation are shown in Figure 2. Again the profile reveals a delayed onset of the SDB with a maximum frequency between day five through eight, followed by a decline. These results suggested that the eyestalks are sites for the production and/or re- lease of a hormonal factor that modulates the activity of CNS pathways controlling courtship display behavior; e.g., by regulating central pattern-generator circuits that coordinate the display. However, since the ligation/ablation procedure also removes a substantial portion of the crab's nervous system (i.e., ganglia located in the eye- stalks), an alternative hypothesis is that the procedure eliminates "inhibitory neural inputs" from these ganglia to courtship display "centers" located elsewhere in the CNS. To distinguish between these alternatives, the optic tracts, which link the eyestalk ganglia with the brain (supraesophageal ganglion), were bilaterally transected in a group of 19 adult, intermolt males (carapace short-width = 1 17.8 mm, SEM ± 2.5). A second group of five adult, intermolt males (carapace short-width =118 mm, SEM ± 4.7) received eyestalk ligations. The behavior of all animals was monitored three times daily over a ten day period. Dissection revealed that the transections were com- plete in 17 of the 19 animals subjected to the procedure, and in virtually all cases, hemolymph circulation to the eyestalks was not compromised as judged by the lack of necrosis in the eyestalk tissues 10 days after the operation. Of the 17 "transected" animals, only a single crab exhibited SDB. In contrast, four of the five "ligated" males HORMONAL MODULATION OF BEHAVIOR GO 100- 0 C/) 0> 80- C !5 lc 60- X Ld - ~o 40- E "c < 20- "o - o^ n - Transected (N = I7) Ligated (N = 5) 8 Days Post-treatment I 10 FIGURE 3. Occurrence of "spontaneous display behavior" (SDB) in adult Callinectes sapidus males following either bilateral eyestalk ligation (ligated) or optic tract transection (transected). were induced to display (Fig. 3). The difference in the frequency of occurrence of SDB between experimental groups is significant (P = 0.003, Fisher exact test). These findings therefore demonstrate that the loss of any inhibitory neural inputs from the eyestalk ganglia is not a factor in the induction of SDB, and support the concept of an eyestalk hormone which acts to modulate this behavior. The latter view is a particularly tenable hypothesis considering that the transection procedure leaves only the integument and hemolymph linking the eyestalks to the animal. The operation therefore approximates (although it does not replace) the implantation of eyestalk tissue as a means of demonstrating the existence of a hemolymph-borne factor. That the androgenic glands may be involved as possible intermediates in the in- duction of SDB is suggested by the hypertrophy they undergo following eyestalk liga- tion. The massive increase in size of the androgenic tissue depicted in Figure 4 is representative of what is consistently observed in adult, intermolt males subjected to the ligation procedure, and is quite apparent under the dissecting microscope as soon as day five following ligation. At least part of the increase in size can be attributed to hypertrophy of individual cells. Control animal tissue was characterized by cords of compact, linearly arranged cells; whereas in eyestalk-ligated males, the cytoplasm was typically more basophilic and greatly increased in volume. Under the light micro- scope, the appearance of the androgenic tissue in males actively exhibiting SDB was indistinguishable from that taken from animals up to 14 days post-ligation, a time at which the frequency of SDB activity has subsided. Hypertrophy was not observed in males in which the optic tracts were transected. To assess the effects of eyestalk ligation on pheromone receptivity, 18 "ligated" males exhibiting SDB were screened for their response to pheromone-containing urine. The urine was presented during a period of quiescence between bouts of SDB. Only one of the eyestalk-ligated males, however, exhibited display behavior with urine presentation, and the response latency suggested that it may have been a bout of SDB. To explore this deficit in receptivity further, a group of seven highly receptive, A R. A. GLEESON ET AL. VD B FIGURE 4. Cross-sections of the posterior vas deferens ( VD) with associated androgenic gland ( AG) from a control (A) and experimental male (B) 1 4 days following bilateral eyestalk ligation. Scale bar equals 200 microns. HORMONAL MODULATION OF BEHAVIOR 7 intact males, which repeatedly exhibited vigorous display activity with urine presen- tations, were tested 24 h after either bilateral eyestalk ligation (two animals) or optic tract transection (five animals). In these trials, the pheromone stimulus did not induce any of the animals to display. These results suggest, first, that disrupting the neural connections between the eyestalks and the supraesophageal ganglion blocks chemosensory activation of court- ship display "centers" in the CNS; thus implying that neural pathways within the eyestalk ganglia are, indeed, important for processing or transmitting pheromone stimulus information. Two lines of reasoning support this view: ( 1 ) males which reli- ably respond to a pheromone stimulus immediately lose their receptivity if the neural connections to the eyestalks are severed; and (2) since the motor program for display behavior remains intact following eyestalk ligation/ablation as revealed by SDB activ- ity, the lack of response to a pheromone stimulus in eyestalk-lesioned males cannot be attributed to the animal's inability to perform the display, implying that chemo- sensory activation of the program is in some way blocked. Interestingly, in several crustaceans bilateral eyestalk lesions have also been shown to profoundly attenuate food-search behavior elicited by chemical stimulation of the antennules (Maynard and Sallee, 1 970; Hazlett, 1 97 1 ). A second point to be made from these data is that the SDB observed in ligated males is not attributable to hyperexcitability of pheromone receptor cells in the antennules (although such hyperexcitability may in fact also oc- cur), but rather involves changes in the excitability of CNS pathways mediating dis- play behavior. This is further supported by the fact that in nine eyestalk-ligated males in which the antennules were also bilaterally ablated, SDB was nevertheless induced. Considering the data together we propose the following relationships as a working hypothesis (Fig. 5). In receptive males, stimulation of antennular pheromone recep- tors normally activates CNS "centers" controlling courtship display; and in the path- way linking the antennular receptors with these "centers," neural connections within the eyestalk ganglia appear to be important components. We propose that the excit- ability of the display "centers," and consequently the threshold for activation by a pheromone stimulus, is modulated by an eyestalk hormone; possibly a neurosecre- tory product released into the hemolymph via the sinus gland. Exactly how this mod- ulation is effected is not known. The action could be direct [e.g., via the action of neuro-depressing hormone (Arechiga et al., 1977, 1979)], and/or entail one or more intermediates. One possibility of the latter is that a hormone(s) from the androgenic glands is involved, as is suggested by the hypertrophy these glands undergo following eyestalk ligation. Indeed, a positive correlation between androgenic gland size and seasonal changes in the testes (Meusy, 1963; Payen, 1973), vas deferens (Adiyodi, 1985), and external male morphology (Carpenter and DeRoos, 1970; Thampy and John, 1973; Dudley and Jegla, 1978) has been noted in other crustaceans. Thus, if an androgenic hormone acts to turn on or increase the excitability of CNS courtship- display "centers," the eyestalk hormone may, in fact, be acting indirectly on these "centers" by modulating androgenic gland activity. Alternatively, the action of the eyestalk hormone may be mediated via other pathways; for example, by regulating the production or activity of a gonad stimulating hormone having CNS actions (Adi- yodi and Adiyodi, 1970;Eastman-ReksandFingerman, 1984; Kulkarni £tf #/., 1984). The temporal profile of the SDB activity is intriguing and will quite likely provide an important clue towards understanding the mechanisms underlying this phenome- non. For example, the latency to onset of SDB may reflect the time required to appro- priately raise the titer of a circulating androgenic hormone and/or reflect the time course of metabolic processes giving rise to the hyperexcitability in the neural ele- ments affected by the hormone. The decline in SDB frequency by day 1 0 post-ligation 8 R. A. GLEESON ET AL. Pheromone Stimulus Pheromone Receptors on Antennules Eyestalk Ganglia Eyestalk V Hormone J ^^^^^™^^^^^^ s ^ f \ i Androgenic > vv Glands ? / CNS Centers Controlling Courtship Display FIGURE 5. Proposed model for activation and modulation of CNS "centers" controlling courtship display behavior. See text for discussion. (Figs. 1 , 2) may result from desensitizing mechanisms within the CNS or from com- pensatory metabolic processes which reduce the circulating titer of hormone. If the androgenic glands are indeed involved in the induction of SDB, the fact that their hypertrophy (and presumably hypersecretion) is evident well past the peak in SDB activity would suggest that mechanisms which compensate for a high titer of circulat- ing androgenic factor are responsible for the reduction in SDB by day 10. Certainly, deciphering the exact mechanisms underlying SDB should provide important insight towards understanding processes regulating the sensory activation of behavior. ACKNOWLEDGMENTS We thank B. W. Ache, W. E. S. Carr, and M. J. Greenberg for their critical review of the manuscript. This work was supported by a grant from the National Science Foundation (#PCM-8038776). LITERATURE CITED ACHE, B. W., AND Z. M. FUZESSERY. 1979. Chemosensory integration in the spiny lobster: ascending activity in the olfactory-globular tract. J. Comp. /V/vs/o/. 130: 63-69. ADIYODI, K. G., AND R. G. ADIYODI. 1970. Endocrine control of reproduction in decapod Crustacea. Biol.Rev.45: 121-165. ADIYODI, R. G. 1985. Reproduction and its control. Pp. 147-216 in The Biology of Crustacea, D. E. Bliss and L. H. Mantel, eds. Academic Press, New York. ARECHIGA, H., C. CABRERA-PERALTA, AND A. HUBERMAN. 1979. Functional characterization of the neurodepressing hormone in the crayfish. J. Neurobiol. 10: 409-422. HORMONAL MODULATION OF BEHAVIOR ARECHIGA, H., A. HUBERMAN, AND A. MARTINEZ-PALOMO. 1977. Release of a neurodepressing hormone from the crustacean sinus gland. Brain Res. 128: 93-108. CARPENTER, M., AND R. DERoos. 1970. Seasonal morphology and histology of the androgenic gland of the crayfish, Orconectesnais. Gen. Comp. Eridocrinol. 15: 143-157. CHARNIAUX-COTTON, H., AND G. PAYEN. 1985. Sexual differentiation. Pp. 217-299 in The Biology of Crustacea, D. E. Bliss and L. H. Mantel, eds. Academic Press, New York. CREWS, D., AND M. C. MOORE. 1986. Evolution of mechanisms controlling mating behavior. Science 231: 121-125. DUDLEY, H. G., AND T. C. JEGLA. 1978. Control of reproductive form in male crayfish. Gen. Comp. Endocrinol.34: 108. EASTMAN-REKS, S., AND M. FINGERMAN. 1984. Effects of neuroendocrine tissue and cyclic AMP on ovarian growth in vivo and in vitro in the fiddler crab, Uca pugilator. Comp. Biochem. Physiol. 79A: 679-684. GLEESON, R. A. 1980. Pheromone communication in the reproductive behavior of the blue crab, Calli- nectes sapidns. Mar. Behav. Physiol. 7: 1 19-134. GLEESON, R. A. 1982. Morphological and behavioral identification of the sensory structures mediating pheromone reception in the blue crab, Callinectes sapidus. Biol. Bull. 163: 1 62- 171. GLEESON, R. A., M. A. ADAMS, AND A. B. SMITH, III. 1984. Characterization of a sex pheromone in the blue crab, Callinectes sapidus: crustecdysone studies. J. Che m. Ecol. 10: 9 1 3-92 1 . HAZLETT, B. A. 197 1 . Non-visual functions of crustacean eyestalk ganglia. Z Vergl. Physiol. 71: 1-13. KULKARNI, G. K., R. NAGABHUSHANAM, AND P. K. JOSHI. 1984. Neuroendocrine control of reproduc- tion in the male penaeid prawn, Parapenaeopsis hardwickii (Miers) (Crustacea, Decapoda, Pen- aeidae). Hydrobiologia 108: 281-289. MAYNARD, D. M., AND A. SALLEE. 1970. Disturbance of feeding behavior in the spiny lobster, Panulirus argus, following bilateral ablation of the medulla terminalis. Z. Vergl. Physiol. 66: 123-140. MEUSY, J.J.I 963. La gametogenese SOrchestia gammarella Pallas, Crustace Amphipode. Bull. Soc. Zool. ' Fr. 88: 197-220. PAYEN, G. G. 1973. Etude descriptive des principals etapes de la morphogenese sexuelle chez un Crustace Decapode a developpement condense, TEcrevisse Pontastacus leptodactylus leptodactylus (Esch- scholtz, 1823). Ann. Embryol. Morphog. 6: 179-206. PERKINS, M. S., AND E. B. WRIGHT. 1969. The crustacean axon: metabolic properties, ATPase activity, calcium binding and bioelectric correlations. J. Neurophysiol. 32: 930-947. SHOREY, H. H. 1976. Animal Communication by Pheromones, Academic Press, New York. 1 76 pp. THAMPY, D. M., AND P. A. JOHN. 1973. Observations on variations in the male sex characters and their relation to the androgenic gland in the shrimp Macrobrachium idae (Heller). Ada. Zool. (Stock- holm) 54: 193-200. TRUMAN, J. W., AND J. C. WEEKS. 1985. Activation of neuronal circuits by circulating hormones in insects. Pp. 38 1 -399 in Model Neural Networks and Behavior, A.I. Selverston, ed. Plenum Press, New York. VAN ENGEL, W. A. 1958. The blue crab and its fishery in Chesapeake Bay. Commercial Fish. Rev. 20: 6- 17. Reference: Biol. Bull. 172: 10-29. (February, 1987) CRUSTACEAN CHEMICAL PERCEPTION: TOWARDS A THEORY ON OPTIMAL CHEMORECEPTION RICHARD K. ZIMMER-FAUST Marine Science Institute, University of California, Santa Barbara, California 93106 ABSTRACT Experiments explored how Crustacea respond to odors that could signal energy and nutrient properties of food. While free amino acids are abundant in living prey, they diffuse rapidly from carrion and are assimilated by biodegradatory bacteria con- comitant with release of ammonia. The ratio of amino acids to ammonia decreases with increasing carrion age and is proposed to signal the relative nutritional (nitrogen) quality of food. Supporting this hypothesis are data showing six bathypelagic and littoral species probing and searching to amino acids, but not to ammonia. Amino acids caused forward ambulation (searching) in some species while ammonia induced only tail-flipping (fleeing) in others. Interactions between amino acids and ammonia were clearly antagonistic, since mixtures combining these substances suppressed both feeding and fleeing. Additional experiments with the spiny lobster, Panulirus inter- ruptus, demonstrate that excitation by the high energy molecule, ATP > ADP > AMP = adenosine, which is the pathway followed in autolytic degradation of ani- mal flesh. Solutions identical in ATP evoked searching and feeding according to their ranked order of adenylate energy charge. Results indicate that Crustacea respond best to odors that may specify food of highest energy and nutrient content, though re- sponses differ between species and probably reflect differences in dietary habits. INTRODUCTION A major problem for animals is how to exploit food resources efficiently. Ability to remotely sense food quality, quantity, and effort needed for food capture is clearly advantageous. This minimizes time choosing food and maximizes net rate of energy or nutrient gain (Hughes, 1979; Erichsen et al, 1980; Orians, 1981). The decision to feed is based on many factors, including internal motivational states (level of starva- tion, social dominance, sex, and reproductive status; McFarland, 1971; Dethier, 1976; Harrison, 1983; Tilson and Hamilton, 1984; Bell et al., 1985) and external environmental qualities, such as presence of predators or competitors (Milinski and Heller, 1978; Nelson and Vance, 1979; Barnard and Brown, 1981; Kacelnik et al., 1981; Lima, 1 985a; Schmitt and Holbrook, 1 985). Foragers are influenced by stimuli specific to food, especially by smell and taste, which are vital to ingestion (Linstedt, 1971; Dethier, 1976; Ache, 1982; Croll, 1983, among others) and to activating and orienting search (Bell and Tobin, 1 982). Chemical senses evolve to increase efficiency of resource use, given the goal of a forager is to minimize time spent acquiring set reward or is to maximize net rate of energy or nutrient gain. Models of optimal foraging usually assume that grazers have complete prior knowledge of resource distributions, or that such knowledge is gained by cost-effec- tive sampling. An implicit assumption is that foragers must distinguish the compara- Received 2 October 1986; accepted 25 November 1986. 10 OPTIMAL CHEMORECEPTION 1 1 live worth of prey, usually by some unspecified mechanism in order to make the "right" energetic decision when grazing in a mixed prey assemblage. How sampling proceeds is poorly understood (but see Krebs et al, 1978; Lima, 1985b), especially with regard to sensory mechanisms and perceptual ability. This investigation determined how Crustacea perceive chemicals that could spec- ify energy and nutrient qualities of food. Marine Crustacea were used because they depend almost exclusively on chemoreception in feeding (Hazlett, 1968; Schembri, 1982; Zimmer-Faust and Case, 1982a) and their behavior is highly stereotyped and easily quantified (Fuzessery and Childress, 1975; Field, 1977; Pearson et al., 1979; Zimmer-Faust and Case, 1 982b). Crustacea respond to a wide range of low molecular weight compounds that are readily synthesized (McLeese, 1970; Shelton and Mackie, 1971;Mackie, 1973;Carr, 197 8; Robertson^ al., 1981; Zimmer-Faust^^/., 1984b), which makes them highly useful in chemical senses research. A verbal model Free amino acids are abundant as osmolytes in tissues of all living aquatic inverte- brates, the principal dietary items of crustacean omnivores (Clark, 1968; Gerard and Gilles, 1972; Clark and Zounes, 1977; Bowlus and Somero, 1979; Zurburg and De- Zwaan, 198 1 ; Yancey et al., 1982). Because amino acids diffuse rapidly from carrion once a prey dies (Rittschof, 1980; Zimmer-Faust and Case, 1982a), they probably specify living or freshly killed prey. Free amino acids can be used as a source of nitro- gen for protein synthesis and comprise as much as 8- 1 0% of the dry weight of inverte- brate flesh. Ammonia, on the other hand, is a nitrogenous waste product of transami- nation in protein catabolism and non-nutritive to Crustacea (Campbell, 1973). Biode- gradatory bacteria selectively assimilate other low molecular weight compounds, including amino acids (Ogura, 1 975), while they produce copious amounts of ammo- nia (Hollibaugh, 1979; Kjosbakken et al, 1983). The ratio of amino acids to ammo- nia decreases with increasing carrion age and decomposition (Kjosbakken et al., 1983; Michel and Zimmer-Faust, unpubl. data); consequently, a corresponding re- duction in predatory feeding might be expected. Interactions between amino acids and ammonia are proposed to signal the relative nutritional (nitrogen) quality of food. Adenosine 5'-triphosphate (ATP) is involved either directly or indirectly with all cellular chemical energy transfers (Atkinson, 1977). Because ATP decays rapidly dur- ing cell death (Hiltz and Dyer, 1970), it is commonly used by ecologists to estimate "living" biomass (Holm-Hansen and Booth, 1966; Sinclair et al., 1979; Ward and Cummins, 1979). Predator sensitivity to ATP alone cannot provide a reliable energy measure; cellular concentrations vary greatly depending on organismal life history stage, activity level, and on long-term seasonal factors (Bamstedt and Skjoldal, 1976; Sjkoldal and Bamstedt, 1976; Giesy and Dickson, 1981). Adenylate energy charge [AEC = (ATP + 0.5 ADP)/(ATP + ADP + AMP)] is a ratio of high energy phosphate bonds expressed as a function of the total phosphoadenylate pool. AEC scales from 0 to 1, with 1 indicating pure ATP (two high energy phosphate bonds) and 0 indicat- ing pure AMP (zero high energy phosphate bonds). In normal healthy organisms (from bacteria to rats), an AEC of 0.8-0.9 is maintained even at the expense of com- promising stored energy reserves. Ratios fall to <0.5 in dead, diseased, or in highly stressed organisms, only after energy reserves have been depleted (Chapman et al, 1971;Balch, 1972; Ridge, 1972; Walker-Simmons and Atkinson, 1977; Ostolovskii, 1978; Buikema, et al 1979; Ivanovici, 1980; Karl, 1980). Because AEC is indicative of the total metabolic energy available to an organism (Atkinson, 1977), its percep- tion by a predator might assist in recognizing prey energetic quality. Predator sensitiv- ity to ATP, as well as to AEC, was therefore expected. 12 R. K. ZIMMER-FAUST FIGURE 1 . Animals representing four of six tested species, showing differences in morphologies and habitats. A. Spiny lobster, Panulirus interruptus; scale bar = 6 cm. B. Yellow crab, Cancer anthonyi (two individuals appear); scale bar = 6 cm. C. Pelagic red crab, Pleuroncodes planipes; scale bar = 2 cm. D. Deep sea mysid, Gnathophausia ingens; scale bar = 2 cm. Not shown are the hermit crab, Pagunis hirsuti- usculus, and the majid crab, Podochela hemiphilla. MATERIALS AND METHODS Species chosen for investigation presented a broad range of dietary preferences and habitats (Fig. 1 and Table I). Following collection, animals were held in 3000 1 aquaria with flowing seawater at ambient temperature (15-1 7°C). A 12:12 D:L cycle (light on:0700 h) was imposed for two weeks. Animals were fed ad libitum on mack- erel (Scomber japonicus) and salmon muscle (Oncorhynchus spp.), then food was withheld for 24 h prior to testing. Chemical solutions were presented to animals placed in plexiglas aquaria (30 X 30 X 15 cm), constructed to allow precise control of stimulus flow characteristics without inhibiting behavior (Zimmer-Faust and Case, 1983). A primary seawater flow (980 ml/min) entered each aquarium from a head- tank maintained under constant hydrostatic pressure, while a secondary flow (120 ml/min) served to inject test substances ( 10 ml/7 s). Dilution associated with injection of a fluorescent dye was 1.02 X 10~3(±0.13 X 10"3 SD) times the original concentra- tion, as measured at the olfactory appendages (antennules) of test animals (see meth- ods of Zimmer-Faust and Case, 1983). Presently reported concentrations are uncor- rected for this dilution. Test procedures were modified slightly for assays with Gnathophausia, since this animal required holding at an ambient temperature of 5°C. Individuals were main- tained in 1 -liter plastic containers and positioned in a constant temperature room (5°C), with seawater changed weekly. G. ingens has been maintained for up to 2.5 years under these conditions (Childress and Price, 1978). Test solutions (2 ml/ 10 s) were added directly to containers and seawater was changed following each trial. OPTIMAL CHEMORECEPTION 13 TABLE I Collection histories of tested animals Order Infraorder Genus species Size (mm) Habitat Depth (m) (x ± SD) Principal dietary components Decapoda Palinura Panulirus interruptus Rock reef Brachyura Cancer anthonyi Sand flat Podochela hemiphilla Pier piling Anomura Pagurus hirsutiusculus Rock intertidal Pleuroncodes planipes Midwater Mysidacea Lophogastrida Gnathophausia ingens Midwater 3-7 67.4 ± 4.0a Animal flesh 25-40 142.2 ±8.4b Animal flesh 2-5 1 3.8 ± 2.0a Algae, animal flesh 0 5.8 ± 0.8" Detritus, animal carrion 100-200 30. 1 ± 2.6a Animal flesh, zooplankton 400-600 32.4 ± 3.4a Animal flesh, zooplankton a Measured as carapace length. b Measured as carapace width. Experimental procedures These were nearly identical to those previously described (Zimmer-Faust el al, 1984b). Animals were tested a maximum of 4 times during a 14-day period, but only once each 72 h. They were put into experimental aquaria 30-60 min prior to testing and observed for 1 min before and for 4 min after chemical presentation. Trial length was chosen to maximize differences in responses to test and control stimuli, as deter- mined in preliminary experiments. All tests were conducted using a double-blind protocol in which the observer was unaware of the composition of solutions intro- duced. Order of chemical presentation was established using a random numbers table with the exception that identical solutions were never repetitively introduced to the same animal. Solutions were prepared immediately before testing using analytical grade reagents and 0.45 /urn filtered seawater, adjusted to pH 7.8. Preliminary experiments evaluated responses to the amino acids, taurine, glycine, and glutamic acid. These were selected because of their proven ability to maximally excite crustacean feeding (McLeese, 1970; Fuzessery and Childress, 1975; Ache, 1 982; Zimmer-Faust et al, 1 984b). Glycine was found most stimulatory to Panulirus, Cancer, Pleuroncodes, and Gnathophausia, while glutamic acid was most stimulatory to Pagurus and Podochela. These compounds were used as stimuli in the follow- ing tests. Responses to amino acids and to ammonia Solutions of 10~2 M amino acid, 10~2Mammonia, 10 2M amino acid plus 10^2 M ammonia, and seawater, were presented to 24 specimens of each species. The in- jected concentration (=10~2 M) produced a dilution-corrected 10~5 M in contact with the animals. This was appropriate given background levels of glycine and glu- tamic acid of 3 X 10~8-3 X 10~7 M, and ammonia of 8 X 10~7-3 X 10"6M, in aquaria seawater with animals present (see Manahan et al, 1983, for chemical analytical methods). Free amino acids are individually maintained at 10"4-10~6 moles/g (dry 14 R. K. ZIMMER-FAUST weight), and ammonia at 10~6 moles/g, in the flesh of invertebrate prey consumed by Crustacea (Clark and Hinkle, 1981; Zurburg and DeZwaan, 1981). Additional tests with Panulirus and Pleuroncodes investigated effects of ammonia concentration on amino acid responses. Solutions of 10~5, 10~4, 10~3, and 1CT2 M ammonia were pre- sented alone, and with 10~2 Mglycine. Twenty animals from each species were tested with solutions. Panulirus was also tested with a mixture of 1 1 amino acids comprising >95% of the total free amino acid pool in abalone muscle, a preferred food (Zimmer- Faust et al, 1984a). This was presented with, and without, 10~2 M ammonia added (n = 20 animals). Responses to adenylates The spiny lobster, Panulirus, was further assayed for responses to ATP, ADP, AMP, and adenosine (ADO) at 10"4, 10~3, and 1(T2 M. Tests were limited to lobsters because of time constraints and because adequate numbers of specimens were lacking for bathypelagic species. Twenty lobsters were tested with each solution. Subsequent trials employed a factorial design to present ATP, ADP, and AMP in all combina- tions. Each substance was maintained at 10"2 M. It was determined whether lobsters are best stimulated by the mixture of highest adenylate energy charge; by the mixture of highest total phosphoadenylate concentration; or by ATP (alone) independently of other phosphoadenylates. The injected concentration of 1(T2 M produced a dilu- tion-corrected 10"5 M in contact with animals. Individual phosphoadenylates are maintained at 1CT4-10~6 moles/g (dry weight) in the flesh of invertebrate prey con- sumed by Crustacea (Hiltz and Dyer, 1970; Ansell, 1977). Experiments with large tanks The above experiments poorly simulated natural conditions. This was especially true for Panulirus and Cancer; these animals were large relative to the size of test- aquaria. For this reason, several assays were repeated with isolated Panulirus and Cancer in large outdoor arenas (1.5 m diameter X 0.6 m water depth), with sandy substrates, shelters, and continuous flowing seawater (8 1/min). Experiments were performed at night under ambient moonlight (intensities = 0.02-0.06 /uW/cm2) with animals placed in arenas 24 h prior. Each animal was tested only once with a solution of 10"5 Mglycine, 10~5M ammonia, 10~5 Mglycine plus 10~5 M ammonia, 10"5M ATP, or seawater. These were pumped to the olfactory region (antennules) at 30 ml/ 1 80 s, through polyethylene tubing (0.32 cm ID) threaded in a hand-held acrylic rod. Sterile cotton was affixed at the tip to diffuse chemicals and to reduce mechanical disturbances caused by stimulus applications. Behavior was observed with an image intensifier and a 75 mm night vision lens (Javelin Electronics Model 221). A positive response was scored when an animal turned, walked towards, then attempted to grasp the cotton tip during a chemical presentation. Addressing the question of mixture interaction Probability of response to a chemical presentation, p(i + j), was a function of chemical [p(i)j and non-chemical [p(j)] factors (e.g., spontaneous activity). Although p(i) and p(j) were assumed independent and binomially distributed, their distribu- tions were not necessarily disjoint; consequently, p(i) could not be directly measured. I used responses to seawater (control) in estimating p(j) and determined p(i), as fol- lows: OPTIMAL CHEMORECEPTION 15 P(i + j) = P(i U j) = p(i) + p(j) - p(i n j) (1) Because / andj were independent, p(i n j) = p(i)p(j), and p(i) was solved algebraically: and, so, Binary mixture interaction between any two chemical agents, a and b, was evaluated by assuming that chemical effects were independent and additive. These assumptions were identical to those made by previous investigators who used "response addition" models to predict responses to mixtures (sensu Hyman and Frank, 1980). In this study the expected probability of response to a mixture, p(i)a+b, was modelled by the equation, P(i)a+b = P(i)a + P(i)b - [p(i)ap(i)b] (3) An independent component index, f(R), was thus derived as: f(R) = p(i)a+b - {p(i)a + p(i)b - [p(i)ap(i)b]} (4) where p(i)a+b, p(i)a and p(i)b were determined empirically. Independence between chemical effects was achieved at f(R) == 0, and p(i)a+b == p(i)a + p(i)b - [p(i)ap(i)b]- Limits were established at -1 < f(R) < 1, given 0 < p(i) < 1. Response synergy oc- curred from 0 to 1 ; suppression from 0 to - 1 ; additivity occurred at zero. This index was similar to that of Hyman and Frank (1980), but with improvements made to account for the binomial nature of present data and to give equal weight to response synergy and suppression. RESULTS Behavioral sequences denning early searching and appetitive phases of feeding are given in Table II, with descriptions taken from previous investigators wherever possible. For this study, it was necessary to derive behavioral sequences only for Podo- chela by watching animals before and after food presentations (= algal turf). I chose cheliped flexion (where applicable) or contact between antennules and mouthparts as criteria for "feeding," since these are exhibited stereotypically by Crustacea (Mackie andShelton, 1972; Snow, 1973; Fuzessery and Childress, 1975; Field, 1977; Pearson et al., 1979; Zimmer-Faust and Case, 1982b; Trott and Robertson, 1984; Zimmer- Faust et al., 1 984a, b) and are of obvious utility either in contacting food, or in main- taining chemoreceptors functioning in orientation to food odor (Hazlett, 1 968, 1971; Reeder and Ache, 1980; Devine and Atema, 1982). For Gnathophausia, I assayed alternating flexion and extension of endopodites on the maxillipeds as an appropriate indicator of feeding (Fuzessery and Childress, 1 975). Locomotory behavior was mon- itored as movement of > 1 carapace length, either forwards or backwards. Responses to amino acids and to ammonia Applications of amino acids, but not of ammonia, caused significant feeding re- sponses relative to seawater (Fig. 2; G-Test for Independence with Williams' correc- tion: G > 10.49, d.f. = 1, P < 0.005, comparisons for all six species). Independent component indices were all negative for mixture interactions (Fig. 3), and amino 16 R. K. ZIMMER-FAUST TABLE II Sequential behavioral patterns in appetitive phases of feeding Animal Behavioral sequences Panulirus (1) Antennule flicking rate is increased. (2) Pereiopods are extended and/or used interruptus* to prod the substratum. (3) The third maxillipeds are flared and set into a lateral motion. (4*) Antennules are vertically deflected and brought into contact with the mouthparts. Cancer anthonyib ( 1 ) Antennule flicking rate is increased. (2) Pereiopods are extended and/or used to prod the substratum. (3) The third maxillipeds are flared and set into a lateral motion. (4) Antennules are vertically deflected and brought into contact with the mouthparts. (5*) Chelipeds are extended and/or used to prod the substratum. Podochela ( 1 ) Antennule flicking rate is increased. (2) The third maxillipeds are flared and hemiphilla* set into a lateral motion. (3*) Chelipeds are extended and/or used to prod the substratum. Pagurus ( 1 ) Antennule flicking rate is increased. (2) The third maxillipeds are flared and hirsutiusculusb set into a lateral motion. (3) Pereiopods are extended and/or used to prod the substratum. (4*) Chelipeds are extended and/or used to prod the substratum, then brought to the mouth. Pleuroncodes ( 1 ) Antennule flicking rate is increased. (2) The third pair of maxillipeds are planipesb rubbed together. (3) The third pair of maxillipeds are outstretched and brought back to the mouth. (4) Pereiopods and chelipeds are extended and/or used to prod the substratum. (5*) Chelipeds are brought back to the mouth. Gnathophausia ( 1 ) Anterior endopodites are moved in a circular motion over the mouth. This ingensb motion appears to be related to filter feeding. (2) Endopodites are partially extended in an unconcerted manner. (3*) Endopodites are extended approximately perpendicular from the body in a concerted motion. (4) The abdomen is arched ventrally, either perpendicular to the body or with the telson almost in contact with the mouth. Sources of behavioral descriptions: aZimmer-Faust et ai, 1984b; bFuzessery and Childress, 1975; cpresent study. Asterisks (*) denote denning acts. acid stimulations of feeding were significantly suppressed by ammonia in Panulirus, Cancer, and Podochela [p(i)a+b < P(Da, and G-Test for Independence with Williams' correction: G > 4.29, d.f. = 1, P < 0.05, all comparisons]. Similar results were ob- tained for Panulirus when the 1 1 -component amino acid mixture was applied. This caused 70% of all tested animals to "feed" and 65% to walk, while only 35% "fed" and 10% walked following addition of ammonia (G-Test for Independence: G > 4.92, d.f. = 1, P < 0.05, both comparisons). The number of times a defining behavioral act was performed in each trial provided a measure of response intensity. For Cancer, Pagurus, and Pleuroncodes, response intensities evoked by amino acids were signifi- cantly depressed by ammonia (Table III; Student's two-tailed Mest for small sample sizes, variances assumed equal: P < 0.08, all comparisons). The intensities exhibited by Panulirus and Podochela were only slightly affected by ammonia (P > 0.20). Littoral species walked forward to amino acids but not to ammonia, while bathy- pelagic species tail-flipped backward (fleeing) to ammonia but not to amino acids (Fig. 2). Mixtures reduced significantly the initiation of walking by Panulirus, and fleeing by Pleuroncodes and Gnathophausia [Fig. 3; p(i)a+b < P(i)a and G-Test: G > 4.14, d.f. = 1, P < 0.05, for all comparisons]. The amino acids and ammonia acted antagonistically though which stimulated while the other suppressed depended on the behavior in question (walking or fleeing, respectively). OPTIMAL CHEMORECEPTION 17 CD C •r-l TJ C o 0. 01 QJ DC 80 60 40 20 0 20 40 80 60 40 20 0 20 40 80 60 40 20 0 20 40 A . Panul irus feeding walk ing B. Cancer feeding SH walk ing Podoche la feeding SW walk ing D. Pagurus feeding E. Pleuroncodes feeding f leeing walk ing F . Gnathophaus la feeding f leeing FIGURE 2. Feeding and locomotory responses to amino acid (AA), ammonia (AM), amino acid + ammonia (MIX), and seawater (SW). Each chemical was presented at 10~2 M. Forward ambulation (walking) was exhibited only by littoral, and tail-flipping (fleeing) only by bathypelagic species. Ammonia was slightly inhibitory at 10 4 M, and significantly so at 10 3 M to glycine-induced feeding by Panulirus and chela probing by Pleuroncodes (Fig. 4; P < 0.05). Ammonia inhibited walking by Panulirus, just as it stimulated fleeing by Pleuroncodes at the same concentrations (=10~3-10~4 M} (Fig. 5). Dose-response data show ammonia consistently inhibitory or repellent over a broad range of concen- trations. Responses to adenylates Adenosine triphosphate was highly stimulatory to Panulirus at 10"2 M (G-Test: G > 6.50, d.f. = 1, P < 0.01, comparisons for feeding and walking), while ADP was only slightly stimulatory (P > 0. 10). Both AMP and ADO were without stimulatory effects (Fig. 6). Stimulation by ATP was not independent of the other phosphoade- nylates, since differences were nearly significant between the responses to ATP, ATP + ADP, ATP + AMP, and ATP + ADP + AMP (Fig. 7A; G-Test for Independence with Williams' correction: G = 7. 16, d.f. = 3, P = 0.06, for feeding). It was proposed 18 R. K. ZIMMER-FAUST 0 Panul irus Cancer Podochela Pagunus Pleuroncodes Gnathophausia -.40 -.20 0 f (R) .20 .40 FIGURE 3. Independent component indices for interactions between amino acids and ammonia. Open bar (D), feeding; closed bar (•), locomotion (walking or fleeing). Negative values indicate suppres- sion. that ADP and AMP might enhance ATP stimulation according to the increase in adenylate concentration. This predicted that proportions of responding animals should be ranked, ATP + ADP + AMP > ATP + ADP == ATP + AMP > ATP. Observed rankings showed no positive association with that predicted (Kendall's one- tailed Tau, with correction for ties: T < -0.40, n = 4, P > 0.80, comparisons for feeding and walking; Accept H0: no association). It was alternatively postulated that ADP and AMP might inhibit ATP stimulation according to the decrease in adenylate energy charge. This predicted that proportions of responding animals should be ranked, ATP > ATP + ADP > ATP + ADP + AMP = ATP + AMP = ADP > ADP + AMP > AMP. Because the observed rankings correlated positively only with those predicted by this last hypothesis, lobster chemosensitivity to adenylate energy charge is indicated (Kendall's one-tailed Tau, with correction for ties: T > 0.827, n = 7, P < 0.02, comparisons for feeding and walking; Reject H0: no association, accept Ha: positive association). Stimulation by ATP was not inhibited significantly by ADP (Fig. 7B; G-Test for Independence with Williams' correction: G < 2.85, d.f. = 1 , P > 0. 10, both compari- sons), though it was by AMP [p(i)a+b < p(i)a, and G-Test: G - 4.65, d.f. = 1, P < 0.05 for feeding]. The stimulation by ADP was independent of AMP. OPTIMAL CHEMORECEPTION 19 TABLE III Response intensities (\ ± SEM) to applied chemical solutions3 Chemical solution Animal Behavior Amino acid Amino acid Seawater + ammonia Ammonia (control) Panulirus interruptus Antennule wiping 8.8 ±2.0 11.4 ±3.5 (n= 17) (n= 10) Cancer anthonyi Cheliped flexion 4.4 ±0.4 2.9 ±0.3 — (n=12) (n = 5) Podochela hemiphilla Cheliped flexion 8.4 ±2.5 7.3 ±3.4 — (n= 12) (n = 5) Pagurus hirsutiusculus Cheliped flexion 34.0 ±5.5 23.0 ±4.4 16.4 ±7.4 13.0 ±1.5 (n= 13) (n=ll) (n = 7) (n = 4) Pleuroncodes planipes Cheliped flexion 6.0 ±0.8 3.9 ±0.5 (n= 15) (n = 10) Gnathophausia ingens Endopodite extension NOT DETERMINED a Data are the numbers of acts performed in 4 min trials. Values appear only where more than one of 24 tested animals responded. Parentheses denote the number of responding animals. Experiments with large tanks Cancer was highly responsive to glycine but less so to the glycine-ammonia mix- ture (Fig. 8A). An independent component index of -0.240 for this interaction was nearly identical to those of -0.142 and -0.367 calculated from tests using smaller aquaria (Fig. 3). Panulirus was also highly responsive to glycine, as well as to ATP (Fig. 8B). The interaction between glycine and ammonia was again inhibitory and produced an index of -0.465, or nearly the same as before (-0.240 and -0.301, see Fig. 3). Experiments with large tanks confirmed the results from tests using smaller aquaria. DISCUSSION Although chemosensory responses in this study were products of chemicals ap- plied individually, significant mixture interaction occurred which caused behavioral suppression. Suppression is commonly reported in psychophysical investigations of odor and taste perception (Moskowitz, 1972;Berglund, 1974;Bartoshuk, 1975, 1977; Carefoot, 1982), though present findings go further in proposing an ecological func- tion. Suppression gives control over the modulation and tuning of feeding according to overall perceived quality of a chemical signal. I propose that interactions between chemical agents function, in part, by tuning responses to relative energy and nutrient properties of odor. Mixture suppression might then benefit a searching organism by reducing its response to lower quality foods. Such behavior would be most advanta- geous under conditions where prey are relatively abundant and food is non-limiting, and where the optimal response is to graze selectively. It might be argued that suppressant interactions between single amino acids and ammonia are relatively meaningless, given that prey flesh is naturally a complex chemical mixture. However, the single stimulants used in this study were those maxi- mizing feeding and locomotory responses. By using binary mixtures in behavioral assays the complexity of ensuing electrophysiological analysis is greatly reduced when 20 R. K. ZIMMER-FAUST en c. 0) 80 70 60 50 40 A . Panul irus Y = 22. 2x (r - -0.97) -0. 108 SW -5 -4 -3 -2 en QJ n a c. Q. ro i — i OJ n u 7 6 B. Pleunoncodes sw -5 -4 -3 -2 Ammonia added M) FIGURE 4. Effects of ammonia on 10~2 A/glycine-induced (A) feeding by Pamdirus, and (B) cheliped flexion (x ± SD) by Pleuroncodes. Responses are adequately modelled as power functions of ammonia concentration. No responses were observed to ammonia presented by itself at 10 5-10~2 M. SW: No am- monia added. considering the neurobiological sites of mixture interaction, i.e., primary receptor or central nervous system. A mixture comprising >95% of the free amino acids in aba- lone muscle was highly stimulatory to Panulirus, but this activity was significantly reduced by the addition of ammonia. Previously, I found a low molecular weight fraction of abalone muscle (<1000 daltons) to be without stimulatory effect to lob- sters (Zimmer-Faust et al, 1984a). Experiments later demonstrated that this results from the presence of chemical suppressants, of which ammonia predominates, and not from a lack of stimulatory agents (Zimmer-Faust et al., 1984b). I interpret results from amino acid and ammonia assays in the following way. Amino acids alone probably trigger attack behavior and intensify within-patch search. Both claw and leg probing were evoked in response to amino acids, and these behaviors are essential to acquiring food in the near-field environment. The amino OPTIMAL CHEMORECEPTION 21 A . Panu 1 irus o sw -5 -4 -3 -2 en c •rt OJ tu 35 30 25 20 15 10 5 0 B . Pleuroncodes \ v sw -5 -4 -3 -2 Ammonia added M) FIGURE 5. Effects of ammonia and glycine on locomotion. A. Walking by Panulirus is stimulated by glycine and suppressed by ammonia. B. Fleeing by Pleuroncodes is stimulated by ammonia and sup- pressed by glycine. Closed circles (•) are responses to ammonia with 10~2 M glycine added; open triangles (A) are responses to ammonia presented by itself. SW: No ammonia added. acid induction of walking appeared to possess a klinokinetic or chemotactic compo- nent, since in large tank experiments, Panulirus and Cancer turned precisely towards a stimulus before moving to grasp the source. Ammonia alone probably causes patch departure. This was demonstrated for bathypelagic species; they tail-flipped away from ammonia. Small tank experiments suggest that littoral species may ignore am- monia, but these experiments did not test animals in an endogenously active state. Large tank experiments confirm that paths of actively moving lobsters (Panulirus) and crabs (Cancer) are unaltered by ammonia (Zimmer-Faust, unpubl. analyses of video data). Endogenous movements would carry these animals to alternative patches in natural habitats. 22 R. K. ZIMMER-FAUST D) C • r-l •a QJ 0) 35 30 25 20 15 10 5 0 ATP AMP -4 -3 -2 Ol c .1-1 ^ I— I ro -4 -3 Concentration added -2 M) FIGURE 6. Dose-responses for (A) feeding, and (B) walking by Panulirus to the adenylates: ATP, ADP, AMP, and adenosine (ADO). Mixtures suppress! ve to the initiation of feeding by Panulirus and Podochela were not inhibitory to the response intensity. The opposite was true for Pagurus since response intensity, but not initiation, was suppressed. Both Panulirus and Podochela typically move long distances while grazing. Their assessments of prey quality and density are likely to occur on the patch-level of resource organization, where a "stop and feed" rule is most advantageous. Pagurus, on the other hand, is a microphagous detritivore that is active yet limited in mobility due to its small size. Given the limita- tions in the patch-selection ability of Pagurus, variable ingestion rates may be the most important animal response affecting feeding energetics. An optimal foraging model predicts that the net time rate of energy gain is maximized when benthic depos- it- and suspension-feeders covary ingestion rates positively with food value (Taghon, 1981). This is because such animals are unable to sample their environments and hence to choose among food items to the same degree that highly mobile predators OPTIMAL CHEMORECEPTION 23 A. B. % Responding 40 20 0 20 40 1.00 0.75 ATP - AMP 0.00 0.00 ADP+AMP ADO AMP sw ATP - ADP ADP - AMP -.40 -.20 .20 .40 f (R) FIGURE 7. A. Feeding and locomotory responses by lobsters to single adenylates and mixtures. Val- ues for adenylate energy charge appear (0-1 scale). B. Independent component indices for binary interac- tions between phosphoadenylates. Each chemical was presented at 10~2 M. Open bar (D), feeding; closed bar (•), walking. can. Benthic marine organisms lacking mobility are dependent on food being brought to them by water currents, by lateral and vertical advections (Taghon et al, 1980), and by microbial regeneration within sediments (Levinton and Lopez 1977; Newell, 1979). By responding to ATP, Panulirus is similar to many of the haematophagous in- sects. Among Diptera, three species of culicine mosquitoes, the blackfly Simulium venustum (Smith and Friend, 1982), the tsetse fly Glossina austeni (Galun and Mar- galit, 1969), and the horsefly Tabanus nigrovittatus (Friend and Stoffolano, 1983) all require ATP or ADP as a phagostimulant. Responses by insects are thought to be adapted to the release of phosphoadenylates, particularly of ATP and ADP, associ- ated with recruitment of blood platelets to open wounds. Apparent sensitivity by Panulirus to adenylate energy charge [AEC -- (ATP + 0.5 ADP)/(ATP + ADP + AMP)] could enable it to assess the total metabolic energy available to a prey (At- kinson, 1977). Chemical excitation of lobsters was ATP > ADP > AMP = adenosine, which is the pathway followed by phosphoadenylates in autolytic degradation of ani- mal flesh (Hiltz and Dyer, 1970). Dephosphorylation of ATP proceeds rapidly during cell death; also, stress related to disease, injury, and starvation elicit physiological defense responses (Selye, 1976) that increase metabolic activity and ATP use (Chap- man et al., 1971; Balch 1972; Walker-Simmons and Atkinson 1977; Ostolovskii 1978;Buikema*tf0/.. 1979). Optimal foraging theory usually predicts that grazers should become more selec- tive at high food densities and less selective at low food densities (Pulliam, 1 974; Pyke et al., 1977; Owen-Smith and Novellie, 1982; Pyke, 1984). A corollary is that food 24 R. K. ZIMMER-FAUST 60 40 20 CD C •M -a c a a. 01 CD DC A . Cancer 18 21 10 10 AA MIX AM SW B. Panulirus 60 40 20 6 10 7 B 6 ATP AA MIX AM SW FIGURE 8. Responses to ATP, glycine ( AA), ammonia (AM), glycine + ammonia (MIX), and seawa- ter (SW), by (A) crabs and (B) lobsters. Animals were placed in large tanks, and each chemical was presented at 10~5 M. The number of animals tested is shown. selectivity should decrease with increasing food deprivation and starvation, if recog- nition or assessment of food availability is modified by hunger levels (Snyderman, 1983). When grazers approach satiation they should behave as if food were abundant and become more selective (Schoener, 1971; Pulliam, 1974; Charnov, 1976). In this study, animals were maintained near satiety and they were without risk of death due to starvation. Therefore, chemical preferences might be expected for odors that sig- nify food of highest energy and nutrient payoffs. Optimal foraging theory predicts that feeding responses should become more broadly tuned as food deprivation in- creases (or as prey densities are lowered), and as energy reserves are spent. That is, we should expect mixture suppression by ammonia and AMP to decrease as food becomes limiting. Mixtures should become more attractive at lower adenylate energy charges and at lower amino acid:ammonia ratios. These predictions have yet to be tested. Caution is urged in developing hypotheses of "optimal" chemoreception. Many factors operate that could confuse experimental design and render test results mean- ingless. For example, optimality only can be tested fairly under the constraint that animals behave given their sensory, motor, and CNS capabilities. It is AMP rather than ATP that acts as a potent chemoattractant to the grass shrimp, Palaemonetes pugio (Carr and Thompson, 1983). Because this animal principally scavenges dead OPTIMAL CHEMORECEPTION 25 animal and detrital materials (Walsh, 1975), the AMP response may be adapted to the dietary mode and need not exclude an optimal chemoreception hypothesis. Differences in adenylate responses by P. interruptus and P. pugio correlate precisely with differences in the electrophysiological response patterns by "ATP-besf ' and "AMP-best" olfactory receptors in the Florida spiny lobster, Pamdirus argus (Derby et ai, 1984; Carr et ai, 1986). The identification of receptor types with response spectra that can account for observed behavior provides a mechanistic argument re- quired to augment theoretical ecological considerations. Ammonia is chemoattractive under some conditions, though this ability depends on the capacity of an organism to assimilate ammonia-nitrogen, or to use ammonia as a cue specific to preferred food. Motile phototropic algae, Chlamydomonas (Sjoblad and Frederiksen, 1981), Dunaliella (Sjoblad et ai, 1978), and Symbiodin- ium (Fitt, 1985), all exhibit positive chemotaxis to ammonia gradients. These organ- isms use ammonia as preferred substrate for inorganic nitrogen during protein syn- thesis. Scavenging animals and detritivores also may be found attracted to ammonia. They typically depend on biodegradatory bacteria as a principal source of carbon and nitrogen (Newell, 1965; Fenchel, 1970; Hargrave, 1970; Lopez et al, 1977; Wetzel, 1977). Previous investigators have argued that because ammonia is excreted by all living marine heterotrophs, it might signal the presence of nearby prey (Borroni et al, 1986). However, such a cue would not allow for prey discrimination, for differen- tiation between predator and prey, or for distinction between living prey and aged carrion. The amino acid to ammonia ratio, on the other hand, provides a continuum on which living and dead prey can both be scaled. Foragers frequently base their preferences on distinguishing prey taxonomic differences. Evidence shows that odor and taste discriminations often depend on unique blendings of chemical substances (Dethier, 1974; Atema et al, 1980). Feeding suppression may therefore arise as a consequence of mixtures that specify inappropri- ate prey images, independently of the perceived energy and nutrient properties. This may explain why investigators sometimes find suppressant interactions among amino acids (McLeese, 1970; Allison and Dorsett, 1977; Johnson and Atema, 1986). Sensory mechanisms evolve to economize resource use. It should be possible to com- bine ecological, evolutionary, and physiological arguments to construct testable hypotheses that predict psychophysical events which influence the chemical percep- tion of food. ACKNOWLEDGMENTS The author expresses his gratitude to J. E. Tyre, T. M. Frank, W. Jorgensen, Drs. W. E. S. Carr, J. J. Childress, D. P. Cook, W. C. Michel, J. F. Case, W. W. Murdoch, B. Robison and S. A. Bernstein. Research was sponsored by a grant from the White- hall Foundation. LITERATURE CITED ACHE, B. W. 1982. Chemoreception and thermoreception. Pp. 369-398 in The Biology of Crustacea, Vol. 3, H. L. Atwood and D. C. Sandeman, eds. Academic Press, New York. ALLISON, P., AND D. A. DORSETT. 1977. Behavioural studies on chemoreception in Balanus hameri. Mar. ' Behav. Phvsiol. 4:205-217. ANSELL, A. D. 1977. The adenosine triphosphate content of some marine bivalve molluscs. / Exp. Mar. Biol. Ecol. 28: 269-283. ATEMA, J., K. HOLLAND, AND W. IKEHARA. 1980. Chemical search image: olfactory responses of yellownn tuna ( Thunmis albacores) to prey odors. /. Chem. Ecol. 6: 457-465. 26 R K. ZIMMER-FAUST ATKINSON, D. E. 1977. Cellular Energy Metabolism and Its Regulation. Academic Press, New York. 293 pp. BALCH, N. 1972. ATP content ofCalanusfinmarchicus. Limnol. Oceanogr. 17: 906-908. BAMSTEDT, V., AND H. R. SKJOLDAL. 1976. Studies on the deepsea pelagic community of Korstjarden, Western Norway. Adenosine phosphates and nucleic acids in Euchaeta norwegia (Copepod) in relation to its life cycle. Sarsia 60: 63-80. BARNARD, C. J., AND C. A. J. BROWN. 1981. Prey size selection and competition in the common shrew, Sorex araneus. Behav. Ecol. Sociobiol. 8: 239-243. BARTOSHUK, L. M. 1975. Taste mixtures: is suppression related to compression? Physiol. Behav. 14: 643- 649. BARTOSHUK, L. M. 1977. Mixtures of substances with similar tastes: a test of a psychophysical model of taste mixture interactions. Sensory Processes 1: 177-186. BELL, W. J., ANDT. R. TOBIN. 1982. Chemo-orientation. Biol. Rev. 57: 219-260. BELL, W. J., C. TORTORICI, R. J. ROGGERO, L. R. KJPP, AND T. R. TOBIN. 1985. Sucrose-stimulated searching behaviour of Drosophila melanogaster in a uniform habitat: modulation by period of deprivation. Anim. Behav. 33: 436-448. BERGLUND, B. 1974. Quantitative and qualitative analysis of industrial odors with human observers. Ann. N. Y.Acad.Sd. 237:35-51. BORRONI, P. F., L. S. HANDRICH, AND J. ATEMA. 1986. The role of narrowly tuned taste cell populations in lobster (Homarus americanus) feeding behavior. Behav. Neurosci. 100: 206-2 12. BOWLUS, R. D., AND G. N. SOMERO. 1979. Solute compatibility with enzyme function and structure: rationales for the selection of osmotic agents and end-products of anaerobic metabolism in ma- rine invertebrates. J. Exp. Zool. 208: 137-153. BUIKJEMA, A. L., S. R. SHERBERGER, G. W. KNAUER, L. A. NEWBERN, J. T. READING, AND J. CAIRNS. 1979. Effects of simulated entrainment on the biology of a freshwater cladoceran. Pp. 809-825 in Energy and Environmental Stress in Aquatic Systems, J. H. Thorp and J. W. Gibbons, eds. DOE Symposium Series 48, CONF-77 1114, Washington, DC. CAMPBELL, J. W. 1973. Nitrogen excretion. Pp. 279-316 in Comparative Animal Physiology, Vol. 1, C. L. Prosser, ed. W. B. Saunders Co., Philadelphia, PA. CAREFOOT, T. H. 1982. Phagostimulatory properties of various chemical compounds to sea hares (Aplysia kurodai and A. dactylomela). Mar. Biol. 68: 207-2 15. CARR, W. E. S. 1978. Chemoreception in the shrimp, Palaemonetes pugio: The role of amino acids and betaine in elicitation of a feeding response by extracts. Comp. Biochem. Physiol. 61 A: 1 27- 131. CARR, W. E. S., AND H. W. THOMPSON. 1983. Adenosine 5'-monophosphate, an internal regulatory agent, is a potent chemoattractant for a marine shrimp. / Comp. Physiol. 153: 47-53. CARR, W. E. S., R. A. GLEESON, B. W. ACHE, AND M. L. MILSTEAD. 1986. Olfactory receptors of the spiny lobster: ATP-sensitive cells with similarities to ?2-type purinoceptors of vertebrates. J. Comp. Physiol. 158: 331-338. CHAPMAN, A. G., L. FALL, AND D. E. ATKINSON. 1971. Adenylate energy charge in Escherichia coli during growth and starvation. J. Bacterial. 108: 1072-1086. CHARNOV, E. L. 1976. Optimal foraging: attack strategy of a manlid.Am. Nat. 110: 141-151. CHILDRESS, J. J., AND M. H. PRICE. 1978. Growth rate of the bathypelagic crustacean Gnathophausia ingens (Mysidacea: Lophogastridae). I. Dimensional growth and population structure. Mar. Biol. 50: 47-62. CLARK, M. E. 1968. A survey of the effect of osmotic dilution on free amino acids of various polychaetes. Biol. Bull. 134: 252-260. CLARK, M. E., AND J. A. M. HINKLE. 1981. Studies on water in barnacle muscle fibres I. The dry weight components of fresh fibres. /. Exp. Biol. 90: 33-4 1 . CLARK, M. E., AND M. ZOUNES. 1977. The effects of selected cell osmolytes on the activity of lactate dehydrogenase from the euryhaline polychaete. Nereis succinea. Biol. Bull. 153: 468-484. CROLL, R. P. 1983. Gastropod chemoreception. Biol. Rev. 58: 293-319. DERBY, C. D., W. E. S. CARR, AND B. W. ACHE. 1984. Purinergic olfactory cells of Crustacea: response characteristics and similarities to internal purinergic cells of vertebrates. J. Comp. Physiol. 155: 341-349. DETHIER, V. G. 1974. The specificity of the labellar chemoreceptors of the blowfly and the response to natural foods. J. Insect Physiol. 20: 1859-1869. DETHIER, V. G. 1976. The Hungry Fly. Harvard University Press, Cambridge, MA. 489 pp. DEVINE, D. V., AND J. ATEMA. 1982. Function of chemoreceptor organs in spatial orientation of the lobster, Homarus americanus: differences and overlap. Biol. Bull. 163: 1 44- 153. ERICHSEN, J. T., J. R. KREBS, AND A. I. HOUSTON. 1980. Optimal foraging and cryptic prey. / Anim. Ecol. 49:271-276. FENCHEL, T. 1 970. Studies on the decomposition of organic detritus derived from the turtle grass Thalassia testudinum. Limnol. Oceanogr. 15: 14-20. OPTIMAL CHEMORECEPTION 27 FIELD, L. H. 1977. A description and experimental analysis of a stereotyped cheliped flexion behaviour in hermit crabs. Behavioural: 147-179. FITT, W. K. 1985. Chemosensory responses of the symbiotic dinoflagellate Symbiodiniwn microadriati- cum (Dinophyceae). /. Phycol. 21: 62-67. FRIEND, W. G., AND J. G. STOFFALANO, JR. 1983. Feeding responses of the horsefly, Tabanus nigrovitta- tus, to phagostimulants. Physiol. Entomol. 8: 377-383. FUZESSERY, Z. M., AND J. J. CHILDRESS. 1975. Comparative chemosensitivity to amino acids and their role in the feeding behavior of bathypelagic and littoral crustaceans. Biol. Bull. 148: 522-538. GALUN, R., AND J. MARGELIT. 1 969. Adenosine nucleotides as feeding stimulants of the tsetse fly Glossina austeni Newst. Nature 222: 583-584. GERARD, J. F., AND R. GILLES. 1972. The free amino-acid pool in Callinectes sapidus (Rathbun) tissues and its role in the osmotic intracellular regulation. J. Exp. Mar. Biol. Ecol. 10: 125-136. GIESY, J. P., AND G. W. DICKSON. 1981. The effect of season and location on phosphoadenylate concentra- tions and adenylate energy charge in two species of freshwater clams. Oecologia 49: 1-7. HARGRAVE, B. T. 1970. The utilization of benthic microflora by Hvalella azteca. J. Anim. Ecol. 39: 427- 437. HARRISON, M. J. S. 1983. Age and sex differences in the diet and feeding strategies of the green monkey, Ceropithec us sabaeus. Anim. Behav. 31: 969-977. HAZLETT, B. A. 1968. Stimuli involved in the feeding behavior of the hermit crab Clibanarius vittatus (Decapoda, Paguridae). Crustaceana 15: 305-31 1. HAZLETT, B. A. 1971. Antennule chemosensitivity in marine decapod Crustacea. J. Anim. Morphol. Phy- siol. 18: 1-10. HILTZ, D. F., AND W. J. DYER. 1970. Principal acid-soluble nucleotides in adductor muscle of scallop (Placopecten magellanicus) and their degradation during postmortem storage on ice. J. Fish. Res. Board Can. 27:83-92. HOLLIBAUGH, J. T. 1979. Metabolic adaptation in natural bacterial populations supplemented with se- lected amino acids. Estuarine Coastal Mar. Sci. 9: 2 1 5-230. HOLM-HANSEN, O., AND C. R. BOOTH. 1966. The measurement of adenosine triphosphate in the ocean and its ecological significance. Limnol. Oceanogr. 11:51 0-5 1 9. HUGHES, R. N. 1979. Optimal diets under the energy maximization premise: the effects of recognition time and learning. Am. Nat. 113: 209-22 1 . HYMAN, A. M., AND M. E. FRANK. 1980. Effects of binary taste stimuli on the neural activity of the hamster chorda tympani. /. Gen. Physiol. 76: 125-142. IVANOVICI, A. M. 1 980. The adenylate energy charge in the estuarine mollusc, Pyruzus ebeninus. I: Labora- tory studies of responses to salinity and temperature. Comp. Biochem. Physiol. 66A: 43-55. JOHNSON, B. R., AND J. ATEMA. 1986. Chemical stimulants for a component of feeding behavior in the common gulf weed shrimp Leander tenuicornis (Say). Biol. Bull. 170: 1-10. KACELNIK, A., A. I. HOUSTON, AND J. R. KREBS. 1981. Optimal foraging and territorial defense in the Great Tit, Parus major. Behav. Ecol. Sociobiol. 8: 35-40. KARL, D. M. 1980. Cellular nucleotide measurements and applications in microbial ecology. Microbiol. Rev. 44: 739-796. KJOSBAKKEN, J., I. STORRO, AND H. LARSEN. 1983. Bacteria decomposing amino acids in bulk-stored Capelin (Mallotus villosus). Can. J. Fish. Aquat. Sci. 40: 2092-2097. KREBS, J. R., A. KACELNIK, AND P. TAYLOR. 1978. Test of optimal sampling by foraging great tits. Nature 275:27-31. LEVINTON, J. S., AND G. R. LOPEZ. 1977. A model of renewable resources and limitation of deposit- feeding benthic populations. Oecologia 31: 177-190. LIMA, S. L. 1985a. Maximizing feeding efficiency and minimizing time exposed to predators: a trade-off in the black-capped chickadee. Oecologia 66: 60-67. LIMA, S. L. 1985b. Downy woodpecker foraging behavior: efficient sampling in simple stochastic environ- ments. Ecologv 65: 1 66- 1 74. LINSTEDT, K. J. 1971. Chemical control of feeding behavior. Comp. Biochem. Physiol. 39A: 553-581. LOPEZ, G. R., J. S. LEVINTON, AND L. B. SLOBODKIN. 1 977. The effect of grazing by the detritivore Orches- tia grillus on Spartina litter and its associated fauna. Oecologia 30: 1 1 1 - 1 27. MACKIE, A. M. 1 973. The chemical basis of food detection in the lobster Homarus gammarus. Mar. Biol. 15. 103-108. MACKIE, A. M., AND R. G. J. SHELTON. 1972. A whole-animal bioassay for the determination of food attractants of the lobster Homarus gammarus. Mar. Biol. 14: 2 1 7-22 1 . MANAHAN, D. T., S. H. WRIGHT, ANDG. C. STEPHENS. 1983. Simultaneous determination of net uptake of 16 amino acids by a marine bivalve. Am. J. Physiol. 244: R832-R838. MCFARLAND, D. J. 1 97 1 . Feedback Mechanisms in Animal Behaviour. Academic Press, London. 279 pp. McLEESE, D. W. 1970. Detection of dissolved substances by the American lobster (Homarus americanus) and olfactory attraction between lobsters. /. Fish. Res. Board Can. 27: 1 37 1 - 1 378. 28 R- K. ZIMMER-FAUST MILINSKJ, M., AND R. HELLER. 1978. Influences of a predator on the optimal foraging behaviour of stickle- backs Gasterosteus aculeatus. Nature 275: 642-644. MOSKOWITZ, H. R. 1972. Perceptual changes in taste mixtures. Percept. Psychophys. 11: 257-262. NELSON, B. B., AND R. R. VANCE. 1979. Diel foraging patterns of the sea urchin Centrostephanus corona- tus as a predator avoidance strategy. Mar. Biol. 51: 25 1-258. NEWELL, R. C. 1 979. Biology of Intertidal Animals. Third edition. Marine Ecological Surveys, Faversham, Kent, UK. 286 pp." NEWELL, R. C. 1965. The role of detritus in the nutrition of two marine deposit feeders, the prosobranch Hydrobia ulvaeand the bivalve Macoma balthica. Proc. Zool. Soc. Lond. 144: 25-45. OGURA, N. 1975. Further studies on decomposition of dissolved organic matter in coastal seawater. Mar. Biol. 31: 101-111. ORIANS, G. H. 1981. Foraging behavior and the evolution of discriminatory abilities. Pp. 389-405 in Foraging Behavior, A. C. Kamil and T. D. Sargent, eds. Garland Press, New York. OSTOLOVSK.II, E. M. 1978. Effect of protein deficiency on nucleotide content and ATPase activity in some rat tissues. Voprosy Pitannia 5: 33-36. OWEN-SMITH, N., AND P. NOVELLIE. 1 982. What should a clever ungulate eat? Am. Nat. 119: 151-178. PEARSON, W. H., P. C. SUGARMAN, D. L. WOODRUFF, AND B. L. OLLA. 1979. Thresholds for detection and feeding behavior in the dungeness crab, Cancer magister. J. Exp. Mar. Biol. Ecol. 39: 65-78. PULLIAM, H. R. 1 974. On the theory of optimal diets. Am. Nat. 108: 59-74. PYK.E, G. H. 1984. Optimal foraging: a critical review. Ann. Rev. Ecol. Syst. 15: 523-575. PYKE, G. H., H. R. PULLIAM, AND E. L. CHARNOV. 1977. Optimal foraging: a selective review of theory and tests. Q. Rev. Biol. 52: 137-154. REEDER, P. B., AND B. W. ACHE. 1980. Chemotaxis in the Florida spiny lobster, Panulirus argus. Anim. Behav. 28:831-839. RIDGE, J. W. 1972. Hypoxia and the energy charge of the cerebral adenylate pool. Biochem. J. 127: 35 1- 355. RITTSCHOF, D. 1980. Chemical attraction of hermit crabs and other attendants to simulated gastropod predation sites. J. Chem. Ecol. 6: 103-1 18. ROBERTSON,!. R., J. A. FUDGE, AND G. K. VERMEER. 1981. Chemical and live feeding stimulants of the sand fiddler crab, Ucapugilator(Bosc). J. Exp. Mar. Biol. Ecol. 53: 47-64. SCHEMBRI, P. J. 1981. Feeding in Ebalia tuberosa (Pennant) (Crustacea: Decapoda: Leucosiidae). J. Exp. Mar. Biol. Ecol. 55: 1-10. SCHMITT, R. J., AND S. J. HOLBROOK. 1985. Patch selection by juvenile black surfperch (Embiotocidae) under variable risk: interactive influence of food quality and structural complexity. /. Exp. Mar. Biol. Ecol. 85: 269-285. SCHOENER, T. W. 1971. Theory of feeding strategies. Ann. Rev. Ecol. Syst. 2: 369-404. SELYE, H. 1976. Stress in Health and Disease. Butterworth Inc, Boston, MA. 334 pp. SHELTON, R. G. J., AND A. M. MACK.IE. 1971. Studies on the chemical preferences of the shore crab, Carcinus maenas (L.). /. Exp. Mar. Biol. Ecol. 7: 41-49. SINCLAIR, M., E. KEIGHAN, AND L. JONES. 1979. ATP as a measure of living phytoplankton carbon in estuaries. J. Fish. Res. Board Can. 36: 180-186. SJOBLAD, R. D., I. CHET, AND R. MITCHELL. 1978. Chemoreception in the green alga Dunaliella tertiolo- lecta. Curr. Top. Microbiol. Immunol. 1: 305-307. SJOBLAD, R. D., AND P. H. FREDERIKSEN. 1981. Chemotactic responses of Chlamydomonas reinhardtii. Mot. Cell. Biol. 1: 1057-1060. SKJOLDAL, H. R., AND U. BAMSTEDT. 1976. Studies of the deepwater pelagic community of Korstjorden, western Norway. Adenosine phosphates and nucleic acids in Meganyctiphanes norvegica (Eu- phausiacea) in relation to the life cycle. Sarsia 61: 1-14. SMITH, J. J. B., AND W. G. FRIEND. 1982. Feeding behavior in response to blood fractions and chemical phagostimulants in the black fly, Simulium venustum. Physio/. Entomol. 7: 219-226. SNOW, P. J. 1973. The antennular activities of the hermit crab, Pagurus alaskiensis (Benedict). J. Exp. Biol. 58: 745-766. SNYDERMAN, M. 1983. Optimal prey selection: the effects of food deprivation. Behav. Anal. Lett. 3: 359- 369. TAGHON, G. L. 1981. Beyond selection: optimal ingestion rate as a function of food value. Am. Nat. 118: 202-214. TAGHON, G. L., A. R. M. NOWELL, AND P. A. JUMARS. 1980. Induction of suspension feeding in spionid polychaetes by high particle fluxes. Science 210: 562-564. TILSON, R. L., AND W. J. HAMILTON III. 1984. Social dominance and feeding pattern of spotted hyena. Anim. Behav. 32: 715-724. TROTT, T. J., AND J. R. ROBERTSON. 1 984. Chemical stimulants of cheliped flexion by the western Atlantic ghost crab Ocypode quadrata (Fabricus). J. Exp. Mar. Biol. Ecol. 78: 237-252. OPTIMAL CHEMORECEPTION 29 WALKER-SIMMONS, M., AND D. E. ATKINSON. 1977. Functional capacities and adenylate energy charge in Escherichia coli under conditions of nutritional stress. J. Bacterial. 130: 676-683. WALSH, B. L. 1975. The role of grass shrimp, Palaemonetes pugio, in a tidal marsh ecosystem. Ecology 56:225-229. WARD, G. M., AND K. W. CUMMINS. 1979. Effects of food quality on growth of a stream detritivore Paratendipes albimanus (Meigen) (Diptera; Chironomidae). Ecology 60: 57-64. WETZEL, R. L. 1977. Carbon resources of a benthic salt marsh invertebrate, Nassarius obsoletus Say (Mol- lusca: Nassariidae). Estuarine Proc. 2: 293-308. YANCEY, P. H., M. E. CLARK, S. C. HAND, B. D. BOWLUS, AND G. N. SOMERO. 1982. Living with water stress: evolution ofosmolyte systems. Science 211: 1214-1222. ZIMMER-FAUST, R. K., AND J. F. CASE. 1 982a. Odors influencing foraging behavior of the California spiny lobster, Panulinis interruptus, and other decapod Crustacea. Mar. Behav. Physiol. 9: 35-58. ZIMMER-FAUST, R. K., AND J. F. CASE. 1982b. Organization of food search in the keip crab, Pugettia producta (Randall). J. Exp. Mar. Biol. Ecol. 57: 237-255. ZIMMER-FAUST, R. K., AND J. F. CASE. 1983. A proposed dual role of odor in foraging by the California spiny lobster, Panulirns interruptus (Randall). Biol. Bull. 164: 341-353. ZIMMER-FAUST, R. K., W. C. MICHEL, J. E. TYRE, AND J. F. CASE. 1984a. Chemical induction of feeding in California spiny lobster, Panulinis interruptus (Randall): responses to molecular weight frac- tions of abalone. J. Chem. Ecol. 10: 957-97 1 . ZIMMER-FAUST, R. K., J. E. TYRE, W. C. MICHEL, AND J. F. CASE. 1984b. Chemical mediation of appeti- tive feeding in a marine decapod crustacean: the importance of suppression and synergism. Biol. Bull. 167: 339-353. ZURBURG, W., AND A. DEZWAAN. 1981. The role of amino acids in anaerobiosis and osmoregulation in bivalves. /. Exp. Zoo/. 215: 315-325. Reference: Biol. Bull. 172: 30-45. (February, 1987) THE BODY PLAN OF THE CYPHONAUTES LARVA OF BRYOZOANS PREVENTS HIGH CLEARANCE RATES: COMPARISON WITH THE PLUTEUS AND A GROWTH MODEL LARRY R. McEDWARD AND RICHARD R. STRATHMANN1 Department of Zoology and Friday Harbor Laboratories, 620 University Road, Friday Harbor, Washington 98250 ABSTRACT The cyphonautes larva of bryozoans is anomalous among ciliary suspension feed- ers with upstream particle capture. This is because the length of the ciliated band that produces the feeding current does not increase disproportionately relative to body length during larval development and growth. Comparisons with previous studies of other upstream collectors (mostly the pluteus of the echinoid Dendraster excentricus or other echinoplutei) demonstrated a striking deficiency in feeding capabilities of the cyphonautes. ( 1 ) In comparison to the pluteus the ciliated band generating the feeding current of the cyphonautes was short both absolutely and relative to larval body size as protein. (2) During larval growth, the length of the ciliated band of the cyphonautes decreased relative to protein in the whole body whereas the length of ciliated band of the pluteus was nearly isometric with body protein. (3) The ratio of metabolic capacity (by electron transport system assay) to protein content was similar for the advanced stage pluteus and cyphonautes. If respiratory rate is proportional to metabolic capacity in these larvae, then the cyphonautes does not compensate for low feeding capacity by reduced respiratory rate. (4) The velocity of the current across the ciliated band and the length of lateral cilia was similar for the cyphonautes and pluteus. Therefore, maximum clearance rates per unit length of ciliated band were not unusually high for the cyphonautes. (5) The cyphonautes was inferior in rate of capture of small (2 /um) spheres relative to 10 ^m spheres in comparison to plutei feeding on the same suspension. (6) The pluteus and cyphonautes were similar in the maximum sizes of spheres ingested. (7) In two experiments the cyphonautes was more selective of spheres flavored by incubation with the alga Dunaliella tertiolecta than was the pluteus. (8) Different patterns of allometric growth of ciliated bands were incorporated in a growth model based on the difference between allometry of gain in organic carbon through feeding and loss of organic carbon through respira- tion. The model predicted that a cyphonautes form required a much greater concen- tration of food than a pluteus form to gain the same organic material in the same time. Thus the cyphonautes proved inferior to the pluteus in quantitative measures of capacity for suspension feeding. The highly conservative differences in larval body plans result in differing capabilities for feeding that are likely to influence larval growth rates and the evolution of life histories. INTRODUCTION The cyphonautes is the only known feeding larval form of bryozoans. The cypho- nautes creates a feeding current with a band of lateral cilia much like the lateral cilia Received 21 August 1986; accepted 25 November 1986. 1 Reprint requests to R. R. Strathmann. 30 LOW FEEDING CAPABILITY OF CYPHONAUTES 3 1 of adult bryozoans, other larval and adult lophophorates and the larvae of echino- derms and enteropneusts (Atkins, 1955; Strathmann, 1971, 1973, 1982; Strathmann etal, 1972; Strathmann and Bonar, 1976; Winston, 1978; McEdward, 1984, 1986b). In other animals with this arrangement of cilia and upstream capture of particles, the ciliary band increases disproportionately as the animal grows, with loops of the band increasing in number or length. In contrast, the two bands of lateral cilia in the cypho- nautes remain simple curved bands; length of the bands of lateral cilia increases in proportion to body length rather than surface area or tissue volume. The cyphonautes has unusual features that might affect quantitative measures of feeding capability. The bands of lateral cilia are within a mantle cavity (Atkins, 1955) and particles are captured by a sieve of laterofrontal cilia upstream from the lateral cilia (Strathmann and McEdward, 1986). Therefore we examined several features of the cyphonautes for comparison with the better known pluteus larvae of echinoids to answer the fol- lowing questions. (1) Does the protein content or metabolic rate of the cyphonautes indicate an unusually low quantity of tissue for a larva of its size and thus account for the short band of lateral cilia? (2) Do velocities of particles crossing the bands of lateral cilia indicate faster feeding currents and hence higher clearance rates per unit length of ciliary band? (3) Are the cyphonautes larvae relatively more efficient than plutei at retaining small particles? (4) Could the cyphonautes larvae capture unusu- ally large particles relative to their body size? The answers to these questions supported predictions based on comparative mor- phology and indicated that the cyphonautes larval form is not well designed for high rates of ingestion from low concentrations of food. Implications for growth were de- rived from a model in which allometric changes in uptake of organic carbon were predicted from allometry of ciliated band length and in which allometric changes in losses of organic carbon were predicted from allometry of respiratory capacity. The results predict that cyphonautes larvae require much higher concentrations of food than do plutei to grow at the same rate. Slow larval growth as a consequence of larval body plan may have influenced the evolution of bryozoan life histories. MATERIALS AND METHODS Eggs and earliest stage cyphonautes larvae were obtained at Friday Harbor, Wash- ington, by letting colonies of Membranipora membranacea spawn into seawater with 0.0001 MEDTA (methods from C. Reed, pers. comm.). Cyphonautes larvae at more advanced stages were collected from the plankton. Larvae were maintained in the laboratory within a few degrees of ambient sea temperature, which ranged from 8°C in early spring to 1 3°C in late summer. The cyphonautes larvae from the plankton could not be identified to species with certainty but were most likely Membranipora membranacea. Those that metamorphosed to ancestrulae could be identified as Membranipora, and the most abundant Membranipora species near the San Juan Islands is presently thought to be M. membranacea. The plutei of the sand dollar Dendr aster excentricus, used in feeding experiments for comparison with cyphonautes larvae, were in most cases reared from eggs and sperm in the laboratory by methods described by Strathmann (1971). Those in the experiments with 5 and 10 /urn spheres were obtained from the plankton. Other larval echinoderms, ophioplutei and bipinnariae, were obtained from the plankton. Dimensions of the larval body were obtained by three-dimensional digitization as described by McEdward (1984, 1985). Body length and apical height were defined as shown in Figure 1 . The measured ciliated band length was doubled to account for the two ciliated ridges. Surface area was taken as twice the projected area of the body outline in lateral view. Relative volume was obtained as the product of maximum thickness (from valve to valve in region X, Fig. 1) and projected area. 32 L. R. McEDWARD AND R. R. STRATHMANN Exponents of the general allometric equation (Y = aXb) were calculated by Bart- lett's nonparametric method (Simpson et al, 1960). The 95% confidence intervals were calculated empirically by nonparametric bootstrapping techniques (Efron, 1979; Diaconis and Efron, 1983). If the isometric exponent fell outside the confidence interval then the difference between the calculated exponent and the isometric expo- nent was statistically significant. Protein content was measured by using the Coomassie brilliant blue G-250 dye binding assay method of Bradford (1976) and Sedmak and Grossberg (1977). Elec- tron transport system (ETS) activity was measured using the methods of Owens and King (1975). The artificial electron acceptor in this reaction is 2-(p-iodophenyl)-3-(p- nitrophenyl)-5-phenyl tetrazolium chloride (INT). Aqueous extracts of larvae were prepared according to the methods of McEdward (1984). Lengths of cilia were measured from cyphonautes larvae taken from a single sam- ple during September 1985. Larvae were sorted live, relaxed with isosmotic MgCl2, and then killed with formalin highly diluted with seawater. Because many cilia lay at an angle to the plane of focus, only the cilium that appeared to be longest was mea- sured for each larva. Because many cilia approached the longest observed in each larva, picking the apparently longest cilium for each larva could only slightly overesti- mate the length of cilia. Poor orientation of many cilia would have resulted in an underestimate of cilia lengths if the mean of a random sample had been taken. Preci- sion for cilium lengths is about ± 1 /mi. Lengths of the locomotory (coronal) cilia were the same on anterior and posterior loops of the corona. Films were taken with a high speed 1 6-mm cinecamera (Redlake Locam) at about 100 or 60 frames/s. A timing light placed a spot on the margin of the film every 0.01 s for more exact calculation of the time between frames. Larvae were held for observation in a cul de sac formed by pieces of broken coverglass between the cover glass and slide. Because the cilia generating the feeding current are within the mantle cavity and currents from the locomotory coronal band dominate the flow outside the mantle cavity, we were not concerned with drag from the walls of the glass cage. The velocity measurements were made on a temperature-controlled stage (Cloney et al., 1970) with the slide cooled to about 12°C. The larvae had been maintained in the laboratory at about 1 0°C. The particles used to test efficiency of capture of small particles were the polysty- rene divinylbenzene spheres that are used in calibrating electronic particle counters. The spheres were suspended in seawater, dissociated by sonication in a water bath, and counted in a hemacytometer before final dilution to test concentrations. Larvae were introduced to a gently stirred suspension with known concentrations of particles of two sizes. After 12 minutes feeding was terminated by preservation in formalin buffered with CaCO3 . The gut contents were counted with differential interference contrast optics. The relative rates of particle capture are an estimate of relative effi- ciency of capture, but because transfer of larvae into the particle suspension disturbs feeding for an undetermined and probably variable time, the absolute rates of capture are not maximum feeding rates. Polystyrene divinylbenzene spheres were also used to test capacity for qualitative selection of particles. For this test spheres of each size were incubated overnight with either seawater or with algal cells (Dunaliella tertiolecta or Thalassiosira weissflogii) that had been centrifuged and resuspended in seawater. Larvae were then introduced to suspensions in which one size of particle had been incubated with algae and the other with seawater only. For a more detailed description of this method, see Rassoul- zadegan ?/ 0/. (1984). Spheres of another polymer (trade name Sephadex) were used to test for maxi- mum sizes ingested. These spheres ranged from ~20 to 60 /mi in diameter. The LOW FEEDING CAPABILITY OF CYPHONAUTES 33 TABLE I Larval dimensions for cyphonautes larvae of four sizes Stage n 1 5 2 16 3 5 4 10 Body length (urn) 155 (±10) 341 (±5) 526 (±15) 649 (±9) Apical height (^m) 138 (±9) 267 (±4) 329 (±7) 394 (±5) Height/length 0.90 (±0.03) 0.79 (±0.01) 0 ,62 (±0.01) 0.61 (±0.00) Thickness (^m) 43 (±2) 68 (±5) 113 (±5) 143 (±4) Band length (^m) 123 (±7) 361 .00 (±13) 577 (±31) 644 (±24) Surface area (104Mm2) 3.6 (±0.4) 14 .4 (±0.4) 28 .8 (±1.2) 43.2 (±1.2) Relative volume 8.0 (±1.1) 49 .9 (±4.2) 163 .9 (±13.9) 308.2 (±10.2) Mean values (±SE). spheres were soaked in seawater with Dunaliella tertiolecta for four hours; the algal suspension was decanted twice, and the spheres were soaked for another hour in seawater. The larvae were introduced to the suspension for 1 9 minutes and then pre- served in buffered formalin. The diameters of spheres in the guts were measured with an ocular micrometer. To check for preservation artifacts, we compared fresh and formalin preserved subsamples of the stock suspension. Preservation and storage for more than a month had no effect on the size frequency distribution of the spheres. For a model of growth based on estimated stage specific gains and losses of organic carbon, the coefficients and exponents of allometric equations were obtained by a model I regression (Sokal and Rohlf, 1981) through the logarithms of mean values for each stage. For isometric relationships the coefficient was the mean of ratios of values. RESULTS Ciliary band length relative to body size and shape The cyphonautes larvae remained roughly triangular throughout growth, but the shape of the triangle changed. Apical height increased slower than body length (Table II). The steady decrease in the height:length ratio (Table I) resulted in longer but squatter larvae in later stages (Figs. 1 , 2). Since thickness changed in proportion to body length, surface area increased more slowly with respect to volume than would be expected if later stages were geometrically similar to early larvae (Table II). The cyphonautes has several bands and groups of cilia. The bands that generate the feeding current through the mantle cavity are the lateral cilia on a pair of ridges separating the upstream (inhalant) and downstream (exhalant) portions of the mantle cavity (Figs. 1, 2). Hereafter "band length" refers to the sum of the lengths of the two ridges because this is the total length of the bands of lateral cilia that generate the feeding current. Band length of the cyphonautes was nearly isometric with body length. Mean band lengths and mean body lengths were 123 ^m and 1 55 ^m for the smallest mea- sured size category and 644 urn and 649 jum for the largest (Table I). Ciliated band length increased in proportion to body length and in proportion to the linear compo- nent of volume (Table II). 34 L. R. McEDWARD AND R. R. STRATHMANN POSTERIOR ANTERIOR e> UJ I 250 CILIATED BAND INCURRENT CHAMBER EXCURRENT CHAMBER CORONA BODY LENGTH FIGURE 1 . Positions of measurements on the cyphonautes. To compare the cyphonautes with the pluteus we used the protein content of the larvae as a measure of larval size. Protein content of the larva increased in proportion to the cube of body length: 79 ng for a cyphonautes 341 yum long (n = 2) and 610 ng for a cyphonautes 649 jitm long (n = 3). Protein content also increased in proportion to larval volume (Table II). Since band length did not increase relative to body length or volume, the band length decreased relative to protein content of the whole cypho- nautes during larval growth (Table II). This was in contrast to the pluteus of D. excen- tricus, in which the ciliated band increased in proportion to larval volume, protein content, and ETS activity of the whole larva during development from the 4 to 8 armed stage (McEdward, 1 984). In this pluteus, the protein content increased relative to ciliated band length only during the 8 armed stage, when the rudiments of post- metamorphic juvenile structures were developing. If larval tissue increases relative to band length, then do metabolic needs increase relative to clearance rate? The ETS assay yielded 1.75 nmols INT h~' larva"1 for the TABLE II Exponents (b) of the general allometric equation (Y = aXb) 95% confidence interval Isometric b Apical height Body length 0.64 0.57-0.71 1.00 Thickness Body length 0.93 0.78-1.10 1.00 Band length Body length 1.03 0.94-1.10 1.00 Protein Body length 2.91 2.38-3.72 3.00 Protein Volume 0.99 0.68-1.34 1.00 Surface area Volume 0.58 0.53-0.64 0.67 Band length Volume 0.37 0.32-0.41 0.33 Band length Volume033 1.11 0.99-1.25 1.00 Band length Protein 0.39 0.21-0.60 1.00 LOW FEEDING CAPABILITY OF CYPHONAUTES 35 P' / • •^ '• FIGURE 2. Change in size and form of a developing cyphonautes (C) of Membranipora and develop- ing pluteus (P) ofDendraster excentricus. All embryos and larvae to the same scale. 36 L. R. McEDWARD AND R. R. STRATHMANN TABLE III Lenglhs of cilia and larvae in urn Length of larva along open edge Lateral cilia Coronal cilia 320 21 38 330 20 40 470 21 48 540 19 57 600 21 57 mean 20 late stage cyphonautes, whose protein content was 610 ng larva ' and body length was 649 ^im. This ratio of metabolic capacity to protein was similar to the ratio for late stage plutei of D. excentricus (protein 600 ng larva"1; ETS 1.44 nmols INT h~' larva^1; larval length 699 yum) (McEdward, 1984). Metabolic capacity is likely to be approximately proportional to metabolic rates. If it is, low metabolic rate does not compensate for the low clearance rate of the cyphonautes. Increase in length of cilia and larvae Coronal cilia increased in length as the larva increased in length, but length of lateral cilia remained constant (Table III). These cilia lengths suggest an increase in volume flow of water per unit length of the locomotory coronal band as the larvae increase in size but little if any increase in volume flow per unit length of the food capturing ciliated ridge as larval size increases. Cilia lengths of food collecting bands of D. excentricus are 20 to 25 ^m long and do not increase during larval development (McEdward, 1984). The cilia producing the feeding currents are therefore similar in length and in constancy of length in both larval forms. Particle velocities at the bands of lateral cilia To compare velocities of feeding currents of the cyphonautes to previous mea- surements on the pluteus, we trapped cyphonautes larvae and filmed the passage of 5 jinn spheres between the two bands of lateral cilia. Under the conditions of observa- tion, the cyphonautes larvae retained and ate few particles and thus had a low clear- ance rate, but they maintained a fast current through the mantle cavity. We did not include any particle tracks in which more than two frames elapsed as the particle crossed the band because particles were sometimes arrested on the upstream side of the band and our object was to use velocities of particles as indicators of water velocity past the bands of lateral cilia. Some particles crossed the band in a single interval between frames (1/60 s) with a mean velocity of 1.7 mm/s (n = 16). If currents were only intermittently fast enough to carry particles across the band in one frame inter- val, this would be a biased estimate of velocity of the feeding current, but estimates with more frames may have included particles momentarily detained by latero-fron- tal cilia or other cilia. For particles crossing the band in two intervals between frames, the mean velocity upstream was 1 .0 mm/s, and the mean velocity downstream was 1.4 mm/s (n = 38). The upstream velocity was significantly lower than the down- stream velocity ( Wilcoxon matched pairs signed ranks test, P < 0.005). Some particles crossed in two intervals, but the midpoint position was obscured; their mean velocity was 1.3 mm/s (n = 22). The grand mean was 1.3 mm/s (n = 76). LOW FEEDING CAPABILITY OF CYPHONAUTES 37 TABLE IV Relative rates ofingestion of spheres of 10 ^m and 2 ^m diameter Number of larvae Total 10 MHI ingested Total 2 Mm ingested Ratio clearance rates (10 Mm/2 Mm) "Cyphonautes 490 to 480 Mm length 6 76 8 57 *Dendraster excentricus 4 armed pluteus 20 567 205 16.6 **Cyphonautes 270 to 490 Mm length 9 45 5 90 **Bipinnaria 1 100 35 29 ***Cyphonautes 260 to 430 Mm length 19 117 37 32 ***Dendraster excentricus 6 armed pluteus 20 757 280 27 * Larvae with 30/Ml of 2 Mm and 5/n\ of 10 Mm spheres for 12 min. ** Larvae with 20/Ml of 2 Mm and 2/V1 of 10 Mm spheres for 15 min. *** Larvae with 30/Ml of 2 Mm and 3/Ml of 10 Mm spheres for 10 min and 2 Mm spheres flavored with Thalassiosira weissflogii. The proportions of 10 Mm and 2 Mm spheres ingested were significantly different from the proportions in suspension (G test, P < 0.00 1 ). Larvae with the same number of asterisks were in the same suspension. Overall, the velocity of currents past the lateral cilia of the cyphonautes larva was similar to velocities of 1 .3 and 1 .7 mm/s past the ciliated band of plutei ofLytechinus pictus at room temperature (Strathmann et al, 1972; Emlet, 1983). The feeding cur- rent of the cyphonautes was as fast or perhaps slightly faster than the current of plutei. Faster currents do not compensate for the short ciliated band of the cyphonautes. Capture efficiency for small particles Are cyphonautes larvae exceptionally efficient in retaining small particles? We compared the relative rates ofingestion of 10 /urn and 2 ^m spheres by cyphonautes larvae and by plutei of Dendraster excentricus and a bipinnaria (Table IV). The cy- phonautes larvae were never much better and usually worse than the echinoderm larvae in the retention of the smaller spheres. Cyphonautes larvae retained the 2 /urn spheres at much lower rates than 10 nm spheres even when the 2 /um spheres were flavored by previous incubation with a suspension of algal cells (Table IV). Capture of smaller food does not compensate for the lower clearance rates of the cyphonautes. Limits on the maximum size of particles captured Could cyphonautes larvae capture and ingest larger particles than those taken by other larvae with upstream capture? To determine the maximum size of particle that could be ingested, we fed echinoplutei, ophioplutei, and cyphonautes larvae a suspen- sion of spheres for which the mean size was near the maximum size that the larvae could ingest (Table V). A sample of 50 spheres from the suspension (first row of Table V) is included for comparison with spheres ingested by the larvae (subsequent rows). Advanced stages of the cyphonautes and plutei with four to eight arms captured parti- cles of about the same mean size. The mean size for the largest 10% of particles in- gested was also similar for the advanced stage cyphonautes and for plutei. Sizes of particles ingested increased with larval size for each type of larva tested, but sizes were comparable for larvae of similar lengths. (The mean midline lengths of 4 armed D. 38 L. R. McEDWARD AND R. R. STRATHMANN TABLE V Maximum size of spheres ingested by larvae feeding in the same suspension for 19 minutes Number larvae Spheres counted Mean ± Mm S.D. i Range Mm Mean of largest 10% ingested Suspension sample 50 45.1 ± 11.8 22.1-70.7 Cyphonautes 330-400 Mm length 6 8 28.2 ± 2.6 24.3-33.5 33.2 Cyphonautes 460-5 80 Mm length 10 30 31.1 ± 5.6 18.8-46.4 42.0 Dendraster excentricus 4 armed pluteus 10 36 30.1 ± 4.4 22.1-38.7 38.2 Dendraster excentricus 8 armed pluteus 10 62 31.6 ± 4.8 21.0-43.1 40.6 Ophiopluteus 6 armed 3 10 28.2 ± 4.1 21.0-34.3 34.3 Ophiopluteus 8 armed 7 39 30.4 ± 6.9 12.1-44.2 42.3 excentricus and 6 armed ophioplutei were 280 /urn and 300 ^m. The mean midline lengths of the 8 armed plutei were 330 yum.) Capture of larger food does not compen- sate for the low clearance rates of the Cyphonautes. Does the width of the mantle cavity ultimately limit the size of particles that can be captured by the Cyphonautes? No particles this large were ingested in this experi- ment. Near the center of the upstream side of the mantle cavity (point X in Fig. 1) the depth of the mantle cavity is 143 /urn in a Cyphonautes larva of 649 /urn body length (Table I), but the ingested particles were less than 50 /urn. Either the maximum width of the incurrent path is much narrower elsewhere, or the width of incurrent openings greatly overestimates maximum sizes ingested. Selection of particles on the basis of quality rather than size Is the Cyphonautes more discriminating than other larvae with upstream retention of particles? Larvae were fed a suspension of 5 and 10 /urn diameter spheres in equal concentration but with spheres of one size flavored by incubation with cells ofDuna- liella tertiolecta. All the larvae tested ingested 10 /urn spheres at higher rates than 5 ^m spheres, but the preferential ingestion of 1 0 /xm spheres was greater when they were flavored than when the 5 ^m spheres were flavored (Table VI). In these experi- ments the effect of flavoring on relative ingestion rates was greater for the cypho- nautes than for the echinoderm larvae. Further experiments would be needed to con- firm the generality of this result because only one algal species was tested and the effect on ingestion rates by the echinoderm larvae was not pronounced. In contrast, in similar experiments by Rassoulzadegan et al. (1984) on larvae of D. excentricus the effect of a flavoring from a culture of algal cells was greater and was sufficient to reverse preference for 5 and 1 0 /urn spheres. Differences in discrimination may exist, but establishing both the existence of differences in selectivity and the consequences of such differences for larvae under natural conditions will require a more extensive set of comparisons. A model relating allometric change inform to growth rates An important consequence of larval feeding is growth, because faster growth re- sults in a shorter planktonic period and presumably therefore less loss of larvae LOW FEEDING CAPABILITY OF CYPHONAUTES 39 TABLE VI Selection of particles on the basis of presence or absence of flavor acquired bv incubation with Dunaliella tertiolecta Number of larvae Flavored bead Total 10 Mm ingested Total 5 Mm ingested Ratio of ingestion rates (10 Mm/ 5 Mm) *Cyphonautes 18 5 Mm 55 36 1.5 **Cyphonautes 24 10 /irn 141 51 2.8 290 to 590 um length *D. excentricus 1 5 jum 78 38 2.1 **D. excentricus 1 10 Mm 90 31 2.9 8 armed pluteus *Bipinnaria 2 5 Mm 31 13 2.4 **Bipinnaria 4 10 Mm 130 37 3.5 ***Cyphonautes -25 5 Mm 198 175 1.1 ****Cyphonautes 26 lOMm 245 58 4.2 500 to 630 um length ***D. excentricus 7 5 Mm 704 419 1.7 ****D. excentricus 7 lOMm 654 288 2.3 8 armed pluteus * and **: Larvae with 5/p.\ of 5 Mm and 5/V1 of 10 Mm spheres for 12.5 min. *** and ****; Larvae with 5/Ml of 5 Mm and S/M! of 10 Mm spheres for 10 min. Larvae with the same number of asterisks were in the same suspension. through transport or predation. Can observations on ciliated band length (here pro- portional to maximum clearance rate) and metabolic capacity (assumed proportional to losses from respiration) be converted to predictions about larval growth at different concentrations of food? We expect allometric changes in feeding capacity to be a major determinant of duration of the planktonic period from first feeding to compe- tence for settling or metamorphosis but the relationship is not simple. Here we ex- plore some of the simplest possible models for the relation between allometry of band length and duration of the larval period. We wish to explore implications of allometric changes in form for duration of the larval period. Our reasoning is similar to that of Putter (1920) and von Bertalanffy (1957) except that we are more willing to sacrifice accurate description of growth by fitted curves (Ricker, 1979) for the sake of analysis of effects of form. For this reason we have used the general form of their model of growth in organic material (w) with time (t) as dw/dt = Bwb -Aw3 (1) but without their restrictive assumptions about the value of the exponents b and a. This permits comparisons based on observed allometric relationships of different lar- val forms. If clearance rates are approximately proportional to lengths of ciliated bands and if respiratory rates are approximately proportional to capacity of the electron trans- port system then rate of growth can be modelled as the difference between two allo- metric equations as in equation ( 1 ) where Bwb is rate of intake of materials and Aw3 is rate of loss of materials. Integrating this expression from the starting organic con- tent to the organic content when the larva is competent to metamorphose gives the minimum planktonic period (T). -r •/Wo dw/(Bwb - Aw3) (2) 40 L. R. McEDWARD AND R. R. STRATHMANN This can be solved by several substitutions. First let u = (A/B)wa"b and rearrange to get T = C fVduVMl-u)] »llQ where C = [(B/A)('-b)/(a-b)]/[B(a - b)] and c = (1 - b)/(a - b) and then let v = u/u, and V = U/UG and rearrange to get f f T/C = Ui [ujV dv]/(l — U]V) — UQ [UQV dv']/(l — u^V) (3) Jo Jo The solution to equation (3) is in Gradshteyn and Ryzhik ( 1 980, p. 286, #3. 1 97.3) in terms of a Beta function and Gauss' hypergeometric function T/C = B(c, l)[u,cF(l,c;c + 1; u,) - UocF(l, c; c + 1 when c> 0. This expression simplifies to 00 \a-bic 'V' r/ A /o\,, a— bin. T « C[(A/B)(w,rb]c Z [(A/B)wrT/(c + n) n=0 - C[(A/B)(w0)a b]c 2 [(A/B)W(rb]7(c + n) (4) n=0 The condition that (1 -- b)/(a - b) > 0 limits the range of values for the allometric exponents for this method of relating larval form to planktonic period. A special case of interest is c = 1 because this occurs when a = 1 , and the allometric exponent for losses through respiration and excretion is likely to be close to 1 for many ciliated larvae. When c = 1, equation (4) simplifies further (Gradshteyn and Ryzhik, 1 980, p. 44, 1 .5 1 3.4) to become T = [1/(A(1 - b))] In [(B - Aw0'-b)/(B - Aw,'-")] (5) When the exponents a and b both equal 1 , the result is simply (6) Equations (4), (5), and (6) apply under restricted conditions. There must be growth throughout the range of sizes. Suspended food must be ingested and assimi- lated at rates proportional to the length of the ciliated band that produces the feeding current. Material must be lost from the body at a rate dependent on size but indepen- dent of concentration of food. A realistic and quantitative prediction would only be possible for the range of concentrations of food that results in nearly constant respira- tion and clearance rates. Both respiration and clearance rates vary with food concen- tration for a variety of suspension feeders. Several types of echinoderm larvae reduce their clearance rates above about 1 000 cells/ml ofDunaliella spp. (Strathmann, 1971; Lucas, 1982). Concentrations of food could also affect respiration (Fenaux et al, 1980) and assimilation by these larvae. The situation is complicated for Dendraster excentricus by the development of longer arms (and therefore a longer ciliated band) at lower concentrations of food (I. Boidron-Metairon, pers. comm.). The following calculations are therefore for gross comparisons between the cyphonautes and plu- teus forms. Application of the model to the cyphonautes of Membranipora and pluteus of Dendraster excentricus requires additional assumptions about measurements and conversion factors. Table VII gives mean values for body protein, the electron trans- port system assay, and length of ciliated band. Table VIII gives coefficients and expo- LOW FEEDING CAPABILITY OF CYPHONAUTES 41 TABLE VII Measures of size and metabolic capacity for the cyphonautes o/Tvlembranipora and pint ens o/TDendraster excentricus Larva Cyphonautes Pluteus 8 arm with Stage New Small Advanced 4 arm 6 arm 8 arm rudiment Protein (^g) 0.013* 0.078 0.61 0.079 0.123 0.169 0.599 ETS(nmolINT/h) — — 1.75 0.093 0.21 0.22 1.44 Ciliated band (mm) 0.12 0.36 0.64 1.45 2.50 4.20 6.90 * Protein value from egg. nents for the allometry of rates of gain and loss in relation to body organic content in terms of organic carbon. To get from Table VII to Table VIII we made the following assumptions and calculations. Organic carbon in the body was assumed to be 0.75 times the protein, a conver- sion taken from an assumed ratio of 17.8% nitrogen in the protein in the whole body and from a ratio of carbon to nitrogen of 4.2 for zooplankton (mean among phyla) (Omori and Ikeda, 1984). This conversion would underestimate organic carbon in later stages if later stage larvae store more fat. The measured protein content per egg was assumed to be equal to the protein in the early stage cyphonautes. Protein proba- bly declines so that organic content of the early cyphonautes was overestimated. As a result growth of the cyphonautes was underestimated, and this biases the compari- son in favor of the cyphonautes. We ignored excretory losses and assumed that the respiratory rate was directly proportional to the ETS assay, which was assumed to be proportional to the protein measurement. For comparison of cyphonautes and pluteus we used the mean of (ETS)/(protein) for the five available points for both types of larvae (Table VII). We assumed 2 nmol INT = 1 nmol O2 (Immers and Runnstrom, 1960; DeVincentiis et al, 1966), a value within the range observed for other zooplankton though toward the lower end of the range (Omori and Ikeda, 1984), and for conversion to organic carbon assumed a respiratory quotient of 1 . For the pluteus we also used the regres- sion through the mean values for the four stages. The coefficients and exponents assumed for isometric and allometric rates of loss are in Table VIII. To convert ciliated band length to a rate of clearance and assimilation, we as- sumed that the volume flow per mm of band was a 20 yum length of cilium times half a tip velocity of 1.7 mm/s times 1 mm and that the assimilation efficiency was 60%, a value within the calculated range for larvae ofMytilus edulis (Sprung, 1984) and above some estimates for gastropod veligers (Pechenik, 1980). The coefficient B in equation (1) was partitioned into the product of B, and the concentration of food (F). For comparison of cyphonautes and pluteus we assumed allometric relationships between ciliated band length and body organic content. For the pluteus we also tried an isometric relationship between band length and body organic content because this relationship is approximately isometric for much of larval development (McEdward, 1984). Coefficients (B,) and exponents (b) for assumed allometric and isometric re- lations are in Table VIII. With these values and equations (4), (5), and (6), we calculated the concentration of food required for growth through the feeding larval stages in three weeks. The cyphonautes required about ten times the concentration of food required by the plu- teus (Table VIII). This result was not changed when rates of loss by the pluteus were 42 L. R. McEDWARD AND R. R. STRATHMANN TABLE VIII Results of a growth model predicting concentration of food required for growth from organic content W, to W2 in ng C in 21 days Larva B, b A a W, W2 F Observed growth Cyphonautes 0.868 0.424 0.364 1 0.00945 0.458 0.27 Pluteus 12.3 0.730 0.364 1 0.0592 0.449 0.025 Pluteus 12.3 0.730 0.584 1.321 0.0592 0.449 0.026 Pluteus 22.1 1 0.364 1 0.0592 0.449 0.021 Same growth Cyphonautes 0.868 0.424 0.364 1 0.045 0.45 0.27 Pluteus 12.3 0.730 0.364 1 0.045 0.45 0.026 The growth equation is dW/dt = B,FWb — AWa with F the concentration of food in /ig C/ml, BI and b estimated from ciliated band length, and A and a estimated from metabolic capacity. See text for the estimates of allometric shifts in loss and gain and other assumptions. assumed to be allometric or when rates of gain by the pluteus were assumed to be isometric, though such changes did produce changes in shapes of growth curves, as expected (see Fletcher, 1 975). The result was also not changed when the same interval of growth was assumed for both Cyphonautes and pluteus. When concentrations of food are limiting, the Cyphonautes should be at a substantial disadvantage relative to the pluteus in rate of growth. Development times in the model were extremely sensitive to concentration of food, however. This sensitivity is highly unrealistic. Rates of development and growth of plutei of Dendraster excentricus do not decrease in proportion to decreased con- centrations of phytoplankton (Paulay et ai, 1985), and rates of growth and develop- ment of larvae of the seastar Acanthaster planci are similar in the food poor waters of the Great Barrier Reef and much higher rations in culture (Olson, 1985). These and many other observations indicate that adjustments in rates of loss or gain of organic material with different concentrations of food are important for these larval forms. Because our model lacks compensatory adjustments to different food supplies, the model may exaggerate the predicted differences in food required for similar growth rates. Ciliated band length and metabolic capacity can be combined for another com- parison of the Cyphonautes and pluteus forms. Volume of water cleared per volume of oxygen consumed is a measure of feeding efficiency from which one can extrapo- late to a concentration of food at which ingestion would equal respiratory losses, a break even point if clearance and respiration rates were constant and assimilation efficiency 100% (Jdrgensen, 1966). With conversion factors similar to those for the growth model except for 100% assimilation efficiency, the late stage Cyphonautes "broke even" at 130 ;ug C/l, the late stage pluteus with a rudiment at 10 ^g C/l, and the earlier 4 to 8 armed stage plutei at about 3 to 4 jug C/l. Because this calculation assumed that respiration does not decline at low ingestion rates, the predicted concen- trations of food required to avert starvation are probably too high. This simpler com- parison also predicts a relative disadvantage for the Cyphonautes form at low concen- trations of food. LOW FEEDING CAPABILITY OF CYPHONAUTES 43 Simple models predict large differences in growth of the cyphonautes and pluteus because of differences in the lengths of ciliated bands that produce the feeding currents. DISCUSSION The cyphonautes larva of bryozoans has an anomalously short band of lateral cilia that generate the feeding current. The current velocity at the bands of lateral cilia, the length of lateral cilia, and the relative rates of capture of large and small food particles are similar to those for plutei, which have a much longer ciliated band at a similar body size. The contrast extends to other suspension feeders with upstream capture. The lateral cilia of the cyphonautes are no longer than those of adult bryozo- ans, phoronids, and brachiopods or those of larval phoronids, brachiopods, enterop- neusts, and echinoderms. The cyphonautes does have one feature as yet unobserved in related suspension feeders with upstream capture. It niters particles with a sieve of laterofrontal cilia (Strathmann and McEdward, 1986). However, it appears that no expansion of capabilities is associated with this sieving mechanism. Because the cy- phonautes has no apparent means of compensating for its low rate of clearing water of suspended particles, it appears to be at a disadvantage relative to other forms in regard to feeding in waters with low concentrations of food. Because the cyphonautes larval form occurs in both ctenostome and cheilostome bryozoans, it is probably the ancestral larval form for the class Gymnolaemata (Niel- sen, 1971), yet most species in this class now lack a feeding cyphonautes stage. One hypothesis for the scarcity of feeding larvae in the group is that the ancestral larval form had a poor capability for feeding and growth under conditions common in coastal waters. Against this hypothesis is the observation that cyphonautes larvae nevertheless have broad geographic and seasonal occurrence and are common in wa- ters low in phytoplankton. Moreover, the eggs of bryozoans with cyphonautes larvae can be quite small. In the San Juan Islands Membranipora membranacea has an egg diameter of about 60 ^m, smaller than that of the local echinoderms (Emlet et al., 1986) and phoronids (Zimmer, 1964; Emig, 1974). Organic content increases with increasing egg size among species with feeding larvae from several phyla (Strathmann and Vedder, 1977), and the protein per egg for Membranipora membranacea (0.013 /ug) is less than that for four species of local echinoids (.024-. 148 /ug) (McEdward, 1 986a). Parental investment per offspring is probably lower for Membranipora mem- branacea with its cyphonautes larva than for co-occuring echinoids and phoronids that use similar larval ciliated bands in larval feeding but have larval forms with ap- parently greater capacity for clearing particles from suspension. With small eggs and low clearance rate, the larval bryozoan is predicted to have an unusually long plank- tonic period with associated risks. Yoshioka's (1982) best fit for a planktonic period from field data off Southern California was about four weeks, not extraordinarily long for a larva. We cannot yet resolve these apparent discrepencies. The combination of poor design for foraging, small initial larval size, and broad geographic occurrence suggests that the cyphonautes larval form has some other re- deeming qualities. Any capabilities in feeding that may have been missed in this study would have to be substantial to compensate for the short ciliated band. Another hy- pothesis is that the cost of slow growth may be low. The only field estimate of mortal- ity rates of the cyphonautes (Yoshioka, 1982) was no lower than those for bivalve veligers and other small larvae (reviewed in Strathmann, 1985), though all of these mortality estimates depended on assumptions about sampling and were subject to large errors. Slow larval growth might be advantageous if there was no advantage to 44 L. R. McEDWARD AND R. R. STRATHMANN more growth before settlement and if there was an advantage to a long precompetent period, but it is difficult to imagine situations in which these two conditions are com- monly met (Strathmann, 1985). Thus possible compensatory benefits from the cy- phonautes form are not yet apparent. A protective shell is not a constraint on length of ciliary band in all larval forms. A shell is combined with an extensible feeding organ in larvae of inarticulate brachio- pods, which have separate lateral cilia and upstream capture of food as does cypho- nautes, and in larvae of gastropod and bivalve molluscs, which differ in that they capture food downstream using a band of compound cilia. Though the cyphonautes form and short ciliated band are conservative, functional or developmental con- straints on this form have not been demonstrated. In conclusion, the cyphonautes larval form has an anomalously poor capacity for suspension feeding. This is predicted to result in slower growth than in other larvae with upstream capture by cilia when food is scarce. A compensating advantage is not yet apparent. ACKNOWLEDGMENTS NSF grants OCE8008310 and OCE8400818, the Kilham Foundation of the Uni- versity of Alberta, and the Friday Harbor Laboratories of the University of Washing- ton supported this study. We are grateful to T. Daniel for integrating growth equations and for other help and advice from B. Hargreaves, C. D. Harvell, R. High- smith, C. G. Reed, M. F. Strathmann, and R. L. Zimmer. LITERATURE CITED ATKINS, D. 1955. The ciliary feeding mechanism of the cyphonautes larvae (Polyzoa Ectoprocta). J. Mar. Biol. Assoc. U.K. 34: 45 1 -466. VON BERTALANFFY, L. 1957. Quantitative laws of metabolism and growth. Q. Rev. Biol. 32: 217-231. BRADFORD, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72: 248-254. CLONEY, R. A., J. SCHAADT, AND J. V. DURDEN. 1970. Thermoelectric cooling stage for the compound microscope. Act a Zoo/. 51: 95-98. DEVINCENTIIS, M., S. HORSTADIUS, AND J. RUNNSTROM. 1966. Studies on controlled and released respi- ration in animal and vegetal halves of the embryo of the sea urchin Paracentrotus lividus. Exp. Cell. Res. 41:535-544. DIACONIS, P., AND B. EFRON. 1983. Computer-intensive methods in statistics. Sci. Am. 248: 1 16-130. EFRON, B. 1979. Bootstrap methods: another look at the jacknife. Ann. Stat. 7: 1-26. EMIG, C. 1974. Observations et discussions sur le developpement embryonnaire des Phoronida. Z. Mor- phol. 77m»77:317-335. EMLET, R. B. 1983. Locomotion, drag, and the rigid skeleton of larval echinoderms. Biol. Bull. 164: 433- 445. EMLET, R. B., L. R. MC£DWARD, AND R. R. STRATHMANN. 1986. Echinoderm larval ecology viewed from the egg. In Echinoderm Studies, Vol. 2, J. Lawrence and M. Jangoux, eds. Balkema, Rotterdam. FENAUX, L., F. RASSOULZADEGAN, AND C. CELLARIO. 1980. Elements du budget energetique des larves de Toursin Paracentrotus lividus (Lam.) (Echinodermata: Echinoidea) Excretion soluble et respi- ration. Pp. 377-379 in Echinoderms: Present and Past, M. Jangoux, ed. Balkema, Rotterdam. FLETCHER, R. L. 1975. A general solution for the complete Richards function. Math. Biosci. 27: 349-360. GRADSHTEYN, I. S., AND I. M. RYZHIK. 1980. Tables of Integrals, Series, and Products. Academic Press, New York. 1160pp. HUNTLEY, M.. AND C. BOYD. 1984. Food-limited growth of marine zooplankton. Am. Nat. 124: 455-478. IMMERS, J., AND J. RUNNSTROM. 1960. Release of respiratory control by 2,4-dinitrophenol in different stages of sea urchin development. Dev. Biol. 2: 90-104. JORGENSEN, C. B. 1966. Biology of Suspension Feeding. Pergamon Press, Oxford. 357 pp. LUCAS, J. S. 1982. Quantitative studies of feeding and nutrition during larval development of the coral reef asteroid Acanthaster planci (L.). J. Exp. Mar. Biol. Ecol. 65: 173-193. LOW FEEDING CAPABILITY OF CYPHONAUTES 45 MCEDWARD, L. R. 1984. Morphometric and metabolic analysis of the growth and form of an echinoplu- teus. J. Exp. Mar. Biol. Ecol. 82: 259-287. MCEDWARD, L. R. 1985. An apparatus for measuring and recording the depth dimension of microscopic organisms. Trans. Am. Microsc. Soc. 104: 194-200. MCEDWARD, L. R. 1986a. Comparative morphometrics of echinoderm larvae. I. Some relationships be- tween egg size and initial larval form in echinoids. J. Exp. Mar. Biol. Ecol. 96: 25 1-265. MCEDWARD, L. R. 1986b. Comparative morphometrics of echinoderm larvae. II. Larval size, shape, growth, and the scaling of feeding and metabolism in plutei. J. Exp. Mar. Biol. Ecol. 96: 267- 286. NIELSEN, C. 197 1 . Entoproct life-cycles and the entoproct/ectoproct relationship. Ophelia 9: 209-34 1 . OLSON, R. R. 1985. In situ culturing of larvae of the crown-of-thorns starfish Acanthaster planci. Mar. Ecol Prog. Ser. 25: 207-210. OMORI, M., AND T. IKEDA. 1984. Methods in Marine Zoolplankton Ecology. Wiley, New York. 332 pp. OWENS, T. G., AND F. D. KJNG. 1975. The measurement of respiratory electron transport-system activity in marine zooplankton. Mar. Biol. 30: 27-36. PAULAY,G., L. BORING, AND R. R. STRATHMANN. 1985. Food limited growth and development of larvae: experiments with natural sea water. J. Exp. Mar. Biol. Ecol. 85: 1-10. PECHENIK, J. A. 1980. Growth and energy balance during the larval lives of three prosobranch gastropods. J. Exp. Mar. Biol. Ecol. 44: 1-28. PUTTER, A. 1920. Wachstumsahnlichkeiten. Pfluegers Arch. Gesamte Phvsiol. Menschen TierelSQ: 298- 340. RASSOULZADEGAN, F., L. FENAUX, AND R. R. STRATHMANN. 1984. Effect of flavor and size on selection of food by suspension feeding plutei. Limnol. Oceanogr. 29: 357-361. RICK.ER, W. E. 1979. Growth rates and models. Fish Physiol. 8: 677-743. SEDMAK, J. J., AND S. E. GROSSBERG, 1977. A rapid, sensitive, and versatile assay for protein using Coo- massie brilliant blue G250. Anal. Biochem. 79: 544-552. SIMPSON, G. G., A. ROE, AND R. C. LEWONTIN. 1960. Quantitative Zoology. 2nd Ed. Harcourt, Brace & Co., New York. 440 pp. SOKAL R. R., AND F. J. ROHLF. 1981. Biometry. W. H. Freeman, San Francisco. 859 pp. SPRUNG M. 1984. Physiological energetics of mussel larvae (Mytilus edulis). IV. Efficiencies. Mar. Ecol. Prog. Ser. 18: 179-186. STRATHMANN, R. R. 1 97 1 . The feeding behavior of planktotrophic echinoderm larvae: mechanisms, regu- lation, and rates of suspension-feeding. / Exp. Mar. Biol. Ecol. 6: 109-160. STRATHMANN, R. R. 1973. Function of lateral cilia in suspension feeding lophophorates (Brachiopoda, Phoronida, Ectoprocta). Mar. Biol. 23: 129-136. STRATHMANN, R. R. 1982. Cinefilms of particle capture by an induced local change of beat of lateral cilia of a bryozoan. J. Exp. Mar. Biol. Ecol. 62: 225-236. STRATHMANN, R. R. 1985 Feeding and non-feeding larval development and life history evolution in ma- rine invertebrates. Ann. Rev. Ecol. Syst. 16: 339-361. STRATHMANN, R. R., AND D. BONAR. 1976. Ciliary feeding of tornaria larvae of Ptychodera flava (Hemi- chordata: Enteropneusta). Mar. Biol. 34: 3 1 7-324. STRATHMANN, R. R., T. L. JAHN, AND J. R. C. FONSECA. 1972. Suspension feeding by marine invertebrate larvae: clearance of particles from suspension by ciliated bands of a rotifer, pluteus, and trocho- phore. Biol. Bull. 142: 505-519. STRATHMANN, R. R., AND L. R. MCEDWARD. 1986. Cyphonautes' ciliary sieve breaks a biological rule of inference. Biol. Bull. 171: 754-760. STRATHMANN, R. R., AND K. VEDDER. 1977. Size and organic content of eggs of echinoderms and other invertebrates as related to developmental strategies and egg eating. Mar. Biol. 39: 305-309. WINSTON, J. E. 1978. Polypide morphology and feeding behavior in marine ectoprocts. Bull. Mar. Sci. 28: 1-31. YOSHIOKA, P. M. 1 982. Role of planktonic and benthic factors in the population dynamics of the bryozoan Membranipora membranacea. Ecology 63: 457-468. ZIMMER, R. L. 1964. Reproductive Biology and Development ofPhoronida. PhD Thesis, Univ. of Washing- ton. Univ. Microfilm, Ann Arbor. 416 pp. Reference: Biol. Bull. 172: 46-60. (February, 1987) RESPONSE OF GREEN HYDRA TO FEEDING AND STARVATION AT FOUR IRRADIANCES G. MULLER-PARKER* AND R. L. PARDY School of Biological Sciences, University of Nebraska, Lincoln, Nebraska 68588-01 18 ABSTRACT The relationship between the productivity of symbiotic algae and growth of the hydra, Hydra viridissima (Florida strain), was investigated in hydra maintained at four irradiances (5, 10, 15, and 30 ME-m~2-s~') and either fed or starved. Although the productivity of fed hydra increased from 0.3 to 1.15 ^g C • mg protein" '-h~' with increase in culture irradiance, there was no significant effect of culture irradiance on population growth or on the protein biomass of individual hydra. The survival of starved hydra was similarly not affected by culture irradiance. Algal-animal biomass parameters changed in response to feeding and culture irradiance. Numbers of symbi- otic algae in fed hydra declined with increased culture irradiance. The protein content of hydra starved for 28 days declined to 10% of initial levels. The relative proportion of algal to animal biomass increased in starved hydra as both algal densities and algal cell volumes were almost twice those of fed hydra. Whereas culture irradiance and feeding alter the ratio of algal to animal biomass, growth of this green hydra is only affected by feeding. INTRODUCTION The green hydra, Hydra viridissima, maintains Chlorella-\ike algae within its di- gestive cells. Algal photosynthesis is an important feature of the symbiosis, as photo- synthetically fixed carbon is translocated from the algae to the animal host (Musca- tine and Lenhoff, 1963; Mews, 1980). Numerous studies show that light sustains green hydra, especially when food is absent or in limited supply. However, the extent to which the productivity of the symbiotic algae influences the growth and survival of green hydra remains unknown. Differences in symbiont productivity can be ob- tained by maintaining green hydra at a series of low irradiances where photosynthesis is light-limited. Using this approach, our objective was to measure the effect of symbi- ont productivity on changes in the biomass of green hydra. Growth rates of fed green and aposymbiotic (=algae-free) hydra maintained in the light showed that light enhances growth of symbiotic hydra, but only when food is limiting (Muscatine and Lenhoff, 1965b). Growth efficiencies derived for green hydra show that the contribution of symbiotic algae is significant, and that this contri- bution increases with a decrease in feeding frequency in the light (Stiven, 1965a). The nutritional importance of symbiont photosynthetic products is also illustrated by starvation experiments in which survival of symbiotic and aposymbiotic hydra was compared. If sufficiently illuminated, symbiotic hydra survive considerably longer periods of starvation than nonsymbiotic species or aposymbiotic animals (Muscatine and Lenhoff, 1965b; Stiven, 1965b; Kelty and Cook, 1976; Rahat and Reich, 1980). Received 25 July 1986; accepted 25 November 1986. * Present address: University of Maryland, C.E.E.S., Chesapeake Biological Laboratory, Solomons, Maryland 20688-0038. 46 RESPONSE OF GREEN HYDRA TO LIGHT 47 Similarly, symbiotic hydra starved in light retain more protein and increase glycogen stores in comparison to aposymbiotic and nonsymbiotic hydra (Cook and Kelty, 1982). Respiratory quotients of symbiotic hydra starved in light indicate that these organisms metabolize mainly carbohydrates, presumed to originate from the symbi- onts; hydra starved in the dark catabolize mainly fat (Pardy and White, 1977). Light is the key factor influencing the ability of symbiotic hydra to survive starvation; in the dark the algae can represent a significant energetic cost to the hydra (Douglas and Smith, 1983). In these light/dark studies, it is assumed that the productivity of symbiotic algae is responsible for the obtained results, although rates of carbon fixa- tion were not provided. To examine the contribution of algal productivity to hydra growth, the productiv- ity and growth of green hydra maintained at different irradiances was measured. Since photosynthesis is dependent on irradiance at low photon flux densities, we hypothe- sized that light may influence the growth and survival of symbiotic hydra cultured under low irradiances. The effect of culture irradiance on the relative proportion of algal to animal biomass of fed green hydra was also measured. Furthermore, although light has been shown to enhance the survival of starved green hydra, it is not known if the total photon flux density available to the algae affects the long-term survival of starving green hydra. We addressed this question by analyzing the survival of starved green hydra acclimated to four different irradiances, and by measuring algal-animal biomass parameters of starved symbiotic hydra. MATERIALS AND METHODS Maintenance of hydra cultures Hydra originating from a single clone of Hydra viridissima (Florida strain) were used for all experiments. Approximately two years prior to these experiments a sym- biosis was established between aposymbiotic hydra and cultured symbionts (NC64A) previously isolated from symbiotic paramecia (Paramecium bursaria). Pardy and Muscatine (1973) showed that this association formed a stable endosymbiosis, and that growth rates of these hydra were nearly identical to those of the original stock of green hydra. These hydra were designated F/NC hydra. F/NC stocks were maintained in M solution (Muscatine and Lenhoff, 1 965a) in glass dishes at 2 1 °C under continu- ous light ranging from 25 to 40 f*E • m 2 • s"1, PAR (Photosynthetically Active Radia- tion; 400-700 nm) provided by banks of fluorescent lamps. Hydra were cultured according to the methods of Loomis and Lenhoff(1956). Animals were fed to reple- tion daily (M-F) with newly hatched Artem ia sp. nauplii, and were rinsed twice daily with fresh culture solution. To determine the effect of irradiance on growth and biomass parameters of F/NC hydra, approximately 1 50 animals from stock culture were placed into each of four glass dishes (28 cm X 1 8 cm X 5 cm) in about 800 ml M solution. Three dishes were covered with one to three layers of fine mesh plastic screen to reduce irradiance; all dishes were covered by a Plexiglas sheet to minimize evaporation of culture medium. Photon flux density in the four dishes was measured prior to addition of hydra and culture medium with a quantum sensor connected to a Li-Cor Model 185 quantum meter. Average photon flux densities in the four experimental dishes were 30, 15, 10, and 5 juE-m~2.s~' (PAR). Experimental dishes were maintained under continuous light at 21°C, and were rotated daily to minimize possible heterogeneity of incident irradiance. 48 G. MULLER-PARKER AND R. L. PARDY Hydra maintained at the four different irradiances were fed four times weekly with Anemia sp. nauplii (M,W,F,Su), and were rinsed daily with M solution. After six months, the animals were subsequently starved for one month to assess the effect of starvation on biomass parameters. Hydra population growth Population growth of fed hydra after three months at the experimental irradiances was measured as follows: six replicate groups of three hydra, each with one bud, were selected from each treatment 24 h after feeding. Groups of animals were placed into petri dishes (3.5 cm X 1 cm) containing 5 ml culture solution and covered with screen material to provide similar irradiances as the parent cultures. Hydra were maintained under the same feeding and cleaning conditions as described for parent cultures. The positions of the four groups of dishes were rotated daily; the six replicate dishes at each irradiance were moved as a unit. Hydranths in each dish were counted daily for 8-10 days. Population growth rate constants and doubling times were calculated according to Loomis ( 1 954). To determine if culture irradiance affected survival of hydra under starvation, fed hydra were placed into individual petri dishes as described above and were starved for more than 80 days. Daily, the number of hydranths in each dish was counted, the culture medium was replaced, and dish positions were rotated. Productivity of hydra To measure photosynthesis of symbiotic algae in hydra acclimated to four irradi- ances, groups of twenty-five hydra (each with one fully developed bud) were incu- bated with NaH14CO3 (0.8 /uCi • ml~') in 5.0 ml M solution supplemented with 5 mM NaHCO3 . Hydra maintained on a regime of 4 feedings per week were last fed 24 hours prior to an experiment. All incubations were carried out at 21°C for one hour at 1 100-1200 h. The productivity of hydra at different irradiances was estimated by placing groups of hydra in beakers covered with various layers of screen. Replicate groups of hydra in beakers covered with foil were also incubated to correct for any dark fixation of NaH14CO3 . Incubation media were sampled at the start of the experi- ments for total 14C activity. At the end of the incubations, hydra were thoroughly rinsed with M solution and then homogenized in distilled water. Aliquots of the ho- mogenate solutions were withdrawn for protein analysis (Lowry et ai, 1951) and determination of algal numbers (see methods under hydra biomass parameters). Or- ganic 14C retained by green hydra was estimated by analyzing three replicate samples (20 yul) of homogenized hydra. Samples were acidified by addition of 100 ^1 of 0.1 TV HC1 under gentle heat to drive offl4CO2. Scintillation fluor (Budget-Solve, Research Products International Corp.; 5 ml) was added, and samples were analyzed in a Pack- ard Tri-Carb liquid scintillation spectrometer (Model 3375). After correction for dark fixation, the amounts of organic I4C retained by hydra were converted to rates of carbon fixation (Vollenweider, 1969). Total CO2 in the incubation medium was calculated from the total alkalinity, measured potentio- metrically (Golterman, 1 969). Productivity data were fit by linear regression to esti- mate photosynthetic efficiencies; all correlation coefficients were greater than 0.95. Hydra biomass parameters The response of symbiotic hydra to different culture irradiances was quantified using replicate samples of 1 00 pooled hydra which were analyzed for total protein, number of algae, and algal chlorophyll. Hydra were sampled 24 h after feeding. Sam- RESPONSE OF GREEN HYDRA TO LIGHT 49 pling of hydra was standardized by selecting individuals with one fully developed bud. As most 30-day starved hydra did not have buds, pooled samples consisted of single individuals. Hydra were homogenized in distilled water with a glass tissue ho- mogenizer. The volume of the homogenate solution was recorded and aliquots re- moved and frozen for later protein analysis by the method of Lowry et al. (1951) using bovine serum albumin as a protein standard. Symbionts were isolated from the remaining homogenate solution and washed in distilled water by repeated centrifuga- tion and resuspension of the algal pellet. The final algal pellet was suspended in a known volume of distilled water and cell concentrations determined by six replicate counts of an aliquot using an A/O Spencer Bright-Line hemacytometer and a Nikon compound microscope at 400X. The total number of symbionts per weight of hydra protein was calculated. The number of symbionts per hydra digestive cell was estimated by placing groups of 25 hydra into maceration fluid (David, 1973) in a small glass vial. After 10 minutes cells were dissociated by gently tapping the vial on a hard surface. Using phase micros- copy at 400X, numbers of algae in 100 randomly selected digestive cells were counted. Chlorophyll content of freshly isolated symbionts was determined by overnight extraction at 4°C of known numbers of algae in absolute methanol. Absorbances of the supernatants of centrifuged methanol extracts were read in a Beckman DB spectrophotometer at 650 nm and 665 nm. The equations of Holden (1976) were used to calculate the weight of chlorophyll per cell. Analysis of variance (ANOVA) was used to determine if culture irradiance had a significant effect on biomass parameters of fed hydra. The Least Significant Differ- ence test was applied at the 5% significance level to data sets for which a significant F-ratio was obtained (Sokal and Rohlf, 1969). Due to the low numbers of starved hydra, it was not possible to obtain replicate samples for the biomass parameters and therefore those data could not be analyzed for statistical significance. Cell volumes of symbionts were determined from measurements of cell diameters obtained using an ocular micrometer at 400X. The mean cell diameter was obtained from measurements of 100 randomly selected algae in hydra homogenized in M solu- tion; algal volumes were calculated assuming spherical cells. RESULTS Population growth of fed hydra and survival of starved hydra maintained at four culture irradiances Population growth rate constants (k) of fed hydra acclimated to four culture irra- diances are given in Table I. There was no significant effect of culture irradiance on k (ANOVA F3,20 = 0.53; P = 0.67), which averaged 0.2238 day"1. Similar results were obtained with a replicate experiment in which k of hydra at the four irradiances averaged 0.22 16 day"1. These results show that, with a 4X/week feeding regime, there is no significant difference in population growth of hydra at culture irradiances rang- ing from 5 to 30 yuE • m~2 • s" ' . To determine if culture irradiance had an effect on the survival of symbiotic hydra under starvation, animals maintained at the four culture irradiances for six months with a 4X/week feeding regime were isolated from parent cultures and starved. Some hydra continued to produce buds, even after 20 days without food. Data sets from five-day intervals were analyzed by the general linear models procedure (SAS, 1982). An analysis of covariance using a split-plot design with unreplicated whole plots 50 G. MULLER-PARKER AND R. L. PARDY TABLE I Population growth rate constants (k) and doubling times (T) ofF/NC hydra fed 4X/week at 4 culture irradiances Culture irradiance k T (ME.m-2.s-') (day-1) (days) 5 0.2084 (±.02 16)a 3.33 10 0.2258 (±.0094) 3.07 15 0.2308 (±.0102) 3.00 30 0.2303 (±.0135) 3.01 n = 6;±S.E. (SAS, 1985) was used to determine if culture irradiance had a significant effect on the rate of decline of numbers of hydra. The preceeding 5-day data set was designated as a covariate. Analysis of covariance indicated that there was no significant difference between the decline in hydra populations maintained at the four culture irradiances (F3 = 2. 1 6; P = 0. 1 02). Days of starvation had a highly significant effect on the decline of numbers of hydra (F, = 15.70; P - 0.0002). Since culture irradiance did not have a significant effect on the survival of F/NC hydra, the data for the four treatments were pooled and the decline in numbers of hydra as a function of days of starvation described by a second order polynomial equation fit to the combined data by the Proc Reg program (SAS, 1985): Number of hydra surviving - 0.002406D - 0.000828D2 + 5.76 (R = 0.802, n == 72) where D = number of days. Figure 1 shows the decline in the number of hydra with days of starvation. There was no real decline in the number of hydra until Day 40 of starvation, and it took 60 days for the population of hydra to decrease to half the initial numbers. Productivity of hydra The productivity of algal symbionts in fed hydra was examined by measuring the amount of 14C retained in hydra exposed to different irradiances. It is not clear if the retention of 14C is a measure of net or gross production by the symbiotic algae since the extent of 14CO2 recycling between animal and algae is unknown. Productivity here refers to the amount of 14C retained by the combined algae and animal tissue after a one hour incubation. To determine if photosynthesis was light-limited at the four culture irradiances, the productivity of hydra from each culture irradiance was measured at several irradiances. The productivity of hydra as a function of irradiance (P-I) is shown in Figure 2. The linear rates indicate that photosynthesis of symbionts was limited by light over the range of irradiances examined. Hydra maintained at a culture irradiance of 10 ^E-m~2-s~' (Fig. 2b) were the most productive hydra at all measured irradiances. The photosynthetic efficiencies of hydra acclimated to the four culture irradiances (the slopes of the P-I curves in Fig. 2) are given in Figure 3. Photosynthetic efficiency declined with an increase in culture irradiance from 10 to 30 ^E • m~2 • s"1, but hydra maintained at 5 /uE-m~2-s~' were only half as efficient as those maintained at 10 m~2 • s"1 at utilizing the available light for photosynthetic carbon production. RESPONSE OF GREEN HYDRA TO LIGHT 51 CO I 3 i i I I I I I I I I I I I I I 10 20 30 40 Days of starvation 50 60 70 80 FIGURE 1 . Decline in number of hydra with days of starvation. Since the survival of hydra at the four irradiances was not significantly different, the data were pooled and a second order polynomial equation fit to the combined data. Dotted lines represent the 95% confidence intervals for the equation (see text). The productivity of green hydra at each acclimated culture irradiance (Pi) showed that hydra maintained at 30 ^E-m~2-s~1 fixed almost four times the amount of car- bon fixed by hydra maintained at 5 ^E-nr^s"1 (Fig. 4b). Hydra population growth rate constants at each culture irradiance were plotted on the same figure (Fig. 4a) as P, to compare the two parameters. Although k was not significantly affected by culture irradiance, the curves in Figure 4 show that population growth rates of hydra (Fig. 4a) and hydra productivity (Fig. 4b) follow, in general, the same pattern. Effect of culture irradiance on biomass parameters of fed hydra The biomass parameters of fed hydra after one month at four culture irradiances are shown in Figure 5. The protein biomass of fed hydra (Fig. 5a) was not significantly affected by culture irradiance (ANOVA F3,,6 = 0.81; P = 0.50). Together with the data in Table I, these results show that culture irradiance does not affect either the growth of individual hydra or that of hydra populations. The population density of symbionts in fed hydra was affected by culture irradi- ance (ANOVA F3,16 = 6.36; P = 0.005). Hydra maintained at the highest irradiance contained significantly fewer symbionts than those maintained at the three lower irradiances (Fig. 5b). Similar results were obtained with numbers of symbionts per digestive cell (Fig. 5d). There was a significant effect of culture irradiance on the num- ber of algae per digestive cell (ANOVA F3,8 = 4.71; P = 0.03), but in this case there 52 G. MULLER-PARKER AND R. L. PARDY 3.0 -• 2.0 c O) "o Q. E u 1.0 >»2.0 o "O o 1.0 (a) (c) 10 (b) 20 30 Irradiance 20 30 FIGURE 2. Productivity of F/NC hydra at different irradiances. Lines were fit to data points by linear regression. Hydra were maintained at culture irradiances of: (a) 5 ^E-m 2-s '; (b) 10 /iE-m~2-s~'; (c) 15 was no significant difference in algal densities of hydra maintained at 1 5 and 30 AiE-nr2.s-'(Fig. 5d). Total chlorophyll per algal cell (Fig. 5c) decreased significantly with increase in culture irradiance (ANOVA F3,15 = 1 5.98; P < 0.00 1 ). Symbionts isolated from hydra 1.10 >, o o> |.08 0) o "S3 I .06 ft o 5 10 15 20 25 30 Culture irradiance (pE m^-s FIGURE 3. Photosynthetic efficiency (a) of F/NC hydra as a function of culture irradiance. Units of aare MgC-OtE-irr2-s~'j i~'-h~'-mg~1 protein. RESPONSE OF GREEN HYDRA TO LIGHT 53 1.0 0.5 (a) - .24 .22 .20 10 15 20 Culture irradiance (juE-i 25 30 FIGURE 4. (a) Population growth rate constants, k, and (b) productivity, PM of F/NC hydra at four culture irradiances. Error bars in (a) are ±S.E.. Units for k same as in Table I and units for P, same as in Figure 2. maintained at 1 5 and 30 ^E • m 2 • s ' contained significantly less chlorophyll than those isolated from hydra maintained at the two lower irradiances (Fig. 5c). Data for chlorophyll a and chlorophyll b gave similar results to those for total chlorophyll, and are not included. There was no significant effect of irradiance on the ratio of chloro- phyll a to chlorophyll b (ANOVA F3,,5 = 0.88; P = 0.47), which averaged 1.19 (±0. 1 7; n = 19, ±S.D.). Cell volumes of algae isolated from fed hydra increased with increase in culture irradiance (Table II). Effect of culture irradiance on biomass parameters of 28-day starved hydra Biomass parameters of hydra starved for 28 days were compared to initial parame- ters obtained from the same populations after two days of starvation (Fig. 6). The protein biomass of starved hydra declined to about 10% of initial levels after 28 days of starvation (Fig. 6a). Algal densities estimated from the number of symbionts per weight of hydra pro- tein in 28-day starved hydra were twice as high as those of 2-day starved hydra (Fig. 6b). Increased numbers of symbionts per digestive cell were also found in 28-day starved hydra (Fig. 6d). Chlorophyll per algal cell also increased after 28 days of starvation (Fig. 6c). There were proportional changes in chlorophyll a and chlorophyll b, with the ratio of chlo- rophyll a:b remaining constant at 1 . 1 2 for algae isolated from all starved hydra. As was found for symbionts from fed hydra, cell volumes increased with increased irradi- ance (Table II). However, cell volumes of symbionts from hydra starved for 30 days were greater than corresponding cell volumes of symbionts from fed hydra (Table II). DISCUSSION The nutritional contribution of symbiont productivity to growth of green hydra was evaluated by measuring changes in biomass of hydra maintained under differ- 54 G. MULLER-PARKER AND R. L. PARDY - (a) o c T 1 10 - a. - 8 - 5 10 15 30 .20 1 o °M5 - I T ii i (c) 0 "S.io - T Q. 1 .05 " 5 10 15 30 Culture irradiance fr Q 3.0 X i T (b) c Q) T 0 1 Q-2.0 2 T en C" 0 - 5 10 15 30 L 20 Q) — T T (d) a> 18 "• | "t/5 |/~ - T "° 14 - 0) 0 f 12 - 5 10 15 30 Culture irradiance FIGURE 5. Final biomass parameters of F/NC hydra maintained at four irradiances for 4 weeks and fed 4X/week with Anemia sp. nauplii. Each bar in 5(a)-5(c) represents the mean of 5 replicate samples of 100 hydra each. Bars in 5(d) are mean values for 3 replicate samples of 100 digestive cells each. Error bars represent standard deviations of the mean. The horizontal brackets connect data sets which were found to be not significantly different at the P > 0.05 level. ent irradiances. The results show that culture irradiances ranging from 5 to 30 /j,E • m • s had no significant effect on both growth rate of fed green hydra and survival of starved green hydra. While the productivity of fed green hydra increased with an increase in culture irradiance, hydra growth was independent of culture irra- diance. These results suggest that the amount of carbon fixed by the symbiotic algae and potentially available to the hydra via translocation or digestion is not transferred TABLE II Cell volumes of symbiotic algae from F/NC hydra either fed or starved for 30 days at four culture irradiances Culture irradiance Algal cell volume 5 10 15 30 From fed hydra 10.4 12.1 15.6 19.0 From starved hydra 18.5 20.6 21.5 27.2 RESPONSE OF GREEN HYDRA TO LIGHT 55 8 'o 6 2 Q. CJ> (a) 5 10 15 30 .40 To 30 £-20 Q. .10 (c; 10 15 30 Culture irradiance §6.0 Tc o> 2 CL o 4.0 o 16 30 (d) 5 10 15 30 Culture irradiance FIGURE 6. The effect of starvation on biomass parameters of F/NC hydra maintained at four irradi- ances. Initial biomass parameters were measured in 2-day starved hydra (D); final biomass parameters in 28-day starved hydra (D). directly into animal biomass of the symbiotic association. Similar results were ob- tained in a previous study on the symbiotic sea anemone Aiptasia pulchella, where changes in sea anemone biomass in fed and starved animals were not related to the productivity of the symbiotic algae at different irradiances (Muller-Parker, 1985). Hydra were maintained at irradiances below those required for light-saturated photosynthesis by the algae. Symbiotic algae were smaller at lower irradiances, and responded to decreases in irradiance with proportional increases in chl a and chl b. The greatest increase in productivity occurred between hydra maintained at 5 and 10 AiE-m~2-s~', and photosynthetic efficiency declined with further increase in culture irradiance. Algal-animal biomass parameters of green hydra changed in response to feeding and culture irradiance. Algal density in fed hydra changed in response to culture irradiance, as both the number of algae per weight of hydra protein and the number of algae per digestive cell were lower for hydra maintained at the highest irradiance (Fig. 5b, d). In spite of lower algal densities, changes in the ratio of plant to animal biomass were less pronounced, since algal cell volumes increased with increase in culture irradiance (Table II). Algal densities of 28-day starved hydra were twice those of fed hydra (Fig. 6b). Since algal cell volumes from these hydra were much greater than those of algae from fed hydra (Table II), the ratio of plant to animal biomass increased greatly with starvation of the host. We have measured the response of one hydra-algal symbiosis to different culture 56 G. MULLER-PARKER AND R. L. PARDY irradiances. Whether other hydra-algal associations show similar responses to changes in culture irradiance is presently unknown. Although the symbiotic algae (NC64A) used in this study were characterized by Mews and Smith (1982) as "low maltose releasers," Mews and Smith (1982) found that there was no difference in the growth rates of hydra (English strain) containing "low maltose releasers" and "high maltose releasers." The amount of carbon fixed by symbionts, the amount of carbon translocated (or digested), and the ratio of algal to animal biomass of a given hydra association may variously affect the response of the symbiotic association to light. Population growth of fed hydra and survival of starved hydra maintained at four culture irradiances Population growth of fed green hydra was independent of photon flux density between 5 and 30 pE • m~2 • s" ' (Table I). These results support those of Epp and Lytle ( 1 969), who showed that light intensity had no effect on the number of buds produced by the Kenilworth strain of green hydra. Two earlier studies have shown a positive correlation between irradiance and growth rates of symbiotic paramecia (Karaka- shian, 1963; Pado, 1965), but food was limiting in both experiments. The effect of irradiance on growth rates of symbiotic paramecia was less pronounced in cultures fed bacteria than in bacteria-free cultures (Pado, 1965). As the productivity of hydra increased with an increase in culture irradiance (Fig. 4b), it appears that growth rates of fed hydra are regulated by factors other than light- dependent carbon fixation. Muscatine and Lenhoff(1965b) showed that the contri- bution of symbiotic algae to hydra growth was dependent on whether or not food was limiting. When hydra were fed to repletion with brine shrimp, growth rates of green and aposymbiotic hydra were identical. When food was limited, growth rates of green hydra exceeded those of aposymbiotic individuals (Muscatine and LenhofF, 1965b). Similar results were obtained with symbiotic and aposymbiotic paramecia (Karaka- shian, 1963). Thus, the similarity in growth rates for the four experimental hydra populations may not be surprising, as equal rations of brine shrimp were provided to all cultures. Although the number of brine shrimp ingested by hydra maintained at the four irradiances was not measured, it is likely that irradiance had no effect on the ability of hydra to capture shrimp since Clayton (1984) found the number of shrimp captured and ingested by green hydra to be the same under both dark and light main- tenance conditions. All other studies on the effect of light on growth of green hydra have dealt with variations in daily photoperiod. No significant difference in budding rate was found for Kenilworth green hydra maintained on either a 12:12 h L:D photoperiod or continuous light (Epp and Lytle, 1969). Douglas and Smith (1984) found that green hydra maintained under continuous light grew more slowly than those maintained on a 12:12 h L:D photoperiod. These studies support the results obtained here, i.e., that there is no positive correlation between total irradiance and growth rates of fed green hydra. Survival of starved green hydra was not affected by culture irradiance. Although symbiotic hydra starved in the light survive longer than those starved in the dark (Stiven, 1965b; Rahat and Reich, 1980), these results indicate that the total photon flux density does not have a significant impact on the length of time symbiotic hydra survive. It would be interesting to determine the effect of very low irradiances (0-5 ME • m~2 • s"1) on the survival and productivity of starved symbiotic hydra. The decline in numbers of hydra with days of starvation (Fig. 1 ) was described by a second order polynomial equation. As some hydra continued to produce buds for RESPONSE OF GREEN HYDRA TO LIGHT 57 up to 20 days of starvation, hydra mortality was somewhat offset by the production of new individuals. Hydra populations did not decline appreciably until the 40th day of starvation. Rahat and Reich (1980) observed buds on green hydra starved for six weeks, and they were able to maintain starved symbiotic hydra in M solution under continuous light for eight months. Unfed hydra cultures were maintained for up to 10 weeks by Douglas and Smith (1984), although survival data were not reported for the starved cultures. Epp and Lytle (1969) showed that there was no significant difference in the number of buds produced by Kenilworth green hydra maintained at two light intensities for 14 days and starved. In their study, bud production stopped after 8 days of starvation in both high and low light cultures. However, the hydra were not pre-acclimated to the high light intensity before the experiment was conducted. Photoperiod had a significant effect on bud production of starved hydra; more buds were produced by Kenilworth green hydra maintained on a 12: 12 L:D cycle than by those maintained under continuous light (Epp and Lytle, 1969). Productivity of green hydra The maximum rate of carbon fixation, 3 yugC-mg hydra protein"1 -h~', was mea- sured at 30 nE • m"2 • s" ' in hydra acclimated to 1 0 yuE • m~2 • s~ ' (Fig. 2b). Normalized to weight of chlorophyll a, this rate is 1 .40 mg C • mg Chi a' ' • h~ ' . As this was a light- limited rate of carbon fixation, it is likely that the productivity of green hydra is at least as high as that of marine symbiotic cnidarians where assimilation numbers range from 1 .0 to 3.9 mg C • mg Chi a~ ' • h" ' (Muscatine, 1 980). The productivity of fed green hydra increased with an increase in culture irradi- ance, with the greatest increase in productivity observed as irradiance increased from 5 to lO/uE-rrT^s"1 (Fig. 4b). As the productivity of hydra increased from 0.3 to 1.15 /ug C-mg protein"1 -h~', and neither population growth nor hydra protein biomass were significantly different at the four irradiances, increased carbon fixation by the symbiotic algae did not result in an increase of hydra biomass. However, the curves in Figure 4 suggest that at low irradiances productivity and growth of green hydra may be related. It is not known if photosynthetically fixed carbon was stored in algal lipid or carbohydrate pools, as these were not measured. Other sinks for photosyn- thetically fixed carbon in symbiotic associations are listed by Muscatine et al. (1984). Algal-derived carbon acquired by the animal may be respired or released as particu- late or dissolved organic carbon. These parameters were not measured in this study. An energy budget constructed for the symbiotic coral Pocillopora eydouxi suggested that about 51% of the photosynthetically fixed energy is used in respiration, 0.9% in growth, and 48% is unaccounted for and probably excreted (Davies, 1984). Figure 2 shows that light-saturated rates of photosynthesis were not obtained over the range of irradiances measured. Phipps and Pardy (1982) found that light-satu- rated photosynthesis in green hydra maintained at 60 ^E-m~2-s~' was obtained at irradiances exceeding 500 ^E-m"2-s"'. In general, photosynthetic efficiency in- creased with decreasing culture irradiance, suggesting that algae in hydra compensate to culture irradiance. Hydra maintained at 10 juE-nT^s"1 were the most efficient at utilizing the available light for carbon fixation (Fig. 3). The low photosynthetic effi- ciency of hydra maintained at 5 /uE • m"2 • s" ' is puzzling, and cannot be explained by any differences in biomass parameters of these hydra (Fig. 5). Effect of culture irradiance on biomass parameters of fed hydra Environmental factors can alter the algal-animal biomass ratio of symbiotic hy- dra. The role of light has previously been addressed by comparing the biomass param- 58 G. MULLER-PARKER AND R. L. PARDY eters of hydra maintained in light and in dark. When hydra are transferred from light to continuous darkness the number of algae per digestive cell goes down (Pardy, 1 974; McAuley, 1981, 1985b; Steele and Smith, 1981; Douglas and Smith, 1984). When hydra are returned to the light, algal densities return to previous levels (Pardy, 1974; Douglas and Smith, 1984). The effect of irradiance on algal densities in hydra has not been previously examined, although at high light intensity symbiotic hydra lose their algae (Pardy, 1976; Steele and Smith, 1981). In this study the density of symbiotic algae in hydra maintained at four irradiances was estimated with two methods; by calculating the number of algae per jug hydra protein and by counting the number of algae in digestive cells. Similar results were obtained with both methods. Hydra maintained at the highest irradiance contained fewer algae than those maintained at the three lower irradiances (Figs. 5b and 5d). Differences in the relative proportion of algal and animal biomass may be less pronounced than suggested by the algal density data, as algal cell volumes in fed hydra maintained at 30 ^E-m~2-s~' were almost twice those of algae from hydra maintained at 5 nE-m~2-s~l (Table II). As cultured symbiotic algae (NC64A) accli- mated to different irradiances exhibited a similar reduction in cell volume with de- crease in culture irradiance (unpubl. data), these data strongly suggest that culture irradiance, and not some influence of the host, determined the cell size of symbionts in situ. Others have found that algae from hydra maintained in the dark were smaller than those from light-maintained hydra (Pardy, 1981; Douglas and Smith, 1984; McAuley, 1985b). Symbiotic algae adapted to low light with an increase in chlorophyll (Fig. 5c) and a decrease in algal cell volume. The cell volume data in Table II suggests that changes in chlorophyll between algae isolated from hydra maintained at the four irradiances would be even greater if chlorophyll were normalized to cell weight or carbon. Effect of irradiance on biomass parameters of starved hydra The protein biomass of F/NC hydra starved for 28 days declined to about 10% of initial levels (Fig. 6a). The final protein biomass of starved hydra was similar in hydra maintained at the four irradiances, but hydra maintained at 30 /iE • irT2 • s~' lost pro- portionately less protein than the other groups. Algal densities in hydra starved for 28 days were twice as high as those of fed hydra (Fig. 6b). The difference between initial and final algal densities was less pronounced in the comparison of numbers of algae per digestive cell (Fig. 6d). A similar increase in the number of symbiotic algae per digestive cell in starved hydra was found by Muscatine and Pool ( 1 979), Douglas and Smith ( 1 984), and McAuley ( 1 985a). Doug- las and Smith (1984) found that this increase in algal density resulted from a decline in protein content and not from an increase in numbers of algae. Algal cells isolated from starved hydra were larger than those isolated from fed hydra (Table II); similar results were obtained by Douglas and Smith (1984) and by McAuley (1985a). In- crease in algal cell size and numbers of algae result in a greater proportion of algal to animal biomass in starved hydra. This study has attempted to define the role of irradiance in the green hydra symbi- osis by comparing growth, productivity, survival to starvation, and algal-animal bio- mass parameters of F/NC hydra maintained at culture irradiances ranging from 5 to 30 /uE-m~2-s~'. The results indicate that irradiance does not have an effect on the biomass of fed or starved hydra, but that algal-animal biomass parameters are influ- enced by irradiance and feeding regime. RESPONSE OF GREEN HYDRA TO LIGHT 59 ACKNOWLEDGMENTS The skilled assistance of Jayne Fleck and Brent Kinsey in maintaining hydra cul- tures is gratefully acknowledged. C. A. Lemen and C. Langefeld provided advice and the statistical analysis of the starvation data presented in Figure 1. Research funded by grants from the National Science Foundation (PCM 83 14590) and the University Research Council to R. L. Pardy. LITERATURE CITED CLAYTON, W. S., JR. 1984. Effects of altered light and feeding regimes on the zooplankton feeding capabil- ity of Hydra viridis. Int. Rev. Hydrobiol. 69: 633-642. COOK, C. B., AND M. O. KELTY. 1982. Glycogen, protein, and lipid content of green, aposymbiotic, and nonsymbiotic hydra during starvation. J. Exp. Zoo/. 222: 1-9. DAVID, C. N. 1973. A quantitative method for maceration of Hvdra tissue. Wilhem Roux Archiv Entwickhmgsrnech. Org. 171:259-268. DA VIES, P. S. 1984. The role of zooxanthellae in the nutritional energy requirements of Pocillopora ev- douxi. Coral Reefs 2: 1 8 1 - 1 86. DOUGLAS, A. E., AND D. C. SMITH. 1983. The cost of symbionts to the host in the green hydra symbiosis. Pp. 631-648 in Endocytobiology, Endosymbiosis and Cell biology', W. Schwemmler and H. E. A. Schenk, eds. W. de Gruyter and Co., Berlin. DOUGLAS, A. E., AND D. C. SMITH. 1984. The green hydra symbiosis. VIII. Mechanisms in symbiont regulation. Proc. R. Soc. Lond. B 221: 291-319. EPP, L. G., AND C. F. LYTLE. 1969. The influence of light on asexual reproduction in green and aposymbi- otic hydra. Biol. Bull. 137: 79-94. GOLTERMAN, H. L., ed. 1969. Methods for Chemical Analysis of Fresh Waters (International Biological Programme Handbook No. 8). Blackwell Sci. Publications, Oxford. 172 pp. HOLDEN, M. 1976. Chlorophylls. Pp. 1-37 in Chemistry and Biochemistry of Plant Pigments, second edition, vol. 2, T. W. Goodwin, ed. Academic Press, New York. KARAKASHIAN, S. J. 1963. Growth of Paramecium bursaria as influenced by the presence of algal symbi- onts. Physiol. Zoo/. 36: 52-68. KELTY, M. O., AND C. B. COOK. 1976. Survival during starvation of symbiotic, aposymbiotic, and non- symbiotic hydra. Pp. 409-414 in Coelenterate Ecology and Behavior, G. O. Mackie, ed. Plenum Press, New York. LOOMIS, W. F. 1954. Environmental factors controlling growth in hydra. /. Exp. Zoo/. 126: 223-234. LOOMIS, W. F., AND H. M. LENHOFF. 1956. Growth and sexual differentiation of hydra in mass culture. / Exp. Zoo/. 132: 555-573. LOWRY, O. H., N. J. ROSEBROUGH, A. L. FARR, AND R. J. RANDALL. 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193: 265-275. McAuLEY, P. J. 1 98 1 . Ejection of algae in the green hydra symbiosis. J. Exp. Zoo/. 217: 23-3 1 . McAULEY, P. J. 1985a. The cell cycle of symbiotic Chlorella. I. The relationship between host feeding and algal cell growth and division. / Cell Sci. 77: 225-239. McAULEY, P. J. 1985b. The cell cycle of symbiotic Chlorella. II. The effect of continuous darkness. J. Cell Sci. 77:241-253. MEWS, L. K. 1980. The green hydra symbiosis. III. The biotrophic transport of carbohydrate from alga to animal. Proc. R. Soc. Lond B 209: 377-401. MEWS, L. K., AND D. C. SMITH 1982. The green hydra symbiosis. VI. What is the role of maltose transfer from alga to animal? Proc. R. Soc. Lond. B 216: 397-4 1 3. MULLER-PARKER, G. 1985. Effect of feeding regime and irradiance on the photophysiology of the symbi- otic sea anemone Aiptasia pulchella. Mar. Biol. 90: 65-74. MUSCATINE, L. 1980. Productivity of zooxanthellae. Pp. 381-402 in Primary Productivity in the Sea, P. G. Falkowski, ed. Plenum Publ. Co., New York. MUSCATINE, L., AND H. M. LENHOFF. 1963. Symbiosis: on the role of algae symbiotic with hydra. Science 142:956-958. MUSCATINE, L., AND H. M. LENHOFF. 1 965a. Symbiosis of hydra and algae. I. Effects of some environmen- tal cations on growth of symbiotic and aposymbiotic hydra. Biol. Bull. 128: 415-424. MUSCATINE, L., AND H. M. LENHOFF. 1965b. Symbiosis of hydra and algae. II. Effects of limited food and starvation on growth of symbiotic and aposymbiotic hydra. Biol. Bull. 129: 3 16-328. MUSCATINE, L., AND R. R. POOL. 1979. Regulation of numbers of intracellular algae. Proc. R. Soc. Lond. 5204: 131-139. MUSCATINE, L., P. G. FALKOWSKI, J. W. PORTER, AND Z. DUBINSKY. 1984. Fate of photosynthetic fixed 60 G. MULLER-PARKER AND R. L. PARDY carbon in light- and shade-adapted colonies of the symbiotic coral Stylophora pistillata. Proc. R. Soc. Lond. B 222: 18 1-202. PADO, R. 1965. Mutual relation of protozoans and symbiotic algae in Paramecium bursaria I. The influ- ence of light on the growth of symbionts. Folia Biol. 13: 1 73- 1 82. PARDY, R. L. 1974. Some factors affecting the growth and distribution of the algal endosymbionts ofHvdra viridis. Biol. Bull. 147: 105-1 18. PARDY, R. L. 1976. The production of aposymbiotic hydra by the photodestruction of green hydra zoo- chlorellae. Biol. Bull. 151: 225-235. PARDY, R. L. 198 1 . Cell size distribution of green symbionts from Hydra viridis. Cytobios 32: 7 1-77. PARDY, R. L., AND L. MUSCATINE. 1973. Recognition of symbiotic algae by Hydra viridis. A quantitative study of the uptake of living algae by aposymbiotic H. viridis. Biol. Bull. 145: 565-579. PARDY, R. L., AND B. N. WHITE. 1977. Metabolic relationships between green hydra and its symbiotic algae. Biol. Bull. 153: 228-236. PHIPPS, D. W., JR., AND R. L. PARDY. 1982. Host enhancement of symbiont photosynthesis in the hydra- algae symbiosis. Biol. Bull. 162: 83-94. RAHAT, M., AND V. REICH. 1980. Survival, budding and osmotrophic nutrition in Hydra viridis. Pp. 465- 469 in Developmental and Cellular Biology of Coe/enterates, P. Tardent and R. Tardent, eds. Elsevier/North-Holland Biomedical Press, Amsterdam. SAS Institute Inc. 1982. SAS User's Guide: Basics, 1982 Edition. SAS Institute Inc., Gary, NC. p. 932. SAS Institute Inc. 1985. SAS User's Guide: Statistics, Version 5 Edition. SAS Institute Inc., Gary, NC. SOKAL, R. R., AND F. J. ROHLF. 1969. Biometry. W. H. Freeman and Co., San Francisco. 776 pp. STEELE, R. D., AND D. C. SMITH. 1981. Factors affecting the reduction of the algal symbiont population in green hydra. J. Zoo/. Lond. 193: 201-214. STIVEN, A. E. 1965a. The relationship between size, budding rate, and growth efficiency in three species of hydra. Res. Popul. Ecol. (Kyoto Univ.) 7: 1-15. STIVEN, A. E. 1965b. The association of symbiotic algae with the resistance of Chloroh ydra viridissima (Pallas) to Hydramoeba hydroxena (Entz). J. Invert. Pathol. 7: 356-367. VOLLENWEIDER, R. A. 1969. Methods for measuring production rates. Pp. 41-127 in A Manual on Meth- ods for Measuring Primary Production in Aquatic Environments (International Biological Pro- gramme Handbook No. 12), R. A. Vollenweider, ed. Blackwell Sci. Publications, Oxford. Reference: Biol. Bull. 172: 61-68. (February, 1987) FLOW-ASSISTED MANTLE CAVITY REFILLING IN JETTING SQUID STEVEN VOGEL Department of Zoology, Duke University, Durham, North Carolina 27706 ABSTRACT During rapid, rearward, repetitive jetting, the mantle of a squid is expanded later- ally by hydrodynamic pressure differences between the region of greatest girth and the refilling apertures beneath its downstream margin; this flow-induced expansion assists refilling between jet pulses. Estimates derived from measurements of pressure distributions on a model in wind tunnels indicate that about 50% of the pressure available for refilling is flow-induced at a speed of 3 m -s"1 while 90% is flow-induced at9m.s~'. INTRODUCTION Rapid rearward (dorsally directed) locomotion in squid is accomplished by ejec- tion of water from the mantle cavity through a ventrally directed funnel. The cavity then refills through a bilateral pair of one-way valves at the anterior end of the mantle, and the ejection process may be repeated. While clearly less efficient than the swim- ming offish (O'Dor, 1 982), this jet propulsion system is capable of propelling animals at impressively high speeds at least in brief episodes. Refilling the mantle cavity demands that the mantle expand its circumference, no ordinary activity for a muscular organ. Two mechanisms have been recognized as contributing to the expansion. Ward and Wainwright (1972) drew attention to a set of very short radial muscle fibers within the mantle, whose contraction could thin the mantle and extend the circumferential muscles. Gosline and Shadwick (1983) showed that, in addition, elastic storage of energy from the expulsion phase played a significant role early in the process of re-expansion. An additional agency for refilling may be involved, a consequence of the pressure distribution predicted by Bernoulli's principle. If a spherical or ellipsoidal body moves through a fluid, a characteristic distribution of pressures develops along its surface, with the most subambient pressures occurring near or ahead of the region of maximum cross-sectional area normal to flow. If the body is reasonably streamlined (as is a squid), flow follows its surface, moving downstream without substantial sepa- ration or reversal, and pressures rise again behind the widest portion. For a jetting squid, maximum girth is adjacent to mantle and mantle cavity; the openings between mantle margin and neck through which refilling occurs are well downstream. Thus pressure inside the mantle cavity should be higher than that outside, creating a net outward force that might augment the actions of radial muscles and elastic connective tissue. [The possibility was mentioned by Ward (1972) but was regarded as inconse- quential inasmuch as squid could refill adequately in still water.] Such flow-induced pressures might reduce the time needed to refill or increase the volume of water in the mantle cavity prior to ejection. Indeed, rapidly moving squid appear to pulse more frequently than those constrained to remain at rest with respect Received 8 August 1986; accepted 25 November 1986. 61 62 S. VOGEL to the local water (Gosline, pers. comm.). Functional roles of analogous flow-induced pressures have been previously investigated in fish (DuBois et al, 1974), in an alga (Vogel and Loudon, 1985), and in scallops (Vogel, 1985). Unfortunately, squid have never been persuaded to perform high speed escape responses with repetitive jetting in any form of confinement. Furthermore, their soft- bodied character, especially when dead, precludes direct measurement of pressures as was done with fish and scallops. Instead, the present study makes recourse to a somewhat abstract physical model — a radially symmetrical body of revolution with a lengthwise row of surface apertures for measurements of the distribution of pressures. Pressures determined on the model are compared to those generated by previously described refilling mechanisms to estimate the contribution of flow-induced pressure differences to the overall process. MATERIAL AND METHODS The model squid (Fig. 1 , top) was based on measurements abstracted and aver- aged from a variety of published and unpublished photographs giving lateral views ofLoligo, mostly L. pealei. Few pictures showed animals in motion, and all reflected full or nearly full mantle cavities. Fins were omitted in the model since they are appar- ently folded flat against the mantle during rapid jetting (Bradbury and Aldrich, 1969; William Kier, pers. comm.). The extent to which the ventral margin of the mantle hooded the head and the smoothness of fairing of mantle and head varied among the photographs; the model assumes a small head, poor fairing, and little hooding — in short, a worst case from the point of view of smooth flow. [It should be noted that Bradbury and Aldrich (1969) showed an unusually conspicuous gap, nearly that of the present model, between mantle margin and head during refilling in Illex illecebro- sus.] At the same time, the model permitted "improvement''1 through the addition of modelling clay. The model, of aluminum, was 293 mm in length and 50 mm in maximum diame- ter. Squid are nearly circular in cross-section; the use of a radially symmetrical model is thus only mildly abnormal. Radial symmetry permitted simple fabrication on a lathe and radically reduced the number of measurements needed to map pressure distributions. Thus a single lengthwise row of 24 holes sufficed; these were 10 mm apart in lateral projection, beginning 10 mm behind the upstream (dorsal) end. Each was normal to the surface, 1 mm in diameter, and penetrated to an axial hole running downstream to an exit within the "arms." An additional hole entered the model be- tween head and mantle margin. For pressure measurements all but one of these aper- tures were occluded by small pieces of plastic adhesive tape. The model was supported by a cylindrical tube that continued the axial hole downstream and by a cylindrical support beneath the widest part of the mantle, diametrically opposite the row of holes. Pressure measurements were made in air at speeds of 8. 75, 16.5, 37.3, and 76.7 m • s~', the lower speeds in the large open circuit tunnel of the Department of Zoology (Duke University) and the higher speeds in the closed circuit tunnel of the School of Engineering. All were differential measurements referred to a static aperture, 1 mm in diameter, 0. 1 m downstream from the leading edge of a flat plate parallel to flow that was mounted adjacent to the model. Where neither compressibility of fluids nor fluid-fluid interfaces complicate mat- ters, the choice of fluid and size of model are quite arbitrary as long as the natural value of the Reynolds number is maintained. Equality of Reynolds number (Re) between reality and model assures dynamic similarity of flows and thus equality of the dimensionless force and pressure coefficients, as discussed by Vogel (1981), Vogel MANTLE CAVITY REFILLING IN SQUID 63 and Loudon (1985), and standard texts in fluid mechanics. Shifting from 20°C seawa- ter to air entails a speed increase factor of 14.5; the use of a model 1.5 times larger than an assumed size of animal mandates a speed decrease of this latter factor. Thus the present speeds in air are functionally equivalent to speeds of 0.91, 1.6, 3.9, and 7.9 m • s^1 in seawater providing that pressures are expressed as pressure coefficients. (A pressure coefficient is simply the measured pressure difference divided by the dy- namic pressure, the latter being the product oif half the density of the medium and the square of the velocity. In practice, pressures were coverted to the coefficients using the density of air and the tunnel velocities; the coefficients were reconverted as needed to equivalent pressures in seawater by substituting the density of seawater and the corresponding speeds in water.) Pressure differences were measured with a barocell electric manometer and capac- itative differential pressure sensor (Datametrics-Dresser, 1 174-B1A-5A1-A1 and 570D-10T-2A1-V1X). The combination had a maximum capacity of 1300 Pa and an accuracy, according to the manufacturer, of better than 0. 1 3 Pa (in practice better than any locally available standard). This accuracy corresponds to a pressure coeffi- cient of less than 0.003 for all present measurements. Tunnel speeds were determined with a Pitot-static tube, 3.2 mm in diameter (United Sensor, Inc., PCC 12-KL) and the same barocell and sensor except for the highest speed, for which a simple water manometer (readable to 0.2 mm) provided the necessarily wider range. Pressure drag (total drag less skin friction) can be viewed as a summation of the streamwise-directed components of the pressures on the surface of a body, where each datum for pressure has been multiplied by the area of surface to which it applies. If skin friction is negligible, it is thus a simple matter to derive values for drag from the pressure measurements; these drag data can be converted to drag coefficients in a manner analogous to the conversion of pressures to pressure coefficients (Vogel, 1981). Thus a general test of procedure and apparatus was carried out by determining the pressure distribution around circular cylinders and from the latter computing drag coefficients — values of 1 .0 were obtained at Reynolds numbers of 8000 and 32,000; a commonly cited value is 1.1 (Hoerner, 1965), of which about 3% is skin friction. Similarly, a sphere at Re == 28,000 gave a drag coefficient of 0.42; the ac- cepted value is 0.47. RESULTS Pressure coefficients Figure 1 (bottom) gives pressure coefficients as a function of position for the model squid. Several points are noteworthy. ( 1 ) These curves are typical of results obtained for streamlined objects (see Dis- cussion). They vary little with changes in Reynolds number (proportional to speed) although the minimum pressure coefficients are slightly more negative at higher Reynolds numbers. Thus the model is well-streamlined— if the decrease in girth is appropriately gentle as flow moves downstream from the point of maximum width then little separation of flow occurs; irregularities further downstream (unfaired man- tle margin, head, and arms) are of little consequence. In fact, building up the head area with modeling clay had so little effect that no systematic exploration was made. (2) The minimum pressure coefficients occur well downstream, not inevitably the case for streamlined bodies of revolution. Clearly location of the pressure mini- mum near the point of maximum girth will be advantageous for flow-induced refill- ing of the mantle cavity. 64 S. VOGEL -0.2 FIGURE 1 . Top: the model squid, with flow from left to right, showing location of the apertures for measurement of pressure. Bottom: pressure coefficient (Cp) as a function of relative distance downstream from the dorsal extremity; the abscissa is scaled to correspond to the drawing above. Extrapolation to a coefficient of unity at the upstream extremity (dotted line) presumes the existence of a normal forward stagnation point. Speeds refer to equivalent values in seawater; all data for the two intermediate speeds fall between those shown for the maximum and minimum speeds. (3) The pressure coeffients for the mid-mantle region are about —0. 1 5 while those for the region around the head from which refilling must take place are about +0. 10; a difference of approximately 0.25 is thus available for refilling the mantle cavity. The corresponding pressure differences in seawater for the four speeds tested are 1 06, 330, 1950, and 8000 Pa. (The wide range of these latter figures reflects multiplication by the squares of the velocities to obtain pressures from the pressure coefficients.) Pressure drag While not directly germane to the present hypothesis, values for pressure drag emerge from these measurements of pressure. Expressed as drag coefficients for the four speeds tested, the latter are (in ascending order of speed) 0.070, 0.060, 0.027, and 0.033, all referred to frontal (maximal cross-sectional) area. Dividing these values by 13.6 (the ratio of total surface to frontal area) shifts the reference to total surface or "wetted" area. Dividing the values instead by 2. 1 3 (the ratio of volume to the two- thirds power, a nominal area, to frontal area) shifts the reference to V2/3. These coefficients are about an order of magnitude below those of cylinders and spheres, cited earlier. There is no obvious discontinuity in the data corresponding to the drop in pressure drag attending the delay of separation as the boundary layer becomes turbulent (Vogel, 1981). Both these features of the data indicate good streamlining with little separation and low total drag for the model. If this crude MANTLE CAVITY REFILLING IN SQUID 65 model gives such a good account of itself then it is likely that real squid moving rearward are well-streamlined. DISCUSSION Evaluation of the contribution of flow-induced pressure differences to mantle cav- ity refilling requires two further items of information: (1) the pressure available for refilling in the absence of any direct hydrodynamic effect and (2) the speeds at which squid move during rearward repetitive jetting. Three separate approaches to estimating the non-hydrodynamic pressure avail- able are currently practical. First, a very rough upper limit is suggested by the capabil- ity of the radial muscle of the mantle. The latter underlies about 10% of the mantle surface (Bone et al., 1981); such muscle should be capable of a stress output of 100 kN-m~2 in the non-isometric contraction by which the mantle wall is thinned, or 10 kN-m~2 referred to mantle area instead of muscle cross-section. Half of the output will be lost since the mantle cannot expand longitudinally, leaving 5 kN • m 2 or 5000 Pa. Indeed, this figure may considerably exaggerate what the muscle can do in the latter part of mantle expansion when it is already substantially contracted. And it is in this hyperexpansion phase that the elastic recoil with which mantle expansion begins is no longer significant (Gosline and Shadwick, 1983). Conversely, at near maximum expansion, flow-induced pressures should be greatest. Second, from observed rates of refilling in non-swimming squid together with the size of the openings through which refilling occurs, Trueman (1980) calculated an inhalent pressure of 1200 Pa. And third, rough but direct measurements of peak refilling pressures are given by Trueman ( 1 980) as 1 500 Pa and by Gosline and Shad- wick (1983) as 1000 Pa. More problematical are the speeds attained by swimming squid during the rear- ward bursts of repetitive jetting. The animals are notoriously ill-adapted to survive in any non-pelagic situation (Summers and McMahon, 1 974); and high-speed, repeti- tive jetting seems never to have been elicited under controlled and reproducible cir- cumstances. Johnson et al. (1972) and Trueman (1975) calculated a speed of 3.5 m • s~'; but the value is based on the drag coefficient of a sphere (0.47), certainly much too high for a well-streamlined form. Replacing this assumption with a coefficient an order of magnitude lower (the coefficient of pressure drag for the model at the highest speed tested here plus an estimate of skin friction from data cited by Hoerner, 1965) increases their value 1005 times, to 11 m-s"1. Beyond that, the evidence is mainly anecdotal. Alexander (1977) notes without disparagement claims of swimming speeds up to 8 m-s~'. Lane (1960) mentions re- ports of squid leaving the water vertically and clearing deck railings 4.5 to 6 m above a calm sea. Assuming gravitational deceleration only, these heights imply exit speeds of 9. 5 to 11 m-s~'. Arata( 1954) notes an airborne trajectory about 2 m in height and 14 m horizontally; ignoring drag and assuming a normal ballistic trajectory gives an exit speed of about 12 m-s"1. Both of these simple calculations, though, may give overestimates: not only do they ignore drag and lift (minor matters for a squid in air) but they neglect the possibility that the last jet of water may be ejected while the squid is in air. And the latter certainly occurs — the photographs of Cole and Gilbert (1970) show that acceleration can continue while an animal is airborne. In short, it seems reasonably certain that a squid can generate refilling pressures of around 1200 Pa. Similarly, it appears that a squid, pursued by a predatory fish, can achieve speeds in the range of 5 to 1 0 m • s~ ' . The present study indicates that the net pressure coefficient available for flow-induced refilling is about 0.25. Figure 2 uses 66 S. VOGEL 100 80 3 Q Z 60 O 40 20 4 6 SPEED, m/s 8 10 12 FIGURE 2. The relative contribution of flow-induced pressure differences to total pressure available for refilling as a function of rearward swimming speed, based on present data. this datum to estimate the relative contribution of hydrodynamics to refilling as a function of swimming speed. It appears that flow-induced pressure accounts for between 50 and 90% of the total pressure available for refilling in squid during rapid repetitive jetting. Thus the phenomenon is a substantial one. It should permit both more rapid refilling and a greater overall inflation of the mantle cavity and hence a greater maximum swim- ming speed. It has the virtue of being increasingly effective as speed increases and will be at its best near full inflation, where elastic recoil is minimal and radial muscles have already shortened considerably. It is unlikely, though, that a squid can save energy through the use of flow-induced pressures. Any force that pulls the mantle outward and contributes to refilling is a force against which the circular muscles will have to contend during the expulsion phase, and the distances of mantle movement outward and inward must be the same. But inasmuch as repetitive rearward jetting constitutes an escape response, speed must be far more relevant than any energetic considerations. The present model is certainly an imperfect squid, although just how imperfect is uncertain in the absense of information on the shape of a squid during repetitive high-speed jetting. Examination of the available data on pressure distribution around streamlined bodies of revolution, though, suggests that a detailed investigation of the effects of taper, size of gap between mantle margin and head, and head size is not warranted. Graphs of pressure coefficients versus location simply don't vary much with either the details of shape or, over a very wide range, the Reynolds number. (The present data span a range of Reynolds numbers from 1 55,000 to 1 ,350,000.) Arnstein and Klemperer (1936) give a curve similar to the present data for a full-size airship for which the Reynolds number, based on length, was over 108. Goldstein (1938) MANTLE CAVITY REFILLING IN SQUID 67 (Fig. 2 1 5a) presents data at Re = 3 X 106 especially similar to the present results. Data for a model of the alga, Halosaccion, at Re == 14,000 (Vogel and Loudon, 1985), are similar as well, as are data obtained using a toy water rocket (Park Plastics Co., Lin- den, New Jersey) whose fins had been removed at Re = 105. Minimum pressure coefficients range only from -0. 1 5 to -0.2. The main consistant variation is that the location of the minimum is further downstream for shapes with less blunt upstream ends. Thus the indifference of the present model to alterations in shape comes as little surprise. And, while a squid of about 0.2 m in length was assumed in calculating equivalent swimming speeds in water, the pressure coefficients vary so little with Reynolds number that the data should apply about as well over a considerable size range. Other, more serious limitations of the present study ought to be borne in mind. Measurements were carried out in steady flow, whereas jetting is an unsteady motion (sensu Daniel, 1984). And the act of refilling will to some unknown extent relieve the predicted pressure differences. The major uncertainty, though, comes from the lack of data on actual swimming speeds; since flow-induced pressure is approximately proportional to the square of speed, mere refinement of model and experimental conditions would be of limited usefulness in the absence of better information on swimming speeds in nature. Nevertheless it is difficult to imagine how substantial flow-induced pressures can be avoided. Indeed, at least one other structural feature of squid is consistent with the occurrence of such pressures during rapid, rearward jetting. The pen more nearly encircles the mantle near the upstream (dorsal) end, where flow will generate com- pressive forces (positive pressure coefficients in Fig. 1 ), and thus perhaps provides a measure of compression resistance in a manner analogous to the skull of a fish. ACKNOWLEDGMENTS Edward Shaughnessy and Marios Pougare kindly made available to me the high speed wind tunnel. And Hugh Crenshaw, Olaf Filers, John Gosline, Matthew Healy, William Kier, Anne Moore, and Stephen Wainwright have made useful comments throughout this project. LITERATURE CITED ALEXANDER, R. McN. 1977. Swimming. Pp. 222-248 in Mechanics and Energetics of Animal Locomo- tion, R. McN. Alexander and G. Goldspink, eds. Chapman and Hall, London. ARATA, G. F. 1954. A note on the flying behavior of certain squids. Nautilus 68: 1-3. ARNSTEIN, K., AND W. KLEMPERER. 1936. Performance of airships. Pp. 6: 49-133 in Aerodynamic The- ory, W. F. Durand, ed. Dover Publications, Inc., New York (reprint, 1963). BONE, Q., A. PULSFORD, AND A. D. CHUBB. 1981. Squid mantle muscle. J. Mar. Biol. Assoc. U.K. 61: 327-342. BRADBURY, H. E., AND F. A. ALDRICH. 1969. Observations on locomotion of the short-finned squid lllex illecebrosus (LeSueur, 1 82 1 ) in captivity. Can. J. Zool. 47: 74 1 -744. COLE, K. S., AND D. L. GILBERT. 1970. Jet propulsion in squid. Biol. Bull. 138: 245-246. DANIEL, T. L. 1984. Unsteady aspects of aquatic locomotion. Am. Zool. 24: 121-134. DuBois, A. B., G. A. CAVAGNA, AND R. S. Fox. 1974. Pressure distribution on the body surface of swim- ming fish. J. Exp. Biol. 60: 58 1-59 1 . GOLDSTEIN, S. 1938. Modern Developments in Fluid Dynamics. Dover Publications, New York (reprint, 1965). 702pp. GOSLINE, J. M., AND R. E. SHADWICK. 1983. The role of elastic energy storage mechanisms in swimming: an analysis of mantle elasticity in escape jetting in the squid, Loligo opalescens. Can. J. Zool. 61: 1421-1431. HOERNER, S. F. 1965. Fluid-dynamic Drag. Author, 2 King Lane, Greenbriar, Brick Town, NJ 08732. 68 S. VOGEL JOHNSON, W., P. D. SODEN, AND E. R. TRUEMAN. 1972. A study in jet propulsion: an analysis of the motion of the squid, Loligo vulgaris. J. Exp. Biol. 56: 155-165. LANE, F. W. i957. Kingdom of the Octopus. Jarrolds, London. 300 pp. O'DOR, R. K. 1982. Respiratory metabolism and swimming performance of the squid, Loligo opalescens. Can. J. Fish. Aquat. Sci. 39: 580-587. SUMMERS, W. C, AND J. J. McMAHON. 1974. Studies on the maintenance of adult squid (Loligo pealei). I. Factorial survey. Biol. Bull. 146: 279-290. TRUEMAN, E. R. 1975. The Locomotion of Soft- Bodied Animals. American Elsevier Publishing Co., New York. 200 pp. TRUEMAN, E. R. 1980. Swimming by jet propulsion. Pp. 93-105 in Aspects of Animal Movement, H. Y. Elder and E. R. Trueman, eds. Cambridge, U.K., Cambridge University Press. VOGEL, S. 1 98 1 . Life in Moving Fluids. Willard Grant Press, Boston. 352 pp. VOGEL, S. 1985. Flow-assisted shell reopening in swimming scallops. Biol. Bull. 169: 624-630. VOGEL, S., AND C. LOUDON. 1985. Fluid mechanics of the thallus of an intertidal red alga, Halosaccion glandiforme. Biol. Bull. 168: 161-174. WARD, D. V. 1972. Locomotor function of the squid mantle. /. Zoo/. Lond. 167: 487-499. WARD, D. V., AND S. A. WAINWRIGHT. 1972. Locomotory aspects of squid mantle. J. Zoo/. Lond. 167: 437-449. Reference: Biol. Bull. 172: 69-82. (February, 1987) METAL REGULATION AND MOLTING IN THE BLUE CRAB, CALLINECTES SAPIDUS: COPPER, ZINC, AND METALLOTHIONEIN DAVID W. ENGEL National Marine Fisheries Service, NO^4A, Southeast Fisheries Center, Beaufort Laboratory. Beaufort, North Carolina 28516 ABSTRACT Blue crab, Callinectes sapidus, hemolymph and digestive glands were examined at different stages of the molt cycle to determine whether molting affected tissue and cytosolic partitioning of copper and zinc and, if so, whether metallothionein was involved. The crabs used in these determinations were not exposed to elevated con- centrations of copper or zinc. Concentrations of hemocyanin, copper, and zinc in the hemolymph all decreased significantly during molt. They were lowest at the soft crab stage (A2) and highest during premolt (D!-D3) and intermolt (C4). The digestive gland copper concentrations also were highest during premolt (D!-D3) and lowest in the papershell stage (BO. Zinc followed the same general pattern in both hemolymph and digestive glands. The cytosolic distributions of copper and zinc were determined in the digestive glands using gel filtration chromotography. The elution profiles showed that the percentages of copper and zinc on metallothionein ranged from 10% copper/ 90% zinc at DI to 100% copper at B, . The estimated concentrations of metallothi- onein were highest during intermolt (Ci) and premolt (D!-D3) and lowest during the papershell (B,). The observed changes in the tissue and cytosolic partitioning of cop- per and zinc are consistent with the physiological changes occurring in the crabs. These observations support the hypothesis that metallothioneins are naturally occur- ring proteins that are actively involved in the synthesis of hemocyanin and zinc regu- lation during the normal processes of growth in blue crabs. INTRODUCTION Molting in crustaceans and the biochemical mechanisms that control the turn- over and synthesis of metallothioneins are complex and possibly interacting physio- logical and biochemical processes. Before this investigation, no attempt had been made to link these two physiological activities. This effort, therefore, was designed to determine if changes in the metal composition and amount of metallothionein in the blue crab, Callinectes sapidus, are correlated with the molt cycle. Molting has been examined in detail in a variety of marine and freshwater crusta- ceans. Skinner (1985) provides an overview of the complex physiological and bio- chemical processes that are involved in crustacean molting and regeneration with an emphasis on hormonal control. Recently, investigators demonstrated that there are changes in hemocyanin concentrations in the hemolymph of lobsters and blue crabs (Hagerman, 1983; Mangum et al., 1985) during the molt cycle and also osmoregula- tory and biochemical changes in molting blue crabs (Towle and Mangum, 1985; Wheatly, 1985). Additionally, there are some reports of changes in metal metabolism Received 12 July 1986; accepted 25 November 1986. 69 70 D. W. ENGEL during molt (Martin, 1975; Adams et al, 1982) which may involve the metal-binding protein metallothionein. Metallothionein has been characterized as a detoxifying protein that is synthe- sized in response to elevated metal concentrations both internally and in the environ- ment. It is a low molecular weight protein (< 10,000 Mr) which binds cadmium, cop- per, zinc, and mercury, and contains high concentrations of cysteine (>20%) (Che- rian and Goyer, 1978). Since initial isolation by Margoshes and Vallee (1957), metallothioneins have been studied extensively in mammals and invertebrates (Ko- jimaandKagi, 1978;NordbergandKojima, 1979;Roesijadi, 1981). These investiga- tions have been concerned primarily with the involvement of metallothionein in detoxification of metals in animals and with the characterization of the protein mole- cule. Currently there is renewed interest in determining "normal" function of metal- lothioneins and the metabolic processes that control constitutive levels of metallothi- oneins in unstressed organisms (Engel and Brouwer, 1984a; Cousins, 1985; George and Viarengo, 1985; Engel and Roesijadi, in press). Metallothioneins have been demonstrated in marine crustaceans exposed to ele- vated levels of cadmium and copper in their water or food (Olafson et al, 1979a, b; Overnell and Trewhella, 1979; Rainbow and Scott, 1979; Overnell, 1982; Wiedow et al., 1982; Engel and Brouwer, 1984b; Engel et al., 1985; Engel and Brouwer, 1986; Overnell, 1986). In the blue crab, metallothioneins have been shown to bind cad- mium, copper, and zinc in both laboratory exposed animals (Brouwer et al., 1984; Engel et al., 1985) and in animals from contaminated environments (Weidow et al., 1982; Engel and Brouwer, 1984b). Gel chromatography of the digestive gland cytosol yields proteins that have an estimated molecular weight of 7,000- 10,000 Mr and bind the three metals. Further purification of these proteins by ion exchange chromatogra- phy demonstrated that there was one protein that binds cadmium and zinc and an- other that binds copper. Further characterization of the cadmium metallothionein showed that it had the characteristics of the mammalian metallothionein (Brouwer et al., 1984). In our investigations with blue crab metallothioneins, we have observed apparent correlations between the physiological condition of the crab and the metal distribution in the cytosol. The present research effort was designed to investigate whether molting contrib- utes to variability in metallothionein concentration and metal composition in blue crabs. These experiments were designed to examine total copper and zinc in the he- molymph and digestive glands and the cytosolic distributions of copper and zinc in the digestive glands of crabs during six different stages of the molt cycle. MATERIALS AND METHODS Animals All blue crabs used in these experiments were collected from estuarine waters in the vicinity of Beaufort, North Carolina. Three groups of crabs were obtained throughout the summer of 1 985 and these animals included premolt, soft, papershell, and intermolt crabs, and were comprised of about 75% immature females. The tem- perature and salinity at the time of collection was 20 to 22°C and 31 to 33%o. The designations of molt stage were according to those described by Mangum (1985) for the blue crab and by Passano (1960) for brachyuran molt stages. Tissue metal measurements The concentrations of copper and zinc were determined in samples of digestive gland and hemolymph from individual blue crabs. Hemolymph samples were col- MOLTING AND METALLOTHIONEIN 71 lected by severing the fifth pereiopod at the meropodite and collecting the fluid in a polyethylene vial. A portion of the hemolymph was taken for metal analysis and the remainder was used for determination of hemocyanin concentration. The crabs were killed by removing the carapace, and the digestive gland was dissected and used for total metal measurements and cytosolic metal determinations. The tissue used for determination of cytosolic distribution of metals was flash-frozen and stored in a freezer at -70°C. Tissue samples used for metal analysis were oven dried at 100°C for 48 hours and wet ashed with concentrated HNO3 at 90°C. Residue was dissolved in 0.25 TV HC1, and concentrations of copper and zinc were measured using flame atomic absorption spectrophotometry. Preparative and measurement techniques were calibrated against the National Bureau of Standards Oyster Reference Material #1566. Measurements of hemocyanin concentrations in the blue crab hemolymph sam- ples were made by Dr. Marius Brouwer (Duke University Marine Laboratory/Bio- medical Center, Beaufort, North Carolina). The clotted hemolymph was homoge- nized, centrifuged, and hemocyanin concentration was determined spectrophoto- metrically using the extinction coefficient of the native undissociated protein (Johnsons al, 1984). Chromatography Individual crab digestive glands were used in each chromatographic separation. Frozen tissue samples were weighed and then placed in 10 ml of 60 mAfTris buffer, pH 7.9, with 2 mM 2-mercaptoethanol in an ice bath and homogenized 30 seconds with Brinkman Polytron1 homogenizer. The homogenate was then centrifuged at 105,000 X g for 90 minutes at 4°C. The resulting supernate was considered to be the cytosolic fraction. Mass balances were conducted on each sample, and aliquots of the homogenate and cytosol and the entire pellet with fat cap were analyzed for metals according to the procedure outlined above. Gel-chromatography of the cytosolic fraction was done on a Sephadex G-75 col- umn (2.6 cm X 66 cm). The elution buffer was 60 mM Tris, pH 7.9, with 2 mM 2- mercaptoethanol. Individual 7-ml fractions were analyzed for absorbance at 250 nm, and for concentrations of copper and zinc were measured directly by flame atomic absorption spectrophotometry. The column was standardized with blue dextran (> 106 Mr), bovine serum albumin (67,000 Mr), ovalbumin (43,000 Mr), chymotryp- sinogen A (25,000 Mr), ribonuclease ( 1 3,700 Mr), and potassium. The determination of molecular weight was calculated from: iVav — Vt - V0 The elution characteristics of the column were: Ve — the elution volume of an un- known or standard, V0 — the elution volume of the high molecular weight material excluded from the column, and Vt — the total elution volume of the ionic component or "salt fraction" as measured by potassium. The peak which eluted with an apparent molecular weight of 7,000-10,000 Mr, was considered to be the metallothionein fraction. Further analysis of the metallothionein peak was performed on a DEAE Sephacel ion-exchange column ( 1 .6 X 10 cm). A gradient was generated from 250 ml, 60 mM ' Mention of a trade name does not constitute an endorsement by the National Marine Fisheries Service, NOAA. 72 D. W. ENGEL z A. B. A C. 0 z , T T < 70 - 2 14 f cc ^ 0. Z 60 - ff 12 - — "? _ T |= 50 1 ^ + 1 t— '^ ul |* 10 0 2 - il f 8 1 40 - o* 8 - m^ 2 - _ E O T- O '._. |~ 30 • 6 - rij z r: 0 x JL < T UJ ^^ f 0 C •r T > 20 Q - * * Q. 4 0. ~ * O 1 — T § 10 - 0 „ 0 2 - Z N UJ I 0 n n D1 D2 D3 A2 B, C4 " D, D2 D3 Aj B, C4 w D, D2 D3 A2 B, C4 MOLT STAGE FIGURE 1 . Hemolymph concentrations of hemocyanin (A), copper (B), and zinc (C), from blue crabs during different stages of the molt cycle. Molt cycle stages were: premolt white sign (D,), pink sign (D2), and red sign (D3); soft shell (A2); paper shell (Bi); and intermolt (C4). Hemocyanin measurements made by Dr. M. Brouwer, Duke Marine Biomedical Center, Beaufort, NC. Each mean represents five or six individual hemolymph samples ± 1 standard error. Tris, pH 7.9, and 250 ml, 300 mA/Tris, pH 7.9, both contained 2 mM 2-mercapto- ethanol. Fractions were analyzed for conductivity and concentrations of copper and zinc. RESULTS Hemolymph and tissue metal concentrations Blue crab hemolymph concentrations of hemocyanin, copper, and zinc changed during the molt cycle. Hemocyanin concentrations during the soft crab (A2) and pa- pershell (B,) stages were significantly different (P < 0.05) from the three premolt stages (D1? D2, D3) (Fig. 1A). The concentrations of copper and zinc followed the same pattern as the hemocyanin and also were different from the intermolt crabs (C4) (Fig. IB, C). The correlation between copper and hemocyanin would be expected, since copper is an integral part of the hemocyanin molecule. The positive correlation between changes in hemocyanin concentration and zinc concentrations suggests that the zinc may be bound to hemocyanin (Fig. 1C). The digestive gland concentrations of copper and zinc were reflective of the physi- ological changes that occur during the molt cycle in the blue crab. Copper concentra- tions in the digestive gland showed significant decreases (P < 0.05) in both the soft and papershell stages (Fig. 2A), and the papershell stage also was lower than the soft crab stage (P < 0.05). The differences were 3.4 and 6.0 times lower than the D, stage of premolt. The observed decreases in copper in the digestive gland are in close corre- lation with decreases of copper in the hemolymph (Fig. IB). This means that copper is being lost or excreted, and not conserved by the digestive gland. In the initial por- tions of the molt cycle, however, it appears that, as the crabs move from intermolt (C4) to premolt (Dj-Ds), there is an increase in copper in the digestive gland, which may be associated with changes in hemocyanin synthesis and degradation, though this retention of copper does not persist. Zinc concentrations showed the same type of variation during the intermolt and premolt periods as copper, and the soft and MOLTING AND METALLOTHIONEIN 73 2.5 A. B. Z -i < § 2.0 . ft n CONCENTR I/kg wet we 01 - •• -• ||r.o < § DC «n T _ *- t).U z| --, UJ s «;n - j i i il jjjb 1.0 . O C3)^ z ^ O^4.0 - i -1- Q. x C^^t oc _r 0 'o 3.0 . O ~L z *~ .5 - -t N-2.0 - 1.0 - 0 0 i D, D2 D3 A2 B, Q D, D2 D3 A2 B1 C4 MOLT STAGE FIGURE 2. Digestive gland concentrations of copper (A) and zinc (B) from blue crabs during different stages of the molt cycle (see Fig. 1). All mean values represent five or six individual samples of digestive gland ± 1 standard error. papershell stages also were different from premolt crabs (D2), (P < 0.05) (Fig. 2B). The concentrations of copper and zinc in the digestive glands of the C4 and D( stage crabs were in good agreement with concentrations measured previously in intermolt blue crabs collected near Beaufort (Engel and Brouwer, 1984b). Cytosolic metal distribution The Sephadex G-75 elution profile for the cytosol prepared from the digestive gland of a DI stage blue crab showed partitioning of the metals with a pronounced 7-10 K Mr metallothionein-like protein that contained zinc and copper (Fig. 3 A). Both zinc and copper also were present in the high molecular weight material (frac- tions 13-22), but only zinc was present in the very low molecular weight material (fractions 42-52). Further characterization of the metallothionein-like protein peak (fractions 32- 38) was done on a DEAE Sephacel ion-exchange column with a Tris gradient 60- 300 mA/Tris, pH 7.9. Three metal containing peaks were resolved, a zinc component in the breakthrough volume and copper and zinc components that eluted at conduc- tivities of 5.5 and 6.0 mS/cm (Fig. 3B). These conductivities represent Tris concentra- tions of 1 10 and 125 mM Tris at pH 7.9, which compare favorably to the data of Brouwer and co-workers (Duke University Marine Biomedical Center, Beaufort, NC, pers. comm.) for cadmium metallothionein in blue crabs. Therefore, the protein peak eluting from the G-75 gel filtration column at an apparent molecular weight of 7-10 K Mr will be referred to as metallothionein (MT). The observed separation between the copper and zinc peaks, however, does not necessarily mean that they are separate copper and zinc-binding proteins, but may simply reflect differences in charge characteristics on the protein. 74 D. W. ENGEL 4.0 > 3.0 § O 2.0 | 1.0 ° FRACTION NUMBER FIGURE 3. Elution profile of the cytosolic fraction of the digestive gland of an intermolt crab showing the partitioning of metals; (A) Sephadex G-75 elution profile column 2.6 X 65 cm with a flow rate of 30 ml/h of 60 mMTris, pH 7.9 with 2 mM-2 mercaptoethanol and (B) DEAE — Sephacel ion-exchange col- umn 1.6 X 1 0 cm using a gradient generate from 250 ml, 60 mM Tris, pH 7.9 and 250 ml, 300 mM Tris, pH 7.9 and both contained 2 mM-2 mercaptethanol. The arrows denote the metallothionein peak that was pooled and subjected to ion exchange chromatography. The peak marked K+ represents the position of the potassium peak (very low molecular weight). .04- The individual elution profiles of digestive gland cytosol for the six different stages of molting blue crabs show differences in metal distributions that can be correlated with the molting process (Fig. 4). The premolt crabs examined (D!, D2, D3) have a very pronounced metallothionein peak, which is dominated by zinc (Fig. 4A, B, C). Zinc is also present in the high (>60 K, Mr), intermediate (~40 K, Mr), and very low (<1 K, Mr) molecular weight fractions. Copper is present only in the high and metallothionein portions of the profile. In the soft crabs (A2), both zinc and copper are present in the metallothionein region (Fig. 4D), and the relative amount of copper has increased. Otherwise, the distributions of copper and zinc in the elution profiles are the same as before. The metal content of the metallothionein in papershell crabs (B,) is entirely copper with no measureable zinc (Fig. 4E), and the zinc is present only in the high, intermediate, and low molecular weight fractions. Hard shell or intermolt crabs (C4) have a very elevated copper metallothionein with a low amount of zinc present in the same fractions (Fig. 4F), and zinc follow the same pattern as seen in the A2 and BI crabs. MOLTING AND METALLOTHIONEIN 75 Estimates of total copper /zinc binding proteins Calculations to estimate the total metallothionein bound metal in the digestive gland cytosol of the different molt stages were made using the modified relationship of Frazier and George (1983). The relationship is as follows: B.P. - EM&1 / SamApf Volume / W* Weight of Tissue 6 10 / (10 ml = 100%) / m grams The molar concentrations of copper and zinc were calculated from the five peak fractions of each metallothionein peak and were adjusted for applied sample volume and wet weight of tissue. It was assumed that individual metallothionein molecules bind only copper or zinc, and that the blue crab metallothionein binds six zinc atoms [e.g., the Scylla protein binds 6 moles of cadmium per mole of protein which is as- sumed to be the same as zinc (Olafson et al, 1979)], or 10 atoms of copper [e.g., conservative estimate from conversations with M. Brouwer (pers. comm)]. Such cal- culations are estimations of concentration and are not meant to be quantitative, but do allow for comparisons between groups of crabs. When the comparisons were made between the different groups of crabs, differ- ences not only were observed in the total amount of protein-bound metal, but also in the relative amount of copper and zinc that was present (Fig. 5). The estimated concentrations of protein- (i.e., metallothionein) bound metal followed roughly the same pattern as total copper and zinc concentrations in tissue (Fig. 5A). When the percentages of metallothionein-bound copper were calculated, they ranged from 7% to 10% of total in the premolt stages, to 35% in the soft crab, to 100% in the paper shell, and 75% at intermolt (Fig. 5B). Such differences show that the cytosolic distri- butions of metals change during molting and that metal metabolism and metallothi- onein must be linked directly with the extensive cellular processes that are occurring. This hypothesis is supported further by the observed changes in the percentages of total digestive gland copper and zinc that are present in the cytosolic fraction (Fig. 5C). These calculations, made from the mass balances, show that the percentages of copper in the cytosol at different stages of molting range from 60 to 80 percent. Dur- ing A 2 and B, , when the copper bound to metallothionein is increasing, the percent- age of zinc in the cytosol drops from ~75 to 34 percent. This decrease can be ac- counted for by zinc in the pellet, indicating that there is an intracellular redistribution of metal, which results in some conservation. Since the zinc is in the pellet, it must be incorporated into relatively inert structures or bound to membranes that were not disrupted by the homogenizing process. DISCUSSION The blue crabs used in these experiments were not exposed to artifically elevated concentrations of metals in the laboratory. Therefore, changes in metal concentra- tions in the hemolymph and the digestive gland and in cytosolic metal distributions were constitutive processes associated with molting. If it is assumed that these changes in metal partitioning are involved in the synthesis of metalloproteins, such as en- zymes and respiratory proteins, then a model of metal metabolism and metallothi- onein function for blue crabs can be developed. The greatest changes in overall metal concentration and cytosolic distribution of metals occurred at or around the time of ecdysis and early soft crab. At this time, 76 D. W. ENGEL in M o o ^ HI o < CO cr O u> m 2.0 1.8 1.6 1.4 1.2 1.0 .8 .6 .4 .2 0 1.6 1.4 1.2 1.0 .8 .6 .4 .2 0 1.8 1.6 1.4 1.2 1.0 .8 .6 .4 .2 0 Absorbance Copper ..Zinc 10 20 30 40 FRACTION NUMBER r-. 50 1.7 1.5 0 E 1.3 o) 3- 1.1 .9 N .7 DC UJ Q. -5 & O .3 .1 .7 .6 .5 O) 3- •4 | N •3 Of UJ •2 8: .1 8 0 1.7 1.5 1.3 1 ^i 1.1 3- • 9 Z N DC UJ 5 fc -» o 3 Q .1 60 FIGURE 4. A series of representative G-75 elution profiles of digestive gland cytosol from all of the different molt stages that were examined [A(D,), B(D2), C(D3), D(A2), E(B,), and F(C4)]. Conditions were the same as in Figure 3A. there were significant decreases in the hemolymph concentrations of copper and zinc, which were correlated with decreases in hemocyanin concentrations. These changes in hemocyanin concentration agree favorably with the data of Mangum et al. (1985). MOLTING AND METALLOTHIONEIN 77 in CM 0 o s«* III o z < CO FRACTION NUMBER FIGURE 4. (Continued) The decrease in copper is understandable, since hemocyanin is a copper protein, but the associated decrease of zinc strongly suggests that it also is bound to hemocyanin. Zatta (1984) demonstrated with Carcinus maenas that each minimal functional sub- unit of hemocyanin has four zinc-binding sites, and that about 90% of the zinc in the hemolymph was not dialyzable. Brouwer and Engel (1982) also suggested that 78 D. W. ENGEL LU O DC Q. O z ^ 0 £ Z 0> DO ® u_ * o * z * o) z LU O o o 1X10" 5X10 -5 1X10 "5 5X10 -6 1X10" B, m 100 i 0 80 0 60 — < CO - i i — 1 £0 60 LU C OQ 40 LU Z Q-" 20 n - ^A ! E 1 i — | i i — | B, MOLT STAGE FIGURE 5. Copper and zinc partitioning on the metal-binding protein (i.e.. metallothioneins) and in the cytosol of the digestive glands of molting blue crabs. (A) Calculated estimates of total binding protein concentrations (clear bar) and copper-binding protein (cross hatched areas). Means ± I standard error (n = 5). (B) Calculated percentage of the total binding protein that binds copper. (C) Percentage of total digestive gland copper and zinc that was present in the cytosol as determined from mass balances on individual digestive glands. MOLTING AND METALLOTHIONEIN 79 hemocyanin is a major vehicle for the transport of zinc in blue crabs. Additional evidence for the association of zinc with hemocyanin in these experiments comes from the close correlation between the increase in both zinc and hemocyanin in the hemolymph of papershell crabs. Observed changes in total copper and zinc in the digestive gland during the molt cycle parallel fluctuations in the hemolymph, and the cytosolic distributions of these metals also show large changes. On a molar basis the changes in copper and zinc in the digestive gland between D3 and B1? are quite similar (Fig. 2) and suggest that there is some linkage between the two metals. This linkage also is associated with the changes in the cytosolic distributions of these metals. The most pronounced changes in cytosolic distribution, however, occur between premolt and papershell where the predominant metal bound to metallothionein changes from zinc to copper. These events are relatively rapid and are associated with other physiological and biochemi- cal events. During this portion of the molt cycle the percentage of copper bound to metallo- thionein increased as the crabs moved toward papershell (Fig. 5B). The percentage of copper associated with the metallothionein peak changed from 10% copper imme- diately prior to shedding, to virtually 100% copper at the papershell stage. It appears, therefore, that metallothionein changes from primarily binding zinc to binding cop- per at the time when the newly molted crabs may be resynthesizing hemocyanin. This correlation supports the hypothesis of Brouwer et al. (1986) that copper metal- lothionein is necessary as Cu+1 donor in the synthesis of hemocyanin in blue crabs and lobsters. Dominance of the copper form of metallothionein continues into the intermolt period and its concentration is high relative to the papershells, but roughly compara- ble to the total concentrations observed during premolt (Fig. 5A). Even though these animals are stable with regard to molting, the predominance of copper metallothi- onein suggests a very active turnover of the hemocyanin in the hemolymph of the blue crabs. Senkbeil and Wriston (1981) reported that the T1/? for hemocyanin in the lobster was 25.5 days. Since blue crabs may have a higher metabolic rate, their hemocyanin turnover rate may be more rapid. Such turnover rates would require an active pool of Cu+1 for degradation, storage, and synthesis of hemocyanin. In the molt cycle of the blue crab, there are significant decreases in copper in both the hemolymph and digestive gland at the time of molting (Fig. IB and 2A). Since it appears that the copper is lost by the animal upon molt, portions of the exuvia were measured for copper. These measurements did not show elevated levels of copper in either the molted exoskeleton or stomach and gut linings. Therefore, we do not know the route of copper excretion in the blue crab at molt. Djangmah (1970) indicated that Crangon vulgaris excreted copper during molt, and must accumulate it again from either water or food. Such large changes in copper in both hemolymph and the digestive gland also have been demonstrated in the crab, Cancer irroratus (Martin, 1 975), and in the crayfish, Austropotamobius pallipes (Adams et al., 1 982). Pathways of metal loss have been examined recently in the shrimp, Penaeus semisulcatus, (Al- Mohanna and Nott, in press), using both atomic absorption spectrophotometry and electronmicroscopy with EDAX. They demonstrated that copper and zinc fluctuated during the molt cycle, and that membrane-bound intracellular inclusions were pres- ent in cells of the hepatopancreas. These granule-containing cells appear to be sloughed at molt, and the granules which are rich in both copper and sulfur may relate back to metallothionein, and constitute a pathway of excretion for metallothi- onein bound metals. 80 D. W. ENGEL While it is possible to relate copper metallothionein to hemocyanin synthesis, it is more difficult to relate zinc metallothionein concentration to any specific synthetic process in the crabs. During the premolt period D,-D3 the predominence of zinc metailothionein is clear, with 90-94% of the total consisting of zinc metallothionein. The decrease in copper metallothionein is probably associated with a decrease in hemocyanin synthesis. Since it is during the premolt period that the new epidermal material is produced, the mobilization of zinc probably correlates with increased need for zinc-dependent enzymes (i.e., carbonic anhydrase) associated with calcification in the new exoskeleton (Henry and Kormanik, 1985). Unlike copper, which varies in concentration in the digestive gland depending upon the environmental condi- tions, zinc remains remarkably stable (Engel and Brouwer, 1984b). Additional data on tissue concentration of zinc in digestive glands of blue crabs from Long Island Sound to Florida further establish the fact that zinc is highly regulated by blue crabs (D. W. Engel, unpub. data). Also during the molt process, the copper concentration varies by a factor of 6 in the digestive gland, but zinc varies only by a factor of 1.7 (Fig. 2). The absolute changes on a molar basis are similar which suggests a linkage in metabolism. The data presented here may alter the perception of metallothionein as primarily a detoxifying protein. The function of metallothionein may be one of a regulatory protein for copper and zinc metabolism in unperturbed crabs. Since most of the early research emphasis was on its ability to bind cadmium and mercury, both toxic metals, it is not surprising that the focus was on detoxification. Cousins (1985), however, discusses some of the functions of metallothionein and indicates that it is also in- volved in stress reactions. Udom and Brady (1980) and Li et al. (1980) demonstrated that mammalian zinc metallothionein has the capability of reactivating apo-carbonic anhydrase, and they suggest that metallothionein has a regulatory or transport func- tion in normal physiology. Brouwer et al. (1986) has hypothesized that copper metal- lothionein is the Cu+1 donor for the synthesis of crustacean hemocyanin, and the results of this investigation support that hypothesis. They also go a step further in that they show that the pathways of zinc and copper, while linked through their binding to metallothionein, follow different pathways of mobilization and turnover. The present studies have established that normal physiological processes, such as growth, can influence metallothionein in regard to metal content and concentration in the blue crab. Further, they show that generalized statements with regard to metal metabolism, even within phylogenetic groups, may not be valid, as shown by our data on blue crabs and lobsters. Finally, the present investigation demonstrates the need for basic physiological and biochemical information in order to understand the processes involved in lexicological assessments of environmental impacts of metals. ACKNOWLEDGMENTS The author thanks Mr. William J. Bowen, III and Lt. (jg) Debra Davis of our Laboratory for their assistance during this investigation. Mr. Peter Berrini, Glouces- ter, NC; Hooper Family Seafood, Smyrna, NC; and Pittman Seafood, Merrimon, NC, supplied the crabs used in this investigation. The author also thanks Dr. Marius Brouwer, Duke University Marine Laboratory/Biomedical Center, Dr. G. Roesijadi, Department of Biology, Pennsylvania State University, and Drs. Brenda Sanders and Kenneth Jenkins, California State University, Long Beach, for their review of this manuscript. This research was supported in part by NOAA's National Ocean Service, Ocean Assessment Division. LITERATURE CITED ADAMS, E., K. SIMKJSS, AND M. TAYLOR. 1 982. Metal ion metabolism in the moulting crayfish (Austropo- tomobius pallipes). Comp. Biochem. Physiol. 72A: 73-76. MOLTING AND METALLOTHIONEIN 81 AL-MOHANNA, S. Y., AND J. A. NOTT. In press. The accumulation of metals in the hepatopancreas of the shrimp Penaeus semisulcatus de Haan (Crustacea: Decapoda) during the moult cycle. Pp. 195- 207 in Proceedings of the First Arabian Conference on Environment and Pollution. Allen Press, Oxford. BROUWER, M., AND D. W. ENGEL. 1982. Stoichiometry and functional consequences of Hg(II) and Cd(II) binding to arthropod hemocyanins. Pp. 289-309 in Physiological Mechanisms of Marine Pollu- tant Toxicity, WW. B. Vernberg, A. Calabrese, F. P. Thurberg, and F. J. Vernberg, eds. Academic Press, New York. BROUWER, M., T. BROUWER-HAEXUM, AND D. W. ENGEL. 1984. Cadmium accumulation by the blue crab, Callinectes sapidus: involvement of hemocyanin and characterization of cadmium-binding proteins. Mar. Environ. Res. 14: 7 1-88. BROUWER, M., P. WHALING, AND D. W. ENGEL. 1986. Copper-metallothioneins in the American lobster, Homarus americanus: potential role as Cu(I) donors to apohemocyanin. Environ. Health Per- spect. 65: 93-100. CHERIAN, M. G., AND R. A. GOYER. 1978. Metallothioneins and their role in the metabolism and toxicity of metals. Life Sci. 23: 1 - 1 0. COUSINS, R. J. 1985. Absorption, transport and hepatic metabolism of copper and zinc: special reference to metallothionein and ceruloplasmin. Physiol. Rev. 65: 238-309. DJANGMAH, J. S. 1970. The effects of feeding and starvation on copper in the blood and hepatopancreas, and on blood proteins ofCrangon vulgaris (Fabricius). Comp. Biochem. Physiol. 32: 709-73 1 . ENGEL, D. W., AND M. BROUWER. 1984a. Trace metal-bind proteins in marine molluscs and crustaceans. Mar. Environ. Res. 13: 177-194. ENGEL, D. W., AND M. BROUWER. 1984b. Cadmium-binding proteins in the blue crab, Callinectes sapi- dus: \aboratory-fie\d comparison. Mar. Environ. Res. 14: 139-151. ENGEL, D. W., AND M. BROUWER. 1986. Copper and cadmium metallothioneins in the American lobster, Homarus americanus. Environ. Health Perspect. 65: 87-92. ENGEL, D. W., AND G. ROESIJADI. In press. Metallothioneins: a monitoring tool. In Physiology and Pollu- tion of Marine Organisms, F. J. Vernberg, F. P. Thurberg, A. Calabrese, and W. Vernberg, eds. University of South Carolina Press, Columbia. ENGEL, D. W., M. BROUWER, AND F. P. THURBERG. 1985. Comparison of metal metabolism and metal- binding proteins in the blue crab and American lobster. Pp. 229-245 in Marine Pollution and Physiology: Recent Advances, F. J. Vernberg, F. P. Thurberg, A. Calabrese, and W. Vernberg, eds. University of South Carolina Press, Columbia. FRAZIER, J. M., AND S. G. GEORGE. 1983. Cadmium kinetics in oysters, a comparative study of Crassos- trea gigas and Ostrea edulis. Mar. Biol. 76: 55-6 1 . GEORGE, S. G., AND A. VIARENGO. 1985. A model of heavy metal homeostasis and detoxification in mussels. Pp. 125-143 in Marine Pollution and Physiology: Recent Advances, F. J. Vernberg, F. P. Thurberg, A. Calabrese, and W. Vernberg, eds. University of South Carolina Press, Colum- bia. HAGERMAN, L. 1983. Hemocyanin concentration of juvenile lobsters (Homarus gammarus) in relation to molting cycle and feeding conditions. Mar. Biol. 77: 11-17. HENRY, R. P., AND G. A. KORMANIK. 1985. Carbonic anhydrase activity and calcium deposition during the molt cycle of the blue crab, Callinectes sapidus. J. Crustacean Biol. 5: 234-24 1 . JOHNSON, B. A., C. BONA VENTURA, AND J. BONAVENTURA. 1984. Allosteric modulation of Callinectes sapidus hemocyanin by binding of L-lactate. J. Biochem. 23: 872-878. KOJIMA, Y., AND J. H. R. KAGI. 1978. Metallothionein. Trends Biochem. Sci. 3: 90-93. Li, T. Y., A. J. KRAMER, C. F. SHAW, III, AND D. H. PETERING. 1980. Ligand substitution reactions of metallothionein with EDTA and apo-carbonic anhydrase. Proc. Natl. Acad. Sci. USA 77: 6334- 6338. MANGUM, C. P. 1985. Molting in the blue crab Callinectes sapidus: a collaborative study of intermediary metabolism, respiration and cardiovascular function, and ion transport. Preface. J. Crustacean Biol. 5: 185-187. MANGUM, C. P., B. R. MCMAHON, P. L. DEFUR, AND M. G. WHEATLY. 1985. Gas exchange, acid-base balance, and the oxygen supply to the tissues during a molt of the blue crab, Callinectes sapidus. J. Crustacean Biol. 5: 188-206. MARGOSHES, M., AND B. L. VALLEE. 1957. A cadmium protein from the equine kidney cortex. / Am. Chem.Soc. 79:4813-4814. MARTIN, J.-L. M. 1975. Le cuivre et le zinc chez Cancer irroratus (Crustace: Decapode): metabolisme compare au cours du cycle d'intermue. Comp. Biochem. Physiol. 51 A: 777-784. NORDBERG, M. AND Y. KOJIMA. 1979. Metallothionein and other low molecular weight metal-binding proteins. Pp. 41-135 in Metallothionein, J. H. R. Kagi, and M. Nordberg, eds. Birkhauser Verlag. Basel. OLAFSON, R. W., A. KERNS, AND R. G. SIMS. 1979a. Heavy metal induction of metallothionein synthesis in the hepatopancreas of the crab, Scylla serrata. Comp. Biochem. Physiol. 62B: 417-424. 82 D. W. ENGEL OLAFSON, R. W., R. G. SIM, AND K. G. BOTO. 1979b. Isolation and chemical characterization of the heavy metal-binding protein metallothionein from marine invertebrates. Comp. Biochem. Physiol. 62B: 407-416. OVERNELL, J. 1982. Copper metabolism in crabs and metallothionein: in vivo effect of copper II on soluble hepatopancreas metal binding components in the crab Cancer pagurns containing various amounts of cadmium. Comp. Biochem. Physiol. 73B: 555-564. OVERNELL, J. 1 986. The occurrence of cadmium in crabs (Cancer pagnrus) and the isolation and properties of cadmium metallothionein. Environ. Health Perspect. 65: 101-106. OVERNELL, J., AND E. TREWHELLA. 1979. Evidence for natural occurrence of (Cadmium, Copper)— metallothionein in the crab, Cancer pagurns. Comp. Biochem. Physiol. 64C: 69-76. PASSANO, L. M. 1 960. Molting and its control. Pp. 473-536 in The Physiology of Crustacea, Vol. 1 , T. H. Waterman, ed. Academic Press. New York. RAINBOW, P. S., AND A. G. SCOTT. 1979. Two heavy metal-binding proteins in the midgut gland of the crab Carcinus maenas. Mar. Biol. 55: 143-150. ROESIJADI, G. 1981. The significance of low molecular weight, metallothionein-like proteins in marine invertebrates: current status. Mar. Environ. Res. 4: 167-179. SENK.BEIL, E. G., AND J. C. WRISTON, JR. 1981. Catabolism of hemocyanin in the American lobster, Homarus americanus. Comp. Biochem. Physiol. 69B: 781-790. SKINNER, D. M. 1985. Molting and regeneration. Pp. 43-146 in The Biology of Crustacea, volume 1, Integument, Pigments, and Hormonal Processes. D. E. Bliss and L. H. Mantel, eds. Academic Press, Inc. New York. TOWLE, D. W., AND C. P. MANGUM. 1985. Ion regulation and transport ATPase activities during the molt cycle in the blue crab Callinectes sapidus. J. Crustacean Biol. 5: 2 1 6-222. UDOM, A. O., AND F. O. BRADY. 1980. Reactivation in vitro of zinc-requiring apo-enzymes by rat liver zinc-thionein. Biochem. J. 187: 329-335. WEIDOW, M. A., T. J. KNEIP, AND S. J. GARTE. 1982. Cadmium-binding proteins from the blue crab (Callinectes sapidus) environmentally exposed to cadmium. Environ. Res. 28: 1 64- 1 70. WHEATLY, M. G. 1985. Free amino acid and inorganic ion regulation in the whole muscle and hemolymph of the blue crab Callinectes sapidus Rathbun in relation to the molting cycle. J. Crustacean Biol. 5: 223-233. ZATTA, P. 1984. Zinc transport in the haemolymph of Carcinus maenus (Crustacea: Decapoda). J. Mar. Biol. Assoc. U.K. 64: 801-807. Reference: Biol. Bull. 172: 83-88. (February, 1987) LOSS AND RECOVERY OF LOCOMOTOR BEHAVIOR AFTER CNS LESIONS IN THE SNAIL MELAMPUS BIDENTATUS KEITH SNYDER AND STACIA MOFFETT Department of Zoology, Washington State University, Pullman, Washington 99164-4220 ABSTRACT To analyze neural pathways underlying the control of locomotion in the snail Melampus bidentatus, we lesioned connectives and commissures involved in loco- motor control. The results suggest the presence of an independent oscillator in each pedal ganglion which is connected primarily via the pedal commissure. The bilater- ally paired cerebropedal and cerebropleural connectives carry redundant information concerned with initiation of locomotion. In the 2-4 weeks after lesions were made, the snails usually regained locomotor coordination. INTRODUCTION Gastropod molluscs are favorable models for investigations of neural control of locomotion because of the relative simplicity of their nervous systems. Several fea- tures common to locomotion of many opisthobranch and pulmonate gastropods have been discovered using CNS lesions. These include: ( 1 ) command cells in cere- bral ganglia that initiate and maintain the locomotor behavior, and (2) pattern gener- ators that organize rhythmic output to muscles. Command centers have been found in cerebral ganglia and pattern generators in the pedal ganglia (Fredman and Jahan- Parwar, 1980, 1983; Jahan-Parwar and Fredman, 1980; Lennard et al, 1980; von der Portend al. , 1982; Getting, 1983; Arshavsky et al., 1985a; Satterlie, 1985). The use of CNS lesions also revealed the gastropod potential for neural repair and associated recovery of specific behaviors (Price, 1977; Arshavsky et al., 1985b; Moffett and Snyder, 1985). The pulmonate gastropod Melampus locomotes by means of a "crawl-step" consisting of several discrete components which make it particularly suitable for behavioral analysis after CNS lesions (Moffett, 1979). The alteration of this behavior following CNS lesions has allowed us to explore the role of particular parts of the nervous system in generation and coordination of locomo- tion. The sequence and timecourse of recovery from the lesions were also recorded. MATERIALS AND METHODS Animals and surgical techniques Snails (Melampus bidentatus} with shell lengths of 6-10 mm were obtained from Wollaston, Massachusetts, and Poquoson, Virginia. Animal maintenance, anesthe- sia, and surgical techniques were as described by Moffett and Snyder ( 1 985). Lesions The primitively unfused nervous system of Melampus contains five paired ganglia and one unpaired visceral ganglion. Five types of lesions were performed: ( 1 ) transec- Received 25 June 1986; accepted 28 October 1986. 83 84 K. SNYDER AND S. MOFFETT tion of the pedal commissure, (2) transection of both cerebral and pedal commissures, (3) bilateral transection of cerebropedal and cerebropleural connectives and statocyst nerve, (4) isolation of the left pedal ganglion by cutting the left cerebropedal and cerebropleural connectives, the statocyst nerve, and the pedal commissure, and (5) excision of the left pedal ganglion (Fig. 1). Behavioral analysis The locomotor behavior of each snail was tested 4 to 7 days after surgery and every 2-4 days thereafter. The snail was placed on a glass plate clamped above an angled mirror which allowed us to view and videotape pedal movements. We deter- mined what portion (%) of the foot surface functioned in locomotion by superimpos- ing a grid on the image of the foot. Even when their crawl was abnormal due to lack of coordination or failure of regions of the foot to participate, snails were able to move along the substratum. Behavioral deficits fell into two categories: ( 1 ) failure of musculature in one region of the foot to participate in locomotion (a musculature control deficit), or (2) a deficit in coordination of locomotor movements in different parts of the foot, such as asym- metrical progression of the pedal wave on the right and left sides of the foot. All observations were made by an experimenter who did not know which surgery the snail had undergone. RESULTS Cerebropedal and cerebropleural connectives cut When cerebropedal and cerebropleural connectives were severed bilaterally (Fig. IB), snails withdrew into their shells upon recovery from anesthesia. About half did not emerge until they could crawl normally. Although these animals were the slowest of all experimental groups to initiate locomotion after surgery (Fig. 2), they exhibited the fastest recovery of normal pedal waves (Fig. 3). Most deficits in metapodial sym- metry in this group involved failure to activate musculature in the posterior third of the foot. They also exhibited abnormal coordination of the propodial region (Fig. IB). Pedal commissure cut All 1 1 snails which had their pedal commissures cut were crawling by day 7 (Fig. 2). All showed lack of control over the posterior portion of the metapodium on one side (5 animals) or both sides (6 animals), but involvement of the entire foot in loco- motion was gradually re-established (Fig. 3). After the entire metapodial musculature became involved in locomotion, the crawl-step was still initiated asymmetrically by contraction of the foot tip on the right or left (Fig. 1C). Once the tip shortened on one side, the pedal wave continued forward either symmetrically or with the initiating side leading. Of 1 1 snails, 4 regained normal locomotion within 32 days. Pedal and cerebral commissure cut When the cerebral commissure was severed in addition to the pedal commissure, side-to-side coordination of pedal locomotion was completely disrupted (Fig. 1 D). This surgery delayed the initiation of spontaneous crawling behavior longer than pedal commissurotomy alone (Fig. 2). The time required to re-establish control over SNAIL LOCOMOTION AFTER CNS LESIONS Control Locomotion 85 O P M Lesions :•->•, \J \J P V Illlllll V Illlllll \) \] V V D F-. FIGURE 1 . Locomotor deficits resulting from CNS lesions. Each line depicts a complete cycle of the crawl-step. Regions on the foot marked by parallel lines are contracting. (A) Control CNS and correspond- ing locomotion. (O = oral veil, P = propodium, M = metapodium). (B) Right and left cerebropedal and cerebropleural connectives cut. Propodium is often elevated but metapodial wave is normal when snails first locomote. (C) Pedal commissure cut. Wave initiated asymmetrically. (D) Cerebral and pedal commis- sures cut. Waves on right and left sides independent. (E) Left pedal ganglion isolation and ablation. Each wave pulls animal toward ablated side, requiring continuous course correction. Arrows indicate initiation of pedal wave. 86 K. SNYDER AND S. MOFFETT 100 O 2 5 CO CO 75 50 25 0 R & L Conn. Cut • PdC Cut • PdC & CC Cut o LPdG Isolated » LPdG Ablated 8 10 12 14 DAYS AFTER SURGERY FIGURE 2. Onset of initial attempts at locomotion in groups of snails recovering from each of the five experimental lesions. Triangle = right and left cerebropleural and cerebropedal connectives cut. Square = Pedal commissure cut. Circle (filled) = pedal and cerebral commissures cut. Circle (open) = left pedal ganglion isolated. Diamond = left pedal ganglion ablated. pedal musculature was similar for snails with pedal commissurotomies and those with pedal plus cerebral commissurotomies (Fig. 3). Animals with both commissures transected initially showed less propodial expansion. Recovery began with the ante- rior portion of the foot and progressed posteriorly. UJ 100 90 CO UJ <~ < * lil 80 70 1 60 O O < 50 A R & L Conn. Cut • PdC Cut • PdC & CC Cut o LPdG Isolated • LPdG Ablated 10 20 30 40 DAYS AFTER SURGERY 50 60 FIGURE 3. Recovery of metapodial wave symmetry in the five experimental groups. Symbols and abbreviations given in Figure 2. SNAIL LOCOMOTION AFTER CNS LESIONS 87 Left pedal ganglion isolated The ganglion was held in place by its pedal nerves after its central connections were cut. The right side of the foot functioned normally, but the left half of the meta- podium failed to participate in locomotion. One animal also exhibited abnormal co- ordination of a small portion of the right posterior metapodium. The six snails that received this lesion were slower to begin crawling than snails with previously de- scribed operations (Fig. 2). The left propodium and the anterior-most portion of the metapodium were the areas in which recovery was first seen. Control over pedal mus- culature was regained in an anterior to posterior progression, and complete pedal waves were seen on the operated left side in all 6 animals by day 40. Left pedal ganglion ablated The most severe operation performed in terms of CNS damage was excision of an entire pedal ganglion (Fig. 1 E). Behavior was similar to that following ganglion isolation. All 25 snails were crawling by day 5, although pedal musculature on the left side was inactive and the right posterior metapodial tip was also inactive in one- fourth of the snails. Recovery resembled that following left pedal ganglion isolation, except that it progressed more slowly and only 5 of 25 snails exhibited complete recovery (Fig. 3). The remaining snails all had some region on the left side of the foot that failed to be activated when tested on postoperative day 60. DISCUSSION Neurons that carry controlling information from cerebral ganglia to pedal ganglia have been classified as "command" neurons in Aplysia (Jahan-Parwar and Fredman, 1978), Pleurobranchaea (Gillette et al, 1978), and Lymnaea (McCrohan, 1984). When both cerebropleural connectives were cut in A. California, normal crawling continued and only the escape locomotion was abolished (Jahan-Parwar and Fredman, 1979). If both cerebropedal connectives were cut, only limited movements were generated; locomotion stopped entirely when all four connectives were severed. Transection of cerebropedal connectives only slows the locomotor rate in Clione (Sat- terlie and Spencer, 1985). In Melampus, signals travelling through either the cere- bropleural or cerebropedal connectives are sufficient to initiate locomotion (Snyder, 1986). In the work reported here, snails with bilateral connective transections did not emerge from their shells and crawl as early as the other groups. The timescale for their onset of locomotion could have allowed for connective regeneration (Sny- der, 1986). In addition to interactions between cerebral and pedal levels, side-to-side coordi- nation is important for locomotion. When the pedal commissure was cut in Clione or Aplysia, the right and left sides of the animal moved asynchronously (Jahan- Parwar and Fredman, 1980; Arshavsky et al, 1985a; Satterlie and Spencer, 1985). In Melampus, cutting the cerebral commissure alone had no detectable effect on loco- motion (Moffett and Snyder, 1985), whereas cutting the pedal commissure alone was almost as devastating as cutting both commissures. Ablation of a pedal ganglion results in the same central lesions as pedal ganglion isolation, but also includes loss of a large population of neurons. The degree of recov- ery we observed following this lesion was therefore quite remarkable, especially con- sidering that the recovery occurred before new neurons are likely to have been gener- ated (Snyder, 1986). For all operations, the rate of behavioral recovery was rapid, considering distance between ganglia and the fact that we had cut neural trunks rather 88 K. SNYDER AND S. MOFFETT than crushing them, so growing axons had no mechanical linkage to follow in finding their targets. Only animals which had the left pedal ganglion removed failed to regain complete control over pedal musculature. ACKNOWLEDGMENTS We wish to acknowledge NSFs MARC Program support of Beverly D. Living- ston, who assisted us in data collection, support of S. M. by R01 NS22896 and K. S. by Sigma Xi Grant-in-Aid of Research and a Grant-in-Aid for Graduate Research from WSU. LITERATURE CITED ARSHAVSKY, Y. I, I. N. BELOOZEROVA, G. N. ORLOVSKY, Y. V. PANCHIN, AND G. A. PAVLOVA. 1985a. Control of locomotion in marine mollusc Clione limacina II. Rhythmic neurons of pedal ganglia. Exp. Brain Res. 58: 263-272. ARSHAVSKY, Y. I, I. N. BELOOZEROVA, G. N. ORLOVSKY, Y. V. PANCHIN, AND G. A. PAVLOVA. 1985b. Control of locomotion in marine mollusc Clione limacina I. Effective activity during actual and fictitious swimming. Exp. Brain Res. 58: 255-262. FREDMAN, S. M., AND B. JAHAN-PARWAR. 1980. Role of pedal ganglia motor neurons in pedal wave generation in Aplysia. Brain Res. Bull. 5: 179-193. FREDMAN, S. M., AND B. JAHAN-PARWAR. 1983. Command neurons for locomotion in Aplysia. J. Neuro- phys.49: 1092-1117. GETTING, P. A. 1983. Mechanisms of pattern generation underlying swimming in Tritonia. II. Network reconstruction. J. Neurophys. 49: 1017-1035. GILLETTE, R., M. P. KOVAC, AND W. J. DAVIS. 1978. Command neurons in Pleurobranchaea receive synaptic feedback from the motor network they excite. Science 199: 798-801. JAHAN-PARWAR, B., AND S. M. FREDMAN. 1978. Control of pedal and parapodial movements in Aplysia. II. Cerebral ganglion neurons. J. Neurophys. 41: 609-620. JAHAN-PARWAR, B., AND S. M. FREDMAN. 1979. Role of interganglionic synaptic connections in the control of pedal and parapodial movements in Aplysia. Brain Res. Bull. 4: 407-420. JAHAN-PARWAR, B., AND S. M. FREDMAN. 1980. Motor program for pedal waves during Aplysia locomo- tion is generated in the pedal ganglia. Brain Res. Bull. 5: 169-177. LENNARD, P. R., P. A. GETTING, AND R. I. HUME. 1980. Central pattern generator mediates swimming in Tritonia. II. Initiation, maintenance, and termination. J. Neurophys. 44: 165-173. McCROHAN, C. R. 1984. Initiation of feeding motor output by an identified interneurone in the snail Lymnaea stagnalis. J. Exp. Biol. 1 13: 35 1 -366. MOFFETT, S. 1979. Locomotion in the primitive pulmonate snail Melampus bidentatus: foot structure and function. Biol. Bull. 157: 306-319. MOFFETT, S., AND K. SNYDER. 1985. Behavioral recovery associated with the central nervous system regeneration in the snail Melampus. J. Neurobiol. 16: 193-209. VON DER PORTEN, K., D. W. PARSONS, B. S. ROTHMAN, AND H. PINSKER. 1982. Swimming in Aplysia brasiliana: analysis of behavior and neuronal pathways. Behav. Neural Biol. 36: 1-23. PRICE, C. H. 1977. Regeneration in the central nervous system of a pulmonate mollusc, Melampus. Cell Tiss. Res. 180: 529-536. SATTERLIE, R. A. 1985. Reciprocal inhibition and postinhibitory rebound produce reverberation in a locomotor pattern generator. Science 229: 402-404. SATTERLIE, R. A., AND A. N. SPENCER. 1985. Swimming in the pteropod mollusc, Clione limacina. II. Physiology. J. Exp. Biol. 116: 205-222. SNYDER, K. A. 1986. Regeneration and locomotor behavior after CNS lesions in Melampus bidentatus. Ph.D. Dissertation, Washington State University, Pullman. Reference: Biol. Bull. 172: 89-107. (February, 1987) THE CONTINUING ENIGMA OF CIGUATERA* DONALD M. ANDERSON AND PHILLIP S. LOBEL Biology Department, Woods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543 ABSTRACT Research on ciguatera fish poisoning has expanded significantly over the last de- cade. In large part, this increase in effort is due to the identification of several benthic dinoflagellates as the toxin producers, a discovery soon followed by a series of field and laboratory studies on their distribution, abundance, growth characteristics, and toxin production. Equally important have been advances in the analytical techniques and equipment needed to chemically characterize the toxins. Much of that work ben- efited significantly from the rapid progress in chemical research on the numerous other toxins produced by marine dinoflagellates. Despite this surge in activity (summarized in the proceedings of four recent con- ferences or workshops: Ragelis, 1984;Salvat, 1985; Anderson el al, 1985; this issue), the general state of knowledge on ciguatera remains relatively poor, both in terms of toxin chemistry and the physiological ecology of the causative dinoflagellates. Some important generalizations are gaining acceptance, but discrepancies and disagree- ments abound. One of the objectives of this review is to place the many recent papers on ciguatera in a current perspective that not only identifies common observations or conclusions, but also accentuates those areas that require more research effort to resolve disagreements or contradictions. INTRODUCTION In many tropical regions, it has long been known that consumption of certain coastal marine fishes can cause human illness and occasional death. The name "ci- guatera" was given to this phenomenon by the Spanish, based on the belief that a marine turban snail (called "cigua" in the Caribbean) was responsible for poisoning settlers in Cuba. Reports of similar fish poisoning in the Pacific date back to the early 1 7th century (Banner, 1 976). Today the term "ciguatera" refers to intoxications resulting from the ingestion of tropical and subtropical finfish, distinct from hista- minic poisonings or those associated with the pufferfish (Halstead, 1967). Morbidity statistics are highly unreliable due to the tendency of many individuals not to report such illnesses, the wide geographic distribution of many islands where the problem is endemic, and the variability in symptomology. Mean annual inci- dence of reported cases in the Pacific island region (excluding Hawaii and Australia) is about 1-4 cases per thousand population (Lewis, 1984; Yasumoto et al., 1984). In the Caribbean, 4.2 cases per thousand were reported from St. Thomas in the Virgin Islands (Olsen et al., 1984). These values can be scaled up by factors of 2-5 using estimates of the fraction of poisonings that are never reported. The resulting estimates and those from other affected areas indicate that ciguatera has been responsible for far more cases of human illness over the past eight years than any other kind of sea- food toxicity associated with consumption of fresh marine organisms (Ragelis, 1984). As many as 10,000-50,000 individuals may be poisoned by ciguatera annually. * For reprints of the entire ciguatera workshop proceedings contact: New England Biolabs Founda- tion, P.O. Box 413, Wenham, MA 01984. Those interested in reprints of individual papers should contact the author directly. 89 90 D. M. ANDERSON AND P. S. LOBEL Although rarely fatal, the intoxications can be extremely debilitating and in some cases can recur sporadically for years after the initial poisoning. However, the most important impact of ciguatera may well lie in its effects on small-scale fisheries for local consumption and for export. This is especially serious in those poor or densely populated islands where fish traditionally have been a primary source of protein. Reviews by Lewis (1984) and Olsen et al. (1984) examined the impacts of ciguatera on marine resource development in the Pacific and Caribbean regions, respectively. In both areas, ciguatera is considered one of the most important constraints to fisher- ies resource development, second only perhaps to the inadequate size of those re- sources relative to the additional demands expected with future population and eco- nomic growth. The symptoms of ciguatera poisoning have been described in great detail in nu- merous publications (Bagnis, 1 968; Bagnis /. 4: 28 1-288. OHIZUMI, Y. 1987. Pharmacological actions of the marine toxins ciguatoxin and maitotoxin isolated from poisonous fish. Biol. Bull. 172: 132-136. OHIZUMI, Y., M. KOBAYASHI, A. KAJIWARA, AND T. YASUMOTO. 1985. Potent excitatory effects of maito- toxin on cardiac and smooth muscle. Pp. 369-374 in Toxic Dinoflagellates, D. M. Anderson, A. W. White, and D. G. Baden, eds. Elsevier, New York. OLSEN, D. A., D. W. NELLIS, AND R. S. WOOD. 1984. Ciguatera in the eastern Caribbean. Mar. Fish. Rev. 46: 13-18. PINCEMIN, J. M. 1971. Telemediateurs chimiques et equilibre biologique oceanique. Troiseme partie. Etude in vitro de relations entre populations phytoplanctoniques. Rev. Int. Oceanogr. Med. 23: 165-196. PRINGSHEIM, E. G. 1912. Die Kultur de Algen. Agar. Bietr. Biol. PJlanzll: 305-333. PROVASOLI, L., J. J. A. MCLAUGHLIN, AND M. R. DROOP. 1957. The development of artificial media for marine algae. Archiv. Mikrobiol. 25: 392-428. RAGELIS, E. P., ed. 1984. Seafood Toxins. Am. Chem. Soc. Symp. Ser. No. 262, Washington, DC. 460 pp. RANDALL, J. E. 1955. A revision of the surgeonfish genus Ctenochaetus, family Acanthuridae, with descrip- tions of five new species. Zoologica 40: 149-166. RANDALL, J. E. 1958. A review of ciguatera, tropical fish poisoning, with a tentative explanation of its cause. Bull. Mar. Sci. Gulf. Caribb. 8: 236-267. RANDALL, J. E. 1961. A contribution to the biology of the convict surgeonfish on the Hawaiian Islands, Acanthurus triostegus sandvicensis. Pac. Sci. 15: 2 1 5-272. RANDALL, J. E. 1967. Food habits of reef fishes of the West Indies. Stud. Trap. Ocean. No. 5: 665-847. RANDALL, J. E. 1974. The effect of fishes on coral reefs. Proc. Sec. Int. Coral Reef Symp. 1: 1 59-166. RANDALL, J. E. 1980. A survey of ciguatera at Enewetabe and Bikini, Marshall Islands, with notes on the systematics and food habits of ciguateric fishes. Fish. Bull. U. S. 78(2): 201-249. RAYNER, M. D. 1970. Marine toxins from the Pacific VII: Recent advances in the pharmacology of cigua- toxin. Pp. 345-350 in Food-Drugs from the Sea, H. W. Youngken, ed. Proceedings, Marine Tech- nology Soc., Washington, DC. RAYNER, M. D., M. H. BASLOW, AND T. I. KOSAKJ. 1969. Marine toxins from the Pacific: not an in vivo anticholinesterase. J. Fish. Res. Board Can. 26: 2208-2210. RAYNER, M. D., AND T. I. KOSAKJ. 1970. Ciguatoxin: effects on Na fluxes in frog muscle. Fed. Proc. Fed. Am. Soc. Exp. Biol. 29: 548. SALVAT, B., ed. 1985. Proceedings of the 5th International Coral Reef Congress, Vol. 4, Antenne Mu- seum—EPHE. 598 pp. SCHEUER, P. J., W. TAKAHASI, J. TSUTSUM, AND T. YOSHIDA. 1967. Ciguatoxin isolation and chemical nature. Science 155: 1267-1268. SHIMIZU, Y., H. SHIMIZU, P. J. SCHEUER, Y. HOKAMA, M. OYAMA, AND J. T. MIYAHARA. 1982. Gam- bierdiscus toxicus, a Ciguatera-causing dinoflagellate from Hawaii. Bull. Jpn. Soc. Sci. Fish. 48: 811-813. STEIDINGER, K. A. 1983. A re-evaluation of toxic dinoflagellate biology and ecology. Pp. 147-188 in Progress in Phycological Research, Vol. 2, F. E. Round and V. J. Chapman, eds. Elsevier, New York. STEIDINGER, K. A., AND D. G. BADEN. 1 984. Toxic marine dinoflagellates, Pp. 20 1 -26 1 in Dinoflagellates, D. L. Spector, ed. Academic Press, Orlando. CIGUATERA FISH POISONING 107 SUNDA, W. G., AND R. R. L. GUILLARD. 1976. Relationship between cupric ion activity and the toxicity of copper to phytoplankton. J. Mar. Res. 34: 5 1 1-529. TACHIBANA, K. 1980. Structural studies on marine toxins. Ph.D. Thesis, University of Hawaii. TACHIBANA, K., M. NUKINA, Y. JOH, AND P. J. SCHEUER. 1987. Recent developments in the molecular structure of ciguatoxin. Biol. Bull. 172: 122-127. TAKAHASHI, M., Y. OHIZUMI, AND T. YASUMOTO. 1 982. Maitotoxin, a Ca2+ channel activator candidate. J. Biol. Chem. 257: 7287-7289. TAKAHASHI, M., M. TATSUMI, AND Y. OHIZUMI. 1983. Ca2+ channel activating function of maitotoxin, the most potent marine toxin known, in clonal rat pheochromocytoma cells. J. Biol. Chem. 258: 10944-10949. TAYLOR, F. J. R. 1985. The distribution of the dinoflagellate Gambierdiscus toxicus in the eastern Carib- bean. Pp. 423-428 in Proceedings of the Fifth International Coral Reef Congress, B. Salvat, ed. Antenne Museum — EPHE. TAYLOR, F. J. R., AND M. S. GUSTAVSON. In press. An underwater survey of the organism chiefly responsi- ble for "ciguatera" fish poisoning in the eastern Caribbean region: the benthic dinoflagellate Gam- bierdiscus toxicus. In Proceedings of the Seventh International Diving Science Symposium, A. Stefanon and N. J. Flemming, eds. TINDALL, D. R., R. W. DICKEY, R. D. CARLSON, AND G. MOREY-GAINES. 1984. Ciguatoxigenic dinoflag- ellates from the Caribbean Sea. Pp. 225-240 in Seafood Toxins, E. P. Ragelis, ed. Am. Chem. Soc. Sym. Ser. No. 262. Washington, DC. UCHIDA, T. 1977. Excretion of diatom inhibiting substance by Prorocentrum micans Ehrenberg. Jpn. J. Ecol.21:l-4. WITHERS, N. W. 1982. Ciguatera fish poisoning. Ann. Rev. Med. 33: 97-1 1 1 . YASUMOTO, T. 1978. Large-scale toxin production by Diolopsalis sp. — biochemical study of maitotoxin and ciguatoxin. So. Pac. Comm. Rep. 785: 7-8. YASUMOTO, T. 1985. Recent progress in the chemistry of dinoflagellate toxins. Pp. 259-270 in Toxic Dinoflagellates. D. M. Anderson, A. W. White, and D. G. Baden, eds. Elsevier, New York. YASUMOTO, T., N. SEINO, Y. MURAKAMI, AND M. MURATA. 1987. Toxins produced by benthic dino- flagellates. Biol. Bull. 172: 128-131. YASUMOTO, T., A. INOUE, R. BAGNIS, AND M. GARCON. 1979a. Ecological survey on a dinoflagellate possibly responsible for the induction of ciguatera. Bull. Jpn. Soc. Sci. Fish. 45: 395-399. YASUMOTO, T., I. NAKAJIMA, Y. OSHIMA, AND R. BAGNIS. 1 979b. A new toxic dinoflagellate found in association with ciguatera. Pp. 65-70 in Toxic Dinoflagellate Blooms, D. L. Taylor and H. H. Seliger, eds. Elsevier, New York. YASUMOTO, T., A. INOUE, T. OCHI, K. FUJIMOTO, Y. OSHIMA, Y. FUKUYO, R. ADACHI, AND R. BAGNIS. 1980. Environmental studies on a toxic dinoflagellate responsible for ciguatera. Bull. Jpn. Soc. Sci. Fish. 46: 1397-1404. YASUMOTO, T., I. NAKAJIMA, E. CHUNQUE, AND R. BAGNIS. 1977a. Toxins in the gut contents of a parrotnsh. Bull. Jpn. Soc. Sci. Fish. 43: 69-74. YASUMOTO, T., I. NAKAJIMA, R. A. BAGNIS, AND R. ADACHI. 1 977b. Finding of a dinoflagellate as a likely culprit of ciguatera. Bull. Jpn. Soc. Sci. Fish. 43: 102 1-1026. 0 YASUMOTO, T., U. RAJ, AND R. BAGNIS. 1984. Seafood Poisoning in Tropical Regions. Lab. of Food Hygiene, Fac. of Agric., Tohoku Univ., Japan. 74 pp. YASUMOTO, T., Y. HASHIMOTO, R. A. BAGNIS, J. E. RANDALL, AND A. H. BANNER. 1971. Toxicity of surgeon fishes. Bull. Jpn. Soc. Sci. Fish. 37: 724-734. Reference: Biol. Bull 172: 108-121. (February, 1987) STUDY OF PRODUCTION AND TOXICITY OF CULTURED GAMBIERDISCUS TOXICUS M. DURAND-CLEMENT Inserm Unite 303, "Meret Sante, "La Darse — B. P. 3, 06230 Villefranche-sur-Mer, France ABSTRACT Gambierdiscus toxicus has been described as the chief producer of ciguateric tox- ins: ciguatoxin (CTX) and maitotoxin (MTX). Dr. R. Bagnis (Malarde Institute, Ta- hiti) provided us with a strain of this benthic dinoflagellate in 198 1 to study its cyto- logical, physiological, and toxicological characteristics. The growth of G. toxicus has been studied under various chemical and physical conditions enabling us to define optimal culture conditions. Since then, we have im- proved the cultivation procedure and have obtained large scale cultures of some clones and strains in different culture media. All the experiments showed that G. toxicus has complex nutritive requirements and a large inertia to response at some non-drastic environmental variations. In- creased growth rates were observed when cultures were treated with antibiotics. The classical extraction procedure of the toxins has been adapted to our algal material. The method was simplified and resulted in enhanced toxinic yield. Our principal results have demonstrated that G. toxicus in culture remained very toxic; 600 to 2000 cells were sufficient to kill a 20 g female mouse within 24 hours (MLD = 1-3 mg/kg). No variation of the degree of toxicity has been observed for three years. A linear relationship between the number of G. toxicus cells (Coulter counted) and the weight of corresponding algal pellet has been found. This leads to an easier evaluation of the quantity of toxic extract required to calculate the minimum lethal dose (MLD). The extraction procedure results in two toxic fractions: a water-soluble (MTX- like) and a lipid-soluble (CTX-like). The latter corresponds to 10 to 25% of the total toxicity. A dose-time to death curve has been established with our CTX-like extracts. INTRODUCTION Most tropical and intertropical coral reef seas contain the benthic dinoflagellate, Gambierdiscus toxicus. From wild G. toxicus, Yasumoto et al. (1977) extracted the two main toxins involved in ciguatera, ciguatoxin (CTX) and maitotoxin (MTX), which are transmitted to fish through the marine food chain (Taylor, 1979; Bagnis, 1981; Shimizu et al., 1982; Withers, 1982). Although G. toxicus was the first dino- flagellate to be linked to the genesis of ciguateric toxins (Bagnis et al., 1977), other toxic dinoflagellates, which are also potential sources of ciguateric toxins, have been isolated from ciguateric areas: Prorocentrum lima in the Pacific Ocean (Yasumoto et al., 1984); and Prorocentrum mexicanum in the Atlantic Ocean (Tindall et al, 1984). A number of dinoflagellates associated with the benthic community of coral reefs may contribute to the complex syndrome of ciguatera fish poisoning (Fukuyo, 198 1 ; 108 GAMBIERDISCUS TOXICUS 109 Nakajima el al, 198 1 ; Steidinger and Baden, 1984), the biogenesis of which remains to be clarified. We adapted a strain of G. toxicus from the Gambier Islands (kindly provided by Dr. R. Bagnis) to culture conditions in our laboratory and obtained mass cultures in order to study the main physiological, cytological, and ultrastructural characteristics of this dinoflagellate (Durand, 1984). We also established the toxicity of the culture under varying environmental conditions. The aim of this paper is to describe the main characteristics of cultured G. toxicus. CULTURE CONDITIONS AND PHYSIOLOGY OF G. TOXICUS Temperature The temperature which supported optimum growth for our strains was 26 ± 1°C. Temperature over 30°C quickly killed the cells, and little growth was observed below 22°C. The temperature range for optimum growth was limited, probably due to the original collection site of the strains (the Gambier Islands), where little variation in seawater temperature occurs. Thus, our G. toxicus strains could be different from those found by Besada et al. ( 1 982) in the Windley Keys (Florida), providing evidence of two distinct ecotypes. However, after numerous transfers over one year, we obtained a strain of G. toxi- cus with satisfactory growth at 20°C. Light By modifying light intensity and photoperiod we demonstrated that the growth rate of G. toxicus was largely influenced by the total amount of illumination received per day (Durand and Puiseux-Dao, 1985). The best conditions were a light intensity of 10 W-irT2 (tubes: Philips TL 65 W and Mazda Fluor TF 65 W) and a light/dark cycle of 10 h/ 14 h. Only a few divisions of G. toxicus were recorded under continuous light, suggesting that G. toxicus, like most dinoflagellates (Loeblich, 1966), has alter- nating light/dark dependence for division. Preliminary experiments showed that G. toxicus cultures can be synchronized by modifying the light/dark alternation as de- scribed for Amphidinium carterae (Galleron, 1976). The low light tolerance of G. toxicus could be related to its epiphytic habits (macroalgal substrates) in nature. G. toxicus exhibited horizontal phototactic migrations at the bottom of the cul- ture flasks in response to light intensity. Cells moved toward the light source in low- light conditions and moved in the opposite direction under high-light intensity. This behavior of G. toxicus in culture demonstrates its great sensitivity to light. We never observed eyespot or stigma structures with transmission optical studies. Culture flasks, maintenance of the cultures, cell counting Clonal cultures were initiated by isolating a single cell under the microscope. Stock cultures were maintained in 250 ml Erlenmeyer flasks. Experiments were conducted in 125 or 250 ml flasks (initial inoculate: 30-50 cells/ml) containing 70- 80 and 1 50-1 70 ml of medium. For large cultures, G. toxicus was cultivated in 1 liter Fernbach flasks or 2 1 Erlenmeyer flasks, since we were able to obtain a greater bio- mass production of this benthic dinoflagellate in these vessels than in large glass car- boys. The cultures were harvested three or four weeks after the initial inoculation. At that time, G. toxicus cells overloaded the bottom of the flasks forming a unique algal 110 M. DURAND-CLEMENT ceSI./ml J 1500 1000 500 1 CAP 10 mgjhil FIGURE 1 . Effect of chloramphenicol (CAP) on Gambierdiscus toxicus growth in MPP. Strain previously treated with CAP. Strain without previous CAP treatment. layer. Cell density reached 4000 cells/ml (algal biomass equals 2 to 4- 106 cells/ 1 liter Fernbach flask). Agitation of the cultures inhibited G. toxicus growth. Just prior to stationary growth, cultures were harvested and washed by centrifuga- tion(1500Xg, 5min). Influence of bacteria on G. toxicus growth Decreasing the bacterial flora improved the growth of G. toxicus cultures in our laboratory, from 500-1000 cells/ml final algal yield, to 4000 cells/ml. We used two methods for this purpose: serial washing in sterile medium and antibiotic treatment. To select the most efficient antibiotics, we tested the cultures with antibiograms (mul- tidisk Sobioda 12GR422), then treated the cultures with either a large dose of antibi- otics (5-20 mg/1) for a few hours in darkness (after which the drug was washed out), or with a smaller amount (0. 1-1 mg/1) added directly to the culture medium. In the first case, bacterial spore germination was promoted by adding 0. 1% neopeptone one or two days prior to treatment. The best results were obtained with rifampicine, mino- cycline, and chloramphenicol at 5 mg/1 for 24 h. (See Fig. 1 for an example of the effect of chloramphenicol on G. toxicus.} The number of bacterial species was de- creased, from more than 50 at the origin, to 3-5. We obtained a batch in which no bacteria could be detected during two months of growth. Presently, the use of chloramphenicol (5 mg/1) results in less than one bacterium per G. toxicus cell. Such cultures produce only a small amount of mucilage relative to contaminated cultures. Culture media and nutrient supplementation assays Adjusting media to study the nutritional requirements of G. toxicus and to en- hance its growth was difficult because of its slow response to environmental modifi- GAMBIERDISCUS TOXICUS 1 1 1 cations. In addition, this large dinoflagellate showed a slow but efficient ability to adapt to various conditions. We tried to cultivate G. toxicus in two different media, enriched seawater and artificial medium. Seawater media. G. toxicus growth in seawater (from Banyuls/sur Mer, France) was slow (generation time: more than six days). For two months, the cells became progressively less pigmented and died. In Lateur medium (Lateur, 1963: autoclaved seawater supplemented with NaNO3, Na2HPO4, and soil extract), the growth remained slow (generation time: about five to six days) but the cultures did not degenerate. One of the first media (Hurtel et a!., 1979) described for G. toxicus culturing was MPP, according to Provasoli and Pintner (Provasoli, 1958: autoclaved seawater sup- plemented with various mineral compounds and vitamins). In this medium, G. toxi- cus growth was satisfactory (division time: about three days). However, growth im- proved when we decreased the amount of enrichment solution (ES) from 2% to 1 .5% and sometimes even 1%. When Tris (5 mg/1) was added, pH was adjusted to 8. 1 5. No modification of G. toxicus growth was detected in cultures not supplemented with vitamins for six months. Artificial media. The first cultures that we grew in ESAW medium (Harrison et al, 1980) showed good growth for eight months and then degenerated suddenly, as reported by Carlson et al. (1984). Similar results ensued from a second attempt of cultures transferred in ESAW. Little amelioration of the algal growth could be ob- tained in this medium, even by removing NaF, silica, and SrCl or by the use of ES supplementation of MPP. Thus, ESAW is not suitable for our G. toxicus cultures. In Shepard's artificial medium (MS: Shepard, 1969), G. toxicus growth was slow and irregular for many months (division time: five to eight days). No promotion of growth was obtained by modifying the salinity or the phosphorus concentration. In contrast, some growth stimulation occurred with the addition of 0.5% of ES (from MPP). At the moment, a strain that we progressively adapted to MS medium shows good growth (division time: four to five days) in this medium. At present, large cultures of G. toxicus grow with a division time of two to four days in a mixture of MPP and MS (1:1). Nutrient supplementation assays. For cultures grown in MPP medium, the follow- ing nutrient supplements did not result in a significant and reproducible improve- ment of culture yield: phosphates (0.1-1 mM), carbonates (0.5 mM), Ca2+, Mg2+, K+ (0.1-1 mM), and glucose ( 1 g/1). We tested various nitrogen sources to evaluate the nitrogen requirements of G. toxicus. The results are summarized below: ( 1 ) Mineral nitrogen: a weak temporary increase of the division rate of G. toxicus was obtained by adding NH4C1 (below 0.5 mM) and NaNO3 (5 mM), but these salts became toxic after a few days. (2) Organic nitrogen: no amino acids tested (0.1 mM) stimulated G. toxicus growth. Addition of urea resulted in an increase of cell yield as shown in Figure 2: after one week, G. toxicus growth was enhanced by 0.5 mMurea, but this concentra- tion became toxic; after three weeks, the largest biomass was obtained with 0. 1 mM urea. We noticed that this stimulatory effect was more efficient while the cells were in active growth phase, suggesting that the algal metabolism must be high for urea to be used by G. toxicus. (3) So/7 extract: a variable but constant improvement of G. toxicus cultures was obtained by adding soil extract to MPP. We saw the following increases of the algal 112 M. DURAND-CLEMENT ceil.lml I >0( 1000 500 .urea 0 0.01 0.1 0.5 1 5 10 FIGURE 2. Effects of urea adjunction on Gambierdiscus toxicus growth in MPP medium. to + 7 days. to + 2 1 days. biomass 20 days after inoculation: X 1.8 for 1 g/1 soil extract, X2.4 for 2 g/1, X2 for 4 g/1 (toxic effect at higher doses). Other workers also obtained this stimulatory effect of soil extract (Carlson et ai, 1984; Yasumoto £•/«/., 1984). (4) Dilution of MPP: diluting MPP with 5% distilled water stimulated G. toxicus growth (XI. 5). This could be related to both a decrease of salinity and a dilution of toxic nutrients. All the assays concerning the nutritional requirements for G. toxicus should be done again in MS medium (with the strain adapted to this medium) which has a well- defined chemical composition in contrast to seawater. G. toxicus possesses a complex physiology, and its study is complicated by the slow division rate and large size of this dinoflagellate, which allow slow responses to non-lethal variations of environmental parameters. In addition, analysis must consider the fact that cultured algae may pre- sent various adaptative behaviors, explaining the differences observed by different workers. PIGMENT COMPOSITION Studies to determine the pigment composition of G. toxicus showed that this dino- flagellate contains unusual pigments (Durand and Berkaloff, 1 985): it contains peridi- nin as the major carotenoid, and both chlorophylls Ci and c2. Peridinin is usually found only in association with chlorophyll c2 (Jeffrey, 1976), and such pigment com- position has been documented only in Prorocentrum cassubicum. MORPHOLOGICAL AND STRUCTURAL FEATURES OF G. TOXICUS We observed cultured G. toxicus with both light and electron microscopes (trans- mission TEM and scanning SEM). A Zeiss standard WL microscope equipped with Nomarski optics was used for light microscopy. GAMBIERDISCUS TOXICUS 1 1 3 For scanning electron microscopy, G. toxicus algal cultures were first pre-fixed by osmium vapors for 10 min, and then fixed by adding osmium (1% for 1 h at room temperature) to the culture medium. After centrifugation, the cell pellet was rinsed with distilled water, dehydrated in an acetone series, and dried using the CO2 critical- point method. The cells were then coated with gold and examined in a Cambridge 600 SEM. In the cultures, as a function of medium composition and light conditions, differ- ent morphological aspects of G. toxicus could be observed: typical armored cells, deformed cells, small motile cells and cysts. Typical cells Figures 3a and b show typical G. toxicus cells. SEM observations were made to define accurately the thecal plate features and ornamentation of this dinoflagellate (Figs. 4a, b). Morphology and thecal plate organization were similar to that reported for wild cells by Adachi and Fukuyo (1979) and by Besada et al. (1982). The mean dorso- ventral diameter, for our cells, was 70 ^m, with a range of 60-90 nm (30-50 ^m for the vertical diameter). G. toxicus possesses the two typical flagella of dinofla- gellates: a longitudinal one beating in the culture medium (about 50 p.m length; see Fig. 3a) and a transversal one. The latter is located in the cingulum and has regular, jerky movements that substantially mixes the medium surrounding the cell. Cell coloration varied, depending on light intensity and culture medium: G. toxi- cus cells appeared orange-brown in seawater, light brown in MPP, and dark brown in MS. Theca. The G. toxicus cell covering consists of a thick theca (about 2 nm as seen in TEM), or amphiesma (Loeblich, 1970), which limits the cytoplasm. This covering is composed of an outer membrane, a well-developed plate layer, a dense pellicle, and a double membrane. The presence of the pellicle has been described by Morill and Loeblich (1981), who used chloral hydratehydriodic acid-iodine staining. This theca is perforated with numerous trichocystic pores. Nucleus. The G. toxicus nucleus is large (30 /xm; see Fig. 3d). It is easily ejected, intact, when the cell is pressed between a slide and a coverslip (Fig. 3c). It contains numerous permanently condensed chromosomes (Fig. 3e) and a large amount of DNA (162 pg/cell). Various nuclear stains revealed, in addition to this dinocaryon, numerous unusual structures that we called "pseudo-nuclear" vesicles: they possess nuclear features as shown by both cytochemical and electron microscopic studies (Durandetal., 1986). Cytoplasmic features. G. toxicus cytoplasm is very dense and contains the usual organelles found in dinoflagellates, including hundreds of trichocysts. When dehy- drated (under the microscope, for instance), one or two brighter areas often appeared (Fig. 4f). They could be linked to the pusule. Deformed cells Cells having abnormal morphological aspects appeared in the cultures under some conditions (Durand and Puiseux-Dao, 1985). The morphology of these cells was deformed in comparison to typical cells: in apical view, their shape was embossed instead of circular; in a lateral view, they were round instead of flattened. The thecal- plate feature of teratogenous forms was completely disorganized as seen by SEM (Fig. 4c). Their size was large (80-100 ^m), their motility reduced, and they did not stick to the bottom of the flasks. They were capable of mitosis, so they cannot be a sexual stage of G. toxicus. 114 M. DURAND-CLEMENT GAMBIERDISCUS TOXICUS 1 1 5 Analysis of the conditions of the appearance and evolution of deformed cells in G. toxicus populations suggested two sources: culture-medium composition and pho- toperiod. Small motile cells and cysts Cells that had abandoned their theca (ecdysis; Fig. 4d) were often observed. Some small, motile, and apparently naked cells were occasionally visible in aged cultures (Fig. 4e). They could correspond to the cells released by ecdysis described above and be a survival behavior as described for other dinoflagellates (Walker, 1982), or to a stage of hypnoid cyst formation denned by Anderson et al. ( 1978). According to these workers, the structure shown in Figures 4g and h (same cell, two different foci) would represent another type of temporary cyst (coccoid). They were detected only once, after increasing the temperature of the culture to over 30°C. Some large (over 100 /urn), dark, and non-motile cells which could be interpreted as hypnocysts, were occasionally observed. Also observed (mostly in seawater) were paired cells which could be interpreted as fused gametes. These observations suggest that sexual reproduction might occur in G. toxicus cultures, but this hypothesis needs to be confirmed by other observations and studies. TOXICITY OF G. TOXICUS CULTURES G. toxicus toxicity was studied to evaluate the influence of culturing on its ability to synthetize toxic compounds. After verifying that G. toxicus was toxic, we tried to improve the toxin extraction procedure; finally, we tested a few parameters which could modify the toxin production of this dinoflagellate. Evaluation of toxicity Toxicity was evaluated using the mouse bioassay described by Yasumoto et al. (1979). Toxic extracts were dried under a nitrogen stream to eliminate any trace of solvent. Each extract was emulsified in 0.4 ml of a Tween 60 solution and injected intraperitoneally into a 19-21 g mouse (Swiss female). Several doses were tested (2 mice per dose) to determine the minimal lethal dose (MLD). MLD is defined as the minimal amount of extract that causes death for a 20 g mouse within 24 h (expressed in mg/kg). To simplify comparison between the toxicity of different extracts, lethality is expressed in mouse-units (one MU equals MLD -H 50), and total lethality is ex- pressed as the number of cells per mouse unit. Toxin extraction procedure First, we extracted ciguatoxins by the classic procedure of Yasumoto et al. (1979) and Bagnis et al. (1980). This procedure was inadequate to our algal material, so we studied the yield of each step of the protocol and modified them progressively. These FIGURE 3. Gambierdiscus toxicus cells, (a) Three typical G. toxicus cells (microcinematography, see the longitudinal flagellum), 320X. (b) A typical G. toxicus cell, 900X. (c) G. toxicus nucleus (Nu) ejected with the cytoplasm while pressing the cell between slide and coverslip, 1 1 50X. (d) A depigmented G. toxicus cell after Feulgen's staining, 800X. (e) Squash of G. toxicus after Feulgen's staining: view of the chromo- somes, 2600X. 116 M. DURAND-CLEMENT ...» ^ V *< <** w i FIGURE 4. (a, b) Typical G. toxicus cells, SEM. The black line represents 20 ^m (Durand and Coute photographs), (c) A deformed G. toxicus cell, SEM. The black line represents 20 /urn. (d) Cell during ecdysis, 600X. (e) Aspect of a small, apparently naked cell, 900X. (f) Two typical cells after small dehydration process, 350X. (g, h) Aspect of a "temporary cyst" of G. toxicus (2 different foci), 600X. GAMBIERDISCUS TOXICUS 1 1 7 TABLE I Toxin extraction procedure for Gambierdiscus toxicus cultures G. toxicus culture (about 106 cells) O Centrifugation (1500 X g— 5 min) o Sonication of the pellet in methanol Extraction of the pellet by methanol ( 10 ml, X3, room temperature) TOTAL CRUDE EXTRACT Evaporation of methanol Partition: water pH 7/isopropylic ether (2V) ^\^ , (IV) AQUEOUS-PHASE ETHER-PHASE assays resulted in the protocol outlined in Table I. The principal modifications were: (1) using centrifugation, instead of filtration on paper filters, for culture harvesting and cell sonication, which permitted us to use small amounts of solvent; and (2) extracting with "cold" (room temperature) methanol and partitioning with isopro- pylic ether. This method is simpler, more rapid, and enhanced the yield of toxin two- to fourfold. To simplify the evaluation of MLD for the initial methanol extract and to save mice, we established a relationship between the cell-pellet weight and the correspond- ing cell number (calculated using a Coulter counter). The equation of the linear rela- tionship is: Y = 0.056 + 0.263X (Y represents the cell number: X represents the weight; n = 4S samples tested). Thus, we established that one million G. toxicus cells weighed 0.36 g. Results For toxicity screening we extracted many batches of G. toxicus cultivated in different media conditions. We also looked for differences between clonal and non- clonal strains. For these experiments cultures were harvested when they reached the stationary growth phase. Results are given in Table II. Analysis of the data shows that differences between toxicities for various culture conditions were not significant. Other preliminary assays suggest that bacteria have little influence on total toxin production of G. toxicus, as the same toxicities were found for contaminated cultures (prior to antibiotic treatment) and "axenized" cul- tures (less than one bacteria per dinoflagellate cell) (Durand, 1987). In addition, we did not find a large increase in toxicity in aged cultures, in contrast to what other workers reported ( Yasumoto, etal., 1979; Bergmann and Alam, 1981). 118 M. DURAND-CLEMENT TABLE II Compilation o/Gambierdiscus toxicus toxicity data Strain Culture medium Cell number (103) Total crude extract Toxicity % Total MU Cells/MU Ether-phase Aqueous-phase Clone 1 MPP 539 800 680 ND ND MS + MPP 324 625 520 8 72 288 290 1000 34 ND Clone 2 MS + MPP 550 550 1000 18 82 218 180 1220 3 ND LATEUR 421 190 2240 8 ND 1843 1860 980 ND ND Clone 3 MPP 307 340 900 3 56 205 180 1140 4 ND MS 203 150 1360 10 83 MS + MPP 1500 1700 880 12 88 769 750 1020 ND ND Clone 4 MS + MPP 243 235 1040 11 60 1640 700 2340 28 71 Clone 5 MPP 326 400 820 ND ND 400 625 640 19 74 MS 336 225 1500 ND ND MS + MPP 1040 420 2480 ND ND 166 170 980 ND ND Strain A MPP 726 480 1520 ND ND MS + MPP 568 470 1200 13 74 2200 1050 2100 18 81 Strain B MPP 1486 1560 960 13 32 2015 4075 500 21 54 MS + MPP 727 725 1000 ND ND ND: non-determined; MU: mouse unit [equals Minimal Lethal Dose (MLD) + 50]; MPP: culturing medium (Provasoli, 1958); MS: Shepard's artificial medium (Shepard, 1968). The mean lethality for our G. toxicus strains was 1200 cells/MU. 106 G. toxicus = 0.36 g = 800 MU; MLD == 1-3 mg/kg. We concluded that, under laboratory culture conditions, G. toxicus remained very toxic — in the same range that was reported for wild populations (taking into account the extraction yield by Yasumoto et al, 1979, and Bagnis et al., 1980). CTX-like and MTX-like toxicities During the toxin extraction procedure, the partition step always resulted in two toxic fractions: a water-soluble fraction (MTX-like) and an ether-soluble fraction (CTX-like). These fractions appeared in a CTX-like/MTX-like ratio of about 0.2. Figure 5 gives the dose/time-to-death curve that was obtained with one CTX-like extract. It must be noted that the minimal death time for this toxic fraction is higher than that reported for ciguatoxin extracted from fish (Tachibana, 1980; Lewis and Endean, 1984). This suggests that the ether phase obtained from G. toxicus cultures GAMBIERDISCUS TOXICUS 119 DOSE (mg/kg) 30 15 125 1O 15 LETHALITY (HR) FIGURE 5. Relationship between dose and death-time for a crude ether phase from Gambierdiscus toxicus cultures (intraperitoneal injection in 20 g mice). is not equivalent to the one obtained from fish. In particular, it could contain MTX (as already suggested by Yasumoto et al, 1979), CTX-MTX complex, or another toxic compound. Further studies are necessary to characterize the toxin content of G. toxicus. CONCLUSION We obtained some stable, well-defined uni-algal cultures of G. toxicus (different clones, in different media). Development of large-scale culturing must, at present, take into account the benthic behavior of this dinoflagellate, and should employ sim- ilar techniques as those used for animal-cell cultures. G. toxicus in culture remains very toxic, thus it could be a superior raw material for providing cleaner and more abundant amounts of toxins: certainly maitotoxin and perhaps ciguatoxin, if further toxin characterization confirms its presence in the ether phase. The relationship between the two toxins remains to be elucidated, as does the importance of G. toxicus in ciguatera outbreaks. As this dinoflagellate is the focus of several studies, we suggest a standardization of its toxicity evaluation in order to help comparisons. ACKNOWLEDGMENTS We thank Dr. R. Bagnis for providing the initial G. toxicus sample and Dr. S. Puiseux-Dao for helpful discussion and for the improvement of this manuscript. We also thank F. Partensky for help in toxicity experiments and C. Guennou for cultures. 120 M. DURAND-CLEMENT This research was supported by a Cordet grant and an Ifremer contract (No. 84m 7QV 84/3179). LITERATURE CITED ADACHI, R., AND Y. FUKUYO. 1979. The thecal structure of a marine toxic dinoflagellate Gambierdiscus toxicus gen. and sp. nov. collected in a ciguatera endemic area. Bull. Jpn. Soc. Sci. Fish. 45( 1 ): 67-71. ANDERSON, D. M., AND D. WALL. 1978. Potential importance of benthic cysts ofGonyaulax tamarensis and Gonyaulax excavata in initiating toxic dinoflagellate blooms. J. Phycol. 14(2): 224-234. BAGNIS, R. 198 1 . L'ichtyosarcotoxisme de type ciguatera: processus biologiques connus et perspectives au seuil des annees 80. Ann. Inst. Oceanogr. Paris 1: 5-24. BAGNIS, R., S. CHANTEAU, E. CHUNGUE, J. M. HURTEL, T. YASUMOTO, AND A. INOUE. 1980. Origins of ciguatera fish poisoning: a new dinoflagellate, Gambierdiscus toxicus, Adachi and Fukuyo, definitively involved as a causal agent. Toxicon 18: 199-208. BAGNIS, R., S. CHANTEAU, AND T. YASUMOTO. 1977. Decouverte d'un agent etiologique vraisemblable de la ciguatera. C. R. Acad. Sci. Paris 28( 1 ): 1 05- 1 08. BERGMANN, J. S., AND M. ALAM. 1 98 1 . On the toxicity of the ciguatera producing dinoflagellate, Gambier- discus toxicus, Adachi and Fukuyo, isolated from the Florida Keys. J. Environ. Sci. Health A16(5): 493-500. BESADA, E. G., L. A. LOEBLICH, AND A. R. LOEBLICH III. 1982. Observations on tropical, benthic dino- flagellates from ciguatera endemic areas: Coolia, Gambierdiscus and Ostreopsis. Bull. Mar. Sci. 32(3): 723-735. CARLSON, R. D., G. MOREY-GAINES, D. R. TINDALL, AND R. W. DICKEY. 1984. Ecology of toxic dino- flagellates from the Caribbean Sea. Effects of macroalgal or growth in culture. Pp. 271-287 in Sea Food Toxins, E. P. Ragelis, ed. ACS Symposium Series, No. 262. DURAND, M. 1984. Etude biologique, cytologique et toxicologique de Gambierdiscus toxicus en culture, dinoflagelle responsable de la ciguatera. These de Seme Cycle. Universite Paris VII, France. DURAND, M. 1987. A study of toxin production by Gambierdiscus toxicus in culture. Toxicon 24: 1 153- 1157. DURAND, M., AND C. BERKALOFF. 1985. Pigment composition and chloroplast organization ofG. toxicus, Adachi and Fukuyo (Dinophyceae). Phycologia 24(2): 217-223. DURAND, M., AND S. PUISEUX-DAO. 1985. Physiological and ultrastructural features of the toxic dino- flagellate G. toxicus in culture. Pp. 61-68 in Toxic Dinoflagellates, Anderson, White, and Baden, eds. Elsevier Science Publishing Co., New York. DURAND, M., A. SQUIBAN, R. RIBIER, R. BAGNIS, AND S. PUISEUX-DAO. 1986. Pseudonuclear vesicles in the toxic dinoflagellate Gambierdiscus toxicus. Biology of the Cell 56: 171-1 80. FUKUYO, Y. 1981. Taxonomical study on benthic dinoflagellates collected in coral reefs. Bull. Jpn. Soc. Sci. Fish. 47(8): 967-978. GALLERON, C. 1 976. Synchronization of the marine dinoflagellate Amphidinium carteri in dense cultures. /Vj.vco/0£/fll2(l):69-73. HARRISON, P. J., R. E. WATERS, AND F. J. R. TAYLOR. 1 980. A broad spectrum artificial seawater medium for coastal and open ocean phytoplankton. J. Phycol. 16: 28-35. HURTEL, J. M., S. CHANTEAU, J. H. DROLLET, AND R. BAGNIS. 1979. Culture en milieu artificiel du dinoflagelle responsable de la ciguatera. Rev. Int. Oceanogr. Med. LV: 29-33. JEFFREY, S. W. 1976. The occurrence of chlorophyll c, and c2 in algae. J. Phycol. 12: 349-354. LATEUR, L. 1963. Une technique de culture pour \\4cetabularia mediterranea. Res. Alg. 3( 1 ): 26-36. LEWIS, R. J., AND R. ENDEAN. 1984. Ciguatoxin from the flesh and viscera of the barracuda, Sphyraena jello. Toxicon 22(5): 805-8 10. LOEBLICH, A. R. III. 1966. Aspects of the physiology and biochemistry of the Phvrrhophvta. Phykos 5(1- 2): 216-255. LOEBLICH, A. R. III. 1970. The amphiesma or dinoflagellate cell covering. Proc. North Am. Paleont. Conv. G. 867-929. MORILL, L. C., AND A. R. LOEBLICH. 1981. The dinoflagellate pellicular wall layer and its occurrence in the division. Pyrrophyta. J. Phvcol. 17: 315-323. NAKAJIMA, I. 198 i . Toxicity of benthic dinoflagellates in Okinawa. Bull. Jpn. Soc. Sci. Fish. 47(8): 1029- 1033. PROVASOLI, L. 1958. Nutrition and ecology of protozoa and algae. Ann. Rev. Microbiol. 12: 279. SHEPARD, D. C. 1969. Axenic culture of Acetabularia in synthetic media. Pp. 49-63 in Methods in Cell Physiology, Vol. 4, D. Prescott, ed. Academic Press, New York. SHIMIZU, Y., H. SHIMIZU, P. J. SCHEUER, Y. HOKAMA, M. OYAMA, AND J. T. MIYAHARA. 1982. Gam- GAMBIERDISCUS TOXICUS 121 bier discus toxicus, a ciguatera causing dinoflagellate from Hawaii. Bull. Jpn. Soc. Sci. Fish. 48(6): 811-813. STEIDINGER, K. A., AND D. G. BADEN. 1984. Toxic marine dinoflagellates. Pp. 201-261 in Dinoflagellates, D. L. Spector, ed. Academic Press, New York. TACHIBANA, K. 1980. Structural Studies on Marine Toxins. Ph.D. dissertation. University of Hawaii. TAYLOR, F. J. R. 1979. A description of the benthic dinoflagellate associated with maitotoxin and cigua- toxin, including observations on Hawaiian material. Pp. 47-56 in Toxic Dinoflagellate Blooms, D. L. Taylor and H. H. Seliger, eds. Elsevier/North Holland, New York. TINDALL, D. R., R. W. DICICEY, T. D. CARLSON, ANDG. MOREY-GAINES. 1984. Ciguatoxic dinoflagellates from the Caribbean Sea. Pp. 225-240 in Sea Food Toxins, E. P. Ragelis, ed. ACS Symposium Series, No. 262. WALKER, L. M. 1 982. Evidence for a sexual cycle in the Florida red tide dinoflagellate, Ptychodiscus brevis (=Gymnodinium breve). Trans. Am. Microsc. Soc. 101(3): 287-293. WITHERS, N. W. 1982. Ciguatera fish poisoning. Ann. Rev. Med. 33: 97-1 1 1 . YASUMOTO, T., I. NAKAJIMA, AND R. BAGNIS. 1977. Finding of a Dinoflagellate as a likely culprit of ciguatera. Bull. Jpn. Soc. Sci. Fish. 43(8): 1021-1026. YASUMOTO, T., I. NAKAJIMA, Y. OSHIMA, AND R. BAGNIS. 1979. A new toxic dinoflagellate found in association with ciguatera. Pp. 65-70 in Toxic Dinoflagellate Blooms, D. L. Taylor and H. H. Seliger, eds. Elsevier/North Holland, New York. YASUMOTO, T., U. RAJ, AND R. BAGNIS. 1984. Seafood poisonings in tropical regions. Symposium on Seafood Toxins in Tropical Regions. Kagoshima, September, 1983. Reference: Biol. Bull. 172: 122-127. (February, 1987) RECENT DEVELOPMENTS IN THE MOLECULAR STRUCTURE OF CIGUATOXIN KAZUO TACHIBANA, MANABU NUKINA, YONG-GOE JOH, AND PAUL J. SCHEUER Department of Chemistry, University of Hawaii at Manoa, Honolulu, Hawaii 96822 ABSTRACT Crystalline ciguatoxin isolated from moray eel (Lycodontis = G ymnothorax ja- vanicus) viscera has an LD50 of 0.45 g/kg (i.p., mice). It has a molecular weight of 1 1 1 1.7 ± 0.2 daltons. 'H NMR studies have shown that it is a polar and highly oxy- genated molecule belonging to the class of polyethers. On basic alumina ciguatoxin is reversibly converted to a chromatographically distinct less polar form, which is equally toxic and elicits typical ciguatoxin symptoms in mice. From parrotfish (Scarus sordidus), which originated on a ciguateric reef on Ta- rawa atoll (Kiribati), we have isolated two toxins that evoke ciguatera symptoms in mice at approximately equal levels. Chromatographic evidence suggests that the two toxins are identical with the two ciguatoxins of different polarity and that the less polar form is the previously described scaritoxin. INTRODUCTION "He who oversees everything also created very many poisonous fish, in this way he punishes those who seek them." — J. Grevin In his book on venoms the 16th century, French physician and poet Jacques Grevin (Grevin, 1 568) intuitively foresaw some of the frustration which has been the hallmark of ciguatera research during the past thirty years. Although ciguatoxin, the principal toxin in ciguateric carnivorous fish, may be superficially compared with the well-known marine toxins tetrodotoxin and the saxi-gonyautoxins, they share few characteristics with ciguatoxin. The single factor responsible for the slow progress in ciguatera research is that ciguatoxin is a slow-acting toxin which is rarely fatal to man. It is rarely fatal not because of its lack of potency — indeed its potency is surpassed only by that of palytoxin (0. 1 5 Mg/kg; Moore and Scheuer, 1971) and maitotoxin (0.2 jug/kg; Ohizumi and Yasumoto, 1983) — consumers of ciguateric fish survive because of its extremely low concentration in fish flesh. Thus, it has been difficult to accumu- late enough toxin for structural research and — for many years — to convince skeptical fellow scientists of the very existence of a well-defined toxic entity. As a slowly acting toxin, ciguatoxin lacks the dramatic impact of a fast-acting toxin, somehow a con- vincing demonstration of the power of a toxin. METHODS AND RESULTS Because of the many variables attending ciguatera outbreaks, among them place, time, species of fish, and reported symptoms in man, we have confined our research to toxin isolated from the moray eel (Lycodontis = Gymnothoraxjavanicus), initially from flesh and viscera, but more recently exclusively from viscera in order to econo- mize on solvents and shipping cost. Geographically, the eels originated from John- 122 CIGUATOXIN MOLECULAR STRUCTURE 123 TOXIN EXTRACTION MINCED EEL VISCERA (10 KG) soak m acetone, RT, 2d f i Her ACETONE FILTRATE cool to -20° overnight filter evaporate <40° AQUEOUS MIXTURE extract with hexane backwash with aq MeOH d t scard hexane cone aq MeOH DEFATTED AQ MIXTURE I sat with NaCI extract w i t h EtOAc EtOAc EXTRACT I dry wi th Na2SO< filter T evaporate BROWN OIL (30 g) FIGURE 1. Ciguatoxin extraction. ston atoll (165°W, 17°N) and more recently from Tarawa atoll (173°E, TSO'N), Re- public of Kiribati. Figures 1 and 2 summarize our procedure for toxin extraction and purification (Tachibana, 1980). Approximately 75 kg of eel viscera, representing some 1 100 kg of toxic eels, yielded 1.3 mg of chromatographically pure toxin. The final HPLC trace (Fig. 3) obtained on a C18 reversed phase silica column is a symmetrical peak. The toxin, LD50 0.45 jug/kg (i.p., mice), is a colorless solid readily soluble in methanol, ethanol, 2-propanol, and acetone, sparingly soluble in chloroform and diethyl ether, and nearly insoluble in water or benzene. Ciguatoxin displays a single UV absorption peak at 215 nm (e 5250). At that wavelength its CD (circular dichroism) molecular ellipticity is -620. The most prom- inent features in the infrared spectrum (FT, solid film) of ciguatoxin are hydroxyl (3450 cm"1) and ether (1080 cm"1) bands. A respectable signal at 1600 cm ' cannot be unequivocally interpreted. Californium-252 plasma desorption mass spectrometry pointed to a likely molec- ular weight of 1 1 1 1.7 ± 0.3 daltons (R. D. Macfarlane and C. McNeil, pers. comm.). Since this technique is unsuitable for high resolution measurements, no molecular formula of ciguatoxin was obtained. On the basis of 1H NMR data, formulas such as C53H77NO24 (111 2.2) or C54H18O24 (1111.2) are reasonable. The bulk of the structural information was derived from extensive 'H NMR ex- periments at 360 and 600 MHz in methanol-d4 or dimethylsulfoxide-d6 . Methanol gives rise to a well-defined spectrum, but provides no clues for exchangeable protons. Ciguatoxin possesses five hydroxyl groups, four carbon-carbon double bonds, and five methyls, all on saturated carbon. The combined structural pieces account for 124 K. TACHIBANA ET AL. TOXIN CHROMATOGRAPHY EXTRACTION RESIDUE (46 g) BioSil A CHCIj - MeOH g radient e lu t ion TOXIC FRACTION (14 g) I basic alumina (14 gj r LESS POLAR g rad ient e I u t ion CHCI,-MeOH — - aq MeOH TOX.N TOXIN <1.«fl) . Sephadex L H ( 315 mg) J CH2ci,- MeOH TOXI N (154 mg) i Sephadex LH CH,CI, - MeOH TOXI N (47 .7 mg) C,8Hi - Flosi I aq MeOH TOXIN (3.2 mg) HPLC Lichrosorb R P - 1 ' aq MeOH (3 X) CIGUATOXIN (0.3 mg) (yield 10 ppb) FIGURE 2. Ciguatoxin purification. nearly 75 hydrogen atoms and imply 54 carbon atoms, but in the absence of a carbon spectrum it is impossible to judge the degree of overlap, which casts doubt on the reliability of the carbon count. Although no satisfactory carbon- 1 3 spectrum of cigua- toxin has been determined because of lack of toxin and/or sufficient instrument time, our sample crystallized in an NMR tube during an attempt to have the carbon spec- trum measured at a mainland applications laboratory. The crystals, unfortunately, are too small to be suitable for x-ray diffraction studies, and our attempts at growing larger crystals have so far failed. DISCUSSION AND CONCLUSION When one considers the broad spectrum of symptoms which ciguatera-intoxi- cated patients describe and the difficulty of isolating a homogeneous toxin from a complex matrix in which it is present at a concentration of approximately 1 X 10"6%, the question of multiple toxins inevitably arises. In the absence of convincing evi- dence to the contrary, we have followed the simplest assumption in our research; i.e., we have assumed that ciguatoxin is a single entity. Yet occasionally we observed (Tachibana, 1980) that in samples of extracts that had been stored for some time, ciguatoxin would be eluted from a silica column with a less polar solvent mixture (chloroform/methanol 97:3) than is the case normally, when the bulk of the toxin is eluted with a 9:1 chloroform/methanol solution. The existence of a less polar form CIGUATOXIN MOLECULAR STRUCTURE 125 Injection Detection Rl UV Ciguatoxin 0 10 20 Time (min) FIGURE 3. Reversed phase HPLC of ciguatoxin. of ciguatoxin can be demonstrated by chromatography on basic alumina of different activity grades. We showed (Nukina et al, 1984) that the two forms of ciguatoxin, while chromatographically distinct, are interconvertible. The two forms have 'H NMR spectra which differ only in minor details and elicit comparable symptoms in mice. An epidemiological survey including detailed case studies in the Gambier islands, where ciguatera intoxication arises principally from eating parrotfishes (Scaridae), led Bagnis et al. ( 1 974) to suggest that either a new toxin or multiple toxins were involved. In her follow-up, Chungue (1977; Chungue et al., 1977) isolated from the flesh of Scams gibbus a toxic mixture which was separable by DEAE cellulose chromatogra- phy into a toxin designated scaritoxin and a more polar toxin which strongly resem- 126 K. TACHIBANA ET AL. 39 OH Me" FIGURE 4. ( i) Molecular structure of okadaic acid. (2) Molecular structure of brevetoxin B. bled ciguatoxin. Scaritoxin was reported to cause severe hind limb paralysis in mice, a symptom not normally observed with ciguatoxin. Recently, we (Joh and Scheuer, in press) examined parrotfish (Scams sordidus) from a toxic reef on Tarawa atoll, Republic of Kiribati. We also isolated two toxins separable on DEAE cellulose. However, by manipulation on basic alumina we were able to interconvert the two toxins. By TLC comparison we showed that scaritoxin and the less polar ciguatoxin (Nukina et al, 1984) are identical, though this finding remains to be confirmed by spectral comparison. When Yasumoto (Murakami et al., 1982) examined the constituents of the toxic dinoflagellate Prorocentrum lima, he made the surprising discovery that on TLC analysis one of the constituents was indistinguishable from ciguatoxin. The substance proved not to be ciguatoxin but okadaic acid (Fig. 4, i ), a compound which we had shortly before reported from a sponge, Halichondria okadai (Tachibana et al., 1 98 1 ). The two compounds differ greatly in size (1111 vs. 804 daltons) and lethality (0.45 vs. 1 92 Mg/kg), but evidently not in polarity because of their similar chromatographic behavior. This was the first clear indication that ciguatoxin belongs to the class of polyethers, as does, inter alia, brevetoxin B (Fig. 4, 2) (Lin et al., 1981). These com- pounds are highly oxygenated long-chain fatty acids, in which most of the oxygen atoms occur as cyclic ether linkages. Okadaic acid (Fig. 4, i), possesses one carboxyl, four hydroxyls, and seven oxa rings, in addition to three carbon-carbon double bonds. This information allows us to extrapolate safely to the ciguatoxin structure, which evidently is a complex polyether possessing five hydroxyls and four carbon- carbon double bonds. This close structural analogy to okadaic acid (Fig. 4, i) and to brevetoxin (Fig. 4, 2) provides a plausible rationale for the interchangeability of different chromatographic forms of ciguatoxin and scaritoxin. The presence in cigua- CIGUATOXIN MOLECULAR STRUCTURE 127 toxin of multiple hydroxyl groups permits formation and destruction of various hy- drogen-bonded forms while preserving the structural integrity of the molecule. Be- cause of its large size, ciguatoxin may well assume two or more secondary shapes which prevail under different conditions of basicity. Perhaps the most intriguing questions posed upon examination of these com- pounds are those dealing with their mechanism of physiological action and with the subtle structural features that give rise to profound differences in lethality and overt symptoms in mammals. ACKNOWLEDGMENT We thank the National Marine Fisheries Service (NA80AA-00101), through a subcontract with the Medical University of South Carolina, for financial support. LITERATURE CITED BAGNIS, R., E. LOUSSAN, AND S. THEVENIN. 1974. Les intoxications par poissons perroquets aux fles Gambier. Med. Trap. 34: 523-527. CHUNGUE, E. \977.Le Complexe Toxinique des Poissons Perroquets. Ph.D. Thesis, Academic de Montpe- lier, Universite des sciences et techniques du Languedoc, Montpelier, France. CHUNGUE, E., R. BAGNIS, N. FUSETANI, AND Y. YASHIMOTO. 1977. Isolation of two toxins from a parrot- fish Scarus gibbus. Toxicon 15: 89-93. GREVIN, J. 1568. Deux Livres des Venins. Christofle Plantin, Antwerp, Belgium. JOH, Y.-G., AND P. J. SCHEUER. In press. The chemical nature of scaritoxin. Marine Fish. Rev. LIN, Y. Y., M. RISK, S. M. RAY, J. CLARDY, D. VAN ENGEN, J. GOLIK, J. C. JAMES, AND K. NAKANISHI. 1981. Isolation and structure of brevetoxin B from the "red tide" dinoflagellate Ptychodiscus brevis (Gymnodinium breve). J. Am. Chem. Soc. 103: 6773-6775. MOORE, R. E., AND P. J. SCHEUER. 197 1 . Palytoxin: a new marine toxin from a coelenterate. Science 172: 495-498. MURAKAMI, Y., Y. OSHIMA, AND T. YASUMOTO. 1982. Identification of okadaic acid as a toxic compo- nent of a marine dinoflagellate Prorocentrum lima. Nippon Suisan Gakkaishi 48: 69-72. NUKINA, M., L. M. KOYANAGI, AND P. J. SCHEUER. 1984. Two interchangeable forms of ciguatoxin. Toxicon 22: 169-176. OHIZUMI, Y., AND T. YASUMOTO. 1983. Contraction and increase in tissue calcium content induced by maitotoxin, the most potent known, marine toxin, in intestinal smooth muscle. Br. J. Pharmacol. 79: 003-005. TACHIBANA, K. 1 980. Structural Studies on Marine Toxins. Ph.D. Thesis, University of Hawaii, Honolulu. TACHIBANA, K., P. J. SCHEUER, Y. TSUKITANI, H. KIKUCHI, D. VAN ENGEN, J. CLARDY, Y. GOPICHAND, AND F. J. SCHMITZ. 1981. Okadaic acid, a cytotoxic polyether from two marine sponges of the genus Halichondria. J. Am. Chem. Soc. 103: 2469-247 1 . Reference: Biol. Bull. 172: 128-131. (February, 1987) TOXINS PRODUCED BY BENTHIC DINOFLAGELLATES TAKESHI YASUMOTO*, NAOKO SEINO**, YASUTAKA MURAKAMI***, AND MICHIO MURATA Faculty of Agriculture, Tohoku University, 1-1 Tsutsumidori-Amamiyamachi, Sendai 980, Japan ABSTRACT Nine species of benthic dinoflagellates collected in subtropical waters were cul- tured, extracted, and tested for mouse lethality, ichthyotoxicity, and hemolytic activ- ity. Hemolytic activity was detectable in all species, but the activities ofAmphidinium carteri, A. klebsi, and Gambler discus toxicus were outstanding. G. toxicus showed the most potent mouse lethality. Two hemolytic constituents of A. carteri were deter- mined to be mono- and di-galactoglycerolipids. Maitotoxin, produced by G. toxicus, was suggested to have a molecular weight of 3402 ± 2 (m/z). Two potent toxins against mice were isolated from Prorocentrum lima and identified as okadaic acid and 5-methylene-6-hydroxy-2-hexen- 1 -okadaate. INTRODUCTION Scientists first noticed the toxigenicity of benthic dinoflagellates when Gambler- discus toxicus was found to produce and to transmit ciguatoxin and maitotoxin to herbivorous fish (Yasumoto et ai, 1977). Subsequent G. toxicus distribution surveys revealed an abundance, in terms of both species and population, of benthic dinofla- gellates in coral reef communities. This observation suggests that toxic metabolites, if any, of these benthic species are taken up by herbivorous fish, as ciguatoxin is, and contribute to the manifestation of the complex symptoms of ciguatera. The actual occurrence of minor toxins in the viscera of herbivorous fish has been confirmed, and the toxigenicity of several benthic species has also been demonstrated previously (Yasumoto et al. , 1976; Nakajima et al. , 1981). The present paper briefly summarizes our knowledge of the toxic benthic species and the chemical natures of their toxins; ciguatoxin, however, has been described separately by Tachibana et al. (1986). MATERIALS AND METHODS Benthic dinoflagellates collected at Okinawa, Japan, were cultured in a nutrient- enriched seawater medium described by Provasoli (1968). The following nine species were tested for the toxin production: Amphidinium carteri, A. klebsi, Coolia monotis, Gambler discus toxicus, Ostreopsis ovata, O. siamensis, Prorocentrum concavum, P. lima, and P. rhathymum. The harvested cells were extracted with boiling methanol. The methanol was re- moved by evaporation, and the residue was then suspended in water and extracted, first with diethyl ether and then with 1-butanol. The residues obtained after evapora- * To whom correspondence should be addressed. **,*** Present addresses: N. Seino, Ohmori Research Institute of Tobishi Pharm. Co. Ltd., 1-16-18 Ohmorinishi, Ohtaku, Tokyo 146, Japan; Y. Murakami, MITSUKAN Co. Ltd. 2-6 Nakamuracho, Handa, Aichi Pref. 475, Japan. 128 TOXINS OF BENTHIC DINOFLAGELLATES 129 TABLE I Toxicigenicity of bent hie dinoflagellates Mouse Hemolytic Species lethality Ichthyotoxicity activity Amphidinium carteri Amphidinium klebsi Coolia monotis Gambierdiscus toxicus Ostreopsis ovata Ostreopsis siamensis Prorocentrum concavum Prorocentrum lima Prorocentrum rhathymum + The relative potency is expressed by increasing the number of +, the potency of undetectable level is expressed by — . tion of the solvents were tested, respectively, for mouse lethality, ichthyotoxicity, and hemolytic activity, as described previously (Nakajima et al., 1981). Three toxins, tentatively named PL toxin- 1, -2, and -3, were isolated from P. lima by successive treatments on columns of silicic acid (CHCl3-MeOH, stepwise), Sephadex LH-20 (CHCl3-MeOH 2:1), LiChroprep RP-2 (Merck, MeOH-H2O 2:1), and ODS (Kyowaseimitsu, MeOH-H2O 4: 1). Five hemolytic compounds (hemolysin-1 to -5) were present in A. carteri. Hemo- lysin-1 and -2 were purified on columns of silicic acid (CHCl3-MeOH, stepwise) and ODS Q-3 (Fujigel, MeOH-H2O 9: 1 ). Hemolysin-3 to -5 were also purified in a similar manner but further purification on a Toyopearl 40 column was necessary (MeOH- H2O, stepwise). Purification of maitotoxin was carried out on columns of silicic acid (CHC13- MeOH, stepwise), Develosil ODS (Nomurakagaku, MeOH-H2O, stepwise), and De- velosil TMS (MeCN-H2O 35:65). 'H NMR and 13C NMR spectra were taken on either a Nicolet NT 360 spectrome- ter or a JEOL FX-100 spectrometer, and mass spectra were taken on a Hitachi M-80 mass spectrometer. RESULTS Bioassays indicate that all the dinoflagellates are toxic by at least one of the assay methods, as shown in Table I. Mouse lethality was observed in A. carteri, A. klebsi, G. toxicus, O. siamensis, P. concavum, and P. lima. The toxicity of G. toxicus to the mouse was outstanding. Hemolytic activity was most prominent in A. carteri and A. klebsi, although observed in all species tested. Potent ichthyotoxicity was observed in A. carteri, A. klebsi, and P. concavum. The chromatographic and spectral analyses of PL toxin-2, the major toxin pro- duced by P. lima, proved it to be okadaic acid (Murakami et al, 1982), a cytotoxic polyether fatty acid derivative previously isolated from sponges (Tachibana et al., 1981). PL toxin- 1 was found to be a mixture of diol esters of okadaic acid. The struc- tures of okadaic acid and the major ester in the PL toxin- 1 fraction are shown in Figure 1 . PL toxin-3 was shown to be a tertiary amine having a molecular weight of 98 1 (mf 2). Elucidation of its chemical structure is under way. 130 T. YASUMOTO ET AL. OH I : R = H H : R = FIGURE 1 . Okadaic acid (I) and one if its diolesters (II) from Prorocentrum lima. Among the five hemolysins of A. carteri, hemolysin- 1 and -2 were more abundant than the other three. The structures of hemolysin- 1 and -2 were determined to be O-j3-D-galactopyranosyl-( 1 - 1 )-3-O-octadecatetraenoyl-D-glycerol and O-«-D-galac- tosyl-( 1 -6)-O-/3-D-galactopyranosyl-( 1 - 1 )-3-O-octadecatetraenoyl-D-glycerol, respec- tively (Fig. 2). The hemolytic activities of hemolysin- 1 and -2 were 80% and 25% of that of the commercial saponin (Merck), respectively. The hemolytic activities of hemolysin-3, -4, and -5 were 100, 9, and 2 times more potent than the commercial saponin, respectively. Their chemical structures were indicated to be entirely different from those of hemolysin- 1 and -2. Further structural work is under way. Maitotoxin judged to be homogeneous by HPLC and TLC was obtained as an amorphous solid. It was extremely lethal to mice (0. 1 3 Mg/kg, ip). Mass spectra sug- gested the molecular weight of 3402 ± 2 (m/z). Chemical and spectral analyses indi- cated the absence of any amino acid or fatty acid moieties in the molecule. Further analyses of the structure are underway. DISCUSSION The occurrence of diverse toxins in benthic dinoflagellates was confirmed. Out of the nine species tested, six produced mouse-lethal toxins, three ichthyotoxins, and all species produced hemolytic substances. Such a high occurrence of toxins is a charac- HO HO HO HO HO HO O :OH HO O Hemolysin I Hemolysin II FIGURE 2. Hemolysin I and II from Amphidinium carteri. TOXINS OF BENTHIC DINOFLAGELLATES 131 teristic feature of benthic dinoflagellates. The biological and ecological significance of the toxins are not clear at present. Whether the toxins deter the growth of other microorganisms and thus benefit the elaborators remains to be tested. The monoacylgalactolipids (hemolysin-1 and -2) are closely related to the known intermediate metabolites of photosynthesis, and therefore are likely to be widely dis- tributed. The compounds may not be involved in ciguatera, but they could be respon- sible, in part, for fish kills during blooms of dinoflagellate species with no known ichthyotoxins. Production of okadaic acid by P. lima is interesting because of its chemical resem- blance to ciguatoxin and its potent diarrheagenicity. As P. lima is widely and densely distributed in coral reefs, there is a possibility that the compound, like ciguatoxin, is taken up by herbivorous fish and thus contributes to the diarrhea which is frequently seen in ciguatera patients. The presence of maitotoxin in the viscera of surgeonfish has already been con- firmed, and the toxin has been suspected of contributing to the diverse ciguatera symptoms seen in patients who have eaten herbivorous fish without first eliminating the viscera. The most characteristic feature of maitotoxin is its high lethality to mice (0. 1 3 MgAg, ipX which is 70 times that of saxitoxin or tetrodotoxin. The specific ac- tion of maitotoxin, to enhance the calcium ion influx through excitable membranes, was first discovered by Ohizumi's group (Takahashi et al., 1982; Ohizumi et al., 1983). Today the compound is being used extensively as a chemical tool in biomedi- cal research. ACKNOWLEDGMENTS The authors are grateful to Dr. I. Nojima of JOEL Co. and Drs. H. Naoki and T. Iwashita of the Suntory Institute for Bioorganic Research for the spectral measure- ments. This work was supported in part by Grant-in-Aid from the Ministry of Educa- tion, Culture, and Science, Japan. LITERATURE CITED MURAKAMI, Y., Y. OSHIMA, AND T. YASUMOTO. 1982. Identification of okadaic acid as a toxic compo- nent of a marine dinoflagellate Prorocentrum lima. Bull. Jpn. Soc. Sci. Fish. 48: 69-72. NAKAJIMA, I., Y. OSHIMA, AND T. YASUMOTO. 198 1 . Toxicity of benthic dinoflagellates in Okinawa. Bull. Jpn. Soc. Sci. Fish. 47: 1029-1033. OHIZUMI, Y., AND T. YASUMOTO. 1983. Contraction and increase in tissue calcium content induced by maitotoxin, the most potent known marine toxin, in intestinal smooth muscle. Br. J. Pharmacol. 79: 3-5. PROVASOLI, L. 1968. Media and prospect for the cultivation of marine algae. Pp. 63-75 in Proceedings of the U. S.-Japan Conference held at Hakone, September 12-15, 1966. Watanabe and Hattori, eds. Japan Society of Plant Physiologists. TACHIBANA, K., P. J. SCHEUER, Y. TSUKITANI, H. KIKUCHI, D. VAN ENGEN, J. CLARDY, Y. GOPICHAND, AND F. J. SCHMITZ. 1981. Okadaic acid, a cytotoxic polyether from two marine sponges of the genus Halichondria. / Am. Chem. Soc. 103: 2469-247 1 . TACHIBANA, KAZUO, MANABU NUKINA, YONG-GOE JOH, AND PAUL J. SCHEUER. 1 986. Recent develop- ments in the molecular structure of ciguatoxin. Biol. Bull. 172: 122-127. TAKAHASHI, M., Y. OHIZUMI, AND T. YASUMOTO. 1982. Maitotoxin a Ca2+ channel activator candidate. J. Biol. Chem. 257: 7287-7289. YASUMOTO, T., I. NAKAJIMA, R. BAGNIS, AND R. ADACHI. 1977. Finding of a dinoflagellate as a likely culprit of ciguatera. Bull. Jpn. Soc. Sci. Fish. 43: 1021-26. YASUMOTO, T., R. BAGNIS, AND J. P. VERNOUX. 1976. Toxicity study on surgeonfishes— II. Properties of the principal water-soluble toxin. Bull. Jpn. Soc. Sci. Fish. 42: 359-365. Reference: Biol. Bull. 172: 132-136. (February, 1987) PHARMACOLOGICAL ACTIONS OF THE MARINE TOXINS CIGUATOXIN AND MAITOTOXIN ISOLATED FROM POISONOUS FISH YASUSHI OHIZUMI Mitsubishi-Kasei Institute of Life Sciences, 11 Minamiooya, Machida, Tokyo 194, Japan ABSTRACT Ciguatoxin and maitotoxin isolated from poisonous fish and toxic dinoflagellates exhibited a powerful excitatory effect on smooth and cardiac muscle. The ciguatoxin- induced excitatory action is due to an increase in Na+ permeability of tetrodotoxin- sensitive Na channels, while the maitotoxin-induced excitatory effect is caused by an increased Ca2+ permeability of the cell membrane. INTRODUCTION A wide range of pharmacologically useful substances has been isolated from ma- rine organisms. Among them, marine toxins such as tetrodotoxin (TTX), saxitoxin and sea anemone toxins, have been extensively studied since they affect a specific physiological function (Catterall, 1980). The first of these toxins, TTX, was isolated from puffer fish. It specifically blocks voltage-sensitive Na channels (Narahashi, 1974). Thus, TTX has been an essential tool in elucidating neurally mediated physio- logical processes. Such neurotoxins as TTX and saxitoxin, which bind with high affinity to Na channels, have proved to be useful probes for purifying and identifying these channels (Catterall, 1984). Ciguatera is a disease caused primarily by the ingestion of any of a variety of fish inhabiting tropical and subtropical seas. The clinical symptoms for ciguatera are prickling of the lips, tongue, and throat; numbness; nausea; vomiting; metallic taste; dryness of the mouth; abdominal cramps; and diarrhea. It is well known that the principal toxin in ciguatera fish is ciguatoxin (CTX), a hydrohobic compound (Ha- shimoto, 1979). CTX is a highly toxic substance; the 50% lethal dose in mice is 0.45 Mg/kg when injected intraperitoneally (i.p.) (Tachibana, 1980). Recently, maitotoxin (MTX), a water-soluble substance, has been isolated from the viscera of a surgeonfish, Ctenochaetus striatus, and from the toxic dinoflagellate Gambierdiscus toxicus; it was found to be the most potent marine toxin known. The minimum lethal dose of MTX in mice is 0.17 yug/kg (i-P-) which is approximately fifty times more potent than that of TTX (Yasumoto, 1980). The chemical structure of MTX has been only partially determined although it is considered to be a non-peptidic compound with a high molecular weight (Yasumoto, 1980). MATERIALS AND METHODS Mechanical response Male guinea-pigs (250-350 g) and male rats (280-300 g) were used. The guinea- pig vas deferens and left atria were excised and mounted vertically in a 20 ml organ bath containing an Krebs-Ringer bicarbonate solution. The procedures for preparing the vas deferens (Ohizumi and Shibata, 1980) and the left atria (Kobayashi et al, 1985a, b) were carried out as previously described. Cardiac myocytes were isolated from the rat using the method of Miyaicoda and Nakamura (1982) and driven by a 132 CIGUATOXIN AND MAITOTOXIN ON MUSCLE 133 TABLE I Effect ofCTXon theNE release from the guinea-pig vas deferens in the presence or absence ofTTX orCa+^ Amount of NE released3 Treatment ng/g tissue/30 min None 39.3 ± 6.3 CTX(l(T6g/ml) 1225.0 ±240.8 TTX(5X 10-7M)b + CTX (10~6 g/ml) 45.0 ± 9.7 Ca++-free medium" + CTX (10-6 g/ml) 82.8 ± 11.9** aMean±S.E. (n = 4). b CTX was applied 1 5 min after treatment with TTX or incubation in Ca++-free medium. ** Significantly different from the CTX-treated group, P < .0 1 . stimulator through a pair of platinum electrodes. Images of the cells were recorded either with a video recording system or with a high-speed movie camera. Assay ofnorepinephrine (NE) Preparation and incubation of the guinea-pig vas deferens were performed as de- scribed previously (Norton et al, 1981). The determination of released endogenous NE was carried out as described by Ohizumi et al. (1983). Tissue Na and Ca content The Na (Ohizumi et al, 1982) and Ca (Ohizumi et al., 1983) concentrations in tissue were determined by the methods described earlier. Na,K-A TPase assay Na,K-ATPase was purified from the porcine brain. The enzyme reaction was car- ried out at 37°C, for 20 min as described previously (Ohizumi and Yasumoto, 1983), and inorganic phosphate was determined according to the method of Martin and Doty (1949). Extraction and purification ofCTXandMTX CTX (Tachibana, 1980) and MTX (Yasumoto, 1980) were extracted and purified as described previously. RESULTS AND DISCUSSION Smooth muscle CTX induced a dose-dependent contraction of the guinea-pig vas deferens at con- centrations above 3 X 10~7 g/ml (Ohizumi et al, 1981). The CTX-induced contrac- tion of the vas deferens was inhibited by phentolamine ( 10~6 M), guanethidine ( 10~' M), reserpine (2 mg/kg/day, twice), TTX (5 X 10"7 M\ or cold storage for 7 days at 4°C, but was not affected by atropine ( 10~6 M) or mecamylamine (3X10 5 M). CTX ( 1-6 X 10~7 g/ml) caused a dose-dependent release of NE from the vas deferens, and the maximum response (1297.4 ng/g tissue) was obtained with a concentration of 6 X 10~6 g/ml. The CTX-induced release of NE was nearly abolished by TTX (Table I) or by cold storage. These data suggest that the CTX-induced contraction of the vas deferens results from an indirect action mediated through NE release from the adrenergic nerve terminals, and that CTX causes an increase in Na+ permeability of the presynaptic membrane, which may play an important role in its releasing action. 134 Y. OHIZUMI The dose-contractile response curves for the action of NE and KC1 on the vas deferens were shifted to the left in a parallel manner by CTX, indicating that CTX caused supersensitivity (Ohizumi et al, 1982). The CTX-induced potentiation was abolished by TTX (5 X 10~7 M) or saxitoxin (5 X 10~7 M); it was inhibited by a Na+- deficient medium, but was not affected by phentolamine (10 6 M). After treatment with reserpine (2 mg/kg/day, twice) or cold storage (4°C, for 7 days), CTX-induced release of NE was completely prevented. CTX still markedly potentiated the response to NE and KC1 in the reserpinized preparation or the cold stored preparation. On the other hand, tissue Na content of the vas deferens was markedly increased by ouabain (1CT5 M), but was not affected by CTX at concentrations between 1 and 5 X 10~7 g/ml. However, in the presence of ouabain ( 10~5 M), Na content was increased 30% or more by CTX (5 X 10"7 g/ml). This increasing effect of CTX on Na content was abolished by TTX (10~6 M) (Ohizumi et al., 1982). These data suggest that CTX causes an increase in Na+ permeability across the TTX-sensitive Na channels of smooth muscle cells, and this may play an important role in its mechanism of poten- tiation. The responses of the guinea-pig vas deferens to MTX were different from those of CTX. MTX (10~9 to 3 X 10~8 g/ml) caused a dose-dependent slower contraction of the vas deferens (second phase) that followed the initial rapid phasic contraction (first phase) (Ohizumi et al., 1983). The second component of the MTX-induced contraction was markedly inhibited by phentolamine (10"6 M) and reserpine (2 mg/ kg/day, twice), whereas the first component remained unaffected. Both components were inhibited or abolished by verapamil ( 10"6 to 10"5 M) or Ca2+-free medium, but were not affected by atropine (10~6 M), chlorpheniramine (10 6 M), or TTX (10~6 M). The tissue Ca content of the vas deferens was increased by MTX (10~9 to 3 X 10~8 g/ml) in a dose-dependent manner. Furthermore, MTX (10~9 to 3 X 10~8 g/ml) caused a dose-dependent release of NE from the vas deferens, which was inhib- ited or abolished by verapamil (3 X 10~6 and 10~5 M) or Ca2+-free medium, but not by TTX (10~6 M}. In Na+-free medium, MTX still caused a marked increase in NE release from the tissue (Table II). It has been reported that MTX elicited Ca-depen- dent excitatory effects on neuronal cells (Takahashi et al., 1982, 1983; Freedman et al., 1984). These observations suggest that the major part of the first component of the MTX-induced contraction is the result of a direct action of MTX on the smooth muscle membrane, whereas the second component is primarily the result of indirect action mediated through the NE release from the adrenergic nerve terminals. The data also suggest that Ca2+, but not Na+ is indispensable for the action of MTX. In the isolated guinea-pig and rat ventricle strips, MTX (10~10 to 4 X 10~9 g/ml) caused a dose-dependent inotropic effect. The MTX-induced inotropic effect was nearly abolished by Co2+ (2 mM) or verapamil (3 X 10"7 M ), but was little affected by propranolol ( 10~6 M), reserpine (2 mg/kg, twice), or TTX (5 X 10~7 M) (Kobayashi et al., 1985). The tissue Ca-content of guinea-pig left atria was increased by MTX (2-3 X 10"8 g/ml), and this increase was markedly inhibited by Co2+ (2 mM) or verapamil (10~5 M). In rat myocardial cells, MTX (10~10 to 10~9 g/ml) induced an increase in the degree and the rate of contraction and subsequent arrhythmogenic actions (Kobayashi et al., 1985). These effects of MTX were antagonized by verapamil ( 10~6 M) and were completely inhibited by Ca2+-free medium. When the concentration of external Ca2+ was increased to 2 mM, MTX (10~8 g/ml) made all the rod cells turn into arrhythmically moving cells and then into rod-shaped cells much faster than was the case in the control medium. These results suggest that the excitatory effect of MTX on heart muscle is caused by a direct action on the cardiac muscle membrane, due mainly to an increase in Ca2+ permeability, possibly through some Ca channels. CIGUATOXIN AND MAITOTOXIN ON MUSCLE 135 TABLE II Effect ofMTX on the NE release from the guinea-pig vas deferens in the presence or absence ofTTX or verapamil Solution Amount of NE released3 Treatment ng/g tissue/30 min Normal solution None 22 + 4 MTX(3X 10-8g/ml) 1350 ±169 MTX(3X 10-8g/ml) + TTX(5X 10~7M)b 1397 ±205 MTX(3X lO'Vml) + verapamil (3 X 10~6M)b 770± 81* + verapamil (10~5M)b 548+ 62* Ca++-free solution0 None 18+ 3 MTX(3X 10"8g/ml) 97 ± 11 Na+-free solutionc None 832 ± 54 MTX(3X 10~8g/ml) 2973 ±134 MTX(3X 10-8g/ml) + verapamil (10-5M)b 2010± 110* a Mean + S.E. (n = 6). b TTX and verapamil were added 1 5 min before the application of MTX. c Tissue were incubated for 15 min in Ca++- or Na+-free solution before the application of TTX or verapamil. * Significantly different from MTX alone in each solution, P < .01 . Finally, scaritoxin, a hydrohobic compound isolated from a ciguatera fish Scams gibbus, has a releasing action of NE and ACh from adrenergic and cholinergic nerve endings, resulting in the contraction of the guinea-pig vas deferens and ileum, respec- tively (Tatsumi et ai, 1985). ACKNOWLEDGMENTS I am greatly indebted to Prof. T. Yasumoto of Tohoku University for his generous supply of MTX and to Prof. J. Scheuer and Dr. K. Tachibana of the University of Hawaii for providing CTX. LITERATURE CITED CATTERALL, W. A. 1980. Neurotoxins that act on voltage-sensitive sodium channels in excitable mem- branes. Ann. Rev. Pharmacol. Toxicol. 20: 1 5-43. CATTERALL, W. A. 1984. The molecular basis of neuronal excitability. Science 223: 653-661. FREEDMAN, S. B., R. J. MILLER, D. M. MILLER, AND D. R. TINDALL. 1984. Interactions of maitotoxin with voltage-sensitive calcium channels in cultured neuronal cells. Proc. Natl. Acad. Sci. USA 81: 4582-4585. HASHIMOTO, Y., ed. 1 979. Pp. 9 1 - 1 05 in Marine Toxins and Other Bioactive Marine Metabolites. Japan Scientific Society Press, Tokyo. KOBAYASHI, M., G. MIYAKODA, T. NAKAMURA, AND Y. OHizuMl. 1 985a. Ca-depcndent arrhythmogenic effects of maitotoxin, the most potent marine toxin known, on isolated rat cardiac muscle cells. Eur. J. Pharmacol. Ill: 121-123. KOBAYASHI, M., Y. OHIZUMI, AND T. YASUMOTO. 1985b. The mechanism of action of maitotoxin in relation to Ca2+ movements in the guinea-pig and rat cardiac muscle. Br. J. Pharmacol. 86: 385- 391. MARTIN, J. B., AND D. M. DOTY. 1949. Determination of inorganic phosphate. Anal. Chem. 21: 965-967. 136 Y. OHIZUMI MIYAKODA, G., AND T. NAKAMURA. 1982. Reversible effects of a high-molecular weight sulfhydryl re- agent on the contractile from adult rat heart. J. Biochem. 92: 1833-1 843. NARAHASHI, T. 1 974. Chemicals as tools in the study of excitable membranes. Physiol. Rev. 54: 8 1 3-889. NORTON, T. R., Y. OHIZUMI, AND S. SHIBATA. 1 98 1 . Excitatory effect of a new poiypeptide (anthopleurin- B) from sea anemone on the guinea-pig vas deferens. Br. J. Pharmacol. 74: 23-28. OHIZUMI, Y., AND S. SHIBATA. 1980. Mechanism of the excitatory action of palytoxin and N-acetylpaly- toxin in the isolated guinea-pig vas deferens. J. Pharmacol. Exp. Ther. 214: 209-2 1 2. OHIZUMI, Y., AND T. YASUMOTO. 1983. Contractile response of the rabbit aorta to maitotoxin, the most potent marine toxin. J. Physiol. (Lond.) 337: 7 1 1-72 1 . OHIZUMI, Y., Y. ISHIDA, AND S. SHIBATA. 1982. Mode of the ciguatoxin-induced supersensitivity in the guinea-pig vas deferens. J. Pharmacol. Exp. Ther. 221: 748-752. OHIZUMI, Y., A. KAJIWARA, AND T. YASUMOTO. 1983. Excitatory effect of the most potent marine toxin, maitotoxin, on the guinea-pig vas deferens. J. Pharmacol. Exp. Ther. 227: 199-204. OHIZUMI, Y., S. SHIBATA, AND T. TACHIBANA. Mode of the excitatory and inhibitory actions of ciguatoxin in the guinea-pig vas deferens. J. Pharmacol. Exp. Ther, 217: 475-480. TACHIBANA, K. 1980. Structural studies on marine toxins. Ph.D. Dissertation, University of Hawaii, Ho- nolulu, Hawaii. TAKAHASHI, M., Y. OHIZUMI, ANDT. YASUMOTO. 1982. Maitotoxin, a Ca2+ channel activator candidate. J. Biol. Chem. 257: 7287-7289. TAKAHASHI, M., M. TATSUMI, AND Y. OHIZUMI. 1983. Ca2+ channel activating function of maitotoxin, the most potent marine toxin known, in clonal rat pheochromocytoma cells. J. Biol. Chem. 258: 10944-10949. TATSUMI, M., A. KAJIWARA, T. YASUMOTO, AND Y. OHIZUMI. 1985. Potent excitatory effect of scaritoxin on the guinea-pig vas deferens, taenia caeci and ileum. J. Pharmacol. Exp. Ther. 235: 783-787. YASUMOTO, T. 1980. Ciguatera. Igaku no Aynmi 112: 886-892. Reference: Biol. Bull. 172: 137-143. (February, 1987) THE USE OF THE MOSQUITO BIOASSAY FOR DETERMINING THE TOXICITY TO MAN OF CIGUATERIC FISH R. BAGNIS, M. BARSINAS, C. PRIEUR, A. POMPON, E. CHUNGUE, AND A. M. LEGRAND Institut Territorial de Recherches Medicates, "Louis Malarde, " B. P. 30, Papeete, Tahiti, French Polynesia ABSTRACT Moderately well-correlated with the mouse bioassay, requiring a much smaller amount of raw fish flesh, and employing a faster chemical extraction procedure, the mosquito bioassay represents significant progress in detecting ciguatoxin in fish mus- cle. We use the mosquito bioassay to evaluate the approximate amount of ciguatoxin present in the remains of 25 fish of 1 7 different species involved in 99 documented cases of ciguatera poisoning. From the data provided by mosquito bioassay and hu- man symptomatology, we determine a dose-effect relationship for ciguatoxin (CTX) to man. Three average pathogenic doses (PD) are established. The minimum PD is estimated to be 0.6 ng CTX/kg, the PD50 to be 2 ng CTX/kg, and the PD100 to be 8 ng CTX/kg. These data suggest a MLD (minimal lethal dose) value of 20 ng CTX/kg. INTRODUCTION Ciguatera is the most frequently reported food-borne disease in French Polynesia. Its sporadic and scattered outbreaks, tied to natural and human phenomena, consti- tute a widespread and serious problem in these islands, both as a health hazard (Fig. 1) and as an obstacle to the exploitation of reef-fish resources (Fig. 2). As accurate an evaluation as possible of ciguatera levels is imperative for use by local authorities. Until a simple, reliable, rapid chemical assay, like the one recently proposed by Hokama (1985), is available, bioassays will continue to be widely used. Usually, small animals are used for bioassays. Thus, the mosquito bioassay (Pompon et ai, 1984), which has a moderate correlation to the cat bioassay and the mouse bioassay (Bagnis et al, 1985), represents significant progress in detecting ciguatoxin in fish muscle. The swiftness of the chemical extraction procedure (Pompon and Bag- nis, 1984) and of the mosquito response, as well as the small amount offish needed, allow us to check the ciguatoxicity of very small fish (Chungue et al, 1984) or of the remains of poisonous meals. The following study aims to establish, from the data taken from the mosquito bioassay and human symptoms, a relationship between the approximate amount of ingested ciguatoxin (Scheuer et ai, 1967) and its clinical effects (Bagnis ?/ al., 1979). MATERIAL AND METHODS Patients Patients either were seen in the consulting room of the Louis Malarde Medical Research Institute in Papeete or were hospitalized in the General Hospital in Papeete. Clinical symptoms were observed in all the poisoned people. 137 138 R. BAGNIS ET AL. C. I . R. C.I . R. 150 100 50 * * * * * * * * * * * * * * — * * * * — • « * * * * * HK__ * * * * — * * * » * * * » * * SOCIETY 5 TUAMOTU 17 GAMBIER 227 MARQUESAS 43 AUSTRAL 1 + # * * * V * * * * * * * * * * * * -* * * * * * » * * * * * * * * • « • . * * * * * * * * * * HOT I — Bn Aus tral 80 * *77| * * * Marquesas * * • « * • • • * • * * * * • * • * * • * 9 Tuamotu-G — ~ — - *~ "Society 1960 1970 1980 years FIGURE 1. Geographical distribution of the incidence of ciguatera per 1000 inhabitants (C.I.R.) in French Polynesia ( 1 960 to 1 984). Fish Fish belonged to 1 7 species at several trophic levels and were caught in various areas of French Polynesia. Each fish sample was responsible for at least one docu- mented case of ciguatera. The average amount of fish flesh eaten by an individual was evaluated by anamnesis and by weighing control rations. Rapid extraction of the ciguatoxin Ciguatoxin extraction procedures are detailed in Figures 1 and 2. A sample (8 g) of ground raw or cooked fish flesh was homogenized with 40 ml of acetone (twice) and centrifuged. The supernatants were concentrated and percolated through a car- tridge of diatomaceous silica (4 g) that was washed successively by n-hexane (45 ml) and diethyl ether (45 ml). After evaporation, the diethyl ether toxic residue was sub- jected to fast, low-pressure chromatography on silicic acid (1 g). Elution was per- formed successively by chloroform (25 ml) and a chloroform-methanol mixture (9:1:25 ml). The resulting eluate was evaporated. From 8 g offish flesh, 1.0 ± 0.2 mg of toxic residue was obtained within 1 h, regardless of the species or the toxicity of the fish. Mosquito bioassay The mosquitoes, weighing 1.6 ± 0.2 mg, belonged to the species Aedes aegypti. They were intrathoracically injected as described by Rosen and Gubler (1974). Each portion of extract obtained from 8 g of flesh was suspended in 1 00 /ul of phosphate- buffered saline (pH 7.4, 0.002 M) containing 0.5% gelatin and 5% heated calf serum (56°C for 30 min). A volume of 0.5 /il of two-fold serial dilutions was injected into eight groups often mosquitoes each. The tested doses ranged from 0.3 1 to 40 X 10~3 g of flesh per mosquito. A control group received diluent. Mortality was observed 1 MOSQUITO BIOASSAY 139 75 50 25 (A) i » (T) (S) Archipelagoes SOCIETY (S) TUAMOTU (T) GAMBIER (G) MARQUESAS (M) AUSTRAL (A) identified fishes 5 805 1 322 898 2 174 45 77/t Acanthurids (Aca. Balistids (Bal . ) Carangids (Car. ) Holocentrids Hoi Kyphosids (Kyp. ) Labrids (Lab.) Lethrinids (Let . ) Lut janids (Lut . ) Mugil ids (Mug . ) Mullids (Mul.)) Murenids (Mur . ) Scarids (Sea. ) • ) Scombrids (Sco. )• Serranids (Ser . ) Siganids (Sig. ) h . Sphyraenids (Sph. (Aca) (Bal) (Car) (Hoi) (Kyp) (Lab) (Let) (Lut) (Mug) (Mul) (Mur) (Sea) (Sco) (Ser) (Sig) (Sph) FIGURE 2. Geographical distribution of ciguateric fish in French Polynesia. h after injection. The method of Bliss (1938) was used to calculate LD50 in mosqui- toes. The fish ciguatoxicity was expressed in equivalent gram offish (fge) or in Aedes mosquito-units (AU). A fish contains 1 AU of ciguatoxin when five mosquitoes in a group often die after each has received extract equivalent to 7 mg offish. Dose-effect relationship for ciguatoxin Using as a baseline the MLD of pure ciguatoxin for mice [i.e., 0.45 g/kg according to Tachibana (1980)], the correspondence between cat bioassay (Bagnis and Fevai, 1971), mouse bioassay (Yasumoto et al, 1977), and mosquito bioassay have been previously established (Bagnis et al., 1985). They are as follows: 1 AU = 17.5 X 10~5 mouse-units (MU) = 8.8 X 10"5 cat units (CU) -3 = 1.6X 10-JngCTX. RESULTS AND DISCUSSION Epidem i ological features Twenty-five fish, eaten by 142 individuals, induced ciguatera symptoms in 99 people (see Table I). Among these, 30 were hospitalized for a period ranging from 24 hours to 10 days. The fish involved in the most severe cases were carnivorous fish from four species: Lethrinus miniatus, Cheilinus undulatus, Aprion virescens, and Sphyraena barracuda. Clinical features The most frequently observed symptoms are shown in Table II. Sensory distur- bances were much more frequent than neuromotor symptoms. 140 R. BAGNIS ET AL. TABLE I Resitlls of the mosquito bioassayfor 25 fish involved in documented cases ofciguatera poisoning, with the corresponding lexicological data provided from the mouse bioassay and the cat bioassay Fish Man Mouse Cat Mosquito (a) (b) (c) (d) (e) (0 (g) Lethrinus miniatus (Tautira) 6/8 3 0.23 0.12 1.02 6.9 1.54 Lethrinus miniatus (Raiatea) 8/10 1.00 0.18 2.70 2.6 0.58 Lethrinus miniatus ( Vairao) 3/5 0.35 0.10 3.11 2.3 0.50 Lethrinus miniatus (Paea) 9/9 3 0.90 0.14 1.64 4.3 0.96 Cheilinus undulatus (Arue) 6/6 6 1.20 0.28 0.89 7.9 1.77 Cheilinus undulatus (Paea) 6/6 2 2.25 0.23 1.44 4.9 1.10 Cephalopholis argus (Tautira) 7/7 5 0.50 0.10 2.30 3.0 0.68 Cephalopholis argus (Teahupoo) 2/2 0.60 0.15 1.61 4.4 0.98 Epinephelus microdon (Mataiva) 4/7 0.18 0.11 2.05 3.4 0.77 Sphyraena barracuda (Moorea) 7/7 0.60 0.12 2.02 3.5 0.78 Aprion virescens( Marquesas) 6/6 6 0.14 0.83 8.4 1.90 Naso unicornis (Mataiva) 2/2 0.52 0.10 2.65 2.6 0.57 Scarusjonesi (Hao) 5/5 2.25 0.14 0.98 7.1 1.60 Scarusjonesi (Mataiva) 3/3 1.80 0.18 1.53 4.6 1.03 Lutjanus bohar (Cook) 3/3 3 0.29 0.18 2.37 3.0 0.67 Lutjanus bohar (Tetiaroa) 6/8 0.45 0.10 2.23 3.1 0.71 Monotaxis grandoculis (Mataiva) 2/2 0.36 1.45 4.8 1.09 Sphyraena barracuda (Rairoa) 5/5 2 0.74 0.10 2.22 3.3 0.74 Lutjanus gibbus (Papeete) 2/5 0.84 6.30 1.1 0.26 Caranx melampygus (Punaauia) 1/6 0.18 0.01 3.14 2.2 0.50 Plectropomus leopardus (Taenga) 1/8 0.09 0.10 4.77 1.5 0.33 Gymnosarda unicolor (Gambier) 2/5 0.30 0.02 6.19 1.1 0.26 Lutjanus gibbus (Papeete) 1/6 0.09 6.66 1.1 0.24 Crenimugil crenilabis (Mataiva) 1/5 0.15 7.95 0.9 0.19 Scarus gibbus (Hao) 1/5 0.10 7.11 1.0 0.22 LEGEND: (a) Number of patients/number of consumers. (b) Number of hospitalized patients. (c) Dose of ciguatoxin in nanograms per gram offish (CTX:ng/fg) estimated by the mouse bioassay. (d) CTX (ng/fg) estimated by the cat bioassay. (e) LD50 in 10~3 fge per 1.6 mg mosquito. (f) LD50 in mosquito-units (AU). (g) CTX (ng:fg) estimated by the mosquito bioassay. Of the 30 hospitalized patients, 24 displayed chiefly cardiovascular symptoms. Bradycardia with a pulse ranging from 40 to 60 beats/min (21% of cases), systolic blood pressure of 100 mm Hg or less (18% of cases), abnormalities of electrocardio- grams ( 1 5% of cases) and cardiac distress (6% of cases) were observed. Twelve patients were hospitalized for other severe symptoms which may or may not have been associ- ated with cardiovascular disturbances: ataxia (10), paresis (2), dyspnea (8), and dys- uria(5). MOSQUITO BIOASSAY 141 TABLE II Frequency of symptoms in 99 documented cases ofciguatera % of patients Symptom with finding* 1. Paresthesia of the extremities 92 2. Circumoral pareshesia 92 3. Burning or pain to skin on contact with cold water 9 1 4. Arthralgia 88 5. Myalgia 86 6. Diarrhea 74 7. Asthenia 66 8. Nausea 56 9. Pruritus 48 10. Headache 45 1 1 . Abdominal pain 44 12. Chills 43 13. Vomiting 40 14. Bradycardia 38 15. Vertigo 33 16. Perspiration 32 17. Ataxia 28 18. Hypotension 28 19. Dental pain 24 20. Tremor 18 21. Salivation 18 22. Watery eyes 17 23. Dyspnea 16 24. Cardiac dysrhythmia 15 25. Neck stiffness 13 26. Dysuria 13 27. Paresis 8 28. Cardiac distress 6 29. Hiccoughs 2 * Percentages in table refer to all 99 documented ciguatera cases. Percentages in text refer to symptoms observed in the 30 hospitalized cases. Toxicological features The data on fish ciguatoxicity, as determined by the mosquito bioassay, are shown in Table I. Also indicated are the corresponding doses of ng CTX, evaluated from the cat bioassay and the mouse bioassay results when available. We could not correlate a high level of ciguatoxicity in a fish to a severe syndrome in all its consumers. For a more accurate analysis, the toxic fish were separated into two classes. Eigh- teen fish, poisonous for at least 50% of their consumers, were labelled >PD50 fish. Their ciguatoxicity ranged from 0.83 X 10"3 fge (inferior toxicity, IT) to 6.3 X 10~' fge (superior toxicity, ST) with a mean toxicity (MT) evaluated to be 1 .83 X 10'3 fge, from 1 . 1 to 8.4 AU (MT = 4.2), and from 0.26 to 1 .9 ng CTX (MT = 1 ). Seven fish, poisonous for less than 50% of the consumers, were labelled LD50 fish) Maximal 2 18 4.5 X 1(T2 0.8 3.6 X 1Q-3 5.7 8.2 Mean 4X 10"' 3.2 2X 1(T2 0.4 1.6X10"3 2.6 2.1 Minimal 6x icr2 0.5 1.5 X 10"2 0.3 0.9 X 10~3 0.5 0.8 (Batchof7/#/., 1981; Baden etai, 1981). Okadaic acid was isolated from the dinoflagellate Prorocentrum lima (Murakami et ai, 1982) as well as from the sponges in the genus Halichondria (Tachibana et ai, 198 1 ), and brevetoxin was extracted from the dinoflagellate causing red tide, Ptycho- discus brevis (Gymnodinium breve) (Lin et ai, 1981; Baden et ai, 1 98 1 ). These poly- ether toxins appear to have an effect on the membranes of nerve and muscle tissues and to induce changes in ion permeability of the cells (Rayner, 1972; Miyahara et ai, 1979;Ohizumi#a/., 1982). A major area of study initiated in the past decade has been the development of sensitive and specific assay methods for the assessment of these toxins in contami- nated fish tissues. Earlier assays included: ( 1 ) the feeding of fish tissues to cats and mongoose (Banner et ai, 1961; Bagnis, 1973); and (2) the injection of crude extracts into mice, chicks, and, more recently, mosquitoes (Kosaki et ai, 1968; Yasumoto et ai, 1971;Kimurae/a/., 1982;Chungue^a/., 1984). These tests lacked both sensitiv- ity and specificity. More recent studies have taken the immunological approach, us- 144 RAPID ENZYME IMMUNOASSAY STICK TEST 145 ing antibodies prepared in sheep and rabbit, following immunization with conjugates of the polyether toxins (Hokama et ai, 1977; Kimura et al, 1982; Hokama et al, 1983, 1984, 1985; Baden effl/., 1984). We have developed and assessed a rapid, simple, specific, and sensitive stick test to detect ciguatoxin and related polyether toxins in contaminated fish tissues. The fishes examined included: ( 1 ) clinically implicated fishes from ciguatera poisonings; (2) fishes from corresponding catches obtained in the same vicinity with the impli- cated fishes; (3) portions of non-toxic fishes that were consumed without incident; and (4) fishes from the nearshore waters of Hawaii where toxicity occasionally occurs. MATERIALS AND METHODS Sources offish samples The following fishes were obtained through the courtesy of the Department of Health (DOH), State of Hawaii: Clinically implicated fish samples from ciguatera poisonings — 1, Acanthocybium solandri; 1, Aprion virescens; 3, Caranx species; 2, Cephalopholis argus; 1 , Sphyraena species; and 1 , grouper extract. Corresponding catch fishes associated with the fishes implicated in ciguatera poisonings — 93, Cteno- chaetus strigosus; 16, Mulloidichthys species; 6, Cephalopholis argus; 3, Cheilimis rhodochrous; and 1 , Aphareus furcatus. The following fishes were obtained through the courtesy of the Department of Land and Natural Resources (DLNR), State of Hawaii: Non-toxic consumed fishes — 43, Seriola dumerili; 6, Elagatis bipinnulatis; 5, Sphyraena species; and 1, Caranx species. Reef fishes from the nearshore waters of Oahu and Hawaii Island — Cheilinus rhodochrous; Kyphosus cinerescens; Ctenochaetus strigosus; Littjanus kasmira; Acanthurus species; and Myripristis species. Non-toxic consumed Thunnus thynnus was obtained from commercial sources (supermarket). Source of purified ciguatoxin Purified ciguatoxin isolated from livers of moray eels (Nukina et ai, 1984) was obtained through the courtesy of Professor P. J. Scheuer of the Department of Chem- istry, University of Hawaii, Honolulu, Hawaii. Stick test reagents The reagents for the stick test were similar to those used in the enzyme immunoas- say reported previously (Hokama et al., 1983). These included: (a) methanol fixative, 0.3% hydrogen peroxide (H2O2) in absolute methanol, prepared fresh daily; (b) Tris buffer A, 0.05 M Tris (hydroxymethyl)aminomethane, pH 7.5 ± 0.05, with 0.1% human serum albumin (HSA) and 0.01% sodium azide; (c) Tris buffer B, prepared as in A, but without HSA and sodium azide; (d) sheep-anti-CTX coupled to horseradish peroxidase (HRP, Type VI, RZ:3:3, Sigma Chemical Co., St. Louis, MO) according to the one-step glutaraldehyde method (Voller et ai, 1980), stored in aliquots of 1 ml at -20°C. until ready for use; (e) substrate, 25 ml of 0.3% H2O2 in Tris buffer B added to 10 mg of 4-chloro-l-naphthol previously dissolved in 0.125 ml of absolute methanol, mixed thoroughly then filtered through Whatman # 1 filter paper, prepared fresh just before use; (f) bamboo sticks (length, 200.0 mm; diameter, 2.5 mm; Mum's Taisei of Hawaii, Honolulu) with the skewer ends coated with Liquid Paper (The 146 Y. HOKAMA ET AL. Gillette Co., Rockville, MD). Any excess Liquid Paper was removed in the prepara- tion of the coat, then the coating was allowed to dry thoroughly before use. Stick test procedure Fish were sampled by inserting the skewer end of the stick coated with the Liquid Paper into the dorsal-anterior and ventral-posterior sections of one or both sides of the fish. Each stick was inserted 5 times into the flesh at 1 s/insertion. Each fish was examined with six sticks; three in the dorsal-anterior and three in the ventral-poste- rior portions of the side of the fish. The stick was air dried and then immersed into the fixative for 1 s without shaking. The excess solution was blotted onto tissue paper and the stick was air dried. The stick was then washed in Tris buffer B thoroughly with gentle shaking for 10 s and the excess solution blotted onto tissue paper. The stick was then immersed for 1 5 to 30 s without agitation into the sheep-anti-CTX- HRP solution previously diluted 1 :200 with Tris buffer A just prior to use. The excess antibody conjugate was blotted onto tissue paper. The stick was then washed in two changes of Tris buffer B, 10 s each with gentle shaking. After the excess buffer was blotted, the stick was immersed into 0.3 ml of the substrate. The tube containing the substrate and the stick was shaken, then incubated for 10 min at room temperature. The intensity of the color was compared with a color scheme as follows: 0, essentially no color; 1.0+, slightly bluish-purple; 1.5 to 2.0+, lightly bluish-purple; 2.0 to 2.5+, moderately bluish-purple; 3.0 to 5.0+, moderately to intensely bluish-purple. The results of these reactions were scored as follows: 0 to 2.0+, negative; 2.1 to 2.4+, borderline; and values greater than 2.5+, positive. The fractional values were the result of averaging several sticks per sample. The reagents for the test were arranged in the following sequence: (a) fixative; (b) Tris buffer B wash 1 ; (c) sheep-anti-CTX-HRP; (d) Tris buffer B wash 2; (e) Tris buffer B wash 3; and (f) substrate. Analysis of purified ciguatoxin by the stick test Purified ciguatoxin in concentrations of 0, 1, 5, and 25 ng/ml in absolute metha- nol were prepared. The coated sticks were immersed for 2 s in the ciguatoxin-metha- nol solutions. The sticks were then air dried and examined by the stick test procedure. The fixation step was not performed for the purified toxin. Several sticks were used for each concentration of the toxin solution. The mean and standard deviation of the stick values were obtained for each toxin concentration and plotted as shown in Fig- ure 1. Examination of the gut contents o/Ctenochaetus strigosus Smears on microslides from the viscera of C. strigosus were prepared from each of the 93 specimens obtained from Kauai as part of the corresponding catches. The smears were suspended in 50% glycerol phosphate buffer saline with a cover slip. Each smear was examined at 400X magnification with the phase and light microscope. Twenty fields per smear were examined and the average number of Gambierdiscus toxicus per field determined. All statistical evaluation was determined by the method of Tallarida and Mur- ray (1981). RAPID ENZYME IMMUNOASSAY STICK TEST 147 Sheep Anti-CTX-HRP 5.0- 4.0- o> _3 D 3.0- 2.0- 1.0- 00 00 10 3.0 5.0 250 ng. Ciguatoxin ml. Methanol FIGURE 1. Analysis of purified ciguatoxin in methanol by the stick test. The y-axis represents the stick test values and the x-axis the concentration of ciguatoxin. The means (O) and standard deviations (vertical lines) are presented. RESULTS Stick test analysis of positive fishes from clinically documented ciguatera poisoning and negative consumed fishes The assessment of the validity of the stick test in the detection of toxic fishes is summarized in Figure 2. Sixteen clinically implicated raw_and cooked (tissues, soup, and gravy) fish samples examined by the stick test gave a x ± S.D. of 3.3 ± 0.7 while 60 non-toxic consumed fishes gave a x ± S.D. of 1.2 ± 0.5 stick test value (Fig. 2). The difference between these two categories was highly significant with P < 0.005. Examination of 2 samples of Thunnus thynnus gave a stick test value of 1.1 ± 0.8. This value was significantly different from the positive values with P < 0.0 1 . Thunnus thynnus, a pelagic fish, has not been implicated in ciguatera poisoning. On the basis of this study, the interpretation of the stick test values has been derived and desig- nated as follows: 0-2.0+, negative; 2.1 -2.4+, borderline; and values greater than 2.5+ as positive (Fig. 2). 148 Y. HOKAMA ET AL. STICK TEST VALUE 5.0- 4.0- 3-0- 2.0- 1.0- 0.0- POSITIVE BORDERLINE NEGATIVE Num ber of Sample 16 N* 96 I 60 360 I 2 30 FIGURE 2. Comparison of clinically documented positive fishes associated with ciguatera poisoning (A) and non-toxic consumed negative fishes (B) by the stick test. (C) Represents the Thunnus thynnus results. The y-axis represents the stick test values and the x-axis the number of samples, n* = the number of determinations. Stick test analysis of corresponding catches of fishes and study of the nearshore reef fishes of Hawaii The results of examining, by the stick test, several species of fishes obtained as corresponding catches with fishes implicated in ciguatera poisonings are shown in Table I. Of the 120 fishes from 4 species examined, 69.2% were considered toxic by the stick test. The positive group also included those fishes in the borderline category. This high percentage is due to Ctenochaetus strigosus, which caused a large outbreak of ciguatera poisoning on Kauai in August, 1984 (see Table II). The gut contents of all 93 samples contained 0. 1 to 1.5 Gambierdiscus toxicus per field at 400X magnifi- cation (Fig. 3). Summary of stick test evaluation of nearshore reef fishes from Oahu, Hawaii Island, and Kauai These studies are summarized in Tables II, III, and IV. All four species of fishes examined from Oahu came from the Leeward side of the nearshore waters (Table III). Fishes caught in these areas generally have the highest frequency of ciguatera poisoning, in contrast to the Windward side, where essentially no ciguatera poisoning RAPID ENZYME IMMUNOASSAY STICK TEST 149 TABLE I Examination of several species of fishes from corresponding catches of ciguatera poisoning in Hawaii, 1984 Stick test Negative Positive Species Total number No. offish Ctenochaetus strigosus (Kole, surgeon fish) 93 18 75 Cepha/opholis argus (Roi, grouper) 6 4 2 Mulloidichthvs species (Weke, goat fish) 16 15 1 Cheilinus rhodochrous (Po'ou, wrasse) 3 0 3 Aphareus furcatus (Wahanui, snapper) 1 0 1 Total 120 37 81 % - 30.8 69.2 has occurred. A high percentage of toxicity is shown by C. strigosus (43.5%) and L. kasmira (36.5%). Note that the herbivore (C. strigosus) gave a higher percentage of toxic samples than either the carnivore (L. kasmira) or the two other species exam- ined (Acanthurus and Myripristis). The positive category includes the borderline samples. Table IV summarizes the results of the samples of three species from Hawaii Is- land. A high percentage of positives is found in C. strigosus (45.0%) as compared to K. cinerescens (0.0%) and C. rhodochrous (20.8%). TABLE II Comparison o/"Ctenochaetus strigosus examined by the stick test from three different islands: Hawaii, Oahu, and Kanai Stick test Negative Positive Source Total number No. offish Kauai* 93 18 % 19.4 75 80.6 Hawaii** 51 28 % 55.0 23 45.0 Oahu*** 108 61 % 56.5 47 43.5 * Samples from corresponding catches of ciguatera poisoning outbreak, August, 1984. ** Samples obtained in areas associated with ciguatera outbreaks from Cephalopho/is argus and Chei- linus rhodochrous, April-August, 1984. *** Samples from Barbers Point and Ewa, routine monthly survey of Harbor development and control site, 1984. 150 Y. HOKAMA ET AL. FIGURE 3. A typical photograph of Gambierdiscus toxicus from the gut of Ctenochaetus strigosus obtained from the nearshore waters of Kauai (400X magnification). The comparison of C. strigosus from the three different islands by the stick test procedure is shown in Table IV. The highest percentage of toxicity was shown in samples from Kauai (80.6%). Samples from Oahu and Hawaii Island were nearly identical, 43.5% and 45.0%, respectively. This is understandable, since the samples from Kauai were part of the corresponding catches of a large ciguatera outbreak due to C. strigosus. The Hawaii Island samples of C. strigosus were caught on the Leeward side in April of 1984 just prior to small outbreaks of ciguatera poisoning that contin- RAPID ENZYME IMMUNOASSAY STICK TEST 151 TABLE III Summary of stick test assessment of fishes from Oahu Stick test Negative Positive Species Total number No. offish Ctenochaetus strigosus 108 61 47 (Kole, surgeon fish) % 56.5 43.5 Lut janus kasmira 96 61 35 (Taape, blue-line snapper) % 63.5 36.5 Acanthurus species 16 19 1 (Palani, surgeon fish) % 76.0 24.0 Myripristis species 12 9 3 (Menpachi, squirrel fish) % 75.0 25.0 Total 241 149 91 % - 61.8 38.2 ued throughout 1984. The fishes implicated included Cephalopholis argus, Cteno- chaetus strigosus, and Cheilinus rhodochrous. The Oahu samples of C. strigosus -were obtained during the autumn of 1984 from the DLNR Barbers Point (Leeward side of Oahu) study. No incidence of ciguatera poisoning was reported during this period (an outbreak occurred in January, 1 985, in this area). In part, this may be attributable to the lack of fishing in this area due to the turbid water conditions caused by the deep harbor dredging and construction at Barbers Point. Microscopic examination of viscera smears from C. strigosus (Kauai) All 93 samples of C. strigosus showed Gambler discus toxicus in their gut contents when smears were examined with the phase microscope at 400X magnification. The range of dinoflagellates/field was 0. 1 to 1 .5. A typical example of G. toxicus is shown in Figure 3. TABLE IV Summary of stick test assessment of fishes from Hawaii Stick test Negative Positive Species Total number No. offish Ctenochaetus strigosus 51 28 23 (Kole, surgeon fish) % 55.0 45.0 Kyphosus cinerescens 28 28 0 (Nenue, rudder fish) % 100.0 0.0 Cheilinus rhodochrous 24 19 5 (Po'ou, wrasse) % 79.2 20.8 Total 103 75 28 % _ 72.8 27.2 152 Y. HOKAMA ET AL. DISCUSSION The results of this study show the feasibility of adapting the enzyme immunoassay to a simple and rapid stick test. This procedure retained its sensitivity and specificity using the conventional heterologous sheep-anti-CTX-HRP (Hokama et al, 1977) or the recently prepared monoclonal antibodies to toxic polyethers (Hokama et al, 1985). The success of the procedure is probably due to the selective adsorption or attraction of the lipid toxins in fish tissues by one or more constituents in the Liquid Paper coated onto the bamboo stick. The latter alone without coating had no activity in the assay. Coated sticks alone have shown essentially little activity with values equal to or below the normal value (less than 2.0+). This is attributable in part to the non-specific binding of the antibody enzyme conjugate. We compared the clinically implicated toxic fishes and the non-toxic consumed fishes with the stick test and demonstrated significant differences in the mean and standard deviation of the stick test values between these two groups with P < 0.005. These same samples tested by the enzyme immunoassay (Hokama et al., 1983, 1984, 1985) agreed with the stick test values presented in this report. Hokama et al., (1984) recently showed that the sheep-anti-CTX-HRP detected ciguatoxin and structurally related polyether toxins such as okadaic acid and brevetoxin in a competitive enzyme immunoassay procedure. The stick test also interacted with purified ciguatoxin in a dose-responsive manner (Fig. 1 ). Examination of the fishes from corresponding catches showed a high level of tox- icity (80.6% positive), especially Ctenochaetus strigosus from the island of Kauai dur- ing August, 1984, when an outbreak of ciguatera poisoning occurred involving 14 individuals. At least one and up to four C. strigosus were consumed by each individ- ual. The viscera of all 93 specimens retrieved from the markets by the DOH contained Gambierdiscus toxicus. Outbreaks of ciguatera poisoning occurred in Hawaii during most of 1984, beginning in the early spring months. A moderate to high level of toxicity was also noted in C. strigosus (45.0%); with a slightly lower level in Cheilinus rhodochrous (2 1 .0%). However, Kyphosus cinerescens essentially showed no toxicity though caught in the same area. The toxicity patterns of Oahu appeared similar to that of the island of Hawaii (Tables II and IV). The radioimmunoassay (Hokama et al, 1977), the enzyme immunoassay (Ho- kama etal, 1983, 1984), and the stick test (Hokama, 1985)allusetheimmunological approach for the detection of ciguatoxin and related polyether toxins directly from fish tissues. The stick test differs from the former two tests in its simplicity and rapid- ity. The stick test also requires no capital equipment, and it can be used at home or in the field. Such a system also may be used for other enzyme immunoassays, perhaps using other coats to selectively adsorb the antigenic or haptenic components onto the stick. Furthermore, the stick test does not require the extraction of tissues, as required for the mouse bioassay ( Yasumoto et al, 1971; Kimura etal, 1982) and the recently described mosquito assay (Chungue et al, 1984). However, the stick test is also capa- ble of examining fish tissue extracts. In addition, the stick test can be performed rap- idly, while retaining its sensitivity and specificity, for mass screening of fishes on a large commercial level with minimal costs. In summation, the stick test is simple, rapid, and inexpensive, with sufficient spec- ificity and sensitivity to evaluate the levels of ciguatoxin and related polyether toxins in fishes suspected of potential ciguatera poisoning. ACKNOWLEDGMENTS We thank R. Shimizu and L. Oda for their excellent technical work. We also acknowledge the support of the Department of Land and Natural Resources, State RAPID ENZYME IMMUNOASSAY STICK TEST 153 of Hawaii, and especially the members of the Division of Aquatic Resources for the procurement of the reef fish samples. We are indebted to the Department of Health, State of Hawaii, for the clinical samples. LITERATURE CITED ADACHI, R., AND Y. FUKUYO. 1979. The thecal structure of a marine toxic dinoflagellate Gambierdiscus toxicns gen. et sp. nov collected in a ciguatera endemic area. Bull. Jpn. Soc. Sci. Fish. 45: 67-7 1 . BADEN, D. G., T. J. MENDE, W. LIGHTER, AND L. WELLHAM. 1981. Crystallization and toxicology of T34: A major toxin from Florida's red tide organism (Ptychodiscus brevis). Toxicon 19: 455-462. BADEN, D. G., T. J. MENDE, J. WALLING, AND D. R. SCHULTZ. 1984. Specific antibodies directed against toxins of Ptychodiscus brevis (Florida's red tide dinoflagellate). Toxicon 22: 783-790. BAGNIS, R. A. 1973. Le ciguatera aux iles Marquises: aspects cliniques et epidemiologiques. Bull. World Health Org. 49: 67-73. BANNER, A. H., S. SASAKI, P. HELFRICH, C. B. ALENDER, AND P. J. SCHEUER. 196 1 . Bioassay of ciguatera toxin. Nature 189: 229-230. CHUNGUE, E., R. A. BAGNIS, AND E. PARC. 1 984. The use of mosquito (Aedes aegypti) to detect ciguatoxin in surgeon fishes (Ctenochaetus striatus). Toxicon 22: 161-164. HOKAMA, Y., A. H. BANNER, AND D. B. BOYLAN. 1977. A radioimmunoassay for the detection of cigua- toxin. Toxicon 15: 317-325. HOKAMA, Y., M. A. ABAD, AND L. H. KJMURA. 1983. A rapid enzyme immunoassay for the detection of ciguatoxin in contaminated fish tissues. Toxicon 21: 817-824. HOKAMA, Y., L. H. KJMURA, M. A. ABAD, L. YOKOCHI, P. J. SCHEUER, M. NUKINA, T. YASUMOTO, D. G. BADEN, AND Y. SHIMIZU. ( 1 984). An enzyme immunoassay for the detection of ciguatoxin and competitive inhibition of related natural polyether toxins. Pp. 307-320 in Seafood Toxins, ACS Symposium Series 262, E. P. Ragelis, ed. American Chemical Society, Washington, DC. HOKAMA, Y. 1985. A rapid simplified enzyme immunoassay stick test for the detection of ciguatoxin and related polyethers from fish tissues. Toxicon 23: 939. KJMURA, L. H., Y. HOKAMA, M. A. ABAD, M. OYAMA, AND J. T. MIYAHARA. 1982. Comparison of three different assays for the assessment of ciguatoxin in fish tissue: Radioimmunoassay, mouse bioassay and in vitro guinea pig atrium assay. Toxicon 20: 907-9 12. KOSAKI, T. J., B. J. STEPHENS, AND H. H. ANDERSON. 1 968. Marine toxins from the Pacific: III. Compara- tive bioassay of ciguatoxin(s) in mouse and chicken. Proc. West. Pharmacol. Soc. 11: 126-128. LIN, Y. Y., M. RISK, S. M. RAY, D. V. ENGEN, J. CLARDY, J. GOLIK, J. C. JAMES, AND K. NAKANISHI. 1981. Isolation and structure of brevetoxin B from red tide dinoflagellate Ptychodiscus brevis (Gymnodinium breve). J. Am. Chem. Soc. 103: 6773-6775. MIYAHARA, J. T., C. K. AKAU, AND T. YASUMOTO. 1979. Effects of ciguatoxin and maitotoxin on the isolated guinea pig atria. Res. Comm. Chem. Pathol. Pharmacol. 25: 177-180. MURAKAMI, Y., Y. OSHIMA, AND T. YASUMOTO. 1982. Identification of okadaic acid as a toxic compo- nent of a marine dinoflagellate Prorocentrum lima. Bull. Jpn. Soc. Sci. Fish. 48: 69-72. NUKINA, M., L. M. KOYANAGI, AND P. J. SCHEUER. 1984. Two interchangeable forms of ciguatoxin. Toxicon 22: 169-176. OHIZUMI, Y., Y. ISHIDA, AND S. SHIBATA. 1982. Mode of ciguatoxin induced supersensitivity in the guinea pig vas deferens. J. Pharmacol. Exp. Therap. 221: 748-752. RAYNER, M. D. 1972. Mode of action of ciguatoxin. Fed. Proc. 31: 1 139-1 145. SCHEUER, P. J., W. TAKAHASHI, J. TSUTSUMI, AND T. YOSHIDA. (1967). Ciguatoxin: isolation and chemi- cal nature. Science 155: 1267-1268. TACHIBANA, K. 1980. Structural Studies on Marine Toxins. Ph.D. Thesis, University of Hawaii, Manoa, Honolulu, Hawaii. TACHIBANA, K., P. J. SCHEUER, Y. TSUKITANI, H. KJKUCHI, D. V. ENGEN, J. CLARDY, Y. GOPICHAND, AND F. J. SCHMITZ. 1981. Okadaic acid, a cytotoxic polyether from two marine sponges of the genus Halichondria. J. Am. Chem. Soc. 103: 2467-2469. TALLARIDA, R. J., AND R. B. MURRAY. 1981. Manual of Pharmacologic Calculations with Computer Programs. Springer- Verlag, New York. 1 50 pp. VOLLER, A., A. BARTLETT, AND D. E. BIDWELL. 1978. Enzyme immunoassay with special reference to ELISA techniques. J. Clin. Pathol. 31: 507-520. YASUMOTO, T., Y. HASHIMOTO, R. A. BAGNIS, J. E. RANDALL, AND A. H. BANNER. 1971. Toxicity of surgeon fishes. Bull. Jpn. Soc. Sci. Fish. 37: 724-734. YASUMOTO, T., I. NAKAJIMA, R. A. BAGNIS, AND R. ADACHI. 1 977. Finding of a dinoflagellate as a likely culprit of ciguatera. Bull. Jpn. Soc. Sci. Fish. 43: 1021-1026. PHARMACOLOGY AND CHEMISTRY TACHIBANA, KAZUO, MANABU NUKJNA, YONG-GOE JOH, AND PAUL J. SCHEUER Recent developments in the molecular structure of ciguatoxin 122 YASUMOTO, TAKESHI, NAOKO SEINO, YASUTAKA MURAKAMI, AND MICHIO MURATA Toxins produced by benthic dinoflagellates 128 OHIZUMI, YASUSHI Pharmacological actions of the marine toxins ciguatoxin and maito- toxin isolated from poisonous fish 132 ASSAY FOR CIGUATOXINS BAGNIS, R., M. BARSINAS, C. PRIEUR, A. POMPON, E. CHUNGUE, AND A. M. LEGRAND The use of the mosquito bioassay for determining the toxicity to man of ciguateric fish , 1 37 HOKAMA, Y., L. K. SHIRAI, L. M. IWAMOTO, M. N. KOBAYASHI, C. S. GOTO, AND L. K. NAKAGAWA Assessment of a rapid enzyme immunoassay stick test for the detection of ciguatoxin and related polyether toxins in fish tissues 144 CONTENTS BEHAVIOR GLEESON, RICHARD A., MICHAEL A. ADAMS, AND AMOS B. SMITH, III Hormonal modulation of pheromone-mediated behavior in a crus- tacean 1 ZIMMER-FAUST, RICHARD K. Crustacean chemical perception: towards a theory on optimal chemore- ception 10 DEVELOPMENTAL ECOLOGY McEowARD, LARRY R., AND RICHARD R. STRATHMANN The body plan of the cyphonautes larva of bryozoans prevents high clearance rates: comparison with the pluteus and a growth model .... 30 GENERAL BIOLOGY MULLER-PARKER, G., AND R. L. PARDY Response of green hydra to feeding and starvation at four irradiances . . 46 VOGEL, STEVEN Flow-assisted mantle cavity refilling in jetting squid 61 ^ -/ ^ ' : ^^ PHYSIOLOGY ENGEL, DAVID W. Metal regulation and molting in the blue crab, Callinectes sapidus: cop- per, zinc, and metallothionein . ^ -f. 69 ^ 7^ ^ SHORT REPORT SNYDER, KEITH, AND STACIA MOFFETT Loss and recovery of locomotor behavior after CNS lesions in the snail Melampus bidentatus r,'. 83 * \ FIRST INTERNATIONAL WORKSHOP ON CIGUATERA Papers from the first International Workshop on Ciguatera held on 5 and 6 June 1986 at the New England Biolabs in Beverly, Massachusetts. Workshop sponsored by the New England Biolabs Foundation and the M.I.T. Sea Grant Program. INTRODUCTION ANDERSON, DONALD M., AND PHILIP S. LOBEL The continuing enigma of ciguatera 89 ORGANISMS AND THEIR ENVIRONMENT DURAND-CLEMENT, M. Study of production and toxicity of cultured Gambierdiscus toxicus . . 108 Continued on Cover Three Volume 172 Number 2 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board RUSSELL F. DOOLITTLE, University of California at San Diego WILLIAM R. ECK.BERG, Howard University ROBERT D. GOLDMAN, Northwestern University C. K. GOVIND, Scarborough Campus, University of Toronto JUDITH P. GRASSLE, Marine Biological Laboratory MICHAEL J. GREENBERG, C. V. Whitney Marine Laboratory, University of Florida MAUREEN R. HANSON, Cornell University JOHN E. HOBBIE, Marine Biological Laboratory LIONEL JAFFE, Marine Biological Laboratory HOLGER W. JANNASCH, Woods Hole Oceanographic Institution WILLIAM R. JEFFERY, University of Texas at Austin GEORGE M. LANGFORD, University of North Carolina at Chapel Hill GEORGE D. PAPPAS, University of Illinois at Chicago SIDNEY K. PIERCE, University of Maryland HERBERT SCHUEL, State University of New York at Buffalo VIRGINIA L. SCOFIELD, University of California at Los Angeles School of Medicine LAWRENCE B. SLOBODKJN, State University of New York at Stony Brook JOHN D. STRANDBERG, Johns Hopkins University DONALD P. WOLF, Oregon Regional Primate Center SEYMOUR ZIGMAN, University of Rochester Editor: CHARLES B. METZ, University of Miami APRIL, 1987 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA Marine Biological Laboratory LIBRARY ?V MAY 1 1 1987 :' \ Woods -Hole, Mass. 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 THE BIOLOGICAL BULLETIN, Marine Bio- logical Laboratory, Woods Hole, Massachusetts. Single numbers, $20.00. Subscription per volume (three issues), $50.00 ($100.00 per year for six issues). Communications relative to manuscripts should be sent to Dr. Charles B. Metz, Editor, or Pamela Clapp, Assistant Editor, at the Marine Biological Laboratory, Woods Hole, Massachusetts 02543. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Copyright © 1987, 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). Very short papers (less than 10 manuscript pages including tables, figures, and bibliography) will be published in a separate section entitled "Short Reports." A limited number of solicited review papers may be accepted after formal review. A paper will usually appear within four months after its acceptance. The Editorial Board requests that manuscripts conform to the requirements set below; those manu- scripts which 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 double spacing (including figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper, 8'/2 by 1 1 inches. Manuscripts should be proofread carefully and errors 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 Editors Style Manual, 4th Edition (Council of Biology Editors, 1 978) 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 addi- tion, authors should supply a list of words and phrases under which the article should be indexed. 2. Figures. Figures should be no larger than 8'/2 by 1 1 inches. The dimensions of the printed page, 5 by 73/4 inches, should be kept in mind in preparing figures for publication. We recommend that figures be about l'/2 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 generally 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 photographic 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. 3. 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 information 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, acknowledgments 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. "•* ' *. . 4. A condensed title or running head of no more than 35 letters and spaces should appear at the top of the title page. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more ~tIian twoTauthors cited as Jones el al., 1980. Personal communications and material 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 LITERATURE 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 include complete titles and inclusive pagination. Journal abbreviations should normally follow those of the U. S. A. Standards Institute (USASI), as adopted by BIOLOGICAL ABSTRACTS and CHEMICAL AB- STRACTS, 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). For- eign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PERIODICALS, may find a booklet published by the Biological Council of the U.K. (obtainable from the Institute of Biology, 4 1 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 additional 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 J. cell. comp. Physiol. ) C. All abbreviated components must be followed by a period, 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 abbrevi- ations invented by the author. For example, use Rit Visindajjelags hlendinga without abbreviation. F. All single word journal titles in full (e.g. Veliger, Ecology. 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 preferred forms rather than WORLD LIST or USASI usage (e.g. Nature, Science, Evolution NOT Nature, Lond., Science, N.Y.; Evolution, Lancaster, Pa.) 6. Reprints, charges. Authors will be charged the excess over $100 of the total of (a) $30 for each printed page beyond 15, (b) $30 for each table, (c) $15 for each formula more complex than a single line with simple subscripts or superscripts, and (d) $ 1 5 for each figure, with figures on a single plate all consid- ered one figure and parts of a single figure on separate sheets considered separate figures. Reprints may be ordered at time of publication and normally will be delivered about two to three months after the issue date. Authors (or delegates or 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). THE Marine Biological Laboratory LIBRARY MAY 1 1 1987 Woods Hole, Mass. BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board RUSSELL F. DOOLITTLE, University of California at San Diego WILLIAM R. ECKBERG, Howard University ROBERT D. GOLDMAN, Northwestern University C. K. GOVIND, Scarborough Campus, University of Toronto JUDITH P. GRASSLE, Marine Biological Laboratory MICHAEL J. GREENBERG, C. V. Whitney Marine Laboratory, University of Florida MAUREEN R. HANSON, Cornell University JOHN E. HOBBIE, Marine Biological Laboratory LIONEL JAFFE, Marine Biological Laboratory HOLGER W. JANNASCH, Woods Hole Oceanographic Institution WILLIAM R. JEFFERY, University of Texas at Austin GEORGE M. LANGFORD, University of North Carolina at Chapel Hill GEORGE D. PAPPAS, University of Illinois at Chicago SIDNEY K. PIERCE, University of Maryland HERBERT SCHUEL, State University of New York at Buffalo VIRGINIA L. SCOFIELD, University of California at Los Angeles School of Medicine LAWRENCE B. SLOBODK.IN, State University of New York at Stony Brook JOHN D. STRANDBERG, Johns Hopkins University DONALD P. WOLF, Oregon Regional Primate Center SEYMOUR ZIGMAN, University of Rochester Editor. CHARLES B. METZ, University of Miami APRIL, 1987 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. in The BIOLOGICAL BULLETIN is issued six times a year at the Lancaster Press, Inc., Prince and Lemon Streets, Lancaster, Penn- sylvania. Subscriptions and similar matter should be addressed to The Biological Bulletin, Marine Biological Laboratory, Woods Hole, Massachusetts. Single numbers, $20.00. Subscription per volume (three issues), $50.00 ($100.00 per year for six issues). Communications relative to manuscripts should be sent to Dr. Charles B. Metz, Editor, or Pamela Clapp, Assistant Editor, Marine Biological Laboratory, Woods Hole, Massachusetts 02543. THE BIOLOGICAL BULLETIN (ISSN 0006-3 185) POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Second-class postage paid at Woods Hole, MA, and additional mailing offices. LANCASTER PRESS, INC., LANCASTER. PA IV Reference: Biol. Bull. 172: 155-160. (April, 1987) THE EFFECTS OF EMBRYOS OF DIFFERENT DEVELOPMENTAL STAGES ON REPRODUCTIVE BEHAVIOR AND PHYSIOLOGY IN BROODING FEMALES OF THE AMPHIPOD CRUSTACEAN GAMMARUS PALUSTRIS BETTY BOROWSKY Osborn Laboratories of Marine Sciences, Boardwalk at West 8:h Street. Brooklyn, New York 11224 ABSTRACT The expression of male reproductive behavior is correlated with female intermolt stage in the amphipod crustacean Gammarus palustris. Previous studies suggested that three factors that vary with female intermolt stage influence males: (1) water- borne pheromones, (2) contact pheromones on females' exoskeleta; and (3) female behavior. The present study was conducted to determine whether a fourth factor, developmental stage of brooded embryos, might also affect male behavior. In addi- tion, the effects of the embryos on the females' own behaviors and physiologies were examined. The contents of the females' brood pouches were altered by removing the broods of females of specific intermolt stages, and presenting the females with embryos of other females. G. palustris females readily place conspecifics' embryos in their pouches. Thus males' responses to females at the same intermolt stage but with broods of different developmental stages could be observed. The data showed that the nature of the females' brood pouches had no measurable effect on male behavior. In addition, the embryos' ages did not affect female intermolt period, reproductive behavior, or reproductive physiology. INTRODUCTION In some species with brood care, the presence of a brood can influence adult repro- ductive behavior and physiology. This has been best documented in the vertebrates and the social insects. For example, the presence of embryos in the nest of the stickle- back, Gasterosteus aculeatus, inhibits male courtship behavior (Sevenster-Bol, 1 962), and the removal of embryos from a nest of the great tit, Parus major, can induce courtship behavior and the production of another brood (Lack, 1966). Further, the presence or absence of eggs clearly influences egg production in at least several social insects (Wilson, 1971). The crustaceans typically exhibit brood care, but the relationship between the embryos and reproductive behavior and physiology has not been examined in most marine species. An exception to this is the observation that the removal of the em- bryos from the brood pouches of female Gammarus lawrencianus (an amphipod) stimulates some aspects of the males' reproductive behavior (Dunham, 1986). This is an interesting discovery because the typical Gammarus life-history pattern (described below) suggests both the feasibility and the advantage of this potential method of regulating reproductive effort. Gammarus females typically produce several broods in succession during the Received 20 October 1986; accepted 16 January 1987. 155 156 B. BOROWSKY warmer months in the temperate zone. Generally, amplexus (precopulation) begins towards the end of the female's intermolt period and continues until the female molts, at which time copulation occurs. Then, within an hour, ovulation occurs, and the male and female separate. All the embryos of the current brood are deposited in the female's brood pouch simultaneously at ovulation. Development occurs in the pouch, and the juveniles hatch and emerge shortly before the female is due to molt again [the times of hatching and emergence are species-specific (Borowsky, 1980b)]. The female remains alone until a few days before her next molt, when amplexus is reinitiated and the cycle is repeated. Thus, all the embryos of a specific brood are at the same stage of development, the stage of development of the embryos is correlated with the female's intermolt period, and males generally initiate amplexus when the brood is in the later stages of development. Further, males should have little difficulty in determining the nature of the brood pouch contents because the pouch is open to the environment (Borowsky, 1 980a), and the embryos could stimulate male behavior either by direct contact, or by water-borne substances. Earlier studies had shown that Gammarus paluslris females will readily place the broods of other conspecifics in their own pouches (Borowsky, 1983). This suggested that the effects of brood pouch contents on male behavior could be examined in detail in this species by exchanging embryos between females of different intermolt stages. It was reasoned that the presence of broods could influence reproductive behavior of both sexes in several ways. First, immature embryos might inhibit these behaviors; second, mature embryos might stimulate them; and third, both of these possibilities might be true. In addition, if the embryos had some effect, it could either be direct, in which case the mere presence of the embryos would influence behavior, or it could be indirect, first modifying the female's physiology over the course of days, and then altering both sexes' behaviors. In the first case, the presence or absence of embryos would alter the animals' behaviors within a short time, but in the second case, some time would be necessary for the physiology of the female to change. MATERIALS AND METHODS All animals employed in the present study were obtained by hand-picking individ- uals from under rocks and debris in the intertidal zone at low tide at Jamaica Bay, New York, on 2 and 4 April 1986. Animals were brought to the laboratory immedi- ately, and couples in amplexus were placed in individual 1 0 cm diameter glass culture dishes in water from the collecting site (about 25 ppt). They were maintained at 20°C with a light cycle of 1 5:9 L:D, and supplied with Ulva lactuca thalli ad libitum. Females in amplexus and carrying embryos in the later stages of development in their brood pouches were classed as "receptive." Previous studies had shown that the intermolt period of G. palustris females is about eight days at 20°C, and that females who had molted two days before rarely entered into amplexus. Therefore, females who had molted and ovulated two days before and were not in amplexus were classi- fied as "non-receptive" in this study. Eight experiments were conducted as follows: non-receptive females were assigned at random to four treatment groups; Experiment I, females were placed on frozen seawater; Experiment II, females whose embryos were removed were presented with the embryos of other non-receptive females; Ex- periment III, females whose embryos were removed were not presented with any embryos; Experiment IV, females were presented with receptive females' embryos. Receptive females were randomly assigned to the other four groups: Experiment V, females were placed on frozen seawater; Experiment VI, females whose embryos were GAMMARUS BROODS' EFFECTS ON BEHAVIOR 157 removed were presented with the embryos of other receptive females; Experiment VII, females whose embryos were removed were not presented with any embryos; and Experiment VIII, females whose embryos were removed were presented with the embryos of non-receptive females (Tables I, II). Each experiment employed ten different pairs of animals (80 tests were conducted altogether). Embryos were removed as follows: females were placed on a bed of frozen seawa- ter to anesthetize them, then the embryos were aspirated from the pouch with a fine pipette inserted between the brood plates. After the embryos were removed, females were placed in a clean dish with fresh seawater and allowed to recover for one day. Depending on the experiment, the females were either presented with their own em- bryos, with other females' embryos, or with no embryos at all (as described above). Females presented with embryos placed them in their brood pouches during the one- day recovery period. The responses of males to females were tested after the recovery period (24 h after treatment) as follows: ten females from each experimental treatment were tested individually by placing the female and a male simultaneously in a clean culture dish with fresh seawater. Each male was obtained by gently separating it from another female immediately before a test. Three observations were made: first, the time from the introduction of the animals to the dish and the occurrence of "grabbing" behavior (the male grasps the female's exoskeleton firmly with its claws); second, the time from grabbing to either the initiation of amplexus or the separation of the couple ("contact time"); and third, the occurrence of amplexus. "Grabbing" was noted because this is the earliest behavior in the typical sequence of reproductive behaviors which indicates unequivocally that the male is aware of the presence of the female, and "contact time" was measured because it represents the time during which the male obtains the information from the female which deter- mines whether he will initiate amplexus (Borowsky and Borowsky, 1985, 1987). Amplexus is stereotyped and is easily distinguished from other behaviors (Bor- owsky, 1984). After the behavioral observations, the couples were maintained in individual dishes until the female either molted or died. This was done for two reasons; first, to verify the viability of removed and replaced embryos by making sure they developed and hatched; and second, to determine whether the presence of embryos whose devel- opmental stage was inappropriate for the female's intermolt period influenced behav- ior later in the female's cycle. Accordingly, after each couple's behavior was observed, the animals were observed daily, noting the day that the brood hatched, the day when amplexus was initiated, and the day of the female's molt. RESULTS There was a significant difference among the eight experiments in the number of times that amplexus was initiated (xi = 52.4, P < 0.001). The difference was due to female class ("receptive" versus "non-receptive"), rather than to the nature of the brood pouch contents. There was a significant difference between all non-receptive females grouped together and all receptive females grouped together regardless of brood pouch contents (Tables I and II, respectively; one of 40 non-receptive females entered into amplexus, while 32 of 40 receptive females entered into amplexus; X2\ = 49.6, P < 0.00 1 ). In contrast, there was no significant difference between females with broods and those without (33 of 60 females with embryos engaged in amplexus, and 10 of 20 females without embryos engaged in amplexus; xf = 0.15, P> 0.05). It is especially noteworthy that there was no significant difference among the four exper- 158 B. BOROWSKY TABLE I Male reproductive behaviors expressed to non-receptive females whose brood pouches had different contents Experiments Decisions Time to decision (s) Number of amplexes x±SD Range x + SD Range (of 20 tests) Time to grab (s) I. Ice-treated: embryos not removed II. Embryos exchanged between non receptive females III. Embryos removed permanently IV. Embryos removed and replaced with receptive females' embryos 122.7 ±127 4-446 11.5 ±12 1-39 0 155.7 ±142 31-516 14.7 ±19 0-59 0 176.0+175 2-442 33.3 + 56 0-164 0 86.3 ± 92 0-285 33.1 ±41 2-147 1 iments performed on each class of female (non-receptive and receptive females, Ta- bles I and II, respectively: xis = 3.08 and 3.50, respectively; Ps > 0.05). Thus, the nature of the contents of the females' brood pouches did not affect the frequency of amplexus. In addition, the lengths of time from the introduction of pairs of animals into observation dishes to the times when males grabbed females was not significantly different among the eight experiments (Tables I and II: one-way ANOVA; F768 = 1.55, P > 0.05). Further, there was no significant difference among the groups in the lengths of time from grabbing to either amplexus or separation (one-way AN- OVA; F772 = 1.76, P > 0.05). Thus, while the decision to engage in amplexus de- pended largely upon whether the female was receptive, the length of time employed to arrive at the decision was not. Another effect of the brood on reproductive behavior could be to modify the fe- male's physiology. The frequency of amplexes increases as female intermolt period advances, but there is variability among individual females' initial day of amplexus (Borowsky and Borowsky, 1985, 1987). It was hypothesized that the absence of a brood might accelerate the time to amplexus, thus enhancing the successful fertiliza- tion of the next brood. However, there was no significant difference in the time from the female's molt to the onset of amplexus between non-receptive females whose broods had been removed permanently and non-receptive females who retained their TABLE II Male reproductive behavior expressed to receptive females whose brood pouches had different contents Time to grab (s) Decisions Time to decision (s) Number of amplexes Experiments X + SD Range X ± SD Range (of 20 tests) V. Ice-treated: embryos not removed 26 ,5± 22 4-79 59 .2 ±44 16-180 7 VI. Embryos exchanged between receptive females 148.1 + 164 9-553 41 .0 + 38 14-147 8 VII. Embryos removed permanently 55. ,7± 44 3-126 41 .1 ± 16 15-76 10 VIII. Embryos removed and replaced with non-receptive females' embryos 103 ,9± 112 4-329 31 .5 ±23 2-74 7 GAMMARUS BROODS' EFFECTS ON BEHAVIOR 159 broods (Experiment III vs. Experiments I and II combined, respectively; ns = 10 and 20, ranges 8-18 and 9-22, and x == 12.8 ± 4 and 13.6 ± 4, respectively: Student's t- test; t28 = -0.4625). Another way the absence of a brood might influence female physiology could be to accelerate the time until the female's next molt. However, the intermolt periods of females without broods were about the same as the intermolt periods of females with broods (Experiment III vs. Experiments I and II combined, respectively: ns = 10 and 17, ranges 12-23 and 12-25 days, and xs = 16.9 ± 4 and 16.9 ± 4, respectively). DISCUSSION The data show that on the basis of the criteria employed in the present study, the nature of the contents of females' brood pouches has no effect on reproductive behav- ior or reproductive physiology in G. palustris. In receptive females, the absence of broods did not increase, and the presence of immature broods did not decrease the frequency of amplexus. In the reciprocal experiments on non-receptive females, the presence of a mature brood did not increase the frequency of amplexus. Finally, the absence of a brood did not significantly hasten the onset of amplexus in non-receptive females without broods nor did it accelerate the time to these females' next molt. These data are consistent with earlier observations of G. palustris reproductive behaviors which suggest that the females' behaviors and contact pheromone(s) on their exoskeleta stimulate amplexus and copulation (Borowsky and Borowsky, 1985, 1987). But the results differ from those of Dunham (1986) who found that G. lawrencia- nus females with empty brood pouches engaged in amplexus more often than females carrying broods, even when they were at a comparable intermolt stage. There are several explanations for the differences. First, two different species were studied. Sig- nificant differences in behavior have been found between other Gammarus species (Borowsky, 1980b). But second, it is possible that the slightly different criteria em- ployed for scoring amplexus accounts for the difference in results. Occasionally a male G. palustris carried a female in the amplexus position immediately after an encounter but dropped it shortly thereafter. Since the function of amplexus is to maintain pairing until fertilization is possible, this temporary amplexus was not con- sidered to be biologically significant and was not scored here. Dunham's observation consisted of noting whether the pair was in amplexus three minutes after the animals were introduced to each other. It would be interesting to test the effects of broods on the reproductive behavior and physiology of the larger Crustacea. The larger species generally reproduce infre- quently [for example the lobster, Homarus americanus, reproduces biennially, (Ai- ken and Waddy, 1980)], and would stand to benefit from a mechanism which com- pensated for the loss of a current brood. ACKNOWLEDGMENTS I thank P. Dunham for stimulating discussions on this subject, R. Borowsky for a critical review of the manuscript, and G. Ruggieri for general advice and support. LITERATURE CITED AIKEN, D. E., AND S. L. WADDY. 1 980. Reproductive biology. Chapter 4. Pp. 2 1 5-276 in The Biology and Management of Lobsters, Vol. I, J. Stanley Cobb and Bruce F. Phillips, eds. Academic Press, New York. 160 B. BOROWSKY BOROWSKY, B. 1980a. Factors that affect juvenile emergence in Gammarus palustris (Bousfield, 1969). / Exp. Mar. Biol. Ecol. 42: 213-223. BOROWSKY, B. 1 980b. Reproductive patterns of three intertidal salt marsh gammaridean amphipods. Mar. Biol. 55: 327-334. BOROWSKY, B. 1983. Placement of eggs in their brood pouches by females of the amphipod Crustacea Gammarus palustris and Gammarus mucronatus. Mar. Behav. Physiol. 9: 319-325. BOROWSKY, B. 1984. The use of the males' gnathopods during precopulation in some gammaridian amphi- pods. Crustaceana 47: 245-250. BOROWSKY, B., AND R. BOROWSKY. 1985. Evidence that female contact pheromones influence the males' reproductive behaviors in the amphipod crustacean Gammarus palustris. Am. Zool. 25: 123A. BOROWSKY, B., AND R. BOROWSKY. 1987. The reproductive behaviors of the amphipod crustacean Gam- marus palustris (Bousfield) and some insights into the nature of their stimuli. J. Exp. Mar. Biol. Ecol. (In press) DUNHAM, P. 1 986. Mate guarding in amphipods: a role for brood pouch stimuli. Biol. Bull. 170: 526-53 1 . LACK, D. 1966. Population Studies of Birds. Oxford University Press, London. Page 24. SEVENSTER-BOL, A. C. A. 1962. On the causation of drive reduction after a consummatory act. Arch. Neerl Zool. 15: 175-236. WILSON, E. O. 1971. The Insect Societies. Harvard University Press, Cambridge, MA. Reference: Biol. Bull. 172: 161-177. (April, 1987) SEXUAL REPRODUCTION AND COLONY GROWTH IN THE SCLERACTINIAN CORAL PORITES ASTREOIDES ELIZABETH A. CHORNESKY1 2 AND ESTHER C. PETERS2 ^Department of Biology and Stone Harbor Marine Laboratory. Williams Hall, Lehigh University, Bethlehem, Pennsylvania 18015*. and2 Department of Invertebrate Zoology, National Museum of Natural History, Smithsonian Institution, Washington. DC 20560 ABSTRACT This study examines patterns of sexual reproduction and colony growth for Pori- tes astreoides Lamarck, an abundant Caribbean reef coral. Five factors influence the reproductive condition of this coral in Jamaica: (i) Season, (ii) lunar day, (iii) polyp location within a colony, (iv) colony size, and (v) colony age. P. astreoides has an unusual mixed breeding system: Approximately half of colonies are hermaphroditic and half are female. Although gonads occur in some colonies throughout the year, there are clear seasonal differences both in the number of reproductive colonies within the population and in colony fecundity. Male gametes are spawned monthly around the time of the full moon. The abundance and maturity of brooded larvae peaks prior to the new moon. Within reproductive colonies, gonads and brooded larvae are more abundant in central polyps than at colony edges. Among female colonies, the onset of reproduction is apparently related to colony size, whereas the fecundity of individual polyps is related to colony age. Hermaphroditic and female colonies differ in the size at which most colonies are reproductive. Rates of vertical and lateral growth for P. astreoides increase with colony size, but not with colony age. These findings demonstrate how the combined effects of several variables can cause individuals within a population to differ greatly in reproductive condition, fe- cundity, and growth rate. INTRODUCTION Since the turn of the century, there have been numerous descriptions of the pat- terns of reproduction and growth among reef corals (e.g., Duerden, 1902; Wood- Jones, 1910; Vaughan, 1915; Marshall and Stephenson, 1933; Atoda, 1947; Harri- gan, 1972;Connell, 1973; Loya, 1976;Gladfelter etal, 1978; Stimson, 1978;Rinke- vichandLoya, 1979a,b; Hughes and Jackson, 1980, 1985;Goreau# 0/., 1981;Kojis and Quinn, 1981, 1982; Harriott, 1983; Highsmith, 1982; van Moorsel, 1983; Bab- cock, 1984; Harrison et al, 1984; Jokiel et al, 1985; Shlesinger and Loya, 1985; Stoddart and Black, 1985; Szmant-Froelich et al, 1985; Wallace, 1985; Kinzie and Sarmiento, 1986; Szmant, 1986). It is becoming clear from some of these studies that co-occurring colonies of a given species can differ greatly in important characteristics such as reproductive state, fecundity, and growth rate (Marshall and Stephenson, 1933; Connell, 1973; Stimson, 1978; Rinkevich and Loya, 1979a, b; Harriott, 1983; Received 14 April 1986; accepted 30 January 1987. * Address for correspondence. 161 162 E. A. CHORNESKY AND E. C. PETERS Kojis and Quinn, 1981, 1984, 1985; Hughes and Jackson, 1985;Stoddart and Black, 1985;Szmant-Froelich, 1985; Wallace, 1985). Such variation can be caused by external events such as environmental stress. For example, coral fecundity may decline due to injury (Kojis and Quinn, 1981), expo- sure to sedimentation or turbidity (Kojis and Quinn, 1984), or competitive encoun- ters (Rinkevich and Loya, 1985). Similarly, rates and patterns of coral growth may be affected by colony breakage (Loya, 1976). Other intraspecific differences in reproduction and growth are more inherent to populations, resulting from relatively predictable factors such as ontogenetic stage, season, or lunar period. For example, in protandrous hermaphroditic species, small and/or young colonies are male while larger and/or older colonies are hermaphroditic (Rinkevich and Loya, 1979a; Kojis and Quinn, 1981). Among species in which the onset of reproduction is related to colony size, including certain gorgonian octocorals (Wahle, 1983) and scleractinian corals (Rinkevich and Loya, 1979b; Kojis and Quinn, 1981, 1985; Babcock, 1984; Szmant-Froelich, 1985), mature colonies will stop producing gametes when reduced in size below the reproductive threshold value, irrespective of their age (Wahle, 1983; Kojis and Quinn, 1985; Szmant-Froelich, 1 985). Polyp fecundity has been shown to vary with colony size (Rinkevich and Loya, 1979b; Kojis and Quinn, 1981; Babcock, 1984), with colony age (Kojis and Quinn, 1985), with polyp size (Harriot, 1983), and with polyp location within a colony (Har- rigan, 1972; Rinkevich and Loya, 1979b; Wallace, 1985). Finally, both the abun- dance and developmental stage of coral gonads or brooded larvae can change tempo- rally with season or with lunar period (e.g., Harriot, 1983; Kojis and Quinn, 1984; Stoddart and Black, 1985). When individuals of a species vary greatly, estimates of reproductive condition, fecundity, or growth rate based on small sample sizes may not accurately reflect pat- terns of reproduction and growth within the population as a whole. Furthermore, interspecific comparisons may be problematic when intraspecific differences equal or exceed those occurring between species. Nevertheless, remarkably few studies have systematically evaluated the levels and possible sources of intraspecific variation in reproduction or growth (e.g., Connell, 1973; Rinkevich and Loya, 1979b; Kojis and Quinn, 1981, 1984, 1985; Harriot, 1983; Babcock, 1984; Hughes and Jackson, 1985; Szmant-Froelich, 1985; Kinzie and Sarmiento, 1986). This paper documents intraspecific patterns of sexual reproduction and growth within a population of the reef coral Porites astreoides Lamarck. Specifically, we describe: (i) Seasonal and lunar patterns in sexual reproduction; (ii) intra-colony vari- ation in polyp fecundity and gender; (iii) the effects of colony age on polyp fecundity; (iv) the relationship between colony size and reproductive condition; and (v) changes in colony growth rate that occur with increasing colony size. MATERIALS AND METHODS Study organism Porites astreoides is among the most common corals on many Caribbean reefs and occurs over a wide range of depths and habitats (Goreau, 1959; Goreau and Wells, 1967). It is generally considered to be among the most fecund species, since juveniles appear early and abundantly in studies of coral recruitment (Bak and Engel, 1979; Rylaarsdam, 1983; Rogers et al, 1984). P. astreoides colonies grow in encrust- ing, platey, and mound-like morphologies. Corals sampled in this study had encrust- CORAL REPRODUCTION AND GROWTH 163 ing to plate-like shapes, as these predominate in the study population at -10 m on the west forereef of Discovery Bay, Jamaica. At this site, instances where injury has subdivided coral colonies into physiologically separate, but genetically identical "daughter" colonies (Hughes and Jackson, 1980, 1985) were usually clearly identifi- able by the color similarity, close proximity, or remaining skeletal connections be- tween adjacent colonies. Previous surveys suggest that most colonies (78%) in this population result from the independent recruitment of larvae rather than from the subdivision of larger colonies (Chornesky, unpub. data). Sexual reproduction Sexual reproduction was evaluated between October 1 98 1 and July 1 982 by histo- logically examining coral tissues for gonads and larvae. Our sampling scheme was designed to reveal both seasonal and lunar patterns in reproductive periodicity. Tis- sues were sampled throughout one lunar period (28 days) during each of four months that reflect the annual range of seawater temperature: July (high); January (low); April (intermediate, increasing); and October (intermediate, decreasing). During each of these four months, samples were collected every five or six days (a total of five sampling days per month). On each sampling day, separate tissue specimens of thirty or more polyps were collected from both the centers and growing edges of five P. astreoides colonies. Colonies sampled on any given day were spatially separated by distances of several meters, and thus can be safely assumed to be of different geno- types. Individual colonies were sampled only once during the study, to avoid any adverse effects of repeated injury on coral reproduction (e.g., Kojis and Quinn, 1 98 1 ). In total, 100 colonies were sampled (4 months X 5 days X 5 colonies). The general condition, size, and maximum skeletal thickness of each colony was recorded. Colony size was determined by calculating the surface area of live coral tissues (the area of an ellipse defined by the two largest, perpendicular diameters laid across the colony's living tissues). Maximum skeletal thickness was measured when possible (59% of colonies sampled). For encrusting and plate-shaped corals, maximum thickness provides a better measure of colony age than surface area. Surface area is often only poorly correlated with colony age since injury along colony perimeters frequently interrupts or inhibits the lateral expansion of corals (Jackson, 1979; Hughes and Jackson, 1980, 1985). In contrast, colony centers are less likely to be injured and upward skeletal growth in this region may proceed relatively uninterrupted. The growth discrepency between colony regions is often reflected in colonies of equal maximum thickness but widely varying surface area (Chornesky, unpub. data). Thus, maximum skeletal thickness is the most reliable historical marker of previous growth, and is used here to estimate minimum colony age. Results of tests for size differences of sampled corals among sampling days and among sampling months by one-way ANOVA were non-significant (among days, F - .98, P > .25, df = 1 9,80; among months, F = . 1 8, P > .75, df = 3,96). Similarly, colony thickness did not differ significantly among sampling months (F = .77, P > .5, df = 3,58). The numbers of colonies for which thickness could be measured did not allow for statistical comparison among sampling days. Colony size and thickness were not significantly correlated for the population of corals sampled (r = .22, P > .05). To minimize the potential effects of uncontrolled interactions and events, we de- liberately sampled colonies with few contacts along their borders with adjacent ani- 164 E. A. CHORNESKY AND E. C. PETERS mals, and with no evidence of disease, lesions, bare spaces, or other injury. Pieces of corals were broken off in situ using a chisel. Specimens were kept immersed in seawa- ter, and then fixed in Helly's fixative within two hours of collection. Fixed tissues were decalcified using a solution of dilute HC1 and EDTA and stored in 70% ethanol (Peters, 1984). Specimens were subsequently embedded in Paraplast (Sherwood Medical) and sectioned at 6 nm. The resulting slides were stained either with hema- toxylin and eosin or with Heidenhain's Aniline Blue Method (Luna, 1968). Three slides were made at different depths within each specimen, one each from the top, middle, and bottom thirds of the tissues. Longitudinal sections of polyps were also prepared when possible. On each cross-section of tissues, the abundance and condition of eggs, spermaries, and developing larvae were quantified by surveying 10 separate, non-overlapping fields at 125X magnification. Fields contained three to four polyps (x = 3.7, S.D. = .71, no. fields counted = 79). Developmental stages of gonads and larvae were estimated by size and morphology (Fig. 1A-F, I). Statistical comparisons between central and peripheral tissues included data from slides cut from mid-depth in the tissues. All other statistical analyses (Sokal and Rohlf, 1969; Steele and Torrie, 1960) were performed on data from the slide of each specimen that revealed the maximum number of gonads. Growth rate Growth rates for P. astreoides colonies were determined by staining colonies in situ with the vital stain Alizarin red S (e.g., Dodge et al, 1984). Alizarin red S is incorporated into a thin layer of skeleton during staining and provides a marker for measuring subsequent skeletal growth. In May of 1981, 10 colonies between 6 and 300 cm2 were enclosed for 24 h in clear plastic bags containing a small amount of stain. Stained corals appeared to be healthy and lacked any evidence of injury or contacts with other sessile animals along their borders. After one year of growth, the colonies were collected and cut in half. Vertical growth was measured at five random locations along the cut face, and lateral growth was measured where five randomly chosen radii intersected the colony perimeter. RESULTS Each factor examined — season, lunar day, polyp location, colony size, and colony age — influenced some aspect of reproduction or growth in Porites astreoides. The resulting patterns are complex; therefore, we have summarized the major findings of the results in Table I. Breeding system Gender and reproductive state of colonies. Colonies of P. astreoides were either female or hermaphroditic. Of the 100 colonies sampled, 28% contained only female gonads, 26% contained both male and female gonads, and 43% contained no gonads. The only colony having exclusively male gonads contained only a single spermary. Thus, the existence of unisexual males in this population is doubtful. Brooded larvae were observed in 50% of female colonies and in 46% of hermaphroditic colonies. Distribution of gender among and within polyps. Within hermaphroditic colonies, individual polyps could be either male, female, or hermaphroditic. The proportions of these three polyp types varied among colonies. Within hermaphroditic polyps, CORAL REPRODUCTION AND GROWTH 165 TABLE I Summarv of results A. BREEDING SYSTEM 1 . Gender: 52% female, 48% hermaphroditic 2. Fertilization: Internal; larvae brooded B. TIMING OF REPRODUCTION 1 . Seasonal: Peaks during April 2. Lunar: Spermaries spawned around time of full moon; larvae mature prior to new moon C. FACTORS AFFECTING POLYP FECUNDITY 1 . Location within a colony: Centers more fecund than colony edges 2. Colony size: Associated with onset of female reproduction 3. Colony age: Correlated with polyp fecundity in females D. FACTORS AFFECTING COLONY GROWTH 1 . Colony size: Correlated with vertical and lateral growth rates male and female gonads were occasionally found on the same mesentery (Fig. ID, E). In both female and hermaphroditic colonies, gonads did not appear restricted to any particular mesenteries, although they were most frequent on incomplete mesen- teries (Fig. 1 H). Up to nine mesenteries might contain gonads within any given polyp. Cross and longitudinal sections of mesenteries yielded similar counts of one or, rarely, two to three eggs per mesentery. In contrast, cross-sectional counts of sperma- ries could underestimate actual spermary abundance. Mesenteries in cross-section contained one to three spermaries. However, up to 14 spermaries, arranged vertically and occasionally overlapping, might be visible in longitudinal sections of mesenteries (Fig. 1 E). There was no pattern to the vertical distribution of male and female gonads within hermaphroditic mesenteries (Fig. IE). Location of gonads within colonies Centers versus edges. Gonads were not distributed evenly throughout colonies. Comparison of central and peripheral tissues showed that the densities per polyp of both eggs and spermaries were greatest at colony centers (one-tailed, paired Mests: For eggs, P = .005 ; for spermaries, P = .0 1 6). Brooded larvae were too rare to perform a statistical test of the relationship between abundance and position within colonies, but appeared to follow the same trend. Within centers. The distribution of gonads among central polyps was patchy. In reproductive colonies, on average only half (51%) of the 10 fields surveyed in each central slide contained gonads (no. slides surveyed = 153). Within edges. Due to the low density of gonads at colony edges, gender might easily be misidentified from samples of only peripheral polyps. Sixty-one percent of hermaphroditic colonies lacked male, female, or both types of gonads in peripheral polyps. Similarly, 36% of female colonies lacked eggs in peripheral regions. Only one of the 54 reproductive colonies contained gonads in peripheral tissues of a type (male) that was absent in the central tissues. Development of gametes and larvae Female and male gamete development were divided into three approximate stages (early, middle, and late; Fig. 1 ). Female. The earliest recognized female gametes had a basophilic cytoplasm and 166 E. A. CHORNESKY AND E. C. PETERS o V ***- 4 * * V v. *?$* ^ - -n<«. . o" .; B i C . . s •% I ••-N •'- '*• •• * . --^x- • f .--/iv>- -;• : *- •- • : ..' H »*i • / -•;>*.i;:.;0;j: I V " •• -' ""•'•. .'JS'T '* * .' \ " \ '? • . '.-'.. •- • / . ,-""•' '.^'•**''" "• V; ' ' ;' •'• T --'• •' *'•' ,. • . ?N .*..' S ,'. •:• , H . * • ' I FIGURE 1 . Gonad and larval development in Porites astreoides. A. Early female gametes (o) develop- ing in two mesenteries. B. Later in development, the cytoplasm of oocytes changes in staining characteris- tics, and lipid droplets accumulate. C. Mature egg (o). Diameters of released eggs were between 1 30 and 1 70 ^m in histological sections. D. Rare view of early oocyte (o) and spermary (s) developing on the same CORAL REPRODUCTION AND GROWTH 167 an enlarged nucleolus. Initially present in the gastrodermis, the oocytes were later enveloped in the mesoglea (Fig. 1 A). Mid-stage oocytes were larger and eosinophilic, with lipid droplets in the yolk (Fig. 1 B). They had a distinctive eccentric nucleus and prominent nucleolus, and were surrounded by a thin layer of mesoglea. Mature eggs had completed vitellogenesis (Fig. 1C). Occasionally, eggs assumed a "dumbell" shape, with the nucleus in one lobe (Fig. 1G). Serial sections suggested that this un- usual shape was a fixation artifact (but see Szmant-Froelich et al, 1985). Male. Male gametes developed from rounded interstitial cells with a large nucleus and little cytoplasm. These cells were found in the gastrodermis, although early sper- matogonia were surrounded by mesoglea (Fig. 1 D). Mid-stage spermaries were distin- guished by the enlargement of irregularly shaped nests of cells with occasional mitotic figures visible (Fig. 1 E). Later spermaries contained smaller spherical secondary sper- matocytes and spermatids. Mature follicles were quite large, and stretched the mesen- teries so that they were covered by only a layer of squammous gastrodermal cells and associated mesoglea (Fig. IF). Mature spermatozoa with eosinophilic tails appeared first in follicle centers and eventually filled them completely. Spermatozoa appeared to burst from mesenteries, and large masses were found in polyp coelenterons and gastrovascular canals. Remnants of spermatozoa occurred in some mesenteries fol- lowing spawning, occasionally accompanied by redeveloping spermaries. Larvae. Fertilization for P. astreoides is internal. Mature eggs were observed being extruded between gastrodermal cells in mesenteries, and masses of sperm were seen surrounding released eggs in the coelenterons of several colonies. While most em- bryos were found in polyp coelenterons and gastrovascular canals, one appeared to lie within a mesentery, as observed by N. I. Goreau in Caribbean Porites spp. (Fadlallah, 1983). Early developing embryos consisted of numerous round blastomeres with ex- ternal red granular yolky cells (HeidenhainVstained) and a vacuolated central area. Zooxanthellae apparently invaded the endoderm of embryos as tissue layers differen- tiated. In later embryos, the interior was filled with foamy endodermal cells, and chromophore (lipofuchsintype pigment) cells and nematocysts appeared in the ecto- derm. The most mature brooded planulae had six mesenteries, a well-formed stomo- daeum, and embryonic chromophore cells and nematocysts in both the ectoderm and endoderm (Fig. II). The timing of reproduction Reproductive events for P. astreoides occurred with both seasonal and lunar peri- odicities. The following analyses of reproductive patterns include only data from colony centers, since they provide the most reliable estimate of colony gender and fecundity. Seasonal periodicity. Although gonads and larvae were present in at least some mesentery. Primary spermatogonia were 4-6 nm in diameter. E. Longitudinal section of mesentery con- taining two oocytes and six spermaries. F. Mature spermaries (s), as tails on spermatozoa begin to appear. A new follicle (f) is also beginning to develop. Individual spermatozoa had 1.5-2 ^m triangular-shaped heads. G. "Dumbbell" shaped egg in mesentery. H. Cross-section through a polyp showing arrangement of eggs (o) and spermaries (s). I. Section through mature planula (pi) with six mesenteries and early filament tissues. Scale bars in A, B, C, E, F, and G, all equal 100 ^m. Scale bar in D = 50 /xm, and in H and I = 200 urn. Photomicrographs were taken using a Zeiss Photomicroscope II, and eggs and sperm were measured using a calibrated ocular micrometer under an oil immersion objective. 168 E. A. CHORNESKY AND E. C. PETERS 100 c a so w a> 'c o o o 100 •o A). Lunar periodicity. Superimposed upon the seasonal changes in reproduction was a pronounced lunar periodicity (Fig. 3). Most spermaries were mature prior to the full moon, and immature spermaries were most common between the full and new moons (Fig. 3A). In contrast, egg development showed no clear lunar pattern (Fig. 3B). Male and female gonadal abundance did not vary predictably with lunar day. Both the abundance (Fig. 3C) and maturity of brooded larvae peaked prior to the new moon. CORAL REPRODUCTION AND GROWTH 169 TABLE II Seasonal variation in the density ofgonads and developing larvae in tissues of Porites astreoides* Female colonies Hermaphroditic colonies Eggs Larvae Eggs Spermaries Larvae Oct.** .08 0.0 .03 .43 0.0 Jan. .40 0.0 .57 .20 0.0 (.22-1.7) (0.0-.03) (.08-1.9) (.05-.54) (0.0-.08) April .51 .07 .38 .20 .03 (.05-2.3) (0.0-.43) (.14-.78) (.05-1.5) (0.0-.62) July .18 0.0 .09 .07 .01 (.03-.59) (0.0-.03) (.03-.54) (.03-.73) (0.0-.14) * Values presented as median number observed per polyp and the (range). ** No ranges given since only one female and one hermaphroditic colony were observed during October. Differences among colonies Gender. Reproductive state was significantly associated with colony size for corals sampled during January, April, and July (x2 = 13.25, df = 4, P < .01) (Fig. 4). Al- though the proportion of hermaphroditic colonies remained constant for all size classes, the proportion of non-reproductive colonies declined and the proportion of female colonies increased with increasing size class (Fig. 4). These trends suggest that hermaphroditic colonies are reproductive at relatively small sizes, whereas small colo- nies lacking gonads are females that will become reproductive when they increase in size. The presence of both females and hermaphrodites among the smallest colonies sampled (7 cm diameter, 38 cm2), however, suggests that some unknown factor(s), in addition to size, may also influence the onset or continuation of female reproduction. Gender and reproductive state were not significantly associated with colony age (thickness). Fecundity. Mean polyp fecundity differed widely among colonies (Table II, range values). To examine the potential sources for this variation, we compared the fecun- dity of all colonies of known size and age (thickness) from the April collecting period. This subsample was chosen because: ( 1 ) April was the month when the maximum number of colonies was reproductive and reproductive colonies were the most fe- cund; (2) gonadal abundance did not vary over a lunar period, allowing comparison among colonies sampled on different lunar days; and (3) previous studies have sug- gested that colony size (Kojis and Quinn, 1981; Babcock, 1984) and/or colony age (Rinkevich and Loya, 1979b; Kojis and Quinn, 1985) may influence the fecundity of coral polyps. The potential influences of colony size and age (thickness) on fecundity were sepa- rated statistically by calculating partial correlation coefficients (Steele and Torrie, 1960). This is equivalent to calculating the correlation between fecundity and size while holding age constant (fs-a), and calculating the correlation between fecundity and age while holding size constant (fa • s). Female colonies showed no significant correlation between fecundity and colony size (rfs.a = -.35, n = 9, P > .05). In contrast, female fecundity was significantly correlated with colony age (thickness) (rfa.s = .91, n = 9, P < .01), in part due to the 170 E. A. CHORNESKY AND E. C. PETERS o € UJ Full Moon New Moon Full Moon (A •o (0 B 5 S * t: il J2 o) Q- .E w 9- co Full Moon New Moon Full Moon 100-1 Full Moon New Moon Lunar Day Full Moon FIGURE 3. Lunar periodicity of reproduction in Porites astreoides. A. Lunar changes in the fre- quency of early (solid line) and late (dashed line) male gonads. Mature spermaries were most abundant prior to the full moon, and immature spermaries were most abundant between the full and new moons. B. Frequencies of immature (solid line) and mature (dashed line) eggs versus lunar day. There was no clear lunar pattern in the development of eggs. In both A and B, data are reported as the mean percent per colony of each stage. C. Most of the larvae observed were in tissues collected between the full and new moons. Figure presents data for January (stars), April (squares), and July (dots) sampling periods. occurrence of a highly fecund small (212 cm2) but old (6.5 cm thick) colony within the sample population (Fig. 5). For hermaphroditic colonies, none of the partial correlations between either colony size or colony age (thickness) and any of three measures of fecundity — eggs per polyp, spermaries per polyp, or (eggs + spermaries) per polyp — was significant (for all six partial correlations, .45 > r > —.51, P > .05, n = 9). Similarly, the ratio of male:female gonads within hermaphroditic colonies (i) was not significantly correlated with either colony size fas- a = -.3 1 , P > .05) or colony age (thickness) (ria.s = .45, P > .05). Growth rate We examined two measures of growth: Lateral extension of the growing edge and vertical deposition of new skeletal material over the pre-existing skeleton (Fig. 6). The CORAL REPRODUCTION AND GROWTH 171 75 SO. 25. (A n jo O 0) N 75 CO |H 0 5O o o o 25 75. 50 HERMAPHRODITIC FEMALE NON-REPRODUCTIVE 0-125 126-250 o Size Class (cm ) > 251 FIGURE 4. The relationship between gender and colony size. Colonies smaller than 125 cm2 in sur- face area were mostly non-reproductive or hermaphroditic. With increasing size, the proportion of female colonies increased while that of non-reproductive colonies decreased. The frequency of hermaphroditic colonies remained approximately equal for all three size classes. median rate of lateral growth was .73 cm/year (range = .2-1.8). The median rate of vertical growth was .3 1 cm/year (range = . 1 3-. 46). Partial correlation coefficients were calculated to separate the potential influences of colony size and colony age (thickness) on growth rates. Rates of both vertical (v) Q. >. 2 I o o o> a u.\ •s ll "" •0 Q. §l2 O O 0) Q. «D 4) LL to 100 2OO 3OO 400 3024 M Surface Area (cm ) 246 Thickness (cm) FIGURE 5. Female fecundity versus colony size and colony age. The fecundity of polyps in female colonies was significantly correlated with colony thickness (an estimate of minimum colony age) but was not correlated with the surface area of live colony tissues (colony size). 172 E. A. CHORNESKY AND E. C. PETERS 4- S o k. o 15 . O 2- 0 > S o fe O 15 0) > 2- B 100 200 300 M Surface Area (cm ) 24 6 8 10 Thickness (cm) E E .10- o h. o 15 t_ 0 *- (0 E E .10- o u_ o 15 w 0 -*- re \ 100 200 300 Surface Area (cm ) 2 4 e 8 10 Thickness (cm) FIGURE 6. Colony growth versus colony size (surface area) and colony age (thickness). Vertical and lateral growth were both significantly correlated with colony surface area (A, C) and not with colony thick- ness (B, D). and lateral (I) growth were significantly correlated with colony size (rvs.a = .75, rls.a = .67, P < .05) (Fig. 6 A, C). In contrast, partial correlations between vertical or lateral growth rate and colony age (thickness) were clearly not significant (rva.s = -.3, rla.s = -.04,/J>.05)(Fig.6B,D). DISCUSSION Periodicity of sexual reproduction Sexual reproduction for the abundant Caribbean reef coral Porites astreoides oc- curs with both seasonal and lunar periodicities in Jamaica. Although some colonies contained gonads throughout the year, the proportion of reproductive colonies in the population and the fecundity of coral colonies varied seasonally. Colonies were maximally reproductive during the spring (April). Superimposed upon this annual pattern was a synchronized, lunar periodicity in the development of male gonads and larvae. Our data suggest male gametes are spawned prior to the full moon, and larvae are released around the time of the new moon. The lack of any clear lunar pattern in CORAL REPRODUCTION AND GROWTH 173 the development of female gonads suggests that eggs may take more than a month to mature and/or may vary widely in developmental rate. The subtle seasonal changes in coral reproduction described here (and elsewhere for other corals, e.g., Harriot, 1983; Kojis and Quinn, 1984; Stoddart and Black, 1985) may have important implications for evaluating the annual reproductive activ- ity of individual colonies. For example, variation among months in the frequency of reproductive colonies (Fig. 2) suggests that individual corals within a population may differ in the duration of their annual reproductive periods. In addition, seasonal differences in coral fecundity suggest that the fecundity of any given coral colony may vary throughout the year. Clearly, estimates of reproductive output based on observations at any single time of year or of only a few colonies would be misleading for this and, perhaps, other corals. Breeding system P. astreoides colonies are either female or hermaphroditic. This is the first descrip- tion of a mixed breeding system of this type for corals — most appear to have predomi- nantly unisexual (dioecious) or hermaphroditic (monoecious) colonies within a pop- ulation (for reviews see Fadlallah, 1983; Harrison, 1985). It differs from Szmant's (1986) description of P. astreoides from Puerto Rico as simultaneously hermaphro- ditic, and is also in striking contrast to Indo-Pacific species of Porites that appear to be largely dioecious (Kojis and Quinn, 1982; Harriott, 1983; Harrison, 1985;Szmant, 1986) [although Kojis and Quinn (1982) did report a very low incidence of hermaph- roditic colonies in the Pacific P. andrewsi]. The simultaneous occurrence of female and hermaphroditic colonies within a population might simply indicate that all colonies are potentially hermaphroditic, but male gonads have been missed in apparently female colonies due to a sampling error. However, the low probability of gender being misidentified from central tissues (1/54) suggests that P. astreoides having only female gonads within centrally located polyps contain only female gonads throughout the colony. Alternately, female colonies might occur within a hermaphroditic population if gender varies seasonally or over a coral's lifespan, and, consequently, corals produce male gametes for only part of the time that they are reproductive (e.g., Fadlallah, 1983). The temporally constant sex ratio in this population of P. astreoides (Fig. 2), however, suggests that seasonal shifts in gender are unlikely. Similarly, there does not appear to be a predictable change in gender associated with colony size and/or age (Rinkevich and Loya, 1979a, b; Kojis and Quinn, 1981). If colonies did switch from hermaphroditic to female with increasing size, the relative frequency of hermaphro- dites would decline in larger size classes and the ratio of male to female gonads within hermaphroditic colonies should be negatively correlated with colony size. Neither are true for P. astreoides. Gender also does not appear to be a function of aging for P. astreoides, as both gender and the ratio of male to female gonads within hermaphro- ditic colonies were not associated with our estimate of colony age (thickness). The combined results suggest, instead, that colonies of P. astreoides are either female or hermaphroditic and that hermaphroditic colonies are generally reproduc- tive at smaller sizes than females (Fig. 4). The apparent lack of true males within this population of P. astreoides suggests that there may be a reproductive disadvantage associated with making only male gametes (e.g., Charnov, 1982). That is, individuals investing reproductive resources exclusively into male gonads would be less successful at producing offspring than 174 E. A. CHORNESKY AND E. C. PETERS hermaphrodites. A mixed breeding system like that of P. astreoides might be selected for, for example, if the likelihood that sperm will fertilize eggs of another colony is unpredictable or temporally variable. The unusual breeding system of P. astreoides also has important genetic implica- tions. Some hermaphroditic corals are capable of self-fertilization (Kojis and Quinn, 1981; Heyward and Babcock, 1986). While cross-fertilization must occur in female colonies of P. astreoides, the proximity in hermaphroditic colonies of eggs and sper- maries on some mesenteries suggests the potential for self-fertilization in this gender. If females and hermaphrodites of P. astreoides do differ in mode of fertilization, their respective offspring will also differ in genetic composition and diversity. Ontogenetic changes and energetic constraints Several life history features of P. astreoides appear to be under independent con- trol; some are associated with colony size and others are associated with colony age (thickness). Moreover, it seems that the influence of at least colony size may differ between genders, as reflected by the difference between female and hermaphroditic colonies in the size at which most colonies become reproductive (Fig. 4). When life history features of colonial animals, such as sexual maturity, are depen- dent upon colony size, they may vary as colonies grow or are reduced in size by external events (e.g., Wahle, 1983, 1985; Hughes and Jackson, 1985; Kojis and Quinn, 1985;Szmant-Froelich, 1985;Karlson, 1986). For P. astreoides, reproductive state, vertical growth rate, and lateral growth rate are associated with colony size. Thus, these attributes potentially may change over a coral's lifespan as colony size increases or decreases. Sexual reproduction and growth are often described as competing biological func- tions that draw on a limited supply of energetic or material resources within the indi- vidual (e.g., Stearns, 1977). It has been predicted that rates of growth in colonial animals should decrease following the onset of sexual reproduction (Williams, 1975; Jackson, 1979; Kojis and Quinn, 1981). However, for P. astreoides, growth rates increase over the range of colony sizes associated with the onset of female reproduc- tion (compare Figs. 4 and 6), suggesting that the presumed costs of gametogenesis do not result in a measurable depression of growth. For female colonies of P. astreoides, polyp fecundity is associated with colony age (thickness) and not with colony size (Fig. 5). This suggests that polyp fecundity is a function of colony aging (e.g., Kojis and Quinn, 1985). Thus total colony fecundity will be related both to the number of reproductive polyps within a colony and to colony age. The former often may be determined largely by colony size. The lack of any correlation between fecundity and either size or age (thickness) among hermaphroditic colonies may reflect the difficulty of quantifying hermaphro- dite fecundity. For example, there are different sampling biases inherent in estimates of egg versus spermary abundance from polyp cross-sections — because eggs and sper- maries differ both in their distributions within mesenteries and in their potential resi- dence times within coral polyps. Moreover, spermaries and eggs probably constitute quantitatively different investments of colony resources, contribute differently to the successful production of offspring, and thus cannot be compared directly. Such con- siderations will be important for accurately ranking or interpreting the fecundity of hermaphroditic colonies having different relative proportions of female and male gonads. Patterns of gonad and larval abundance within individual colonies of P. astreoides CORAL REPRODUCTION AND GROWTH 175 may reflect energetic and/or ontogenetic constraints. Low densities at colony edges might occur because the proximate allocation of colony resources to growth along the edge reduces the local resources available for gamete production (e.g., Williams, 1975). Alternatively, lower fecundity along the edges of colonies might be an evolved "strategy" to ensure that valuable gonads are not located in colony tissues most likely to experience injury due to competition, predation, or growth into unsuitable habitats (e.g., Jackson, 1979). Finally, if polyp fecundity is influenced by polyp age (Fadlallah, 1982; Wallace, 1985), higher fecundity in central polyps might simply reflect their relative age and stability. Although these three hypotheses are not mutually exclusive, we favor the last for P. astreoides, since it is consistent with the positive correlation observed between the fecundity of female polyps and colony age (thickness). Conclusion Patterns of sexual reproduction and growth described here for P. astreoides are complicated by the interaction of the following factors: (1) Temporal patterns of re- production; (2) "pre-programmed" differences among colonies in fecundity appar- ently associated with aging; (3) size-related variation among colonies in reproductive condition and growth rate; and (4) differences between colony regions in fecundity. In large part, this complexity is related to the colonial mode of construction which allows reproductive activity to differ among polyps of a colony, and the life history features associated with size and age to vary independently. ACKNOWLEDGMENTS The comments of L. Hayek, H. B. McCarty, J. Render, C. M. Wahle, and an anonymous reviewer improved the manuscript. We thank C. M. Wahle for his valu- able insight and criticism and for many stimulating discussions of colonial life histo- ries, E. Sides for essential support and assistance in the field, P. P. Yevich for provid- ing facilities for the histological preparations, R. Ramsay for help with field collec- tions, and L. Hayek for consultation concerning statistical methods. This research was funded by grants to E.A.C. from the National Science Foundation (DEB 8 105 172) and Sigma Xi and by a research fellowship to E.A.C. from the Division of Biological Sciences of The University of Texas at Austin. Both authors were sup- ported by Smithsonian Postdoctoral Fellowships during the final stages of this work. Publication #377 of the Discovery Bay Marine Laboratory, University of the West Indies. LITERATURE CITED ATODA, K. 1947. The larva and postlarval development of some reef-building corals. I. Pocillopora dami- cornis c. espitosa (Dana). Sci. Reports Tohoku Univ. 4th Ser. 18: 24-47. BABCOCK, R. C. 1984. Reproduction and distribution of two species ofGoniastrea (Scleractinia) from the Great Barrier Reef Province. Coral Reefs 2: 187-195. BAK, R. P. M., AND M. S. ENGEL. 1979. Distribution, abundance and survival of juvenile hermatypic corals (Scleractinia) and the importance of life history strategies in the parent coral community. Afar. Biol. 54:341-352. CHARNOV, E. L. 1982. The Theory of Sex Allocation. Princeton University Press, Princeton, NJ. 355 pp. CONNELL, J. H. 1 973. Population ecology of reef-building corals. Pp. 205-245, in Biology and Geology of Coral Reefs, II. Biologv, O. A. Jones and R. Endean, eds. Academic Press, New York. DODGE, R. E., S. C. WYERS, H. R. FRITH, A. H. KNAP, S. R. SMITH, C. B. COOK, AND T. D. SLEETER. 1984. Coral calcification rates by the buoyant weight technique: Effect of alizarin staining. J. Exp. Mar. Biol. Ecol. 75: 217-232. 176 E. A. CHORNESKY AND E. C. PETERS DUERDEN, J. E. 1902. West Indian madreporarian polyps. Mem. Nat. Acad. Sci. 8: 401-597. FADLALLAH, Y. H. 1982. Reproductive ecology of the coral Astrangia lajolaensis in a kelp forest habitat. Oecologia 55: 379-388. FADLALLAH, Y. H. 1983. Sexual reproduction, development and larval biology in scleractinian corals. Coral Reefs 2: 129-150. GLADFELTER, E. H., R. K. MONAHAN, AND W. G. GLADFELTER. 1978. Growth rates of five reef-building corals in the northeastern Caribbean. Bull. Mar. Sci. 28: 728-734. GOREAU, N. I., T. J. GOREAU, AND R. HAYES. 1981. Settling, survivorship and spatial aggregation in planulae and juveniles of the coral Porites porites (Pallas). Bull. Mar. Sci. 31: 425-435. GOREAU, T. F. 1959. The ecology of Jamaican reefs. I. Species composition and zonation. Ecology 40: 67-90. GOREAU, T. F., AND J. W. WELLS. 1967. The shallow-water scleractinia of Jamaica: Revised list of species and their vertical distribution range. Bull. Mar. Sci. 17: 442-453. HARRIGAN, J. 1972. The planula larva of Pocillopora damicornis; lunar periodicity of swarming and sub- stratum selection behavior. Ph.D. thesis, University of Hawaii, 3 1 3 pp. HARRIOTT, V. J. 1983. Reproductive ecology of four scleractinian species at Lizard Island, Great Barrier Reef. Coral Reefs 2: 9- 18. HARRISON, P. L. 1985. Sexual characteristics of scleractinian corals: Systematic and evolutionary implica- tions. Proc. Fifth Int. Coral Reef Symp. 4: 337-342. HARRISON, P. L., R. C. BABCOCK, G. D. BULL, J. K. OLIVER, C. C. WALLACE, AND B. L. WILLIS. 1984. Mass spawning in tropical reef corals. Science 223: 1 1 86- 1 1 89. HEYWARD, A. J., AND R. C. BABCOCK. 1986. Self- and cross-fertilization in scleractinian corals. Mar. Biol. 90:191-195. HIGHSMITH, R. 1982. Reproduction by fragmentation in corals. Mar. Ecol. Prog. Ser. 1: 207-226. HUGHES, T. P., AND J. B. C. JACKSON. 1980. Do corals lie about their age? Some demographic conse- quences of partial mortality, fission and fusion. Science 209: 7 1 3-7 15. HUGHES, T. P., AND J. B. C. JACKSON. 1985. Population dynamics and life histories of foliaceous corals. Ecol. Monogr. 55: 141-166. JACKSON, J. B. C. 1979. Morphological strategies of sessile animals. Pp. 499-555 in The Biology and Systematics of Colonial Organisms, G. Larwood and B. Rosen, eds. Academic Press, New York. JOKIEL, P. L., R. Y. ITO, AND P. M. Liu. 1985. Night irradiance and synchronization of lunar release of planula larvae in the reef coral Pocillopora damicornis. Mar. Biol. 88: 167-174. KARLSON, R. H. 1986. Disturbance, colonial fragmentation, and size-dependent life history variation in two coral reef cnidarians. Mar. Ecol. Prog. Ser. 28: 245-249. KJNZIE, R. A., Ill, AND T. SARMIENTO. 1986. Linear extension rate is independent of colony size in the coral Pocillopora damicornis. Coral Reefs 4: 1 77- 181. KOJIS, B. L., AND N. J. QUINN. 198 1 . Aspects of sexual reproduction and larval development in the shallow water hermatypic coral, Goniastrea australensis (Edwards and Haime, 1 857). Bull. Mar. Sci. 31 : 558-573. KOJIS, B. L., AND N. J. QUINN. 1 982. Reproductive strategies in four species of Porites (Scleractinia). Proc. Fourth Int. Coral Reef Symp. 2: 145- 1 5 1 . KOJIS, B. L., AND N. J. QUINN. 1984. Seasonal and depth variation in fecundity of Acropora palifera at two reefs in Papua New Guinea. Coral Reefs 3: 165-1 72. KOJIS, B. L., AND N. J. QUINN. 1985. Puberty in Goniastrea favulus age or size limited? Proc. Fifth Int. Coral Reef Symp. 4: 289-293. LOYA, Y. 1 976. The Red Sea coral Stylophora pistillata is an r strategist. Nature 259: 478-480. LUNA, L. G. 1968. Manual of Histologic Techniques of the Armed Forces Institute of Pathology. McGraw- Hill, New York. MARSHALL, S. M., AND T. A. STEPHENSON. 1933. The breeding of reef animals part 1. The corals. Sci. Rep. Great Barrier Reef Exp. 3: 219-245. PETERS, E. C. 1984. A survey of cellular reactions to environmental stress and disease in Caribbean sclerac- tinian corals. Helgol. Meeresunters. 37: 1 13-137. RINKEVICH, B., AND Y. LOYA. 1979a. The reproduction of the Red Sea coral Stylophora pistillata, I. Gonadsand planulae. Mar. Ecol. Prog. Ser. 1: 133-144. RINKEVICH, B., AND Y. LOYA. 1979b. The reproduction of the Red Sea coral Stylophora pistillata, II. Synchronization in breeding and seasonality of planulae shedding. Mar. Ecol. Prog. Ser. 1: 145- 152. RINKEVICH, B., AND Y. LOYA. 1985. Intraspecific competition in a reef coral: Effects on growth and repro- duction. Oecologia 66: 100-105. ROGERS, C. S., H. C. FITZ, III, M. GILNACK, J. BEETS, AND J. HARDIN. 1984. Scleractinian coral recruit- CORAL REPRODUCTION AND GROWTH 177 ment patterns at Salt River Submarine Canyon, St. Croix, U.S. Virgin Islands. Coral Reefs 3: 69- 76. RYLAARSDAM, K. W. 1983. Life histories and abundance patterns of colonial corals on Jamaican reefs. Mar. Ecol. Prog. Ser. 13: 249-260. SHLESINGER, Y., AND Y. LOYA. 1985. Coral community reproductive patterns: Red Sea versus the Great Barrier Reef. Science 228: 1 333- 1 335. SOK.AL, R., AND F. ROHLF. 1969. Biometry. W. H. Freeman and Co., San Francisco. 776 pp. STEARNS, S. 1977. The evolution of life history traits: a critique of the theory and a review of the data. Ann. Rev. Ecol. Sys. 8: 145-171. STEELE, R. G. D., AND J. H. TORRIE. 1960. Principles and Procedures of Statistics (with Special Reference to the Biological Sciences). McGraw-Hill Book Company, Inc., New York. 48 1 pp. STIMSON, J. S. 1978. Mode and timing of reproduction in some common hermatypic corals of Hawaii and Enewetak. Mar. Biol. 48: 173-184. STODDART, J. A., AND R. BLACK. 1985. Cycles of gametogenesis and plantation in the coral Pocillopora damicornis. Mar. Ecol. Prog. Ser. 23: 153-164. SZMANT-FROELICH, A. 1985. The effect of colony size on the reproductive ability of the Caribbean coral Montastrea annularis (Ellis and Solander). Proc. Fifth Int. Coral ReefSymp. 4: 295-300. SZMANT-FROELICH, A. REUTTER, AND L. RIGGS. 1985. Sexual reproduction of Favia fragum (Esper.): Lunar patterns of gametogenesis, embryogenesis and plantation in Puerto Rico. Bull. Mar. Sci. 37: 880-892. SZMANT, A. 1986. Reproductive ecology of Caribbean reef corals. Coral Reefs 4: 43-54. VAN MOORSEL, G. W. N. M. 1983. Reproductive strategies of two closely related stony corals (Agaricia, Scleractinia). Mar. Ecol. Prog. Ser. 13: 273-283. VAUGHAN, T. W. 1 9 1 5. The geological significance of the growth-rate of the Floridian and Bahaman shoal- water corals. J. Wash. Acad. Sci. 5: 59 1-600. WAHLE, C. M. 1983. The roles of age, size and injury in sexual reproduction among Jamaican gorgonians. ' Am. Zool. 23:961. WAHLE, C. M. 1985. Habitat-related patterns of injury and mortality among Jamaican gorgonians. Bull. Mar. Sci. 37: 905-927. WALLACE, C. C. 1 985. Reproduction, recruitment and fragmentation in nine sympatric species of the coral genus Acropora. Mar. Biol. 88: 217-233. WILLIAMS, G. C. 1975. Sex and Evolution. Princeton University Press, Princeton, NJ. 200 pp. WOOD-JONES. 1910. Coral and Atolls. Lovell Reeve &Co., LTD., London. 392 pp. Reference: Biol. Bull. 172: 178-186. (April, 1987) THE FORMATION OF CLONAL TERRITORIES IN EXPERIMENTAL POPULATIONS OF THE SEA ANEMONE ACTINIA TEN E BROS A DAVID J. AYRE1 Department of Zoology, University of Western Australia, Nedlands, Western Australia, 6009, Australia ABSTRACT Fine and gross-scale clumping of electrophoretically identical anemones (geno- typic clumping) was detected 1 1 months after multi-clonal groups of adults had been transplanted into rock pools on two shores. The locomotory separation of non-clone- mates following conflicts appeared to be the primary cause of gross-scale genotypic clumping of adults, and the fine-scale genotypic clumping of adults and recruits. The gross-scale clumping of adults and recruits must reflect the effects of localized asexual recruitment. In contrast, genotypic clumping was not detected within a third popula- tion with lower recruitment, located on a relatively smooth shore. In that population, adults were typically restricted to small individual depressions and movement was rarely detected. These data support the hypothesis that intergenotypic aggression may play an important role in determining the genotypic structure of populations, and indicate that the importance of this factor may be partially determined by the topog- raphy of the shore. INTRODUCTION Within several species of sea anemones, interclonal aggression is thought to play a major role in the formation of clonal aggregations (Francis, 1 973a, b; Purcell, 1 977; Sebens, 1982; Ayre, 1983a). Laboratory trials (Francis, 1973b; Ayre, 1982) and field observations (Ottaway, 1978) have shown that intraspecific aggression leads to the locomotory separation of adult anemones. Nevertheless, the importance of aggres- sion in the field has chiefly been inferred from observations of seemingly static clonal boundaries. The most striking examples have been documented for the fissiparous Anthopleura elegantissima, where anemone-free zones have remained between clones for periods of up to four years (Francis, 1973a). However, in these cases it has not been possible to separate the effects of aggression from those of other factors, such as passive growth and dispersal, which may determine clonal distributions. Further- more, the role of aggression has been obscured by recent reports that aggressiveness may be reduced or absent in pairs of non-clonemate Metridium senile of different sex (Kaplan, 1983) and for pairs of non-clonemate M. senile (Purcell and Kitting, 1982) and Anthopleura xanthogrammica (Sebens, 1984) within genetically diverse aggregations. Actinia tenebrosa is an asexually viviparous, solitary species which is also believed to reproduce sexually, producing widely dispersed, genetically diverse colonists (Black and Johnson, 1979; Ayre, 1983a; 1984a). A. tenebrosa is dioecious (Ottaway, 1979) and it appears likely that each clone is unisexual (Ayre, 1 984c). The majority of recruits into natural populations on stable rock platforms are the asexually produced Received 2 October 1 986; accepted 15 January 1987. 1 Present address: Department of Biology, University of Wollongong, P.O. Box 1 144, Wollongong, N.S.W. 2500, Australia. 178 CLONAL TERRITORIES 179 juveniles of resident adults (Black and Johnson, 1979; Ayre, 1983a). Brooded juve- niles may settle at least several meters from their brood parents and consequently, local populations typically consist of at least several intermingled clones (Ayre, 1983a, 1984a, b). Nevertheless, clones may have restricted (clumped) distributions and it has been suggested that interclonal aggression is used to maintain and extend clonal territories (Ayre, 1983a). Adult A. tenebrosa fight using specialized nemato- cyst-bearing structures called acrorhagi (Ottaway, 1978). These can be used to attack adult or juvenile non-clonemates within tentacle contact (Ayre, 1982). Each fight may last for as little as a few minutes. In some cases acrorhagial scarring persists and is visible for several days. In the field, resident adults are most often able to repel adult immigrants (Ottaway, 1978). In the laboratory, even mild attacks may significantly increase the level of juvenile mortality (Ayre, 1982). This paper documents the active behavioral formation of clumps of genotypically identical anemones within experimental populations of A. tenebrosa and provides an indirect test of the hypothesis that intergenotypic aggression is directed only against non-clonemates of the same sex. MATERIALS AND METHODS Experimental design All experiments were conducted on exposed, sloping limestone shores within three natural populations of A. tenebrosa on the coast of Rottnest Island, Western Australia (32°S; 115°30E). All adult anemones (>10 mm column diameter; Ayre, 1984b) were found within a band of shore approximately 1 m high, in the lower and mid-intertidal zone. The experimental sites were on shores of two topographic types. Two sites were within areas of small rockpools on shores at Green Island and Salmon Point. Pools ranged up to 70 cm in diameter and some individual pools contained more than 50 adults. The third site was an area of smooth shore within one of three populations at Strickland Bay (Ayre, 1983b). These adults were typically isolated within small depressions (2 to 3 cm diameter). To test the hypothesis that the formation of clumps of genotypically identical adult A tenebrosa has a behavioral basis, multi-clonal groups of 40 adults were hap- hazardly transplanted onto each of twenty-seven 1-m long strips of shore. This was a reciprocal transplantation experiment in which adults were transplanted within and between the three populations (Ayre, 1985). Three study sites were established within each population, and at each site a 1 -m strip of shore was allocated to anemones from each of the three populations. Transplants were performed in early November 1980, following the removal of all resident anemones from each of the 27 transplant strips. All adults were transplanted into the zone from which anemones had been removed, but were never intentionally returned to their original positions. All transplanted adults rapidly attached themselves to the limestone shore and all were in place for at least one hour prior to exposure to wave action. Detailed topographic maps, based on sketches and photographs, were prepared for each strip. To ensure that the trans- planted adults could be distinguished from recruits that had grown to > 10 mm (new adults), the locations and sizes of all transplanted adults, new adults, and juveniles were monitored at approximately two-month intervals. After 1 1 months the positions of all surviving adults and new recruits were recorded and all anemones were then collected. Adults were stored, live, at 4°C for up to five days prior to dissection, and adults and all juveniles were then stored at -20°C pending electrophoresis. The sexes of adults were determined on the basis of the color and morphology of their gonads (Ayre, 1984b). 180 D. J. AYRE Electrophoresis All anemones were assayed for the five polymorphic enzymes catalase (CAT, EC 1.11.1 .6), malate dehydrogenase (MDH, EC 1.11.1 .37), mannose phosphate isomer- ase (MPI, EC 5.3.1.8), phosphoglucomutase (PGM, EC 2.7.5.1), and superoxide dis- mutase (SOD, EC 1.15.1.1), as described by Ayre (1983a). Anemones were consid- ered to be non-clonemates if they had different genotypes at one or more loci. Electro- phoretically identical anemones might be either clonemates or non-clonemates. Analysis Clumping of electrophoretically identical (genotypic clumping) anemones was examined on two spatial scales (Ayre, 1983a). Within each population, fine-scale ge- notypic clumping, on a scale consistent with the predicted effects of intergenotypic aggression, was tested by comparing the proportions of anemones with only genotypi- cally identical neighbors within a two-cm radius (i.e., within tentacle contact; Ayre, 1982) and a 2-4 cm radius. It was assumed that if aggression was occurring then a disproportionate number of near neighbors should be electrophoretically identical. Gross-scale genotypic clumping within each whole transplant strip, resulting from differential dispersal or mortality, was tested for by comparing the mean distances between all pairs of electrophoretically identical anemones and all pairs of electro- phoretically different anemones. This analysis was conducted for each multi-locus genotype represented by two or more anemones. Genotypes were considered to be clumped if the mean distance between pairs of anemones with that genotype was less than the mean distance between them and all other anemones. These results were then pooled for all strips within each population, and the proportions of clumped and unclumped genotypes were compared with the expectation that half of all geno- types would have clumped distributions due to chance alone. These tests for geno- typic clumping should be conservative since it may not be possible to distinguish all clones using only five variable loci and the areas of the transplant strips may be small relative to the limits of juvenile dispersal. RESULTS Survival and recruitment At the conclusion of the experiments, the mean numbers (x ± S.E.) of surviving adults per strip were similar within each of the three populations ranging from 1 7 ± 2 at Strickland Bay and Salmon Point, to 2 1 ± 2 at Green Island. As was expected, all adults were restricted to the shelter of pools and depressions whereas many juve- niles settled on exposed shore (Fig. 1 ). Settlement and recruitment varied considera- bly within and between colonies, and 445, 187, and 51 recruits were collected from Green Island, Salmon Point, and Strickland Bay, respectively. Fine- scale genotypic clumping Genotypic clumping on a scale consistent with the predicted effects of intergeno- typic aggression was detected on the rough shore within both the Green Island and Salmon Point populations (Table I). Disproportionately more anemones had only genotypically identical neighbors within a 2 cm radius compared with the propor- tions within a 2- to 4-cm radius. Significant levels of fine-scale clumping were detected within the groups of transplanted adults collected from Salmon Point and within the groups of transplanted adults and recruits collected from both Green Island and CLONAL TERRITORIES 181 a FIGURE 1 . The distribution of all adult and juvenile Actinia tenebrosa 1 1 months after a multi-clonal group of 40 adults had been transplanted haphazardly into rock pools (outlines shown as lines) on a 1- meter wide strip of shore at Green Island. Each anemone is represented by a letter denoting its size and age, and 5-locus genotype. A total of 1 8 different genotyopes are represented. This strip was chosen at random from the set of 27 experimental sites. Transplanted adults = enclosed upper case letter; adult recruits = upper case letter-Juvenile recruits = lower case letter. Salmon Point (Table I). In contrast there was no such clumping of electrophoretically identical adults at Strickland Bay. In fact, a greater proportion of anemones had elec- trophoretically different neighbors within a 2-cm radius as compared with a 2- to 4- cm radius (Table I), although this difference was not statistically significant. There was some evidence of fine-scale genotypic clumping of transplanted adults and re- cruits at Strickland Bay, however these data provided only a weak test for such clump- ing as only 10 recruits were found within 4-cm of transplanted adults (Table I). Gross-scale genotypic clumping Each of the 27 transplant strips was found to support anemones with at least 3, and typically 5 or more, 5-locus genotypes. Genotypes were intermingled on the shore, and many juveniles were detected outside the limits of adult distributions (Fig. 1 ). Nevertheless, for transplanted adults, a significant majority of genotypes were judged to be clumped within the Green Island and Salmon Point strips (Table II), although these results were not statistically significant for either individual popula- tion. In contrast only half of the 45 groups of genotypically identical adults on the smooth shore at Strickland Bay were judged to be clumped, the proportion expected 182 D. J. AYRE TABLE I Comparison of the numbers of adult Actinia tenebrosa with only genot ypically identical adult neighbors1 (S) or at least one genotypically different neighbor (D) within a 2-cm and 2 to 4-cm radius Separation <2 cm 2 to 4 cm Population Pairing S D S D X2(n,i Green Island Adult-adult 18 12 17 28 2.74 <0.05 Green Island Adult-recruit 65 28 32 68 26.2 <0.001 Salmon Point Adult-adult 19 4 13 17 6.83 <0.005 Salmon Point Adult-recruit 14 13 5 21 7.62 <0.005 Strickland Bay Adult-adult 12 21 15 12 1.50 N.S. Strickland Bay Adult-recruit 5 0 2 3 Fisher's Exact Test 0.083 Comparisons are made for pairs of transplanted adults, and transplanted adults and adult and juvenile recruits. Data are presented for anemones within three experimental populations, which had been estab- lished 1 1 months earlier by haphazardly transplanting multi-clonal groups of 40 adult anemones onto cleared areas of shore. a Based on 5 locus-genotype. N.S. P> 0.05. by chance alone (Table II). In addition, at Green Island, a significant majority of the groups of electrophoretically identical transplanted adults and recruits were judged to be clumped (Table II, P < 0.001). In contrast only half of such groups were clumped within the Salmon Point and Strickland Bay populations (Table II), how- ever, this may reflect the underlying unclumped distribution of genotypes represented by transplanted adults. Furthermore, the present experimental design provides only a weak test for gross-scale genotypic clumping involving recruits since each test was restricted to only a 1-m2 experimental strip. In natural populations juveniles may be dispersed passively over several meters and genotypic clumps are often spread over at least several square meters of shore (Ayre, 1983a). Therefore, this difference between populations may be a real environmental effect but cannot be considered evidence that the same processes do not operate on all three shores. Effects of sex The distribution of the 59 pairs of electrophoretically distinct sexually mature adults separated by <4 cm did not support the hypothesis that aggression is directed only towards non-clonemates of the same sex. The proportions of electrophoretically different, same-sex and mixed-sex pairings separated by <2 cm and 2-4 cm were not significantly different (Table III). This implies that if intergenotypic aggression is a major determinant of the distribution of genotypes on these shores then the initiation of conflicts is not affected by similarity of sex. DISCUSSION The rapid formation of genotypic clumps within experimental populations of adult A. tenebrosa supports the hypothesis that clonal distributions may be, at least partially, determined by the effects of interclonal aggression and subsequent migra- CLONAL TERRITORIES 183 TABLE II The incidence of gross genotypic" clumping within experimental groups of Actinia tenebrosa in each of three populations Genotype Group Population Clumped Not-clumped X2(l),l Transplanted adults Green Island 18 11 N.S. Salmon Point 15 7 N.S. Strickland Bay 23 22 N.S. Transplanted adults Green Island 50 19 13.04 <0.001 and recruits Salmon Point 23 22 N.S. Strickland Bay \2 14 N.S. Data are presented for adult anemones which had been haphazardly transplanted 1 1 months prior to sampling, and for those adults and all adult and juvenile recruits. Chi-squared analyses were used to deter- mine if a significant majority of genotypes were clumped. 8 For each 5-locus genotype, anemones were judged to be clumped if the mean distance between all anemones with that genotype was less than the mean distance between all anemones with different genotypes. N.S. not significant P > 0.05. tion (Ayre, 1983a). Similarly, genotypic clumping has been described within natural populations of A. tenebrosa (Ayre, 1983a) where interclonal conflicts have been ob- served. These findings, together with experimental demonstrations that these clones are locally adapted (Ayre, 1985), support the theoretical prediction that inter-clonal competition should restrict clones to those parts of the habitat to which they are best adapted (Williams, 1975). In this study both fine and gross-scale genotypic clumping developed within at least some of the multi-clonal groups of adults which had been established 1 1 months earlier by haphazard transplantation. Genotypic clumping on both scales is most simply explained by the differential migration of genotypes following interclonal con- flicts, or by the preferential association of clonemates. Fine-scale clumping of adults occurred on a scale that was consistent with the predicted, aggressive repulsion of non-clonemate near neighbors. Disproportionately more pairs of anemones within TABLE III Comparison of the numbers of pairs ofgenotypically different ' adult Actinia tenebrosa with the same or different se.\ which were separated by <2 cm and 2 to 4 cm Separation Pairings <2 cm 2 to 4 cm Same sex 8 15 N.S. X (n.i=0.32 Male-female 10 26_ Results were pooled for groups of anemones which had been haphazardly transplanted into 3 popula- tions 1 1 months prior to sampling. ' Based on 5 locus-genotype. N.S. P> 0.05. 184 D. J. AYRE the range of tentacle contact were genotypically identical than those just outside the range of contact (2-4 cm). Such fine-scale genotypic clumping could also be ex- plained by the differential migration of adult clonemates into favored micro-habitats, but this seems inconsistent with the overlapping gross-scale distribution of genotypes detected in this study (Fig. 1) and within natural populations (Ayre, 1983a). Gross- scale genotypic clumping on rough shores could have resulted from the aggressive repulsion and subsequent migration of non-clonemates, differential migration of ge- notypes into particular microhabitats, or differential mortality of clones within mi- crohabitats. It is not possible to distinguish between the first two hypotheses. How- ever, the third hypothesis is weakened by the fact that, although mortality rates were similar in each population, no evidence of gross genotypic clumping was detected at Strickland Bay. At Strickland Bay pedal locomotion is severely restricted by the topography of the shore, since anemones are almost all restricted to small, individual pools. As might be predicted, because adult anemones on that shore were rarely within tentacle contact, no intraspecific fighting was observed or acrorhagial scarring detected during a three-year study of both the experimental and natural populations at that location. In contrast, 12 fights between pairs of adults were inferred from the presence of acrorhagial scarring during the present 1 1 -month study of experimental populations at Green Island and Salmon Point. In eight of these cases one adult in each pair died or was lost from the study site within the following two months, which may be compared with an average of 7% mortality per 2-month period for all adults in this experiment (Ayre, 1985). At Green Island, statistically significant gross-scale genotypic clumping of trans- planted adults and adult and juvenile recruits was superimposed upon the already clumped distributions of the groups of transplanted adult clonemates. This result strongly supports the prediction that, within natural populations, clonal territories are extended by the combined effects of interclonal aggression and the passive dis- persal of asexually produced juveniles (Ayre, 1983a). Passive dispersal should always play a major role in extending clonal boundaries. However, the fine-scale clumping of adults and juveniles, detected at Green Island and Salmon Point, supports the hypothesis that this effect may be strengthened by intergenotypic aggression (Ayre, 1983a). Such fine-scale genotypic clumping of adults and juveniles should be ex- pected to develop on all types of shore as a result of attacks on juvenile non-clone- mates which settle close to adults. Nevertheless, it is clear from this study and earlier work (Ayre, 1985) that environmental factors affecting populations can greatly in- fluence the rate of expansion of clones through effects on asexual fecundity, recruit- ment, and the frequency of intergenotypic conflicts. The conclusions of this and earlier studies of the genetic structure of populations of A. tenebrosa (Black and Johnson, 1979; Ayre, 1983a, 1984a, b) are in apparent contrast with the outcome of studies of the northern hemisphere species Actinia equina var mesembryanthemum. Several authors have claimed that in British popu- lations there is little fine or gross-scale clumping of clonal genotypes and that asexu- ally produced juveniles are in fact widely dispersed (Orr et al, 1982; Quicke and Brace, 1983; Brace and Quicke, 1 985, 1 986a, b). The latter conclusion is based princi- pally on the finding that single locus genotype frequencies approach the values pre- dicted for Hardy- Weinberg equilibria, an outcome which is more consistent with the recruitment of sexually reproduced recruits. In addition, Brace and Quicke (1986b) have concluded that for A. equina the principal role of intraspecific aggression is "to act to enhance adult survivorship through the securement of space . . ." Although Brace and Quicke (1985) emphasize the apparent differences between populations of the two species, they recognize that in their studies environmental differences be- CLONAL TERRITORIES 185 tween sites may account for differences in the importance of both localized asexual recruitment and inter-genotypic aggression. In fact they conclude that in the single area in which genotypic clumping was detected anemones may have been less suscep- tible to dislodgement resulting from wave action. It could also be argued that only two of their four study sites contained anemones at sufficiently high densities (<2.5 cm mean nearest neighbor distance) to provide a meaningful test of the effects of inter-genotypic aggression, and that at least three of their study sites were too small (0. 1 8-0.59 nr) to provide a useful test for gross-scale genotypic clumping. This appar- ent contrast of life-history tactics may therefore simply reflect the decision to study populations of A. equina in which low levels of asexual recruitment and interclonal aggression are favored, and the use of different methods of analysis in the Australian and British studies. It is certainly true that interpopulation variation with respect to many life-history characters has been documented for other anemones, including the contributions of sexual and asexual reproduction to recruitment, as inferred from allozyme data, for A. tenebrosa (Ayre, 1984a) and Metridium senile (Hoff- mann, 1986). The present results do not support the idea that the aggressive response of adult A. tenebrosa to adult non-clonemates (Ayre, 1982) is reduced when pairs of adults are of different sex. Similar proportions of same sex and mixed sex pairs were detected within and immediately beyond the range of tentacle contact. Since only 59 suitable pairings were recorded, it is possible that similarity of sex may have some influence on the aggressiveness of pairs. This finding contrasts with Kaplan's ( 1 983) clear dem- onstration that mixed sex pairs of Metridium senile would not fight. Sebens (1984) also failed to find any evidence that the incidence of aggression is affected by similar- ity of sex in populations of Anthopleura xanthogrammica. These results, together with other reports of variation in the sensitivity of histocompatibility reactions in coelenterates (Resing and Ayre, 1985) and the finding that factors such as age (Hi- daka, 1985) and prolonged contact with non-clonemates may modify their expres- sion (Purcell and Kitting, 1982; Sebens, 1984), highlight our limited understanding of these processes. Although the sensitivity of the histocompatibility response has not been thor- oughly tested for A. tenebrosa, it is clear that intergenotypic aggression plays an im- portant role in determining the genotypic structure of populations of A. tenebrosa on some shores, where it must therefore play an important role in inter-clonal competi- tion. The finding that genotypic clumping did not develop on the smooth shore at Strickland Bay implies that interclonal aggression accelerates and enhances the effects of passive asexual dispersal on the development of clonal territories. The reciprocal transplant study which generated the data described in the present study revealed that clones of this species were highly locally adapted. With regard to asexual fecundity, however, the importance of inter-genotypic aggression as a component of clonal fit- ness has not been determined. It seems likely that aggressiveness should be most strongly selected for on shores which support moderate or high densities of mobile adults. ACKNOWLEDGMENTS I thank Mike Johnson for advice and encouragement throughout this study, Kerry Ayre for skilled field assistance, and Joe Neal for assistance with data analysis. Kerry Ayre, Jim Campbell, Jennie Chaplin, Alison Hunt, Dale Roberts, and Rob Whelan provided comments which substantially improved this manuscript. This study was undertaken while the author held a University of Western Australia Re- 186 D. J. AYRE search Studentship and is a contribution from the Marine Biological Laboratory of the University of Western Australia. Accommodation was provided by the Rott- nest Island Biological Research Station. Typing, graphics, and computing facilities were provided by the Australian Institute of Marine Science and the University of Wollongong. LITERATURE CITED AYRE, D. J. 1982. Inter-genotype aggression in the solitary sea anemone Actinia tenebrosa. Mar. Biol. 68: 199-205. AYRE, D. J. 1983a. The effects of asexual reproduction and intergenotypic aggression on the genotypic structure of populations of the sea anemone Actinia tenebrosa. Oecologia 57: 158-165. AYRE, D. J. 1983b. The distribution of the sea anemone Actinia tenebrosa in south-western Australia. West. Aust. Nat. 15: 136-140. AYRE, D. J. 1984a. The effects of sexual and asexual reproduction on geographic variation in the sea anemone Actinia tenebrosa. Oecologia 62: 222-229. AYRE, D. J. 1984b. The effects of environment and population density on the sea anemone Actinia tene- brosa. Aust. J. Mar. Freshwater Res. 35: 735-746. AYRE, D. J. 1984c. The effects of reproductive mode and colonising ability on the genetics of populations of the sea anemone Actinia tenebrosa. Ph.D. thesis, University of Western Australia. 108 pp. AYRE, D. J 1985. Localized adaptation of clones of the sea anemone Actinia tenebrosa. Evolution: 39: 1250-1260. BLACK, R., AND M. S. JOHNSON. 1979. Asexual viviparity and population genetics of Actinia tenebrosa. Mar. Biol. 53: 27-3 1 . BRACE, R. C, AND D. L. J. QUICKE. 1985. Further analysis of individual spacing within aggregations of the anemone Actinia equina. J. Mar. Biol. Assoc. U.K. 65: 35-53. BRACE, R. C., AND D. L. J. QUICKE. 1986a. Dynamics of colonization by the beadlet anemone, Actinia equina. J. Mar. Biol. Assoc. U.K. 66: 21-47. BRACE, R. C., AND D. L. J. QUICKE. 1986b. Seasonal changes in dispersion within an aggregation of the anemone, Actinia equina, with a reappraisal of the role of intraspecific aggression. J. Mar. Biol. Assoc. U.K. 66: 49-70. FRANCIS, L. 1973a. Clone specific segregation in the sea anemone Anthopleura elegantissima. Biol. Bull. 144: 64-72. FRANCIS, L. 1 973b. Intraspecific aggression and its effects on the distribution of Anthopleura elegantissima and some related sea anemones. Biol. Bull. 144: 73-92. HIDAKA, M. 1985. Tissue compatibility between colonies and between primary polyps of Pocillopora da- micornis. In Coral Population Biology. P. L. Jokiel and R. H. Richmond, eds. HIMB Tech Rpt 37. HOFFMANN, R. J. 1986. Variation in contributions of asexual reproduction to the genetic structure of populations of the sea anemone Metridium senile. Evolution 40: 357-365. KAPLAN, S. W. 1983. Intrasexual aggression in Metridium senile. Biol. Bull. 165: 416-418. OTTAWAY, J. R. 1978. Population ecology of the intertidal anemone Actinia tenebrosa I. Pedal locomotion and intraspecific aggression. Aust. J. Mar. Freshwater Res. 29: 787-802. OTTAWAY, J. R. 1979 Population ecology of the intertidal anemone Actinia tenebrosa II. Geographical distribution, synomony, reproductive cycle and fecundity. Aust. J. Zool. 27: 273-290. ORR, J., J. P. THORPE, AND M. A. CARTER. 1982. Biochemical genetic confirmation of the asexual repro- duction of brooded offspring in the sea anemone Actinia equina. Mar. Ecol. Prog. Ser. 7: 227- 229. PURCELL, J. E. 1977. Aggressive function and induced development of catch tentacles in the sea anemone Metridium senile (Coelenterata: Actinaria). Biol. Bull. 153: 355-368. PURCELL, J. E., AND C. L. KITTING. 1982. Intraspecific aggression and population distributions of the sea anemone Metridium senile. Biol. Bull. 162: 345-359. QUICKE, D. L. J., AND R. C. BRACE. 1983. Phenotypic and genotypic spacing within an aggregation of the anemone Actinia equina. J. Mar. Biol. Assoc. U. K. 66: 49-70. RESING, J. M., AND D. J. AYRE. 1985. The usefulness of the tissue grafting bioassay as an indicator of clonal identity in scleractinian corals (Great Barrier Reef: Australia). Proc. 5th Int. Coral Reef Congr.6:15-S\. SEBENS, K. P. 1982. Asexual reproduction in Anthopleura elegantissima (Brandt )(Anthozoa: Actiniaria). Ecology 63: 343-444. SEBENS, K. P. 1984. Agonistic behaviour in the intertidal sea anemone Anthopleura xanthogrammica. Biol. Bull. 166: 457-472. WILLIAMS, G. C. 1975. Sex and Evolution. Princeton University Press, Princeton, NJ. Reference: Biol. Bull. 172: 187-201. (April, 1987) THE ALLOMETRY OF DEPOSIT FEEDING IN CAPITELLA SPECIES I (POLYCHAETA:CAPITELLIDAE): THE ROLE OF TEMPERATURE AND PELLET WEIGHT IN THE CONTROL OF EGESTION THOMAS L. FORBES AND GLENN R. LOPEZ Marine Sciences Research Center, State University of New York, Stony Brook, New York 11794 ABSTRACT This study investigates the relationships between egestion rate, body size, and environmental temperature in the opportunistic marine polychaete Capitella species I. Measurements were made of (1) fecal pellet weight, (2) pellet production rate, and (3) pellet standing stock within the gut of live worms. Egestion rate experiments were conducted with worms ranging in size from 0.27 mm3 to 2.62 mm3 (approx. 1.0-15 mm in length). Pellet production rate measurements were made at 1 5°, 20°, and 25°C. Individual fecal pellet weight was related to worm body volume to the 0.70 power. Fecal pellet production rate in Capitella sp. I was independent of body size. The number of pellets maintained on average within the guts of larger animals is at least equal to that of smaller animals. Fecal pellet production rate increased exponentially with increasing temperature between 15°C and 25°C, with an overall Q]0 value of 2.49. Power functions relating changes in egestion rate (/ug sediment h"1) and size- specific egestion rate [/ug dry weight sediment (mm3 worm)"1 h~'] to body size were fit to the data. These curves show that egestion rate scales as body volume to the 0.70 power, indicating that larger Capitella sp. I specimens process relatively less sediment per unit body volume than smaller worms. Capitella sp. I individuals thus control sediment processing rate during ontogeny by reducing the relative pellet weight as they grow. Estimates of the scaling of external surface area in Capitella sp. I show that surface area scales as worm size to the 0.77 power, paralleling the scaling of feeding rate to body size. Data on oxygen uptake rates at different temperatures for Capitella sp. I (Cammen, 1985) are re-examined in light of our results and the implications for a coherent metabolic strategy (sensu Newell, 1980) are discussed. We hypothesize that a physiological surface such as the external respiratory surface or the absorptive area of the gut surface may limit growth and anabolism in Capitella species I. INTRODUCTION Few studies have addressed the effects of temperature and body size on egestion rates within single species of deposit-feeding invertebrates (Hargrave, 1972; Hylle- berg, 1975; Kudenov, 1982). This study was designed to measure changes in feeding rate versus body size at different temperatures for small, actively growing Capitella species I. This species has been shown to be one of a complex of sibling species, all formerly classified as Capitella capitata, each of which exhibits distinctive life history characteristics (Grassle and Grassle, 1976, 1977, 1978; Grassle, 1979). Throughout Received 28 May 1986; accepted 29 January 1987. Contribution No. 539 from the Marine Sciences Research Center, State University of New York, Stony Brook. 187 188 T. L. FORBES AND G. R. LOPEZ this paper we adopt the convention of referring to the worm by sibling species where it has been determined, and as Capitella capitata when referring to studies where the species is unknown. Capitellid polychaetes typically occur, occasionally in great numbers, in organi- cally enriched nearshore marine environments (e.g., Grassle and Grassle, 1974; Mc- Call, 1977; Pearson and Rosenberg, 1978; Rhoads et al, 1978; Thistle, 1981). The genus Capitella has been termed an 'enrichment opportunist' by Pearson and Rosen- berg (1978) and laboratory studies have documented that small individuals of Capi- tella sp. I can attain daily weight-specific growth rates as high as 21% (Tenore and Chesney, 1985). These characteristics make Capitella sp. I an ideal animal in which to study the allometry of feeding during the period of rapid growth and maturation. Data on the allometry of feeding or egestion rate among species of deposit feeders indicates that feeding rate increases as a function of body size to a power less than one (Hargrave, 1972; Cammen, 1980). This means that larger animals will process relatively less sediment per unit weight than smaller animals. Hargrave ( 1 972) discov- ered, in comparisons primarily among species of lacustrine and marine deposit feed- ers, that feeding rate scaled to approximately the 2/3 power of body weight. In a compilation of data encompassing 1 9 deposit-feeding species and spanning 3 phyla, Cammen (1980) found that organic matter ingestion was a function of body weight to the 0.74 power. These results suggest that egestion or feeding rate may be related to a physiological surface that scales to body size to approximately the 0.7 to 0.8 power (Pauly, 1981). Studies of the allometry of egestion rate within a single species of polychaete, [the so-called 'dynamic' or ontogenetic allometry of Calder (1984)] indicate that the allometric exponent of feeding rate versus body size can be quite variable (Cadee, 1979; Hobson, 1967; Nichols, 1974; Kudenov, 1982; Dobbs and Scholly, 1986). These differences in the size-scaling of feeding rate may reflect differing metabolic strategies between species of infaunal polychaetes. Several studies have measured the effect of temperature on feeding rate in poly- chaetes (Gordon, 1966; Cadee, 1976, 1979; Kudenov, 1982; Dobbs, 1983). Of these most were principally concerned with estimates of sediment reworking rate in the field, and the effect of temperature change was followed seasonally (Cadee, 1976, 1979; Kudenov, 1982). The effect of temperature may have been confounded with changes in food level and/or quality, worm density, and other random variables. Laboratory microcosm studies by Dobbs (1983) and Gordon (1966) have found that egestion rate varied linearly with temperature in Clymenella torquata and Pectinaria gouldii, respectively. Polychaetes of the genus Capitella are often found in organic-rich environments which frequently have very low to nonexistent interstitial oxygen partial pressures (e.g., Wells and Warren, 1975; Pearson and Rosenberg, 1978 and references therein). This suggests the possibility that Capitella species I may at times be limited energeti- cally more by oxygen than by food supply in nature. Thus, an animal's ability to digest and absorb food would be greater than its ability to procure oxygen to effi- ciently metabolize food. One might also postulate an alternative hypothesis that growth and anabolism in Capitella sp. I are limited by food or energy availability (Tenore, 1975, 1977a, b, 1981; Tenore etal, 1979; Tenore and Hanson, 1980). Un- der this hypothesis, at high food levels, anabolism would ultimately be limited by a surface area such as that of the gut absorptive area or the worm's ability to digest and absorb food or essential nutrients. The hypotheses of food or oxygen limitation need not be mutually exclusive, and indeed, interactions between the two may be impor- tant to animals in nature. A critical assumption of this type of analysis is that animals FEEDING ALLOMETRY IN CAPITELLA 189 are built 'reasonably,' and that the evolutionary process will tend to minimize energy expenditure where possible. For example, ifCapitella sp. I were energetically limited during growth by the allometry of gut or respiratory surface area, then the processing of additional sediment above an amount that could be effectively digested and ab- sorbed, or metabolized by the oxygen reaching the mitochondria, would be an added cost with no additional benefit. One might then expect natural selection to reduce this added cost of feeding as body size increases through an egestion rate that scales in proportion to the relevant physiological surface(s). We have performed experiments which indicate that egestion rate (/*g h~') in Ca- pitella sp. I scales as body size to the 0.70 power and increases exponentially as a function of acute temperature change. Data are also reported which directly measure the size-scaling of one potentially growth-limiting surface, that of external surface area, which scales as body size to the 0.77 power. These results are then compared to studies of respiration as a function of temperature in Capitella sp. I and other polychaetes. In addition, we discuss the results in view of the hypothesis that a physio- logical surface may limit growth or net energy gain in Capitella species I. MATERIALS AND METHODS Sediments The sediment for all egestion rate experiments was obtained on 8 November 1 983 from an intertidal mudflat located near the Flax Pond Marine Laboratory, Long Is- land, New York. The top 1-2 cm of mudflat sediment was sieved to <61 ^im on a Nitex screen, thoroughly homogenized, transferred to one pint freezer containers, and frozen at — 20°C. Organic content of the sediment was estimated from weight loss on ignition (475°C, 6 h) after drying (60°C, 48 h) to be approximately 10%. Laboratory culture conditions Small juvenile and adult Capitella species I individuals (obtained from Dr. K. Tenore) were used in all egestion rate experiments. The worms were maintained in mass culture in 720 cm2 aerated plastic containers layered with 1-3 cm clean fine sand (<300 yum) at 15°C in a recirculating seawater aquarium. Animals were fed a 7 g nT2 daily ration of Gerbers Mixed Cereal (Tenore, 1981). The cereal was first ground with a mortar and pestle and then stirred into the overlying culture water. Salinity in the mass culture varied between 25%o and 35%o . The live worms used for the determination of the number of fecal pellets per gut and the number of fecal pellets per unit animal volume had been maintained in mass culture (28%o S, 20°C) on Flax Pond silt-clay (<6 1 nm) surface sediment. Determination of body-size relationships To determine the relation between body size and fecal pellet weight, between 200 and 800 fecal pellets (depending on size) were counted and filtered onto preweighed Nuclepore® membrane filters, rinsed with several ml 3% NH4COOH to remove salts, and dried for 24 hours at 55°C. The filters were then weighed on a Cahn 26 microbal- ance to ± 1 ^g to measure pellet dry weight. Thus each point represents a value for several hundred individual pellets. To measure worm dry weight, preserved animals were placed on preweighed Nuclepore® filters, dried for 24 hours (55°C), and weighed to±l jig. Animal volumes for egestion rate experiments (uncorrected for gut lumen) were 190 T. L. FORBES AND G. R. LOPEZ measured by drawing the animals with the aid of a camera lucida and assuming a two-dimensional projection of a cylinder (Self and Jumars, 1978). To estimate the number of fecal pellets per unit body volume, live worms were removed from mass culture on silt-clay sediment (20°C) and rapidly transferred to glass petri dishes (4 cm diam.) where they were relaxed in isosmotic MgCl2 solution. The number of fecal pellets in each worm gut is readily determined by flattening the worm under a glass cover slip. Fecal pellets defecated during the relaxation period were added to those counted within the gut. Body volumes of relaxed animals were determined in an analogous manner to those of preserved worms using a video camera mounted on a dissecting microscope. The precision of volume measurement for both preserved and relaxed worms is within 5%. In order to compare live and preserved worms we have assumed that fixation does not alter body volume. Measurements of the projected area of the worms used in the feeding rate experi- ments were made by camera lucida drawings of the preserved worms. Feeding rate versus temperature experiments The worms were removed from mass culture and transferred without food to the experimental temperature ( 1 5°, 20°, 25°C) for 24 hours prior to the start of each eges- tion rate experiment. Thus, the feeding rates measured at 20° and 25°C were based on acute temperature changes. Egestion rate experiments were run for either 6 (I, II, III, V) or 26 (IV) hours. The experimental salinities were 24%o (III, IV) and 30%o (I, II, V). Upon completion of each experiment animals were cleaned of adhering debris, fixed in a seawater solution of 20% buffered formaldehyde for 24 hours, and then transferred to 70% ETOH. All animals had a preserved body volume of between 0.27 and 2.62 mm3. These body volumes correspond to worm lengths of approximately 1.0 to 15 mm. Feeding rate experiments (I through V) were conducted on worms placed singly in small (4 cm diameter) glassware dishes in 2-3 mm of Flax Pond silt- clay sediment. The animals were then allowed to feed undisturbed at the experimen- tal temperature for 6 or 26 hours. None of the animals ever pelletized all of the avail- able sediment. Feeding rate was determined by sieving (85 ^m) and counting all fecal pellets defecated by an individual worm over the duration of the feeding experiment. Food density was manipulated by adding a known volume of non-nutritive diluent (clean glass beads) to sediment (1 beads:4 settled sediment, v/v). The glass beads added were in the range of particle diameters (3-85 ^m) ingestible by Capitella sp. I of experimental size. Feeding rate was expressed as fecal pellets animal"1 h~'. Body size versusfeeding rate A power function was fit to the data to describe the relationship between individ- ual fecal pellet weight and worm volume. Plots of feeding rate (^g sediment processed h"1) and worm volume at the three experimental temperatures were made by multi- plying the number of fecal pellets produced by an individual worm times the expected fecal pellet weight calculated from the relation between pellet weight and worm vol- ume. Worm volume-specific [/ug dry weight (DW) sediment (mm3 worm)"1 h~'] eges- tion rate plots were made by dividing feeding rate (;ug h~ ') by worm volume. Geomet- ric mean (GM) regressions were used to construct all functional relationships while an ordinary predictive least-squares regression was used for the relation between pre- served animal volume and dry weight (Ricker, 1973; 1984 and references therein). FEEDING ALLOMETRY IN CAP1TELLA 191 o 3.6 3.4 3.2 3 2.8 2.6 2.4 2.2 2 1.8 1.6 1.4 1.2 0.8 0.6 0.4 0. 0 DO 200 400 WORM DRY WEIGHT (ug) 600 FIGURE 1 . Relationship between estimated preserved volume and dry weight for Capitella species I. V = 0. 127 + 0.006DW, r = 85. 1%, n = 40; where V = worm volume (mm3) and DW = worm dry weight (Mg). RESULTS Body- size relationships There is a positive linear relationship between individual worm dry weight and estimated body volume (Fig. 1 ). Individual fecal pellet weight was allometrically re- lated to worm volume (Fig. 2). Large worms form relatively lighter fecal pellets. The relation between projected area and worm dry weight is shown in Figure 3. GM re- gression of the data indicates that projected area Capitella sp. I scales as dry weight to the 0.77 power. The number of pellets per unit live worm volume was found to decrease exponentially as body size increased (Fig. 4A). However, the relationship between the absolute number of pellets per worm and body volume was positive but nonsignificant (r = .275, P > 0.05, n = 21) (Fig. 4B). Thus, the number of pellets maintained on average within the guts of larger animals is at least equal to that of smaller animals. This is true in spite of the fact that the smallest worms measured had approximately ten times more pellets per unit body volume than the largest ani- mals (Fig. 4A). The average number of pellets within the gut over the entire range of worm body sizes was 30.2. Feeding rate versus temperature The results of egestion rate experiments I through V are given in Table I. Mean egestion rates ranged from approximately 1 3 pellets animal"1 h~' at 1 5°C to 34 pellets animal"1 h"1 at 25°C. Egestion rate (pellets animal"1 h"1) did not vary with animal size (r = -.109, P > 0.20, n = 42). Differences in salinity or duration of the experi- ment had no influence on mean egestion rate (Table I). Experiments II and V were carried out under identical conditions with the exception that in experiment V the sediment was diluted with a known volume of glass beads. The worms showed no 192 T. L. FORBES AND G. R. LOPEZ "^ 5 l- x S 4 ^ 3 2 LU 0. 5 LU U_ 0 > 0 I 2 3 WORM VOLUME (mm3) FIGURE 2. Allometric relation between individual fecal pellet weight and preserved Capitella sp. I volume. PW = 2.270V0701, r2 = 93.1%, n = 17; where PW = individual pellet weight (/*g) and V = worm volume (mm3). difference in egestion rate between experiments II and V. Thus, Capitella sp. I did not alter its egestion rate in response to changes in food density. An exponential relationship exists between egestion rate and temperature (Fig. 5). The calculated Q(0 values for egestion rate between 1 5°C and 20°C and between 20°C and 25°C are 1.52 10 (— 12 D 1 1 - a 10 _ o CM E 9 - _£ 8 , a a LU 7 D a cc 6 a a Q UJ 5 QDD of o h— 1 1 1 4 t? aP° UJ 0 3 [JD ^fl OL 2 - °*° 1 - n 1 1 1 1 1 200 400 WORM DRY WEIGHT (ug) 600 FIGURE 3. Allometric relation between worm projected area and dry weight. A = 0. 109DW0 772, r2 78.7%, n = 40; where A = projected area (mm2) and DW = worm dry weight FEEDING ALLOMETRY IN CAP1TELLA 193 100 90 0.05, n = 2 1 ). 194 T. L. FORBES AND G. R. LOPEZ TABLE I Summary ofegestion rate experiments Expt. Date °C S%0 Egestion rate Mean (S.D.) (n) Time (h) I 29 Nov 83 25 30 34.12(12.99)(11) 6 II 1 1 Feb 84 15 30 13.50 (7.657) (10) 6 III 27 Sept 83* 20 24 17.10(5.167)(6) 6 IV 27 Sept 83* 20 24 16.70 (4.433) (6) 26 27 Sept 83** 20 24 16.90 (4.600) (12) — V 19 Nov 83* 15 30 13.94 (5.284) (10) 6 Egestion rates are in fecal pellets per animal hour (mean) (S.D.) (n). * Glass beads added. ** 27 Sept 83 experiments combined. duction rate was independent of body size. The relationship between volume-specific egestion rate ([^g (mm3 worm)"1 h~'j) and body size is shown in Figure 7. These data were fit to allometric power functions and the parameters are reported in Figures 6 and 7. Egestion rate is related to body size to the 0.70 power, indicating that large worms feed at a lower rate per unit body volume than small worms (Fig. 7). The fact that all curves for the three experimental temperatures have an identical slope when log-transformed indicates that a large animal has a greater absolute change in egestion rate with increasing temperature whereas a smaller worm has a greater relative change in egestion rate. DISCUSSION Body size relationships Like many other deposit-feeding invertebrates, egestion rate in Capitella species I (fig DW sediment h"1) is related to body size by an allometric exponent less than one 50 45 - 40 0. - 30 LU £ 25 rr z 20 o H , c (ft I D LJ S 10 5 i i i i i i i i i i i i 10 12 14 16 18 20 22 24 26 28 30 TEMPERATURE °C FIGURE 5. Changes in fecal pellet production rate with temperature. FP h ' = 0.94exp(0. 15T), r2 = 33.1%, n = 43; where T = °C. Error bars ± 1 S.D. FEEDING ALLOMETRY IN CAP1TELLA 195 IDVJ 0 25°C 140 0 ,-. 0 J= 120 0 0> Q 3 1 1 1 100 0 LU i- 0 cr 80 0 + z o 60 o + + «rc + a ° CO 0 + a UJ o 40 + a ° LJ -0 + a 20 a .4 Q i i i i i i i i i i i i 0.2 0.6 1.0 1.4 1.8 2.2 2.6 WORM VOLUME (mm3) FIGURE 6. Calculated curves showing changes in estimated egestion rate versus worm volume at the three experimental temperatures. The parameters for the allometric functional relationships at the three experimental temperatures are a(15) = 31.14; a(20) = 38.36; a(25) = 77.45; b = 0.701. Where ^g D\V sediment h~' = a(T)Vb. V = worm volume (mm3) and a(T) = temperature dependent constant. (Hargrave, 1972; Cammen, 1980 and references therein). The change in the specific feeding rate ofCapitella sp. I [yug DW sediment (mm3 worm)"1 h~'] with body size is due entirely to two observed relationships: ( 1 ) Fecal pellet production rate does not '£ o ro CO 120 100 80 E E CT 60 5 40 tr z o 20 0 0 0 0 0 0 '-> * * * • • 25°C 0 "D + D D + + + + 20°C ° D a D a I5°C i i i i i i 1 1 1 1 1_ i 0.2 0.6 1.0 1.4 1.8 WORM VOLUME (mm3) 2.2 2.6 FIGURE 7. Calculated curves showing changes in estimated relative egestion rate with worm volume. The parameters for the allometric functional relationships at the three experimental temperatures are a'( 15) = 31.14; a'(20) = 38.36; a'(25) = 77.45; b = -0.299. Where [^g dry weight sediment (mm3 worm) 'h '] = a'(T)Vb . V = worm volume (mm3) and a'(T) = temperature dependent constant. 196 T. L. FORBES AND G. R. LOPEZ change as a function of body size, and (2) smaller animals produce relatively heavier pellets (Fig. 2). This indicates that worms control egestion rate without changing pellet production rate, and therefore gut residence time, as they grow. Measure- ments made on individual Capitella species I during growth indicate that relative pellet weight is altered primarily by changes in pellet dimensions during ontogeny (Ms in prep.). Gut surface area and volume. One potentially limiting physiological surface is that of the gut. It is possible that gut surface area limits the capacity of the worm to absorb food material. Nothing is presently known of how gut surface area might change with body size in Capitella sp. I. Direct measurements of gut surface area over a range of body sizes would be of great interest in this regard. In Capitella species I, fecal pellets are formed within the first few setigers after ingestion of sediment (T. F. pers. obs.). Unless the digestion and absorption process occurs within the first few setigers, the worms must absorb food material from in- gested sediment already in the form of compact fecal pellets. If there are physico- chemical constraints (e.g., rate of diffusion of enzymes etc.) pellet size (or weight) might reach an absolute limit in larger animals. However, recent data indicate that this is not the case and that the allometric relationship between pellet weight and worm body volume is maintained up to animal volumes of at least 10 mm3 (Ms in prep.). If the relative fullness (i.e., fecal pellets/unit volume gut) of the gut of large worms is at least as great as that of small animals, then the exponential decline of relative pellet number with increasing body size (Fig. 4A) indicates that the percentage of the total worm volume occupied by the gut decreases as worms increase in size, meaning that gut volume will be related to total worm volume raised to a power less than one. Additional data on pellet standing stocks within the gut of worms between the size of metamorphosis and approximately 1 mm3 would be of great interest. Animals in this size range all have pellet standing stocks below the overall mean of 30 fecal pellets per gut. If this trend is real, and given the constant relationship between body size and pellet production rate, the average gut residence times of these smaller ani- mals will be shorter than those of larger animals. These data further suggest that the growth of very small animals (less than 1 mm3 body volume) may be limited by their ability to pack sediment into their gut and effectively absorb food. This is important because the time needed for digestion and absorption of food may impose additional constraints on the net energy intake of small worms, given their potentially shorter gut passage times. These small worms may be processing relatively greater amounts of sediment and retaining it for a shorter period within the gut. Larger animals do not show this trend, and growth in worms larger than approximately 1 mm3 may be limited by other factors, such as oxygen uptake ability. In light of these considera- tions, absorption efficiency estimates for very small Capitella sp. I will be of great interest. Due to these potential constraints, recently metamorphosed worms may be particularly vulnerable to stresses imposed by food limitation or unfavorable temper- atures. Metabolic stress may be further aggravated by the fact that oxygen solubility in sea water (35 %o S) decreases quite markedly with increasing temperature (approxi- mately 30% from 10° to 30°C, DeJours, 1981). Small deposit feeders may not be scaled-down versions of adults. Juvenile animals may face fundamentally different problems and solve them in different ways. External surface area. In order to obtain an estimate of external surface area in Capitella sp. I, we have assumed that external or respiratory exchange surface is pro- portional to worm projected area. In a study of the external surface of the aquatic oligochaete Tubifex tubifex, Kaster and Wolff (1982) found the posterior region of FEEDING ALLOMETRY IN CAPITELLA 197 the worm (last 25 segments) to be about three times as surface-rich as the same length of worm at the anterior end. Unfortunately, no attempt was made to determine the size-scaling of this relationship. However, for the present discussion of the allometric arguments, and for the estimation of the scaling of external (respiratory) surface area in Capitella sp. I, it is not necessary to assume that the entire surface of the worm is used as an exchange surface, only that the proportion used as an exchange surface does not change during ontogeny. Given these assumptions, external surface area in Capitella sp. I will scale as body volume to the 0.77 power (Fig. 3). The similarity of the allometric exponents between the feeding rate and surface area functions indi- cates that feeding rate is proportional to, and indeed may be constrained by, external surface area. Two of the most likely ways by which external surface area might influ- ence the scaling of egestion rate are through (a) limitation of transepidermal uptake of dissolved nutrients (DOM, vitamins) and/or (b) limitation of oxygen uptake for aerobic respiration. (a) Uptake of dissolved nutrients. Capitella capitata has been shown to possess active uptake mechanisms for the transport of dissolved primary amines from inter- stitial water at a rate which approaches that calculated for metabolic oxygen con- sumption (Stephens, 1975). The uptake of dissolved nutrients probably does not amount to a large component of the nutritional strategy of Capitella sp. I because the worms shrink and lose weight exponentially when maintained in seawater without sediment (Cammen, 1985; Forbes and Lopez, unpub.). However, the possibility that some essential nutrient or 'vitamin' obtainable only through transepidermal uptake in dissolved form does limit growth or net energy gain, and thereby influences the scaling of feeding rate, cannot be ruled out. (b) Uptake of dissolved oxygen. External surface area may limit oxygen uptake and therefore anabolism. A relationship of this type may have evolved under condi- tions of generally high food availability and chronically low oxygen tensions. Man- gum ( 1 976) states that "the greatest deprivation . . . that has accrued from the exploi- tation of the soft bottom was the loss of a microenvironment that remains constant in oxygen and virtually equilibrated to the atmosphere." The cellular hemoglobin of C. capitata has a very high affinity for oxygen (P50 = 3 mm Hg; Wells and Warren, 1975) and sulfide has been shown to be a larval settlement cue for Capitella species I (Cuomo, 1985). These data indicate that the genus Capitella is adapted to exploit organic-rich low oxygen habitats. However, ex- periments by Warren (1977) and studies reviewed by Pearson and Rosenberg (1978) suggest that Capitella capitata is not exceptionally tolerant of low oxygen or high sulfide concentrations. Interstitial oxygen partial pressures from the organic sedi- ments occupied by Capitella capitata are very low (8-12 mm Hg; Wells and Warren, 1975), suggesting that tube irrigation may be important in this species. Wells (1949) was able to simultaneously measure irrigation and proboscideal ac- tivity in the polychaete Arenicola marina and found an inverse relationship with a periodicity of three to four minutes. That is, animals decreased feeding activity during irrigation episodes. A similar suspension of feeding activity occurs in Capitella species I during burrow irrigation (T.F. pers. obs.). These data suggest that cessation of feed- ing during periods of low ambient O2 levels may be one mechanism by which de- creased oxygen availability could act to suppress growth or anabolism. Coelomic space. Another potential constraint affecting the feeding rate to body size relationship may be that animals must decrease the relative amount of space occupied by the gut within the coelom as they grow, and thereby decrease the relative pellet weight as they increase in size. This could be due to a demand for coelomic space not related to energy procurement or growth, such as the requirement for the 198 T. L. FORBES AND G. R. LOPEZ production and development of eggs. Because sex determination in Capitella species I is probably a digametic system, with females as the heterogametic sex, all animals are either potential hermaphrodites or true females (Petraitis, 1985). Therefore, one would expect any constraints on total gut volume or width to operate on all animals, even functioning males, because they retain the potential to become hermaphrodites in the absence of females. Feeding rate and temperature The calculated overall Qi0 value of 2.49 (Fig. 5) for the change in fecal pellet production rate with temperature from 15° to 25°C agrees well with values obtained for oxygen consumption in polychaetes (Coyer and Mangum, 1 973) over similar tem- perature ranges. In general however, the values obtained for active and resting metab- olism in polychaetes are quite variable. For example, Coyer and Mangum (1973) found Q10 values ranging from 1.81 to 4.56 over 12.5° to 27.5°C for the increase in resting metabolism ofGlycera dibranchiata. In a recent study, Cammen (1985) estimated metabolism in Capitella species I by a variety of methods (oxygen uptake, weight loss, metabolic carbon loss), over a wide range of body sizes. GM regressions were calculated for experiments performed at 10°, 20°, and 30°C. Allometric coefficients and exponents calculated for each regres- sion were also given (Cammen, 1985; Table I). It is thus possible to calculate, using data from the O2 uptake method, metabolic carbon loss for a 200 ^g individual be- tween 15° and 25°C. This calculated loss will be directly proportional to oxygen up- take. The Q,0 value calculated from Cammen's data from 15° to 25°C is 2.82. The agreement with the change of feeding rate with temperature calculated in this study (Q10 = 2.49) is thus quite close. This is somewhat surprising given the fact that in one study the temperature change was acute (this study) and in the other more gradual acclimation occurred (Cammen, 1985). Increases in metabolic rate with temperature seem to be closely paralleled by concomitant increases in feeding rate. It does not appear [based on calculations made from data in Cammen (1985), for a 200 jug DW worm] that Capitella sp. I is able to show a compensatory response in metabolic rate with increases in temperature. That is, metabolic carbon loss increases exponentially with temperature, even in worms acclimated to experimental temperatures for 1 1 to 16 days. One can also calculate from Cammen's (1985) O2 uptake data that tempera- ture has a greater effect on the metabolic carbon loss (i.e., metabolism) of smaller worms. The rate-temperature (RT) curves of metabolic carbon loss increase exponen- tially with increasing temperature for an animal of given size. More interesting, how- ever, is the fact that the rate of increase (i.e., the exponential coefficient of tempera- ture, K) is greater for smaller animals. We have plotted the calculated values of K versus body size over a wide range of body sizes in Figure 8. Capitella sp. I may thus face an energetic 'bottleneck' during ontogeny due to the fact that the effect of unfavorable temperature or food conditions may have a much greater impact on the small animals. However, under conditions of abundant food, the high metabolic rate of small animals may result in the very high growth rates observed in the laboratory (21% per day for small worms; Tenore and Chesney, 1985). The inability to adjust metabolic energy expenditure with increasing temperature, particularly evident in small worms, is indicative of an 'exploitative' metabolic strat- egy (Newell, 1979, 1980). Briefly, species adopting an exploitative strategy are adapted to an abundant food supply, and exhibit high rates of energy turnover and production. Energy conservation brought about by metabolic compensation under conditions of high food availability may not be the most efficient strategy for animals FEEDING ALLOMETRY IN CAPITELL.4 199 o 0.12 O.I I O.I 0.09 0.08 0.07 0.06 0.05 0.04 0.03 D 0 9 mg D 2.0 mg Q6.0 mg 24 ESTIMATED WORM WET WEIGHT (mg) FIGURE 8. Calculated changes in the exponential coefficient K with body size. Where MR = C [exp(KT)]; MR = ^ig carbon released day~' worm'1, T = temperature °C. C = constant. Data from Cam- men, 1985. tending to grow rapidly and reproduce quickly (Newell, 1979). Animals adapted to exploiting food-rich habitats may be unable to make adjustments in metabolic sys- tems that are designed to make the most out of environmentally favorable situations. Homeostatic changes in such metabolically important parameters as the activities of key enzyme systems, or the number of mitochondria per cell, in response to fluctuat- ing environmental conditions may not be possible for some of the so-called opportu- nistic or 'weedy' species. Shumway et al. (1983) have found that the opportunistic bivalve Mulinia lateralis maintains elevated activity even under conditions of anoxia. The clams maintain a high level of metabolism and do not seem capable of adjust- ment in the face of low oxygen levels. M. lateralis appears to be adapted to high levels of food availability, and when food level declines, characteristic mass mortalities may occur (Shumway and Newell, 1984). It may be that species such as Capitella sp. I and Mulinia lateralis, by adopting physiologies that maximize energy turnover rates at high food levels, are then incapable of adjustment in response to a decrease in food or oxygen supply. Summary Egestion rate (/ug h~') in Capitella species I scales as body size to a power less than one. This is due entirely to two observed relationships: ( 1 ) Fecal pellet production rate does not change as a function of body size, and (2) smaller worms produce rela- tively heavier pellets. We discuss several hypotheses that may explain why such rela- tionships have evolved. For example, egestion rate may be functionally related to a limiting physiological surface such as: ( 1 ) Gut surface area, or (2) external respiratory surface area. Relative pellet size ([^g (mm3worm)"']) also may be limited by the need for coelomic space for the development of the ovary. Changes in measured egestion rates (fecal pellets IT1) closely parallel those of met- 200 T. L. FORBES AND G. R. LOPEZ abolic carbon loss over similar temperatures. However, there is a greater rate of in- crease with temperature in smaller animals (data from Cammen, 1 985). This suggests the possibility of an energetic 'bottleneck' during ontogeny. Smaller animals may be affected to a greater degree (both positively and negatively) by the prevailing food and temperature regime. ACKNOWLEDGMENTS We thank K. Tenore for supplying the Capitella species I starter culture. This manuscript has benefitted greatly from the critical comments and discussion pro- vided by L. Cammen, A. Carlson, R. Cerrato, F. Dobbs, V. Forbes, S. Siddall, G. Taghon, and two anonymous reviewers. This work was supported by NSF grant OCE-8501 140 to G. R. Lopez. LITERATURE CITED CADEE, G. C. 1976. Sediment reworking by Arenicola marina on tidal flats in the Dutch Wadden Sea. Neth. J. Sea Res. 10: 440-460. CADEE, G. C. 1979. Sediment reworking by the polychaete Heteromastus filiformis on a tidal flat in the Dutch Wadden Sea. Neth. J. Sea Res. 13: 441-456. CALDER, W. A. Ill 1984. Size, Function, and Life History. Harvard University Press, Cambridge, MA. 431 pp. CAMMEN, L. M. 1980. Ingestion rate: An empirical model for aquatic deposit feeders and detritivores. Oecologia 44: 303-310. CAMMEN, L. M. 1985. Metabolic loss of organic carbon by the polychaete Capitella capitata (Fabricius) estimated from initial weight decrease during starvation, oxygen uptake, and release of I4C by uniformly-labeled animals. Mar. Ecol. Prog. Ser. 21: 163-167. COYER, P. E., AND C. P. MANGUM. 1973. Effect of temperature on active and resting metabolism in polychaetes. Pp. 173-180 in Effects of Temperature on Ectothermic Organisms: Ecological Im- plications and Mechanisms of Compensation. W. Wieser, ed. Springer- Verlag, New York. CUOMO, M. C. 1985. Sulphide as a larval settlement cue for Capitella sp. I. Biogeochemistry \: 169-181. DEJouRS, P. 198 1 . Principles of Comparative Respiratory Physiology. Elsevier/North-Holland Biomedical Press, New York. 265 pp. DOBBS, F. C. 1983. Monitoring defecation activity of infaunal deposit feeders. Mar. Ecol. Prog. Ser. 33: 225-264. DOBBS, F. C. AND T. SCHOLLY. 1986. Sediment processing and selective feeding by Pectinaria koreni (Polychaeta: Pectinariidae). Mar. Ecol. Prog. Ser. 29: 165-176. GORDON, D. C. 1966. The effects of the deposit feeding polychaete Pectinaria gouldii on the intertidal sediments of Barnstable Harbor. Limnol. Oceanogr. 11: 327-332. GRASSLE, J. F., AND J. P. GRASSLE. 1974. Opportunistic life histories and genetic systems in marine ben- thic polychaetes. / Mar. Res. 32: 253-284. GRASSLE, J. F., AND J. P. GRASSLE. 1977. Temporal adaptations in sibling species of Capitella. Pp. 177- 189 in Ecology of Marine Benthos, B. C. Coull, ed. Belle Baruch Library of Marine Science No. 6., University of South Carolina Press, Columbia. GRASSLE, J. F., AND J. P. GRASSLE. 1978. Life histories and genetic variation in marine invertebrates. Pp. 347-364 in Marine Organisms, B. Battaglia and J. Beardmore, eds. Plenum Publishing Corp., New York. GRASSLE, J. P. 1979. Polychaete sibling species. Pp. 25-32 in Aquatic Oligochaete Biology, R. O. Brink- hurst and D. G. Cook eds. Plenum Press, New York. GRASSLE, J. P., AND J. F. GRASSLE. 1976. Sibling species of the marine pollution indicator Capitella (Polychaeta). Science 192: 567-569. HARGRAVE, B. T. 1972. Prediction of egestion by the deposit-feeding amphipod Hyalella azteca. Oikos 23:116-124. HOBSON, K. D. 1967. The feeding and ecology of two North Pacific Abarenicola species (Arenicolidae, Polychaeta). Biol. Bull. 133: 343-354. HYLLEBERG, J. 1975. The effect of temperature on egestion rate in mud snails (Gastropoda: Hydrobiidae). Oecologia 21: 279-289. RASTER, J. L., AND R. J. WOLFF. 1982. A convoluted respiratory exchange surface in Tubifex tubifex (Tubificidae). Trans. Am. Microsc. Soc. 101(1): 91-95. FEEDING ALLOMETRY IN CAPITELLA 201 KUDENOV, J. D. 1982. Rates of seasonal sediment reworking inAxiothella rubrocincta (Polychaeta: Mal- danidae). Mar. Biol. 70: 181-186. MANGUM, C. P. 1976. Primitive respiratory adaptations. Pp. 191-278 in Adaptation to Environment: Es- says on the Physiology of Marine Animals, R. C. Newell ed. Butterworths, Boston. McCALL, P. L. 1977. Community patterns and adaptive strategies of the infaunal benthos of Long Island Sound. J. Mar. Res. 35: 221-266. NEWELL, R. C. 1979. Biology of Intertidal Animals. Marine Ecological Surveys LTD, United Kingdom. 781 pp. NEWELL, R. C. 1980. The maintenance of energy balance in marine invertebrates exposed to changes in environmental temperature. Pp. 561-582 in Animals and Environmental Fitness. R. Giles, ed. Pergamon Press, New York. NICHOLS, F. H. 1974. Sediment turnover by a deposit-feeding polychaete. Limnol. Oceanogr. 19: 945- 950. PAULY, D. 1981. The relationships between gill surface area and growth performance in fish: A generaliza- tion of von Bertalanffy's theory of growth. Meeresforshung 28: 25 1-282. PEARSON, T. H., AND R. ROSENBERG. 1978. Macro-benthic succession in relation to organic enrichment and pollution of the marine environment. Oceanogr. Mar. Biol. Ann. Rev. 16: 229-31 1. PETRAITIS, P. S. 1985. Females inhibit males propensity to develop into simultaneous hermaphrodites in Capitella species I (Polychaeta). Biol. Bull. 168: 395-402. RHOADS, D. C., P. C. McCALL, AND J. Y. YINGST. 1978. Disturbance and production on the estuarine seafloor. Am. Sci. 66: 577-586. RICKER, W. E. 1973. Linear regressions in fishery research. /. Fish. Res. Board Can. 30: 409-434. RICKER, W. E. 1984. Computation and uses of central trend lines. Can. J. Zool. 62: 1897-1905. SELF, R. F., AND P. JUMARS. 1 978. New resource axes for deposit feeders? /. Mar. Res. 36: 627-64 1 . STEPHENS, G. C. 1975. Uptake of naturally occurring primary amines by marine annelids. Biol. Bull. 149: 397-407. SHUMWAY, S. E., AND R. C. NEWELL. 1984. Energy resource allocation in Mulinia lateralis (Say), an opportunistic bivalve from shallow water sediments. Ophelia 23: 101-118. SHUMWAY, S. E., T. M. SCOTT, AND J. M. SHICK. 1983. The effects of anoxia and hydrogen sulphide on survival, activity and metabolic rate in the coot clam, Mulinia lateralis (Say). J. Exp. Mar. Biol. Ecol.ll: 135-146. TENORE, K. R. 1975. Detrital utilization by the polychaete Capitella capitata. J. Mar. Res. 33: 261-274. TENORE, K. R. 1977a. Growth of the polychaete Capitella capitata cultured in different levels of detritus from various sources. Limnol. Oceanogr. 22: 936-941. TENORE, K. R. 1977b. Utilization of aged detritus derived from different sources by the polychaete Capi- tella capitata. Mar. Biol. 44: 51-55. TENORE, K. R. 1981. Organic nitrogen and the caloric content of detritus: I. Utilization by the deposit- feeding polychaete Capitella capitata. Estuarine Coastal Shelf Sci. 12: 39-47. TENORE, K. R., R. B. HANSON, B. DORNSEIF, AND G. WIEDERHOLD. 1979. The effect of organic nitrogen supplement on the utilization of different sources of detritus. Limnol. Oceanogr. 24: 350-355. TENORE, K. R., AND R. B. HANSON. 1 980. Availability of detritus of different types and ages to a polychaete macroconsumer, Capitella capitata. Limnol. Oceanogr. 25: 553-558. TENORE, K. R., AND E. J. CHESNEY. 1 985. The effects of interaction of food supply and population density on the bioenergetics of the opportunistic polychaete, Capitella capitata (Type I). Limnol. Ocean- ogr. 30: 1188-1195. THISTLE, D. 1981. Natural physical disturbances and communities of marine soft bottoms. Mar. Ecol. Prog. Ser. 6: 223-228. WARREN, L. M. 1977. The ecology of Capitella capitata in British waters. J. Mar. Biol. Assoc. U.K. 57: 151-159. WELLS, G. P. 1949. The behavior ofArenicola marina L in sand, and the role of the spontaneous activity cycles. J. Mar. Biol. Assoc. U. K. 28: 465-478. WELLS, R. M. G., AND L. M. WARREN. 1975. The function of cellular haemoglobin in Capitella capitata (Fabricius) and Notomastus latericeus (Capitellidae: Polychaeta). Comp. Biochem. Physiol. 51 A: 737-740. Reference: Biol. Bull. 172: 202-21 1. (April, 1987) THE PREDOMINANCE OF CLONES IN POPULATIONS OF THE SEA ANEMONE ANTHOPLEURA ASIATIC A (UCHIDA) HARUHIKO FUJII Mukaishima Marine Biological Station, Faculty of Science, Hiroshima University, Onomichi, Hiroshima 722, Japan ABSTRACT To assess the population structure of the sea anemone Anthopleura asiatica, sex ratios and acrorhagial responses were examined among ten populations in the Seto Inland Sea of Japan. In spite of the fact that all specimens were capable of eliciting an acrorhagial response on contact with other individuals of the same species, no response was found in certain populations. These populations were always composed of individuals of the same sex. This result indicates that examination of the response was effective in distinguishing clones of the anemone. On the basis of sex and acror- hagial response, each of the populations seems to consist of a single or a few clones, and moreover, one clone appears to be distributed over 5 km through six separate locations. Diversity of clones among locations may imply sexual recruitment in the past. The anemones probably prosper asexually in local populations while dispersal to new habitats is facilitated by sexual reproduction. The causes of the clones' pre- dominance and the applicability of the Strawberry-Coral Model to the anemone are discussed. INTRODUCTION Some sea anemones can reproduce in a variety of ways both sexually and asexu- ally (Chia, 1976). It is interesting to study how these anemones use their sexual and asexual reproductive modes in their life cycles, because such studies may provide useful information which helps to answer the questions about the roles of sex (Wil- liams, 1975). Genotypic structure of a population is a consequence of reproductive modes. However, few studies have examined sea anemone population structure: An- thopleura elegantissima (Francis, 1973a; Sebens, 1982), Haliplanella luciae (Shick, 1 976; Shick and Lamb, 1 977), Metridium senile (Hoffmann, 1 976; Shick et al, 1979; Bucklin and Hedgecock, 1982) and Actinia tenebrosa (Black and Johnson, 1979; Ayre, 1983, 1984). Aggressive behavior using acrorhagi borne under tentacles has been known in sea anemones for some time (Abel, 1954; Bonnin, 1964; Francis, 1973b; Bigger, 1976, 1980; Brace and Pavey, 1978; Ottaway, 1978; Williams, 1978; Ayre, 1982). Acror- hagial responses are elicited by contact with other anemones of the same species and with other cnidarian species. On the other hand, it is known that no response occurs among clonemates, which have the same genotype. Therefore, the response between anemones must be useful to recognize clones and to assess the structure of popula- tions (Sebens, 1982), although the usefulness of this distinguishing method may be limited (Bigger, 1980; Ayre, 1982). Anthopleura asiatica, a common intertidal sea anemone in western Japan, repro- Received 25 September 1986; accepted 6 January 1987. Contribution no. 278 from the Mukaishima Marine Biological Station. 202 CLONES IN A. ASIATICA POPULATIONS 203 duces asexually by longitudinal fission, as well as sexually (Uchida, 1958, 1965). On the rocky shore of Matsugahana Pt. near Mukaishima Marine Biological Station, Japan, the anemones live in tidepools with their pedal discs in contact with each other (Fujii, 1985). When I examined anemone gonads, in 1983, to assess the reproductive cycle of anemones inhabiting the rocky shore, the noticeably mature specimens were always males (Fujii, in prep.). The present study examines the sex ratio and acrorhagial responses of A. asiatica specimens at ten locations in the central Seto Inland Sea of Japan (Table I, Fig. 1), in order to assess the population structure of the anemone. The effectiveness of the acrorhagial response in distinguishing clones, the population structure, and the appli- cability of the Strawberry-Coral Model (Williams, 1975) to the species, are discussed on the basis of these results. MATERIALS AND METHODS Between August and September 1983, collections ofAnthopleura asiatica individ- uals were made from ten locations which differed in size and habitat type: Kamiebu- jima Islet, Matsugahana Pt., Kannonzaki Pt., Sasajima Islet, Shijushima Islet, Iwashi- jima Is., Ategijima Islet, Toyoshima Is., Takaikamishima Is., and Aoshima Is. (Fig. 1, Table I). Although I also visited other locations on these and other islands, A. asiatica was only found at these ten locations. At three of them, Kannonzaki Pt., Toyoshima Is., and Aoshima Is., the anemones were found only in one or a few tidepools, while at the others, they were distributed over more than ten meters of shoreline. The anemones often form aggregations of contiguous individuals. In each location, the specimens were collected randomly from several aggregations which had been chosen throughout the habitats. The specimens from each aggregation were transported separately to the laboratory in labeled plastic bags and kept in seawater in labeled plastic cups (8 cm in basal diameter, 5 cm in depth) until they were exam- ined. The seawater was changed daily. Acrorhagial responses were examined using the specimens of the three aggrega- tions which had been furthest away from each other in each location. Each anemone was allowed to attach to a piece of oyster shell in the cup, so that it could be moved easily. After two days, firmly attached anemones were paired with the specimens from the other two aggregations at the same location. A 1 -second touch technique (Bigger, 1980) was applied to test the response. The pair's tentacles were lightly touched to each other for one second every fifteen seconds. Successive contacts were made ten times between the same tentacle tips, or until an acrorhagial response occurred. Over three repetitions were done in every three combinations. If no response was elicited in a pair, the pair was placed with their tentacles in contact and observed for more than one hour. The acrorhagial response also was examined between anemones from different locations, using specimens from the three aggregations. Twenty-nine combinations were tested with at least five repetitions of each, using different pairs (Fig. 3). Collected anemones, including those used for the response test, were all anaesthe- tized in the cups with a 50:50 solution of seawater and 7.5% (w/v) magnesium chlo- ride. Well-relaxed anemones were cut longitudinally and their gonads were examined under a stereomicroscope. In November 1983, another collection of anemones was made from five of the ten locations, Kamiebujima Islet, Matsugahana Pt., Sasajima Islet, Shijushima Islet, and Ategijima Islet, in order to examine the acrorhagial responses again. Specimens were collected from 6, 20, 13, 5, and 9 aggregations, respectively (five specimens at 204 H. FUJII SETO INLAND SEA FIGURE 1 . The map of the ten locations. each aggregation). The responses were examined with every possible combination of the aggregations in each location and those of the location pairing their representa- tives. If no response occurred between their representatives, the rest of the specimens from different locations were set in contact with each other in a vessel and observed to confirm the results. RESULTS The sex of 408 specimens of A. asiatica was determined from 62 aggregations and they were found to be dioecious: 76 females, 28 1 males, 5 1 immature (Table II). Ripe individuals were easily sexed by the color and form of the gonads. Females' gonads CLONES IN A. ASIATIC A POPULATIONS 205 TABLE 1 Description of locations Locations Ranges (area or length) Habitat types ( 1 ) Kamiebujima Islet 20m t = P = r (2) Matsugahana Pt. 2000 m2 >P (3) Kannonzaki Pt. 0.35m2 t (4) Sasajima Islet 50m r^>p (5) Shijushima Islet 12m P>t (6) Iwashijima Is. * 22m t (7) Ategijima Islet 40m p ^* r — t (8) Toyoshima Is. 9m2 t (9) Takaikamishima Is. * 250 m^ p> r = t (10) Aoshima Is. 6 m" t Relative abundance of inhabited habitants are represented by inequality signs, t: tidepool, p: pebble, r: rock. *After this study, the anemones were found at the next shores outside the reported ranges. were filled with light blue eggs, and the white vesicles in the males were filled with sperm. The sex ratios in most locations were extremely skewd (Table II). Only a single sex was found in eight of the ten locations. All specimens from six locations, Kamie- bujima Islet, Mastugahana Is., Kannonzaki Pt., Sasajima Islet, Shijushims Islet, and Ategijima Islet, were males, while all specimens from two locations, Toyoshima Is. and Aoshima Is., were females. Although both sexes were found in the other two locations, Iwashijima Is. and Takaikamishima Is., even there, the specimens from an aggregation were always of one sex (Table II). No acrorhagial responses were elicited in some combinations of the specimens; their tentacles were interlaced for over one hour without any response. But all anemo- nes used in this study apparently had the potential to display the acrorhagial response because they always elicited the responses after several artificial tentacle contacts (from one to four times) in the other combinations. The process of the response in A. asiatica was similar to those in the other acrorhagi-bearing anemones (Bonnin, 1 964; Francis, 1 973b; Bigger, 1 980): ( 1 ) tentacle withdrawing after contact; (2) expan- sion of acrorhagi; (3) leaning to the opponent; (4) application of acrorhagial ecto- derm; and (5) recovery. In each of the eight locations where all specimens were the same sex, no acror- hagial response was found in any pair of the three aggregations (Fig. 2). Additional examinations of acrorhagial response were carried out in five of the eight locations, combining more aggregations, but no response occurred in any of them (Table III). In fact, the responses were always found in two of the three combinations in the two locations where there were both males and females (Fig. 2), but even there, the specimens of the two aggregations which did not respond to each other were always the same sex. As for acrorhagial responses between anemones from different locations, twenty- nine combinations were examined (Fig. 3). In this test the responses between males (Ategijima Islet vs. other male locations) and those between females (Aoshima Is. vs. Iwashijima Is.) were also observed. The acrorhagial response between anemones seems to be independent of their sex. The responses were elicited in every combina- tion of the locations except those of the locations along Mukaishima Is., where all 206 H. FUJII TABLE II Distributions of sexes in the ten locations Locations Aggregations i j k 1 m n Total (1) Kamiebujima M 5 5 3 5 5 10 5 5 M 43 F 0 0 0 0 0 0 0 0 F 0 I 0 0 2 0 0 0 2 0 I 4 (2) Matsugahana M 6 5 4 5 5 5 5 5 55545 25 9M 98 F 0 0 0 0 0 0 0 000000 0 OF 0 I 0 0 1 0 0 0 0 000010 8 27 I 37 (3) Kannonzaki M 4 M 4 F 0 F 0 I 2 I 2 (4) Sasajima M 5 5 5 5 5 M 25 F 0 0 0 0 0 F 0 I 0 0 0 0 0 I 0 (5) Shijushima M 7 7 7 7 7 9 M 44 F 0 0 0 0 0 0 F 0 I 0 0 0 0 2 1 I 3 (6) Iwashijima M 0 4 5 3 M 12 F 8 0 0 0 F 8 I 0 0 0 0 I 0 (7) Ategijima M 8 11 6 5 5 M 30 F 0 0 0 0 0 F 0 I 0 0 0 0 0 I 0 (8) Toyoshima M 0 0 0 0 0 M 0 F 4 5 5 5 2 F 21 I 1 0 0 0 3 I 4 (9) Takaikamishima M 5 5 5 0 0 0 0 5 5 M 25 F 0 0 0 5 5 5 5 00 F 20 I 0 0 0 0 0 0 0 0 0 I 0 (10) Aoshima M 0 0 0 0 M 0 F 5 5 5 12 F 27 I 0 0 1 0 I 1 M: male, F: female, I: immature. specimens were males. In order to make sure of the results, particularly the exception, additional examinations were done combining five locations; four of them along Mu- kaishima Is. No response was also found among the four locations while the responses were always elicited in combinations between the fifth location, Ategijima Islet, and one of the other four (Table III). DISCUSSION Sea anemones bearing acrorhagi display aggressive behavior on contact with the same species or with other cnidarian species (Bigger, 1980). This study shows that A. asiatica responds aggressively to other individuals of the species with the same pattern of behavior as other acrorhagus-bearing anemones (Abel, 1954; Bonnin, 1964; Fran- cis, 1973b; Bigger, 1976, 1980; Ottaway, 1978; Williams, 1978; Ayre, 1982). Acrorhagial response seems to be useful to the assessment of the genetical struc- CLONES IN A. ASIATICA POPULATIONS 207 SETO INLAND SEA FIGURE 2. Acrorhagial responses among three aggregations in each location. R: the responses oc- curred in two of three combinations of aggregations. NR: no response was elicited. Signs of sexes represent the sex compositions of the locations. ture of anemone populations because anemones are able to distinguish between clonemates and non-clonemates (Francis, 1973a, b, 1976; Bigger, 1976, 1980; Ayre, 1982), although their distinguishing ability may be limited. Ayre (1982) failed to pro- duce aggressive responses on contact between some genetically dissimilar individuals of Actinia tenebrosa. When we use the acrorhagial response to distinguish clones, it is important to know to what extent we can trust the response. A clone of dioecious animals is composed of a single sex. Therefore distribution of the sexes can be a useful indication of asexual reproduction, especially in sessile animals such as anemones (Ford, 1964; Sebens, 1982). The exceedingly skewed sex ratios in the locations (Table II) suggest that the anemones reproduce asexually there. 208 H. FUJII TABLE III Results of the second acrorhagial response test within and between five locations (1) (2) (4) (5) (7)_ (1) Kamiebujima Is. + C(6.2) (2) Matsugahana Pt. + C(18.2) (4) Sasajima Islet + C(13.2) (5) Shijushima Islet + C(5.2) (7) Ategijima Islet C(9.2) — : no responses were elicited in the combination, +: the responses were always elicited. The number of tested combinations within locations were represented by C(N.2) = (N!/(N - 2)!)/2, where N is the number of collected aggregations. As for the test between locations, at least five repetitions were done. All specimens of ( 1 )-(5) were mingled in a vessel without any response. The acrorhagial responses of anemones are thought to be independent of their sexes (Ayre, 1982; Sebens, 1984). In this study, A. asiatica responded aggressively to both male and female specimens from different clones. Therefore, assuming the test of the response was ineffective in distinguishing clones, one would expect to find individuals of both sexes among the anemones that did not respond to each other. However, anemones that did not respond were always of the same sex (Fig. 2, 3). This indicates that A. asiatica 's ability to distinguish clones is very good. Asexual reproduction has a great influence on genetical structure of a population. Sea anemones which reproduce asexually have less genotypic diversity in a popula- tion than would be expected in a sexual population with free recombination and random matings (Shick and Lamb, 1977; Black and Johnson, 1979; Ayre, 1983; Hoffmann, 1986; Shick and Dowse, 1985). The results of this study indicate that the populations of A. asiatica have extremely low diversities. Eight of the ten locations seem to consist of single clones, although some minor clones may have been missed. Even the other two locations seem to consist of a few dominant clones. Moreover, this study suggests that one of the clones may dominate at six locations, covering a range of over five kilometers (Fig. 3). This might be explained by a tidal current passing through the locations, but more evidence will be needed to substantiate this idea. It has been reported that the sea anemones, A. elegantissima and Metridium senile, form clonal aggregations (Ford, 1964; Francis, 1973; Hoffmann, 1976; Sebens, 1982) and a clone of them can be distributed over a relatively large area in their populations (Shick et al, 1979; Sebens, 1 982). It has also been shown that the popula- tions of the sea anemone H. luciae, introduced on the Atlantic coast of North Amer- ica, are composed of one or a few clones (Shick, 1976; Shick and Lamb, 1977; Shick et al., 1979). However, the formation of discrete colonies by a single clone had not been reported until now. As Shick et al. (1979) suggested for H. luciae, single clone predominances in A. asiatica may be caused by one or more of the following: ( 1 ) the clone best adapted to the habitats was selected; (2) there were few immigrants and new settlements so that the few lucky anemones arriving on unoccupied habitats could prosper and predomi- nate there, or (3) each clone at a location had excluded others aggressively. Ayre (1985) revealed through a field transplantation experiment that populations of A. CLONES IN A. ASIATIC A POPULATIONS 209 FIGURE 3. Acrorhagial responses among locations. No response occurred between the locations con- nected by broken lines. The responses occurred between the locations connected by solid lines. tenebrosa consist of highly locally adapted clones. The anemones showed higher asex- ual fecundity when they were transplanted into their original habitats. In the center of the Inland Sea, populations ofAnthopleura asiatica seem to consist of one or a few clones. Even if the clones are the best adapted to their locations, they probably do not have narrow but rather wide adaptational ranges, because each location consists of a variety of habitat types (e.g., tidepool, rock, and boulder at different tidal heights). The clones of the colonizing anemone, H. luciae, are adapted to a wide range of temperature and salinity (Shick, 1976; Shick and Lamb, 1977; Shick and Dowse, 1985). Shick et al. ( 1 979) revealed that in H. luciae, the preadaptations of some clones to low temperature allowed their predominance. Although such a severe condition could not be found out in the Seto Inland Sea, there may be a different kind of limiting 210 H. FUJII condition which functions as a bottleneck. A variety of clones would be expected in an unstable habitat (Sebens, 1982; Ayre, 1984). Because the predominance of a single or a few clones occurs even on the boulder beaches, which seem to be unstable, it also seems plausible that the predominant clones started as a few lucky founders which encountered unoccupied habitats. A low rate of new recruitment may allow the clones to prosper in diverse and unstable habitats. Aggressive interactions of anem- one are thought to play an important role in intra-specific competition for space (Francis, 1973a, b Ayre, 1983). But it is still uncertain whether one clone can exclude others completely from its habitats through aggressive interactions. In fact. Bigger (1980) suggested that acrorhagial response can be viewed as one of a set of ecological facters which may maintain a genetic heterogeneity. In any case, they may serve to repel the other invading clones once a clone occupies a habitat (Sebens, 1982). The present study suggests not only asexual proliferation of A. asiatica in the local populations but also their sexual settlement in the past. On the basis of the aggressive response, each location seems to be occupied by different clones, with the exception of the locations around Mukaishima Is. (Fig. 3). This diversity was probably brought about through sexual reproduction. A. asiatica seems to use sexual reproduction for wide dispersal while using asexual reproduction in local proliferation. The life-history pattern of A. asiatica appears to fit the predictions of the Straw- berry-Coral Model which Williams (1975) proposed to overcome the theoretical dis- advantage of sex. The model assumes that clonal proliferation of sessile organisms is limited by their habitat space, and that there is an intense selection, so that only the progeny with sufficiently high local fitness can win in each habitat. In the model, diverse sexual progeny are predicted to have a greater probability of establishing new colonies outside the limits of original habitats, because of the different demands of new habitats. In spite of the conformity in the life history pattern, some situations of A. asiatica may be inconsistent with the assumptions of the model. For the start, a clone of the species seems to inhabit a variety of habitat types. From this situation, it is likely that the clones were not the fittest but just the lucky ones (intense competition may be rare) and/or that each clone has a wide adaptational range (i.e., variability may be less important). Thus the model may not be applicable in each case. Aggressiveness may be another problem (Ayre, 1983), because it does not seem to be dependent on environmental conditions. If the most aggressive clone was selected in every habitat, the introduction of variation through sexual reproduction might be disadvantageous. ACKNOWLEDGMENTS I am grateful to Dr. J. M. Shick, Dr. T. Hoshino, and Dr. Y. Hirano for their useful advise on the manuscript. I also thank Mr. M. Schauman and Miss S. Ingle for revising the manuscript, and Mr. T. Abo for his assistance in collecting specimens. LITERATURE CITED ABEL, E. F. 1954. Ein Bietrag zur Guftwirkung der Aktinien und Funktion der Randsachchen. Zoo/. Anz. 153:259-268. AYRE, D. J. 1982. Inter-genotype aggression in the solitary sea anemone Actinia tenebrosa. Mar. Biol. 68: 199-205. AYRE, D. J. 1983. The effects of asexual reproduction and inter-genotypic aggression on the genotypic structure of populations of the sea anemone Actinia tenebrosa. Oecologia 57: 1 58- 165. AYRE, D. J. 1984. The effect of sexual and asexual reproduction on geographic variation in the sea anem- one Actinia tenebrosa. Oecologia 62: 222-229. CLONES IN A. ASIATICA POPULATIONS 21 1 AYRE, D. J. 1985. Localized adaptation of clones of the sea anemone Actinia tenebrosa. Evolution 39: 1250-1260. BIGGER, C. H. 1976. The ecological response in Anthopleura krebsi: intraspecific and interspecific recogni- tion. Pp. 127-136 in Coelenterate Ecologv and Behavior, G. O. Mackie, ed., Plenum Press, New York. BIGGER, C. H. 1980. Intraspecific and interspecific acrorhagial aggressive behavior among sea anemones: a recognition of self and nonself. Biol. Bull. 159: 1 17-134. BLACK, R., AND M. S. JOHNSON. 1979. Asexual viviparity and population genetics of Actinia tenebrosa. Mar. Biol. 53: 27-3 1. BONNIN, J. R. 1964. Recherches sur la "reaction d'aggression" et sur le fonctionnement des acrorrhages & Actinia equina L. Bull. Biol. 98: 225-250. BRACE, R. C., AND J. PAVEY. 1978. Size-dependent dominance hierarchy in the anemone Actinia equina. Nature 273: 752-7 '53. BUCKLIN, A., AND D. HEDGECOCK. 1982. Biochemical genetic evidence for a third species ofMetridium (Coelenterata: Actinaria). Mar. Biol. 66: 1-7. CHIA, F. S. 1976. Sea anemone reproduction: patterns and adaptive radiations. Pp. 261-270 in Coelenter- ate Ecology and Behavior, G. O. Mackie, ed., Plenum Press, New York. FORD, C. E. 1964. Reproduction in the aggregating sea anemone Anthopleura elegant issima. Pac. Sci. 18: 138-145. FRANCIS, L. 1973a. Clone specific segregation in the sea anemone Anthopleura elegantissima. Biol. Bull. 144: 64-72. FRANCIS, L. 1973b. Intraspecific aggression and its effect on the distribution of Anthopleura elegantissima and some related sea anemones. Biol. Bull. 144: 73-92. FRANCIS, L. 1976. Social organization within clones of the sea anemone Anthopleura elegantissima. Biol. Bull. 150:361-376. FUJII, H. 1985. Locomotion and distribution of sea anemone, Anthopleura oriaf /oz (Uchida), in microhab- itat. Special Publication of the Mukaishima Marine Biological Station 1 67- 1 72. HOFFMANN, R. J. 1976. Genetics and asexual reproduction of the sea anemone Metridium senile. Biol. Bull. 151:478-488. HOFFMANN, R. J. 1986. Variation in contributions of asexual reproduction to the genetic structure of populations of the sea anemone Metridium senile. Evolution 40: 357-365. OTTAWAY, J. R. 1978. Population ecology of the intertidal anemone Actinia tenebrosa I. Pedal locomotion and intraspecific aggression. Aust. J. Mar. Freshwater Res. 29: 787-802. SEBENS, K. P. 1982. Asexual reproduction in Anthopleura elegantissima (Brandt) (Anthozoa: Actinaria): seasonality and spatial extent of clones. Ecology 63: 434-444. SEBENS, K. P. 1984. Agonistic behavior in the intertidal sea anemone Anthopleura xanthogramica. Biol. Bull. 166: 457-472. SHICK, J. M. 1976. Ecological physiology and genetics of the colonizing actinian Haliplanella luciae. Pp. 137-146 in Coelenterate Ecology and Behavior, G. O. Mackie, ed., Plenum Press, New York. SHICK, J. M., AND A. N. LAMB. 1977. Asexual reproduction and genetic population structure in the colo- nizing sea anemone Haliplanella luciae. Biol. Bull. 153: 604-6 17. SHICK, J. M., R. J. HOFFMANN, AND A. N. LAMB. 1979. Asexual reproduction, population structure and genotype-environment interactions in sea anemones. Am. Zool. 19: 699-713. SHICK, J. M., AND H. B. DOWSE. 1985. Genetic basis of physiological variation in natural populations of sea anemones: intra- and inter-clonal analyses of variance. Pp. 465-479 in Proceedings of the Nineteenth European Marine Biology Symposium, P. E. Gibbs, ed., Cambridge University Press, U.K. UCHIDA, T. 1958. Notes on some Japanese sea anemones. J. Fac. Sci. Hokkaido Univ., Ser. VI, Zool. 14: 111-119. UCHIDA, T. 1965. New Illustrated Encyclopedia of the Fauna of Japan. Hokuryu-kan, Tokyo. WILLIAMS, G. C. 1975. Sex and Evolution. Princeton University Press, Princeton. WILLIAMS, R. B. 1978. Some recent observations on the acrorhagi of sea anemones. J. Mar. Biol. Assoc. U. K. 58: 787-788. Reference: Biol. Bull. 172: 212-224. (April, 1987) APPLICATION OF AN IMPROVED METHOD OF TWO-DIMENSIONAL ELECTROPHORESIS TO THE SYSTEMATIC STUDY OF HORSESHOE CRABS JUN-ICHI MIYAZAKI*. KOICHI SEKIGUCHI**, AND TAMIO HIRABAYASHI* ' institute of Biological Sciences, The University ofTsukuba, Niihari-gun Sakura-mura. Ibaraki 305, Japan; and ** Business Department, Jobu University, Isesaki-shi, Gunma 372, Japan ABSTRACT Protein constituents of skeletal and cardiac muscles of extant horseshoe crabs, Tachypleus tridentatus, T. gigas, Carcinoscorpius rotundicanda, and Limulus poly- phemus, were examined by two-dimensional gel electrophoresis in order to investi- gate the applicability of this technique in systematics and to obtain more data on phylogenetic relationships of the four species. Electrophoretic patterns were analyzed by both comparison of spot positions only and that of presumed protein groups (posi- tions and shapes). Our phenetic analyses confirmed the disparate position of the North American species, L. polyphemus, and the similarity of the three Asian species. Our cladistic analysis suggested a sister-group relationship of T. tridentatus and C. rotundicauda. This suggestion, however, was based on a small number of characters. On the basis of this electrophoretic study, we suggest that conventional native (non-denaturing) gel electrophoresis for enzymes is optimal for measuring genetic differences among organisms at lower taxonomic levels, but two-dimensional gel electrophoresis is best for overall protein comparison. INTRODUCTION There are four currently recognized extant species of horseshoe crabs, Tachypleus tridentatus, T. gigas, Carcinoscorpius rotundicauda, and Limulus polyphemus. T. hoeveni, described by Pocock (1902), was considered by Waterman (1958) to be a synonym of T. gigas. Previous phylogenetic studies on those four species have utilized external mor- phology, hybridization tests, and amino acid sequencing (Pocock, 1902; Sekiguchi and Sugita, 1980; Shishikura et al, 1982; Srimal et ai, 1985). These studies have shown that Limulus polyphemus, which inhabits the east coast of North America, is disparate from the three southeast Asian species. The classical or morphological view places two of these Asian species in the genus Tachypleus and considers Carcinoscorpius to be a monotypic genus (Pocock, 1902; Sttfrmer, 1952). Contradictory biochemical, embryological, and genetical data, how- ever, suggest that T. tridentatus is more similar to C. rotundicauda than to T. gigas (Sekiguchi and Sugita, 1980; Shishikura # al, 1982; Srimal et al, 1985). Since previous biochemical studies on horseshoe crabs have dealt only with two proteins, coagulogen and hemocyanin (Shishikura and Sekiguchi, 1978; Sugita and Received 14 July 1986; accepted 28 January 1987. 1 To whom all correspondence should be addressed. 212 SYSTEMATIC STUDY OF HORSESHOE CRABS 2 1 3 Sekiguchi, 1981; Shishikura et al, 1982; Srimal et al, 1985), the number of protein characters examined has been too small for reasonably deducing phylogenetic rela- tionships. Because two-dimensional gel electrophoresis can detect more subtle pro- tein differences than conventional electrophoresis, it can provide information on many kinds of proteins. However, the usefulness of the data for phylogenetic analyses has not yet been examined (Wiley, 198 1). In this study, proteins from cardiac and skeletal muscles of the four species of horseshoe crabs were separated by two-dimensional gel electrophoresis ( 1 ) to exam- ine its applicability to phylogenetic analysis, (2) to obtain new data on phylogenetic relationships among the three Asian species, and (3) to re-examine the disparate taxo- nomic position of the North American species. We also compared the applicability of the two-dimensional gel electrophoresis method with that of the conventional method which utilizes a native (non-denaturing) gel. Previous studies have suggested that the two-dimensional method gives a smaller magnitude of both intraspecific and interspecific differences than does the latter method (Brown and Langley, 1979; Mc- Conkey et al, 1979; Walton et al, 1979; Racine and Langley, 1980; Aquadro and Avise, 1981;McConkey, 1982). MATERIALS AND METHODS Tachypleus tridentatus was collected in Imari, Japan. Tachypleus gigas and Carci- noscorpius rotundicauda were obtained from Bangsaen, Thailand. Limulus polyphe- mus was purchased from the Marine Biological Laboratory, Woods Hole, Massachu- setts, USA. The tubular heart (cardiac muscle) and a pair of skeletal muscles that connect the prosoma with gills were dissected from each specimen and cleaned quickly in filtered seawater to prevent contamination from other tissues and hemolymph. The heart was cut into anterior, central, and posterior parts for comparison of protein compo- nents in respective sections. The heart parts and skeletal muscles were cut into small pieces and stored at — 80°C. Two-dimensional electrophoresis was carried out essentially according to the method improved by Hirabayashi (1981). One piece (0.1 g) of each muscle was ho- mogenized with a Dounce homogenizer in 4 ml of an extraction medium of 8 M urea, 1 AfKI, 0.1 mMNa-tosyl-L-lysylchloromethanehydrochloride, 1 0% 0-mercap- toethanol, and 0.1 MTris-HCl, pH 7.5. This homogenate was dialyzed against 7 M urea for 3-4 hours and centrifuged at 30,000 X g for 30 minutes with a Beckman J A- 14 rotor. For single species runs, 60 n\ and 40 /ul of the supernatants from cardiac and skeletal muscles, respectively, were loaded on the first dimension isoelectric focusing gels. When samples from two species were compared, a 1 : 1 mixture of 80 ^1 (40 n\ from each heart) or 50 M! (25 ^1 from each skeletal muscle) was also loaded. Isoelectric focusing was carried out at 500 V for 24 hours with cardiac muscle and 30 hours with skeletal muscle. The remaining supernatants and precipitates were each mixed with a hot 1% agarose solution containing 2% SDS, 10% glycerol, 5% /3-mercaptoethanol, 0.02% bromophenol blue, and 0.05 M Tris-HCl, pH 6.8. Each mixture was chilled in a glass tube and stored as reference for total supernatant or precipitate components of the second dimension SDS-polyacrylamide gel electrophoresis. The second dimen- sion, with a concentration gradient of acrylamide (12-20%), was carried out at 30 mA, when the bromophenol blue tracking dye was in the stacking gel, and then con- tinued at 60 mA until the dye reached the lower end of the running gel. For optimal comparison of two-dimensional patterns, we electrophoresed ex- 214 J.-I. MIYAZAKI ET AL. tracts from each of two species and a mixture of the two at the same time ("triplet method," Fig. 1). A close examination of the mixture pattern (Fig. Ib) was indispens- able to confirm overlapping (or sharing) of spots suspected to be similar from individ- ual patterns (Fig. la, c). Preliminary examination of cardiac muscle with the "triplet method" did not show significant differences among the three male and one female specimens of T. trident at us and among the anterior, central, and posterior parts of a single heart of the animal. However, we used only the anterior part, since the sample was large enough. For interspecific analyses, we used six "triplet" combinations to compare patterns among the four species. In all, about 100 obvious spots were exam- ined on each pattern of cardiac and skeletal muscles for each of the four species. We only used those spots that could be consistently matched or clearly defined. The two-dimensional electrophoresis method, applied herein, has two advantages over those used in previous studies. First, the substitution of agarose for acrylamide in the first dimension gel allowed us to analyze proteins of higher molecular weights (up to about 250 KD). Second, we used four kinds of ampholytes (rather than one or two) in the first dimension for wider pH coverage (pH 3.5-9.5). These changes, plus a concentration gradient of acrylamide in the second dimension, permitted us to analyze about 100 proteins, presenting many more characters for systematic study than other methods of electrophoresis. RESULTS Typical two-dimensional electrophoresis patterns of cardiac muscle of the four species are shown in Figures 1 and 2. Ninety-seven of about 100 major spots for each of the four species were "shared" (that is, having overlapping positions) by at least two species. These spots were numbered in the order of their relative molecular weights (Fig. 3). Those shown in boxes below Nos. 5 and 6 (Fig. 2) may have been derived from hemolymph proteins, probably hemocyanin. Hemolymph could not be removed completely, even though the muscle was thoroughly washed in seawater. The sharing (overlapping) of cardiac muscle spots by the four species is shown in Table I, where the numbers correspond to those in Figure 3. The numbers of spots shared by any two of the three Asian species were almost the same and roughly twice those shared by each Asian species and the North American one (Table II). When only the spot position was considered, we regarded two similar-appearing spots separately even if both were presumed to be derived from the same kind of proteins by their similar morphology and positions relative to other spots. Nos. 20 and 2 1 , for example, were counted separately for analysis of the position only (Table I), although they may have been counterparts. Next we considered such spots (e.g., Nos. 20 and 21) as belonging to the same protein group and counted all examined spots as representing protein groups (Table III). In some instances, no spot was shared for the presumed protein group, e.g., Nos. 98-108 in Table III. This resulted in a difference of five shared spots for cardiac mus- cle between the two approaches (see Tables I and III). Tabulation of shared cardiac muscle protein spots (Table IV) gave a similar result to that when only the spot position was considered (Table II), the usual method of analyzing two-dimensional electrophoretic patterns. Therefore, both approaches sup- port the view that the three Asian species are more similar to one another than to the North American species. Cardiac muscle was used in this study both because of its importance to life and the homogeneity of its cell population. For similar reasons, we also examined skeletal SYSTEMATIC STUDY OF HORSESHOE CRABS 215 1st. dim. 2nd. dim. f « - FIGURE 1 . The "triplet patterns" of cardiac muscle. For optimal comparison of two-dimensional patterns, we electrophoresed extracts from each species (a, c) and a mixture of the two (b). a, Tachyplens tridentatus: c, Limuluspolyphemus. Two lanes at the right-hand end of each pattern are references for total supernatant (left) and precipitate (right) components. 216 J.-I. MIYAZAKJ ET AL. Cardiac muscle b FIGURE 2. The typical two-dimensional electrophoresis patterns of cardiac muscle, a, Tachypleus gigas; b, Carcinoscorpius rotundicauda. The spots in boxes were probably derived from hemolymph pro- teins. muscle that connects the prosoma with the gills. With skeletal muscle, 97 of about 1 00 major spots on two-dimensional electrophoresis patterns (Fig. 4) were recognized as shared (i.e., having overlapping positions) between at least two species, and shared spots by each pair of species were counted as before (see also Table II). Re-examina- tion for shared spots in protein groups (i.e., similar spot morphology and positions) gave a similar result (Table IV), although it differed by 22 spots. Skeletal muscle pat- terns also confirm the greater similarity among the Asian species when compared to the North American species. Using the results in Table IV, we quantified the similarity among the four species PH9 200K 2nd. dim. 40K- - 2~ J38"27 2.526 A)S o^ • 444342 4J 38 34--. ^ 484950 CO "•' "-* ^^V V w ^ i 'j^ v^o 646362 ^ „ «55 5439 6O61 so 35 ^W - - 51^253 57 ^55 543952 ,fc_. ^ 75 74 70 92 go 95 96 66 71 67 89 87 w 8381 V • 9091 94 97 FIGURE 3. The schematic drawing of 97 spots of cardiac muscle. These spots, numbered in the order of their relative molecular weights, were shared by at least two of the four species. SYSTEMATIC STUDY OF HORSESHOE CRABS 217 ua _j !E r) ri r i O NO ON NO ri O Z C/3 4- + + + 4- + ~ OS CO ON r i oo r I "i s: c a s - •^ G ~C &o c ^ t-; ri + O IB •^ n. 8 + + + + s + + r- + + + O o o <] t> n i- CN CN + + O ON + + + o r- + + + o tfi o o < > n 5§ I o (N OO + + + r- + + o CN O o n 5l + * + * fl> C a >, O CN r^ + + ? + + + o o o < > oj u •5 "^ *o c ON o + + < ^ + + + o P o + + < 0 o < > n E *"* ' ' 3 « t^ ^ OO + + + + in + + + CN o + + < ON o < D> n 5| r- + + + + ^ + + r- + + + + OO ON o < D> a " o C '^ o c £ o + + < 3 + + + + O r- + + + r- ON + o < + $3 ~ 2 + o + < CN + + + + S o + + < SO ON + + a2 "> C3 ^ tu ° "o ° CO - + + o < 5 + + + o 00 o + + < m ON + + + H ™ + + O + + + £ + o < + ON + + + o ^ s: <*. u -g ° CN + o + m + + + o « + + + + ON + + + + "5 "2 c ^••s ^ __ oo m + + + + vC o + + + s + + o O ON + + 0 3 « a g .y Jjj c/^ r« ^^ 5 00 + + £' + + + o CN + + + o ON OO + o + < r« cs.E ^ "o c -* -| r~- + + + + 2, + + + o 3 + + + 00 oo + + + o 8]§l ^ Oj O ~~ --J vO + + + o r? + o < + o + + + o oo + + S a< s£ o. J2 S £, •n + + + + CN + + + + ON + + + o SO 00 + o + + r-, -C !)_, I <* o + + < M + + + o OO + + + + in 00 + O + 4 " 0 & c ^ -i O .~ Z Z Z Z 5 1 g i s ^ •% Z "is "o "cs o ^ .0 O. **-> Cfl i ^ 1 u 2 5 S § Q -is — 1 11 5? o 1 2 'C a CO l.lll i: 5: C S, 'o OJ a wa 11:11 '5 OJ a VI c "S "5 ^ .1 1 is1 'o u o. oo -Si S S — ? -2f o "o ~ «>C ^ ^ 100 nuclei arranged in concentric circles near the periphery of the cell (Fig. 7). 238 R. S. ANDERSON -. i »? V • * :V p 4T • \ * - FIGURE 1 . Typical appearance of hemocyte monolayer 1 h after adhesion to the glass substrate. FIGURE 2. A hemocyte containing numerous irregular yellow granules (arrow). FIGURE 3. Hemocyte with large (2X5 p.m) clumped yellow-brown inclusions (arrow). FIGURE 4. Early inflammatory foreign-body type cell after 3 h culture, arrow indicates one of five closely packed nuclei. FIGURE 5. Typical FBC after 5 h in culture containing 18 nuclei; one of many yellow-brown gran- ules, similar to those in Figures 2 and 3, is indicated by the arrow. FIGURE 6. Langhans type cells in a 6 h hemocyte culture. Scale bar = 20 /^m for Figures 1-6. MULTINUCLEATED GIANT CELLS IN BIVALVES 239 The FBC also continued to enlarge during 24-48 h in culture. In these prepara- tions, FBC are often connected by numerous cytoplasmic strands (Fig. 8). The more mature FBC frequently contained foreign material such as amorphous masses (Fig. 8), yellow-brown particles (Fig. 9) or cell-like structures (Fig. 10). The cell-like inclu- sions were typically surrounded by a conspicuous vacuole; however, the yellow gran- ules and amorphous material were not usually enclosed by vacuoles. Multinucleated giant cells of either kind were rarely seen in >48 h hemocyte monolayer cultures; possibly they lysed or became less adherent, and were lost during medium changes. Giant cell formation was a highly variable phenomenon in Mercenaria. Groups of 10-24 clams were examined at a given time; usually 50-60% of the individual organisms per batch would yield MGC-producing hemolymph samples, but this number could be as low as 5-10% in other batches. The actual number of cells in- volved in giant cell formation represented a small portion of the total hemocytes present in any particular monolayer preparation. Whereas a giant cell might contain >50 nuclei, it was unusual to find more than 5-10 giant cells per coverslip, and each preparation consisted of many thousands of hemocytes. Polykaryon formation was followed in 20 batches of clams collected at roughly monthly intervals, but its fre- quency did not correlate with any obvious parameters such as water temperature, photoperiod or size of organism. The clams were obtained from commercial sources, so nothing was known of their prior history, exposure to environmental stressors, etc. Attempts were made to augment giant cell formation in vitro, by the addition of foreign particles to the media overlying the hemocyte cultures, or in vivo, by injecting the particles directly into the adductor muscle hemolymph sinus, and taking hemo- lymph samples at intervals thereafter. The results of these studies will not be pre- sented in detail because it was impossible to correlate statistically the frequency of giant cell formation with either of the treatments. Metal particles were phagocytosed in vitro to some extent, for example, carbon, graphite, aluminum, brass, or iron parti- cles immediately adhered to the hemocytes and were ingested by 2-4 h. Silica and titanium dioxide were infrequently actually phagocytosed, but remained bound to hemocyte membranes. Considerable uptake of 2.02 ^m diameter latex beads was seen, but phagocytosis of 1.10 nm diameter beads was minimal. Both FBC and LC were seen in many of the above-mentioned cultures, but at about the same frequency as in the untreated control preparations. Yeast particles were rapidly ingested in vitro and enclosed in large phagocytic vacuoles. In about half of the studies, phagocytic stimulation of the hemocytes by yeast cell exposure caused a marked increase in giant cell formation; however, repeated attempts to reproduce this phenomenon gave in- consistent results. The rate of giant cell formation in vivo by hemocytes taken from clams injected 1-48 h previously with particulates was comparable to that seen in controls. Hemocytes from these animals often had cell surface-bound and/or ingested particles. In some cases, e.g. 6 h after injection of 50 ^g graphite in 0.5 ml SSW, the foreign material was recovered in clumps enveloped by cellular capsules. If con A was present in the medium, the hemocytes underwent membrane alter- ations apparently leading to clumping and eventual fusion. The course of events fol- lowing exposure to any given con A concentration was essentially the same, regardless of length of exposure (2-48 h). The typical appearance of an early (6 h) untreated hemocyte preparation is shown in Figure 1 1 ; the cells are rather evenly distributed over the glass substrate. However, if the cells from the same animal were exposed to Con A in the medium, the hemocytes showed a marked tendency to clump and form syncytial masses (Fig. 12). Exposure of hemocytes to 60-250 jug/ml con A produced numerous FBC and LC, as well as cell aggregations, in the hemocyte preparations from >80% of the clams 240 * • % R. S, ANDERSON v< -\ « * * '• f FIGURE 7. A mature LC in a 48 h preparation containing at least 80 nuclei. FIGURE 8. Two FBC (24 h) connected by cytoplasmic strands, the lower cell contains unidentified amorphous material (arrow). FIGURE 9. A mature FBC (48 h) with randomly arranged nuclei and several yellow-brown inclusions (arrows). FIGURE 10. A FBC (24 h) containing cell-like structure (arrow) in a large vacuole. Scale bar = 20 for Figures 7-10. MULTINUCLEATED GIANT CELLS IN BIVALVES 241 ..• • ! i f I- FIGURE 1 1 . Untreated adherent hemocytes after 6 h; cells are randomly dispersed over the prepara- tion. Scale bar = 200 ^m. FIGURE 1 2. Cells from sample hemolymph sample as in Figure 1 1 . Adherent cells were exposed to 250 Mg/ml con A for 1 h and maintained for 5 h in the absence of con A. Many foci of aggregation are present throughout the preparation. Scale bar = 200 ^m. FIGURE 1 3. Typical FBC in preparation exposed to con A (62.5 ng/m\, 24 h). FIGURE 14. A FBC with highly ruffled membranes in a 24 h culture exposed to 125 ng/m\ con A. Scale bar = 20 ^m for Figures 1 3 and 14. 242 R. S. ANDERSON studied. Cellular adhesion to glass was increased by exposure to con A, as shown by the unusually dense cell pellets remaining after vigorous washing with SSW. The morphology of MGC produced in the presence of con A was generally similar to that in spontaneously produced polykaryons (Fig. 13); however, occasionally FBC with highly ruffled membranes were present (Fig. 14). At higher con A concentrations (>250 ^g/ml), hemocyte clumping and disruption were pronounced. The cells ap- peared to be extensively lysed, and the preparations contained filamentous networks and flocculent masses containing entrapped nuclei and debris. Careful examination of 500 hemocytes in each of 10 monolayer preparations showed no evidence of mitotic activity; similarly, no sign of nuclear division was seen in any of 20 randomly selected giant cells. The apparent mitotic figure in Figure 13 is an artifact which is not seen in other focal planes of this cell. Since mitosis is very rare in hemocytes, either immediately after withdrawal or during our in vitro studies, the role of endomitosis in Mercenaria giant cell formation is probably minimal. DISCUSSION It is now accepted that MGC in laboratory rodents and other vertebrates arise from fusion of mononuclear phagocytes, based on direct observation (Sutton and Weiss, 1966; Bayliss, 1976), incorporation of labeled macrophage nuclei into MGC (Mariano and Spector, 1974), and the production of hybrid enzymes in MGC exu- dates from chimeric mice composed of two strains, each homozygous for one of the two isoenzymic forms of glucose-6-phosphate isomerase (Murch el ai, 1982). The blood phagocytes of most invertebrates examined so far show a number of morpho- logical and biochemical similarities to the mononuclear leukocytes of vertebrates (Anderson et al, 1973; Anderson, 1975; 1977; 1981). The ability of Mercenaria he- mocytes to form several classical types of MGC is taken as another example of the early phylogenetic origins of macrophage-like cells. Multinucleated giant cells seen in these clam hemocyte cultures were virtually identical to those described in verte- brates. Several lines of indirect evidence point to fusion as the means of MGC forma- tion in Mercenaria. Mitotic figures were almost never seen in individual hemocytes or in polykaryons. Polykaryon formation was rare at low cell densities, when the individual hemocytes were not in contact with each other. Concanavalin A was shown to promote hemocyte aggregation and MGC formation. It is not clear that FBC and LC are distinct cell types, many intermediate forms exist in both Mercenaria and in vertebrates. Early ideas that FBC were formed only in response to the presence of insoluble foreign substances and that LC were restricted to granulomata associated with certain infections have proven to be subject to numer- ous exceptions. Most evidence supports the view that LC and FBC are actually different states of the same cell (Roizman, 1962). While there are reports that LC are the precursors of FBC (van der Rhee et al., 1978; 1979), it is generally thought that LC arise from FBC, following microtubule-dependent reordering of the nuclei and other organelles (Mariano and Spector, 1974; Adams, 1976; Chambers and Spector, 1982). The results of this study suggest that a similar transformation of FBC to LC may occur in Mercenaria, because LC usually appeared in culture later than FBC, and because the LC:FBC ratio generally increased between 1 2-48 h. Long-term study of MGC in Mercenaria was difficult because the cells apparently became nonadherent or lysed after several days in culture. Murine MGC were also reported to have a short lifetime (Papadimitriou and Sector, 1971), with the actual life span inversely proportional to the number of nuclei per cell (Papadimitriou and Walters, 1979). The exact stimulus for MGC formation by Mercenaria hemocytes is unknown, MULTINUCLEATED GIANT CELLS IN BIVALVES 243 but several suggestions have been proposed to account for macrophage fusion in higher animals. There are reports that fusion is directed by lymphokines released after antigenic stimulation (Galindo, 1972; Parks and Weiser, 1975), but many of the agents most commonly used to induce MGC formation are poorly antigenic. Indeed, thymectomy was shown to have little effect on MGC production (Papadimitriou, 1976). In this study, it is unlikely that soluble blood cell products, either secreted into the hemolymph or released from accidental hemocyte lysis during hemolymph collection and subsequent handling, mediated MGC formation. All hemocyte prepa- rations were maintained in media containing pooled, cell-free hemolymph, which presumably would contain hemocyte products, but MGC formation was seen in only certain samples. The evidence strongly suggests that the capability to spontaneously form MGC was restricted to individual Mercenaria, and was not merely a result of in vitro culture of hemocytes. It was suggested that MGC were produced as a result of recognition of cell surface abnormalities on ageing or otherwise altered macrophages (Mariano and Spector, 1974), but various experimental alterations of macrophage surfaces failed to lead to the formation of polykaryons (Chambers, 1977a). Another possible theory (Chambers, 1977b) suggests that MGC form after the simultaneous attachment of several macrophages to the same ingestible material. The cells then fuse in the course of their efforts to ingest the same particle; cell membrane fusion normally accompanies the phagocytic process. Attempts were made to promote MGC formation by Mercenaria hemocytes in vivo and in vitro by exposure to a variety of foreign particles. Whereas many kinds of particles were phagocytosed, in few cases were there clear increases in MGC forma- tion caused by competition by the cells for the material in question, or as an incidental consequence of the cells ingesting material in close proximity to other hemocytes, as suggested by Chambers and Spector (1982). The plant lectin con A promoted MGC formation in Mercenaria hemocyte cul- tures; both FBC and LC were more abundant after con A treatment. Herman and Stulberg (1962), Ptak et al. (1970), Smith and Goldman (1971), and Chambers (1977c, d) using lectins or antimacrophage serum, showed that if human or rodent macrophages are brought into contact before ingestion of surface-bound material, subsequent incubation will result in cell fusion. These molecules which induce MGC formation after interaction with the membrane may mimic the action of natural moi- eties generated in the inflammatory environment, and may account for the apparent lack of foreign material in certain MGC when examined under the microscope. Irregular, yellow-brown granules, resembling those of brown or serous cells, were seen free in the cytoplasm of FBC, or still contained in apparently intact cells within the giant cells. It is interesting that cell-like inclusions can also be found in murine MGC (see Fig. 5 in Mariano and Spector, 1974). The central location of this material in clam FBC of all sizes suggests a possible role in MGC initiation, possibly following phagocytosis. The presence of foreign material in the cytoplasm suggested to early workers that augmented phagocytosis was a major function of MGC. However, more recent work showed that giant cells have a lower phagocytic potential than an equiva- lent mononuclear cell population. In fact, it was shown that both reduction in phago- cytic activity and loss of cell surface receptors accompany the development of MGC (Papadimitriou^ al, 197 5; Chambers, 1977e). The granulomatous inflammatory response in vertebrate animals has received continuous scientific study for many years. Comparable information on invertebrates is not as comprehensive. Mammalian granulomas can be defined as a compact, orga- nized collections of mature mononuclear phagocytes (Adams, 1976); frequently other elements of the mononuclear system are also present, including epithelioid cells 244 R. S. ANDERSON and multinucleated giant cells. Giant cells are commonly associated with inflamma- tory responses to bacterial and viral infections, implanted foreign material, and/or neoplasms. Preliminary evidence presented here suggests that giant cells are also found in invertebrates, and that they are probably derived, as is the case in higher animals, from fusion of phagocytic leukocytes. ACKNOWLEDGMENTS The author gratefully acknowledges the expert technical assistance of Miss Ann M. Murtha, and wishes to thank Mrs. Billie Little for typing the manuscript. The study was partially funded by grant OCE-8016186 from the National Science Foun- dation. LITERATURE CITED ADAMS, D. O. 1976. The granulomatous inflammatory response. A m. J. Pathol. 84: 164-191. ANDERSON, R. S. 1 975. Phagocytosis by invertebrate cells in vitro: biochemical events and other character- istics compared with vertebrate phagocytic systems. Pp. 153-180 in Invertebrate Immunity, K. Maramorsch and R. E. Shope, eds. Academic Press, New York. ANDERSON, R. S. 1977. Biochemistry and physiology of invertebrate macrophages in vitro. Comp. Patho- biol.3:l-2\. ANDERSON, R. S. 1981. Comparative aspects of the structure and function of invertebrate and vertebrate leukocytes. Pp. 629-641 in Invertebrate Blood Cells, Vol. 2, N. A. Ratcliffe and A. F. Rowley, eds., Academic Press, London. ANDERSON, R. S., B. HOLMES, AND R. A. GOOD. 1973. Comparative biochemistry of phagocytizing insect hemocytes. Comp. Biochem. Physiol. 46B: 595-603. BAYLISS, O. B. 1976. The giant cell in cholesterol resorption. Brit. J. Exp. Pathol. 57: 610-618. BERMAN, L., ANDC. S. STULBERG. 1962. Primary cultures of macrophages from normal human peripheral blood. Lab. Invest. 11: 1322-1331. BORREL, A. 1 893. Tuberculose pulmonaire experimentale. Ann. Inst. Pasteur Paris 7: 593-627. CHAMBERS, T. J. 1977a. Failure of altered macrophage surface to lead to the formation of polykaryons. J. Pathol. 122: 185-189. CHAMBERS, T. J. 1977b. Fusion of macrophages following simultaneous attempted phagocytosis of gluter- aldehyde-fixed red cells. J. Pathol. 122: 7 1-80. CHAMBERS, T. J. 1977c. The mechanism of fusion of hamster macrophages induced by antimacrophage serum./ Pathol. 122: 163-173. CHAMBERS, T. J. 1977d. Fusion of hamster macrophages induced by lectins. /. Pathol. 123: 53-61. CHAMBERS, T. J. 1977e. Studies on the phagocytic capacity of macrophage polykaryons. J. Pathol. 123: 65-77. CHAMBERS, T. J., AND W. G. SPECTOR. 1982. Inflammatory giant cells. Immunobiology 161: 283-289. CHENG, T. C. 1981. Bivalves. Pp. 233-300 in Invertebrate Blood Cells, N. A. Ratcliffe and A. F. Rowley, eds. Academic Press, London. CHENG, T. C., AND P. C. GALLOWAY. 1970. Transplantation immunity in mollusks: the histoincompati- bility ofHelisoma duryi normalewilh allografts and xenografts. / Invert. Pathol. 15: 177-192. GALINDO, B. 1972. Antigen-mediated fusion of specifically sensitized rabbit alveolar macrophages. Infect. Immunol. 5: 583-594. LEWIS, M. R., AND W. H. LEWIS. 1926. Transformation of mononuclear blood cells into macrophages, epithelioid cells, and giant cells in hanging-drop blood-cultures from lower vertebrates. Contrib. Embryo!. 18:95-120. MARIANO, M., AND W. G. SPECTOR. 1974. The formation and properties of macrophage polykaryons (inflammatory giant cells). J. Pathol. 113: 1-19. MOORE, M. N., AND D. M. LOWE. 1977. The cytology and cytochemistry of the hemocytes of Mytilus edulis and their response to experimentally injected carbon particles. J. Invert. Pathol. 29: 1 8- 30. MURCH, A. R., M. D. GROUNDS, C. A. MARSHALL, AND J. M. PAPADIMITRIOU. 1982. Direct evidence that inflammatory multinucleate giant cells form by fusion. J. Pathol. 137: 177-180. PAPADIMITRIOU, J. M. 1976. The influence of the thymus on multinucleate giant cell formation. / Pathol. 118:153-156. PAPADIMITRIOU, J. M., T. A. ROBERTSON, AND M. N-I. WALTERS. 1975. An analysis of the phagocytic potential of multinucleate foreign body giant cells. Am. J. Pathol. 78: 343-353. MULTINUCLEATED GIANT CELLS IN BIVALVES 245 PAPADIMITRIOU, J. M., AND W. G. SPECTOR. 1971. The origin, properties, and fate of epithelioid cells. J. Pat hoi. 105: 187-203. PAPADIMITRIOU, J. M., AND M. N-I. WALTERS. 1979. Macrophage polykarya. Crit. Rev. Toxicol. 6: 21 1- 255. PARKS, D. E., AND R. W. WEISER. 1975. The role of phagocytosis and natural lymphokines in the fusion of alveolar macrophages to form Langhans giant cells. J. Reticuloendothel. Soc. 17: 2 1 9-228. PTAK, W., Z. PORWIT-BOBR, AND Z. CHLAP. 1970. Transformation of hamster macrophages into giant cells with antimacrophage serum. Nature 225: 655-657. VAN DER RHEE, H. J., W. HILLEBRANDS, AND W. TH. DAEMS. 1978. Are Langhans giant cells precursors of foreign-body cells? Arch. Dermatol. Res. 263: 13-21. VAN DER RHEE, H. J., C. P. M. VAN DER BURGH-DE WINTER, AND W. TH. DAEMS. 1979. The differentia- tion of monocytes into macrophages, epithelioid cells, and multinucleated giant cells in subcuta- neous granulomas. Cell Tissue Res. 197: 355-378. ROIZMAN, B. 1962. Polykaryocytosis. Cold Spring Harbor Symp. Quant. Biol. 27: 327-342. RUDDELL, C. L., AND S. R. WELLINGS. 1971. The ultrastructure of the oyster brown cell, a cell with a fenestrated plasma membrane. Z. Zellforsch. 120: 17-28. SMITH, C. W., AND A. S. GOLDMAN. 1971. Macrophages from human colostrum. Multinucleated giant cell formation by phytohmagglutinin and concanavalin A. Exp. Cell Res. 66: 317-320. SPARKS, A. K., AND G. B. PAULEY. 1 964. Studies of the normal postmortem changes in the oyster Crassos- trea gigas (Thunberg). J. Insect Pat hoi. 6:78-101. SUTTON, J. S., AND L. WEISS. 1966. Transformation of monocytes in tissue culture into macrophages, epithelioid cells, and multinucleated giant cells. J. Cell Biol. 28: 303-33 1 . VIRCHOW, R. 1858. Reizungund Reizharkeit. Virchows Arch. 14: 1-63. Reference: Biol. Bull. 172: 246-263. (April, 1987) LOW TEMPERATURE EVOKES RAPID EXOCYTOSIS OF SYMBIOTIC ALGAE BY A SEA ANEMONE R. GRANT STEEN1 AND L. MUSCATINE Department of Biology, University of California, Los Angeles, California 90024 ABSTRACT Darkness evokes expulsion of zooxanthellae from the sea anemone Aiptasia pul- chella, but brief exposure to low temperature (4°C, 4 h) increases the expulsion rate four-fold. Ninety-eight percent of the zooxanthellae are expelled within four days. Low temperature incubation had no detectable effect on host animal survival or be- havior, but the effect on zooxanthellae was profound. Low temperature reduced the rate of photosynthesis, increased the rate of release of fixed carbon, reduced the num- ber of viable cells and decreased the rate of cell division. Electron micrographs re- vealed that low temperature elicited the appearance of a thickened cell envelope and novel electron-dense inclusions, tentatively identified as crystallized lipoprotein. Im- mediately after low temperature incubation zooxanthellae move towards the apex of the host cells and are released to the coelenteron by exocytosis. The mechanisms by which low temperature may evoke exocytosis of zooxanthellae (e.g., disassembly of host cell microtubules; increased concentration of cytosolic calcium ions) are dis- cussed. INTRODUCTION Symbiotic dinoflagellates (^zooxanthellae) are found in tropical reef cnidarians (e.g., sea anemones, corals) in very high concentration relative to the surrounding water column. Under normal conditions neither the algae nor the host outgrows the other, suggesting that the algal population density is somehow regulated (Muscatine et al, 1985). Yet, when environmental conditions on coral reefs are naturally per- turbed (lowered salinity, abnormally high or low water temperature, sub-aerial expo- sure, and UV irradiation), zooxanthellae are lost from corals (Goreau, 1964; Jaap, 1979, 1985; Egana and Disalvo, 1982; Verrill, 1902, and Mayer, 1914, in Porter et al., 1982; Glynn, 1983, 1984; Lasker et al., 1984; Harriott, 1985; Fisk and Done, 1985). Additional types of stress conditions, brought to bear on symbiotic cnidarians by investigators in the laboratory or in the field, also evoke loss of zooxanthellae. These conditions include constant light, elevated temperature, elevated salinity (Reimer, 1971; see also Steele, 1976, 1977), prolonged darkness (Yonge and Nicholls, 1931; Franzisket, 1970; Kevin and Hudson, 1979), shading (Rogers, 1979), and an- oxia (Yonge and Nicholls, 1931; O. Hoegh-Guldberg, pers. comm.). Given the fundamental importance of zooxanthellae to reef corals and other sym- biotic cnidarians (Muscatine, 1980), it is surprising that very little is known of the mechanisms by which zooxanthellae population densities are controlled, or how the various environmental stresses evoke elimination of zooxanthellae. Steen (1985) recently noted that exposure of the tropical symbiotic anemone Aip- Received 2 October 1986; accepted 23 January 1987. 1 Present address: Department of Radiology, Johns Hopkins University Medical School, Baltimore, Mary land 2 1205. 246 EXOCYTOSIS OF ALGAE 247 tasia pulchella to low temperature (4°C) for one hour in darkness, followed by main- tenance at 25°C in darkness, resulted in rapid and substantial loss of zooxanthellae over the next few days. The host was otherwise unaffected. In view of the rapidity of the response and the ability of the host to survive, we have used low temperature treatment as a bioassay for the study of mechanisms by which zooxanthellae are elim- inated from their host. This paper describes aspects of the kinetics and cell biological features of the loss of zooxanthellae from a sea anemone in darkness and after brief exposure to low temperature. The results show that although darkness evokes loss of zooxanthellae, low temperature accelerates loss four-fold, increases the permeability of zooxanthel- lae to photosynthate, and alters zooxanthellae ultrastructure. The mechanism by which low temperature evokes rapid exocytosis of zooxanthellae is discussed. MATERIALS AND METHODS Organisms and maintenance The tropical sea anemone Aiptasia pulchella (Clone B; Muller-Parker, 1984) was used for all experiments. Anemones were normally maintained in large finger bowls in 800 ml seawater in a Model 808 Precision Incubator at 25°C on a 12h:12h light: dark photoperiod (70 juEin-m~2-s~'). Animals were fed freshly hatched Anemia nauplii twice weekly. After each feeding, seawater and uneaten Anemia were dis- carded and replaced with fresh seawater. When dark conditions were required, a sim- ilar incubator was used, without lights, and the temperature was adjusted to 25°C. Anemones were not fed during the experiments. Counting numbers of zooxanthellae in sea anemones Samples of five anemones of similar size were homogenized in a glass-Teflon Pot- ter homogenizer (10 ml), and the homogenate centrifuged in an IEC-HNS table cen- trifuge at 700 X g for 3 min to pellet the algae. The animal homogenate was decanted and the pellet preserved in seawater containing 4% formaldehyde. Numbers of algae per anemone were determined from hemacytometer counts of samples of formalin- fixed cells. Algal pellets were saved and used later for determination of mitotic index (see below). Viability of zooxanthellae expelled from anemones Viability of zooxanthellae was assessed from levels of fixation of 14C and from observations on dye exclusion. Expelled clumps of zooxanthellae were collected with a Pasteur pipette from bowls in which anemones were maintained. When sufficient cells were collected, the clumps were gently homogenized in several ml seawater and 5 replicate samples of 1 ml each were placed in separate 1 2 ml conical graduated centrifuge tubes. Samples were taken for cell counts. Then, 0. 1 ml of a stock solution of NaH14CO3 containing one )uCi per ml was added to each tube. After mixing the contents of each tube, 25 jul of the stock solution was immediately taken for estimation of Added Activity. Tubes were incubated for 2 h at 300 ^Ein • m2 • s"1 at 25°C and mixed again after one h. After incubation, tubes were centrifuged at 1000 X g for 4 min to pellet the algae, and the labelled seawater supernatant was decanted and saved. The algae were resuspended in 0.2 ml distilled water, the suspension was transferred to glass scintillation vials, acidified with 0. 1 ml 1 TV HC1, and evaporated to dryness with low heat on a warming 248 R. G. STEEN AND L. MUSCATINE plate. Samples of labelled supernatant were treated similarly. Then 10 ml scintillation cocktail was added to each vial. For Added Activity, samples were added to 10 ml of scintillation cocktail, and the vials were capped immediately. Vials were counted in a Beckman LSC Model 100 scintillation counter, and corrected for quench by the external standards ratio method. Results are expressed as /ug C fixed- 106 cells"1 -h"1. For the dye exclusion assay, expelled zooxanthellae were collected, and placed on a standard microscope slide. A drop of 0.25% (w/v) Evans Blue (Gaff and Okong'o- Ogola 1971) was added and the percentage of cells permeated by the dye after 5 min was established from observation of five replicate samples of 200 cells at 450X. Mitotic index Mitotic index was determined for zooxanthellae in the same samples used to mea- sure zooxanthellae population density. The number of cells undergoing cytokinesis was determined for 10 samples of 100 cells, each viewed at 450X on a hemacytometer grid to facilitate counting. Specific growth rates and loss rates of zooxanthellae Specific growth rates of zooxanthellae in situ were calculated from n = l/N(N/t) where N = the standing stock of zooxanthellae and (N/t) the increment of new cells added per day (Strickland and Parsons, 1965). By analogy, specific expulsion rates were calculated from MX- l/N(Nx/t) where Nx is the number of algae expelled, calculated from N at to — t + 10 days, and Nx/t, the average increment of cells expelled per day (O. Hoegh-Guldberg, pers. comm.). Morphology and ultrastructure For transmission electron microscopy whole anemones and discharged pellets of algae were fixed in 4% formaldehyde and 1% glutaraldehyde in 0.1 M phosphate buffer according to McDowell ( 1 978). Tentacle crowns were removed and subdivided into small pieces, rinsed in buffer, and dehydrated in a series of ethanol solutions (30%, 50%, 70%, 90%, and three changes of absolute ethanol), 30 min at each step. Tissues were transferred to absolute ethanol:propylene oxide ( 1 : 1 ) for 1 h, then into propylene oxide:medcast epoxy (1:1) and finally into 100% medcast epoxy for infil- tration overnight. Tissues were embedded in fresh 100% medcast and the material was hardened according to the manufacturer's instructions. Thick sections of tentacles and pellets were cut with glass knives, stained with methylene blue and mounted on standard microscope slides. These were viewed and in some cases photographed with a Olympus BH-2 compound microscope using Pan- chromatic BW film (ASA 200). Some tentacles were deliberately cut in cross section to facilitate determination of the intracellular locus of individual algae with respect to the long axis of their host endodermal cell. This locus is expressed in terms of Rf, the ratio of the distance from (a) the base of the endoderm cell to the leading edge of the alga, to (b) the base of the endoderm cell to the center of the tentacle lumen. The latter criterion was used because the apex of each endoderm cell usually could not be discerned. EXOCYTOSIS OF ALGAE 249 Thin sections of tentacles or algal pellets were cut with diamond or glass knives, collected on 200 mesh copper grids coated with parlodion and carbon film, and stained with uranyl acetate and lead citrate. Grids were examined with a Philips 300 Electron microscope operating at 60 kV. For scanning electron microscopy, excised tentacles and discharged pellets were fixed in 4% formalin in seawater. Tentacles were slit longitudinally with a scalpel blade and all specimens were critical-point dried in a Polaraon Critical-Point Drier, mounted on stubs, sputter-coated with gold, and viewed with an ISI-DS- 1 30 scanning electron microscope. RESULTS The effect of darkness on maintenance of zooxanthellae population density Since our bioassay was carried out in cold and darkness, it was necessary to quan- tify and thereby control for loss of algae from A. pulchella due to darkness alone. Groups of A. pulchella that were starved for one day were incubated for 4 h in dark- ness at 25°C. Thereafter they were maintained in darkness without feeding for up to 20 days. Five anemones were sampled at zero time (immediately after treatment) and at two-day intervals, and the number of zooxanthellae remaining in each anemone was determined. The results are shown in Figure 1 . In these anemones, the numbers of algae decreased steadily, and over ten days the standing stock of algae decreased by 58%. (MX - 0.06 d~'). Clumped pellets of zooxanthellae were found in the experi- mental containers in the vicinity of individual anemones during the experiment, indi- cating that the loss of substantial numbers of zooxanthellae was due to expulsion from the anemone. Digestion of some algae could have occurred as well, but this possibility was not investigated. Effect of a brief exposure to low temperature on the maintenance of zooxanthellae population density Anemones incubated at 14°C for 4 h in darkness and then maintained at 25°C in darkness lost zooxanthellae at a faster rate (MX = -08 d"1) than controls, losing about 72% of their algae in 10 days (Fig. 1 ). But when anemones were incubated at 4°C in darkness and then 25°C in darkness, loss of zooxanthellae was even more rapid. After only 4 days the population of algae had decreased by 98% (MX == -24 d '). Further decrease in numbers was only slight over the next 20 days. The combination of dark- ness and low temperature incubation caused relatively rapid loss of roughly 98-99% of the algae associated with A. pulchella. Low temperature incubation had no obvious visible effects on the host anemone behavior. They exhibited normal posture (column erect, tentacles extended) and feeding behavior, and continued to produce buds by pedal laceration for at least six months after low temperature treatment. However, in one case, several tentacle tips, 1-2 mm long and filled with zooxanthellae, were recovered from incubation dishes. Presumably these had sloughed or broken off during or after low temperature treatment. Effect of light on loss of zooxanthellae after low temperature incubation To determine if the loss of zooxanthellae after low temperature treatment in dark- ness was affected by light, a group of anemones was incubated at 4°C in darkness as usual, but then maintained in the light (70 MEin-nT2-s~') for 10 days. Anemones were sampled at two-day intervals and numbers of zooxanthellae per anemone deter- 250 R. G. STEEN AND L. MUSCATINE 120 - 100 _o ~4> O X o o V .0 e 0 468 Time (days) 10 20 FIGURE 1. Number of zooxanthellae remaining in Aiptasia pulchella vs. time in dark at 25°C after an initial 4-h incubation at 25°C (•), 14°C (•), and 4°C (A); and time in light at 25°C after an initial 4-h incubation in dark at 4°C (A). Vertical bars indicate standard error of the mean. mined. Figure 1 shows that such individuals lost abundant zooxanthellae immedi- ately and at approximately the same rate as those that were chilled and maintained in darkness, but the rate of loss decreased slightly after 2 days until, by day 6, about 91% of the original population had been expelled. Then, over the next 4 days, the population began to increase at a specific growth rate of about 0.44 d"1. Viability of cold-treated zooxanthellae To determine if low temperature incubation affected the viability of zooxanthel- lae, we tested the ability of freshly isolated and expelled zooxanthellae to photosyn- thetically fix 14C and to exclude dye. The results are given in Table I. Zooxanthellae freshly isolated (FIZ) from control anemones (25°C, dark) fixed 14C at normal rates (cf. Muller-Parker, 1984), and about 98-99% of these zooxanthel- lae were viable as judged from ability to exclude dye. FIZ from cold (4°C) treated anemones assayed immediately or after a 4-h recovery period also fixed 14C but at much lower rates. Of these cells, 9 1 .8-93.6% were judged viable. Extending the recov- ery period at 25°C to 24 h resulted in a modest increase in photosynthetic rates, but the rates were still below control levels and viability was reduced to 82.5%. FIZ customarily release a fraction of their fixed 14C to the external medium (Mus- catine, 1980). Those from A. pulchella (Clone B) normally release less than 10% of EXOCYTOSIS OF ALGAE 251 TABLE I Photosynthetic rate (x± S.D.) and % viability (dye exclusion) of zooxanthellae freshly isolated (FIZ) and/or expelled (EZ)from control (25°C dark) and chilled (4°C dark) Aiptasia pulchella Cell-specific photosynthetic rate Host Recovery period treatment before assay Cells assayed Photosynthetic rate (/ugC -106 cells' '-IT') Viable cells (MgC-106 viable cells" '-h'1) 25°Cdark 4h FIZ 9.63 ± 1.57 98.7 9.76 4. 15 ±0.53 98.5 4.22 x = 6.99 ± 3.92 4°C dark 4h FIZ 0.51 ±0.05 91.8 0.56 4 h 25°C 4 h, dark FIZ 0.70 ± 0.09 93.6 0.74 4 h 25°C 24 h, dark FIZ 2.89 + 0.21 82.4 3.50 EZ 1.53±0.13 82.3 1.86 x= 1.42 ± 1.02 the total 14C fixed in a one-hour incubation (O. Hoegh-Guldberg, pers. comm.). Our controls released 28.8% under similar conditions (n = 5). In contrast, zooxanthellae from cold-treated anemones released an average of 58.8 ± 3.4% (n = 5) of the total 14C fixed. Therefore, the cold treatment seemed to reduce rates of photosynthesis, increase the percentage of fixed 14C released, and reduce numbers of viable cells. Expelled zooxanthellae (EZ) were similar to FIZ with respect to 14C fixation. Algae expelled up to four hours after cold treatment showed greatly reduced rates of photo- synthesis and about 6 1 % viability. Those expelled during a recovery period of 4-24 hours at 25°C showed some increase in rate of photosynthesis and level of cell via- bility, but the mean cell-specific rate of photosynthesis (1.42 ± 1.02 ^gC-106 cells"1 -IT1) was still significantly lower than that of freshly isolated controls (6.99 ± 3.92; P < 0.05, Student's /-test). Release of fixed 14C by EZ was not assayed. Effect of low temperature on cell division The mean MI for zooxanthellae from A. pulchella (Clone B) maintained normally is 0.76% (Wilkerson et al, 1983), and ranges from 0.38 to 1.54% (Muller-Parker, 1984). Figure 2 shows that mitotic index of zooxanthellae from controls incubated at 25°C ranged from 1.65% at time zero to 4.6-5.1% after 2-4 days in darkness and then declined again to 2% after 10 days. In contrast, algal cells from anemones incu- bated at 14° or 4°C had lower mitotic indices, ranging between 1 .25 and 0.25%. These results indicate that darkness gave rise to a transient increase in MI and that brief exposure to low temperature may inhibit such an increase. Zooxanthellae from anemones incubated in light at 25°C after low temperature incubation in darkness exhibit a higher mitotic index, suggesting that light may reverse the inhibitory effect of low temperature on cell division. Morphological correlates of expulsion of zooxanthellae Intracellular loci of zooxanthellae. Zooxanthellae are confined to the endoderm cells of A. pulchella. Most endoderm cells contain 1-3 algae, each within an animal cell vacuole (Glider et al., 1980). To gain insight into morphological correlates of 252 R. G. STEEN AND L. MUSCATINE X OJ 73 c 0 468 Time (days) 10 20 FIGURE 2. Mitotic index (%) of zooxanthellae in Aiptasia pulchella vs. time in dark (or light) at 25°C after initial treatment as in Figure 1 . Symbols and vertical bars as in Figure 1 . expulsion, the loci of algae along the long axis of host endoderm cells, before and after exposure to darkness and low temperature, were determined from Rf measure- ments of 80-200 cells in thick sections of tentacles. Results are shown in Figure 3. Zooxanthellae in cells of controls maintained in darkness for 4 h at 25°C were evenly distributed along the long axis of host cells. During prolonged maintenance in dark- ness, the population shifted toward the apical end of the host cells. In cold-treated anemones, this shift was detectable even at time zero. Within 24 hours after the cold treatment, the pattern appears reversed, but it is likely that the zooxanthellae near the luminal face of the host cell had already been expelled. The apical shift is restored after two days and considerably exaggerated after four days. These data are consistent with the interpretation that a change in intracellular locus precedes and accompanies the discharge of zooxanthellae. Uhrastructure of zooxanthellae in situ and during expulsion. Zooxanthellae in situ in anemones maintained in light at 25°C are shown in Figure 4a. The cell envelope consists of two to five sets of double membranes between the presumptive animal cell vacuolar and zooxanthellae plasma membranes. The average thickness of the envelope is about 60 nm. The chloroplast is lobed, with a prominent chloroplast membrane and lamellae in the three-thylakoid configuration. The pyrenoid (Fig. 4a, inset) is congruent with the chloroplast stroma but apparently not traversed by thyla- koids. It is capped by a thick sheath, presumably consisting mainly of starch. The nucleus contains condensed chromosomes. Mitochondria are present and adjacent to the pyrenoid there is a large amorphous body and crystal-like electron translucent profiles. Zooxanthellae in situ in the anemones incubated at 4°C and sacrificed at day 0 EXOCYTOSIS OF ALGAE 253 20 '10 0 Control Day 0 TrTHtfl- 20 10 0 Control Day 10 , — , 20 1 20^ 10-1 H IT" ]Q\ "0 2 .4 6 .8 10 C Rf 20^ 01 °10^ 0 20 °/ °10^ 0 4° Day 0 inn 4° Day 1 4° Day 2 4° Day 4 0 .2 .4 .6 .8 1.0 FIGURE 3. Number of zooxanthellae (%) vs. relative position along the long axis of host cells in control (25°C, dark 4 h) and chilled (4°C, dark 4 h) anemones. Dashed line indicates mean Rf. (immediately after cold treatment) are shown in Figure 4b. In some cells the average thickness of the cell envelope has increased and there is loss of resolution of cell envelope membranes. The most striking change is manifested by the appearance of electron-dense inclusions and an increase in the density of the pyrenoid starch cap. After one day (Fig. 5a), cell envelope thickness has increased to about 120 nm. The dense inclusions are less granular and now exhibit a dense core. Dense core inclusions are also present in the host cell. Electron-dense inclusions are still evident in zooxan- thellae in anemones sacrificed after two days (not shown). Release of zooxanthellae from tentacle cells. The inner surface of tentacles is lined with flagellated endoderm cells, each containing zooxanthellae. In tentacles of control anemones, the zooxanthellae are situated proximally within the host cells and the endodermal epithelium is relatively smooth (not shown). As a result of chilling, the zooxanthellae move distally within the host cells, causing the distal end of the host cell to bulge into the tentacle lumen and to assume the spherical shape of the zooxan- thellae they contain. Low power scanning electron micrographs of the tentacle endo- derm immediately after chilling give the general appearance of a tentacle lined with extracellular zooxanthellae (Fig. 6a). However, at this time most of the zooxanthellae are still within the host cells. Figure 6b shows the flagella at the apex of the extended host cells and in one case a torn host cell shows that it persists as a thin enclosure for a zooxanthellae cell. About 8 h after chilling, zooxanthellae can be seen emerging from host cells, apparently by exocytosis (Fig. 6c). By 12 h the terminal stages of exocytosis and release of zooxanthellae can be observed (Fig. 6d). By 20 h most of the zooxanthellae have been discharged from the epithelium. Those few that remain are also in late stages of exocytosis (Fig. 7a). Algae are discharged from the coelenteron as individual cells or as pellets of cells. The pellets are irregularly spherical, and also contain bacteria and assorted fibrils EXOCYTOSIS OF ALGAE 255 and sheets of material of unknown origin and composition (Fig. 7b). Transmission electron micrographs of pellets (Fig. 5b) show that some zooxanthellae are apparently intact and exhibit a thick cell envelope, while others are in various stages of disintegra- tion. Each is surrounded by one or more thin membranes raised above the surface of the cell. Figure 7c shows that the algae are generally spherical but enclosed by a loose wrinkled sheath probably corresponding to the elevated membranes shown in Fig. 5b). In cells in situ this sheath is torn apart and shed as the two daughter cells undergo cytokinesis (Fig. 7d) and is therefore interpreted as a mother cell envelope. DISCUSSION The effect of darkness on maintenance of zooxanthellae population density The results of this investigation show that when the symbiotic anemone Aiptasia pulchella (Clone B) is maintained in darkness, without feeding, it progressively loses zooxanthellae. These results confirm the observations of Muller-Parker (1984), that darkness causes loss of zooxanthellae from A. pulchella (Clone B), and extend them by showing that at least 50% of the zooxanthellae are lost in 10 days. Previous studies have demonstrated that symbiotic anemones and corals lose zoo- xanthellae in prolonged darkness (Yonge and Nicholls, 1931; Buchsbaum, 1968; Franzisket, 1970; Kevin and Hudson, 1979; Rogers, 1979; Muller-Parker, 1984), but neither the time course nor the mechanism of loss has been rigorously investigated. More recently, Glider (1983) demonstrated that symbiotic Aiptasia pallida main- tained on a 12 h light: 12 h dark photoperiod continuously extrude pellets of zooxan- thellae but in significantly greater numbers during the dark period. In the case of A. pulchella, loss of zooxanthellae in darkness could occur by expul- sion or digestion. The appearance of numerous clumps of zooxanthellae in the vicin- ity of individual anemones shows that loss is due, at least in part, to expulsion. Data from Figure 3 show that algae are initially distributed more or less evenly along the long axis of the host cell. In darkness, they move toward the distal end of the host cell and are then released from host cells by exocytosis (Fig. 6c, d). In the coelenteron they form large pellets which are then egested. Egestion of regularly formed pellets of zooxanthellae by A. pallida and by Hawaiian zoanthids has been described by Reimer (1971). The mechanism which evokes intracellular migration and exocytosis of zoo- xanthellae in darkness is unknown but may be correlated with diminished algal pho- tosynthesis. In this connection, we note that the maintenance of symbiotic Chlorella sp. in the cells of green hydra may depend on the sustained translocation of photosyn- thetically fixed carbon from the algae to the host cells (Hohman et al, 1982). There is evidence for presumed digestion of zooxanthellae in some symbiotic cnidarians (Fitt and Trench, 1983) but the possibility was not investigated in this study. The remainder of this discussion will address loss of zooxanthellae by expul- sion only. FIGURE 4. (a) Transmission electron micrograph of a zooxanthella in Aiptasia pulchella maintained at 25°C on a 12L:12D photoperiod. ce, multilayered cell envelope; cp, chloroplast; m, mitochondrion n, nucleus with condensed chromosomes; p, pyrenoid, 20,383X. Inset, chloroplast-pyrenoid complex, 17,976X. (b) Zooxanthella from an anemone maintained 4 h at 4°C in dark. Electron-dense inclusions at arrows. 11,194X. 256 R. G. STEEN AND L. MUSCATINE '«* B EXOCYTOSIS OF ALGAE 257 The effect of low temperature on maintenance of zooxanthellae population density Brief exposure to low temperature increased the rate of loss of zooxanthellae slightly at 14°C but four-fold at 4°C (Fig. 1). We speculate that low temperature could potentiate exocytosis by increasing the rate of microtubule depolymerization (Mel- konian et al, 1980; Dustin, 1984). For example, in green hydra, endoderm cells form endocytic vacuoles upon uptake of symbiotic Chlorella sp. These vacuoles migrate away from the site of uptake (Muscatine et al., 1975). Migration is thought to be mediated by microtubules since the movement of vacuoles containing algae is inhib- ited by colchicine (Cook, unpub., in Muscatine et al., 1975) or vinblastine (McAuley and Smith, 1 982). Treatment of green hydra with vinblastine, which promotes micro- tubule disassembly, causes the algae, which are normally situated at the base of the host cell, to gradually shift back toward the apex of the cell (L. Muscatine, unpub.). Low temperature could also stimulate exocytosis by causing an increase in cytosolic calcium ions (Campbell, 1983). Increased cytosolic calcium has been implicated in a wide range of systems involving stimulus-secretion coupling (See, for example, Doug- las, 1973;DahlandHenquin, 1978; Spearman and Butcher, 1982). Effect of light on loss of zooxanthellae after low temperature incubation Maintenance of A. pulchella in light after low temperature incubation in darkness did not significantly affect loss of zooxanthellae for up to six days. This suggests that the effect of low temperature incubation is not quickly or easily reversed by the resto- ration of photosynthesis or translocation, or the putative reassembly of microtubules. After six days, the standing stock of algae begins to increase, suggesting that the effect of low temperature is no longer manifested. Viability of cold-treated zooxanthellae Zooxanthellae expelled by A. pallida are photosynthetically active (Glider, 1 983), as are zooxanthellae freshly isolated from A. pulchella after incubation for 4 h in darkness at 25°C (Table I). However, incubation of A. pulchella at 4°C caused a de- crease in photosynthetic rate and impaired the ability of freshly isolated or expelled cells to exclude dye. This effect of low temperature was only partially reversible at 24 h, again suggesting that metabolism was substantially altered. This conjecture is supported by data showing that release of fixed 14C by chilled cells was higher than release by controls. Table I also shows that, since the percentage of viable cells in samples of expelled zooxanthellae was less than the percentage in freshly isolated zooxanthellae, the cells impaired by chilling may be preferentially expelled. Interestingly, the viability of zooxanthellae freshly isolated from controls was 98.6%. These cells were separated from the host by homogenization and centrifuga- tion. However, if cells are isolated by excising a tentacle and gently pressing it with a glass rod, then 100% of the cells are able to exclude the Evans Blue dye. This implies that either a small proportion of the zooxanthellae are damaged by the homogeniza- tion method or that non- viable cells are derived from other parts of the anemone. FIGURE 5. (a) Transmission electron micrograph of a zooxanthella from an anemone incubated 4 h in dark at 4°C, then 24 h in dark at 25°C. Electron-dense inclusions at arrows. 18,846X. (b) Transmission electron micrograph of a pellet of zooxanthellae discharged after 5 days from anemones treated as in 5a. 7780X. 258 R. G. STEEN AND L. MUSCATINE EXOCYTOSIS OF ALGAE 259 Effect of low temperature on cell division The MI of zooxanthellae increased from the normal range of 0.38% to 1.54% to 4.6% to 5.1% after several days in darkness. The reason for this increase is not yet known, but a similar result was obtained by Glider (1983). He found that MI was greater in darkness in zooxanthellae both in situ in A. pallida and in expelled pellets. Given that algae are expelled in darkness, the effect of increased division rate in dark- ness is to offset the rate of expulsion. Since low temperature decreases MI (Fig. 2) it is possible that accelerated loss arises merely from a reduction in the rate of replace- ment. Calculations addressing this possibility (Appendix 1) suggest that it is highly unlikely that accelerated loss arises solely from inhibition of division (replace- ment) rate. Morphological correlates of expulsion of zooxanthellae Transmission electron micrographs of chilled zooxanthellae in situ show a thick- ened cell envelope. The significance of the thickened envelope is not yet understood but it does resemble the "pellicle" which develops when zooxanthellae are grown in culture (Schoenberg and Trench, 1980; Glider, 1983). The nature of the electron- dense entities in zooxanthellae cytoplasm which do not appear in controls is un- known. Puppione et al (1982) suggest that triglyceride-rich lipoprotein enriched in saturated fats, upon exposure to temperatures sufficiently low to induce phase transi- tion, forms crystallized particles enclosed within a phospholipid monolayer. If zoo- xanthellae contain such materials in membranes, and if they crystallize at low tem- perature, membrane function might be perturbed. Such a perturbation might explain the observed increased release of photosynthate by chilled cells. In this connection, Harnischfeger and Jarry (1982) report that cooling Chlorella emersonii to around -5°C causes a reversible alteration in thylakoid fluidity which in turn increases the permeability of thylakoids to cations. Glider (1983) observed transmission electron micrographs of host cells during normal expulsion of zooxanthellae. Algae at the distal end of host cells were covered with a thin layer of cytoplasm connected to the cell by a narrow tissue bridge. This profile was interpreted as imminent "budding" rather than classical exocytosis. The animal tissue covering was thought to be digested in the coelenteron since it was not present in zooxanthellae pellets. We suggest that Glider (1983) observed the apical "bulging" of host cells caused by the extreme distal migration of zooxanthellae. While it is possible that some zooxanthellae are released by "budding," our scanning elec- tron micrographs (Figs. 6b-d) offer direct evidence to support the interpretation that zooxanthellae are discharged by exocytosis. Further, the loose, wrinkled sheath which we observed in pellet zooxanthellae (Fig. 7c) more likely corresponds to a mother cell envelope, since it seems to be intimately associated with the cells during cytokinesis. Studies now in progress show that brief exposure to low temperature evokes exo- cytosis of zooxanthellae in A. pallida from Bermuda and Aiptasia sp. from Java, but not in several scleractinian corals from Bermuda. The specificity of this phenomenon is now under investigation. FIGURE 6. (a) Inner surface of Aiptasia pulchella tentacle immediately after chilling, showing endo- derm and abundance of zooxanthellae. 234X. (b) As above, showing that zooxanthellae are still within host cells. Host cell flagella and torn host cell revealing zooxanthella surface at arrows. 2400X. (c) As above, 8 h after chilling, showing zooxanthellae within and emerging from host cells (at arrows). 4750X. (d) As above, 1 2 h after chilling, showing zooxanthella freed from exocytotic cup (arrows). 3250X. 260 R. G. STEEN AND L. MUSCATINE JL^SiSfc^ EXOCYTOSIS OF ALGAE 261 ACKNOWLEDGMENTS We thank Nedra Simmons for technical assistance, Barbara Muller and Monica Eiserling for assistance with scanning and electron microscopy, and Dr. Charles R. Taylor for computation of Appendix 1 . Supported by a UCLA Biomedical Research Support Grant and by NSF OCE 85 10518. APPENDIX 1 Let N, = standing stock of algae per anemone at time t dN, = number of algae produced - number of algae expelled dt per unit time per unit time Let m = number of algae produced per alga per unit time, and let x = probability of an algae being expelled per unit time. Then, f = N,(m - x) and since dN, 1 M, then dt N, fj. = m — x Our task is to compute X25 and M4 = m4 - x4, and to compare x25 and x4. From Figure 1 we know Nt, so we can calculate n at 25° and 4°C from N, = Noe"1 This yields M25 = —.08, and ^4 = -.92 Then, m= l/tdln(l +f), where td = duration of cytokinesis and f = the average daily mitotic index. Assuming that td is the same at 25° and 4°C, and taking the value of td = 0.46 days (Wilkerson et al, 1983), then from Figure 2 f25 = 3.33% and f4 = 0.54%, and m25 = 0.71 and m4 = 0.12 FIGURE 7. (a) Endodermal surface of tentacle ofAiptasia pulchella, 20 h after chilling, showing a pair of zooxanthellae in late stage of exocytosis from host cell, and paucity of zooxanthellae elsewhere. 591 Ox. (b) Pellet of zooxanthellae discharged by chilled anemone. 24 IX. (c) Zooxanthellae in discharged pellet, each with a wrinkled envelope and connected by mucus strands. 3590X. (d) Zooxanthellae in tenta- cle lumen of anemone treated 4 h, 4°C in dark, showing various aspects of zooxanthellae mother cell envelope during cytokinesis. 2430X. 262 R. G. STEEN AND L. MUSCATINE Therefore, X25 = m25 — (J-25 = .07 - (-.08) = .15 and X4 = .01 -(-.92) = .93 Therefore the probability of a zooxanthella being expelled after chilling is six times higher than that at 25°C. LITERATURE CITED BUCHSBAUM, V. 1968. Behavioral and physiological responses to light by the sea anemone Anthopleura elegantissima as related to its algal endosymbionts. Ph.D. Thesis, Stanford University, Palo Alto, CA. CAMPBELL, A. K. 1983. Intracellular Calcium: Its Universal Role as Regulator. John Wiley, New York. 556 pp. DAHL, G., AND J.-C. HENQUIN. 1978. Cold-induced insulin release in vivo: Evidence for exocytosis. Cell Tissue Res. 194:387-398. DOUGLAS, W. W. 1973. Involvement of calcium in exocytosis and the exocytosis-vesiculation sequence. Biochem. Soc. Symp. 39: 1-28. DUSTIN, P. 1984. Microtubules. Springer- Verlag, Berlin. 482 pp. EGANA, A. C., AND L. H. DiSALVO. 1982. Mass expulsion of zooxanthellae by Easter Island corals. Pac. Sci. 36:61-63. FITT, W. K., AND R. K. TRENCH. 1983. Endocytosis of the symbiotic dinoflagellate Symbiodinium micro- adriatricum Freudenthal by endodermal cells of the scyphistomae of Cassiopeia xamachana and resistance of the algae to host digestion. J. Cell Sci. 64: 1 95-2 1 2. FISK, D. A., AND T. J. DONE. 1985. Taxonomic and bathymetric patterns of bleaching in corals, Myrmidon Reef (Queensland). Proc. Fifth Int. Coral Reef Cong. 6: 149-1 54. FRANZISKET, L. 1970. The atrophy of hermatypic reef corals maintained in darkness and their subsequent regeneration in light. Int. Rev. Hydrobiol. 55: 1-12. GAFF, D. F., AND O. OKONG'O-OGOLA 1971. The use of non-permeating pigments for testing the survival of cells. J. Exp. Bot. 22: 756-758. GLIDER, W. V., D. W. PHIPPS, JR., AND R. L. PARDY. 1980. Localization of symbiotic dinoflagellate cells within tentacle tissue of Aiptasia pallida (Coelenterata, Anthozoa). Trans. Am. Micros. Soc. 99: 426-438. GLIDER, W. V. 1983. The biology of the association of Symbiodinium microadriaticum with Aiptasia pallida: An anemone-alga symbiosis. Ph.D. Thesis, University of Nebraska, Lincoln. GLYNN, P. W. 1983. Extensive 'bleaching' and death of reef corals on the Pacific coast of Panama. Environ. Conserv. 10: 149-154. GLYNN, P. W. 1984. Widespread coral mortality and the 1982/1983 El Nino warming event. Environ. Conserv. 11: 133-146. GOREAU, T. F. 1964. Mass expulsion of zooxanthellae from Jamaican reef communities after Hurricane Flora. Science 145: 383-386. HARNISCHFEGER, G., AND H. JARRY. 1982. Changes in the fluorescence emission spectrum ofChlorella emersonii induced by cold treatment; a possible regulative feature of energy uptake. Z. Natur- forsch. 37: 448-451. HARRIOTT, V. J. 1985. Mortality rates of scleractinian corals before and during a mass bleaching event. Mar. Ecol. Prog. Ser. 21: 81-88. HOHMAN, T. C., P. L. MCNEIL, AND L. MUSCATINE. 1982. Phagosome-lysosome fusion inhibited by algal symbionts of Hydra viridis. J. Cell Biol. 94: 56-63. JAAP, W. C. 1979. Observations on zooxanthellae expulsion at Middle Sambo Reef, Florida Keys. Bull. Mar. Sci. 29:414-422. JAAP, W. C. 1985. An epidemic zooxanthellae expulsion during 1983 in the lower Florida Keys coral reefs: Hyperthermic ecology. Proc. Fifth Int. Coral Reef Cong. 6: 143-148. KEVIN, K. M., AND R. C. L. HUDSON. 1979. The role of zooxanthellae in the hermatypic coral Plesiastrea un'illei (Milne-Edwards and Haime) from cold waters. /. Exp. Mar. Biol. Ecol. 36: 157-1 70. LASKER, H. R., E. C. PETERS, AND M. A. COFFROTH. 1984. Bleaching of reef coelenterates in the San Bias Islands, Panama. Coral Reefs 3: 183-190. EXOCYTOSIS OF ALGAE 263 McAULEY, P. J., AND D. C. SMITH. 1982. The green hydra symbiosis. VII. Conservation of the host cell habitat by the symbiotic algae. Proc. R. Soc. Land. B 216: 41 5-426. MCDOWELL, E. M. 1978. Fixation and processing. Pp. 1 13-139 in Diagnostic Electron Microscopy, B. Trump and S. Jones, eds. John Wiley, New York. MELKONIAN, M., K.-H. KROGER, AND K.-G. MARQUARDT. 1980. Cell shape and microtubules in zoo- spores of the green alga Chlorosarc inopsis gelatinosa (Chlorosarcinales): Effects of low tempera- ture. Protoplasma 104: 283-293. MULLER-PARKER, G. 1984. Photosynthesis-irradiance responses and photosynthetic periodicity in the sea anemone Aiptas ia pulchella and its zooxanthellae. Mar. Biol. 82: 225-232. MUSCATINE, L. 1980. Productivity of zooxanthellae. Pp. 381-402 in Primary Productivity in the Sea. P. G. Falkowski, ed. Plenum Press. New York. MUSCATINE, L., C. B. COOK, R. L. PARDY, AND R. R. POOL. 1975. Uptake, recognition, and maintenance of symbiotic Chlorella by Hydra viridis. Symp. Soc. Exp. Biol. 29: 175-203. MUSCATINE, L., L. R. MCCLOSKEY, AND Y. LOYA. 1985. A comparison of the growth rates of zooxanthel- lae and animal tissue in the Red Sea coral Stvlophora pistillata. Proc. Fifth Int. Coral Reef Symp. 6: 119-123. O'BRIEN, T. L., AND C. R. WYTTENBACH. 1 980. Some effects of temperature on the symbiotic association between zoochlorellae (Chlorophyceae) and the sea anemone Anthopleura xanthogrammica. Tram. Am. Microsc. Soc. 99: 221-225. PORTER, J., J. F. BATTEY, AND G. JASON SMITH. 1 982. Perturbation and change in coral reef communities. Proc. Natl. Acad. Sci. 79: 1678-1681. PUPPIONE, D. L., S. T. KUNITAKE, R. L. HAMILTON, M. L. PHILLIPS, V. N. SCHUMAKER, AND L. D. DAVIS. 1982. Characterization of unusual intermediate density lipoproteins. J. Lipid Res. 23: 263-290. REIMER, S. A. 1971. Observations on the relationships between several species of tropical zoanthids (Zoan- thidea, Coelenterata) and their zooxanthellae. J. Exp. Mar. Biol. Ecol. 1: 207-2 14. ROGERS, C. S. 1979. The effects of shading on coral reef structure and function. J. Exp. Afar. Biol. Ecol. 41:269-288. SCHOENBERG, D. A., AND R. K. TRENCH. 1980. Genetic variation in Symbiodinium (=Gymnodinium) microadriaticum Feudenthal, and specificity in its symbiosis with marine invertebrates. II. Mor- phological variation in Symbiodinium microariaticurn. Proc. R. Soc. Lond. B 207: 429-444. SPEARMAN, T. N., AND F. R. BUTCHER. 1982. Calcium and exocytosis — the rat parotid as a model. Pp. 1 2 1 - 1 46 in The Role of Calcium in Biological Systems. Vol. II, L. J. Anghileri and A. M. Tuffet- Anghileri, eds. CRC Press, Boca Raton, Florida. STEELE, R. D. 1976. Light intensity as a factor in the regulation of the density of symbiotic zooxanthellae in Aiptasia tagetes. J. Zool. Lond. 179: 387-405. STEELE, R. D. 1977. The significance of zooxanthellae-containing pellets extruded by sea anemones. Bull. Mar. Sci. 27:591-594. STEEN, R. G. 1985. In vivo nuclear magnetic resonance analysis of a symbiotic sea anemone: Host bioener- getics and interactions with the symbiont. Ph.D. Thesis, University of California, Los Angeles. STRICKLAND, J. D. H., ANDT. R. PARSONS. 1965.. 4 Practical Handbook of Sea Water Analysis. Bull. 167 Fish. Res. Bd. Canada, 2nd Edition, Queen's Printer. Ottawa, Canada. WILKERSON, F. P., G. MULLER-PARKER, AND L. MUSCATINE. 1983. Temporal patterns of cell division in natural populations of endosymbiotic algae. Limnol. Oceanogr. 28: 1009-1014. YONGE, C. M., AND A. G. NICHOLLS. 193 1 . Studies on the physiology of corals. V. The effect of starvation in light and in darkness on the relationship between corals and zooxanthellae. Sci. Rep. Great Barrier ReefExped. 1 928-1929 1:177-211. \ • •'- .-i •'•- ••"•""'} - \ CONTENTS •\ DEVELOPMENT AND REPRODUCTION BOROWSKY, BETTY The effects of embryos of different developmental stages on reproduc- tive behavior and physiology in brooding females of the amphipod crus- tacean Gammarus palustris 155 CHORNESKY, ELIZABETH A., AND ESTHER C. PETERS Sexual reproduction and colony growth in the scleractinian coral Porites astreoides 161 ECOLOGY AND EVOLUTION AYRE, DAVID J. The formation of clonal territories in experimental populations of the sea anemone Actinia tenebrosa : . . . . 178 FORBES, THOMAS L., AND GLENN R. LOPEZ The allometry of deposit feeding in Capitella species I (Polychaeta: Cap- itellidae): The role of temperature and pellet weight in the control of egestion 187 FUJII, HARUHIKO The predominance of clones in populations of the sea anemone An- thopleura asiatica (Uchida) 202 MlYAZAKI, JUN-ICHI, KOICHI SEKIGUCHI, AND TAMIO HlRABAYASHI Application of an improved method of two-dimensional electrophoresis to the systematic study of horseshoe crabs 212 WENNER, ADRIAN M., DAVID M. HUBBARD, JENIFER DUGAN, JULIE SHOF- FNER, AND KERRY JELLISON Egg production by sand crabs (Emerita analogd) as a function of size and year class (Decapoda, Hippidae) 225 , ~j ^ : , ^ « GENERAL BIOLOGY ANDERSON, ROBERT S. Polykaryon formation by Mercenaria mercenaria hemocytes 236 PHYSIOLOGY STEEN, R. GRANT, AND L. MUSCATINE Low temperature evokes rapid exocytosis of symbiotic algae by a sea anemone 246 Volume 172 Number 3 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board RUSSELL F. DOOLITTLE, University of California at San Diego WILLIAM R. ECKBERG, Howard University ROBERT D. GOLDMAN, Northwestern University C. K. GOVIND, Scarborough Campus, University of Toronto JUDITH P. GRASSLE, Marine Biological Laboratory MICHAEL J. GREENBERG, C. V. Whitney Marine Laboratory, University of Florida MAUREEN R. HANSON, Cornell University JOHN E. HOBBIE, Marine Biological Laboratory LIONEL JAFFE, Marine Biological Laboratory HOLGER W. JANNASCH, Woods Hole Oceanographic Institution WILLIAM R. JEFFERY, University of Texas at Austin GEORGE M. LANGFORD, University of North Carolina at Chapel Hill GEORGE D. PAPPAS, University of Illinois at Chicago SIDNEY K. PIERCE, University of Maryland HERBERT SCHUEL, State University of New York at Buffalo VIRGINIA L. SCOFIELD, University of California at Los Angeles School of Medicine LAWRENCE B. SLOBODKIN, State University of New York at Stony Brook JOHN D. STRANDBERG, Johns Hopkins University DONALD P. WOLF, Oregon Regional Primate Center SEYMOUR ZIGMAN, University of Rochester Vr Editor: CHARLES B. METZ, University of Miami JUNE, 1987 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA Marine Biological Laboratory LIBRARY jUL 7 1987 Woods Hole, Mass. 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 THE BIOLOGICAL BULLETIN, Marine Bio- logical Laboratory, Woods Hole, Massachusetts. Single numbers, $20.00. Subscription per volume (three issues), $50.00 ($100.00 per year for six issues). Communications relative to manuscripts should be sent to Dr. Charles B. Metz, Editor, or Pamela Clapp, Assistant Editor, at the Marine Biological Laboratory, Woods Hole. Massachusetts 02543. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Copyright © 1987, 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). Very short papers (less than 10 manuscript pages including tables, figures, and bibliography) will be published in a separate section entitled "Short Reports." A limited number of solicited review papers may be accepted after formal review. A paper will usually appear within four months after its acceptance. The Editorial Board requests that manuscripts conform to the requirements set below; those manu- scripts which 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 double spacing (including figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper, 8'/2 by 1 1 inches. Manuscripts should be proofread carefully and errors 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 Editors Style Manual, 4th Edition (Council of Biology Editors, 1 978) 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 addi- tion, authors should supply a list of words and phrases under which the article should be indexed. 2. Figures. Figures should be no larger than 8'/2 by 1 1 inches. The dimensions of the printed page, 5 by 7% inches, should be kept in mind in preparing figures for publication. We recommend that figures be about 1 '/2 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 generally 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 photographic 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. 3. 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 information 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, acknowledgments 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. 4. A condensed title or running head of no more than 35 letters and spaces should appear at the top of the title page. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more than two authors cited as Jones et al., 1980. Personal communications and material 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 LITERATURE 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 include complete titles and inclusive pagination. Journal abbreviations should normally follow those of the U. S. A. Standards Institute (USASI), as adopted by BIOLOGICAL ABSTRACTS and CHEMICAL AB- STRACTS, 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). For- eign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PERIODICALS, may find a booklet published by the Biological Council of the U.K. (obtainable from the Institute of Biology, 4 1 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 additional 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. Physio! . NOT /. cell. comp. Physiol. ) C. All abbreviated components must be followed by a period, 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 abbrevi- ations invented by the author. For example, use Rit I 'isindajjelags Islendinga without abbreviation. F. All single word journal titles in full (e.g. Veliger, Ecology. 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 preferred forms rather than WORLD LIST or USASI usage (e.g. Nature, Science, Evolution NOT Nature, Land., Science, N.Y.; Evolution, Lancaster, Pa.) 6. Reprints, charges. Authors will be charged the excess over $100 of the total of (a) $30 for each printed page beyond 1 5, (b) $30 for each table, (c) $ 1 5 for each formula more complex than a single line with simple subscripts or superscripts, and (d) $ 1 5 for each figure, with figures on a single plate all consid- ered one figure and parts of a single figure on separate sheets considered separate figures. Reprints may be ordered at time of publication and normally will be delivered about two to three months after the issue date. Authors (or delegates or 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' creditors' errors). >• ». i • ^ THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board RUSSELL F. DOOLITTLE. University of California at San Diego WILLIAM R. ECKBERG, Howard University ROBERT D. GOLDMAN, Northwestern University C. K. GOVIND, Scarborough Campus, University ofToronto JUDITH P. GRASSLE, Marine Biological Laboratory MICHAEL J. GREENBERG, C. V. Whitney Marine Laboratory, University of Florida MAUREEN R. HANSON, Cornell University JOHN E. HOBBIE, Marine Biological Laboratory LIONEL JAFFE, Marine Biological Laboratory HOLGER W. JANNASCH, Woods Hole Oceanographic Institution WILLIAM R. JEFFERY, University of Texas at Austin GEORGE M. LANGFORD, University of North Carolina at Chapel Hill GEORGE D. PAPPAS, University of Illinois at Chicago SIDNEY K. PIERCE, University of Maryland HERBERT SCHUEL, State University of New York at Buffalo VIRGINIA L. SCOFIELD, University of California at Los Angeles School of Medicine LAWRENCE B. SLOBODK.IN, State University of New York at Stony Brook JOHN D. STRANDBERG, Johns Hopkins University DONALD P. WOLF, Oregon Regional Primate Center SEYMOUR ZIGMAN, University of Rochester Editor: CHARLES B. METZ, University of Miami JUNE, 1987 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. Marine Biological Laboratory LIBRARY JUL 71987 Woods Hole, Mass. The BIOLOGICAL BULLETIN is issued six times a year at the Lancaster Press, Inc., Prince and Lemon Streets, Lancaster, Penn- sylvania. Subscriptions and similar matter should be addressed to The Biological Bulletin, Marine Biological Laboratory, Woods Hole, Massachusetts. Single numbers, $20.00. Subscription per volume (three issues), $50.00 ($100.00 per year for six issues). Communications relative to manuscripts should be sent to Dr. Charles B. Metz, Editor, or Pamela Clapp, Assistant Editor, Marine Biological Laboratory, Woods Hole, Massachusetts 02543. THE BIOLOGICAL BULLETIN (ISSN 0006-3 185) POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, MA 02543. Second-class postage paid at Woods Hole, MA, and additional mailing offices. LANCASTER PRESS, INC., LANCASTER. PA IV CONTENTS No. 1, FEBRUARY 1987 BEHAVIOR GLEESON, RICHARD A. MICHAEL A. ADAMS, AND AMOS B. SMITH, III Hormonal modulation of pheromone-mediated behavior in a crus- tacean ! ZIMMER-FAUST, RICHARD K. Crustacean chemical perception: towards a theory on optimal chemore- ception 10 DEVELOPMENTAL ECOLOGY MCEDWARD, LARRY R., AND RICHARD R. STRATHMANN The body plan of the cyphonautes larva of bryozoans prevents high clearance rates: comparison with the pluteus and a growth model .... 30 GENERAL BIOLOGY MULLER-PARKER, G., AND R. L. PARDY Response of green hydra to feeding and starvation at four irradiances . . 46 VOGEL, STEVEN Flow-assisted mantle cavity refilling in jetting squid 61 PHYSIOLOGY ENGEL, DAVID W. Metal regulation and molting in the blue crab, Callinectes sapidus: cop- per, zinc, and metallothionein 69 SHORT REPORT SNYDER, KEITH, AND STACIA MOFFETT Loss and recovery of locomotor behavior after CNS lesions in the snail Melampus bidentatus 83 FIRST INTERNATIONAL WORKSHOP ON CIGUATERA Papers from the first International Workshop on Ciguatera held on 5 and 6 June 1986 at the New England Biolabs in Beverly, Massachusetts. Workshop sponsored by the New England Biolabs Foundation and the M.I.T. Sea Grant Program. INTRODUCTION ANDERSON, DONALD M., AND PHILIP S. LOBEL The continuing enigma of ciguatera 89 ORGANISMS AND THEIR ENVIRONMENT DURAND-CLEMENT, M. Study of production and toxicity of cultured Gambler discus toxicus . . 108 vi CONTENTS PHARMACOLOGY AND CHEMISTRY TACHIBANA, KAZUO, MANABU NUKINA, YONG-GOE JOH, AND PAUL J. SCHEUER Recent developments in the molecular structure of ciguatoxin 122 YASUMOTO, TAKESHI, NAOKO SEINO, YASUTAKA MURAKAMI, AND MICHIO MURATA Toxins produced by benthic dinoflagellates 128 OHIZUMI, YASUSHI Pharmacological actions of the marine toxins ciguatoxin and maito- toxin isolated from poisonous fish 1 32 ASSAY FOR CIGUATOXINS BAGNIS, R., M. BARSINAS, C. PRIEUR, A. POMPON, E. CHUNGUE, AND A. M. LEGRAND The use of the mosquito bioassay for determining the toxicity to man of ciguateric fish 137 HOKAMA, Y., L. K. SHIRAI, L. M. IWAMOTO, M. N. KOBAYASHI, C. S. GOTO, AND L. K. NAKAGAWA Assessment of a rapid enzyme immunoassay stick test for the detection of ciguatoxin and related polyether toxins in fish tissues 144 No. 2, APRIL 1987 DEVELOPMENT AND REPRODUCTION BOROWSKY, BETTY The effects of embryos of different developmental stages on reproduc- tive behavior and physiology in brooding females of the amphipod crus- tacean Gammarus palustris 155 CHORNESKY, ELIZABETH A., AND ESTHER C. PETERS Sexual reproduction and colony growth in the scleractinian coral Porites astreoides 161 ECOLOGY AND EVOLUTION AYRE, DAVID J. The formation of clonal territories in experimental populations of the sea anemone Actinia tenebrosa 178 FORBES, THOMAS L., AND GLENN R. LOPEZ The allometry of deposit feeding in Capitella species I (Polychaeta: Cap- itellidae): The role of temperature and pellet weight in the control of egestion 187 FUJII, HARUHIKO The predominance of clones in populations of the sea anemone An- thopleura asiatica (Uchida) 202 MlYAZAKI, JUN-ICHI, KOICHI SEKIGUCHI, AND TAMIO HlRABAYASHI Application of an improved method of two-dimensional electrophoresis to the systematic study of horseshoe crabs 212 WENNER, ADRIAN M., DAVID M. HUBBARD, JENIFER DUGAN, JULIE SHOFF- NER, AND KERRY JELLISON Egg production by sand crabs (Emerita analoga) as a function of size and year class (Decapoda, Hippidae) 225 CONTENTS vii GENERAL BIOLOGY ANDERSON, ROBERT S. Polykaryon formation by Mercenaria mercenaria hemocytes 236 PHYSIOLOGY STEEN, R. GRANT, AND L. MUSCATINE Low temperature evokes rapid exocytosis of symbiotic algae by a sea anemone 246 No. 3, JUNE 1987 MARINE BIOLOGICAL LABORATORY — CALENDAR OF CENTENNIAL EVENTS 265 INVITED ARTICLE DOOLITTLE, RUSSELL F. The evolution of the vertebrate plasma proteins 269 DEVELOPMENT AND REPRODUCTION DINEEN, JOSEPH F., JR. The larval stages of Lithotrya dorsalis (Ellis & Solander, 1786): a bur- rowing thoracican barnacle 284 SAFRANEK, Louis, AND CARROLL M. WILLIAMS Studies of the ecdysiotropic activity of juvenile hormone in pupae of the tobacco hornworm, Manduca sexta 299 ECOLOGY AND EVOLUTION PETRAITIS, PETER S. Immobilization of the predatory gastropod, Nucella lapillus, by its prey, Mytilus edulis 307 GENERAL BIOLOGY SMITH, PETER R., EDWARD E. RUPPERT, AND STEPHEN L. GARDINER A deuterostome-like nephridium in the mitraria larva of Owenia fusi- formis (Polychaeta, Annelida) 315 TELFORD, MALCOLM, RICH Mooi, AND ANTONY S. HAROLD Feeding activities of two species ofClypeaster (Echinoides, Clypeaster- oida): further evidence of clypeasteroid resource partitioning 324 PHYSIOLOGY LlM, A. L. L., AND Y. K. IP Membrane transport of phosphate by Hymenolepis diminuta 337 MELLON, DEFOREST, JR., AND ELIZABETH GREER Induction of precocious molting and claw transformation in alpheid shrimps by exogenous 20-hydroxyecdysone 350 viii CONTENTS SHORT REPORTS MARCUM, JAMES A. A trypsin inhibitor from the coelomic fluid of the sea star Asterias forbesi 357 O'SULLIVAN, J. B., R. R. McCONNAUGHEY, AND M. E. HUBER A blood-sucking snail: the Cooper's nutmeg, Cancel/aria cooperi GAAB, parasitizes the California electric ray, Torpedo californica AYRES 362 INDEX TO VOLUME 172 . 367 Reference: Bio/. Bull. 172: 265-268. (June, 1987) MARINE BIOLOGICAL LABORATORY CALENDAR OF CENTENNIAL EVENTS The Marine Biological Laboratory was founded in 1888. The centennial of this founding will be recognized by the laboratory through a series of special events. This series will include special lectures, seminars, symposia, and workshops. The centen- nial celebration will begin in the summer of 1987 and extend through 1988. These events should encompass many fields of biology and provide "state-of-the- art" knowledge about them. Collectively, they should be of general interest to biolo- gists. Therefore, a cumulative calendar of these events will be published in The Bio- logical Bulletin beginning with this issue. — The Editor GAMETE DIALOGUE IN FERTILIZATION: FROM SEA URCHIN TO HUMAN* A MARINE BIOLOGICAL LABORATORY CENTENARY SYMPOSIUM IN HONOR OF ALBERTO MONROY 23-25 August 1987 Marine Biological Laboratory Woods Hole, Massachusetts 02543 Sunday, August 23 12 noon-7 pm: Registration. Informal Reception 5-7 pm. Monday, August 24 8:45-9:00 am: Opening remarks by C. B. Metz (U. Miami). 9:00-12:00 noon: Session I: Responses of sperm to the egg's environment. G. S. Kopf, Chairperson. Speakers: G. S. Kopf (U. Pennsylvania) M. Hoshi (Tokyo Institute of Technology) D. L. Garbers(Vanderbilt University) D. P. Wolf (Oregon Regional Primate Research Center) S. Meizel (U. California, Davis) 1 :30-4:30 pm: Session II: Reactions comprising sperm/egg recognition and binding. B. Shapiro, Chairperson. Speakers: B. Shapiro (U. Washington) R. De Santis (Stazione Zoologica, Napoli) E. Carroll (U. California, Riverside) B. Storey (U. Pennsylvania) N. Cross (U. California, Davis) Program subject to change. 265 266 MBL CENTENNIAL CALENDAR Tuesday, August 25 9:00-12 noon: Session III: Reactions comprising sperm/egg plasma membrane fu- sion. D. Myles, Chairperson. Speakers: D. Myles (U. Connecticut) T. Chambers/D. McCulloh (U. Miami) J. Hedrick (U. California, Davis) R. Harrison (AFRC Cambridge) P. Saling (Duke University) 1 :30-4:30 pm: Session IV: Egg responses to sperm/egg fusion. G. Schatten, Chairper- son. Speakers: G. Schatten (U. Wisconsin) L. Jaffe (U. Connecticut) M. Whitaker(U. London) R. M. Schultz (U. Pennsylvania) G. Guidice (U. Palermo) 4:30-5:00 pm: Summation of meeting by D. Epel (Hopkins Marine Station). 5:00-5:15 pm: Closing remarks by L. Mastroianni (U. Pennsylvania). The Alberto Monroy memorial symposium is sponsored by the National Institute of Child Health and Human Development, The Rockefeller Foundation, and the Marine Biological Laboratory. The symposium has been coordinated by Luigi Mastroianni, Jr., Chair; Jerry L. Hedrick; Gregorys. Kopf; and Bayard T. Storey. For further information please write: Ms. Debbie Coffin Division of Reproductive Biology Medical Labs 3 1 1 Department of Obstetrics and Gynecology University of Pennsylvania Philadelphia, PA 19104-6080 THE SECOND SYMPOSIUM ON REGULATORY ACTIONS OF GROWTH FACTORS IN REPRODUCTION AND EARLY DEVELOPMENT A SYMPOSIUM HELD IN CELEBRATION OF THE CENTENNIALS OF THE NATIONAL INSTITUTES OF HEALTH (1887-1987) AND THE MARINE BIOLOGICAL LABORATORY (1888-1988) 1-2 September 1987 Marine Biological Laboratory Woods Hole. Massachusetts 02543 Participants in the Symposium will review and assess the current state of knowledge of the molecular mechanisms of the regulation of reproductive and early develop- mental events. The Symposium will focus on the interactions and interrelationships of growth factors, reproductive hormones, and their receptors in the regulation of cell function and responsiveness to environmental signals. Autocrine and paracrine modes of growth factor action in the reproductive system or during early develop- ment of the embryo will be discussed. In addition, consideration will be given to the extent and possible significance of extensive homologous sequences in the structures of relevant growth factors, oncogenes, reproductive hormones, and receptors. MBL CENTENNIAL CALENDAR 267 Interested scientists are invited to submit a 200-word abstract of work to be presented. The deadline for abstract submission is 1 July 1987 The Symposium is sponsored by the Reproductive Sciences Branch, Center for Popu- lation Research, National Institute of Child Health and Human Development. For further information, please call or write to: Dr. William A. Sadler Chief, Reproductive Sciences Branch National Institute of Child Health and Human Development 7910 Woodmont Avenue, Room 7C33 Bethesda, Maryland 20892 Telephone: (30 1)496-65 15 OPTICAL APPROACHES TO THE DYNAMICS OF CELLULAR MOTILITY A MARINE BIOLOGICAL LABORATORY CENTENARY SYMPOSIUM IN HONOR OF ROBERT D. ALLEN 5-8 October 1987 Marine Biological Laboratory Woods Hole. Massachusetts 02543 Monday, October 5 8:00 pm: Wine and cheese party and demonstrations of optical equipment by manu- facturers. Tuesday, October 6 Morning: Introductory remarks by N. Kamiya (Okazaki) Morning session: Optical techniques in cell biology. D. L. Taylor and S. Inoue, Chair- persons. Speakers: G. Nomarski (Antony, France) S. Inoue (Marine Biological Laboratory) D. L. Taylor (Carnegie Mellon) D. A. Agard (U. California Medical School, San Francisco) T. Yanagida (U. Osaka) Afternoon session: Amoeboid movement and cytoplasmic streaming. J. Condeelis, Chairperson. Speakers: N. Kamiya (Okazaki) J. Condeelis (Albert Einstein) M. Schliwa (U. California, Berkeley) J. Hartwig (Mass. General, Harvard University) G. Oster (U. California, Berkeley) Evening session: Talks and demonstrations by equipment manfuacturers (Swope Center). Speakers: R. Wicks (PMI Corp.) 268 MBL CENTENNIAL CALENDAR Speakers Zeiss representative (cont 'd): Remainder of program to be arranged Wednesday, October 7 Morning session: Mitosis. L. I. Rebhun, Chairperson. Speakers: Y. Hiramoto (Tokyo) E. W. Salmon (U. North Carolina, Chapel Hill) A. Bajer (U. Oregon, Eugene) L. I. Rebhun/R. Palazzo (U. Virginia) W. C. Cande (U. California, Berkeley) Afternoon session: Posters. Evening session: Motility in other systems. K. Edds, Chairperson. Speakers: N. S. Allen (Wake Forest) R. Hard (SUNY, Buffalo) J. Hayden (Sienna College) J. Lafountain (SUNY, Buffalo) C. Izzard (SUNY, Albany) Thursday, October 8 Morning session: Microtubule dynamics. R. Sloboda, Chairperson. Speakers: R. Sloboda/S. Gilbert ( Dartmouth College) J. Travis (Vassar College)/S. Bowser (NY State Dept. of Health, Albany) T. Tchen (Wayne State) P. Satir (Albert Einstein) H. Hotani (Kyoto University) Afternoon session: Axoplasmic transport. D. Weiss, Chairperson. Speakers: D. Weiss (Munich)/G. Langford (U. North Carolina, Chapel Hill) R. Smith (Edmonton, Canada) A. Breuer (Cleveland Clinic) B. Schnapp (NIH/Marine Biological Laboratory) R. Weisenberg (Temple University) Evening: Dinner and concert. Abstracts may be submitted for inclusion in the poster session and will be published in Cell Motility and Cytoskeleton. Information on preparation of abstracts will be sent with application forms for registration. Final date for abstract submission will be 1 August 1987. Registration and Housing: The registration fee is $75. Housing is limited and will be assigned in order of receipt of application. Deadline for housing and meal plan is 1 August 1987. Forms and information may be obtained by writing: Robert Day Allen Symposium PO Box 477 Woods Hole, MA 02543 Reference: Biol. Bull. 172: 269-283. (June, 1987) THE EVOLUTION OF THE VERTEBRATE PLASMA PROTEINS* RUSSELL F. DOOLITTLE Department of Chemistry, University of California, San Diego, La Jolla, California 92093 ABSTRACT The appearance of vertebrate animals some 450 million years ago was heralded by the invention of a number of new proteins that are found in the blood plasma, including albumin and other transport proteins, the non-enzyme blood clotting and complement proteins, and a host of protease inhibitors. Comparisons of present day amino acid sequences from various species allow us to look back to how and when many of these proteins originated. INTRODUCTION It is particularly gratifying for me to present this Friday Evening Lecture on the evolution of vertebrate plasma proteins, since it is a project I began many years ago here at the MBL. At the time I was a graduate student at Harvard in a laboratory devoted to the study of human blood proteins, and, after having spent a summer in Woods Hole working in the area of comparative physiology, it was only natural that I should consider a comparative study of blood proteins from other vertebrates. The question immediately presented itself as to where, in the biological world, the charac- teristic proteins of blood are found, and from this gradually emerged the more pro- found questions as to when and how they evolved. A typical mammalian blood plasma — the fluid in which blood cells are sus- pended— contains upwards of 600 protein components, as can be shown by high resolution two-dimensional electrophoresis (Anderson et a!., 1984). Some of these materials are doubtless only altered forms of particular proteins that result from vari- ous processing events; still, the multitude of individual types is impressive. In mam- mals, a single protein, albumin, accounts for half the protein mass of the plasma, however, and five more — haptoglobin, fibrinogen, transferrin, a-1-antitrypsin and a-2-macroglobulin — make up another quarter; lipoproteins and immunoglobulins constitute another large fraction (Table I). This evening I am going to concentrate on the origins of some of these most abundant proteins, although I will touch upon a few of the less abundant ones in passing. Regarding the phyletic distribution of the plasma proteins, we can state flatly that many of these proteins, but certainly not all of them, are found in the blood plasmas of all the major classes of vertebrates, from fish to mammals. It has been alleged by some that albumin is absent from the blood of fish, although some workers have reported that it is present in assorted teleosts in small amounts (for a review, see Doolittle, 1984). There is general agreement that amphibians, reptiles, and birds all have albumins that are similar to the mammalian type. Only a few of the vertebrate plasma proteins have counterparts among the invertebrates, and, as we shall see, albu- Received 3 March 1987; accepted 23 March 1987. * This article is reproduced from the Friday Evening Lecture delivered by Dr. Doolittle at the Marine Biological Laboratory on 8 August 1 986. 269 270 R. F. DOOLITTLE TABLE I Most abundant proteins found in mammalian blood plasma Protein Abundance* (gm/liter) Internal duplications Carbohydrate Albumin 45 Yes No Immunoglobulins 15 Yes Yes Lipoproteins 10 Yes Yes Haptoglobin 6 Yes Yes Fibrinogen 3.5 Yes Yes Transferrin 3 Yes Yes a,-Antitrypsin 3 No Yes a2-Macroglobulin 2.5 No Yes 87.5 * These eight (sets of) proteins account for more than 95% of the mass of plasma proteins in mammals. The values given are approximate and vary somewhat from species to species and among individual species. min is not among them. We'll return to the matter of occurrence shortly, but there are a few definitional matters to be settled first. FUNCTIONS OF THE PLASMA PROTEINS The plasma proteins can be grouped functionally. Quite apart from albumin, there are numerous transport proteins, including transferrin (iron transport) and the lipoproteins. Indeed, any potential metabolite that is poorly soluble in plasma is likely to have a corresponding protein for rendering it soluble so that it can be transported from one place to another. In addition to the transport proteins, the plasma contains numerous polypeptide hormones and growth factors, many in precursor forms awaiting proteolytic activa- tion upon the appropriate signal. There are also numerous defense proteins, including the immunoglobulins and the associated complement system, and there is a set of proteins whose function is to prevent loss of the blood: the coagulation proteins. The latter is a special interest of mine, and I will be giving it an undue amount of attention this evening. The complement and blood coagulation systems, and some other schemes in the plasma, depend upon a complex plan of proteolytic activation for their normal opera- tion. Indeed, much of the extracellular way of life in multicellular creatures is regu- lated by a judicious and limited proteolysis. Whereas intracellular regulation depends largely on ATP-dependent phosphorylation (kinases) and subsequent hydrolysis (phosphatases), extracellular regulation derives its energy potential at the time of pro- tein biosynthesis. It is then "leaked out" at some later time by a gradual and succes- sive proteolysis. As a consequence, blood plasma is rich with proteases and protease precursors. There are very many related serine proteases involved, and some sulfhydryl proteases. In order to keep the system stable, there are large amounts of general and specific protease inhibitors. Two of these, a-2-macroglobulin and «-l-antitrypsin, appear on our list of the most abundant plasma proteins (Table I), and they will figure heavily in our evolutionary discussion. How NEW PROTEINS ARE FORMED At this point let me make some sweeping statements about the evolution of pro- teins in general. First, most new proteins evolve from old proteins as a result of gene EVOLUTION OF VERTEBRATE PLASMA PROTEINS 271 TABLE II Divergence times of some major vertebrate groupings* Millions of Vertebrate groupings years ago Old World monkeys/humans 28 ± 2 Artiodactyls/primates 80 ±10 Primates/rodents 80 ±10 Birds/mammals 200 ± 20 Reptiles/mammals 200 ± 20 Amphibia/mammals 320 ± 30 Bony fish/mammals 375 ± 35 Cyclostomes/mammals 450 ± 50 * Groupings are based on the fossil record. duplication. DNA has a propensity for duplicating itself under all sorts of conditions, and segments of all sizes are constantly being randomly and tandemly repeated. There are many known mechanisms for this excessive internal duplication, including unequal and homologous crossing over and other genetic delinquencies, but the im- portant thing for the present discussion is that these duplications are observed in all living forms. If the region of duplicated DNA is short and inside the boundaries of the gene for a particular protein, then the protein is lengthened by the process, some portion of it now showing up twice or more. If the DNA duplication encompasses the entire gene — starts and stops included — then for a brief moment two genes will exist where there used to be one, and two separate products may be issued. Only one of these need be subject to natural selection, of course, and the other will be free to change, mostly as a result of errant base substitutions leading to single amino acid replacements. Usually such duplicons are destined for the genomic scrap-pile, since sooner or later a mutation will occur that will decommission the whole enterprise, but occasionally some new function — with benefit to the organism — will be encoun- tered by chance. And with that, a bona fide "new" protein is eligible for the registry. It is related to the "old protein" of course, and depending on the rate of amino acid replacement along their divergent courses, the two will, for a long time, be recogniz- able merely on the basis of their amino acid sequences alone. By "a long time" I am implying time frames of a 100 million to several billion years (Table II). SETTING MOLECULAR CLOCKS There are two independent means at our disposal for finding when a "new pro- tein" made its appearance during evolution. We have already alluded to one of these, which simply amounts to surveying the biological community to see which creatures have the protein in question. If all mammals have some protein — e.g., lactalbumin, which is found in milk — and none of the other classes of vertebrates have this protein, then we can presume that lactalbumin was "invented" some time around or just before the appearance of mammals, about 200 million years ago. If all vertebrates except fish have some particular protein, then we would expect that protein must have evolved about 375 million years ago, since that is when fish and the other verte- brates last shared a common ancestor (Table II). The second way to gauge when a gene duplication leading to a new protein oc- curred, involves the comparison of amino acid sequences and the use of the "molecu- lar clock." The method can also be used to find when an internal duplication leading 272 R. F. DOOLITTLE HEMOGLOBIN PRESENT -200 -400 -600 -800 -1000 Primates Bony Fishes Jawless Fishes Invertebrates Ol o c in s_ s o. o. TPA — v-ooo- UK -ooo- FX — oo- FIGURE 8. Schematic comparison of several vertebrate proteins that contain homologous "modules" resulting from exon shuffling. EGFP, epidermal growth factor precursor; LDLR, low-density lipoprotein receptor. FN, fibronectin; C9, complement component 9; TPA, tissue plasminogen activator; UK, uroki- nase; FX, blood clotting Factor X. The different types of peptide segments are denoted by the different symbols, asterisks (*) are the active sites of serine proteases, and "ms" marks membrane-spanning segments (from Doolittle, 1985a). tion, complete versions of which can give rise to new, modified gene products, while partial or incomplete versions give rise to elongated, internally duplicated sequences. The second proliferative force is exon shuffling, an operation that has led to a rela- tively small set of polypeptide segments being present in many different proteins (Ta- ble IV). Most of these segments likely act as recognition units for binding the proteins TABLE IV Some animal proteins that have sequence segments in common A. EOF- type: Epidermal growth factor precursor Tumor growth factors Low density lipoprotein receptor Factor IX Factor X Protein C Tissue plasminogen activator Urokinase Complement C9 Notch protein (Drosophila) Lin-12(Nematode) Thrombospondin B. Fibronectin "finger": Fibronectin Tissue plasminogen activator C. C9-type: Complement C9 Low density lipoprotein receptor (Fibronectin?) Notch (Drosophila) Lin-12(Nematode) D Proprotease ' 'kringle ' ': Plasminogen Tissue plasminogen activator Urokinase Prothrombin E. $-2 type: ft-glycoprotein Complement Factor B Complement Factor H Factor XIII b-chain 282 R. F. DOOLITTLE TABLE V Some known protein families found in vertebrate blood plasma A. Albumin a-fetoprotein (fetal albumin) Vitamin-D-binding protein (group-specific component) B. Immunoglobulins (assorted) /32-microglobulin C. Fibrinogen a chain Fibrinogen p chain Fibrinogen y chain D. Coagulation serine proteases Fibrinolysis serine proteases Complement serine proteases Miscellaneous serine proteases Haptoglobin E. a2-macroglobulin Pregnancy zone protein Complement C3 Complement C4 Complement C5 F. /32-Glycoprotein Complement Factor B (nonenzyme portion) Complement Factor H Factor XIII b-chain G. Retinol-binding protein a,-microglobulin H. Antithrombin III ai-antitrypsin a rantichymotrypsin Angiotensinogen I. Ceruloplasmin Factor V Factor VIII J. Lipid-binding proteins (A, B, C,etc.) K. Kininogens £i-Microglobulin Acute phase proteins Thiol protease inhibitors L. Transthyretin(prealbumin) Glucagon Glycentin M. jtf-Thromboglobulin Platelet factor 4 N. Serum amyloid P-component C-reactive component differentially to various cells. In any case, the net result is that the six hundred electro- phoresis components observed in mammalian plasma will ultimately be grouped into a relatively small number of families (Table V). When did all these events transpire? Doubtless they occurred throughout the his- tory of the vertebrates, but the fact is that much of the inventive action must have been a necessary precondition to the evolution of vertebrates themselves. Many of the sequence comparisons we have discussed this evening indicate that the original gene duplication leading to some of the principal plasma proteins must have occurred among invertebrate ancestors. The point is made further by the fact that the lamprey, one of our most distant vertebrate relatives, has so many plasma proteins in common with mammals. The fact that many of these proteins have not yet been identified among invertebrates or protochordates remains a mystery, but it also provides an opportunity for much more exploration. I have no doubt that that matter will resolve itself as more sequence data are collected, particularly among the protochordates and echinoderms. ACKNOWLEDGMENTS I am grateful to all the many students and associates who have worked with me on these problems over the past 25 years. I thank Karen Anderson for assistance in preparing this manuscript. LITERATURE CITED ANDERSON, N. L., R. P. TRACY, ANDN. G. ANDERSON. 1984. High resolution two-dimensional electro- phoretic mapping of plasma proteins. Pp. 221-270 in The Plasma Proteins, Second ed, Vol. IV, F. W. Putnam, ed. Academic Press, New York. EVOLUTION OF VERTEBRATE PLASMA PROTEINS 283 BANYAI, L., A. VARADI, AND L. PATTHY. 1983. Common evolutionary origin of the fibrin-binding struc- tures of fibronectin and tissue-type plasminogen activator. FEES Lett. 163:37-4 1 . BOHONUS, V. L., R. F. DOOLITTLE, M. PONTES, AND D. D. STRONG. 1986. Complementary DNA se- quence of lamprey fibrinogen beta chain. Biochemistry 25: 65 1 2-65 1 6. BROWN, J. R. 1976. Structural origins of mammalian albumin. Fed. Proc. 35: 2141-2144. DOOLITTLE, R. F. 1979. Protein evolution Pp. 1-1 18 in The Proteins, 2nd ed., Vol. IV, H. Neurath and R. L. Hill, eds. Academic Press, New York. DOOLITTLE, R. F. 1983. Angiotensinogen is related to the antitrypsin-antithrombin-ovalbumin family Science 222: 4\l-4\9. DOOLITTLE, R. F. 1 984. Evolution of the vertebrate plasma proteins. Pp. 3 1 7-360 in The Plasma Proteins. Second ed., Vol. IV, F. W. Putnam, ed. Academic Press, New York. DOOLITTLE, R. F. 1985a. The genealogy of some recently evolved vertebrate proteins. Trends Biochem Sci. 10: 233-237. DOOLITTLE, R. F. 1985b. More homologies among the vertebrate plasma proteins. Biosci. Rep. 5: 877- 884. DOOLITTLE, R. F., J. L. ONCLEY, AND D. M. SURGENOR. 1962. Species differences in the interaction of thrombin and fibrinogen. /. Biol. Chem. 237: 3123-3127. DOOLITTLE, R. F., ANDG. M. FULLER. 1972. Sodium dodecyl sulfate polyacrylamide gel electrophoresis studies on lobster fibrinogen and fibrin. Biochim. Biophys. Ada 263: 805-809. DOOLITTLE, R. F., D. F. FENG, ANDM. S. JOHNSON. 1984. Computer-based characterization of epidermal growth factor precursor. Nature 307: 558-560. EVANS, B. R., K. W. K. WATT, AND R. F. DOOLITTLE. 1985. Characterization and amino terminal se- quence of lamprey transferrin. Fed. Proc. 69: 3799 (abstr). FULLER, G. M., AND R. F. DOOLITTLE. 1 97 1 . Transformation of lobster fibrinogen into fibrin. Biochemis- try 10: 1311-1315. HUEBERS, H. A., A. W. MARTIN, ANDC. A. FINCH. 1984. A mono-sited transferrin from a representative deuterostome: The ascidian Pyura stolonifera. Fed. Proc. 43: 2058 (abstr). KUYAS, C, M. RILEY, J. BUBIS, AND R. F. DOOLITTLE. 1983. Lamprey plasma albumin is a glycoprotein with a molecular weight of 1 75,000. Fed. Proc. 42: 2085 (abstr). MULLER-ESTERL, W., H. FRITZ, J. KELLERMAN, F. LOTTSPEICH, W. MACHLEIDT, AND V. TURK. 1985. Genealogy of mammalian cysteine proteinase inhibitors. Common evolutionary origin of stefins, cystatinsand kininogens. FEBS Lett. 191: 221-226. NY, T., F. ELGH, ANDF. LUND. 1984. The structure of the human tissue-type plasminogen activator gene: Correlation of intron and exon structures to functional and structural domains. Proc. Natl. Acad. Sci. USA 81: 5355-5359. PONTES, M., X. Xu, D. GRAHAM, M. RILEY, AND R. F. DOOLITTLE. 1987. cDNA sequences of two apolipoproteins from the lamprey. Biochemistry 26: 1611-1617. QUIGLEY, J. P., AND P. B. ARMSTRONG. 1985. A homologue of a-2-macroglobulin purified from the hemolymph of the horseshoe crab Limulus polyphemus. J Biol. Chem. 260: 12715-12719. RIPOCHE, J., A. J. DAY, A. C. WILLIS, K. T. BELT, R. D. CAMPBELL, AND R. B. SIM. 1 986. Partial character- ization of human complement factor H by protein and cDNA sequencing: homology with other complement and non-complement proteins. Biosci. Rep. 6: 65-72. SOTTRUP-JENSEN, L., T. M. STEPANIK, T. KRISTENSEN, P. B. LONBLAD, C. M. JONES, D. M. WlERZBICKI, S. MAGNUSSON, H. DOMDEY, R. A. WETSEL, A. LUNDWALL, B. F. TACK, ANDG. H. FEY. 1985. Common evolutionary origin of a-2-macroglobulin and complement components C3 and C4. Proc. Natl. Acad. Sci. USA 82: 9-13. STRONG, D. D., M. MOORE, B. A. COTTRELL, V. L. BOHONUS, M. PONTES, B. EVANS, M. RILEY, AND R. F. DOOLITTLE. 1985. Lamprey fibrinogen gamma chain: Cloning, cDNA sequencing, and general characterization. Biochemistry 24: 92- 10 1 . SUDHOFF, T. C., D. W. RUSSELL, J. L. GOLDSTEIN, M. S. BROWN, R. SANCHEZ-PESCADOR, ANDG. I. BELL. 1 985. Cassette of eight exons shared by genes for LDL receptor and EGF precursor. Science 228: 893-895. VEHAR, G. A., B. KEYT, D. EATON, H. RODRIGUEZ, D. P. O'BRIEN, F. ROTBLAT, H. OPPERMANN, R. KECK, W. I. WOOD, R. N. HARKINS, E. G. D. TUDDENHAM, R. M. LAWN, ANDD. J. CAPON. 1984. Structure of human factor VIII. Nature 312: 337-342. YANG, F., J. L. BRUNE, S. L. NAYLOR, R. L. CUPPLES, K. H. NABERHAUS, AND B. H. BOWMAN. 1985. Human group-specific component (Gc) is a member of the albumin family. Proc. Natl. Acad. Sci. USA 82: 7994-7998. Reference: Biol. Bull. 172: 284-298. (June, 1987) THE LARVAL STAGES OF LITHOTRYA DORSALIS (ELLIS & SOLANDER, 1786): A BURROWING THORACICAN BARNACLE JOSEPH F. DINEEN, JR. Horn Point Environmental Laboratories, Center for Environmental and Estuarine Studies, University of Maryland, Cambridge. Maryland 2 161 3 ABSTRACT Lithotrya dorsalis is a member of the only genus of thoracican barnacles known to burrow and is widely distributed throughout the tropical western Atlantic. It occurs primarily in high energy intertidal environments. L. dorsalis undergoes the typical thoracican larval sequence. Six naupliar stages are followed by the cyprid stage. These larvae were reared in the laboratory and their stages are described for the first time. Newly hatched stage I nauplii are typically 360 ^m in total length; larval size increases to 1 100 yum by the 6th instar. The most distinguishing characteristic of Lihtotrya dorsalis nauplii is the presence of unusually long, spinulated posterior shield spines in stages IV through VI. Complete larval development (stage I nauplius to cyprid) averaged 18 days and ranged from 12 to 23 days. Scanning electron micrograph de- scriptions of the cyprid cuticle of this animal, showing a unique striated appearance, are also included. INTRODUCTION First described by Sowerby ( 1 822), the genus Lithotrya (family Scalpellidae, sub- family Lithotryinae) is further treated by Darwin (1851 ), Sewell ( 1 926), Otter ( 1 929), and Cannon (1947). L. dorsalis is considered to be the only species occurring in the western Atlantic (Zevina, 1981). Found exclusively in carbonate substrata, this spe- cies is an abundant and ubiquitous constituent of exposed tropical coastlines (Gins- burg, 1953; Newell et ai, 1959; Ahr and Stanton, 1973; Southward, 1975; Focke, 1 977; Spivey, 1981). Settling gregariously, L. dorsalis is capable of substantial bioero- sion. Bore holes up to 7 cm in length, with a typical oval aperture as large as 5 X 8 mm have been reported (Ahr and Stanton, 1973). These authors described average boring densities in limestone beach rock off Puerto Rico to be one tube per cm2 of rock with up to 30% of rock excavated in some samples. It has been found subtidally (75-80 ft) (Pequegnat and Ray, 1 974) and in fringing reef environments (MacGeachy and Stearn, 1976). However, its larval history has remained virtually unknown. Complete larval descriptions of pedunculate barnacles are listed by Lang (1976). Dalley (1984) has since successfully reared the naupliar stages of Conchoderma auri- tum (L.). Of these pedunculate barnacles, only four are scalpellids (Thoracica: Scal- pellidae): Capitullum mitella L. (= Mitella mitella = Pollicipes mitella) (Yasugi, 1 937); Calantica spinosa Quoy and Gaimard (= Pollicipes spinosus) (Batham, 1 946); Scalpellum scalpellum L. (Kaufmann, 1965); and Pollicipes polymerus Sowerby (Lewis, 1975). The burrowing habit of Lithotrya makes it unique not only among the scalpellids but among the thoracican cirripeds in general. Description of the larval stages of L. dorsalis make it possible to identify and stage specimens from tropical plankton tows, thereby enhancing future ecological studies. Received 20 August 1986; accepted 23 March 1987. 284 LARVAL STAGES OF L. DORSALIS 285 Furthermore, documentation of the plate ontogeny and maturation of the post-cy- prid, a direct consequence of this study, may have implications for the geological age of the Lithotryinae and the burrowing rate of newly settled cyprids, respectively. Cyprid appearance is often cited in the literature as being unhelpful in differenti- ating species because of its interspecific similarity and intraspecific variability in size. Standing (1980) compared seven species of barnacle cyprids at the light microscopy level and concluded that shape (side view) is a more reliable indicator of species than size and further concluded that carapace sculpturing may prove to be a useful tool for differentiating species at the cyprid level. Carapace sculpturing, along with size, was previously used to differentiate cyprids of Balanus glandula from Balanus cario- sus, respectively (Strathman and Branscomb, 1979). Both Lang (1979) and Standing (1980) alluded to the fact that the use of the scanning electron microscope (SEM) may be a valuable tool in extending our knowledge of the external morphological features of the cyprid larva and thereby rendering these features of some diagnostic value at the interspecific level. SEM descriptions of the external cuticle of the cyprid larva of L. dorsalis are included here in an attempt to encourage similar descriptions of this significant stage in future larval studies. Complete sets of larvae are deposited with the American Museum of Natural His- tory, New York, New York; and the National Museum of Natural History, Washing- ton, DC. MATERIALS AND METHODS Sampling took place in September and October, 1983, and again, at the same time of year, in 1984. Using a hammer and chisel, pieces of limestone reef rock con- taining adult barnacles were removed from the intertidal zone of Indian Key, Florida, and placed in buckets containing ambient seawater. This rock was returned to the laboratory the same day and vigorously aerated. Over the next several days, the rock was further chiseled into smaller pieces. Gravid individuals, easily recognized by the bright orange pigmentation of the ovigerous lamellae (paired egg bearing structures) seen through the peduncular cuticle, were removed intact and placed individually into small culture bowls. The capitulum was subsequently disjoined and the lamellae were coaxed from the peduncle by applying gentle pressure. Culturing techniques The basic rearing techniques of Freiberger and Cologer ( 1 966) were followed. De- pending on the degree of maturation, stage I nauplii would either hatch immediately or after incubation of the lamellae for up to 10 days. Maturation in incubated cultures was gauged by the increasing darkness of the naupliar eyespot. Incubating lamellae were placed in 0.45 ^m millipore filtered seawater (MFSW), 30 ppt salinity at 30°C, in the dark. Cultures were observed and seawater was changed daily. Stage I nauplii were transferred at varying densities (2-8 nauplii per ml medium) to stendor dishes containing MFSW reduced to 28 ppt salinity. Cultures were placed in a tempera- ture— controlled chamber at 30°C on a shake table at gentle speed. Either a 12: 12 or 18:6 light:dark schedule was used. Various algal diets were tried but the complete larval sequence was only obtained using the mixed algal diets of Thalassiosirajluvia- tilus/Tetraselmis suecica or Chaetoceros gracilis/Isochrysis sp. Algae used as food were grown in f/2 medium in batch culture, centrifuged, resuspended in MFSW, and dispensed at initial densities of 2 X 105 cells/ml. The antibiotics Streptomycin sulfate, Polymixin B sulfate (both at 12 ng/m\) and Penicillin G (25 Mg/ml) (Sigma) were 286 J. F. DINEEN, JR. TABLE I Lithotrya dorsalis: Measurements of larvae* Total length (^m) Shield length (^m) Width ( urn) Stage x SD Range x SD Range x SD Range I 362 25 333-421 186 15 170-219 II 546 10 540-564 202 9 190-220 238 14 220-263 III 749 35 681-795 243 13 220-257 258 10 237-267 IV 899 44 799-952 345 17 303-360 323 17 284-335 V 1035 40 936-1070 446 13 435-472 409 19 366-438 VI 1141 37 1070-1171 510 15 484-536 464 22 424-493 cyprid 518 15 496-542 240 8 233-252 * n = 10 in stages I-V and cyprid, n = 8 in stage VI. x = mean; SD = standard deviation. used (Landau and D'Agostano, 1 977). Glassware was autoclaved daily. Cultures were transferred by pipet every 24 h. Several individuals from each culture were removed daily and preserved in either 70% ethanol or 5% buffered formalin in seawater. An ocular micrometer was used to approximate the stage of these individuals. Represen- tatives from each stage were later stained with chlorazol black in a 10% solution of glycerin in 70% ethanol. The alcohol was allowed to slowly evaporate. Larvae, now more manipulable, were then either mounted whole or dissected with fine tungsten needles. Stages were verified and drawings were made using camera lucida on a Wild M20 compound microscope. Specimen preparation — SEM Cyprids prepared for scanning electron microscopy were initially fixed in buffered glutaraldehyde and dehydrated through a graded series of ethanols. Then the cyprids were either placed in trichlorotriflouroethane (Nott, 1969) which was allowed to slowly evaporate under a bell jar through which nitrogen gas was gently bubbled, or were transferred to acetone and then critical point dried (Samdri pvt.3, Tousimis). Using a method similar to Waller ( 198 1 ), these larvae were individually glued with poly vinyl acetate to one end of a small platinum wire. The other end of this wire was twisted into a loop which was then attached to a glass cover slip using silver paint. The cover slip was then affixed to a mounting stub using double sided tape. The animals were then sputter coated with gold/palladium under vacuum (Hummer 5, Technics). This method allows manipulation of the specimen by bending the wire with fine forceps. Larvae treated this way were examined with either a JSM-U3, JEOL or a 1000A, AMRay scanning electron microscope at a gun potential of 1 5 to 20 kV. RESULTS Larval cultures Lithotrya dorsalis exhibited the typical thoracican larval sequence. Six naupliar stages were followed by the cyprid stage. Average survival from naupliar stage I to cyprid in 29 larval cultures was 12.5% (range = .54-46.1%). Cultures with initially lower densities (1-2 nauplii/ml medium) yielded higher percent survival. Complete larval development (stage N I to cyprid) averaged 18 days and ranged from 12 to 23 days. Several cultures molted in synchrony, the majority did not. Sizes of larval stages are given in Table I. Shield lengths were approximated for naupliar stages II and III because of the lack of a distinct posterior margin in these stages. LARVAL STAGES OF L DORSALIS 287 FIGURE 1 . Lithotrya dorsalis. Dorsal view of shield outlines of naupliar stages I-VI. The nauplii^ The naupliar stages as seen from above are illustrated in Figure 1. First stage nauplii (Fig. 1, I) are characterized by frontolateral horns directed posteriorly with the posterior region of the cephalic shield tapering distinctly into a rudimentary tho- racic spine dorsal to a small, forked thoraco-abdominal process. A median naupliar eye is evident and remains throughout naupliar development. A unilobed labrum with one median tooth is present. Stage I nauplii have the full complement of naupliar appendages, the uniramous antennules, biramous antennae, and mandibles, provided with simple setae. It is a nonfeeding stage and molts to stage II within one or two hours post hatching. The thoraco-abdominal process and dorsal thoracic spine become elongated and spinulated at stage II (Fig. 1, II) and remain this way throughout naupliar develop- ment. The thoraco-abdominal process is distinctly and deeply forked and contains one pair of series 1 spines (Fig, 2, II). A row of fine, slightly arched hairs appear on each side of the proximal portion of the labrum. An additional tooth appears at either 1 Descriptive terminology after Lang (1979) with the exception of the term "thoraco-abdominal" pro- cess suggested by W. A. Newman. 288 J- F. DINEEN, JR. end of the posteriolateral margin (Fig. 3, II). Frontal filaments are apparent at stages II- VI. Stage III nauplii (Fig. 1, III) are somewhat similar to stage II nauplii but the tho- raco-abdominal process and dorsal thoracic spine have become proportionately longer relative to the cephalic shield. The posterior margin of the labrum now bears five teeth and two to three spines appear laterally (Fig. 3, III). As is typical of cirriped nauplii in general, the frontolateral horns at stages III-VI appear split at the tip. Stage III nauplii appeared in culture as early as day 3. Stage IV nauplii (Fig. 1, IV) are easily distinguished from previous stages by the presence of unusually long, barbed posterior shield spines and a distinct posterior shield margin. In active, swimming larvae, the posterior shield spines, present throughout stages I V-VI, curve prominently away from the main body axis. One pair of series 1 spines as well as 1 median and 1 pair of series 2 spines (Fig. 2, IV) are evident on the thoracic portion of the thoraco-abdominal process of stage IV nauplii [Brown and Roughgarden ( 1 985) refer to series 2 spines as thoracic spines.] Increasing proportionately in size, stage V nauplii (Fig. 1 , V) possess several spines on the lateral posterior margin of the cephalic shield. One pair each of series 1 and 3 spines as well as one median and two pairs of series 2 spines are present on the abdom- inal process (Fig. 2 V). The earliest day of appearance for stage V nauplii was day 7. Approximately 24 hours post-molt, a pair of compound eyes become visible in stage VI nauplii (Fig. 1, VI). A variable number of spines are present on the anterior as well as lateral margins of the cephalic shield. Six pairs of series 2 spines and one pair each of series 1 and 3 spines are present on the thoraco-abdominal process (Fig. 2, VI). Thoraco-abdominal process and caudal spines The thoraco-abdominal process and corresponding caudal spines for naupliar stages I-VI are shown in relation to the dorsal thoracic spine and posterior shield spines in Figure 2. The thoraco-abdominal process in all stages is shorter than the dorsal thoracic spine. The distal end of the thoraco-abdominal process is deeply forked in naupliar stages II- VI. The appearance of caudal spine series 1, 2, and 3 on this process closely follows the sequence established for other cirripeds (Moyse, 196 1 ; Lang, 1979). However, examination of several stage V nauplii revealed 3 pairs of series 2 spines instead of the usual 2 as drawn. The thoraco-abdominal process along with the dorsal thoracic spine (stages II-VI) and the posterior shield spines (stage IV- VI) are heavily spinulated. There appears to be no pattern to this secondary ornamen- tation. Labrum The unilobed labrum (Fig. 3), characteristic of all lepadomorph nauplii, is dis- tinctly indented at approximately midlength in naupliar stages I and II ofLithotrya dorsalis (Fig. 3, I & II). The proximal half of the stage I labrum is noticeably wider than the distal portion. By naupliar stage III, the median tooth now appears clearly associated with the labral gland (Fig. 3, III). Two to three spines are now present on the lateral edges. On the dorsal (inside) surface a row of very fine hair-like processes appear along either side of the labral gland duct. At stages IV- VI (Fig. 3, IV, V, & VI), very thick, hair-like projections appear to surround all but the median tooth on the posterior margin. By stage V, the median tooth becomes less prominent while the remaining teeth become more pronounced. By stage VI, the median tooth is reduced to an inconspicuous knob and appears disassociated from the now indistinct labral gland. LARVAL STAGES OF L DORSAL1S 289 FIGURE 2. Lithotrya dorsalis. Ventral and lateral views of the dorsal thoracic spine (DTS) and tho- raco-abdominal process (TAP) relative to the posterior shield spines (PSS) in naupliar stages I-VI. Caudal spine series are designated by Arabic Numbers (1,2, and 3). The appendages The uniramous antennules (Fig. 4), biramous antennae (Fig. 5), and mandibles (Fig. 6) are shown for all 6 naupliar stages. The entire setae are not drawn, in most cases, due to their extremely long and delicate nature. Using the basic setation formu- lae of Newman (1965) as modified by Lang (1979), these appendages are further described in Table II. The number of setae appearing within a particular stage was relatively constant, reinforcing the idea that setation of an appendage (barring the consequences of dissec- tion) is an accurate indication of naupliar stage (Lang, 1979; Dalley, 1984). The tran- sition from simple to plumose setae appeared to be somewhat variable, particularly on the antennal and mandibular endopodites of stages IV-VI. 0 1 mm FIGURE 3. Lithotrya dorsalis. Labrum of naupliar stages I-VI. 290 J. F. DINEEN, JR. FIGURE 4. Lithotrya dorsalis. Antennule of naupliar stages I-VI. The cyprid The cyprid (Fig. 7 A) is relatively small (averaging 5 1 8 /j,m total length) and similar in overall appearance to other previously described cyprid larvae at the level of light microscopy. A bivalved carapace, compound and median eyes, six pairs of biramous natatory appendages (Fig. 7B), caudal furca (Fig. 7C), and a pair of antennules (Fig. 7D) modified to act as exploratory and attachment organs are present. The anterior margin of the cyprid is evenly rounded. The posterior ventral margin tapers more sharply than the dorsal margin. As a consequence, the posterior end appears more ovoid (Fig. 7A). The living cyprid is somewhat translucent in both the anterior and posterior extremities. The bulk of the body is filled with golden-colored oil cells. There is an irregularly shaped, orange pigment area in the mid to posterior region. The compound eyes have a distinct dark red coloration. The setation of the cyprid appendages (Fig. 7B, C, D) is remarkably consistent with that described for other species (Batham, 1946; Walker and Lee, 1976). The use of the scanning electron microscope in describing barnacle cyprids has been limited. The cyprid antennule of Semibalanus balanoides L. was described by Nott ( 1 969). Walker and Lee ( 1 976) studied the surface structure of the S. balanoides cyprid using SEM. Svane (1986) showed a settling Scalpel/urn scalpellum cyprid (low magnification) using SEM. Walker (1985) described the external features of the cy- prid of the rhizocephalan, Sacculina carcini Thompson. Other rhizocephalan cyprids described with the SEM are Lernaeodiscus porcellanae (Ritchie and Heteg, 1981; Htfeg, 1985a)and Clistosaccus paguri (\\6zg, 1985b). It is perhaps premature to speculate about the reliability of cyprid carapace sculp- turing as revealed by the SEM in differentiating species until more such descriptions are available for meaningful comparisons. But at the very least, such SEM descrip- tions should supplement other characteristics (size, shape, color) currently being used with varying amounts of success. SEM investigation of the cyprid larva of Lithotrya dorsalis reveals distinct sculp- turing over the entire carapace surface (Fig. 8 A, C; 9 A). Sculpturing consists of cari- nae arranged in a somewhat parallel configuration and interspersed with numerous cuticular setae (Fig. 9B), pores (Fig. 9C) and small papillae (Fig. 9D). The parallel LARVAL STAGES OF L. DORSALIS 291 FIGURE 5. Lithotrya dorsalis. Antenna of naupliar stages I-VI. 292 J. F. DINEEN, JR. FIGURE 6. L.thotrya dorsalis. Mandible of naupliar stages I- VI. LARVAL STAGES OF L. DORSALIS 293 TABLE II Setation formulae for nauplii o/Lithotrya dorsalis Stage Antennule Antenna Mandible VI S P P PSPS SP S PS S 4P 8P PPSPP SSP FSFF PFSC G P 5P 5S PSPS PSCP PPP G V S S P PSPS SP S P S 3P 7PS PPSPS SSP FSPF PFSC G P 5P 5S PSPS PSCP PPP G IV S P PSPS SP P S 2P 7P PPSPS SSP F SF PFSC G P 4P 4S PSP PSCP PPP G III S PSPS SP P S 2P 5P pp P S P F F PPSC G P 3PS 3S PSS P CP PPP G 11 SSPS SP P S SP 4PS PS S S P F F SP C G P 3PS 3S P S P CP PP G I ssss ss S S S 4S SS S S S S S SS G S 3S 3S S S S S SS G arrangement of these carinae or ridges give the L. dorsalis cyprid cuticle a striated appearance quite distinct from the pitted appearance of the Semibalanus balanoides cyprid cuticle described by Walker and Lee (1976) and the somewhat hairy appear- ance of the Lernaeodiscus porcellanae (Ritchie and H#eg, 1981) and Sacculina car- cini (Walker, 1 985) cyprids due to the long setae present on their cuticles. The cuticu- lar pores and setae of L. dorsalis and S. balanoides cyprids however, appear similar in shape and size. In L. dorsalis, pores (Fig. 9C) are round, small (approximately 5- 8 /urn) and numerous. Cuticular setae (Fig. 9B) are also numerous, particularly in the posterior region of the carapace. These setae extend from a pore-like process and vary slightly in length (5-8 /j,m). Dorsal and ventral marginal carinae sometimes continue uninterrupted for the entire length of the cyprid. Towards the mid-lateral section of the carapace (Fig. 9A) the carinae become shorter and appear to follow more closely the contour of the animal causing some degradation of their parallel arrangement. The distinct dorsal junction of the carapace valves including the dorsal hinge region is shown in high magnification (Fig. 8B). DISCUSSION Rearing techniques Survivorship to the cyprid stage was relatively poor. Many permutations of diet, temperature, salinity, and light were tried with varying degrees of success. It was hoped that antibiotics could be avoided in the culture medium by lowering the tem- FIGURE 7. Lithotrya dorsalis. Carapace shield outline of cyprid (A); natatory appendage (B); caudal furca (C); and antennule (D). 294 J. F. DINEEN, JR. FIGURE 8. Lithotrya dorsalis. SEM Dorsal (A) and ventral (C) view of bivalved cyprid carapace; high magnification view of carapace junction including hinge region (B). (Scale bars: A = 50 ^m; B = 10 ^m; C = 100/urn). perature from 30 to 22°C but larvae subsequently became sluggish and died off. Am- bient seawater temperature when animals were sampled in September, 1983, was 32°C. This sluggish behavior may have resulted from thermal shock or perhaps a lower feeding efficiency at the decreased temperature (Scheltema and Williams, 1982). By lowering the salinity to 28 ppt in combination with antibiotics, entangle- ment of nauplii, particularly in mature cultures when appendage setation became LARVAL STAGES OF L. DORSALIS 295 FIGURE 9. Lithotrya dorsalis. SEM Posterior lateral view of cyprid carapace (A) and cuticular struc- tures: Seta (B). pore (C) and papillae (D). (Scale bars: A = 50 ^m; B = 10 ^m; C and D = 1 . more elaborate, was avoided. In an attempt to increase the cyprid yield, while keeping the logistics of culture maintenance at a minimum, initial densities were increased from 1 or 2 to 6 or 8 nauplii per ml culture medium. Some of these high density cultures reached stage IV nauplius with no apparent deleterious effects. At this time, however, the algae being used as food (Thalassiosirafluviatilus/Tetraselmis suecicd) became contaminated. By the time a new diet was selected (Chaetoceros gracilis/ Isochrysis sp.), most cultures had been without food for 36 h. This apparent starva- tion resulted in substantial mortality, especially in older cultures. Younger cultures survived somewhat better, perhaps due to yolky material still present in these early naupliar stages. Because of the unexpected mortality, it was impossible to compare survivorship between low and high density cultures statistically. However, even though percent survival to the cyprid stage was lower in the "higher density" cultures, it was encouraging that actual cyprid yields in these cultures were higher than in the "low density" cultures. Suspecting the possible release of chemical cues, several unsuccessful attempts to initiate metamorphosis of stage VI nauplius to the cyprid stage were made by placing stage VI nauplii in "conditioned" culture medium where metamorphosis had pre- viously taken place. The size of gravid individuals at the time of sampling (September and October) was small; their carinal-rostral axes averaged 3.4 mm. The larger individuals sampled at this time (carinal-rostral axes as large as 7.8 mm) were not gravid. This may have 296 J. F. DINEEN, JR. TABLE III Measurements ofscalpellid cyprid larvae Length Width Species (Mm) (Mm) Author Capitullum mitella 1050 400 Calantica spinosa 850 385 Yasugi, 1937 Scalpellum scalpellum 686 329 Batham, 1946 Pollicipes polymems 425 232 Kaufmann, 1965 Lithotrya dorsalis 518 240 Lewis, 1975 Dineen (present study) been a consequence of sampling at the end of the breeding season. Reproductive patterns in Lithotrya dorsalis have not been documented. Attempts to induce repro- duction in the laboratory by increasing food availability and temperature failed. Comparison ofscalpellid larvae Adaptive radiation in the scalpelloids has given rise to the more advanced sessile barnacles [Suborders: Brachylepadomorpha (now extinct), Verrucomorpha, and Ba- lanomorpha (Newman, 1982)]. It is noteworthy that the basic sequential pattern of larval development of this relatively primitive scalpellid barnacle is remarkably sim- ilar to that of other more advanced cirripeds described to date. When comparing the larval stages of Lithotrya dorsalis with the four other de- scribed scalpellid species, the most striking similarity occurs between the cephalic shield outlines of L. dorsalis and Capitullum mitella (Yasugi, 1937). Although the length of L. dorsalis nauplii is greater at all stages, the same basic shield shape, along with the elongated and spinulated thoraco-abdominal process and dorsal thoracic spine and in particular, the elongated posterior shield spines in larval stages IV-VI, are present in both species. Rudimentary posterior shield spines appear in C. mitella naupliar stages II and III. Nauplii of the three other scalpelloids described, Calantica spinosa (Batham, 1946), Scalpellum scalpellum (Kaufman, 1965), and Pollicipes po- lymerus (Lewis, 1975) have relatively abbreviated thoraco-abdominal processes and dorsal thoracic spines and appear to lack posterior shield spines entirely. The unilo- bate labra of L. dorsalis, P. polymems, and C. mitella are similar in outline, having a blunt posterior margin with numerous teeth, particularly in later stages. This same region in the naupliar labrum of stage VI C. spinosa appears narrowly rounded, with a single middle tooth. Labra of stage I nauplii of both L. dorsalis and C. mitella widen in the proximal region. Unfortunately, setation formulae for C. mitella are not given. Cyprid sizes of these five scalpellids are compared in Table III. The similarity between Lithotrya dorsalis and Capitullum mitella nauplii stimu- late two questions. (1) Are these two species more closely aligned than presently thought? [Zevina ( 198 1 ) places them in separate subfamilies: Lithotryinae and Polli- cipinae, respectively). (2) More fundamentally, how reliable are naupliar characteris- tics as indicators of phylogenetic affinities among the Cirripedia? ACKNOWLEDGMENTS I am most grateful to Mary E. Rice for allowing me use of the facilities at the Smithsonian Marine Station — Link Port, Florida, and for her continuous guidance and support throughout this study. I especially thank Hugh Reichardt (SMS-LP) LARVAL STAGES OF L. DORSALIS 297 without whose advice, assistance, and sense of humor, this study would never have come to completion. Terry West, Sandy Branscomb, and Matt Landau provided me with many helpful suggestions regarding barnacle rearing techniques as did Bob Eif- ert, who also provided me with algal cultures. I thank Tim Maugel and especially Linda Morin for their technical assistance with the SEM. I thank colleagues at the University of Maryland, Ed Bird, Loren Coen, and Miriam Smyth for their encour- agement. This study is part of a dissertation to be submitted to the University of Maryland in partial fulfillment for the Doctor of Philosophy Degree. In that regard, I would like to acknowledge the support of my committee — Drs. Shirley Pomponi, Richard Smucker, Dennis Taylor, Larry Ward and Melbourne Carriker (CMS-Uni- versity of Delaware). A special thanks to Ruth D. Turner, Museum of Comparative Zoology, Harvard University, for her help and friendship and to Mark and Diane Littler, for the use of facilities at the National Museum of Natural History, Washing- ton, DC. I thank W. A. Newman, W. H. Lang, M. J. Grygier, and two anonymous reviewers for their very helpful comments on the manuscript. This work was partially supported by grants from the Graduate School, University of Maryland, the Horn Point Environmental Lab, the Lerner-Gray Fund, and Sigma Xi Foundation. This is Contribution No. 1 73 from the Smithsonian Marine Station at Link Port, Florida. LITERATURE CITED AHR, W. M., AND R. J. STANTON. 1973. The sedimentologic and paleoecologic significance of Lithotrya, a rock-boring barnacle. J. Sediment. Petrol. 43: 20-23. BATHAM, E. J. 1946. Pollicipes spinosus Quoy and Gaimard. II: Embryonic and larval development. Trans. R. Soc. New Zealand 15: 405-418. BROWN, S. K., AND J. ROUGHGARDEN. 1985. Growth, morphology, and laboratory culture of larvae of Balanus • glandiila (Cirripedia: Thoracica). J. Crust. Biol. 5: 574-590. CANNON, H. G. 1947. On the anatomy of the pedunculate barnacle Lithotrva. Phil. Trans. R. Soc. Land. Ser.B233: 12-136. DALLEY, R. 1984. The larval stages of the oceanic pedunculate barnacle Conchoderma auritum (L.) (Cir- ripedia, Thoracica). Crustaceana 46: 39-54. DARWIN, C. 1 85 1 . A monograph on the subclass Cirripedia with figures of all the species. Pp. 332-393 in The Lepadidae : or, Pedimculated Cirripedes. Ray Soc. Publ., London. ELLIS, J., AND D. SOLANDER. 1786. The Natural History of Many Curious and Uncommon Zoophytes Collected From Various Parts of the Globe. Benjamin White and Son, London. FOOKE, J. W. 1977. The effect of a potentially reef-building vermetid-coralline algal community on an eroding limestone coast, Curacao, Netherlands Antilles. Proc. 3rd Int. C oral Reef Symp. (RSMS). Pp. 239-245. FREIBERGER, A., ANDC. P. COLOGER. 1966. Rearing acorn barnacle cyprids in the laboratory for marine fouling studies. Nav. Eng. J. 78: 88 1-890. GINSBURG, R. N. 1953. Intertidal erosion on the Florida keys. Bull. Mar. Sci. 3: 55-69. H0EG, J. T. 1985a. Cypris settlement, kentrogon formation and host invasion in the parasitic barnacle Lernaeodiscus porcellanae(Mu\\er) (Crustacea: Cirripedia: Rhizocephala). Ada Zool. 66: 1-45. H0EG, J. T. 1985b. Male cypris settlement in Clistosaccus paguri Lilljeborg (Crustacea: Cirripedia: Rhizo- cephala). J. Exp. Mar. Biol. Ecol. 89: 221-235. KAUFMANN, R. 1 965. Zur Embryonal- und Larvalentwicklung von Scalpellum scalpellum L. (Crust. Cirr.) miteinem Beitrag zur Autokologie dieser Art. Z. Morph. Okol. TiereSS: 161-232. LANDAU, M., AND A. D'AGOSTINO. 1977. Enhancement of laboratory cultures of the barnacle Balanus eburneus Gould using antibiotics. Crustaceana 33: 223-224. LANG, W. H. 1976. The larval development and metamorphosis of the pedunculate barnacle Octolasmis mulleri (Cooker, 1902) reared in the laboratory. Biol. Bull. 150: 255-267. LANG, W.H.I 979. Larval development of shallow water barnacles of the Carolinas (Cirripedia: Thoracica) with keys to the naupliar stages. NOAA Tec/in. Rep. NMFS Circular 421: 1-39. LEWIS, C. A. 1 975. Development of the gooseneck barnacle Pollicipes polymerus (Cirripedia: Lepadomor- pha): Fertilization through settlement. Mar. Biol. 32: 141-153. MACGEACHY, J. K., AND C. W. STEARN. 1976. Boring by macro-organisms in the coral Montastrea annu- laris on Barbados reefs. Int. Rev. Hvdrobiol. 61: 715-745. 298 J. F. DINEEN, JR. MOYSE, J. 1 96 1 . The larval stages of Acasta spongites and Pvrgoma anglicum (Cirripedia). Proc. Zool. Soc. Lond. 137:371-392. NEWELL, N. D., J. IMBRIE, E. G. PURDY, AND D. L. THURBER. 1959. Organism communities and bottom .:ies, great Bahama bank. Bull. Am. Mus. Nat. Hist. 1 17: 1 80-2 1 1 . NEWM -•-, N, W. A. 1 965. Prospectus on larval cirriped setation formulae. Crustaceana 9: 5 1 -56. NEWMAN, W. A. 1982. Cirripedia. Pp. 197-220 in The Biology of Crustacea, Vol. 1, L. Abele, ed. Academic Press, New York. NOTT, J. A. 1969. Settlement of barnacle larvae: Surface structure of the antennular attachment disc by scanning electron microscopy. Mar. Biol. 2: 248-25 1 . OTTER, G. W. 1929. Rock destroying organisms in relation to coral reefs. Sci. Rep. Great Barrier Reef Expd. 1:324-350. PEQUEGNAT, L. H., AND J. P. RAY. 1974. Crustacea and other arthropods. P. 259 in Biota of the West Flower Garden Bank, T. J. Bright and L. H. Pequegnat, eds. Gulf Publishing Company, Houston, Texas. RITCHIE, L. E., AND J. T. H&EG. 1981. The life history of Lernaeodiscus porcellanae (Cirripedia: Rhizo- cephala) and co-evolution with its porcellanid host. /. Crust. Biol. 1: 334-347. SCHELTEMA, R. S., AND I. P. WILLIAMS. 1982. Significance of temperature to larval survival and length of development in Balanus eburneus (Crustacea: Cirripedia). Mar. Ecol. Prog. Ser. 9: 43-49. SEWELL, R. B. S. 1926. A study ofLithotrya nicobarica Reinhardt. Rec. Indian Mus. 28: 269-330. SOUTHWARD, A. J. 1975. Intertidal and shallow water Cirripedia of the Caribbean. Stud. Fauna Curacao 46: 1-53. SOWERBY, G. B. 1822. The genera of recent and fossil shells, for the use of students in Conchology and Geology, with original Plates by James Sowerby. Pt. VIII. London. SPIVEY, H. R. 198 1 . Origins, distribution, and zoogeographic affinities of the Cirripedia (Crustacea) of the Gulf of Mexico. J. Biogeog. 8: 153-176. STANDING, J. D. 1980. Common inshore barnacle cyprids of the Oregonian faunal province (Crustacea: Cirripedia). Proc. Biol. Soc. Wash. 93(4): 1 184-1203. STRATHMAN, R. R., AND E. S. BRANSCOMB. 1979. Adequacy of clues to favorable sites used by settling larvae of two intertidal barnacles. Pp. 77-89 in Reproductive Ecology of Marine Invertebrates, S. E. Stancyk, ed. Univ. South Carolina Press, Columbia. SVANE, I. 1986. Sex determination in Scalpellum scalpellum (Cirripedia: Thoracica: Lepadomorpha), a hermaphroditic goose barnacle with dwarf males. Mar. Biol. 90: 249-253. WALLER, T. R. 1981. Functional morphology and development of veliger larvae of the european oyster Ostrea edulis Linne. Smithson. Contrib. Zool. 328: 1 -70. WALKER, G., AND V. E. LEE. 1976. Surface structures and sense organs of the cypris larva of Balanus balanoides as seen by scanning and transmission electron microscopy. J. Zool. Lond. 178: 161- 172. WALKER, G. 1985. The cypris larvae of Sacculina carcini Thompson (Crustacea: Rhizocephala). J. Exp. Mar. Biol. Ecol. 93: 1 3 1 - 1 45. YASUGI, R. 1937. On the free swimming larvae ofMitella mitella L. Botany Zool. Tokyo 5: 792-796. ZEVINA, G. B. 1 98 1 . Barnacles of the suborder Lepadomorpha (Cirripedia, Thoracica) of the World Ocean. Part 1 : Family Scalpellidae. Guides to the fauna of the USSR. Zoological Institute of the Academy of Sciences of the USSR. 127: 1-406. Leningrad. Reference: Biol. Bull. 172: 299-306. (June, 1987) STUDIES OF THE ECDYSIOTROPIC ACTIVITY OF JUVENILE HORMONE IN PUPAE OF THE TOBACCO HORNWORM, MANDUCA SEXTA LOUIS SAFRANEK AND CARROLL M. WILLIAMS The Biological Laboratories, Harvard University, Cambridge, Massachusetts 02138 ABSTRACT Topical administration of a juvenile hormone analog (JHA) to pupae of the to- bacco hornworm accelerated the initiation of adult development and proved to be an ecdysiotropic stimulus even in the absence of the brain. Nevertheless, individuals developing in response to JHA were never normal. In fact, doses of JHA too low to accelerate development often provoked typical abnormalities in the resulting moths. Allatectomy of either diapause-destined or non-diapause pupae failed to alter the time course of adult development. In line with this finding, corpora allata (CA) from both diapause-destined and non-diapause pupae were without activity in two juvenile hormone (JH) bioassays. These results suggest that JH is not involved in the normal initiation of adult development. Indeed, it must be absent for normal development to proceed. Evidently, the ecdysiotropic effect of exogenous JH in pupae reflects a sensitivity to JH retained from the larval period. INTRODUCTION Increased secretion of ecdysone by the prothoracic glands (PG) is believed to con- stitute the proximate stimulus for the termination of pupal diapause and initiation of adult development. This augmented synthesis is believed to depend on the produc- tion of an ecdysiotropic hormone by the pupal brain, the so-called prothoracicotropic hormone (PTTH). Thus, in many well-documented instances, removal of the brain can significantly delay or even prevent the onset of adult development (for review, see Safranek and Williams, 1980). In previous work from this laboratory we have examined the regulation of adult development in pupae of the tobacco hornworm, Manduca sexta. In these studies we have documented the critical role of the brain and PG in the normal initiation of adult development in both diapausing and non-diapausing pupae (Safranek and Wil- liams, 1980). But in other studies our findings conflict with the classical model. For example, we have observed the initiation of adult development in brainless or decapi- tated pupae as well as in pupae lacking PG (Safranek and Williams, 1980; Safranek et ai, 1986). We have also demonstrated the ability of juvenile hormone (JH) or its analogs (JHA) to hasten the termination of pupal diapause and the initiation of adult development (Safranek et al, 1980). These findings suggested that the endocrine sys- tem governing adult development in the hornworm might be significantly different from that defined by the classical model. In the present study we focus especially on Received 24 October 1986; accepted 23 March 1987. Abbreviations: CA = corpora allata; JH = juvenile hormone; JHA = juvenile hormone analog; LD = long-day; SD = short-day; PG = prothoracic gland; PTTH = prothoracicotropic hormone; RIA = radioimmunoassay. 299 300 L. SAFRANEK AND C. M. WILLIAMS the ability of exogenous JH to accelerate the termination of pupal diapause and exam- ine its possible role in the normal onset of adult development. MATERIALS AND METHODS Hornworms were reared as described previously (Safranek and Williams, 1980) under either a short-day (SD, 12L:12D) photoperiod or a long-day (LD, 17L:7D) photoperiod at 25°C. The day of pupation is termed Day 1 of the pupal stage, and the first seven days of the pupal stage Week 1 . Operations and ligations were performed as described previously (Safranek and Williams, 1980). Allatectomy was performed through a small horizontal incision in the vertex of the head within 12 h after pupal ecdysis. The corpora cardiaca were left in situ in this procedure. Sham operations on the CA were performed with complete visualization of these glands. In all experi- ments the initiation of development was recognized under the dissecting microscope by tracheal apolysis in the pupal wings. The JHA "Hydroprene" (ZR-512, Zoecon Corp., technical grade) was used in certain experiments. For topical application, it was dissolved in acetone and dis- pensed onto the thorax with a 100 ^1 Hamilton syringe on a repeating dispenser. The endocrine activity of individual CA was evaluated by the use of the "black larval assay" as described by Safranek and Riddiford (1975). JH secretion was also bioassayed by the implantation of CA into the heads of non-diapausing LD pupae within 24 h after pupal ecdysis: After the completion of development, the moths were scored for JH-induced morphological aberrations of the type described by Riddiford and Ajami( 1973). Ecdysteroid levels were measured by radioimmunoassay as previously described (Carrow et al., 198 1 ) using 20-OH-ecdysone as the standard. RESULTS Effects of juvenile hormone application on diapausing pupae Groups of 50 diapause-destined SD pupae received one of three graded doses of JHA in a single topical application on the first day after pupal ecdysis. A control group of 50 similar pupae was treated with the acetone solvent only. The initiation of adult development was monitored at weekly intervals. As shown in Figure 1 , the two lower doses of JHA failed either to accelerate or to retard development over and beyond that of the controls. By contrast, about half of the individuals that received the highest dose of 200 yug terminated diapause prema- turely; the rest, which were not distinguished by their sex, did not develop notably faster than the controls. Additional experiments demonstrated that all individuals initiated development within 3 weeks of pupation after a single administration of 400 /ug (n = 20) or after daily applications of 20 ^g over 7 days (n = 15). Those pupae initiating development during the first 6 weeks in response to the high dose of 200 ^g uniformly displayed pronounced JH-related abnormalities of the type previously described by Riddiford and Ajami (1973). These abnormalities were most reliably apparent in the compound eyes where the occurrence of facet-free cres- cents regularly provided clear evidence of an individual's exposure to JHA. Even when the dose was too low to accelerate the termination of diapause, as in the case of pupae receiving a single dose of 20 jug, those individuals which initiated development within the first 4 weeks after pupation showed typical JH-induced abnormalities, often limited to the compound eyes. ECDYSIOTROPIC ACTIVITY OF JUVENILE HORMONE 301 100 *- 80 c 0> E a. 2 o> I 60 ^40 o^ > O E 3 o 20 0 I I 46 8 10 12 Weeks After Pupation 14 16 FIGURE 1 . Initiation of development by groups of short day, diapause-destined pupae receiving topi- cal application of JHA on the day of pupal ecdysis. Dosages of hormone per pupa were 200 ng (•), 20 ng (V), 2 jug (D). Hormone was delivered in 2 jil acetone and one group of pupae received topical application of 2 fil acetone only (A). Pupae were examined for tracheal apolysis at weekly intervals. The cumulative percentage of each group demonstrating apolysis is plotted as a function of time after pupation. The lines are drawn by inspection. Ecdysiotropic effects of JHA We next inquired whether the acceleration of development witnessed in prepara- tions treated with JHA was accompanied by an elevation of the ecdysteroid liter. Three groups were established: intact diapause-destined pupae, sham-operated dia- pause-destined pupae, and diapause-destined pupae from which the brain had been extirpated on Day 2. To individuals in each of these groups we administered daily topical applications of 20 ^g of JHA beginning on pupal Day 3 and continuing until the onset of adult development. Hemolymph was collected on the day that tracheal apolysis was first observed and assayed for ecdysteroids by RIA. Ecdysteroid levels were also measured in hemolymph from two JH A-free control groups, one consisting of 4-week-old diapausing pupae, the other of individuals that had been in diapause for at least 4 weeks and had undergone tracheal apolysis during the 24 h preceding hemolymph collection. Intact JHA-treated pupae initiated development 5-14 days after initiation of 302 L. SAFRANEK AND C. M. WILLIAMS TABLE I Effects of JHA administration on the ecdysteroid liter at the outset of adult development Procedure Stage Number Ecdysteroid titer* SD pupa, intact Pupal diapause, 4 weeks 12 0.23 ±0.05 SD pupa, intact Tracheal apolysis 12 19±7 SD pupa, intact + JHA Tracheal apolysis 12 43 ± 12 SD pupa, sham + JHA Tracheal apolysis 5 49 ± 15 SD pupa, — brain + JHA Tracheal apolysis 10 30 ± 15 [ The ecdysteroid liter is expressed in Mg/ml 20-OH-ecdysone equivalents. treatment, 50% by the seventh day. Sham-operated individuals developed 7-18 days after initiation of treatment, 50% by the ninth day. All brainless individuals devel- oped 5-35 days after initiation of treatment, 50% by the eleventh day. The ecdyste- roid liters of these groups and of the control pupae are shown in Table I. Manifestly the JHA treatment had a marked ecdysiotropic effect on all individuals. The average ecdysteroid liter achieved at Ihe oulsel of developmenl was Iwo orders of magnilude higher lhan lhal lypical of diapausing pupae. Moreover, Ihe ecdysleroid levels noled in Ihese groups were higher even lhan Ihose lypically allained by inlacl unlrealed pupae al a similar developmenlal slage. This was Irue even of brainless JHA-lrealed pupae. No correlalion was seen belween Ihe ecdysleroid liler and Ihe day of Ireatment on which developmenl was firsl noled. Effects of allatectomy The CA were removed from 50 diapause-deslined SD pupae 6-12 h after pupal ecdysis. Anolher 50 were sham-operaled, while an addilional 100 served as unoper- aled conlrols. As shown in Figure 2, all Ihree groups underwenl a lypical and essen- lially idenlical diapause. Once apolysis was inilialed, Ihe individuals of all Ihree groups developed inlo morphologically normal adulls in Ihe usual period of 3 weeks. A duplicale experimenl on Ihree groups of 25 pupae provided subslanlially Ihe same resulls. The experimenl was repealed on Ihree groups of polenlially non-diapausing LD pupae. The resulls are summarized in Figure 3. Unoperaled conlrols inilialed devel- opmenl in an average of 4. 1 days, sham-operaled conlrols in 5.5 days, and allaleclom- ized pupae in 6.6 days. The one day delay of allaleclomized pupae relalive lo conlrols could have been due lo Ihe slighlly grealer injury involved in Ihe allaleclomy. Il could also reflecl Ihe pulalive role of Ihe CA as neurohemal organs in Ihis species (Nijhoul, 1975;GibbsandRiddiford, 1977; Agui et al., 1980;Carrow^«/., 1981). Assay of the JH activity of pupal corpora allata Secrelion of JH by pupal CA was assayed in bl larvae and in non-diapausing pupae as described in Malerials and Melhods. To Ihis end, groups of 1 2 non-diapaus- ing pupae received CA from Day 1 or Day 3 non-diapausing pupae or from Day 1 or Day 14 diapause-deslined pupae. Each assay pupa received one pair of CA. All individuals subsequenlly developed inlo morphologically normal adulls — a resull documenling Ihe inaclivily of Ihe pupal CA. By conlrasl, implanlalion of only a ECDYSIOTROPIC ACTIVITY OF JUVENILE HORMONE 303 IOO 80 1 CL £> > 60 40 20 0> 3 E 3 o 4 6 8 10 12 Weeks After Pupation 14 16 FIGURE 2. Lack of effect of allatectomy on the termination of pupal diapause. Diapause-destined (SD) pupae were operated 6-12 h after pupal ecdysis and were examined for initiation of development at weekly intervals. The three groups consisted of 50 allatectomized pupae (•), 50 sham-operated pupae (•), and 100 unoperated controls (A). The cumulative percentage of each group showing apolysis is plotted as a function of time after pupation. The line is drawn by inspection. single CA from a Day 1 fifth-instar LD larva into each of 12 additional non-diapaus- ing pupae produced a range of typical JH-related aberrations in all resulting adults. In additional tests, individual CA from the same classes of donors as above were implanted for bioassay into appropriate bl larvae. The average score of each group of 12 pupal CA was less than 0.5, indicating essentially no detectable activity. By contrast, the CA from a group of 1 2 early fifth-instar larvae averaged an essentially maximal score of 2.8. Thus, whereas CA from larvae at the outset of the fifth instar were highly active in both assays, none of the pupal CA possessed any detectable JH activity when similarly tested. DISCUSSION The foregoing findings leave little doubt that elevation of the ecdysteroid liter is the ultimate cause of the onset of adult development in pupae subjected to juvenoid 304 L. SAFRANEK AND C. M. WILLIAMS < ° Q_ Q_ K 0 UJ GO 10 5 0 Allalectomized rrj Sham-operated m Unoperated n 8 DAYS TO INITIATION OF ADULT DEVELOPMENT FIGURE 3. Effect of allatectomy on the initiation of development by non-diapausing (LD) pupae. Operations were performed 6-12 h after pupal ecdysis. Individuals were examined daily for tracheal apo- lysis. Each of the three groups comprised 15-18 pupae. exposure, just as in the case of untreated pupae. Thus, as illustrated in Table I, the ecdysteroid titer in JHA-treated preparations at the outset of adult development achieved average levels 100 times that typical of untreated diapausing pupae, and in fact exceeded that encountered in untreated individuals spontaneously initiating adult development. Since this same result was observed in brainless pupae, the brain was manifestly unnecessary for mobilizing ecdysone sufficient to provoke the devel- opmental response. Moreover, since the ecdysteroid titer at the outset of JHA-pro- voked development was actually greater than that witnessed normally, the accelera- tion of development occurring after JHA administration cannot be attributed to an enhanced tissue sensitivity to ecdysteroids, as has been suggested (Sehnal et al., 198 1 ). The most parsimonious hypothesis is that the administered juvenoid acted di- rectly on the PG to provoke the secretion of ecdysone. Nevertheless, this possibility has been questioned in the literature. Thus, the suggestion has been made that JH acts indirectly by mobilizing a diffusable factor from the fat body which in turn enhances ecdysone secretion by the PG; a direct action on the PG has been specifically denied (Gruetzmacher et al., 1984). Other experiments have been interpreted to indicate an effect of JH on PG through an indirect mechanism requiring the PG to be in situ (Sehnal et al., 198! ). The question of direct or indirect action of JH on the PG re- mains, in our opinion, unresolved; we shall consider it in further detail in a subse- quent communication. Though JH can accelerate the initiation of adult development of diapausing Man- duca pupae and has been shown to have pronounced ecdysiotropic effects at this stage, the hormone appears to play no role in normal adult development or in the maintenance or termination of diapause. Indeed, JH must be absent for normal de- ECDYSIOTROPIC ACTIVITY OF JUVENILE HORMONE 305 velopment to proceed. Thus as we have seen, the initiation of adult development could not be stimulated by JHA without also producing morphological aberrations in the resultant adults. Indeed, abnormalities could be brought about by doses of JH that were too low to accelerate development. In addition, as summarized in Figures 2 and 3, allatectomy produced virtually no delay in the development of either dia- pausing or non-diapausing pupae. Finally, pupal CA proved to be completely inactive in two different bioassays for JH. In conjunction with our previous results (Safranek and Williams, 1980) these findings make clear that the pupal brain's ecdysiotropic effects are not mediated through activation of JH production by the pupal CA. The present findings confirm and extend previous observations of Bradfield and Denlinger (1980) that allatectomy failed to alter the duration of the Manduca pupal diapause. Moreover, in that same study JH could not be detected in the hemolymph derived from diapausing pupae of several different ages — a finding consistent with the inactivity of pupal CA documented in the present investigation. The current re- sults also agree with prior findings documenting the lack of a role for JH in the initia- tion of adult development in silkworm pupae (Ichikawa and Nishiitsutsuji-Uwo, 1959; Williams, 1959, 1961; Gilbert and Schneiderman, 1959). One remaining question concerns the raison d'etre for the marked sensitivity to juvenoids during the pupal stage. Though the application of JHA can accelerate the termination of pupal diapause, the aberrant development induced by even low levels of JH makes clear that at least the early phases of adult morphogenesis require a JH- free environment. The pupal epidermis has long been known to retain from the larval stage the ability to respond to exogenous JH at the outset of adult development. Sim- ilarly the ecdysiotropic effects of JH administered during the pupal stage may reflect a normally unused sensitivity carried over from the larval period when JH is typically present and may have significant effects on ecdysone secretion. ACKNOWLEDGMENTS We acknowledge early support of these investigations by the N.I.H. and the subse- quent generous support of Rohm and Haas. LITERATURE CITED AGUI, N., W. E. BOLLENBACHER, N. A. GRANGER, AND L. I. GILBERT. 1980. Corpus allatum is release site for insect prothoracicotropic hormone. Nature 285: 669-670. CARROW, G. M., R. L. CALABRESE, AND C. M. WILLIAMS. 1981. Spontaneous and evoked release of prothoracicotropin from multiple neurohemal organs of the tobacco hornworm. Proc. Natl. Acad. Sci. USA 78: 5866-5870. BRADFIELD, J. Y., AND D. L. DENLINGER. 1980. Diapause development in the tobacco hornworm: A role for ecdysone or juvenile hormone? Gen. Comp. Endocrinol. 41: 101-107. GIBBS, D., AND L. M. RIDDIFORD. 1977. Prothoracicotropic hormone in Manduca sexta: Localization by a larval assay. J. Exp. Biol. 66: 255-266. GILBERT, L. L, AND H. A. SCHNEIDERMAN. 1959. Prothoracic gland stimulation by juvenile hormone extracts of insects. Nature 184: 1 7 1- 1 73. GRUETZMACHER, M. C., L. I. GILBERT, N. A. GRANGER, W. GOODMAN, AND W. E. BOLLENBACHER. 1 984. The effect of juvenile hormone on prothoracic gland function during the larval-pupal devel- opment of Manduca sexta: An in situ and in vitro analysis. J. Insect Physiol. 30: 33 1 -340. ICHIKAWA, M., AND J. NISHIITSUTSUJI-UWO. 1959. Studies on the role of the corpus allatum in the eri- silkworm, Philosamia cynt/iia ricini. Biol. Bull. 116: 88-94. NUHOUT, H. F. 1975. Axonal pathways in the brain-retrocerebral neuroendocrine complex of Manduca sexta (L.) (Lepidoptera: Sphingidae). ////. J. Insect Morphol. Embryo!. 4: 529-538. RIDDIFORD, L. M., AND A. M. AJAMI. 1973. Juvenile hormone: Its assay and effects on pupae of Manduca sexta. J. Insect. Physiol. 19: 749-762. 306 L. SAFRANEK AND C. M. WILLIAMS SAFRANEK, L., AND L. M. RIDDIFORD. 1975. The biology of the black larval mutant of the tobacco horn- worm, Manduca sexta. J. Insect Physiol. 21: 1931-1938. SAFRANEK. L., AND C. M. WILLIAMS. 1980. Studies of the prothoracicotropic hormone in the tobacco horn worm, Manduca sexta. Biol. Bull. 158: 141-153. SAFRA A:K, L., B. CYMBOROWSKI, AND C. M. WILLIAMS. 1980. Effects of juvenile hormone on ecdysone- dependent development in the tobacco hornworm, Manduca sexta. Biol. Bull. 158: 248-256. SAFRANEK, L., C. R. SQUIRE, AND C. M. WILLIAMS. 1986. Precocious termination of diapause in neck- and abdomen-ligated pupal preparations of the tobacco hornworm, Manduca sexta. Biol. Bull. 171: 126-134. SEHNAL, F., P. MAROY, AND J. MALA. 1981. Regulation and significance of ecdysteroid litre fluctuations in lepidopterous larvae and pupae. J. Insect Physiol. 27: 535-544. WILLIAMS, C. M. 1959. The juvenile hormone. I. Endocrine activity of the corpora allata of the adult cecropia silkworm. Biol. Bull. 116: 323-328. WILLIAMS, C. M., 1961. The juvenile hormone. II. Its role in the endocrine control of molting, pupation, and adult development in the cecropia silkworm. Biol. Bull. 121: 572-585. Reference: Bio/. Bull. 172: 307-314. (June, 1987) IMMOBILIZATION OF THE PREDATORY GASTROPOD, NUCELLA LAPILLUS, BY ITS PREY, MYTILUS EDULIS PETER S. PETRAITIS Biology Department, University of Pennsylvania, Philadelphia, Pennsylvania 19104-6018 ABSTRACT Slow moving and sessile organisms cannot escape predators and must rely on other defenses. Laboratory experiments show that mussels (Mytilus edulis) immobi- lize predatory gastropods with byssal threads. This study assesses the risk of immobili- zation faced by the gastropod, Nucella lapillus, under natural conditions. Nearly 30% of the N. lapillus within mussel beds are trapped, and the per day risk to a predator in a mussel bed is between 0.010 and 0.038. The mussels nearest to the predator attach 20 or more byssi to the body whorl of its predator's shell. Byssi are retracted, flipping the predator over. Nucella lapillus are rarely found in mussel beds (0. 10 indi- viduals per 0.25 m2 in beds and 2.38 per 0.25 m2 out of beds). It is not clear if N. lapillus are avoiding mussel beds for other reasons, such as exposure to predators, or are seeking alternative prey which do not co-occur with mussels. Nevertheless, mus- sels pose a danger to predatory gastropods. INTRODUCTION Predation poses one of the greatest risks for slow-moving and sessile organisms and often limits the distribution of many such species (Connell, 1961; Paine, 1974; Menge, 1976; Vermeij, 1978; Peterson, 1979). While some clonal sessile marine in- vertebrates can survive partial destruction by predators (e.g. Harvell, 1984), non- clonal individuals must inhibit access by predators or live in areas where predators are otherwise excluded (Schmitt, 1981;Moran, 1985; Warren, 1985). Mytilus is one of the most successful and ubiquitous genera on rocky shores and forms extensive beds that appear to be completely defenseless against predators. Other shelled marine invertebrates co-exist with predators by virtue of reinforced shells (Vermeij, 1978) or have distasteful tissue (Thayer, 1985), yet M. edulis (the common blue mussel) has neither defense. Paradoxically, M. edulis flourishes in asso- ciation with predators, and in sheltered bays in New England, mussels occur with a common predatory gastropod, the dogwhelk Nucella lapillus. Dogwhelks readily take mussels and can easily kill them by drilling through the mussel's relatively thin shell (Menge, 1978a, b, 1983; Hughes and Dunkin, 1984). This study assesses the effectiveness of a novel antipredator defense of Mytilus edulis against Nucella lapillus. Groups of mussels immobilize dogwhelks with byssal threads, which are secreted by a gland in the mussel's foot and are normally used by the mussel to attach to the surface. This defense by M. edulis against predators has been observed on the east coast in response to the presence of the oyster drill Urosal- pinx cinerea (Carriker, 198 1) and on the west coast of North America in response to attacks by Nucella emarginata and N. lamellosa (Wayne, 1 980). Similar responses by Received 15 October 1 986; accepted 17 March 1987. 307 308 P. S. PETRAITIS M. californianus to N. canaliculate, have been observed on the west coast of North America (T. H. Suchanek, pers. comm.). Studies with Mytilus edulis have been done only in the laboratory. Carriker's (1981) observations were made at 1 5°C, and U. cinerea were inactive at this tempera- ture. Wayne (1980) has documented a series of specific responses by mussels to whelks. Besides attachment of byssal threads, Wayne observed a mussel would open and close its valves and would extend its mantle. The responses were highly stereo- typed and occurred only in the presence of N. lamellosa and N. emarginata. The whelks were assumed to be at risk because they used their radulae to strike at the mussel's foot during attempts to attach a byssus (Wayne, 1 980). However, there has been no assessment of the risk to individual whelks. This report documents the distribution of N. lapillus with respect to M. edulis and the risk of being trapped by byssal threads under natural conditions. MATERIALS AND METHODS Observations were made between June 1983 and July 1986 just below Mean Low Water (-0.3 to 0.0 MLW) at five sites within protected bays on Swans Island, Maine (44° 10' N, 68° 25' W). Four sites (Long Cove, Mill Pond, Mill Pond South, and Mill Pond Point) on the eastern shore of Burnt Coat Harbor, and a fifth site on the eastern shore of Mackerel Cove were used. Long Cove was used only in 1983. Mill Pond South and Mill Pond Point sites were less than 200 m apart; the other sites in Burnt Coat Harbor were 500-1000 m apart. The substratum at all sites was a mixture of granite outcrops, boulders, and cobbles. Barnacles (Semibalanus balanoides) and mussels (Mytilus edulis) were the most common sessile organisms, and algae covered less than 5% of the surface (see Petraitis, 1987, for a more detailed description of sites). To document the distribution of Nucella lapillus in relation to mussel beds, the amount of surface covered by mussels and the abundance of N. lapillus inside and outside mussel beds was estimated. In June 1 986, transect lines (88- 1 25 m long) were run parallel to the shore, and the species present were noted at 1 m intervals. On 23- 24 June 1986, abundance of N. lapillus was sampled with 50 X 50 cm quadrats which were placed outside or inside mussel beds at each site. Mill Pond Point was also sam- pled on 3 June 1984. Data from the quadrats were bimodal, so Kruskal-Wallis tests (Sokal and Rohlf, 1981), which were corrected for ties, were used to test for differences. Nucella lapillus that were trapped within mussel beds were tallied over a four year period. At low tide, mussel beds were searched for dogwhelks, and each live dogwhelk was scored as trapped or free. Dogwhelks seen on the margins of the mussel beds were also counted, and some of these individuals were checked for evidence of byssi. No site was surveyed more than once a year. Two manipulations estimated the rate at which dogwhelks were trapped. In the first experiment, dogwhelks were caged in 1 5 X 45 X 50 cm baskets for 3 weeks with mussels covering the bottom of the basket. Baskets were made of stainless steel mesh with openings of about 5X5 mm. Clumps of mussels were carefully transferred into the cages three to five days before the start of the experiment. On 16 July 1983, three baskets with 25 dogwhelks per basket were started at Long Cove. Between 22 and 23 June 1 986, four baskets with 20 dogwhelks per baskets were started at Mackerel Cove, Mill Pond, and Mill Pond Point. Additional treatments with 20 dogwhelk shells per basket or 20 Littorina littorea (a herbivorous gastropod commonly found in mussel beds) per basket were used in 1986 to check if mussels were preferentially trapping IMMOBILIZATION OF NUCELLA 309 TABLE I Mean number of Nucella lapillus /w 0.25 m2 in and out of mussel beds Sampling Within mussel beds Outside mussel beds Location Mean S.D. n Prop. Mean S.D. n Prop. Mill Pond South 0.00 0.00 20 1.00 0.15 0.49 20 0.90 2.05 Mill Pond Pt. 84 0.23 0.51 26 0.81 6.03 7.36 29 0.17 25.58 Mill Pond Pt. 86 0.15 0.49 20 0.90 3.26 4.19 19 0.32 14.38 Mill Pond 0.05 0.22 20 0.95 1.90 3.57 20 0.55 8.84 Mackerel Cove 0.05 0.22 20 0.95 0.57 0.75 21 0.57 7.93 n gives the number of quadrats, and prop, is the proportion of quadrats without N. lapillus. Column T shows test values for Kruskal-Wallis tests. live N. lapillus. Three baskets of these additional treatments were set out at each site. Data from 1986 were analyzed as a two-way analysis of variance of treatments by sites. The three sites served as blocks, and the analysis was a mixed model with repli- cation. The data, proportions of individuals trapped per cage, were arcsine trans- formed for the analysis (Johnson and Klotz's transformation, see Sokal and Rohlf, 1981). The second manipulation was a mark-recapture study. Dogwhelks were marked with red paint and placed in the middle of a mussel bed. On 2 June 1984, 100 marked dogwhelks were released at Mackerel Cove, and individuals were recovered 4 days later. On 24 June 1986, 238 individuals were marked and half were released at Mack- erel Cove and half at Mill Pond Point. Recaptures were done three days later. At the same time in 1 986, 68 marked empty Nucella shells were set in mussel beds at Mack- erel Cove and Mill Pond Point to assess the rate at which empty shells were normally trapped and incorporated into mussel beds. RESULTS Nucella lapillus were rare in mussel beds, even though mussels were usually the most common sessile organism (Table I, Fig. 1). The overall average density of dog- whelks within mussel beds was 0. 10 per 0.25 m2 while the density outside beds was 2.38 per 0.25 m2. For each site except Mill Pond South, the difference was significant (Table I). In total, dogwhelks were found in only 12 of the 106 quadrats which were within beds, but in 58 of the 109 quadrats outside beds. The density outside beds is within the range of densities found at these sites at other times (Petraitis, 1987) but below the range of densities reported for more exposed locations (Lubchenco and Menge, 1978). Just over 27 percent of all live dogwhelks found within beds were tightly bound with byssal threads (average of entries in Table II). Over the four year period, only 293 individuals were found within beds, although 4186 dogwhelks were recorded on the margins of mussel beds. Although each site was searched as nearly as possible with the same amount of effort, the number of dogwhelks seen and trapped varied year to year and site to site (Table II). Position and number of the byssal threads on the trapped Nucella lapillus were very distinctive. The three or four mussels closest to, but not necessarily touching, 310 P. S. PETRAITIS Mill Pond South N=94 Mill Pond N=88 CD > O o 60 40 20 0 Bare ME SB CC Other c CD O CD CL Mill Pond Point N=125 Mackerel Cove N=96 60 40 20 0 FIGURE I. Summary of transect data. Abbreviations for species are: ME for Mytilus edulis, SB for Semibalanus balanoides, CC for Chondrus crispus, and OTHER for other algae. N is the number of points sampled. the dogwhelk attached 20 or more byssal threads to the body whorl of the dogwhelk. The threads appeared to be retracted, flipping the dogwhelk over so its foot did not touch the surface. Dogwhelks were often seen extending the foot in an attempt to right and free itself, although the inverted position and the collective strength of the threads made escape seem unlikely. Outside mussel beds, the proportion of dogwhelks with remains of byssal threads was highly variable but indicated many encounters between dogwhelks and mussels. From lowest to highest, the proportion of dogwhelks with byssi: 0.05 at Mackerel Cove (n : 137), 0.40 at Mill Pond Point (n = 311), 0.52 at Mill Pond South (n == 1 89), and 0.75 at Mill Pond (n == 1 07). There were significant effects of treatment and site in the analysis of variance of the 1 986 caging data (Fig. 2). However, only 3% of the total variance was due to differences among the three sites (see Sokal and Rohlf, 1981, for estimation of vari- ance components), and byssal threads were attached more often to live Nucella lapil- lus and empty shells of N. lapillus than to Littorina littorea. Empty shells were very frequently trapped, but these shells were normally found near the bottom of the mus- IMMOBILIZATION OF NUCELL.4 3 1 1 TABLE II Proportion of live Nucella lapillus trapped by byssal threads of mussels Dates of sampling Location 9-14Jul 1983 4-6 Jun 1984 2-5 Jul 1985 18-26 Jun 1986 Long Cove 0.42 ND ND ND (ND) Mill Pond S. and Pt. ND 0.24 0.07 0.52 (29/842) (14/229) (64/564) Mill Pond 0.42 0.16 0.36 0.71 (12/62) (25/955) (11/491) (17/177) Mackerel Cove 0.00 0.15 0.15 0.25 (47/97) (39/149) (40/151) (8/411) ND stands for no data. Ratios in parentheses are the total number of live dogwhelks (trapped and free) found in mussel beds over the total number seen in and out of beds while sampling. The proportion is based on the number found in the mussel bed, e.g., for Mill Pond in 1983, 0.42 = 5/12. sel clump and were attached with very few byssi. In contrast, the live N. lapillus which were trapped were usually on the top of the mussel clump and covered with many byssal threads. The low incidence of trapping L. littorea may be unrelated to the ability of mussels to distinguish between dogwhelks and periwinkles. Periwinkles are much more mobile than dogwhelks, and thus periwinkles may not be sitting in one spot long enough for a mussel to attach a byssus. Based on the caging study, the overall per day risk was 0.014 with a quarter of the dogwhelks trapped after three weeks. A dogwhelk had a half-life of 50 days, and an individual had less than a 1 in 5 chance of surviving its active season if it remained within a bed during summer and early fall (about 1 20 days). The average per day risk was 0.038 (S.D. = 0.020) for the Nucella lapillus in the mark-recapture study (Table III). Dogwhelks were trapped very quickly, and stood a 50:50 chance of being trapped in 2 to 4 weeks. Empty shells of dogwhelks which were marked and placed in mussel beds were also trapped but at a lower rate (Table III). For live A", lapillus, the estimates may be high because it was assumed that all unac- counted dogwhelks left the mussel bed. If these individuals remained in the bed but were undiscovered, then the average risk was reduced to 0.010 per day. Success at finding marked shells and dogwhelks did not differ; in 1986, 75.0% of the empty shells and 70.6% of N. lapillus were recovered. A two by two Chi-square test of independence, using data from 1986 in Table III, was not significant (x2 = 0.85 with one degree of freedom). DISCUSSION Observations and manipulations show that Nucella lapillus runs a substantial risk of being immobilized by byssal threads ofMytilus edulis. Given the level of risk, one would expect dogwhelks to avoid mussel beds even though mussels are a common prey item of dogwhelks (Menge, 1976; Hughes and Dunkin, 1984). Although sam- pling confirms this notion (Table I), it is possible that N. lapillus are not so much avoiding mussel beds but are searching for an alternative prey, such as the barnacle Semibalanus balanoides, outside of the mussel beds. It is also possible that predators 312 P. S. PETRAITIS 100 r- 80 - CD Q. Q. 03 C CD O i_ CD Q- 60 40 20 0 N=3 12 NL NL Shell FIGURE 2. Means ± 95% confidence limits with sample sizes (N) for percentage of individuals trapped by byssi in the caging experiments. The leftmost bar gives 1984 data; the rest for data collected in 1 986. The treatments are abbreviated as: NL for Nucella lapillits, LL for Littorina littorea, and SHELL for shells of Nucella lapillus. Blocks were ignored in calculating the means and confidence limits for 1 986 data; one observation is missing from the LL treatment. Based on analysis of variance of 1986 data, F = 26.77 for the effect of treatments (tested over interaction mean square which is 134.58 and has 4 degrees of freedom) and F = 4.44 for the effect of sites (tested over error mean square which is 60.37 and has 20 degrees of freedom). Main effects are significant at the 5% level, and there is no significant interaction. quickly remove N. lapillus from mussel beds or that mussel beds are an unsuitable habitat for other reasons. Nevertheless, the rarity of Nucella lapillus in mussel beds is striking because mus- sels cover more than 50% of the surface at all sites except Mackerel Cove (Fig. 1 ). If the proportion of N. lapillus that are outside mussel beds is adjusted for the area covered by mussels, then less than 10% of all dogwhelks are found within mussel beds. For example, at Mackerel Cove 36% of the surface is covered by mussels, and there are 0.05 dogwhelks per 0.25 m2 within beds and 0.57 dogwhelks per 0.25 m2 outside mussel beds (see Table I and Fig. 1 for data). Adjusting the densities for the proportion of area occupied by mussels, this means 95% of the dogwhelks are not in mussel beds (i.e., 0.95 = (0.64)(0.57)/[(0.64)(0.57) + (0.36)(0.05)]). The behavior of Nucella lapillus suggests large individuals of Mytilus edulis are dangerous prey. Nucella lapillus that have previously taken mussels drill well away from the mussel's foot and byssi, slightly ventral and anterior to the posterior adduc- tor muscle (Hughes and Dunkin, 1984). While Hughes and Dunkin (1984) suggest this position gives the highest caloric return, it is also likely to be the least dangerous. IMMOBILIZATION OF NUCELLA 313 TABLE III Mark and recaptures of live individuals and shells of Nucella lapillus Marked Recaptures Risk Location Object n Trapped Free Not in bed (#/day) Mackerel Cove 1984 Snails 100 4 43 4 0.022 Mackerel Cove 1986 Snails 119 10 50 29 0.061 Shells 68 0 49 0 0.000 Mill Pond Pt. 1986 Snails 119 3 31 45 0.031 Shells 68 3 50 0 0.019 Column labeled "Not in bed" gives the number found on the margins and outside of the mussel beds. Risk is the exponential rate of trapping. In addition, these "experienced" Nucella lapillus generally prey upon smaller mussels even though they do not provide the highest caloric return (Hughes and Dunkin, 1984), possibly because a small mussel cannot ward off an attack and thus involves less risk. Cooperative defense by group-living animals is well-known (e.g., Alexander, 1974; Buss, 1981), but its possible importance for sessile non-clonal organisms such as Mytilus edulis is not appreciated. Because Nucella lapillus is usually immobilized by several mussels, solitary mussels may be much more likely to be taken by N. lapil- lus. Thus mussel beds may provide a common, group protection and may explain the absence of solitary mussels in areas where dogwhelks are common. There are other possibilities. For example, mussels are also the prey of seastars, crabs, fishes, and shore birds (Mason, 1972; Menge, 1976; Edwards = >i.-',3 be ''"•<-• '^^T^ . ^v :~ X-'^"iiki::- A '' - •• wK Wr • 'X- ., > >>^^H^--A - - FIGURE 3. TEMs of nephridium and nephridioduct of mitraria larva ofOweniafusiformis. A. Trans- verse section through distal most portion of nephridium near point of union with nephridioduct. Note lysosome-like bodies (ly) within cytoplasm of podocytes. Scale = 1 jum. B. Transverse section through medial portion of nephridioduct. Note adhaerens junction (arrow) between apposed surfaces of doughnut- shaped cell and numerous endocytotic vesicles (arrowheads) within its cytoplasm. Cilia (ci) situated within its lumen extend from duct cells. Scale = 0.5 p.m. C. Ciliary basal body (bb), rootlet (it), and accessory centriole (ac) of a nephridioduct cell. Note microvilli (mv) within duct lumen and endocytotic vesicles (arrowheads) within cytoplasm of duct cell. Arrow indicates direction of nephridiopore. Scale = 0.5 ^m. D. Parasagittal section through distal most portion of nephridioduct. Note endocytotic vesicles (arrowheads) within duct cell cytoplasm. Arrow indicates direction of nephridiopore. Scale = 1 p.m. ac, acessory centri- ole; bb, basal body, be, blastocoel; ci, cilia; ecm, basal extracellular matrix; in, intestine; lu, lumen of nephridioduct; ly, lysosome-like body; mv, microvilli; po, podocyte; rt, rootlet. (Ruppert and Smith, 1985; in prep.). Based on these definitions, each mitraria ne- phridium is a giant protonephridium. Presumably the ECM between the podocyte pedicels forms the filtration surface while the podocyte and duct cilia produce the filtration pressure gradient. Blastocoelic fluid may be ultrafiltered across the ECM between the pedicels. Numerous endocytotic vesicles in the podocyte perikarya sug- gest that modification of the ultrafiltrate occurs within the nephridium. Additional NEPHRIDIUM OF THE MITRAR1A LARVA 321 modification may occur along the nephridioduct before the fluid is discharged at the nephridiopore. As suggested by Ruppert and Balser (1986) for nephridia of deuteros- tome larvae, osmotic fluid recovery across the body wall of the mitraria may be estab- lished by proteins in the blastocoelic fluid. The pore canal-hydropore complexes contribute during metamorphosis to the metanephridial systems of adult hemichordates and possibly asteroid echinoderms (Ruppert and Balser, 1986), whereas the nephridium of the mitraria undergoes histol- ysis (Wilson, 1932) and is superseded by a pair of metanephridia in the sixth adult segment (setiger five). Ultrastructurally, the mitraria nephridium is identical to those of larval enterop- neusts and asteroids, but are they homologs or analogs? Consideration must be given to comparative larval anatomy and functional constraints on nephridial design in larval body organization to provide a tentative answer to this question. We have pre- viously discussed the correlation of cilia-driven filtration nephridia (protonephridia) with the absence of blood vessels and the physical basis for the lack of circulatory systems in small animals, such as many larvae (Ruppert and Smith, 1985; in prep.). On the basis of functional considerations, protonephridia are expected in small animals. Comparison of the mitraria nephridium with that of a tornaria, with which it seems directly comparable, reveals histological similarities and anatomical dissimi- larities. The nephridia of both larvae are lined by an epithelium composed of mono- ciliated podocytes and myocytes. The myocytes form an apical muscle band that traverses the blastocoel and inserts on the episphere, but the point of insertion is different in the two larvae. In the mitraria, insertion is posterior to the cerebral gan- glion, whereas in the tornaria it is on the ganglion. Moreover, the mitraria nephridia are situated postorally and the nephridioducts open to the exterior through the first pair of segmental somites, whereas the nephridium and its dorsal duct in the tornaria are situtated preorally. In addition, the mitraria nephridia are paired while those of the tornaria are unpaired, although paired ducts, apparently atavisms, occur fre- quently in laboratory reared larvae (Gemmill, 1914; Ruppert, unpub. obs.). We conclude that the striking cytological and histological similarities between the mitraria nephridia and those of asteroid and enteropneust larvae reflect a cellular homology, the common possession of monociliated podocytes. Like myocytes or neurons which also occur widely in metazoans, podocytes per se are of limited useful- ness in the recognition of relationships. On the other hand, the anatomical differences cited above suggest mitraria nephridia, as organs, may not be homologous to those of hemichordates and echinoderms; at best they might be serial homologs. However, this does not weaken the archicoelomate affinities of Owenia because the mitraria nephridia compare favorably to the monociliated larval protonephridia of phoronids. Although Wilson (1932) rejected the correspondence of mitraria nephridia with typi- cal protonephridia (head kidneys) of other polychaete trochophores, we suggest that the two are homologous and that the differences encountered between them may be explained in terms of larval body size (Ruppert and Smith, in prep.). Owenia is viewed as being primitive within the extant Polychaeta because it pos- sesses several characters which are believed to be plesiomorphic within the Metazoa. These include an unspecialized monociliated epidermis (Gardiner, 1978, 1979), a basiepidermal nervous system (Gardiner, 1978, 1979), and muscle cells with a single rudimentary cilium (Gardiner and Rieger, 1980). Do the larval nephridia further corroborate the view that Owenia is primitive among the polychaetes? Polychaetes are, in basic structural design, vermiform, soft-bodied, segmented 322 P. R. SMITH ET AL. coelomates. Based on recent discussions of function, this implies that adult poly- chaetes rely on a hydrostatic skeleton for locomotion (Clark, 1 964), a blood vascular system for a through-flow internal transport system (Ruppert and Carle, 1983), and a metanephridial system for excretion (Ruppert and Smith, in prep.). If an unseg- mented small-bodied larva, such as a trochophore, is included in the basic polychaete life cycle, then on functional grounds it should have a protonephridial system (Rup- pert and Smith, in prep.). Therefore, it is possible that the stem species of the Poly- chaeta had a larval protonephridium and an adult metanephridium. In that case, neither the protonephridium nor the metanephridium is primitive to the other, rather they are co-primitive organs in polychaetes, each of which may be modified in partic- ular lineages (Smith and Ruppert, in press). O. fiisiformis possesses the nephridial design postulated above for the adult and larva of the polychaete stem species. The segmented adult of O. fiisiformis possesses a blood vascular system and metanephridia and the mitraria larva possesses a proto- nephridial system. Both the metanephridia of the adult and the protonephridia of the mitraria are composed of monociliated cells. Rieger (1976) has shown that the pres- ence of an epithelium composed of monociliated cells is a plesiomorphic feature within the Metazoa. This suggests that O. fiisiformis has preserved the original ne- phridial designs of the adult (a monociliated metanephridial duct) and the larva (a protonephridium composed of monociliated podocytes) of the polychaete stem spe- cies, thereby corroborating Owenia's position as the most primitive of the extant polychaetes. ACKNOWLEDGMENTS We are grateful to Dr. J. M. Turbeville for generously providing some of the em- bedded specimens used in this study. This study was supported by NSF Grant No. BSR-8408500 to E. E. Ruppert. LITERATURE CITED CLARK, R. B. 1964. Dynamics in Metazoan Evolution: The Origin oftheCoelom and Segments. Clarendon Press, Oxford. 3 1 3 pp. GARDINER, S. L. 1978. Fine structure of the ciliated epidermis on the tentacles of Owenia fusiformis (Polychaeta, Oweniidae). Zoomorphologie9l: 37-48. GARDINER, S. L. 1979. Fine structure of Owenia fiisiformis (Polychaeta: Oweniidae) and its significance for annelid phylogeny. Ph.D. Thesis, University of North Carolina, Chapel Hill. GARDINER, S. L., AND R. M. RIEGER. 1980. Rudimentary cilia in muscle cells of annelids and echino- derms. Cell Tissue Res. 213: 247-252. GEMMILL, J. F. 1914. The development and certain points in the adult structure of the starfish Asterias rubens. L. R. Soc. Land Phil. Trans. B 205: 2 1 3-294. GOODRICH. E. S. 1945. The study of nephridia and gential ducts since 1895. Q. J. Microsc. Sci. 86: 1 13- 392. HEIMLER, W. 1981. Untersuchungen zur Larvalentwicklung von Lanice conchilega (Pallas) 1766 (Poly- chaeta, Terebellomorpha) Teil II: Bau und Ultrastruktur der Trochophora-Larvae. Zoo/. Jb. Anal. 106: 236-277. HOLBOROW, P. L. 1971. The fine structure of the trochophore of Harmothoe imbricata. Pp. 237-246 in the The Fourth European Marine Biology Symposium. Cambridge University Press, Cambridge. MORGAN, T. H. 1894. The development of Balanoglossus. J. Morphol. 9: 1-86. PEMERL, J. 1965. Ultrastructure of the protonephridium of the trochophore larva ofSerpula vermicularis (Annelida: Polychaeta). Am. Zoo/. 5: 666-667. REMANE, A. 1963. The enterocoelic origin of the coelom. Pp. 78-90 in The Lower Metazoa: Comparative Biology and Phylogeny, E. C. Dougherty, ed. University of California Press, Berkeley. REMANE, A. 1967. Die Geschichte der Tiere. Pp. 589-677 in Die Evolution der Organismen, G. Herber, ed. Gustav Fischer, Stuttgart. NEPHRIDIUM OF THE M1TRARIA LARVA 323 RlEGER, R. M. 1976. Monociliated epidermal cells in Gastrotricha: significance for concepts of early meta- zoan evolution. Z. Zool. Syst. Evolutionsforsch. 14: 198-226. RUPPERT, E. E., AND E. J. BALSER. 1986. Nephridia in the larvae of hemichordates and echinoderms. Biol.Bull. 170: 188-196. RUPPERT, E. E., AND K. J. CARLE. 1983. Morphology of metazoan circulatory systems. Zoomorphology 102: 193-208. RUPPERT, E. E., AND P. R. SMITH. 1985. A model to explain nephridial diversity in animals. Am. Zool. 25: 40A. SIEWING, R. 1976. Probleme und neue Ergebnisse in der Grosssytematik der Wirbellosen. Verh. Dtsch. Zool. Ges. 65: 59-83. SIEWING, R. 1980. Das Archicoelomatenkonzept. Zool. Jb. Anal. Ontog. 103: 439-482. SIEWING, R. 1981. Problems and results of research on the phylogenetic origin of the coelomata. Pp. 123- 1 60 in Origine Dei Grandi Phyla Dei Metazoi. Academica Nazionale Lincei, Rome. SMITH, P. R., AND E. E. RUPPERT. Nephridia. In The infrastructure of the Polychaeta, W. Westheide and C. O. Hermans, eds. Gustav Fischer Verlag. Stuttgart (in press). WESSING, A., AND A. POLENZ. 1 974. Structure, development and function of the protonephridia in trocho- phores of Pomatoceros triqueter (Annelida, Polychaeta, Sedentaria). Cell Tissue Res. 156: 21- 33. WILSON, D. P. 1932. On the mitraria larva ofOwenia fusiformis delle Chiaje. Phil. Trans. R. Soc. Loud. B 221:231-334, WILSON, R. A., AND L. A. WEBSTER. 1 974. Protonephndia. Biol. Rev. 49: 1 27- 1 60. Reference: Biol. Bull. 112: 324-336. (June, 1987) FEEDING ACTIVITIES OF TWO SPECIES OF CLYPEASTER (ECHINOIDES, CLYPEASTEROIDA): FURTHER EVIDENCE OF CLYPEASTEROID RESOURCE PARTITIONING MALCOLM TELFORD, RICH MOOI, AND ANTONY S. HAROLD* Department of Zoology, University of Toronto, Toronto, Ontario, Canada M5S 1A1 ABSTRACT Clypeaster rosaceus and C. subdepressus have typical clypeasteroid feeding mech- anisms and collect participate material primarily by means of oral surface accessory podia with terminal suckers. They lack the specialized food collecting podia seen in sand dollars, and the poorly differentiated food grooves carry only part of the food material. Maps of surface ciliary currents are provided and it is shown that they are not part of the feeding mechanism. Feeding mechanisms are discussed with reference to clypeasteroid phylogenetics. Clypeaster rosaceus occurs in association with sea- grasses ( Thalassia testudinum) and feeds extensively on dead leaves. Clypeaster sub- depressus occurs on coarse biogenic sands and selects the upper size fractions (66% greater than 400 ^m) for ingestion. Both are essentially epibenthic in habit and lack specialized spines to keep sand and debris from the body surface. Instead, they secrete an extensive sheet of mucus which prevents particles from falling between the spines. The suckered accessory podia of C. rosaceus have a mean tip diameter of about 1 80 Mm, those of C. subdepressus about 1 30 /urn. Although Encope michelini may occur in mixed flocks with C. subdepressus or Leodia sexiesperforata, they all feed on different fractions of the sediment. Probabilities of podial-particle encounters are insufficient to account for these differences and it is suggested that the basis of resource partition- ing between sympatric species is due primarily to differences in size of food collec- ting podia. INTRODUCTION The Clypeasteroida is a large, diverse group of epibenthic and shallow burrowing echinoids ranging in size from the tiny fibulariids (<15 mm) to the huge plate-like clypeasterids (>180 mm). They are abundant in many tropical seas and extend into cold temperate regions (Mortensen, 1948). In a recent series of papers the morphol- ogy and feeding behavior of these urchins has been investigated. At the outset, Mooi and Telford (1982) accepted the prevailing hypothesis that sand dollars such as Echi- narachnius parma (Lamarck), ingested fine particles sieved through the aboral spine canopy. However, particles substantially larger than could be accommodated by the sieve mechanism (100 yum) were found in the gut, and this suggested that some other mechanisms might be involved. In a study of Echinocyamus pusillus (O. F. Miiller) Telford et al. (1983) developed a method for watching feeding from underneath, us- ing a horizontally mounted microscope with an inclined mirror. By this means, direct observations of podial and lantern tooth activities were made of clypeasteroids in their natural orientation. Using similar methods in a re-examination of Echinarach- Received 17 November 1986; accepted 19 March 1987. * Present address: Department of Biology, Memorial University, St. John's, Newfoundland, Canada A1B3X9. 324 FEEDING OF CLYPEASTER SPECIES 325 nius parma intended to determine the possible contribution of the oral surface to feeding, Ellers and Telford ( 1 984) concluded that most, and perhaps all, food gather- ing was accomplished by oral surface podia. This cast the first serious doubt on the sieve hypothesis. According to Telford et al. (1985), another major objection to the sieve mechanism is the physical impossibility of ciliary currents transporting material as proposed by Goodbody (I960), Seilacher (1979) and, indeed, as we ourselves thought at one time (Mooi and Telford, 1982). An entirely different mechanism has been proposed by Telford et al. (1985), based on the podial collection of particles which may subsequently be fractured by the lantern teeth. Although Kier (1974) had previously commented on the presence of "crushed" material in clypeasteroid guts, and Timko (1976) specifically described lantern activity in Dendraster excentricus (Eschscholtz), the lantern was thought to be inactive during feeding in Mellita quin- quiesperforata (Leske) (Lane, 1977). More recently, Telford and Mooi (1986) have described the restriction of clypeasteroid species to specific substrate types and within limited particle size ranges. There is sufficient difference in the types and sizes of podia (Mooi, 1986a, b) to form the basis for the resource partitioning which appears to take place between sympatric species (Telford and Mooi, 1986). The above studies have included the typical, little-specialized Echinarachnius parma (Echinarachniidae); three species, Encope michelini L. Agassiz, Leodia sex- iesperforata (Leske), and M. quinqiiiesperforata (Mellitidae), specialized for hydrody- namically active environments (Telford, 1983); and one species, Echinocyamus pusillus of a family (Fibulariidae) specialized by miniaturization. The research has concentrated particularly on two areas: (a) Those structures and activities implicated in the supposed sieve hypothesis, notably spine morphology and spacing, ciliary cur- rents and sites of mucus secretion; and (b) those involved in the proposed podial feeding mechanism, notably podial diversity, distribution, and dimensions. Our data have pointed to a widespread reliance on the podial mechanism. In this paper we extend our observations to the family Clypeasteridae. This family is part of the sister group to all other clypeasteroid families (Smith, 1984). Therefore, members of the Clypeasteridae are very important because their feeding behavior might reflect the ancestral condition for the rest of the Clypeasteroida. In spite of their great abundance, size, diversity, and probable ecological significance, no previous studies of feeding in any species of Clypeaster have been reported. Our study deals with C. rosaceus (Linnaeus) and C. subdepressus (Gray), both abundant in the Gulf of Mexico, Caribbean, and adjacent waters. We show that their feeding mechanism is fundamentally the same as that of other clypeasteroids, and relies on podial collec- tion and the crushing action of the lantern. Further evidence from species distribu- tions and podial data are offered in support of our hypothesis of resource partitioning. Finally, an account of ciliary currents is given, which indicates their uniformity and ubiquity among the Clypeasteroida and confirms that, as in other species, they are not involved in food transport. MATERIALS AND METHODS Specimens of C. rosaceus were collected at several sites in the Florida Keys (Pigeon Key, Bahia Honda, Long Key) in the summers of 1982 and 1984, at depths of 2-5 m. Patches of Thalassia testudinum were found at all collection sites. At some sites, the turtle grass was so dense that only small, isolated areas of sand were visible. Clypeaster subdepressus was collected 2-3 miles off Pigeon Key, at depths of 10-15 m. The substrate consisted of coarse carbonate sand with conspicuous coral and shell debris and without significant plant cover. For live observation, specimens were maintained in the laboratory in running seawater on 10- 1 5 cm of sediment from the 326 M. TELFORD ET AL. collection sites. Other specimens were fixed in the field by injection of 20% formalin in seawater and immediate immersion in 10% formalin in seawater. These specimens were used for examination of gut contents. For measurement of podial tip diameters, specimens were relaxed by slow addition of ethanol and fixed as described above. Forty-eight hours after fixing, specimens were transferred to 3% formalin in seawater for storage. The natural buffering capacity of seawater is sufficient to prevent etching of specimens in dilute formalin for periods of several weeks. Substrate samples were fixed by addition of concentrated formalin to give a final dilution by volume of 5% in seawater. Additional bathymetric and distributional data were obtained from the collection records of the Florida Department of Natural Resources (FDNR, Marine Laboratory, Tampa), from collections in the United States National Museum (USNM, Smithsonian Institution, Washington), and from those in the Museum of Comparative Zoology (MCZ, Harvard University, Cambridge, Massachusetts). Representative samples of primary and miliary spines from the aboral surface, oral surface, ambitus, and infundibulum were drawn with the aid of a camera lucida. Podia were classified as accessory, food groove and large food groove types (Mooi, 1 986a, b) and tip diameters of relaxed specimens were measured by eyepiece microm- eter. Plastic thick sections (approximately 1 ^m) of spines and podia were prepared using metachromatic staining methods (toluidine blue, cresyl violet) to determine the distribution of mucus secreting cells, as described by Mooi ( 1986a). Surface ciliary currents were made visible by carmine particles or black ink and mapped as described by Telford et al. (1985). Observations of feeding activity on the oral surface were made using an inclined front-silvered mirror and stereomicroscope mounted horizontally on a swinging boom, as previously described (Telford et al, 1983; 1985; Telford and Mooi, 1986). Analysis of particle size-frequency distribu- tions of natural substrates and of material collected from the food grooves and infun- dibulum, was performed using a slight refinement of the method described by Telford and Mooi (1986). Small samples were strewn on a microscope slide and all particles within several fields of view were drawn using the camera lucida. The drawings were then spread on a "Summagraphics" M 1 103 digitizing tablet and their greatest orthog- onal diameters were measured and recorded directly into a computer. A 1-mm stage micrometer was drawn with the camera lucida at the same enlargement for calibra- tion. After determination of mean dimensions, the particles were sorted into size classes (<24.9, 25-49.9, 50-99.9, 100-199.9, 200-399.9, 400-799.9, >799.9 fim). Elongation (Leeder, 1 982) was calculated as the ratio of lesser diameter to greater diameter. This particle size analysis cannot follow standard sedimentological proce- dures because the samples collected from individual animals are far too small for sieve analysis. Substrate and food particle size-frequency distributions were com- pared by Chi square. All statistical determinations followed the procedures of Sokal and Rohlf( 1981). RESULTS The spines of these Clypeaster species are differentiated into two principal types, primaries and smaller miliaries interspersed between them. Both types vary in size, depending on their location on the test (Fig. 1 ). Histological examination showed that primary and miliary spines possess mucus secreting cells along their shafts and at their tips. In C. rosaceus the primary spines are alike in shape in all regions of the body. They range continuously in size from the small aboral spines (1.7 mm, length/ width ratio 7.7) to the large infundibular spines (5.2 mm, 1/w 12.6). Those of C. subdepressus are mostly shorter and more slender. The aboral primaries differ from the others in having slightly inflated tips. They are about 1 mm in length, with a FEEDING OF CLYPEASTER SPECIES 327 Gt lillll.ll'llllllH, ^,,,m » « n_ FIGURE 1. Spine types of Clypeaster rosaceus (above scale bar) and C. subdepressus (below scale bar). (A) aboral primary, (B) oral primary, (C) primary from around periproct (D) primary from inside infundibulum, near peristome, (E) aboral miliary, (F) oral miliary, (G) miliary from inside infundibulum, (H) miliary from food groove. length/shaft-width ratio of 9.6 and a length/tip-width ratio of 6.6. In both species, the primary spines bear two well-defined rows of cilia which extend up to 1 mm from the base of the shaft. The miliary spines bear similar rows of cilia extending almost to the tip. The rows of cilia are easily discernible in living specimens as well as histological and SEM preparations. Unlike members of any other genus which we have exam- ined, both species possess very small miliary spines, approximately 0.6 mm long, within the food grooves. Elsewhere, the miliaries range from 0.9 to 1 .2 mm in length in C. subdepressus and are slightly longer in C. rosaceus. Mooi ( 1 986a, b) has described the histology and diversity of nonrespiratory podia in all clypeasteroid families. There are two principal types in Clypeaster, both of which bear terminal suckers. They are (1) accessory podia and (2) modified accesso- ries or food groove podia. Examination of sectioned material showed these two types 328 M. TELFORD ET AL. TABLE I Mean diameters of podial tips (^m) for two species of Clypeaster C. rosaceus C. subdepressus Mean ±S.D. n Mean ±S.D. Aboral accessories 90 199.9 10.35 90 145.0 7.86 Oral accessories 90 181.5 8.58 90 129.8 -~- 8.78 Food groove (fg) 60 166.1 10.27 60 140.3 "-9.42 Large fg podia 12 234.2 24.39 12 196.2 23.45 of podia to be well supplied with mucus secreting cells. Almost the entire oral surface supports accessory podia, with only narrow zones in the center of each interambula- crum lacking them. On the aboral surface they are distributed in triangular patches tapering from the ambitus into the petaloids. Accessory podia are highly extensible and very mobile. The food grooves are flanked by modified accessory (food groove) podia and have a narrow central area without podia along the suture line. Food groove podia are stubbier than the other accessories, but are none the less mobile and may extend to more than double their resting length. In adults the buccal podia are the first of some 10-15 pairs of considerably enlarged podia flanking each food groove adjacent to the mouth. These podia are extremely extensible and capable of reaching well beyond the tips of the infundibular spines. Table I provides measurements of tip diameters of accessory podia from three ambulacra as well as large food groove podia for one adult specimen of each species. Ciliary currents in species of Clypeaster, although similar to those described for other clypeasteroids, are sufficiently different in some details to warrant reporting. Over the aboral surface the flow is essentially centrifugal (Fig. 2A). The petaloids are an exception where a counter current system exists between external ciliary flow and water- vascular fluid of the respiratory podia. On the oral surface of C. subdepressus (Fig. 2B) flow converges on the three anterior-most food grooves and follows them centripetally towards the mouth. Posterior to the peristome there is a very small zone of centripetal flow and an extensive area of flow convergent on the midline. This latter flow traverses the two posterior food grooves. Around the periproct (Fig. 2C) there is another area of convergent flow. In all areas of convergent flow there must, of course, be a change of direction away from the test (shown by circled dots in Fig. 2) to satisfy the principle of continuity (Vogel, 1981). As centripetal currents flow towards the center they, too, are constantly converging and must, therefore, have some flow exiting the spine field and directed away from the test. Within the infundib- ulum the centripetal flow reaches almost to the peristome and passes along the shafts of the circum-oral spines. As a result of this convergence, there is a downward flow towards the substrate, in the center of the infundibulum. The decreased velocity of flow in these areas of directional change can lead to a build-up of ink or carmine particles. In naturally oriented Clypeaster, the bulk of these particles fall away from the test onto the substrate below the infundibulum. Along the posterior interambula- crum and at the posterior margin of the periproct, convergent flows are mutually deflected towards the substrate, with similar loss of paniculate matter suspended in these flows. On the oral surface of C. rosaceus (Fig. 2D), the centripetal flow in the broad infundibulum is more pronounced, including all five food grooves. The genus Clypeaster is the only one in which ciliary flow along the food grooves has been ob- served, and, as already remarked, it is the only one with miliary spines within the food grooves. Irregularities in the sediment surface and activity of the spines ensure FEEDING OF CLYPEASTER SPECIES 329 FIGURE 2. Ciliary currents on Clypeaster subdepressus (A-C) and C. rosaceus (D). (A) aboral sur- face, (B) oral surface, (C) area around periproct (boxed in B), (D) oral surface. In A, B, and D, large arrows indicate strong ciliary flow, smaller arrows weaker flow. In B to D, circles with central dot indicate downward flow towards the substrate in living specimens (towards the viewer in these diagrams). that neither species is ever likely to be so closely in contact with the substrate as to seriously impede escape flow around the margin. The two species considered here are very different in shape and behavior. When resting on a more or less flat surface (Fig. 3), C. rosaceus contacts the substrate around the edge and the broad, deep infundibulum places the mouth high above the surface. The rounded ambitus is not obscured by sediment particles but the aboral surface may be partially covered by dead filaments of Thalassia or other material (Fig. 3 A), held in place by accessory podia. Clypeaster subdepressus is much flatter on the oral surface (Fig. 3B). Numerous particles picked up by podia around the ambitus are passed upward to the aboral surface by the combined action of spines and podia. On the aboral surface, particles move principally towards the posterior with a slight 330 M. TELFORD ET AL. FIGURE 3. Diagramatic longitudinal sections through (A) Clypeaster rosaceus and (B) C. subdepres- sus showing epibenthic habit, degree of aboral surface cover, and relative positions of the peristomes. A, anus; I, infundibulum; L, lantern; S, test support pillar; T, tooth; Th, piece of dead Thalassia. convergence on the midline. So far as we could tell, passage of material over the aboral surface was in no way related to the feeding activity of C. subdepressus. The aboral surface is largely covered by a thick mucus sheath secreted by the podia and spines. When the animals are resting or moving slowly, this sheath may be almost stationary but at times of greater activity it is moved steadily towards the rear by the action of spines, and it is replenished at the front and sides. Small sand grains and other fine particulate materials are trapped in the mucus, move with it, and are re- turned to the sediment as the sheet drops off the posterior edge. When complete, very few particles penetrate the sheet. Those which were pushed through with forceps, to fall among the spine bases, were rapidly moved to the ambitus by the combined ac- tion of pedicellariae and ciliary currents. The pedicellariae are so numerous that every spine is accessible to at least half a dozen of them. During feeding, both species collected particulate material by means of oral sur- face accessory podia. Particles were then passed from podium to podium towards the mouth, sometimes without reference to the food grooves and sometimes along them. Transported particles moved up the slope of the infundibulum which, especially in C. subdepressus, became quite crowded with food material. The food grooves are rather unspecialized, the spines and podia flanking them being little differentiated from those adjacent to them. Small particles ( 100-200 /urn) tended to become incor- porated into indistinct mucus-bound strands; larger particles frequently moved alone but with adherent debris. On arrival at the peristome, particles were steered and pro- pelled into the mouth by the large food groove podia and the circum-oral spines. Ingested particles were pulverized by the lantern teeth, which could be glimpsed peri- odically during observations of feeding. However, some particles, particularly large ones such as mollusc shells, were occasionally rejected. This was achieved by simply releasing the particle, or sometimes it was swiftly and methodically passed back to the ambitus against the prevailing centripetal particle movement. Food material ingested by C. rosaceus included fragments of corals, coralline al- gae and large amounts of dead Thalassia. No attempt was made to determine the particle size-frequency distributions of substrate and ingestate of C rosaceus because the presence of so much Thalassia vitiates the method. The material collected by C. subdepressus was analyzed and compared to the natural substrate (Fig. 4). The surrounding sediment was devoid of particles below 50 ^m, had over 64% between 100 and 400 /xm, and 35% larger than that. The mean grain size was 337.4 ± 200.9 /urn with a mean elongation value of 0.69. This represents a poorly sorted sediment from which the very fine material has been removed by current activity. Within the food grooves and infundibulum, only 33% were in the 100-400 ^m fraction and 26% were greater than 800 ^m, compared to only 5% in the substrate. The mean grain size was much larger, 582.2 ±417.6 /um. These differences are statistically highly signifi- cant (P <^ 0.001), but the elongation value (0.71) was not significantly different. The FEEDING OF CLYPEASTER SPECIES 331 sediment • . o o c in t c NOIilSOdlNO C infundi bulum — 799.9 Mm, respectively. intestinal contents of C. subdepressus consisted mostly of a chalky paste with a few larger fragments. When the material is passed as feces, it is loose, not mucus bound and not particularly adherent. The stomach contents of C. rosaceus included pulver- ized material and conspicuous amounts of shredded leaves. DISCUSSION Clvpeaster species of the western Atlantic are found on biogenic sediments such as algal sands and shell debris (Serafy, 1979). According to Kier and Grant (1965), C. subdepressus occurs on substrates where it can burrow without interference from turtle grass roots. We never found this species properly burrowed. Usually individuals were extensively covered by substrate material but were always conspicuous, and never beneath the level of the surrounding substrate surface (Fig. 3B). The unspecial- ized spination of Clvpeaster species and secretion of the mucus sheet both point to their essentially epibenthic habit. In C. subdepressus the aboral primary spines (Fig. 1 A) are only slightly expanded at their tips, unlike the club spines of the truly burrow- ing mellitids (Telford el al, 1985; Telford and Mooi, 1986). The miliary spines (Fig. IE) have simple rounded tips and lack any crown-like expansions or terminal sacs (Mooi, 1986c). Both C. rosaceus and C. subdepressus produce copious amounts of mucus from spines and aboral podia. Mucus secretion by podia is a common occur- rence among clypeasteroids (Mooi, 1986a, b). Goodbody (1960) suggested that the miliary spines of Leodia were sites of mucus secretion, but Mooi (1986c) has shown that only Clvpeaster has mucus secreting cells in its miliary spines. The presence of mucus secretory cells in the primary spines further distinguishes the genus. In C. subdepressus the mucus covers much of the animal and forms a sheet or tent which traps fine and large particles. The mucus sheet prevents material from entering the spine field and is analogous in function to the canopy of club spines and miliary sacs found in the mellitids (Telford et al., 1985; Telford and Mooi, 1986; Mooi, 1986c) and the mucus envelope of spatangoids (Smith, 1 984). Beneath this protective cover- ing, ciliary currents sweep away any stray particles that might chance to penetrate it. Observing an undisturbed animal from underneath, with its aboral mucus sheet in place, there are very few particles to be seen in the ciliary currents. The interspine spaces are kept scrupulously free of debris. Although extensive, the mucus cover on 332 M. TELFORD ET AL. C. rosaceus was not as complete as that of C. subdepressus. Clypeaster rosaceus was never found covered by sand, but it does cover itself with large pieces of shell and plant debris (Fig. 3A), as is the habit of many regular echinoids. Presumably its ele- vated shape and concave oral surface are adaptations to life within the beds of turtle grass. The wide, deep infundibulum provides space for the easy movement of long pieces of Thalassia and other food material, free of interference from living turtle grass and spines. In the papers already cited, we have shown that the feeding mechanisms of several clypeasteroid species share features in common with other echinoid groups. Food material is collected and transported by the podia, in a manner analogous to that of spatangoids. In contrast to the spatangoids, both regular echinoids and clypeasteroids possess a fully functional Aristotle's lantern. The regular echinoids use the lantern to bite and scrape, whereas clypeasteroids use it to crush food material. The species of Clypeaster described in this paper are no exception. The great bulk of their food consists of particles collected and transported by oral surface accessory podia and it is thoroughly crushed by their large and powerful lanterns. We obtained no evidence that either species made use of fine material trapped in mucus. When carmine parti- cles or black ink were being used to trace ciliary currents, it was apparent that some of the material reached the mouth area although it was often caught in mucus long before that. Some of it also became adherent to sand grains and was subsequently ingested. The mucus sheet covering most of the aboral surface traps fine material, all of which is returned to the substrate as the sheet trails off the posterior edge. As we have shown elsewhere (Telford et al, 1985; Telford and Mooi, 1986) use of carmine and, to a lesser extent ink, can be misleading. A load of fine particles sufficient to be clearly visible vastly exceeds the amount normally encountered among the spines. It is possible that this stimulates mucus secretion. Furthermore, carmine particles are known to stick to almost any surface, including sand grains. We suspect that adhesion to the podia and some of the material in the food grooves is just such a passive effect, as we found with Mellita (Telford et al., 1985), Leodia and Encope (Telford and Mooi, 1986). We cannot say categorically that ciliary borne material is not ingested, in fact some probably is. We can unequivocally state, however, that there is absolutely no evidence of its systematic collection; that remarkably little fine material is appar- ent in the sediment or in the ciliary currents and, finally, that the overwhelming bulk of the material ingested is collected by the podia. Field observations on these species by Kier and Grant (1965) are fully compatible with our findings. Comparison of the size-frequency distributions of particles in the sediment and those from the infundibulum of C. subdepressus (Fig. 4) shows that large particles are collected with greater frequency than their occurrence in the sand. The oral surface accessory podia, responsible for collection of food, have suckered tips with a mean diameter of 1 30 nm (Table I). There appears to be a complex relationship between podia and particles which determines the composition of the material collected. Sev- eral factors are involved, including podial-particle encounters, handling efficiency, particle desirability, and the requirements for subsequent processing. The probability of a single podium encountering a grain of one particular size class can be estimated as the contribution of that size class to total sediment area. Thus, in a sample of natural sediment we would expect 54% of them to be in size class 5 (Fig. 4), for which the mean dimension is 300 /urn. The product of grain proportion (0.54) and grain area (3002) divided by the total area of all the grains is 0.2208. If particles were picked up in simple proportion to the frequency with which a podium encounters that grain size, then 22.1% should be in class 5, 40.8% in class 6 and 26.0% in class 7. Large particles are indeed favored, but simple encounter frequency alone cannot explain the observed size distribution of particles in the infundibulum. The relationship be- FEEDING OF CLYPEASTER SPECIES 333 TABLE II Percent by number of particles in natural sediment for habitat o/"Clypeaster subdepressus and estimated percentage by mass Observed Estimated Particle size % by number % by mass Class 3 1.0 0.0 Class 4 10.0 0.2 Class 5 54.0 9.1 Class 6 31.0 44.6 Class 7 4.0 46.0 Each size class represents twice the linear dimensions of the preceeding one. Mass is proportional to length3, or eight times the mass of the next lower particle size class. tween podial and grain dimensions is probably critical in particle handling. Thomas and Hermans (1985) have described the elegant duo-gland adhesive system responsi- ble for attachment and release of the podia of a starfish. Mooi ( 1 986a, b) has described the structure and action of the terminal suckers of clypeasteroid tube feet. Both pro- cesses, adhesion-release and suction, appear to operate in the collection and transport of particles. Our data suggest that particles of about five times the podium diameter might be optimal, and that below this size, effectiveness of the pick-up and transporta- tion mechanisms diminishes. Larger particles could be handled cooperatively by sev- eral podia at once but above about twelve times the sucker diameter, effectiveness again falls off. No doubt particles are not equally attractive for reasons other than their size. Ellers and Telford (1984) observed that the presence of diatoms dramati- cally stimulated feeding in E. parma, and Telford el al. (1985) recorded a dispropor- tionate accumulation of diatoms and foraminiferans in the food grooves of M. quin- quiesperforata. It appears that sensory structures in the tube feet (Mooi, 1986a, b) enable clypeasteroids to discriminate between particles. Finally, the preferred size of particles collected might also be related to the dimensions and mechanics of the lan- tern apparatus where subsequent processing must take place. It should be stressed that the data presented here consist of actual linear measure- ments of two dimensions and are therefore not directly comparable to sieve data. We have avoided using the standard sieve particle size classes, usually expressed in units, because that would invite spurious comparisons. It is apparent in Table II that when our data are transformed to approximate mass data (proportional to length3), an entirely different impression of sediment composition is obtained. Grains are picked up individually during feeding and in our opinion frequency of particle size occurrence is more important in their selection than the accumulated mass passing through a screen, which is the form of data normally reported in sieve analyses. Bathymetric range, geographical distribution, and the factors discussed above could provide the basis for resource partitioning by clypeasteroid species (Telford and Mooi, 1986). Clypeaster subdepressus has a bathymetric range of 5-210 m (Serafy, 1979) but most commonly occurs well below 50 m (Table III). Kier and Grant ( 1 965) found it as shallow as 5 m, but mostly around 12 m, as we did ourselves. It inhabits coarse biogenic sand and we have been unable to find a reference to any Clypeaster species occurring on siliceous, terriginous sediments. This does not mean that some deep water species, such as C. lamprus H. L. Clark, could not occur on non-biogenic sediments. Although we found C. subdepressus co-occurring with Encope michelini at several localities, we never found it in association with Leodia sexiesperforata. However, the latter two species were often found in mixed flocks from 3-10 m on 334 M. TELFORD ET AL. TABLE III Bathymetric ranges, substrate preferences, andpodial diameters of clypeasteroid species Natural substrate Mean sizes (^m) Common Size Species depth (m) Type Sorted (Mm) Food Podia C. subdepressus 10-20 c P 200-800 580 130 E. parma 0-50' s m 100-400 2 90 M. quinquiesperforata 0-4 s w 100-400 170 70 E. michelini 3-10 e m 100-400 180 100 L. sexiesperforata 2-5 c w 100-400 140 70 E. pusillus 10-20 s P >1000 350 60 1 Stanley and James (1971). 2 Method for collecting these data not developed at time of observation. Unless otherwise indicated, original data can be found in reports of Telford and co-workers: See litera- ture citations. Substrate types were categorized as biogenic carbonates (c), terrigenic siliceous sediments (s), or either (e). Degree of sorting was ranked as poor (p), moderate (m) or well (w). The particle size range listed includes 70-90% of the total number of particles in the sample. Mean food size refers to material collected by the podia and transported towards the mouth directly or via the food grooves. rather finer carbonate sands. As shown in Table III, Leodia generally occurs shallower than 5 m, Encope deeper. These bathymetric data from our field experience are very similar to those reported in the literature (Mortensen, 1948; Serafy, 1979; inter alia). For example, Kier and Grant (1965) found Leodia in 3-8 m depths, and they specifi- cally commented on the preference of C. subdepressus for greater depths. It is quite apparent from these data and other collection records (USNM, MCZ, and FDNR) that L. sexiesperforata and C. subdepressus have almost completely separate bathy- metric and substrate size ranges. Encope michelini occupies an intermediate range. It overlaps with L. sexiesperforata when that species is approaching its preferred depth limit and on relatively coarse sand, as we found at Torch and Long Keys, Florida. When E. michelini is itself near its preferred habitat limits, it overlaps with C. subdepressus, as at Pigeon Key, Florida. Detailed field data for all of these species are scarce, especially substrate analyses, and further evidence regarding their spatial separation is urgently needed. None the less, the data presented here and in Telford and Mooi (1986) indicate that within their broad geographical ranges, habitat prefer- rences tend to keep the species separate (by bathymetric range, substrate type, and size, etc.). Where species do co-occur (L. sexiesperforata with E. michelini and E. michelini with C. subdepressus), differences in podial sizes affecting particle collec- tion and transport, as well as lantern sizes affecting ingestion and crushing, cause them to feed preferentially on different particle fractions. This is of considerable eco- logical and biogeographical interest and has previously received very little attention. We have summarized all of the available data on common depth of occurrence, as well as particle and podial dimensions, in Table III. Generally, the mean food particle size is around two to five times the podial diameter and drawn from the most abun- dant particle size classes in the surrounding substrate. The exception is the minuscule Echinocyamus pusillus which has the smallest podia but lives in the coarsest substrate from which it selects relatively scarce small particles during feeding. This species does not actually swallow many of its food particles, it scrapes them with the lantern teeth and then releases them (Telford et al, 1983). Unlike other clypeasteroid species, Echi- nocyamus pusillus lives in frequently disturbed environments, where it nestles be- tween pebbles. FEEDING OF CLYPEASTER SPECIES 335 Populations of clypeasteroids are often dense and must play a significant role in reworking the sediment and recycling nutrients. The absence of very fine particles in many of the substrates examined suggests that clypeasteroid fecal material, released on or close to the surface, must be swept away by water currents or more rapidly dissolved as a result of increased surface area/volume ratios. Clypeaster rosaceus is specialized for existence among Thalassia plants and for a diet mostly of dead leaves. It is sometimes found adjacent to turtle grass patches in open sandy areas with En- cope, where its diet is augmented by relatively large fragments ofHalimeda and other coralline algal debris, pieces of coral and even bits of mollusc shell. Thus C. rosaceus also appears to feed mostly on resources not specially sought after by scutelline species with small podia. According to Kier and Grant (1965) it may also co-occur with C subdepressus. Whether or not the two species then select essentially the same particle fractions is not known. This study of feeding in Clypeaster species is also important because of its phylo- genetic implications. The Clypeasteridae is part of the sister group (Clypeasterina) to the remaining members of the order Clypeasteroida (Smith, 1984). The feeding mechanism in Clypeaster is simpler and less refined than that of scutelline sand dol- lars such as E. parma (Ellers and Telford, 1984) or the mellitids (Telford et al, 1985; Telford and Mooi, 1986). In Clypeaster the food grooves are simple, straight furrows and carry only a fraction of the food. In addition, they are unique in having small miliary spines along their length. In Echinarachnius and, more particularly, the Mel- litidae, they are precisely defined, extensively branched, and convey all of the food to the mouth. Associated with the food grooves is the greater specialization of the podia. In Clypeaster there is little differentiation of the podia flanking the food grooves and those in adjacent areas. In the scutellines specialized podia actually occupy the food grooves and are solely responsible for moving food towards the mouth. Oral surface accessory podia in Clypeaster are not morphologically distinguishable from those of the ambitus or the aboral surface. In scutellines the oral surface bears barrel-tipped podia specialized for collection and transport of food material (Mooi, 1986a, b; Tel- ford et al., 1985; Telford and Mooi, 1986). In the most highly derived scutellines (the mellitids) these barrel-tipped podia are further divided into those specialized for gathering particles from the sediment and those for transporting them to the food grooves. Feeding by podial collection of sand also occurs in cassiduloids such as Cas- sidulus caribbearum (Lamarck) (Gladfelter, 1974) and Apatopygus recens (Milne- Edwards) (Higgins, 1974), which do so without the aid of food grooves. Since the cassiduloids represent an outgroup for the clypeasteroids (Smith, 1984), this feed- ing mechanism is plesiomorphic for the Clypeasteridae and culminates in the apo- morphic, highly specialized podia and food grooves of the Mellitidae. The absence of such specialized structures in Echinocyamus (Fibulariidae) (Telford et al., 1983) is an apomorphy associated with miniaturization and is not indicative of the ancestral condition. ACKNOWLEDGMENTS This work has been supported by the Natural Sciences and Engineering Research Council of Canada through Operating Grant #A4696. We are indebted to the Direc- tors and Staff of the Sea World Shark Institute, Long Key, Florida, and the Pigeon Key Laboratory, University of Miami, for use of laboratory facilities and boats; also to the curatorial staff of the United States National Museum, the Museum of Com- parative Zoology, and the Florida Department of Natural Resources Marine Labora- tory, for loans of specimens. Their assistance was greatly appreciated. We also thank our colleague Hugh Griffith for his help in the field and discussions of the work in progress. 336 M. TELFORD ET AL. LITERATURE CITED ELLERS, O., AND M. TELFORD. 1984. Collection of food by oral surface podia in the sand dollar, Echina- rachnius parma (Lamarck). Biol. Bull. 166: 574-582. GLADFELTER, W. B. 1978. General ecology of the cassiduloid urchin Cassidulus caribbearum. Mar Biol 41: 149-160. GOODBODY, I. 1 960. The feeding mechanism in the sand dollar Mellila sexiesperforata (Leske). Biol. Bull. 119:80-86. HIGGINS, R. C. 1 974. Observations on the biology ofApatopvgus recens (Echinoidea: Cassiduloida) around New Zealand. / Zool Lond. 173: 505-5 16. KIER, P. M. 1974. Evolutionary trends and their functional significance in the post-Paleozoic echinoids. J. Paleontol. 48 (suppl) Paleontol. Soc. Mem.: 5: 1-95. KIER, P. M., AND R. E. GRANT. 1965. Echinoid distribution and habits. Key Largo coral reef preserve, Florida.. Smithsonian Misc. Collect. 149(6): 1-68. LANE, J. M. 1 977. Bioenergetics of the sand dollar, Mellita quinquiesperforata (Leske, 1 778). Ph.D Disser- tation, Department of Biology, University of South Florida, Tampa. LEEDER, M. R. 1982. Sedimentology: Process and Product. George Allen and Unwin, London. 344 pp. Mooi, R. 1986a. Non-respiratory podia of clypeasteroids(Echinodermata, Echinoides): I. Functional anat- omy. Zoomorphology 106: 21-30. Mooi, R. 1986b. Non-respiratory podia of clypeasteroids (Echinodermata, Echinoides): II. Diversity. Zoo- morphology 106: 75-90. Mooi, R. 1986c. Structure and function of clypeasteroid miliary spines (Echinodermata, Echinoides). Zoomorphology 106: 212-223. Mooi, R., AND M. TELFORD. 1982. The feeding mechanism of the sand dollar, Echinarachnius parma (Lamarck). Proc. Int. Echinoderms Conf., Tampa Bay, 1981. Pp. 51-57. MORTENSEN, T. 1948. A Monograph of the Echinoidea. IV. 2. Clvpeasteroida. C. A. Reitzel, Copenhagen. 47 1 pp. SEILACHER, A. 1979. Constructional morphology of sand dollars. Paleobiology 5(3): 19 1-22 1 . SERAFY, D. K. 1979. Memoirs of the Hourglass Cruises: Echinoids (Echinodermata: Echinoidea). Florida Department of Natural Resources, St. Petersburg, Florida. 120 pp. SMITH, A. B. 1984. Echinoid Palaeontology. George Allen and Unwin, London. 190 pp. SOKAL, R. R., AND F. J. ROHLF. 1981. Biometry. 2nd. ed. W. H. Freeman and Co., San Francisco, CA. 859 pp. STANLEY, D. J., AND N. P. JAMES. 197 1 . Distribution of Echinarachnius parma (Lamarck) and associated fauna on Sable Island Bank, Southeast Canada. Smithsonian Contr. Earth Sci. 6: 1-24. TELFORD, M. 1 983. An experimental analysis of lunule function in the sand dollar Mellita quinquiesperfor- ata. Mar. Biol. 76: 125-134. TELFORD, M., A. S. HAROLD, AND R. Mooi. 1983. Feeding structures, behavior and microhabitat of Echinocyamus pusillus (Echinoidea: Clypeasteroida). Biol, Bull. 165: 745-757. TELFORD, M., R. Mooi, AND O. ELLERS. 1985. A new model of podial deposit feeding in the sand dollar, Mellita quinquiesperforata (Leske): The sieve hypothesis challenged. Biol. Bull. 169: 43 1 -448. TELFORD, M., AND R. Mooi. 1986. Resource partitioning by sand dollars in carbonate and siliceous sedi- ments: Evidence from podial and particle dimensions. Biol. Bull. 171: 197-207. THOMAS, L. A., AND C. O. HERMANS. 1985. Adhesive interactions between the tube feet of a starfish, Leptasterias hexactis, and substrata. Biol. Bull. 169: 675-688. TIMKO, P. L. 1976. Sand dollars as suspension feeders: A new description of feeding in Dendraster excentri- cus. Biol. Bull. 151:247-259. VOGEL, S. 198 1 . Life in Moving Fluids. Willard Grant Press, Boston. 352 pp. Reference: Biol. Bull. 172: 337-349. (June, 1987) MEMBRANE TRANSPORT OF PHOSPHATE BY HYMENOLEPIS DIMINUTA A. L. L. LIM AND Y. K. IP Parasitology Laboratory, Department of Zoology, National University of Singapore, Kent Ridge, Singapore 05 1 1 ABSTRACT The absorption of phosphate ion by Hymenolepis diminuta involved diffusion at all substrate concentrations tested. Mediated transport predominated at lower sub- strate concentrations and it exhibited saturation kinetics with Vmax and K,,, values of 0.056 yumol ethanol-extractable inorganic phosphate/g ethanol extracted dry wt/2 min and 1.19 mM, respectively, obtained from the Lineweaver-Burk transformation. The Vmax and K,,, values determined by the Hanes-Woolf plot were 0. 1 36 /umol etha- nol-extractable inorganic phosphate/g ethanol extracted dry wt/2 min and 3.64 mM, respectively. Absorption was influenced by changes in pH, temperature, and sodium ion concentration. Stable inorganic phosphate inhibited transport competitively. Various organic phosphates and common metabolic poisons introduced had no effect on transport. However, transport was inhibited by molybdate. From these data, it can be deduced that a separate transport system exists for phosphorus in H. diminuta. INTRODUCTION Intestinal parasites acquire inorganic substances largely, but not exclusively, from the intestinal contents of their hosts (von Brand, 1973). Read (1950) showed that Hymenolepis diminuta accumulated appreciable amounts of 32P in vivo only when the labelled inorganic phosphate (Pi) was given orally to the host under conditions ensuring that it reached the worm before being absorbed by the host tissues. However, a slow phosphate accumulation also occurred when it was administered to the host (rats) intraperitoneally (Read, 1950). Inorganic phosphate absorbed by parasites is generally incorporated rapidly into various organic compounds (Lesuk and Ander- son, 1 940; Scheibel £>/#/., 1968; von Brand etal, 1969). Von Brand (1973) stated that there was reason to assume that several inorganic substances, including Pi, entered cestodes by diffusion since Daugherty (1957) discovered that the Qi0 of 32Pi uptake by H. diminuta was very low. However, the authors were unable to find such informa- tion in the original article cited (Daugherty, 1957). Parasites usually grow at a much faster rate than free-living organisms, indicating that a large amount of Pi would be incorporated into phospholipids during membrane formation. Moreover, parasitic worms, like H. diminuta, produce enormous amounts of eggs containing many phos- phorylated compounds. Hence, it would be inefficient for the growth and develop- ment of the parasite if it was to obtain such an important ion through diffusion. Therefore, the present investigation was undertaken to further elucidate the nature Received 13 August 1986; accepted 24 March 1987. 337 338 A. L. L. LIM AND Y. K. IP of the transport mechanism involved in phosphate absorption in the rat tapeworm, H. diminuta. MATERIALS AND METHODS Hymenolepis diminuta was obtained from Carolina Biological Supply Co. (Bur- lington, North Carolina) as cysticercoids in adult Tenebrio sp. Male rats weighing 100 to 125 g were infected with 30 cysticercoids. Before and after infection, the rats were provided with water and food ad lib. Worms were flushed from the excised gut 10 days postinfection with Krebs-Ringer saline (KRT) containing 120 mM NaCl, 4.8 mM KC1, 2.6 mM CaCl2, 1.2 mM MgSO4, and 25 mM tris (hydroxymethyl)- aminomethane-maleate buffered at pH 7.4 (Read et al., 1963). For all experiments where the ionic product of calcium phosphate can easily exceed its solubility product, calcium was deleted from the KRT buffer used for preincubation and incubation. Identification oj the ethanol-extractable radio-phosphate Experiments examined the nature of the ethanol-extractable inorganic phosphate after a two-minute incubation period. Samples of 0.5 g worms were preincubated in KRT buffer, followed by two minutes incubation in 0.2 mM labelled phosphate me- dia. Upon removal from the incubation medium, the sample was rinsed rapidly in three changes of KRT buffer, blotted dry, and homogenized in six volumes of cold ethanol (80%). It was then centrifuged and the precipitate removed. To the superna- tant fluid, 0.05 ml of 25% barium acetate was added to every milligram of inorganic phosphate in solution and the pH brought to 8.3 with sodium hydroxide. The result- ing mixture was chilled and the precipitated phosphates collected by centrifugation, washed once with 70% ethanol, dried, and made into thick pastes with water. Dowex 50X8-200 (H+) was added until dissolution was complete. The supernatant fluid was separated from the resin, adjusted to pH 4.5 with 5 N sodium hydroxide, reduced to a small volume of 25 /ul, and used for paper chromatography and high voltage paper electrophoresis. The solvent system for paper chromatography was methyl cellosolve, pyridine, glacial acetic acid, and water (8:4:4: 1 ), while 0.2 M borate buffer at pH 9.5 was used for high voltage paper electrophoresis. Electrophoresis was performed at 20V per centimeter for 45 min. Radioactive spots were located with a Berthold Beta Camera LB 292 (Berthold, Germany) and further determined by liquid scintillation counting of 0.5 cm strips of the processed paper. Incubation studies Standard incubation conditions were two minutes at 37°C in a water bath shaking at 100 oscillations per minute. All samples were incubated in radioactive monoso- dium phosphate adjusted within a range of 0.04 to 1.50 /j.Ci/nmo\ depending on the nature of the experiment. For the concentration study (0.1 to 3.0 mM), a constant amount of 3H-polyethylene glycol (PEG 4000, NEN) of 1 .39 j*Ci was also introduced into the incubation media adjusted to 0.3 /uCi/ml. For all the other experiments, radioactive PEG was deleted from the incubation medium. Upon removal from the incubation medium the worms were rinsed rapidly in three changes of KRT buffer, blotted free of excess moisture on filter paper, and extracted overnight in tubes con- taining two milliliters of 70% ethanol. During the subsequent 24-h extraction, the tubes were agitated several times and one milliliter aliquots of the ethanol extracts were mixed with five milliliters of Biofluor (NEN) and radioactivity assayed using a Packard Tri-Carb 300 liquid scintillation spectrometer. Dry weights of ethanol- PHOSPHATE TRANSPORT BY H. DIMINUTA 339 0 020 0 010 234 Time ( min ) FIGURE 1. Absorption of 0. 1 mM radiophosphate by Hymenolepis diminuta as a function of time (min). V = jumol/g ethanol-extracted dry wt/2 min. Each point is an average of three determinations, r2 equals 0.87 extracted samples were obtained by heating worms at 1 10°C for five hours. Studies were also performed using balanced electrolyte solution (BES) containing: Na+ 133 mM, K+ 5 mM, Mg++ 1.2 mM, Cl~ 130 mM, HCO3~ 10 mM, mannitol 25 mM, and equilibrated with 5% carbon dioxide in nitrogen to pH 7.0 (Podesta et ai, 1977b); and balanced saline solution (BSS) containing: NaCl 120 mM, KC1 5 mM, MgCl2, NaHCO3 1 mM, 275 to 280 mosmol/kg, and equilibrated with 5% carbon dioxide in nitrogen to pH 7.4 (Ip and Fisher, 1982). Calcium and phosphate were deleted from the original formulae. Efficiency of ethanol extraction After ethanol extraction, worm samples were rinsed in 70% ethanol and digested in 0.4 ml of 60% perchloric acid and 0.8 ml of 30% hydrogen peroxide. The mixture was incubated at 70°C until colorless, cooled, and radioactivity assayed by liquid scintillation counting to determine the efficiency of extraction. Monosodium of 32P-phosphate and 3H-polyethylene glycol were obtained from Amersham International pic (Amersham, United Kingdom). Reagent grade chemi- cals and stable phosphates were obtained from commercial sources. Uptake velocities were expressed in micromoles of ethanol-extractable radioac- tive inorganic phosphate/g ethanol-extracted dry wt/2 min. Student's / test was used to evaluate differences between means. Whenever appli- cable, linear regression was performed using the Apple II microcomputer. RESULTS The absorption rate of 0. 1 mM radiophosphate by H. diminuta expressed as mi- cromoles of ethanol-extractable phosphate per gram ethanol extracted dry weight was constant over a period of four minutes (Fig. 1). Therefore, two minutes was chosen as a suitable incubation time for all incubation studies. 340 A. L. L. LIM AND Y. K. IP TABLE I Efficiency ofethanol extraction of absorbed radioactive phosphate Ethanol-extractable radiophosphate Conditions — — — — - — - — Total radiophosphate absorbed 0. 1 mM. 2 min 0.608 0.3 mM. 2 min 0.542 0.5mA/. 2 min 0.513 0.7 mM. 2 min 0.533 1.0mA/. 2 min 0.464 1.5mA/. 2 min 0.548 2.0 mM, 2 min 0.519 2.5 mM, 2 min 0.561 3.0 mM, 2 min 0.509 0.1mM0.5min 0.521 0.1 mM. 1.0 min 0.513 0.1 mM, 2.0 min 0.489 0.1 mM, 10.0 min 0.539 Each value is an average from two experiments. Identification of the ethanol-extractable radio-phosphate Information on the identity of the ethanol-extractable phosphate was obtained by determining the ratio of radioactive inorganic phosphate to radioactive total phos- phate in the ethanol extract. Results obtained by the various methods employed indi- cated that 95% of the extracted phosphate remained as the inorganic form. Efficiency ofethanol extraction Efficiency of radioactive phosphate extraction by ethanol from H. diminuta is presented in Table I. It was found to be consistently about 50% of the total amount of radiophosphate absorbed disregarding the length of the incubation period and the concentration of the external phosphate. Effect of different media on phosphorus absorption The absorptions of radiophosphate at various substrate concentrations tested in the BES and BSS were consistently lower than those in KRT (Table II). TABLE II Effect of different media on radiophosphate absorption in Hymenolepis diminuta Rates of uptake in different buffers (^moles/g ethanol- Radiophosphate extracted dry wt/2 min) concentration (mM) KRT* BES** BSS* 0.1 0.0071 0.0033 0.0045 0.3 0.0201 0.0118 0.0107 0.5 0.0331 0.0127 0.0222 1.0 0.0547 0.0179 0.0308 2.0 0.0718 0.0217 0.0660 * Krebs-Ringer tris-maleate buffer (Read el a/., 1963). ** Balanced electrolyte solution (Podesta et al, 1977b). *** Balanced saline solution (Ip and Fisher, 1982). PHOSPHATE TRANSPORT BY H. DIMINUTA 341 0 25 - 0 20 - 0 15 - 0 10 - 0 05 - FIGURE 2. Absorption of radiophosphate by Hymenolepis diminula as a function of phosphate con- centration (0.01-3 mA/). Each point is an average value from two experiments. S = phosphate concentra- tion in mA/, V = /^mol/g ethanol-extracted dry wt/2 min. • = Total uptake; O = mediated component; • = diffusive component. Effect of substrate concentration The absorption process was a straight line relationship over a substrate concentra- tion range of 0.03 mA/ to 20 mA/ suggesting diffusion might be involved. At lower substrate concentrations (0.01-3 mA/") uptake was nonlinear and appeared to be a combination of mediated uptake and diffusion (Fig. 2). Subtracting the diffusive com- ponent obtained by deleting sodium from the incubation medium from the overall rate, the mediated process was unveiled with calculated Vmax and Km values of 0.056 jumol ethanol-extractable radioactive inorganic phosphate/g ethanol-extracted dry wt/2 min and 1.19 mA/, respectively (Figs. 3, 4), obtained from the Lineweaver-Burk plot. Using Hanes-Woolf plot, the corresponding values of 0.136 yumol ethanol-ex- tractable radioactive inorganic phosphate/g ethanol-extracted dry wt/2 min and 3.64 mA/ were obtained. From studies with 3H-PEG, it was found that after rinsing and blotting dry three times, only 0.5 to 0.6 /j,\ of the incubation medium was carried over as unstirred layer per sample of worm (0.09-0.10 g ethanol-extracted dry weight). Since this was only a minute volume and it did not significantly affect the calculated results, the unstirred layer was not labelled in the subsequent incubation studies. With increasing concentrations of stable phosphate in the incubation medium, the initial rate of uptake of 0.5 mA/ labelled phosphate decreased and then levelled off (Fig. 5). This result supported the hypothesis that mediated transport was involved and fur- ther demonstrated a component of uptake that was not inhibited by increases in stable phosphate concentrations. 342 A. L. L. LIM AND Y. K. IP FIGURE 3. Lineweaver-Burk plot of radiophosphate absorption by Hymenolepis diminuta. V /umol/g ethanol-extracted dry wt/2 min; S = phosphate concentration in mA/. r2 = 0.96. Effect ofNa+ Worms were preincubated in various KRT media deficient in sodium, potassium, calcium, magnesium, chloride, nitrate, and sulfate ions. They were rinsed in three [S]/V 20 [S]mM FIGURE 4. Hanes-Woolf plot of radiophosphate absorption by Hymenolepis diminuta. V = ^mol/g ethanol-extracted dry wt/2 min; S = substrate concentration in mA/; r2 = 0.6 1 . PHOSPHATE TRANSPORT BY H. DIMINUTA 343 0 06 0 04 0 02 [l]mM 10 FIGURE 5. Effect of increasing stable phosphate concentration on uptake of 0.5 mA/ radiophosphate by Hymenolepis diminuta. Each point is an average value from two experiments. V = jumol/g ethanol- extracted dry wt/2 min; I = stable phosphate concentration in mM. changes of the respective ion deficient buffer, blotted dry, and then incubated in 0.1 mM radiophosphate in the absence of the respective ions. Absorption of labelled phosphate was affected by Na+ but not by the other ions tested. Therefore, worms were incubated in 0. 1 mA/ phosphate with various sodium ion concentrations (using cho- line as a replacement for Na+). Mediated phosphate uptake by H. diminuta was so- dium-dependent and a hyperbolic function of Na+ concentration of the ambient me- dium (Fig. 6). Effect of glucose and methionine The absorption of 0. 1 mA/ radiophosphate was not affected by the presence of either 1 mA/ glucose or 1 mM methionine in the incubation medium. The rates of absorption in the presence of glucose and methionine were 0.016 ± 0.002 (n = 5) and 0.018 ± 0.002 (n = 5) /umol ethanol-extractable Pi/g ethanol-extracted dry wt/2 min respectively, which were not significantly different from the control value of 0.019 ± 0.003 (n = 5, P> 0.01). Effect of organic phosphates Fructose-6-phosphate (F6P), glucose- 1 -phosphate (G 1 P), guanosine triphosphate (GTP), uridine diphosphate (UDP), NADP, and ATP (concentration 10 mM) were used to determine the specificity of the transport system. Absorption of 0. 1 mM radio- phosphate was not affected by NADP and ATP but was inhibited by F6P, G 1 P, GTP, and UDP. However, when the worms were incubated in the latter organic phosphate esters together with ATP, no signs of inhibition were observed (Table III). Effect of metabolic poisons and molybdate Sodium floride, potassium cyanide, parachloromercuribenzoic acid, iodoacetate, iodoacetamide, ouabain, dinitrophenol, and phlorizin were examined to determine 344 A. L. L. LIM AND Y. K. IP 13 11 FIGURE 6. Effect of Na+ concentration on mediated absorption of 0. 1 mM radiophosphate by Hy- menolepis diminuta. Each point is an average value from two experiments. V = ^mol/g ethanol-extracted dry wt/2 min. their effects on 0.1 mM labelled phosphate absorption. None of the above poisons had any effect on the uptake of phosphate by H. diminuta when they were added to both the preincubation and incubation media. Molybdate, however, inhibited the mediated uptake of radiophosphate by 72% when compared to the control value. Effect ofpH and temperature To study the effect of pH on the uptake of phosphate by H. diminuta, the worms were preincubated in KRT buffer at pH 7.4 and incubated for two minutes in 0.1 TABLE III Effect of various organic phosphates on the uptake ofO. 1 mM radiophosphates by Hymenolepis diminuta Inhibitor(s) (10mA/) % Inhibition of mediated process F6P GIF GTP UDP NADP ATP F6P + ATP G1P + ATP GTP + ATP UDP + ATP 30 28 23 41 0 0 0 0 0 0 PHOSPHATE TRANSPORT BY H. DIM1NUTA 345 0 020 - 0 016 - 0 012 - V 0 008 - 0-004 - FIGURE 7. Effect of pH on the rate of 0.1 mM radiophosphate absorption by H. diminuta. Each point is an average value from two experiments. V = ^mol/g ethanol-extracted dry wt/2 min. mA/ radiophosphate at various pHs (5.2 to 8.5) (Fig. 7). The velocity of the absorptive process increased rapidly and peaked at pH 6. The effect of temperature on phosphate uptake was determined by incubating the worms in 0. 1 mM radiophosphate in the presence or absence of external sodium at various temperatures. Incubation in Na+ yielded results for both the diffusive and mediated processes (Fig. 8). However, incu- bation in Na+ deficient KRT buffer abolished the mediated process, producing results only for the linear component. Hence, the difference between the two velocities of the uptake expressed results of temperature effect on the mediated process. Absorption of phosphate was optimal at 37°C (Fig. 9). DISCUSSION The results of the short interval incubation experiments showed that 10-day-old Hymenolepis diminuta could absorb inorganic phosphate (Pi) from its surroundings. Ethanol could consistently extract only 50% of the radioactive phosphate absorbed (Table I) disregarding the incubation conditions. Even for an incubation period as short as 30 s, a similar large proportion of the absorbed phosphate was incorporated into the calcareous corpuscles and nucleic acids of the parasite (Ip, unpub. results). Since the time course study indicated that the free pool of inorganic phosphate within the worms, which was ethanol-extractable, did not attain a steady state relationship with the other internal phosphate pools and the external radioactive phosphate within four minutes of incubation indicating condition of initial uptake velocity (Table I, Fig. 1 ), the authors expressed the absorption rate in terms of micromoles of ethanol extractable radioactive phosphate per gram ethanol extracted dry weight per two min- utes. More than 95% of the ethanol-extractable phosphate remained unmetabolized in the worm tissues. 346 A. L. L. LIM AND Y. K. IP 0-16 - 0 08 30 Temp(°C) 60 FIGURE 8. Effect of temperature on rate of total absorption and Na-independent uptake of 0. 1 mM radiophosphate by Hymenolepis diminuta. Each point is an average value from two experiments. V = jumol/gethanol-extracted dry wt/2 min. • = Total uptake; A = Na-independent uptake. Krebs-Ringer tris-maleate (KRT) buffer was used in all incubation studies be- cause of its ubiquity in other membrane transport studies of//, diminuta. It has been proposed that maleate in KRT may affect transport processes (Roth et ai, 1976). However, there are no clear indications that maleate specifically inhibits any trans- 0 006 v 0 004 0 002 30 40 50 Temp(°C) 60 FIGURE 9. Effect of temperature on rate of mediated uptake of 0. 1 mM of radiophosphate by Hymen- olepis diminuta. V = mediated uptake of radiophosphate in ^mol/g ethanol-extracted dry wt/2 min. PHOSPHATE TRANSPORT BY H. DIMINUTA 347 port process in H. diminuta, although Podesta et al. (1977b) reported an apparent reduction in the influx of glucose using KRT as the incubation media. Furthermore, the uptake rates of phosphate in KRT were considerably higher than those of bal- anced electrolyte solution (Podesta et al., 1977b) and balanced saline solution (Ip and Fisher, 1982) (Table I). The adsorption of radioactive substance to the 'unstirred layer' of the brush border membrane of//, diminuta was minimum after rinsing and blot-drying, and did not impose serious errors on analyses of results for phosphate absorption. There was argument that the 'unstirred layer' would contribute errors to transport determination not only due to the adsorption phenomenon but also to the requirement of substance to diffuse across such barriers before becoming available to the transport mechanism (Podesta, 1977). Since the 'unstirred layer' is naturally pres- ent in the gut of the host in vivo, it was not the authors' intention to report Vmax and K,,, values of the mediated phosphate transport process under an artificial situation of no 'unstirred layer' effect, but to demonstrate a distinct phosphate transport mech- anism different from those involved in other transport phenomena in H. diminuta under standard in vitro incubation conditions. Phosphate uptake in H. diminuta consisted of two components: A saturable, Na+ dependent component which obeys Michaelis-Menton kinetics and a Na+ indepen- dent process which shows a linear relationship with increasing extracellular Pi con- centration (Fig. 2). As suggested by Atkins and Gardner (1977), direct fit of the two term function v = (Vmax{S}/Km + {S}) + k{S} (where v = micromoles ethanol- extractable radioactive Pi/g ethanol/extracted dry wt/2 min, Vmax = maximum veloc- ity, K,M = Michaelis-Menten constant, and k = the rate constant of the linear compo- nent) to the data obtained from concentration studies was performed. Subtraction of the rate of uptake of the Na+ independent component (diffusion) from the overall rate yielded a corrected rate for the saturable process with Vmax and Kw values of 0.056 Atmol ethanol-extractable radioactive Pi/g ethanol-extracted dry wt/2 min and 1.19 mM, respectively, as obtained from the Lineweaver-Burk plot and correspond- ing values of 0. 1 36 jumol ethanol-extractable radioactive Pi/g ethanol extracted dry wt/2 min and 3.64 mA/as determined by the Hanes-Woolf plot. Such a discrepancy can be due to the inaccuracy of the Lineweaver-Burk determination of kinetic con- stants as there are relatively few points at the high end of the 1/(S) scale, and it is these points that are most heavily weighted in such determinations. Moreover, small errors in the determination of Vmax are magnified when reciprocals are taken. The results of the present investigation contradict previous suggestion that absorption of Pi was due to diffusion (von Brand, 1973). Although diffusion occurred at all substrate con- centrations tested, mediated transport predominated at substrate concentrations be- low 1 .5 mA/ (Fig. 2). Mediated phosphate absorption was affected by the concentra- tion of Na+ in the ambient medium. This finding is consistent with that reported on the transport of inorganic phosphate transport by SV3T3 cultured cells (Brown and Lamb, 1975). The cations and anions K+, Ca++, Mg++, Cl~, SO4~~, and NO3" had no effect on transport activity. The sodium requirement, as demonstrated in numerous other systems (Read et al., 1974; Schultz et al., 1974; Crane, 1977), suggests that sodium and phosphate may be cotransported by a catalytic protein 'carrier' system in H. diminuta similar to that proposed by Crane (1965) for glucose transport in mamma- lian intestine. Metabolic poisons, in particular ouabain, had no inhibitory effect on the rate of phosphate absorption in H. diminuta. Such a Na+-coupled active transport system usually requires an ouabain-sensitive Na+-K+ activated ATPase (Schultz et al., 1974; Crane, 1977). The observation that ouabain has no effect on glucose trans- port in H. diminuta (Lee et al., 1963; Dike and Read, 197 Ib) raises doubt to the 348 A. L. L. LIM AND Y. K. IP presence of such an enzyme in this parasite and its involvement in phosphate trans- port in the present study. However, H. diminuta is impermeable to ouabain (Gal- logly, 1972) and the ouabain-sensitive Na+-K+ activated ATPase may not be local- ized in the surface membrane(s) of the brush border. Recently, the Na+ extrusion mechanism was found to be localized below the tegument (Podesta et al, 1977a) and this hypothesis is sustained by Lumsden and Murphy (1980). As such, the absence of any effect of ouabain on phosphate uptake in H. diminuta is expected. The mediated transport of phosphate was inhibited by molybdate, possibly the result of structural damage of the surface catalytic protein carrier molecules. Arme and Read (1970) and Dike and Read (197 la) showed that phosphatase activity on the surface membrane of//, diminuta was totally inhibited by ATP. The absence of inhibitory effect when ATP was added to the incubation media in the presence of either F6P, G1P, GTP, or UDP indicated that these organic phosphates had no direct effect on inorganic phosphate absorption. The apparent inhibitory effect of these organic phosphates was a result of increasing inorganic phosphate con- centration in the external media after hydrolysis of the respective phosphate esters by the surface enzyme. Since phosphate transport was not affected by a variety of organic phosphate esters, glucose, and methionine, it may be concluded that phosphate trans- port in H. diminuta occurs through a different mechanism with high substrate speci- ficity. The absorption of phosphate by the worm was sensitive to temperature and pH of the ambient medium. Optimal absorption occurred at 37°C. At lower tempera- tures, the surface enzymes required for mediated phosphate uptake are probably inac- tivated and at higher temperatures, the decreased rate is likely due to damage of the cestode integument. It is significant that maximum absorption occurred at pH 6 at which the major inorganic phosphate exists as H2PO4" in solution. Hence, diffusion can play a major part in phosphate uptake since this phosphate is less negatively charged as compared to the other possible phosphate ions. ACKNOWLEDGMENTS This research was supported in part by grants RP 59/83 and RP 145/82 from the National University of Singapore. LITERATURE CITED ATKINS, G. L., AND L. G. GARDNER. 1977. The computation of saturable and linear components of intestinal and other transport kinetics. Biochim. Biophys. Ada 468: 127-145. ARME, C, AND C. P. READ. 1 970. A surface enzyme in H. diminuta (Cestoda). J. Parasitol. 56: 514-516. VON BRAND, T. 1973. Biochemistry of Parasites. Academic Press, New York. P. 1 3. VON BRAND, T., M. U. NYLEN, G. N. MARTIN, F. K. CHURCHWELL, AND E. STITES. 1969. Cestode calcareous corpuscles: Phosphate relationships, crystallization patterns and variations in size and shape. Exp. Parasitol. 25: 291-310. BROWN, K. D., AND J. F. LAMB. 1975. Na-dependent phosphate transport in cultured cells. J. Phvsiol. 251:58-59. CRANE, R. K. 1 965. Na+-dependent transport in the intestines and the other animal tissues. Fed. Proc. 24: 1000-1006. CRANE, R. K. 1977. The gradient hypothesis and other models of carrier mediated active transport. Rev. Physiol. Biochem. Pharmacol. 78: 99-159. DAUGHERTY, J. W. 1957. Intermediary protein metabolism in helminths. IV. The active absorption of methionine by the cestode, H. diminuta. Exp. Parasitol. 6: 60-67. DIKE, S. C., AND C. P. READ. 197 la. Tegumentary phosphohydrolases of H. diminuta. J. Parasitol. 57: 81-87. PHOSPHATE TRANSPORT BY H. DIMINUTA 349 DIKE, S. C, AND C. P. READ. 197 Ib. Relation of tegumentary phosphohydrolases and sugar transport in H. dimimila. J. Parasitol. 57: 1251-1255. GALLOGLY, R. L. 1972. ATPase and membrane transport in the rat tapeworm H. diminuta. Ph.D. Thesis, Rice University, Houston, TX. IP, Y. K., AND F. M. FISHER JR. 1 982. Membrane transport of inositol by Hvmenolepis diminuta (Cestode). J. Parasitol. 68: 53-60. LEE, D. L., A. H. ROTHMAN, AND J. B. SENTURIA. 1963. Esterases in Hvmenolepis and Hvdatigera. Exp. Parasitol. 14: 285-295. LESUK, A., AND R. J. ANDERSON. 1940. Concerning the chemical composition of Cysticercus fasciolaris. II. The occurrence of cerebroside containing dihydrosphingosine and hydrolecithin in cysticercus larvae. J. Biol. Chem. 139: 457-469. LUMSDEN, R. D., AND W. A. MURPHY. 1980. Morphological and functional aspects of the cestode surface. In Cellular Interactions in Symbioses and Parasitism, C. B. Cook, P. W. Pappas, and E. D. Ru- dolph, eds. The Ohio State University Press, Columbus, OH. P. 321. PODESTA, R. B. 1977. Hvmenolepis diminuta: Unstirred layer thickness and effects on active and passive transport kinetics. Exp. Parasitol. 43: 12-24. PODESTA, R. B., W. S. EVANS, AND H. E. STALLARD. 1977a. Hvmenolepis diminuta and Hvmenolepis microstorna: Effect of ouabain on active non-electrolyte uptake across the "epithelial" syncytium. Exp. Parasitol. 43: 25-38. PODESTA, R. B., H. E. STALLARD, W. S. EVANS, P. E. LUSSIER, D. J. JACKSON, AND D. F. METTRICK. 1977b. Hymenolepis diminuta. Determination of unidirectional uptake rates for nonelectrolytes across the surface "epithelial" membrane. Exp. Parasitol. 42: 300-317. READ, C. P. 1950. Acquisition of isotopically labelled inorganic phosphate by the tapeworm, Hymenolepis diminuta, with some remarks on the host-parasite relationship. J. Parasitol. 36: 34-40. READ, C. P., A. H. ROTHMAN, AND J. E. SIMMONS, JR. 1963. Studies on membrane transport, with special reference to parasite-host integration. Ann. N. Y. Acad. Sci. 113: 154-205. READ, C. P., G. L. STEWART, AND P. W. PAPPAS. 1 974. Glucose and sodium fluxes across the brush border of//, diminuta (Cestoda). Biol. Bull. 147: 146-162. ROTH, K. S., S. M. HWANG, AND S. SEGAL. 1976. Effect of maleic acid on the kinetics of methyl-D- glucoside uptake by isolated rat renal tubules. Biochim. Biophys. Acta 413: 3 1 7-32 1 . SCHEIBEL, L. W., H. J. SAZ, AND E. BuEDiNG. 1968. The anaerobic incorporation of 32P into adenosine triphosphate by Hymenolepis diminuta. J. Biol. Chem. 243: 2229-2235. SCHULTZ, S. G., R. A. FRIZZEIT, and N. NELLANS III. 1974. Ion transport by mammalian small intestine. Ann. Rev. Phvsiol. 36: 51-91. Reference: Biol. Bull. 172: 350-356. (June, 1987) INDUCTION OF PRECOCIOUS MOLTING AND CLAW TRANSFORMATION IN ALPHEID SHRIMPS BY EXOGENOUS 20-HYDROXYECDYSONE DEFOREST MELLON, JR. AND ELIZABETH GREER Department of Biology, University of Virginia, Charlottesville, Virginia 22901 ABSTRACT Exogenous micromolar concentrations of 20-hydroxyecdysone were used to ac- celerate the molt cycle in the snapping shrimp, Alpheus heterochelis. Shrimps were exposed to the hormone by adding it to the seawater in which they were cultured. Five days of continuous exposure of an animal to a hormone concentration of 5 Mgm/ ml shortened the winter molt cycle length by 18 days, or 65%. Claw transformation, a commonly observed phenomenon in alpheid shrimps, also was accelerated at these concentrations, and postmolt animals exhibited appropriate modifications in exter- nal claw morphology and, in some cases, restructuring of the claw closer muscle ac- cording to schedules identical with those of animals having normal molt cycle lengths. These studies strengthen the hypothesis that the cellular events which under- lie claw transformation are modulated by endogenous ecdysteroids. INTRODUCTION The insect growth hormone, 20-hydroxyecdysone, promotes molting in a variety of crustaceans, whether directly injected into subject animals (Krishnakumaran and Schneiderman, 1969, 1970; Freeman and Bartell, 1976; Rao et al, 1973; Tighe-Ford and Vaile, 1972) or added to the medium of cultured larvae (Cheung, 1974). Nor- mally during the crustacean molting cycle, alpha ecdysone is thought to be synthe- sized and released from the Y-organ, and then converted to 20-hydroxyecdysone in target tissues (Skinner, 1985). The widespread and dramatic effects of these ste- roids has promoted the acceptance of ecdysones as generalized arthropod growth hormones. We were interested in determining whether exogenous 20-hydroxyecdysone would be effective in shortening the molt cycle and in accelerating claw transforma- tion in the snapping shrimp, Alpheus heterochelis. These intertidal crustaceans have extremely asymmetric claws — a large snapper and a small pincer — that take part in a remarkable morphological transformation: If the large snapper claw is caused to undergo autotomy, the pincer claw transforms into a snapper, while a new pincer regenerates at the original snapper site (Przibram, 1901; Wilson, 1903). Claw trans- formation, however, involves more than growth and change in external morphology. The two muscles in the propodite — the opener and the closer — enlarge considerably during transformation. Their motor neurons also become proportionally larger (Mel- lon et al. , 1981; Mellon and Smith, unpub. obs. ) and their physiological and biochem- ical properties are modified (Stephens and Mellon, 1979; Quigley and Mellon, 1984). Moreover, a dramatic instance of cellular death accompanies transformation of the claw. In the pincer the closer muscle is a composite of two different fiber types: There is a central band of large fast type fibers that is flanked on either side by smaller Received 22 December 1986; accepted 17 March 1987. 350 ECDYSONE INDUCED MOLTING IN SHRIMPS 351 diameter slower type fibers (Stephens and Mellon, 1979; O'Connor et al., 1982). As the muscle transforms, the fast fibers die, while the slow fibers are modified to become very slow fibers (Quigley and Mellon, 1986). Muscle fiber death is apparent within a week following the first molt after onset of transformation; thus, both the change in claw morphology and the transformation of muscle fiber appear to be tied to the molt cycle, and both may be under hormonal control. To test this possibility, we exposed snapper-less shrimps to exogenous 20-hydroxyecdysone. The results show that the molt cycle can be reduced by as much as 65% in response to external ecdysone. Mor- phological claw transformation also was apparent in all of our experimental animals, while degeneration of the fast closer muscle fibers group occurred in two of the three experimental groups. MATERIALS AND METHODS Experimental animals were recently molted, medium sized (4-5 cm) snapping shrimps (Alpheus heterochelis). Shrimps were housed individually in plastic contain- ers in a 4 foot by 5 foot tray. Seawater entered the tray at one end, passed through the shrimp cages, and exited at the other end of the tray. Within a day of molting, animals were removed from the culture system and isolated individually in plastic containers with 100 ml artificial seawater (Forty Fathoms, Marine Enterprises Inc., Towson, Maryland). The animals were exposed to exogenous hormone by adding to an individual animal's plastic container 100 microliters of a stock solution containing either 0.05 mg/ml or 0.5 mg/ml of 20-hydroxyecdysone (Sigma, E-2003). The stock solution was originally prepared by dissolving the hormone in 5% ethanol to the desired concen- tration. Three different experimental protocols were used in our experiments. In the first protocol, shrimps were taken on the day following a molt, caused to autotomize the snapper claw, and placed immediately in the hormone bath. Hormone exposure was limited to five days, following which the animals were put into plain artificial seawater. The second protocol involved removal of the snapper on the day following a molt and then waiting five days before exposure to hormone. In the third protocol, animals were left undisturbed for six days after molting at which time their snapper claws were removed. The animals were then placed in the hormone bath. Regardless of the protocol, the animals were fed on the tropical fish food Tetramin every other day after which the bath water was changed. Molt cycles of animals ex- posed to the hormone were compared with those of animals kept in our culture sys- tem and from which the snapper claw had been removed two days following a molt. These control animals were obtained throughout the year, which was arbitrarily di- vided into winter (October-March) and summer (April-September). Our experi- ments were performed during February and March or during October and Novem- ber. Data from the two groups were pooled. Room temperatures during the experi- mental periods varied between 22 and 24°C. Room lighting was diurnally supplemented through large glass windows. Seasonal variations in day length thus may have been sensed by the experimental animals. The molt cycle of crustaceans has been divided into stages by various workers (Skinner, 1985). These differ qualitatively even across different groups of decapods. While there may be a correlation with the various molt cycle stages and the level of ecydsone in the hemolymph, this has not been measured in any shrimp species. In the crayfish Orconectes, however, both in vivo and in vitro measurements of ecdysone production suggest that the hemolymph liter begins to rise during the early stages of 352 DEF. MELLON JR. AND E. GREEK premolt and reaches a peak in late premolt (Jegla et al, 1983). The blood hormone level then falls to low levels just prior to ecdysis. We measured the molt cycle lengths in 36 snapper-less summer shrimps, 28 snap- per-less winter shrimps, and in 43 ecdysone-treated snapper-less winter animals. Molt cycle length was taken as the number of days between two successive ecdyses. Immunohistochemical procedures for examining biochemical and structural changes in the closer muscles of transforming claws have been described previously (Quigley and Mellon, 1986). Briefly, transforming claws were removed from animals eight days following the first molt after snapper claw autotomy. The claws were im- mediately frozen in melting isopentane and sectioned on a cryostat at 30 micrometers thickness. Frozen sections were mounted on glass coverslips and either stored at -80°C, or they were immediately reacted with monoclonal antibodies developed in our laboratory against fast and slow shrimp muscle myosin heavy chains. Muscle degeneration was assessed by microscopic examination of the stained, antibody-re- acted sections. Fast fiber histolysis was presumed to have occurred when the fast- specific antibody reaction could no longer be detected in the central regions of the closer muscle. Usually, the absence of fast muscle fibers was also indicated by a tissue- free space in the central muscle. In any case, the reaction of adjacent frozen sections with antibody to slow fiber myosin assured that the tissues and reaction mixtures were normal. RESULTS Figure 1 illustrates various stages in the morphological transformation of pincer to snapper. Not all stages are observed in any one animal during transformation. The earliest obvious morphological clue that transformation of the claw is occurring is the appearance (after a molt) of a mesial ridge on the dactylus. This ridge represents the earliest expression of the plunger on the fully transformed dactyl. We used the occurrence of any of the stages in Figure 1 as an indication that morphological trans- formation of the claw was occurring. Our data, shown in Figure 2, compare molt cycle lengths in snapper-less summer and winter shrimps, and in ecdysone-treated snapper-less winter animals. On aver- age, winter (October-March) molt cycles in transforming shrimps lasted about 29 days, while transforming summer (April-September) animals molted about every 23 days. This finding is similar to other studies of seasonal molt cycle change (e.g., Free- man and Bartell, 1975). The underlying physiological reasons for the longer length of molt cycle in winter animals are not known; room temperatures in general were about two degrees higher in the summer, but other factors, such as day length, may play a significant role in determining growth and molting. It is clear that treatment with 20-hydroxyecdysone at the concentrations we em- ployed significantly shortened the usual winter molt cycle. Treatment in the experi- mental groups of animals was started either in February or early October, and experi- ments continued for two months. Normal, snapper-less animals from this time period would have had a molt cycle length of about 29 ± 3.8 days; those treated for five days with 20-hydroxyecdysone at the lower concentration we employed had a molt cycle of about 2 1 ± 4.8 days, while at the higher concentration the effect was more dramatic and the molt cycle length was only 10.5 ± 1.5 days. Three different experimental protocols were used in exposing the experimental animals to the higher concentration of exogenous hormone. Figure 3 is a diagram illustrating these application schedules and the consequent effects upon a "normal" molt cycle length of 29 days. Shrimps undergoing protocol #1 molted, on average, 10.5 days after their previous molt, thus shortening their normal intermolt interval by 65%. ECDYSONE INDUCED MOLTING IN SHRIMPS 353 PINCER SNAPPER FIGURE 1 . Morphological stages in the transformation of a pincer claw to a snapper claw. As a rule no single animal successively exhibits all of the stages illustrated, and in some instances, the stage exhibited at the first molt during transformation is considerably advanced. The caret indicates the earliest stage observed in the formation of the plunger found on the dactyl of the fully formed snapper. Shrimps in the other two protocols molted roughly 14 days following their previ- ous ecdysis, shortening their intermolt interval by 52%. Thus, the maximum reduc- tion in molt cycle length was obtained by beginning hormone treatment immediately following the previous ecdysis. It was of interest whether hormone-induced precocious molting and the accompa- nying claw transformation would elicit the modifications in claw closer muscle struc- ture that normally occur in snapper-less shrimps. Therefore the claws of all surviving, transforming experimental animals were subjected to immunohistochemical analy- sis, as described previously (Quigley and Mellon, 1986). Table I indicates the incidence of muscle degeneration in each of the three experi- mental groups. While two of the groups exhibited a preponderance of muscle remod- eling, the group in which hormone treatment was initiated directly after a molt gave no evidence of closer muscle degeneration even though, in every case, the external morphology of the claw was clearly changing. DISCUSSION A previous study (Freeman and Bartell, 1976) demonstrated that 20-hydroxyec- dysone will shorten the molt cycle in another caridean shrimp, Palaemonetes pugio. Some differences are apparent between the results of that study and of our own exper- 354 DeF. MELLON JR. AND E. GREER 28 32 fm 28 - 3 6 17 cn Q r - i- (D2O . Z UJ CO CO I .T CYCLE S ffl - CONTROL R CONTROl 17 9 I I _J O 8 - WINTER SUMME O) in o T™ . CO 4 CM 9 & O o _i O 4 O p 8 u p - 0 Q o (£ ££ o: n Q_ o. FIGURE 2. Bar graphs illustrating mean molt cycle duration in several groups of shrimps: Winter controls; summer controls; experimental animals in protocol 1 exposed to either 0.5 micrograms per milli- liter or 5 micrograms per milliliter hormone concentration, and the pooled data from experimental animals in protocols 2 and 3, both of which were exposed to hormone concentrations of 5 micrograms per milliliter. Brackets at the top of each bar are ±one standard deviation, and numbers of animals in each group are shown above the brackets. iments. In the study by Freeman and Bartell (1976), 20-hydroxyecdysone was in- jected into the experimental animals. Four different dosages were tried, and the lowest effective dosage (0.5 /ugm/animal) shortened the intermolt cycle by 67%. Only 19% of this group survived the succeeding molt, however, whereas the overall survival rate in our three experimental groups was 72%. If the two shrimp species are physiologi- cally comparable, it appears that simply bathing in the hormone is as effective a treat- ment as injection and is less traumatic to the experimental animals. It is clear from our studies that exogenous 20-hydroxyecdysone accelerates not only the molt cycle in snapper-less Alpheus, but also the morphogenetic events that bring about transformation of the claw from a pincer to a snapper. This includes, in some cases, the programmed death of the fast group of muscle fibers in the claw closer. Fast fiber death was not seen in any of the experimental animals in protocol #1, in which hormone treatment was started one day following the preceding molt. The underlying physiological reasons for this failure are not known. In many decapod crustaceans the concentration of ecdysteroids in the hemolymph begins to rise in the premolt phase of the molting cycle, reaches a peak in mid premolt, and falls precipitously in late premolt, just prior to ecdysis (Jegla et al, 1983; Skinner, 1985). ECDYSONE INDUCED MOLTING IN SHRIMPS 355 \ nwnmvynit \ PROTOCOL 2 HORMONE PROTOCOL 3 HORMONE WINTER CONTROLS FIGURE 3. Relative molt cycle durations in winter control animals and in the three experimental protocols. Periods of hormone exposure are indicated, and arrows delineate the time point at which snapper claws were removed in each group. If the hemolymph liters of ecdysone in Alpheus approximate this schedule, then it is apparent that the experimental animals in the first protocol may have been subjected to two distinct pulses of ecdysone, as opposed to the single, rather prolonged pulse which probably occurred in protocols #2 and #3. Since onset as well as reduction in blood hormone levels may act as a trophic signal (e.g., Morton and Truman, 1985), the two pulses could have unpredictable consequences. Until the relative blood con- centrations of ecdysone are precisely known throughout the intermolt period in Al- pheus, however, further speculation cannot be justified. It is sufficient to conclude that, with an appropriate waiting period following molting, 5 days of treatment with exogenous ecdysone will shorten the mean normal molt cycle by more than 1 8 days, or by about 65%, and will proportionally accelerate the normal processes of claw transformation, including remodeling of the closer muscle. The procedure thus may prove to be useful as a tool in future experimental manipulations. The mechanism of uptake (if any) of ecdysone from the bathing medium by the shrimps is not known. The gills of these intertidal shrimps are probably permeable to water and some inorganic ions, but it does not necessarily follow that molecules the size of steroids can pass into the blood through this route. Possibly, the hormone is ingested when the animals feed. TABLE I Effects of experimental hormone treatments on closer muscle restructuring in transforming claws Total number Number of Number of closer Number of of animals successful molts* muscles examined restructured muscles Protocol 1 30 27 10 0 Protocol 2 6 6 3 2 Protocol 3 14 5 3 2 * Attrition was due either to death of the animal or to autotomized pincer claws. 356 DEF. MELLON JR. AND E. GREEK Other possibilities must also be considered. We have observed that snapping shrimps held in communal culture systems, while physically separated from one an- other, tend to molt concurrently, suggesting the ecdysone or some other pheromone excreted into the culture water may be synchronizing their individual cycles. Such synchronization may be important to these gregarious, but cannibalistic crustaceans. In any event, the concentration of pheromone in the water must be very small. If this is the case, the shrimps may possess specialized external chemoreceptors that respond specifically to ecdysone, and through which molting could be triggered. ACKNOWLEDGMENTS We are grateful to Dr. W. Otto Friesen for a gift of 20-hydroxyecdysone, and to W. D. Nelms Creekmur for technical assistance. The illustrations in Figure 1 were executed by Ms. Donna Bennett. This research was partially supported by an in- house grant from the University of Virginia. LITERATURE CITED CHEUNG, P. J. 1974. The effect of ecdysterone on cyprids of Balanus eburnem Gould. J. Exp. Mar. Biol. Ecol. 15:223-229. FREEMAN, J. A., ANDC. K. BARTELL. 1975. Characterization of the molt cycle and its hormonal control in Palaemonetes pugio (Decapoda., Caridea). Gen. Comp. Endocrinol. 25: 517-528. FREEMAN, J. A., ANDC. K. BARTELL. 1976. Some effects of the molt-inhibiting hormone and 20-hydroxy- ecdysone upon molting in the grass shrimp, Palaemonetes pugio. Gen. Comp. Endocrinol. 28: 131-142. JEGLA, T. C., C. RULAND, G. KEGEL, AND R. KELLER. 1983. The role of the Y-organ and the cephalic gland in ecdysteroid production and the control of molting in the crayfish, Orconectes limosus. J. Comp. Physiol. 152: 91-95. KRISHNAKUMARAN, A., AND H. A. SCHNEIDERMAN. 1969. Induction of molting in Crustacea by an insect hormone. Gen. Comp. Endocrinol. 12: 515-518. KRISHNAKUMARAN, A., ANDH. A. SCHNEIDERMAN. 1970. Control of molting in mandibulate and chelic- erate arthropods by ecdysones. Biol. Bull. 139: 520-538. MELLON, DEF., J. A. WILSON, ANDC. E. PHILLIPS. 1981. Modification of motor neuron size and position in the central nervous system of adult snapping shrimps. Brain Res. 223: 1 34-140. MORTON, D. B., AND J. W. TRUMAN. 1985. Steroid regulation of the peptide-mediated increase in cyclic GMP in the nervous system of the hawkmoth, Manduca sexta. J. Comp. Physio/. 157: 423-432. O'CONNOR, K., P. J. STEPHENS, AND J. M. LEFEROVICH. 1982. Regional distribution of muscle fiber types in the asymmetric claws of Californian snapping shrimp. Biol. Bull. 163: 329-336. PRZIBRAM, H. 1901. Experimented Studien uber Regeneration. Arch. Entwicklungsmech. 11: 321-345. QUIGLEY, M. M., AND DEF. MELLON. 1984. Changes in myofibrillar gene expression during fiber-type transformation in the claw closer muscles of the snapping shrimp, Alpheus heterochelis. Dev. Biol. 106: 262-265. QUIGLEY, M. M., AND DEF. MELLON. 1 986. Myofiber death plays a role in determining fiber type composi- tion in the claw closer muscles of the snapping shrimp, Alpheus heterochelis. J. Exp. Zool. 239: 299-305. RAO, K. R., S. W. FINGERMAN, AND M. FINGERMAN. 1973. Effects of exogenous ecdysones on the molt cycles of fourth and fifth stage American lobsters, Homarus americanus. Comp. Biochern. Phys- iol. 44 A: 1105-1120. SKINNER, D. M. 1985. Molting and Regeneration. Pp. 43-146 in The Biology of Crustacea, Vol. 9, D. E. Bliss and L. H. Mantel, eds. Academic Press, New York. STEPHENS, P. J., AND DEF. MELLON. 1979. Modification of structure and synaptic physiology in trans- formed shrimp muscle. J. Comp. Physiol. 132: 97-108. TIGHE-FORD, D. J., AND D. C. VAILS. 1972. The action of crustecdysone on the cirripede, Balanus bala- noides (L). /. Exp. Mar. Biol. Ecol. 9: 19-28. WILSON, E. B. 1903. Notes on the reversal of asymmetry in the regeneration of chelae in Alpheus hetero- chelis. Biol. Bull. 4: 197-210. Reference: Biol. Bull. 172: 357-361. (June, 1987) A TRYPSIN INHIBITOR FROM THE COELOMIC FLUID OF THE SEA STAR ASTERIAS FORBESI JAMES A. MARCUM The Marine Biological Laboratory, Woods Hole, Massachusetts 02543; Department of Pathology, Beth Israel Hospital, Harvard Medical School, Boston, Massachusetts 02115; and Department of Biology, Massachusetts Institute of Technology. Cambridge, Massachusetts 02139* ABSTRACT Cell-free coelomic fluid obtained from the sea starAsteriasforbesi rapidly neutral- ized, in a dose-dependent fashion, trypsin derived from several different animal spe- cies. Approximately 50 to 300 ^g of bovine trypsin was inhibited per mg of total sea star coelomic fluid protein. Crude preparations of sea star and sea urchin trypsins isolated from gastric tissue, as well as purified human plasmin, were also neutralized upon addition to echinoderm coelomic fluid. In contrast, bovine a ! -chymotrypsin and thrombin were not inactivated by the coelomic fluid inhibitor. The trypsin inhib- itor was only mildly sensitive to heat or acid treatments and exhibited a molecular weight of ~6500. The protease inactivator was also detected in cell lysates derived from washed coelomocytes collected in N-ethyl maleimide, as well as in cell-free coe- lomic fluid collected by non-surgical means. INTRODUCTION Trypsin represents a constitutive protease of most organisms, including bacteria (Winter and Neurath, 1 970; Okabe and Noma, 1 974; Russo and YadofT, 1 978; Gold- berg et al., 1980; Zwilling and Neurath, 1980). Trypsin inhibitors also have been detected within a wide variety of organisms, including invertebrates (Greene et al., 1976; Tschesche, 1976; Tschesche and Dietl, 1976; Holzman and Russo, 1978). The invertebrate trypsin inhibitors inactivate sundry enzymes, including chymotrypsin. In the present communication, the detection and partial characterization of a sea star coelomic fluid inhibitor, which inactivates trypsin but not chymotrypsin, is de- scribed. MATERIALS AND METHODS All chemicals and reagents were purchased from either Sigma Chemical Co. (St. Louis, Missouri) or Fisher Scientific (Pittsburgh, Pennsylvania) and were reagent quality or better. Synthetic substrates were obtained from Helena Laboratories (Beaumont, Texas). Bovine thrombin was purchased from Parke-Davis (Morris Plains, New Jersey). Sea stars (Asterias forbesi) weighing ~150 gm were obtained daily from the Supply Department of the Marine Biological Laboratory (Woods Hole, Massachusetts) and used only once. Coelomic fluid was obtained by surgically amputating the tips of sea stars' arms and draining the fluid into chilled, pyrogen-free glass tubes. After centrifugation at 100 X g for 5 min at 4°C, the supernate was decanted and represented cell-free coe- Received 23 December 1986; accepted 4 March 1987. ' Present address. 357 358 J. A. MARCUM 100 10 20 30 40 COELOMIC FLUID (/Jl) FIGURE 1. Inhibition of bovine trypsin by sea star cell-free coelomic fluid. Increasing amounts of coelomic fluid were pre-incubated with 0. 1 ^g of bovine trypsin for 10 min at 22°C (410 n\, final volume). Residual proteolytic (- • -) activity was quantitated with the chromogenic substrate S2222 as described in Materials and Methods. Similar studies were conducted with coelomic fluid after heating to 100°C for 5 min (- D -) and after acidification to pH 2.5 for 1 h at 24°C and neutralization (- A -). lomic fluid. The pellet, consisting of coelomocytes, was suspended in pyrogen-free, distilled water (one ml per animal). The cellular pellet was disrupted by frequent and vigorous shaking for 2 h at 22°C. Cellular debris was removed by centrifugation at 1000 X g for 10 min at 22°C. The supernate was decanted and represented the cell lysate. Phase contrast microscopy revealed that >90% of the cells were lysed under these conditions. To measure the inhibitor activity of sea star cell-free coelomic fluid or cell lysate, residual enzymatic activity was quantitated employing chromogenic substrates. The assay was initiated by mixing coelomic fluid or cell lysate (5-400 /ul) with 10 ^1 of enzyme for 10 min at 22°C (410 ^1 total volume). Controls were constructed by sub- stituting the environmental buffer, 0.45 MNaCl in 50 mM Tris-HCl, pH 7.5, for the test sample. Chromogenic substrate (100 iA) was added to the above solution at a final concentration of 0.5 mM, and the reaction mixture was incubated for 5 min at 22°C and quenched with 200 n\ of glacial acetic acid. Amidolysis was quantitated by read- ing the absorbance at 405 nm. The fraction of inactivated enzymatic activity was calculated by dividing residual amidolytic activity in either coelomic fluid or cell lysate by total enzymatic activity determined in the absence of these samples. Protein concentrations were quantitated by the procedure of Bradford (1976) us- ing bovine serum albumin as standard (Sigma, fraction V). RESULTS AND DISCUSSION Addition of bovine trypsin (0. 1 ng) to cell-free coelomic fluid resulted in rapid neutralization of the enzyme in a dose-dependent fashion (Fig. 1). As little as 5 n\ of the sea star coelomic fluid inhibited ~40% of the added protease. The specific activity of the inhibitor under these conditions ranged from 50 to 300 ^g of trypsin inhibited per mg of total coelomic fluid protein (8 animals were tested). As shown in Figure 1, SEA STAR TRYPSIN INHIBITOR 359 O H CD X IOO- 80- 60- 4O- 20- o O CD o CT> > CM I I — r— 25 15 20 FRACTION NO T" 30 35 FIGURE 2. Gel permeation of sea star cell-free coelomic fluid. Coelomic fluid (200 n\) was gel filtered on Sephadex G-50 (0.6 cm X 1 15 cm) equilibrated in 0.45 MNaCl in 50 mMTris-HCl, pH 7.5. Fractions of 0.9 ml were collected at a column flow rate of 9 ml/h. Inhibition of bovine trypsin by eluate fractions was determined by measuring residual enzymatic activity with S2222 as outlined above. Molecular weight standards included carbonic anhydrase (29,000), lyzozyme ( 1 4,300), and aprotinin (6500). the inhibitor is only mildly sensitive to heat and acid treatments. A molecular weight of about 6500 was estimated by gel permeation chromatography using Sephadex G- 50 (Fig. 2). The above molecular characteristics of the coelomic fluid inhibitor are similar to those described for other small molecular weight inhibitors (Fritz and Krejci, 1976; Tschesche, 1976; Tschesche and Dietl, 1976). Crude preparations of proteolytic enzymes were obtained from the gastric tissues of A. forbesi and the sea urchin Strongylocentrotus droebachiensis. Briefly, the gut was excised surgically and disrupted with a glass pestle and homogenizer (Radnoti Glass Technology, Inc., Monrovia, California 91016). Cellular debris was removed by centrifugation at 1000 X g for 20 min at 4°C, and the supernate containing the gut enzyme(s) was decanted. These gut proteases represented trypsin-like enzymes based upon the following criteria: ( 1 ) Enzymatic activity was inhibited completely by bo- vine pancreatic trypsin inhibitor and soybean trypsin inhibitor, and (2) the proteases hydrolyzed synthetic peptide substrates containing carboxy-terminal arginyl residues (data not shown). Admixture of protease from either sea star or sea urchin with sea star coelomic fluid resulted in rapid inhibition of the respective enzymatic activity (Table I). Inhibition of these trypsin-like enzymes was dose-dependent, with ~40% of the enzymatic activity being inactivated by 5 /A of coelomic fluid (data not shown). Plasminogen was purified from human plasma (Deutsch and Mertz, 1970) and acti- vated with streptokinase (Marcum et al, 1982). Plasmin was rapidly neutralized upon addition to sea star coelomic fluid (Table I). In contrast, bovine a i -chymotrypsin was not neutralized by the sea star coelomic fluid inhibitor (Table I). As little as 10 ng of the protease was not inactivated upon addition to sea star coelomic fluid (200 ^1). The above data indicates that the specific- ity of the protease inhibitor is distinct from the Kunitz inhibitor which neutralizes trypsin and chymotrypsin (Kunitz and Northrop, 1936), but similar to the Kazal inhibitor which inactivates trypsin but not chymotrypsin (Kazal et al, 1948). The majority of invertebrate inhibitors described to date are Kunitz-type inhibitors, whereas only the bdellins isolated from leeches are Kazal-type inhibitors (Fritz and 360 J. A. MARCUM TABLE I Inactivation of enzymes by coelomic fluid inhibitor Enzyme1 % Inhibition Bovine trypsin 100 Sea star trypsin 1 00 Sea urchin trypsin 100 Human plasmin 100 Bovine archymotrypsin 0 Bovine thrombin 0 ' Residual proteolytic activity was quantitated utilizing chromogenic substrate S2222 for measuring 0.1 ng bovine pancreatic trypsin (12,400 BAEE units/mg), S2238 for 0.1 unit bovine thrombin (56 NIH units/mg), S2586 for 0.01 to 1 ng bovine pancreatic chymotrypsin (50 BTEE units/mg), S2444 for 5.6 ^g sea star gut-extracted protease(s) (197 BAEE units/mg), and S2238 for 3.3 p.g sea urchin gut-extracted protease(s) (149 BAEE units/mg). In the case of human plasmin ( 1 ng), enzymatic activity was measured employing I25l-labeled a-casein as substrate (Highsmith and Rosenberg, 1977). Krejci, 1976). It must be noted, however, that Kazal inhibitors inactivate thrombin, whereas the sea star inhibitor did not neutralize a crude preparation of bovine throm- bin (Table I). To determine whether the inhibitor is present in coelomocytes, coelomic fluid was collected in N-ethyl malemide (5 mM, final concentration), a potent stabilizer of plasma membranes (Bryan el ai, 1964). Analysis of washed coelomocytes collected under the above conditions revealed that lysate preparations of these cells also con- tained the trypsin inhibitor. However, the specific activity of lysates obtained from washed coelomocytes (5.89 /ug trypsin inhibited/mg total protein) was ~ 20-fold less than the specific activity of the cell-free coelomic fluid (90 ^g trypsin inhibited/mg total protein). These results suggest that the trypsin inhibitor is a natural component of coelomocytes and possibly synthesized by these cells and secreted into the coelo- mic fluid. Experiments conducted with cell-free coelomic fluid, which was collected by me- chanical manipulation without damage to sea stars, revealed that trypsin inhibitor levels within the above fluid were comparable to amounts quantitated for coelomic fluid obtained by surgical means. Bovine trypsin was inhibited in a dose-dependent fashion, with ~40% of the protease inactivated by 5 M' of the coelomic fluid (data not shown). The specific inhibitor activity of coelomic fluid collected by non-surgical methods (250 ng trypsin inhibited/mg total protein) was comparable to those ob- tained by surgical removal of the tips of the sea star arm (200 /ag trypsin inhibited/ mg total protein). In conclusion, an inhibitor which inactivates trypsin but not chymotrypsin has been detected in the coelomic fluid of the sea star Asterias forbesi. The above data suggest that the trypsin inhibitor within the coelomic fluid of this invertebrate resem- bles the Kazal-type inhibitors of vertebrates. Although the physiological function of the protease inhibitor remains unclear, it may involve scavenger activity in neutraliz- ing trypsin, which appears in the coelomic fluid, or host defense mechanisms. ACKNOWLEDGMENTS I thank Drs. J. Levin, L. Lorand, and W. Troll for their assistance during these experiments and Drs. David L. Beeler and Gregory A. Marchildon for critically read- SEA STAR TRYPSIN INHIBITOR 361 ing the manuscript. This work was supported by a Frederik B. Bang Fellowship (Ma- rine Biological Laboratory, Woods Hole, Massachusetts). LITERATURE CITED BRYAN, F. T., C. W. ROBINSON, JR., C. F. GILBERT, AND R. D. LANGDELL. 1964. N-ethylmaleimide inhibitions of horseshoe crab hemocyte agglutination. Science 144: 1 147-1 148. DEUTSCH, D. G., AND E. T. MERTZ. 1970. Plasminogen: Purification from human plasma by affinity chromatography. Science 170: 1095-1096. FRITZ, H., AND K. KREJCI. 1976. Trypsin-plasminogen inhibitors (bdellins) from leeches. Methods Enzv- mol. 45: 797-806. GOLDBERG, A. L., K. H. S. SWAMY, C. H. CHUNG, ANDF. S. LARIMORE. 1981. Proteases in Escherichia coli. Methods Enzymol. 80: 680-702. GREENE, L. J., M. H. PUBOLS, AND D. C. BARTELT. 1976. Human pancreatic secretory trypsin inhibitor. Methods Enzymol. 45: 8 1 3-825. HIGHSMITH, R. F., AND R. D. ROSENBERG. 1977. A rapid and sensitive proteolytic assay for human plas- minogen and plasmin using radioiodinated «-casein. Thrombin. Res. 11: 131-140. HOLZMAN, T. F., AND S. F. Russo. 1978. Isolation of an inhibitor of bovine trypsin from the tissues of the purple sea star Pisaster ochraceus. Comp. Biochem. Biophys. 60B: 329-33 1 . KAZAL, L. A., D. S. SPICER, AND R. A. BRAHINSKY. 1948. Isolation of a crystalline trypsin inhibitor- anticoagulant protein from pancreas. J. Am. Chem. Soc. 70: 3034-3040. KUNITZ, M., AND J. H. NORTHROP. 1936. Isolation from beef pancreas of crystalline trypsinogen, trypsin, a trypsin inhibitor, and a trypsin-inhibitor compound. J. Gen. Physiol. 19:991-1007. MARCUM, J. A., R. F. HIGHSMITH, AND D. L. KLINE. 1982. Streptokinase-dependent delayed activation of horse plasminogen. Biochim. Biophys. Ada 702: 19-27. OKABE, H., AND A. NOMA. 1974. Studies on the starfish — III. Activities and distribution of enzymes in the starfish Asterina pectinifera. Comp. Biochem. Physiol. 49B: 599-604. Russo, S. F., AND D. J. YADOFF. 1978. Purification and characterization of trypsin-like enzymes from the purple sea star Pisaster oc hraceus. Comp. Biochem. Physiol. 60B: 453-457. TSCHESCHE, H. 1976. Trypsin-kallikrein inhibitors from cuttlefish (Loligo vulgaris). Methods Enzymol. 45: 792-797. TSCHESCHE, H., ANDT. DIETL. 1976. Trypsin-kallikrein inhibitors from snails (Helix pomatia). Methods Enzymol. 45: 772-785. WINTER, W. P., AND H. NEURATH. 1970. Purification and properties of a trypsin-like enzyme from the starfish Evasterias trochelii. Biochemistry 24: 4673-4679. ZWILLING, R., AND H. NEURATH. 1 98 1 . Invertebrate proteases. Methods Enzymol. 80: 633-664. Reference: Bio/. Bull. 172: 362-366. (June, 1987) A BLOOD-SUCKING SNAIL: THE COOPER'S NUTMEG, CANCELLARIA COOPERI GABB, PARASITIZES THE CALIFORNIA ELECTRIC RAY, TORPEDO CALIFORNICA AYRES J. B. O'SULLIVAN1, R. R. McCONNAUGHEY2, AND M. E. HUBER2 1 Monterey Bay Aquarium, Monterey, CA 93940-1085, and2 Scripps Institution of Oceanography, A-007, LaJolla, California 92093 ABSTRACT Feeding habits of the over 200 living species of Cancellariidae are largely un- known. The Cooper's nutmeg, Cancellaria cooperi, was seen by divers on the dorsal surface of Pacific electric rays. Torpedo californica. C. cooperi was observed to parasit- ize electric rays in the laboratory. The snails made small cuts on the ray's ventral surface and inserted their proboscises into the wounds. Snails were also observed to insert their proboscises into the mouth, gill slits, and anus, as well as previously exist- ing wounds. After insertion of the proboscis, the snails appeared to suck blood from the ray. In Y-maze experiments, the snails actively sought out electric rays, but not other common California bottom fishes. Host location appeared to be by chemosen- sory means. In the absence of electric rays, snails remained buried in the sand without moving for at least 1 2 days. Field observations indicate that snails may travel as much as 24 meters in search of rays. INTRODUCTION The family Cancellariidae comprises over 200 species of marine neogastropod mollusks inhabiting subtidal and bathyal sand and mud bottoms (Harasewych and Petit, 1982, 1984). Mechanisms of feeding in the Cancellariidae are almost entirely unknown (Harasewych and Petit, 1982, 1984; Petit and Harasewych, 1986). Early workers suggested that the snails feed on "soft-bodied microorganisms" (Graham, 1966; Olsson, 1970), but the snails' radular structure, relatively simple digestive sys- tem, long extensile proboscis, and accessory salivary glands are associated with a car- nivorous diet (Harasewych and Petit, 1982, 1984). These and other features suggest that they are piercing fluid feeders (Harasewych and Petit, 1982; Petit and Harase- wych, 1986). The snails also possess organs associated with distance chemoreception (Harasewych and Petit, 1982, 1984; Kohn, 1983). Previous reports of feeding are limited to single observations of cancellariids found on a bivalve mollusk, a gastro- pod, and pieces offish and squid eggs in an aquarium (Cernohorsky, 1 972; Talmadge, 1972; Garrard, 1975). Here we report that Cancellaria cooperi GABB, the Cooper's nutmeg, parasitizes the Pacific electric ray, Torpedo californica AYRES, by sucking the ray's blood. This is the first known report of gastropod mollusks parasitic upon fishes. MATERIALS AND METHODS Field observations and collections were made during scuba dives at depths of 20- 22 m at an artificial reef ("Torrey Pines # 1 " Lat 32°53'12"N Long 1 1 705ff5(TW; Call- Received 21 November 1986; accepted 20 March 1987. 362 C COOPERI PARASITIZES ELECTRIC RAYS 363 fornia Fish and Game Bulletin 45319-800 4-77-20m.OSP). Individuals of C. cooperi of 5-8 cm total shell length were collected by hand at a depth of 20.5 m. Snails were maintained in a 1 20-1 aquarium equipped with running refrigerated seawater — about 14°C — with about 8 cm of sand on the bottom. Rays weighing 1.4-9.1 kg were collected with a hand net by the authors or with gill nets by local fisherman. They were maintained in a 1 200-1 aquarium equipped with running seawater until use in experiments. For laboratory observations of behavior, snails were placed in a second 1 20-1 aquarium, also provided with running seawater and 8 cm of sand. Rays were placed in this aquarium with the snails. In some cases, rays were first transferred to a 160-1 aquarium and anesthetized by the addition of approximately 0.025 g/1 of Tricane Methanesulfonate (Crescent Research Chemicals, Paradise Valley, Arizona). Experiments were conducted in a plastic Y-maze with walls 5 cm high. The base of the 1Y' was 49.5 cm long and each arm was 25 cm long. Seawater from a tank containing an electric ray or other fish was siphoned into one arm, the "stimulus" arm, of the maze at a rate of 0.7 1/min.; seawater from a tank containing no animals was siphoned into the other, "control" arm at the same rate. The stimulus arm was chosen for each trial by a coin toss. The subject snail was placed in an experimental tray at the base of the kY' and allowed to acclimate for 3 min. Snails that did not emerge from their shells after 3 min were allowed to acclimate until they did so, up to a total acclimation period of no more than 5 min. By definition, a "response" occurred when the subject snail moved 5 cm into either arm within 30 min following the acclimation period. In positive responses, snails entered the stimulus arm; in neg- ative responses, they entered the control arm. The maze was washed after each trial. To test the alternate hypotheses that snails locate rays by chemoreception or via the rays1 electric field, we performed Y-maze experiments as described above, except that in one treatment — the "ray water" treatment — seawater from a ray's aquarium was transferred by bucket to a reservoir, then siphoned into the stimulus arm of the maze; plain seawater was siphoned into the other arm. In the control treatment, plain seawater was siphoned into both arms of the maze. Movements by snails in the maintenance aquarium were monitored by placing toothpicks upright in the sand directly adjacent to buried snails and checking the position of the snails relative to that of the toothpicks daily for 12 days. Movements during the intervening periods could be detected because the snails leave distinct trails in the sand. RESULTS During scuba dives, we observed Cancellaria cooperi individuals resting on the dorsal surface of partially buried Pacific electric rays (Torpedo californicd). As many as seven snails were seen on a single ray, and 16 of the 23 individuals of this rare snail which we collected were found in the presence of rays. When maintained in an aquarium with an electric ray, C. cooperi individuals approached the ray and repeatedly touched its surface with extended tentacles. Within a few minutes, the proboscis was extended and used to probe the ray's surface. In most cases, snails made small cuts on the ray, presumably with the radula which is well suited for piercing (Fig. 1 ), and inserted the proboscis into the resulting wound. Sometimes the proboscis was inserted in the ray's mouth, gill slits, anus, or into open wounds, if already present. The proboscis was usually inserted on the ventral surface of the ray even if the ray was anesthetized and placed on its back. After being inserted, the proboscis appeared to pulsate, as if pumping fluid. This behavior was maintained 364 J. B. O'SULLIVAN ET AL. FIGURE 1. Lateral view of the central radular teeth of Cancellaria cooperi. Distal end of the radular ribbon is shown. No lateral teeth are present. Scale bar represents 20 ^m. Scanning electron micrograph courtesy of M. G. Harasewych. for periods of up to 40 minutes. On a few occasions, snails which were subsequently forced by rough handling to retract into their shells exuded a red fluid which appeared to be ray blood upon microscopic examination. Rays showed no apparent response to the snails1 activities. In Y-maze experiments, snails were attracted to water siphoned from tanks con- taining rays (Table I). Snails did not approach other benthic fishes common to the habitat (Table II), though recently we observed C. cooperi feeding on the angel shark, Squatina californica, in the laboratory. Snails were attracted to ray water even if it was not siphoned directly from the ray's aquarium to the maze (Table III). Snails in the aquarium also moved to a micro- scope slide which had been scraped over the ray's surface to collect mucus, though TABLE I Results of Y-maze experiments with Cancellaria cooperi, using the California electric ray, Torpedo californica as the stimulus Snail no. + Response NR 1 4 2 1 2 2 4 1 3 4 1 2 4 4 2 1 5 4 3 0 Total 18 12 5 There was no significant difference in behavior among snails (rows by columns (7-test). Positive re- sponses were significantly more frequent than negative responses (two-tailed binomial P < 0.01 ). C. COOPERI PARASITIZES ELECTRIC RAYS 365 TABLE II Results ofY-maze experiments with various common California hot torn fishes used as stimuli Species + Response NR Torpedo californica (electric ray) 21 6 1 Urolophus halleri (stingray) 0 8 0 Platyrhinoidis halleri (thornback ray) 0 8 0 Eptatretus stoutii (hagfish) 0 8 0 Paralichth ys californicus (halibut) 0 8 0 Differences in snail response among fish species are statistically significant (rows by columns G = 49.8, they did not respond when pieces of fresh ray flesh, with skin attached, were placed in the aquarium. When maintained in aquaria in the absence of rays, snails remained buried in the sand without moving for the entire 1 2 days of observation, but snails moved at speeds of over 14 cm min~' when approaching rays in the maze. In the field, we have retraced the trails left by moving snails for distances of as much as 24 meters from rays. DISCUSSION We conclude that Cancellaria cooperi is a specialized suctorial parasite of the Pacific electric ray, Torpedo californica, and perhaps of certain other bottom fishes such as Squatina californica. Though rare, the snails have been observed on electric rays by other divers (M. Tegner, Scripps Inst. Oceanography, pers. comm.; R. Kiwala, Monterey Bay Aquarium, pers. comm.). Location of rays is by chemosensory means, rather than by detection of the ray's electric field, since snails are attracted to ray water even if not directly connected by siphon to the ray tank. The chemical(s) sensed by the snails are apparently contained in the ray's surface mucus. The snails probably remain buried in the sand for long periods until rays come TABLE III Results ofY-maze experiments with Cancellaria cooperi in which seawaterfrom a ray's tank was transferred by bucket to a reservoir, then siphoned into the stimulus arm of the maze; plain seawater was siphoned into the other arm. In control experiments, plain seawater was siphoned into both arms of the maze Response Treatment + NR Ray water 1 0 Control 0 8 0 The differences in response between treatments are statistically significant (rows by columns G = 1 5.9, P<0.01). 366 J. B. O'SULLIVAN ET AL. within chemosensory distance. By minimizing movement, snails presumably also minimize expenditure of energy in the absence of a food supply. The snails are capa- ble of chemoreception over long distances, and can move rapidly to locate the ray once it is sensed. The ray, like the angel shark, is well suited as host for the snails since both fishes commonly remain on the bottom, partially buried, for long periods of time. The ray's apparent lack of response is in keeping with its behavior: Rays often will not respond even when prodded by divers. Furthermore, at least some cancellari- ids possess a hypobranchial gland which may be involved in toxin production (Hara- sewych and Petit, 1984). It is possible that the secretions of the hypobranchial gland of C. cooperi act as a local anesthetic and/or anticoagulant. This appears to be the first published report of a gastropod mollusk parasitizing a fish (C. Hickman, University of California, Berkeley, pers. comm.; R. Robertson, Academy of Natural Sciences, Philadelphia, pers. comm.). ACKNOWLEDGMENTS We are indebted to M. G. Harasewych, N. Holland, W. Newman, R. E. Petit, the San Diego Natural History Museum, A. Shanks, D. Swails, S. Swails, T. Tomlinson, and W. Wright for their contributions to this research. LITERATURE CITED CERNOHORSKY, W. O. 1972. Marine Shells of the Pacific, Vol. 2. Pacific Publications, Sydney, Australia. GARRARD, T. A. 1975. A revision of the Australian Cancellariidae (Gastropoda: Mollusca). Rec. Aust. Mus.3Q: 1-62. GRAHAM, A. 1966. Fore-gut of marginellid and cancellariid prosobranchs. Stud. Trop. Oceanogr. Miami 4: 134-151. HARASEWYCH, M. G., AND R. E. PETIT. 1 982. Notes on the morphology of Cancel/aria reticulata (Gastro- poda: Cancellariidae). The Nautilus 96: 104-1 13. HARASEWYCH, M. G., AND R. E. PETIT. 1984. Notes on the morphology of Olssonella smithii (Gastro- poda: Cancellariidae). The Nautilus 98: 37-44. KOHN, A. J. 1983. Feeding biology of gastropods. Pp. 4-15 in The Mollusca, Vol. 5, Physiology Pt. 2, A. S. Saleuddin and K. M. Wilbur, eds. Academic Press, New York. OLSSON, A. A. 1970. The cancellariid radula and its interpretation. Palaeontogr. Am. 7: 19-27. PETIT, R. E., AND M. G. HARASEWYCH. 1986. New Phillipine Cancellariidae (Gastropoda: Cancellaria- cea), with notes on the fine structure of the nematoglossan radula. The Veliger 28: 436-443. TALMADGE, R. 1972. Pinky. Of Sea and Shore 3, Gambles, WA. Pp. 189 & 200. INDEX Absorption of phosphate, 337 Actinia tenebrosa, 178 ADAMS, MICHAEL A., see Richard A. Gleeson, 1 Age effects on reproduction, 1 6 1 Aiptasia pulchella, 246 Albumin, 269 Algal symbiosis, 46 Allen symposium, 267 Allometry of deposit feeding in Capitella species I (Polychaeta: Capitellidae): The role of temper- ature and pellet weight in the control of eges- tion.The, 187 Allometry, 30, 187 Alpheid shrimps, 350 Amino acid, 269 Amino acid sequences, 269 Amphidinium carteri, 128 Amphidinium klebsi, 128 Amphipod, 155 ANDERSON, DONALD M., AND PHILLIP S. LOBEL, The continuing enigma of ciguatera, 89 ANDERSON, ROBERT S., Polykaryon formation by Mercenaria mercenaria hemocytes, 236 Androgenic gland, 1 Annelida, 315 Anthopleura asiatica, 202 Anthozoan, 161 Application of an improved method of two-dimen- sional electrophoresis to the systematic study of horseshoe crabs, 2 1 2 Assessment of a rapid enzyme immunoassay stick test for the detection of ciguatoxin and related polyether toxins in fish tissues, 144 Asteria forbesi, 357 AYRE, DAVID J., The formation of clonal territories in experimental populations of the sea anem- one Aclina tenebrosa, 1 78 B BAGNIS, R., M. BARSINAS, C. PRIEUR, A. POMPON, E. CHUNGUE, AND A. M. LEGRAND, The use of the mosquito bioassay for determining the toxicity to man of ciguateric fish, 1 37 Barnacle, 284 BARSINAS, M., see R. Bagnis, 137 Behavior, 307 Behavioral recovery, 83 Benthic dinoflagellates, 89 Bioassay methods, 89 Blood-sucking snail: the Cooper's nutmeg, Cancel- laria cooperi GABB, parasitizes the California electric ray, Torpedo Califarnica AYRES, A, 362 Blue crab, 1,69 Body plan of the cyphonautes larva of bryozoans prevents high clearance rates: comparison with the pluteus and a growth model. The, 30 BOROWSK.Y, BETTY, The effects of embryos of different developmental stages on reproductive behavior and physiology in brooding females of the amphipod crustacean Gammarus palus- iri.s, 155 Brain repair, 83 Breeding systems, 161 Broods, 155 Bryozoa, 30 Burrowing barnacle, 284 Cancellaria cooperi parasitizes electric rays, 362 Clvpeaster subdepressus, 324 CNS lesions, 83 CTX in fish tissues, 144 California electric ray, 362 Callinectes sapidus, 1 , 69 Cancellaria cooperi GABB, 362 Capitella species I, 187 Centennial calendar, 265 Central nervous system, 1 Cestode, 337 Chemical stimulation, 10 Chemoreception, 10 Chlorella. 46 CHORNESKY, ELIZABETH A., AND ESTHER C. PE- TERS, Sexual reproduction and colony growth in the scleractinian coral Forties astreoides, 161 CHUNGUE, E., see R. Bagnis, 1 37 Ciguatera, 108, 122,137 Ciguatera fish, 1 32 Ciguatera fish poisoning, 89 Ciguatera Workshop, papers, 89-153 Ciguateric risk to man, evaluation using mosquito bioassay, 137 Ciguatoxin, 89 Ciguatoxin and maitotoxin on muscle, 132 Ciguatoxin detection, 144 Ciguatoxin molecular structure, 122 Ciguatoxin structure, 122 Cirriped, 284 Cladistic analysis, 2 1 2 Clams, 236 Claw transformation, 350 Clonal territories, 178 Clones in Anthopleura asiatica populations, 202 Clvpeaster rosaceus, 324 Clypeasteroida, 324 Cnidarian, 161 Coelomic fluid, 357 367 368 INDEX TO VOLUME 172 Colonial animal, 161 Continuing enigma of ciguatera, The, 89 Coolia monotis, 128 Cooper's nutmeg, 362 Copper, 69 Coral reefs, 89 Coral reproduction and growth, 1 6 1 Crustacea, 1, 10 Crustacean chemical perception: towards a theory on optimal chemoreception, 10 Cyphonautes, 30 Cyprid, 284 D Decapod, 225 Deposit feeding, 187 Deuterostome-like nephridium in the mitraria larva of Owenia fiisijormis (Polychaeta, Annel- ida), A, 315 DINEEN, JOSEPH R, JR., The larval stages of Litho- trya dorsalis (Ellis & Solander, 1786): a bur- rowing thoracican barnacle, 284 Dinoflagellates, 89 DOOLITTLE, RUSSELL F., The evolution of the ver- tebrate plasma proteins, 269 DUGAN, JENIFER, see A. M. Wenner, 225 DuRAND-CLEMENT, M., Study of production and toxicity of cultured Gambierdiscus toxicus, 108 E Ecdysiotropic activity of juvenile hormone, 299 Ecdysone induced molting in shrimps, 350 Ecdysteroids, 350 Echinoides, 324 Effects of embryos of different developmental stages on reproductive behavior and physiology in brooding females of the amphipod crustacean Gammarus pal list ris, The, 155 Egestion rate, 1 87 Egg production, 225 Egg production by sand crabs (Emerita analoga) as a function of size and year class (Decapoda, Hippidae), 225 Emerita analoga, 225 ENGEL, DAVID W., Metal regulation and molting in the blue crab, Callinectes sapidus: copper, zinc, and metallothionein, 69 Enzyme immunoassay for CTX and related poly- ethers, 144 Epiphytes, 89 Evolution of the vertebrate plasma proteins. The, 269 Excretion, 3 1 5 Exocytosis, 246 Exocytosis of algae, 246 Exogenous 20-hydroxyecdysone, 350 Eyestalk hormone, 1 Feeding activities of two species of Clypeaster (Echinoids, Clypeasteroida): Further evidence of clypeasteroid resource partitioning, 324 Feeding allometry in Capitella, 1 87 Feeding of Clypeaster species, 324 Feeding, 10 Fibrinogen, 269 Filter feeding, 30 Fish poison, 122 Fish poisoning, 89 Flow, 6 1 Flow-assisted mantle cavity refilling in jetting squid, 61 Foraging, 10 FORBES, THOMAS L., AND GLENN R. LOPEZ, The allometry of deposit feeding in Capitella spe- cies I (Polychaeta: Capitellidae): the role of temperature and pellet weight in the control of egestion, 187 Formation of clonal territories in experimental populations of the sea anemone Actinia tene- brosa, 178 FUJII, HARUHIK.O, The predominance of clones in populations of the sea anemone Anthopleura asiatica (Uchida), 202 Gambierdiscus toxicus, 89- 1 53 Gamete dialogue in fertilization: from sea urchin to human, 265 Gammarus broods' effects on behavior, 1 55 GARDINER, STEPHEN L., see Peter R. Smith, 3 1 5 Gastropod, 83 Giant cells, 236 GLEESON, RICHARD A., MICHAEL A. ADAMS, AND AMOS B. SMITH, III, Hormonal modulation of pheromone-mediated behavior in a crusta- cean, 1 GOTO, C. S., see Y. Hokama, 144 GREER, ELIZABETH, see DeForest Mellon Jr., 350 Growth model, 30 H HAROLD, ANTONY S., see Malcolm Telford, 324 Hemocytes, 236 Hemolysin-1, 128 Hemolysin-2, 128 Hermaphrodite, 161 Hippidae, 225 HIRABAYASHI, TAMIO, see J-I. Miyazaki, 212 HOKAMA, Y., L. K. SHIRAI, L. M. IWAMOTO, M. N. KOBAYASHI, C. S. GOTO, AND L. K. NAKAGAWA, Assessment of a rapid enzyme immunoassay stick test for the detection of ci- guatoxin and related polyether toxins in fish tissues, 144 INDEX TO VOLUME 172 369 Hormonal modulation of pheromone-mediated be- havior in a crustacean, 1 Hormone, juvenile, 299 Horseshoe crab, 212 HUBBARD, DAVID M., see A. M. Wenner, 225 HUBER, M. E., see J. B. O'Sullivan, 362 Hydra viridissima, 46 20-hydroxyecdysone, exogenous, 350 Hymenolepis diminuta. 337 Immobilization of Nucella, 307 Immobilization of the predatory gastropod, Nucella lapillus, by its prey, Mytilus edit/is. 307 Induction of precocious molting and claw transfor- mation in alpheid shrimps by exogenous 20- hydroxyecdysone, 350 Inflammation, 236 Invertebrate, 315 Ion channel, 132 IP, Y. K., see A. L. L. Lim, 337 IWAMOTO, L. M., see Y. Hokama, 144 JELLISON, KERRY, see A. M. Wenner, 225 JOH, YONG-GOE, see Kazuo Tachibana, 122 Juvenile hormone, 299 K KOBAYASHI, M. N., see Y. Hokama, 144 Lamprey, 269 Larva, 30, 3 1 5 Larval description, 284 Larval stages of Lilhotrya dorsalis, 284 Larval stages of Lithotrya dorsalis (Ellis & So- lander, 1786): a burrowing thoracican barna- cle, 284 LEGRAND, A. M., see R. Bagnis, 1 37 LIM, A. L. L., AND Y. K. IP, Membrane transport of phosphate by Hymenolepis diminuta, 337 LOBEL, PHILLIP S., see Donald M. Anderson, 89 Lobster, 10 Locomotion, 61, 83 Loligo, 6 1 LOPEZ, GLENN R., see Thomas L. Forbes, 187 Low feeding capability of cyphonautes, 30 Low temperature evokes rapid exocytosis of symbi- otic algae by a sea anemone, 246 Mantle cavity refilling in squid, 6 1 MARCUM, JAMES A., A trypsin inhibitor from the coelomic fluid of the sea star Asterias forbesi, 357 Marine biotoxins, 1 37 Marine invertebrates, 307 McCONNAUGHEY, R. R., see J. B. O'Sullivan, 362 MCEDWARD, LARRY R., AND RICHARD R. STRATHMANN, The body plan of the cypho- nautes larva of bryozoans prevents high clear- ance rates: comparison with the pluteus and a growth model, 30 MELLON, DEFOREST, JR., AND ELIZABETH GREER, Induction of precocious molting and claw transformation in alpheid shrimps by exoge- nous 20-hydroxyecdysone, 350 Membrane transport of phosphate by Hymenolepis diminuta, 337 Mercenaria mercenaria, 236 Metal regulation and molting in the blue crab, Cal- linectes sapidus: copper, zinc, and metallothio- nein, 69 Metallothionein, 69 Mixture suppression, 10 MIYAZAK.I, JUN-ICHI, KOICHI SEKIGUCHI, AND TAMIO HIRABAYASHI, Application of an im- proved method of two-dimensional electro- phoresis to the systematic study of horseshoe crabs, 2 1 2 Modal analysis, 225 MOFFETT, STACIA, see Keith Snyder, 83 Mollusca, 61 Molting and metallothionein, 69 Monroy symposium, 265 Mooi, RICH, see Malcolm Telford, 324 Moray eel, 122 Mosquito bioassay, 1 37 MULLER-PARKER, G., AND R. L. PARDY, Re- sponse of green hydra to feeding and starvation at four irradiances, 46 Multinucleated giant cells in bivalves, 236 MURAKAMI, YASUTAKA, see T. Yasumoto, 128 MURATA, MICHIO, seeT. Yasumoto, 128 MUSCATINE, L., see L. Grant Steen, 246 Mytilus edulis, 307 N NAKAGAWA, L. K., see Y. Hokama, 144 Nauplii, 284 Nephridium, 315 Nephridium of the mitraria larva, 3 1 5 Neural regeneration, 83 Nucella lapillus, 307 NUKINA, MANABU, see Kazuo Tachibana, 122 M Macroalgae, 89 Maitotoxin, 89, 128, 132 Manduca sexta, 299 o O'SULLIVAN, R. R. MCCONNAUGHEY, AND M. E. HUBER, A blood-sucking snail: the Cooper's nutmeg. Cancel/aria cooperi GABB, parasit- 370 INDEX TO VOLUME 172 izes the California electric ray. Torpedo califor- nica AYRES, 362 Odor perception, 10 OHIZUMI, YASUSHI, Pharmacological actions of the marine toxins ciguatoxin and maitotoxin isolated from poisonous fish, 1 32 Optical approaches to the dynamics of cellular mo- tility, 267 Optimal chemoreception, 10 Optimal foraging, 10 Ostre ops is ovata, 128 Ostreopsis siamensis, 128 PL toxin- 1, 128 PLtoxin-2, 128 Pamilirus, 10 PARDY, R. L., see G. Muller-Parker, 46 PETERS, ESTHER C, see Elizabeth A. Chornesky, 161 PETRAITIS, PETER S., Immobilization of the preda- tory gastropod, Nucella lapillus. by its prey, Mytilus edulis, 307 Pharmacological actions of the marine toxins cigu- atoxin and maitotoxin isolated from poison- ous fish, 132 Phenetic analysis, 2 1 2 Pheromone, 1 Phosphate, 337 Phosphate transport, 337 Phosphate uptake, 337 Phylogeny, 212 Plasma proteins, 269 Pluteus, 30 Podia, food collecting, 324 Polychaeta, 187,315 Polykaryon formation by Mercenaria mercenaria hemocytes, 236 POMPON, A., see R. Bagnis, 1 37 Porites astreoides Lamarck, 1 6 1 Precocious molting, 350 Predominance of clones in populations of the sea anemone Anthopleura asiatica (Uchida), The, 202 PRIEUR, C., see R. Bagnis, 1 37 Prorocentrum concavwn, \ 28 Prowcentrum lima, 128 R Rapid enzyme immunoassay stick test, 144 Recent developments in the molecular structure of ciguatoxin, 122 Reproductive behaviors, 155 Resource partitioning, 324 Response of green hydra to light, 46 Rocky intertidal shores, 307 RUPPERT, EDWARD E., see Peter R. Smith, 315 SAFRANEK, Louis, AND CARROLL M. WILLIAMS, Studies of the ecdysiotropic activity of juvenile hormone in pupae of the hormone tobacco horn worm, Manduca sexta, 299 Sand crab egg production, 225 SCHEUER, PAUL J., see Kazuo Tachibana, 122 Scleractinia, 161 Sea anemone, 1 78, 202, 246 Sea star, 357 Sea star trypsin inhibitor, 357 Second symposium on regulatory actions of growth factors in reproduction and early develop- ment, 266 SEINO, NAOKO, see T. Yasumoto, 128 SEKJGUCHI, KOICHI, see J-I. Miyazaki, 212 Sexual reproduction and colony growth in the scler- actinian coral Porites astreoides, 1 6 1 SHIRAI, L. K., see Y. Hokama, 144 SHOFFNER, JULIE, see A. M. Wenner, 225 Size effects on reproduction and growth, 161 SMITH, AMOS B., Ill, see Richard A. Gleeson, 1 SMITH, PETER R., EDWARD E. RUPPERT, AND STE- PHEN L. GARDINER, A deuterostome-like ne- phridium in the mitraria larva ofOwenia fnsi- formis (Polychaeta, Annelida), 3 1 5 Snail locomotion after CNS lesions, 83 Snail, blood-sucking, 362 SNYDER, KEITH, AND STACIA MOFFETT, Loss and recovery of locomotor behavior after CNS le- sions in the snail Melampus bidentatm. 83 Squid, 61 STEEN, R. GRANT, AND L. MUSCATINE, Low tem- perature evokes rapid exocytosis of symbiotic algae by a sea anemone, 246 Stick enzyme immunoassay forCTX, 144 STRATHMANN, RICHARD R., see Larry R. McEdward, 30 Studies of the ecdysiotropic activity of juvenile hor- mone in pupae of the tobacco hornworm, Manduca sexta, 299 Study of production and toxicity of cultured Gam- bierdiscus toxic us, 108 Suspension feeding, 30 Swimming, 61 Systematic study of horseshoe crabs, 2 1 2 TACHIBANA, KAZUO, MANABU NUKINA, YONG- GOE JOH, AND PAUL J. SCHEUER, Recent de- velopments in the molecular structure of cigu- atoxin, 122 Tapeworm, 337 TELFORD, MALCOLM, RICH Mooi, AND ANTONY S. HAROLD, Feeding activities of two species of Clypeaster (Echinoides, Clypeasteroida): fur- ther evidence of clypeasteroid resource parti- tioning, 324 Temperature effects, 187 INDEX TO VOLUME 172 371 Tobacco hornworm, see Manduca sexta, 299 Torpedo californica AYRES, 362 Toxin production, 108 Toxins produced by benthic dinoflagellates, 1 28 Transport, 337 Tropical plankton, 284 Trypsin inhibitor from the coelomic fluid of the sea stsrAsteriasforbesi, A, 357 Trypsin, 357 Two-dimensional electrophoresis, 2 1 2 u infrastructure, 3 1 5 Use of the mosquito bioassay for determining the toxicity to man of ciguateric fish. The, 1 37 VOGEL, STEVEN, Flow-assisted mantle cavity re- filling in jetting squid, 6 1 w WENNER, ADRIAN M., DAVID M. HUBBARD, JEN- IFER DUGAN, JULIE SCHOFFNER, AND KERRY JELLISON, Egg production by sand crabs (Em- erita analoga) as a function of size and year class (Decapoda, Hippidae), 225 WILLIAMS, CARROLL M., see Louis Safranek, 299 YASUMOTO, TAKESHI, NAOKO SEINO, YASUTAKA MURAKAMI, AND MICHIO MURATA, Toxins produced by benthic dinoflagellates, 128 Year classes, 225 ZIMMER-FAUST, RICHARD K., Crustacean chemi- cal perception: towards a theory on optimal chemoreception, 10 Zinc, 69 Zooxanthellae, 246 581*5 06 CONTENTS MARINE BIOLOGICAL LABORATORY — CALENDAR OF CENTENNIAL EVENTS 265 INVITED ARTICLE DOOLITTLE, RUSSELL F. The evolution of the vertebrate plasma proteins 269 DEVELOPMENT AND REPRODUCTION DINEEN, JOSEPH F., JR. The larval stages of Lithotrya dorsalis (Ellis & Solander, 1786): a bur- rowing thoracican barnacle 284 SAFRANEK, Louis, AND CARROLL M. WILLIAMS Studies of the ecdysiotropic activity of juvenile hormone in pupae of the tobacco hornworm, Manduca sexta , . 299 ECOLOGY AND EVOLUTION PETRAITIS, PETER S. Immobilization of the predatory gastropod, Nucella lapillus, by its prey, Mytilus edulis : . .,.^.' 307 GENERAL BIOLOGY SMITH; PETER-R.; EDWARD JE, RUPPERT, AND STEPHEN L. GARDINER A deuterostome-like nephridium in the mitraria larva of Owenia fusi- formis (Polychaeta, Annelida) 315 TELFORD, MALCOLM, RICH Moor, AND ANTONY S. HAROLD Feeding activities of two species of Clypeaster (Echinoides, Clypeaster- oida):4iirther evidence of clypeasteroid resource partitioning 324 PHYSIOLOGY : ..' ' LIM, A.cL, L.., AND.Y. K. IP Membrane transport of phosphate by Hymenolepis diminuta 337 MELLON, DEFOREST, JR., AND ELIZABETH GREER Induction of precocious molting and claw transformation in alpheid shrimps by exogenous 20-hydroxyecdysone 350 ( ; ^r^ "5 ^ * ^ SHORT REPORTS MARCUM, JAMES A. A trypsin inhibitor from the coelomic fluid of the sea star Asterias forbesi 357 O'SULLIVAN, J. B., R. R. MCCONNAUGHEY, AND M. E. HUBER .A blood-sucking snail: the Cooper's nutmeg, Cancellaria cooperi GAAB, parasitizes the California electric ray, Torpedo californica AYRES 362 INDEX TO VOLUME 172 . 367 MBL WHOI LIBRARY UH 1BEB V