THE BIOLOGICAL BULLETIN FEBRUARY 1999 2 3 1999 Editor Associate Editors Section Editor Online Editors Editorial Board Editorial Office MICHAEL J. GREENBERG Louis E. BURNETT CHARLES D. DERBY MICHAEL LABARBARA RUDOLF A. RAFF SHINYA INDUE, Imaging and Microscopy JAMES A. BLAKE, Keys to Marine Invertebrates of the Woods Hole Region WILLIAM D. COHEN, Marine Models Electronic Record and Compendia PETER B. ARMSTRONG ANDREW R. CAMERON ERNEST S. CHANG THOMAS H. DIETZ RICHARD B. EMLET DAVID EPEL GREGORY HINKLE MAKOTO KOBAYASHI DONAL T. MANAHAN MARGARET McFALL-NGAi MARK W. MILLER TATSUO MOTOKAWA YOSHITAKA NAGAHAMA SHERRY D. PAINTER J. HERBERT WAITE RICHARD K. ZIMMER-FAUST PAMELA CLAPP HINKLE VICTORIA R. GIBSON CAROL SCHACHINGER PATRICIA BURNS The Whitney Laboratory, University of Florida Grice Marine Biological Laboratory, College of Charleston Georgia State University University of Chicago Indiana University Marine Biological Laboratory ENSR Marine & Coastal Center, Woods Hole Hunter College, City University of New York University of California, Davis California Institute of Technology University of California, Davis Louisiana State University Oregon Institute of Marine Biology, Univ. of Oregon Hopkins Marine Station, Stanford University Cereon Genomics, Cambridge, Massachusetts Hiroshima University of Economics, Japan University of Southern California Kewalo Marine Laboratory, University of Hawaii Institute of Neurobiology, University of Puerto Rico Tokyo Institute of Technology, Japan National Institute for Basic Biology, Japan Marine Biomed. Inst., Univ. of Texas Medical Branch University of California, Santa Barbara University of California, Los Angeles Managing Editor Staff Editor Editorial Assistant Subscription & Advertising Secretary Published by MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASSACHUSETTS Cover The hydrozoan Podocoiyne carnca (Phylum Cni- daria), consists in its colonial benthic phase of two principal elements: polyps and stolons. The polyps are the feeding members of the colony and grow to a length of about 1 mm. The stolons are branched, vascular tissues that connect the gastric cavities of polyps; these gastrovascular vessels are about 60-70 /im in external diameter. Polyps are propagated asexually through the differentiation of cells in the stolon. Typically, hydractiniid colonies, like those of Podocoiyne , comprise several hundred polyps, all interconnected by a network of branched stolons that covers the gastropod shells occupied by hermit crabs (diagram, top; reproduced from G. J. Allman. 1872. A monograph of the gymnoblastic or tubularian hydroids. Part II. Ray Society, London. 155-450). Individual polyps with a bit of attached stolon can be transplanted onto microscope slides where they generate clonal replicates of colonies for experimental use. Such a cloned colony, growing on the edge of a slide, is shown on the left (photo- graph by Neil Blackstone). A vegetative polyp in the center of the picture is flanked by polyps that are budding off medusae (containing orange struc- tures). The relatively small size and transparency of these hydractiniid hydroids, as well as the ease with which replicate colonies can be produced, make them excellent systems with which to study the physiology of vascular fluid transport. The colony- wide distribution of nutrients and dissolved gases is now thought to play a critical role in the morpho- genesis and life history of the colony. In this issue. Dudgeon et al. use videomicroscopy and automated image analysis to characterize the feed- ing-related dynamics of the gastrovascular system, particularly the oscillations of polyps and stolons that occur during and after feeding. The photomicro- graph at right (differential interference contrast, 100X; photograph by Neil Blackstone) shows a branching stolon of Podocoiyne camect in its open state; that is, the lumen of the stolon is expanded, permitting the export of fluid from the polyps into the colonial gas- trovascular system. In contrast, the lumen of the nearby stolon tip is closed. This investigation is the first quantitative treatment of cnidarian feeding behav- ior at high temporal resolution. It suggests that a cnidarian colony can be reasonably and readily treated as a system of coupled nonlinear oscillators distributed in space. CONTENTS VOLLIMI-: 196. No. 1: FEBRUARY 1999 PHYSIOLOGY Dudgeon, Steve, Andreas Wagner, J. Rinias Vaisnys, and Leo W. Buss Dynamics of gastrovascular circulation in the hvdro- zoan Podocaiyne ranifii: the one-polyp case 1 Kelly, Robert H., and Paul H. Yancey High contents of trimethylamine oxide correlating with depth in deep-sea teleost fishes, skates, and decapod crustaceans IS Beaven. Amy E., and Kennedy T. Paynter Acidification of the phagosome in ('.raMtistn-a i'/>'g/>i/ru hemocytes following engulfment of zymosan 26 Smith, Andrew M., Tonya J. Quick, and Rachel L. St Peter Differences in the composition of adhesive and non- adhesive mucus from the limpet Lottiu limatula .... 34 Kawaii, Satom, Keiji Yamashita. Mitsuyo Nakai, Miyuki Takahashi, and Nobuhiro Fusetani Calcium-dependence of settlement and nematocyst discharge in actinulae of the hydroid Tnbulurin mesembryanthemum 45 RESEARCH NOTE Tapley, David W., Garry R. Buettner, and J. Malcolm Shick Free radicals and chemiluminescence as products of the spontaneous oxidation of sulfide in seawater, and their biological implications 52 DEVELOPMENT AND REPRODUCTION Froggett, Stephan J., and Esther M. Leise Metamorphosis in the marine snail Ilyanassa obsoleta, ves or NO?. . 57 Stewart-Savage, J.. Bradley J. Wagstaff. and Philip O. Yund Developmental basis of phenotypic variation in egg production in a colonial ascidian: primary oocyte production versus oocyte development 63 Schwarz. Jodi A., David A. Krupp, and Virginia M. Weis Late larval development and onset of symbiosis in the scleractinian coral [''img/a srularia 70 80 ECOLOGY AND EVOLUTION Lopez, Jose V., Ralf Kersanach, Stephen A. Rehner, and Nancy Knowlton Molecular determination of species boundaries in corals: genetic analysis of the Montastraea annulrtris complex using amplified fragment length polymor- phisms and a microsatellite marker Bingham, Brian L., and Nathalie Reyns Ultraviolet radiation and distribution of the solitary ascidian Corella inflata (Huntsman) 94 NEUROBIOLOGY AND BEHAVIOR Cromarty, S.I., J. Mello, and G. Kass-Simon Time in residence affects escape and agonistic behav- ior in adult male American lobsters . . 105 ULTRASTRUCTURE Hirose, Euichi, Satoshi Kimura, Takao Itoh, and Jun Nishikawa Tunic morphology and cellulosic components of py- rosomas, doliolids, and salps (Thaliacea, Urochor- data) . 113 THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory, 7 MBL Street, Woods Hole, Massachusetts 02543. Subscriptions and similar matter should be addressed to Subscription Manager, THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, 7 MBL Street, Woods Hole, Massachusetts 02543. Subscription per year (six issues, two volumes): $205 for libraries; $95 for individuals. Subscription per volume (three issues): $102.50 for libraries: S47.50 tor individuals. Back and single issues (subject to availability): $40 for libraries; $20 for individuals. Communications relative to manuscripts should be sent to Michael J. Greenberg. Editor-in-Chief, or Pamela Clapp Hinkle. Managing Editor, at the Marine Biological Laboratory, 7 MBL Street. Woods Hole, Massachusetts 02543. Telephone: (508) 289-7428. FAX: 508-457-1924. E-mail: pclapp@mbl.edu. http://www.mbl.edu/BiologicalBulletin/ The home page for the electronic companion to THE BIOLOGICAL BULLETIN the Marine Models Electronic Record and other BIOLOGICAL BULLETIN publications is available on the World Wide Web at the address shown above. THE BIOLOGICAL BULLETIN is indexed in bibliographic services including Index Medicus and MEDLINE. Chemical Abstracts. Current Contents, Elsevier BIOBASE/Current Awareness in Biological Sciences, and Geo Abstracts. Printed on acid free paper, effective with Volume 180. Issue 1, 1991. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, 7 MBL Street. Woods Hole, MA 02543. Copyright 1999. by the Marine Biological Laboratory Periodicals postage paid at Woods Hole. MA. and additional mailing offices. ISSN 0006-3185 INSTRUCTIONS TO AUTHORS The Biological Bulletin accepts outstanding original research reports of general interest to biologists throughout the world. Papers are usually of intermediate length (10-40 manuscript pages). A limited number of solicited review papers may be accepted after formal review. A paper will usually appear within four months after its acceptance. Very short, especially topical papers (less than 9 manuscript pages including tables, figures, and bibliography) will be published in a separate section entitled "Research Notes." A Research Note in The Biological Bulletin follows the format of similar notes in Nature. It should open with a summary paragraph of 150 to 200 words comprising the introduction and the conclusions. The rest of the text should continue on without subheadings, and there should be no more than 30 references. References should be referred to in the text by number, and listed in the Literature Cited section in the order that they appear in the text. Unlike references in Nature. references in the Research Notes section should conform in punctuation and arrangement to the style of recent issues of The Biological Bulletin. Materials and Methods should be incorpo- rated into appropriate figure legends. See the article by Loh- mann et al. (October 1990. Vol. 179: 214-218) for sample style. A Research Note will usually appear within two months after its acceptance. The Editorial Board requests that regular manuscripts con- form to the requirements set below; those manuscripts that do not conform will be returned to authors for correction before review. 1. Manuscripts. Manuscripts, including figures, should be submitted in quadruplicate, with the originals clearly marked. ( Xerox copies of photographs are not acceptable for review pur- poses.) The submission letter accompanying the manuscript should include a telephone number, a FAX number, and (if possible) an E-mail address for the corresponding author. The original manu- script must be typed in no smaller than 12 pitch or 10 point, using double spacing (including figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper, 8 by 1 1 inches. Please, no right justification. 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. 5th Edition (Council of Biology Editors, 1983) and to American spelling. Unusual abbreviations should be kept to a minimum and should be spelled out on first reference as well as defined in a footnote on the title page. Manuscripts should be divided into the following components: Title page. Abstract (of no more than 200 words). Introduction, Materials and Methods. Results, Discussion. Acknowledgments, Literature Cited, Tables, and Figure Legends. In addition, authors should supply a list of words and phrases under which the article should be indexed. 2. Title page. The title page consists of a condensed title or running head of no more than 35 letters and spaces, the manuscript title, authors' names and appropriate addresses, and footnotes listing present addresses, acknowledgments or contribution num- bers, and explanation of unusual abbreviations. 3. Figures. The dimensions of the printed page. 7 by 9 inches, should be kept in mind in preparing figures for publication. We recommend that figures be about 1 times the linear dimensions of the final printing desired, and that the ratio of the largest to the smallest letter or number and of the thickest to the thinnest line not exceed 1:1.5. Explanatory matter generally should be included in legends, although axes should always be identified on the illustra- tion 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 and cited, in order, in the text. The author's name and an arrow indicating orientation should appear on the reverse side of all figures. Color: The Biological Bulletin will publish color figures and plates, but must bill authors for the actual additional cost of printing in color. The process is expensive, so authors with more than one color image should consistent with editorial concerns, especially citation of figures in order combine them into a single plate to reduce the expense. On request, when supplied with a copy of a color illustration, the editorial staff will provide a pre-publi- cation estimate of the printing cost. 4. Tables, footnotes, figure legends, etc. Authors should follow the style in a recent issue of The Biological Bulletin in preparing table headings, figure legends, and the like. Because of the high cost of setting tabular material in type, authors are asked to limit such material as much as possible. Tables, with their headings and footnotes, should be typed on separate sheets, num- bered with consecutive Roman numerals, and placed after the Literature Cited. Figure legends should contain enough informa- tion 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 contri- bution numbers, and explanation of unusual abbreviations. All such footnotes should appear on the title page. Footnotes are not normally permitted in the body of the text. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more than two authors cited as Jones et al.. 1980. Personal communications and material in prep- aration 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 ABSTRACTS, with the minor differences set out below. The most generally useful list of bio- logical journal titles is that published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of Serials; the most recent issue). Foreign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PERIODICALS, may find a booklet published by the Biological Council of the U.K. (obtainable from the Institute of Biology, 41 Queen's Gate, London, S.W.7, England, U.K.) useful, since it sets out the WORLD LIST abbreviations for most biological journals with notes of the USASI abbreviations where these differ. CHEMICAL ABSTRACTS publishes quarterly supplements of addi- tional 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 abbreviations invented by the author. For example, use Rit Vfsindafjelags Islendinga without abbreviation. F. All single word journal titles in full (e.g., Veliger, Ecol- ogy. Brain). G. The order of abbreviated components should be the same as the word order of the complete title (i.e., Proc. and Trans. placed where they appear, not transposed as in some BIOLOGICAL ABSTRACTS listings). H. A few well-known international journals in their 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, page proofs, and charges. Authors of articles in black and white (no color figures) receive their first 50 reprints (without covers) free of charge. Color reprints and additional black-and-white reprints may be purchased; authors will receive order forms. Reprints normally will be delivered about 2 to 3 months after the issue date. Authors (or delegates for foreign authors) will receive page proofs of articles shortly before publi- cation. They will be charged the current cost of printers' time for corrections to these (other than corrections of printers' or editors' errors). Other than these charges for authors' alterations, Ttie Biological Bulletin does not have page charges. Rctca-ncc: Bin/. Bull. 196: 1-17. (Februarv. 1999) Dynamics of Gastrovascular Circulation in the Hydrozoan Podocoryne carnea: the One-Polyp Case STEVE DUDGEON 1 -*, ANDREAS WAGNER 2 , J. RIM AS VAISNYS 3 ' 4 , AND LEO W. BUSS 3 - 5 Department of Biology, California State University, Northridge, California 91330-8303: Department of Biology, The University of New Mexico, 167A Castetter Hall, Albuquerque, New Mexico 87131-1091; 3 Departments of Ecology and Evolutionary Biology, ^Electrical Engineering, ^Geology & Geophysics, Yale University, New Haven, Connecticut 1)6520 Abstract. Time-lapse video microscopy and image anal- ysis algorithms were used to generate high-resolution time series of the length and volume of a single hydrozoan polyp before and after feeding. A polyp of Podocoryne carnea prior to feeding is effectively static in length and volume. At 20C, feeding elicits 8-millihert/ (mHz) oscillations in polyp length and volume. A polyp connected to a colony by a single stolon displayed an abrupt transition from low- amplitude. 8-mHz oscillations to large-amplitude, 6-mHz oscillations at 1 .5-2 h after feeding. The transition was preceded by a substantial decrease in polyp volume and increase in length which coincided with the export of food items from the digestive cavity of the polyp into the colonial gastrovascular system. In contrast. 8-mHz oscillations of a polyp isolated from a colony continued for 12.7 h after feeding, at which time particulates from the digestive cavity were exported into the hydrorhiza and a 4-mHz subhar- monic became briefly dominant. Regular oscillatory behav- ior was terminated by regurgitation at comparable intervals post-feeding in coupled and isolated polyps. These obser- vations are compatible with the hypothesis that the presence of nutrients in the digestive cavity induces polyp oscilla- tions and that release of nutrients into the gastrovascular system similarly induces unfed polyps to oscillate, thereby distributing the contents of the fed polyp throughout the colony. Introduction The gastrovascular system is the only physiological sys- tem of hydrozoans whose behavior is known to be mani- Received 20 February 1998; accepted 25 November 1998. * E-mail: steve.dudgeon@csun.edu fested colonywide. The system transports fluid between the digestive cavities of polyps through the lumens of the endodermal canals between polyps, resulting in the colony- wide exchange of nutrients and dissolved gases. Recent studies have shown that perturbation of gastrovascular transport has marked effects upon colony ontogeny and life history. Specifically, the production of polyps and the fre- quency of stolon branching and anastomosis are acceler- ated, and the age at which medusae are produced is altered, in Podocoryne carnea (Sars, 1846), by perturbations of energetic metabolism that reduce the volumetric flow rate through stolons (Blackstone and Buss, 1992, 1993; Black- stone, 1997, 1998). Moreover, surgically manipulating the relative sizes of stolons within a colony of Hydractinia symbiolongicarpus (Buss and Yund, 1989) is sufficient to stably convert a runner-like colony into a sheet-like colony and vice versa (Dudgeon and Buss. 1996). Control of vascular morphology by response to internal hydromechanical signals is increasingly well-known in ver- tebrate systems (Bevan et ai, 1995). Murray (1926) pro- posed that tree-like vascular designs minimize the total energy expended in propelling the fluid and maintaining the tissues. The predicted optimum is one in which the wall shear stress is constant throughout (Zamir, 1977; Sherman. 1981; LaBarbera, 1990). Several genes are known to be differentially expressed upon perturbation of wall shear stress in a fashion that adjusts vessel radii to values that restore a systemwide constant shear stress (Bevan et ai, 1995). Similar design optimizations and flow-dependent gene expression may underlie the response of hydrozoan colonies to altered patterns of gastrovascular transport. However, the task of identifying the relevant hydrome- chanical features and the patterning elements that respond to S. DUDGEON ET AL such features requires an understanding of how fluid circu- lates within a colony. Unlike the vertebrates, which have a vascular system with a single pump that propels j unidirectional flow of fluid within a dichotoim usly branching tree, a hydrozoan colony can be compoM : of thousands of individual polyps con- nected to one another by a complex array of anastomosing stolons witiii'i which fluid may flow alternately in any direction I nderstanding how an array of pumps and vessels generate a time- and space-varying distribution of metabo- lites and hydrodynamic signals is a considerable challenge. Our approach has been to characterize the dynamics of a single polyp, in the hope that the behavior of an isolated unit will prove simple enough to allow us to develop a mathe- matical model. Polyp models, when coupled by suitably developed models of the stolon, may eventually permit systematic analysis of the consequences of the arrangement of polyps in various geometries. To this end, we have documented the feeding behavior of single polyps of the colonial hydroid Podocoryne camea. We present time se- ries of the length and volume of a single isolated polyp and contrast this behavior with that of a single polyp coupled to a colony, both before and after feeding. Materials and Methods Animals and their maintenance The hydrozoan Podocoryne cornea produces encrusting colonies of a typical filiform form. Our observations are restricted to young colonies bearing only gastrozooids (hereafter called polyps). Polyps extend upright atop the stolons, which adhere to the substratum. The polyp is the sole component of the system that can exchange fluid with both the external medium (via the mouth) and the rest of the gastrovascular system (via a contractible opening between its gastric cavity and the stolon or stolons coextensive with it). Exclusive of epithelial conductance, the gastrovascular system is the only known colonywide conducting system in this species (i.e.. neither a nerve net nor muscle fibers occur in the stolons; Stokes. 1974; Schierwater et al., 1992). All colonies of P. carnea were asexually propagated from a single clone (P34) collected at the Peabody Museum Field Station, Guilford, Connecticut, in September 1989. Colo- nies were grown on the surface of either glass slides (25 X 20 X 1 mm) or coverslips (484 mm 2 ) in 40-1 aquaria containing artificial seawater (REEF CRYSTALS. Aquar- ium Systems, Mentor, Ohio) at 18 1C. Animals were fed to repletion twice per week on a diet of 3- to 5-day-old brine shrimp (Anemia salina (Linnaeus, 1758)) nauplii. Colonies were propagated asexually by surgically explanting single polyps onto the surface of a glass slide or coverslip. Ex- planted polyps were held in place by a loop of thread until the growth of stolons attached them to the surface. Observational protocol Colonies were not fed for 2 days prior to treatment and subsequent observation. Polyps growing on the edge of a glass slide were standardized to lengths of 100 /J,m, and stolons were severed as necessary to establish two treat- ments. For observation of isolated polyps, all stolons con- necting the chosen polyp to the colony were severed so as to retain a segment of stolon roughly 800 /^m long with two blind ends and the polyp positioned near the center. Severed stolons heal and become occluded instantly (Berrill, 1953). For observations of coupled polyps, all but one stolonal connection to the colony was severed. In this case, the polyp was situated about 300-400 p.m from the blind end of the stolon. The size of the colony varied between replicates, but was in all cases vastly larger than the internal volume of the chosen polyp. Following surgery, the colony was maintained under standard conditions for 1 2 h prior to the start of observa- tions. The colony to be observed was then placed within a temperature-controlled chamber in 10 ml of 0.45-p.m fil- tered seawater at 20 0.1 C and viewed at 100X using a Zeiss Axiovert 35 inverted microscope. The polyp was positioned so as to provide a longitudinal profile extending from the mouth to the base of the polyp. Illumination was arranged to optically filter the tentacles, leaving only the body column visible (i.e., the field was flooded with light sufficient to render the polyp outline black and the remain- der of the field uniformly white). Polyp behavior was vid- eotaped, typically for 1 .5-2 h prior to feeding, using a Dage MTI camera connected to the microscope and a videocas- sette recorder. The polyp was then removed from the cham- ber, hand-fed a single newly hatched brine shrimp nauplius, returned to the chamber immediately after ingesting the food item, and videotaped for the following 24 h. Image analysis Images were recovered from the videotaped record of polyp behavior using a PCVISION frame grabber and OP- TIMAS image analysis software. A series of programs, written in the OPT1MAS macro language, were used to extract polyp length and diameter at multiple points along the longitudinal axes from each binary image of a polyp's outline. Figure 1 and its legend illustrate the steps by which length and width measurements are generated from an im- age. Polyp volume was estimated from these measurements of polyp length and diameters, using the extended Simp- son's rule (Press et al.. 1992). The macros used and an ex- tensive discussion of the reasoning which led to their de- velopment are available at http://www.csun.edu/~sd51881. There are three sources of error in the procedures we employ: ( 1 ) errors associated with sampling the coordinates that compose the outline of the polyp, (2) errors associated with the assumption that the polyp is rotationally symmetric GASTROVASCULAR CIRCULATION Figure 1. Steps used by image-processing algorithm following inversion of the binary image (black pixels to white and vice versa). (A) Establishing 11 evenly spaced transects along the longitudinal polyp axis; (B) detecting and marking the edge of the polyp at both ends of each transect line; (C) determining the midpoint of each transect line along the body column with respect to the marked edges of the polyp and connecting the midpoint segments along the longitudinal polyp axis; and (D) establishing the line normal to the midpoint-to- midpoint line segment at each transect the length of these normal lines constitute the diameter measurements. (E) Example illustrating the switch between horizontal and vertical scans by the algorithm when a bend of the polyp exceeding 53 is detected. (Fl Fitting a 4th order polynomial function to the set of midpoint coordinates for determining polyp length. about the axis of the focal plane, and (3) errors associated with bending of the polyp outside the focal plane. The first two sources of error were estimated from an analysis of sampling efficiency and by extensive experimentation on volume fluxes observed in colonies of known stolonal volume; they constitute an error of < 5%, which corresponds to an error in estimates of volume amplitude of 0.7 nl. The relevant experiments and data upon which this error estimate is based are available at http://www.csun.edu/~sd5 1 88 1 . With respect to the remaining source of error, our algorithms cannot detect the bending of polyps outside the focal plane. Such bends, however, consti- tuted only a small fraction of the overall record (5.9% and 2.6% in the coupled and isolated records, respectively). These data points are spurious and are denoted as such by tickmarks along the abcissa in plots of the time series we present below. Time-series aiwlvsis Using the algorithms described above, one of the four replicates in each treatment was analyzed to generate a high-resolution time series of polyp length and estimated volume. Measurements were made at 8-s intervals from the onset of feeding to 436:09 and 1474:12 min:s post-feeding for coupled and isolated treatments, respectively. Time-series data were analyzed using Mathematica (Ver- sion 2.2. Wolfram Research, Inc.) on a Hewlett-Packard Apollo 9000 workstation. From each raw time series of length and volume, we calculated a low- and high-pass- filtered version (Priestley, 1981). A low-pass-filtered time series is one from which short-term fluctuations have been removed. The low-pass-filtered time series was computed as a sliding running average, using a uniform window S. DUDGEON ET AL spanning 201 points and centered about the point in ques- tion. The (complementary) high-pass-filtered time series was obtained by subtracting the low-pass-filtered series from the original series, giving a series with a mean of zero from which long-term drifts had been removed. To detect trends within the time series, the high-pass- filtered series was further analyzed by examining a succes- sion of windowed Fourier transformations. The entire high- pass-filtered time series was divided into segments ("windows") comprising 20 min of observational data, with two consecutive windows overlapped by 15 min. Fourier coefficients were estimated for each of the windows sepa- rately using a Fast Fourier transformation, and the power spectrum was calculated. Efficient techniques for analyzing time-series data de- pend on having a complete time series that has been sam- pled at equal time intervals. If data points are missing, an error proportional to the number of missing points is intro- duced into the estimation of the series spectrum. In the series analyzed here, missing observations are rare; they contribute 1.90% and 0.39% to the variability in the spec- trum in the coupled and isolated treatments, respectively. "Aliasing," a frequent problem in the spectral analysis of discretely sampled time series (Priestley, 1981), does not appear to bias the spectra we present below. This conclusion was reached on the basis of an exploratory analysis in which data sampled at a higher temporal resolution did not display any qualitative changes in the spectral composition. The observation that none of the spectra derived at the chosen sampling density have any peaks in the high-frequency range support this conclusion. Repeatability In addition to the high-resolution time series, data were collected from the remaining three replicate video records for each treatment at the same 8-s resolution for 20-tnin intervals to determine whether patterns observed in the high-resolution analysis were repeatable. These 20-min in- tervals were chosen haphazardly, with the added criterion that the images of the polyp were of good quality with respect to contrast, brightness, and definition of the outline against the background. In both treatments, power spectra and amplitudes of length and volume oscillations were calculated for at least 1 (and up to 5) 20-min intervals in both the pre- and post-export phases of digestion for each replicate polyp. From each power spectrum, we identified the frequency of each peak and calculated its signal-to-noise ratio as the ratio of peak height to the maximum height of noise in the plot (i.e.. the highest point remaining on the landscape after excluding the set of peaks). We report the frequency and the signal-to-noise ratio for only those frequencies > 3.3 and 10 mHz (millihertz). The lower frequency limit is set by the duration of the record, and the higher limit is set to avoid reporting bends and contractions as frequency signals. The upper limit of a 10-mHz frequency also excludes harmonics (if present) at twice and three times the principal frequency. To estimate amplitude, length or volume measurements from the same 20-min window were subdivided into four segments of 5 min each (i.e., approximately 2 oscillation cycles). Each of these shorter segments was viewed as an unordered sample of length or volume measurements. The difference between length (volume) at the 97.5 percent quantile and length (volume) at the 2.5 percent quantile in a segment was used as an estimate of amplitude. The mean amplitude of the four 5-min segments was used as the amplitude of length (volume) for the 20-min interval. This measure not only reflects short-term polyp oscillations, it also effectively excludes trends in length (volume) within the window and the effect of short polyp contractions. Sliape variation We do not present extended time series of the width of each polyp cross-section. Rather, to characterize changes in polyp shape, 25 widths along the body column of the polyp were measured every 8 s for two 1 5-min intervals for each treatment. The first interval was taken at 90 min post- feeding in both treatments and the second at about 180 and 700 min post-feeding in the coupled and isolated treatments, respectively. The coefficient of variation of each of the widths was calculated and plotted against the position of that width along the longitudinal polyp axis. Stolon observations A series of observations were made on stolon behavior in an attempt to correlate features of the polyp time series with events observed within the stolons. The rationale for doing so is that the mouth of the polyp remains closed over the time periods analyzed here, hence any changes in polyp volume (exclusive of experimental error) must represent exchange with the stolon. Three replicate records were made of both isolated and coupled treatments, established as described above, but with the focal plane established adja- cent to the polyp-stolon junction at 400X. We do not present detailed time series for stolonal oscillations here (see Buss and Vaisnys, 1993); rather, we use these video- tapes to describe events observed in stolons at times corre- sponding to principal features in the polyp record. These descriptions are supplemented with limited mea- surements. From the videotapes, we measured the onset of oscillations in stolon diameter after feeding and the time at which the polyp initiated export of particulates from its digestive cavity into the stolon. In addition, from frame- grabbed images we measured lumen diameters, amplitudes and frequency of stolon oscillations for three consecutive cycles preceding and following export. Measures were ob- GASTROVASCULAR CIRCULATION tained before and alter feeding, but prior to export, at 30 inin post-feeding in the stolons of coupled polyps and at 30 and 150 min post-feeding in the stolons of isolated polyps. Post-export measures were obtained at 150 min post-feed- ing in the coupled treatment and at 1 h after export in the isolated treatment. To facilitate comparisons between sto- lons with different diameters (Blackstone and Buss, 1992). stolon measurements were standardized to measures of the periderm-to-periderm width. Results Pre-feeding behavior and return Jo pre-feeding conditions Prior to feeding, polyps behaved similarly in both iso- lated and coupled treatments. A representative record ap- pears in Figure 2A, showing that polyp length and volume remained constant prior to feeding, exclusive of occasional volume-conserving contractions in length (e.g., at t = 22 min). Feeding was terminated by regurgitation of undi- gested materials, which occurred at 18-22.5 h post-feeding in both treatments. After regurgitation, the polyp regained near-original values of length and volume (Fig. 2B; see also min 1350 in Fig. 6B). Contraction pulses and asymmetric polyp bends occurred frequently before and after regurgita- tion. Polyps connected to a colony The raw time series for both polyp length and volume is shown in Figure 3. The presentation of the time series is simplified by treatment of three different intervals: 0-15 minutes post-feeding. After ingesting the brine shrimp, the polyp contracted from about 950 to 400 /urn in length, and its volume increased from about 7 to 18 nl (cf. Figs. 2A and 3). In the first 15 min following feeding, the polyp displayed a trend of increasing length and decreasing volume, but otherwise lacked regular behavior. At 8 min post-feeding, volume decreased sharply, from about 18 to 14 nl. coinciding with the repositioning of the brine shrimp nauplius within the digestive cavity. Coincident with the rapid decrease in volume was an apparent stabilization of volume (at ca. 12-14 nil and length (at ca. 550 /am). Regular oscillations in both polyp length and volume began shortly after this repositioning. 15-125 minutes post feeding. The oscillatory behavior, as well as the plateau in length and volume established by 15 min post-feeding, was retained until about 100 min post- feeding. Throughout this period of oscillation, changes in polyp shape were largely restricted to variation in the width of the hypostomal region of the polyp (Fig. 4A). At 100 min, regular oscillations were interrupted as volume under- went another substantial decrease, dropping from about 14 nl to a minimum of about 5 nl at 1 10-120 min, only to increase to 8 nl at 125 min. Prior to the volume decrease at 100 min, the polyp was opaque. In the interval between 95 and 105 min, the polyp became increasingly transparent, indicating an export of contents of the digestive cavity. Since no material was seen leaving the mouth, the exchange must have occurred be- tween the polyp and the colony gastrovascular system. 125-436 minutes post-feeding. The rapid decline in vol- ume immediately preceding this interval was followed by a change in length dynamics and a return to regular oscilla- tory dynamics in volume. Length, which had reached a plateau at 550 /u,m. began to increase to a new plateau at 750 ;nm. The volume oscillations commencing at 125 min were of substantially greater amplitude than those which pre- ceded it, with changes in volume often exceeding 100% with each cycle. These pronounced oscillations were ac- companied by a change in the regions of the polyp showing the greatest variation in width. In contrast to the preceding interval, during which most shape change was restricted to the hypostome, the large-amplitude oscillations characteriz- ing this time interval displayed the largest coefficient of variation in the mid-gastric region of the body column (Fig. 4B). Large-amplitude volume oscillations continued until about 225 min, after which they gradually declined in am- plitude from up to 8 nl at the beginning of the interval to less than 2 nl at the end of the record. Polyp length over the interval spanning 170-200 min showed a gradual lengthen- ing trend from 750 to 900 /nm, after which the plateau at 900 ^im in length persisted for several hours. Both contraction pulses and polyp bending became increasingly common as the amplitude of volume contractions attenuated. Oscillations in both length and volume of the polyp throughout the feeding cycle were characterized by a single dominant frequency component (Fig. 5). The dominant component, however, shifted from an initial dominant fre- quency of about 8 mHz (corresponding to 1 cycle every 2 min) to 6 mHz at 125 min. This shift to the lower frequency between 125 and about 200 min coincided with the large- volume oscillations of the polyp (cf. Fig. 3). Weak harmon- ics of two and three times the dominant frequency were also present in the length spectrum during the first 100 min. Subharmonics were not evident. Because most of the vol- ume spectrum was dominated by the massive oscillations between 125 and 200 min, oscillations before and after this period, although present, left only faint traces in the spec- trum of Figure 5. No appreciable contributions to either length or volume oscillations came from frequencies greater than 24 mH/. The frequency and amplitude of oscillations over the course of a feeding cycle were repeatable with respect to both length and volume among replicates of the coupled polyp treatment (Table I). For all replicates, a single fre- quency predominated both before and after the export of the contents of the gastric cavity. Moreover, the predominant S. DUDGEON ET AL. a a> a "o I. 01 a c. ^ "o . noo- - 24 A. lOftO - ^Y^ ..^-'vv*-'-' 1 " ^SW^* 930- /" r - 20 OAA / Volume T ^1 800 700 - / Length + Bend - 16 600 - - 12 500 - 400- ..>". ... '"'''/'/'"> .;, .', . /v' 1 "VvVX, tf*fc - 8 300- - >nn - -H- + - 4 -35 -30 -25 -20 -15 -10 -5 Time (minutes prior to feeding) 1200 - - 24 B. 1100 - ___ ^_ ~^~ ~^j-^ 1000 - I ^ r^^ - 20 900- ^ R 1 / 800 - / - 16 700 - 600- - 12 500 - 'itf*A ,.rV\-4^ , :-,,.i v . -M ^j%.| . >.$*$& M 400 - i ' .' \; > ' - 8 :> y 300 - + 200 10 90 1100 1110 1120 . i . . 1130 11 40 Time (minutes after feeding) | "o > a "o a. a | "o a "o c- Figure 2. Time series of length (solid line) and volume (dashed line) dynamics of a polyp coupled to a colony by one stolon (A) prior to feeding and (B) at the end of digestion showing the return to initial conditions. R denotes regurgitation. Tickmarks along the abcissa represent spurious data points (see text for further discussion) corresponding to rapid bends drawing the polyp outside the focal plane. Large-amplitude variations in length that are not accompanied by a tickmark represent contraction pulses and are not spurious. frequency shifted from the pre-export level (range of means among replicate polyps: for length, 8.3-9.1 mHz; for vol- ume, 7.9-9.1 mHz) to a lower post-export level (range of means: length, 7.1-7.5 mHz; volume, 6.6-7.5 mHz) in all replicate polyps. Polyps isolated from a colony The raw time series for both polyp length and volume (Fig. 6) is conveniently summarized in three intervals: 0-75 minutes post-feeding. As in the case of the coupled GASTROVASCULAR CIRCULATION 1100 1000 24 20 16 12 e o o o &. 200 40 80 120 160 200 240 280 320 360 400 440 Time (minutes after feeding) Figure 3. Time series of length (solid line) and volume (dashed line) dynamics of a polyp coupled to a colony by one stolon from to 436. 1 5 min after ingestion of a single brine shnmp nauplius. Tickmarks along the abcissa represent spurious data points (see text for further discussion) corresponding to rapid bends drawing the polyp outside the focal plane. Large-amplitude variations in length that are not accompanied by a tickmark represent contraction pulses and are not spurious. polyp, feeding resulted in a contraction in polyp length and an increase in volume (not shown). A detailed record of the onset of oscillations immediately after ingestion from a replicate isolated polyp record is shown in Figure 7. As in the coupled treatment (Fig. 3, min 0-20), behavior was irregular for a short time after feeding, and regular oscilla- tions began about 10-15 min post-feeding. 15-763 minutes post-feeding. Like the coupled treatment, the isolated polyp increased in length and volume after feeding (Fig. 6A). As in the coupled case, volume reached a plateau about 40 min. after feeding, whereas the plateau in length was not attained until roughly 150 min post-feeding. The variation in polyp shape 90 min after feeding mirrors that characterizing the early post-feeding period in the cou- pled treatment, with the greatest variation in polyp width occurring in the hypostome (Fig. 4C). In the isolated case, however, shape also varied substantially at the base of the polyp. Figure 8 compares widths at the polyp base during the period of polyp lengthening (Fig. 8A) with that occur- ring after the plateau in length has been reached at 150 minutes post-feeding (Fig. 8B). The latter record reveals that the polyp base alternated every other length cycle in maximal and minimal width, which made every other length cycle of greater amplitude (Fig. 8B). Next, for about 9 h the polyp displayed a trend of grad- ually increasing length and gradually decreasing volume (Fig. 6A). Regular oscillations in length and volume con- tinued for the entire period. This record is in marked con- trast to the coupled treatment, where similar behavior ter- minated abruptly at about 1 10 min post-feeding with a large decrease in volume and the onset of large-amplitude volume oscillations. Neither rapid declines in volume nor large- amplitude volume oscillations comparable to those in the coupled case were observed in the isolated case. 763-1474 minutes post-feeding. 763 minutes after feed- ing, the polyp underwent a 7-min interval of rapid, repeated length contractions, accompanied by variation in volume dynamics (Fig. 6B). From 770 to 790 minutes post-feeding, both length and volume regained values comparable to those that preceded the event at 763 min. The digestive cavity of the polyp had previously been largely opaque: after the event, regions of the cavity became increasingly transparent. This change coincided with a marked alteration in the shape of the polyp as it oscillated. In Figure 9 these S. DUDGEON ET AL. A. Couplrd polyp " minutes after feeding C. Isolated polyp 90 minutes after feeding 1.00 0.80 0.60 '* 20 0.00 5 10 15 20 25 30 35 B. Coupled polyp appro*. 180 minutes post-feeding 35 30 25 20 15 10 5 D. Isolated polyp approi. 700 minutes post-feeding o 0.80 0.60 0.20 0.00 1.00 0.60 0.40 0.20 2, 1 o 5 10 15 20 25 30 35 35 30 25 20 15 10 5 Coefficient of variation Figure 4. Coefficients of variation from samples of width measures taken every 8 s over a l.vmin timecourse at each of 25 loci along the longitudinal axis of a polyp (polyp-stolon junction at 0.00, mouth of polyp at 1.00) for (A) a coupled polyp 90 min after feeding, (B) a coupled polyp 180 min after feeding, (C) an isolated polyp 90 min after feeding, and (D) an isolated polyp 700 min after feeding. polyp shapes are superimposed on the corresponding record of length and basal width. With every other length maxima the polyp alternated between maintaining a bolus of fluid in the center of the digestive cavity and maintaining two such boli one in the subtentacular region and one in the basal body column separated by a constriction in the polyp's center. The same pattern of oscillation is evident in Figure 4D. Variation was maximal in the basal and subtentacular regions of the body column during this interval. The polyp continued to display regular oscillations in length and volume for an additional 10 h after the event at 763 min (Fig. 6B). Just as in the coupled polyp, contraction pulses and polyp bending became increasingly common in later stages of the record. At 1338-1343 min the polyp regurgitated undigested materials through the mouth and regained a length and volume comparable to those observed prior to feeding (Fig. 6B). Regurgitation in the isolated treatment differed from that in the coupled case (Fig. 2B). In the isolated case, the return to initial conditions was far more abrupt. The isolated polyp showed a dominant, and remarkably stable. 8-mHz oscillation frequency and a weaker harmonic at 16 mHz that persisted throughout the feeding cycle (Fig. 10). A subharmonic of ~ 4 mHz (i.e., corresponding to events that occurred every other cycle) emerged nearly 150 min after feeding, at the point when basal widths began alternating every other cycle. This subharmonic coexisted with the dominant frequency for the duration of the record. In the period following the event at 763 min (while the polyp displayed movement of fluids between the subten- tacular and basal regions of the body column with every other cycle; Fig. 9), the 4-mHz frequency became dominant for 200 min. No appreciable contributions to either length and volume spectra came from frequencies greater than 24 mHz. The principal features of behavior in isolated polyps were repeatable among replicates (Table I). The principal oscil- lation frequency varied among replicates (range; 6.7 to 9.0 mHz) but, unlike the frequency in the coupled-polyp treat- ment, did not shift consistently downward during the post- feeding period. Also, subharmonic frequencies were de- tected only in the isolated polyps. Finally, the amplitude of GASTROVASCULAR CIRCULATION a) Sliding Window Spectrum (Length) 350 300 f 250 - 200 u I 150 100 50 10 20 30 40 50 60 Frequency (mHz) b) Sliding Window Spectrum (Volume) 350 300 1 250 - 200 0) I 150 100 50 10 20 30 40 50 60 Frequency (mHz) Figure 5. Contour plot, based on the time-series data in Figure 3. of sliding window spectral analysis of (a) length and (b) volume of the coupled polyp. Abcissa represents frequency of oscillation, ordinate rep- resents the sliding window of time (in minutes), and contour lines represent the height of peaks (coming out of the page) that signify the relative importance of a given frequency underlying the cyclic behavior. volume oscillations was consistently much lower than the larger amplitude volume oscillations characteristic of post- export coupled treatments. (Fig. 3; Table I). Stolon observations Feeding was associated with a closing of the polyp- stolonal junction and a momentary cessation of gastrovas- cular flow in colonies that had been experiencing fluid movement prior to feeding. As the polyp lengthened after the initial contraction, the polyp-stolon junction reopened. The stolon lumen began to oscillate in diameter at 13 and 8 miii (on average) in the coupled and isolated treatments respectively: export into the stolon accompanied polyp con- traction and import into the polyp accompanied polyp lengthening. The interval between feeding and the onset of stolonul oscillations did not significantly differ from the interval between feeding and the onset of regular oscilla- tions in polyp length and volume (Student's t test; t = 0.27, df = 8, P = 0.80). Stolon observations made after this initial period are most conveniently treated separately for the two experimental treatments. Pol\p connected to colonv. The frequency of stolon con- tractions in the first hour after feeding did not differ signif- icantly from the frequency of polyp oscillations (Table II. t = 0.23, df = 5. P = 0.83). Neither was a significant difference detected in the average lumen diameter of stolons before feeding and in the first hour after feeding (Table II; Student's t test; lumen diameter, / == 1.19, df = 4. P = 0.30). However, after feeding, stolons contracted signifi- cantly more frequently and with greater amplitude than they did before feeding (amplitude, t = 3.13, df = 4. P = 0.03; frequency, t == 17.43, df == 4, P < 0.01). Despite the observed export of fluid from the polyp into the stolon, export of paniculate matter was only rarely observed in the stolon at this time; on the few occasions when particles were evident, they were few and did not exceed 2 /urn in diam- eter. The period of interruption in polyp oscillations and de- crease in volume, seen at 100 min in the coupled-polyp record, was correlated with events observed on average at 79 min in replicate stolon records. In each stolon replicate. the stolon became greatly expanded with fluid imported from the colony, and the contents of the fed polyp were observed leaving the polyp in a dense stream of large particles (up to 100 /urn in length and 15 /urn in diameter). Stolonal oscillations changed markedly in the period after export: they were smaller in amplitude and lower in fre- quency than before export, and the average lumen diameter of the stolens was larger (Table II; lumen diameter t = 2.88, df = 4, P = 0.04; amplitude, t = 1 1.46, df = 4, P < 0.01; frequency. / == 6.11. df == 4, P < 0.01). Notably, the frequency of stolon oscillations following export mirrored the shift in the frequency of polyp oscillations. Before export, both polyps and stolons oscillated at a frequency of ~8 mHz, whereas after export both oscillated at a fre- quency of -6 mHz the same frequency as in stolons prior to feeding. Polyp isolated from colony. As in the coupled case, stolon oscillations did not significantly differ in frequency from those observed in polyps (t = 0.44, df = 5. P = 0.68). Similarly, stolon oscillations prior to export were more frequent and of greater amplitude than stolon oscillations 10 S. DUDGEON ET AL. Table I Repeatability of length and w/,/ < dynamics fur coupled and isolated polyps based on ohsen'ing three replicate polyps in each treatment: values represent means standard erro . (in parentheses) Polyp Length Polyp Volume Pre- or Principal Principal Post- Frequency Signal : Subharmonic Signal : Amplitude Frequency Signal : Subharmomc Signal : Export (mHz) Noise ratio (mHz) Noise ratio (nl) (mHz) Noise ratio (mHz) Noise ratio Couplet! Pre Post polyps 84.42 (3.71) 92.43(18.33) 8.57(0.27) 7.27(0.13) 3.69(1.13) 2.66(0.70) 4.36(1.98) 6.19(1.59) 8.30(0.41) 7.06(0.25) 2.80(0.69) 3.31 (0.55) Isolated polyps Pre 93.56 (3.95) 7.89(0.67) 5.36(1.04) 4.14* 2.60* 1.54(0.08) 7.84(0.51) 2.08(0.12) 3.95t (0.1 6) 2.28(0.12) Post 65.11 (8.43) 7.50(0.42) 2.11 (0.31) 5.14(0.77) 2.30(0.13) 1.84(0.24) 7.93 (0.64) 2.40(0.86) 3.32(0.01) 2.08(0.33) Pre- and post-export phases of coupled polyps were distinguished visually on the basis of shape variation of the polyp body column while it was oscillating (see Fig. 4a,b); those of isolated polyps were inferred from the coexistence of strong subharmonic frequencies in length or volume and time of occurrence (see text and Fig. 9). Signal-to-noise ratio represents a ratio of a peak height of a frequency to the maximum height of noise in the spectrum. * Values lacking standard errors indicate that only one replicate displayed a subharmonic. t Value was based on two replicates instead of three. prior to feeding, but no difference in average lumen diam- eter was detected (Table II; lumen diameter / = 0.42, df = 4, P = 0.69; amplitude, t == 3.26, df == 4. P = 0.03; frequency, t = 5.10, df = 4, P < 0.01 ). Unlike the coupled treatment, however, these oscillations displayed similar av- erage lumen diameters and similar amplitudes and frequen- cies for 10+ hours post-feeding (Table II). Continuous observation of stolons showed that the isolated polyp, de- spite the continued fluid exchange with the stolon, exported little paniculate matter into the stolon over this extended interval. Three replicate isolated stolon films showed export of a dense stream of large particulates at an average time of 807 min (13.5 h) after feeding, an interval comparable to the event observed in the high-resolution isolated polyp record at 763 min (12.7 h). The average lumen diameter of the stolon after export was not significantly different from that before export, but the oscillations were of smaller amplitude (i.e., stolons remained expanded and filled with particulates throughout oscillation cycles; lumen diameter t = 1.14. df = 4. P = 0.32; amplitude, t = 8.40. df = 4, P < 0.01 ). After export, the frequency of oscillation of the stolon was comparable to that prior to feeding, as well as to that of the stolon in the coupled polyp case after export; but it was distinct from the isolated polyp signal pre- and post-export (Table II). Discussion Although it has long been known from anecdotal ac- counts that hydrozoan polyps undergo periodic changes in shape after ingesting a food item, the results presented above represent, to our knowledge, the first quantitative treatment of this behavior. These data reveal three distinct phases of post-feeding behavior. Phase 1 corresponds to the onset of oscillatory behavior immediately following inges- tion. phase 2 to the subsequent period during which the polyp oscillates with limited exchange of particulates with the gastrovascular system, and phase 3 to the interval initi- ated by the export of particulates from the polyp into the gastrovascular system and terminated by regurgitation. In what follows, we elaborate details of each phase, contrast differences among coupled and isolated polyps, and hypoth- esize physiological mechanisms for the transitions between behaviors. Phase 1. In both isolated and coupled treatments, inges- tion does not immediately elicit oscillations by the polyp (Figs. 3, 7). Rather, oscillations begin 5-15 min after in- gestion and are characterized by a gradual increase in am- plitude up to a value which thereafter (phase 2) remains constant. These findings bear on the mechanism that trig- gers the oscillatory behavior. One obvious candidate for such a trigger is the change in the internal dimensions of the polyp, as might be sensed by stress or strain receptors. Alternatively, the polyp might sense the presence of nutri- ents or their correlates (e.g., the liter of digestive enzymes). If the former were the case, one would expect polyps to oscillate immediately following ingestion, whereas the lat- ter would imply that oscillation would be delayed by the length of time required for digestion to release nutrients and digestive enzymes. Moreover, if the polyp does respond to some product of digestion, one might expect the titer of such a product to increase gradually as digestion proceeds and, as seen in Figures 3 and 6A for length, the oscillations to increase in amplitude as the titer increases, up to some threshold set by the size of the polyp. Phase 2. At the end of phase 1 , oscillations are constant in amplitude and their frequency does not differ between GASTROVASCULAR CIRCULATION II isolated and coupled treatments (Table I). During phase 2, polyps of both treatments display comparable patterns of shape variation (Fig. 4A, C) and similar frequencies and relative amplitudes of stolon oscillations (Table II). Al- though the polyp exchanges fluid with the stolon throughout this interval, paniculate exchange is only rarely observed in either treatment. The similarity of the coupled and isolated polyp records during this period suggests that coupled pol- yps are behaving as autonomous elements. The treatments, however, differ markedly in the duration of phase 2. Isolated polyps retain this behavior for 13 h (Fig. 6, Table II). whereas coupled polyps undergo an abrupt transition to phase 3 at 1 .5-2 h (Fig. 3, Table II). The difference in duration between the two treatments may reflect a simple mechanical limit: a minimum pressure dif- ferential between the polyp and the hydrorhi/,a may be required to move large particulates through the polyp-stolon junction. If so, the long duration of phase 2 in the isolated polyp may reflect the time required to solubilize food items to the extent that pressure differentials generated by a single polyp are sufficient to drive particulates through the junc- tion. Conversely, the short duration of phase 2 in the cou- pled case may reflect the far greater pressure differentials associated with colonywide behavior (see below). Beginning in phase 2. isolated polyps display a trend of gradually decreasing volume (e.g.. Fig. 6, ca. 0.3 nl/h) that continues until regurgitation. This gradual decrease in vol- ume likely reflects the transfer to endodermal cells that is associated with digestion (Schierwater et al., 1992): the increase in length of the stolon lumen that is associated with tip growth; and perhaps, leakage. Estimating the extent of the latter will require monitoring of endodermal cell volume and stolon length, which we have not attempted here. Phase 3. Phase 3 is initiated by the export of dense streams of particulate material from the polyp into the stolon. Phase 3 differs between the isolated and coupled cases, as might be expected from the differences between the gastrovascular systems to which the polyps are export- ing. In the coupled case, fluid is exchanged with an entire colony. The volume of the colonial gastrovascular system is very large relative to that of the fed polyp, and the colony possesses many other polyps which themselves oscillate to drive large fluid volumes to effect exchanges with the fed polyp. In contrast, the isolated polyp is exporting to a gastrovascular system that lacks other polyps and whose total volume is but a fraction of its own. These differences are interpreted to underlie both the differences and the commonalities in phase 3 between the isolated and coupled treatments. In the coupled case, phase 3 is marked by an abrupt decline in polyp volume (Fig. 3). In the stolon records this decline is correlated with a large import of fluid from the colony and subsequent export of the contents of the fed polyp into the colonial gastrovascular system (Table II). Export is followed by the onset of regular high-amplitude oscillations in volume that have a frequency distinct from that displayed in phase 2 (Fig. 5) and identical to that of the stolon oscillations in phase 3 (Table II). The repeated ap- pearance of these features in all coupled records (Table I), their absence in all isolated records (Table I), and the temporal correlation of these events with distinctive signa- tures in stolon records (Table II) lead us to interpret phase 3 in the coupled case as a colonywide exchange distinct from the autonomous behavior displayed during phase 2. Phase 3 is also accompanied by changes in the way that polyp shape varies when oscillating (Fig. 4), and this tran- sition is likewise attributable to the large-volume fluxes associated with exchange of particulates with the colony. Prior to export, the prey fills the gastric cavity; hence, most of the variation in volume is attributable to changes in the dimensions of the hypostome. The gastric cavity is emptied during export to the stolon: thereafter, its shape is presum- ably limited only by its own extensibility. Finally, with multiple polyps contributing to volume exchange with the fed polyp, the frequency of oscillation may be expected to be determined not by the polyp's autonomous rhythm, but by a frequency characteristic of fluxes through the stolon (Table II). In the isolated case, phase 3 is similarly marked by the export of particulate matter from the polyp into the stolon, but without the large volume fluxes observed in the coupled case. Since the volume of the gastrovascular system in the isolated case is only a fraction of that of the polyp, the absence of large-volume fluxes and associated variations in polyp shape to accommodate such fluxes is expected. An- other notable difference between isolated and coupled cases is that the oscillation frequency of isolated polyps does not change during phase 3 to the frequency characteristic of the stolon (Fig. 10). as it does in the coupled case (Fig. 5). Phase 3 in the isolated case retains the same two frequen- cies. ~4 and ~8 mHz, that were established at the end of phase 1 and retained throughout phase 2, although the predominant frequency shifts to the subharmonic in phase 3 (Fig. 10. Table I). These differences are likewise interpreted as a consequence of the differences in volume fluxes be- tween the two cases. In the coupled case, the principal force driving fluid movement is the activity of the colony com- municating with the fed polyp through the stolon. In the isolated case, the fed polyp remains the principal driving force in exchange with the stolon. The shift between the subharmonic and principal frequencies likely derives from similar considerations. Prior to export, the gastric cavity is rich in particulates, which constrains variation in shape; following export this constraint is released and shape vari- ation associated with the 4-mHz phase 3 signal predomi- nates (Figs. 4D. 9; Table I). Finally, the fact that the stolon in the isolated case oscillates at a frequency different from that of the isolated polyp and identical to that of the post- 12 S. DUDGEON ET AL. 1100 1000 9W 7001 600 500 400 300 200 '^ 40 80 JS D. 5* 1100 a. 1000 900 800 700 600 500 400 300 200 120 160 200 240 280 320 B. 360 760 800 840 880 920 960 1000 1040 1080 Time Figure 6. Time series of length (solid line) and volume (dashed line) dynamics of an isolated polyp after ingestion of a single brine shrimp nauplius from (A) to 740 min and (B) 740 to 1474.2 min. R denotes regurgitation. Tickmarks along the abcissa represent spurious data points (see text for further discussion) corresponding to rapid bends drawing the polyp outside the focal plane. Large-amplitude variations in length that are not accompanied by a tickmark represent contraction pulses and are not spurious. GASTROVASCULAR CIRCULATION 13 r ; ' '$!$'ij! i ni ' '''IfjEj : ' : 16 14 12 10 8 400 440 480 520 560 600 640 680 r + ++-*- -HH+- -H-4- -H- HH- + 4- 720 E 3 I 18 16 14 12 10 8 6 4 o 0. 1120 1160 1200 1240 1280 1320 1360 1400 1440 1480 (minutes after feeding) Figure 6. (Continued) 14 S. DUDGEON ET AL. _i a 700 650 - 600 550 - 500 450 - 400 350 - 300 8 12 Time (minutes after feeding) 16 20 Figure 7. Time series, in minutes after ingestion. of polyp length showing the commencement of regular oscillations of an isolated polyp. - Length export coupled case strongly suggests the existence of hy- drorhizal-specific dynamics (Table II). Indeed, previous ob- servations have established that isolated stolons (i.e., stolons without polyps) in Hydractinia symbiolongicarpus exhibit endogenous oscillatory dynamics (Buss and Vais- nys, 1993). In both treatments, the contraction pulses and rapid polyp bends become increasingly frequent during phase 3 (Figs. 3, 6), consistent with known suppression of contraction pulses during digestion (Passano and McCullough, 1962. 1964; Josephson and Mackie, 1965; Shibley, 1969; Stokes, 1974). The interaction between the contraction pulse system and the digestive oscillations we characterize here bears further attention. Winfree (1970) has shown that a key feature of certain nonlinear oscillators is the capacity for the oscilla- tion to be terminated by perturbations occurring at specific 925 : 900 - 875 - 850 ' 825 - 800 - 775 - - Length 250 - 210 ? - 170 ~ 130 1 n a 90 121 2121 21212 50 85 850 855 860 865 870 875 Time (minutes after feeding) - Length 230 235 240 245 250 Time (minutes after feeding) 255 Figure 8. Development of the subharmonic associated with basal width alternating every other length cycle in the isolated polyp record. (A) Polyp length (filled circles) and width (open circles) at the basal-most section for 12 cycles from a representative period (60 to 86 min) when the polyp is lengthening from to 150 min post-feeding. (B) Polyp length (filled circles) and width (open circles) at the basal-most section for 12 cycles from a representative period (230 to 255 min) > 150 min following the development of the 4-mH/ subharmonic. B. 2 Figure 9. Polyp oscillations during the dominant period by the 4-mHz subharmonic' following the 763-min event in the isolated polyp. (A) The length (filled circles) and basal-most width (open circles) of the isolated polyp for 1 2 cycles from 85 1 to 878 min after feeding that is representative of the period when the subharmonic dominates the power spectrum. (B) Schematic illustrations of the pattern of shape change of the bolus of transparent fluid in the digestive cavity during three consecutive length maxima of the polyp. Numbers above the polyp correspond to those along the abcissa in (A) that indicate the appearance of the polyp at the maximal polyp length of each cycle. Schematics redrawn from frame-grabbed im- ages during this interval. Arrows within boli inside polyp gastric cavity in (B) indicate the direction of movement of the bolus during the subsequent oscillation cycle. GASTROVASCULAR CIRCULATION 15 a) Sliding Window Spectrum (Length) c) Sliding Window Spectrum (Volume) 525 450 1 375 - 300 u | 225 150 75 10 20 30 40 50 Frequency (mHz) 10 20 30 40 50 Frequency (mHz) 60 b) Sliding Window Spectrum (Length) d) Sliding Window Spectrum (Volume) 1125 1050 975 900 825 750 675 600 1125 1050 975 900 825 750 675 600 10 20 30 40 50 Frequency (mHz) 60 10 20 30 40 50 Frequency (mHz) 60 Figure 10. Contour plot, based on the time-series data presented in Figure 6, of sliding window spectral analysis of the isolated polyp for length from (a) 1 to 600 min. and (b) 601 to 1200 min, and for volume from (c) 1 to 600 min, and (d) 601 to 1200 min after ingestion. Abcissa represents frequency of oscillation, ordinate represents the sliding window of time (in minutes), and contour lines represent the height of peaks (coming out of the page) that signify the relative importance of a given frequency underlying the cyclic behavior. phase relationships. Taddei-Ferretti and Cordelia (1976) have shown that contraction pulses can be experimentally annihilated in the predicted fashion. It is conceivable that feeding annihilates contraction pulses and, similarly, that the contraction pulses which reappear in phase 3 annihilate digestive oscillations by this mechanism. Phase 3 is terminated at comparable times in both the isolated and coupled treatments by regurgitation through the mouth, followed by a return of length and volume to values characteristic of pre-feeding conditions. The dynamics of regurgitation differ between isolated and coupled polyps (c/ Figs. 2B and 6B). Whereas regurgitation in the coupled case is not a conspicuous feature of either the length or volume record, regurgitation in isolated polyps is abrupt. The iso- lated polyp exhibits several rapid, large contractions in length and an associated decline to half of its previous volume. Coupled polyps regurgitate less material that ap- pears more finely paniculate, whereas the contractions of isolated polyps are associated with the export of larger pieces of undigested debris. We attribute the differences between treatments in the behavior of polyps and the nature of the material regurgitated to the vastly larger colonial gastrovascular system with its many additional polyps from which undigested particles could be regurgitated. 16 S. DUDGEON ET AL. Table II Characteristics of oscillations of stolons in the isolated and coupled polyp treatments: values represent the mean and standard errors (in parentheses) of three replicates in each treut/ticnl Isolated Polyp Coupled Polyp Onset of Oscillations (min) 8.34 (4.99) 13.19 (11.87) Export to Hydrorhiza (min) Lumen 807.33 (203.95) Frequency Lumen 78.67 (12.50) Frequency Stolon Oscillations Diameter Amplitude (mHz) Diameter Amplitude (mHz) Pre-feeding 0.29 (0.03) 0.07 (0.03) 5.69 (0.29) 0.24 (0.02) 0.11 (0.03) 5.63 (0.13) Pre-export (30 min post-ingestion) 0.32 (0.06) 0.23 (0.04) 8.48 (0.46) 0.20(0.03) 0.19 (0.02) 8.40(0.10) ( 150 min post-ingestion) 0.37 (0.05) 0.25 (0.02) 8.64 (0.43) Post-export (Time of export + 60 minutes) (150 minutes post-ingestion) 0.45 (0.05) 0.05 (0.01) 6.00 (0.2 1 ) 0.32 (0.03) 0.07 (0.01) 6.07 (0.37) Data for onset of oscillations and export to hydrorhiza are presented in minutes after ingestion. Means and standard errors for lumen diameter, amplitude, and frequency estimated from the sample of means of each replicate determined from measures taken over three consecutive cycles. Lumen diameter and amplitude are dimensionless indices calculated using the following formulas: Lumen diameter = (max + min lumen diameter)/(2 x periderm diameter) Amplitude = max - min lumen diameter/periderm diameter A theoretical model. These findings suggest a simple conceptualization of the feeding response of an isolated hydrozoan polyp. The principal phases of behavior reflect differing input-output relationships between the polyp and either the external (via the mouth) or internal (via the polyp-stolon junction) environment. Inputs in the form of food items elicit oscillatory behavior (phase 1 ), which leads to the output of fluid, but few particulates, to the stolonal system (phase 2). The final phase begins with the export into the stolon of particulates (phase 3) and is terminated by export of undigested material (regurgitation) from the mouth and subsequent return to initial conditions. These data lead us to suggest that an input of food releases elicitors (e.g., nutrients or digestive enzymes) whose action triggers an underlying biochemical system (e.g.. ion potentials at neuromuscular junctions), and that the oscillation of that system is reflected in corresponding oscillations in length and width. An ordinary differential equation model of a nonlinear oscillator that can undergo a supercritical Hopt bifurcation as the concentration of an elicitor is increased has been shown to be capable of reproducing principal features of the isolated polyp data presented here (Wag- ner el ai. 1998). Elaboration of such a single-polyp model is easily imagined based on the hypothesis that elicitors are circulated in gastrovascular fluids during phase 2 and that when these elicitors reach a threshold value they trigger oscillations of adjacent polyps, thereby generating the behavior described above for a coupled polyp. These considerations suggest that a spatially dis- tributed system of coupled nonlinear oscillators is a rea- sonable abstraction of the gastrovascular system of a hydrozoan colony. Acknowledgments We thank Dr. Arvydas Matiukas for his development of early versions of the algorithms for this project. Dr. Harald Freund for advice in developing the volume analysis algo- rithm, and Jim Bonacum for technical assistance. We thank Neil Blackstone and an anonymous reviewer for improving the quality of this manuscript. This research was supported by the National Research Council Twinning Program and the National Science Foundation (OCE-93-15082) to Leo Buss. Literature Cited Berrill, N. J. 1953. Growth and form in gymnoblastic hydroids. VI. Polymorphism within the Hydractiniidae. J. Morphology 92: 241-272. Bevan, J. A., G. Kaley, and G. M. Rubanyi (eds.) 1995. Flow-Depen- dent Regulation of Vascular Function. Oxford Univ. Press, Oxford, UK. Blackstone, N. W. 1997. A dose-response relationship for experimental heterochrony in a colonial hydroid. Biol. Bull. 193: 47-61. Blackstone, N. W. 1998. Physiological and metabolic aspects of exper- imental heterochrony in colonial hydroids. J. Evol. Biol. 11 (in press). Blackstone, N. W., and L. W. Buss. 1992. Treatment with 2,4-dmitro- phenol mimics ontogenetic and phylogenetic changes in a hydractiniid hydroid. Proc. Nail. Acad. Sci. USA 89: 4057-4061. Blackstone, N. W., and L. W. Buss. 1993. Experimental heterochrony in hydractiniid hydroids: why mechanisms matter. J. Evol. Biol. 6: 307-327. Buss, L. W., and J. R. Vaisnys. 1993. Temperature stress induces dynamical chaos in a cnidarian gastrovascular system. Proc. Roy. Soc. Land. B 252: 39-41. Buss, L. W., and P. O. Yund. 1989. A sibling species group of Hydrac- tinia in the northeastern United States. J. Mar. Biol. Assoc. UK 69: 857-875. Dudgeon, S. R., and L. W. Buss. 1996. Growing with the flow: on the GASTROVASCULAR CIRCULATION 17 maintenance and malleability of colony form in the hydroid Hvdrac- tinin. Am. Nat. 147: 667-691. Josephson, R. K., and G. (). Mackie. 1965. Multiple pacemakers and the behavior of the hydroid Tubularia. J. v/>. Biol. 43: 293-332. I.uBarbera. M. 1990. Principles of design of fluid transport systems in Zoology. Science 249: 992-1000. Murray. C. D. 1926. The physiological principle of minimum work applied to the angle of branching of arteries. Proc. Nail. At at!. Sci. USA 12: 835-841. Passano. L. M., and C. B. McCullough. 1962. The light response and the rhythmic potential of Hydra. Proc. Null. Acad. Sci. USA 48: 1376-1382. Passano, L. M., and C. B. McCullough, 1964. Coordinating systems and behavior in Hydra. II. The rhythm potential system. J. Exp. Biol. 42: 205-231. Press, \V. H., S. A. Teukolsky, VV. T. Vetterling, and B. P. Flannery. 1992. Numerical Recipes in C. Cambridge University Press. New York. Priestley, M. B. 1981. Spectral Analysis anil Time Series. Academic Press. New York. Schierwater, B., B. Piekos, and L. W. Buss. 1992. Hydroid stolonal contractions mediated by contractile vacuoles. J. Exp. Biol 162: 1-21. Sherman, T. F. 1981. On connecting large vessels to small. J. Gen. Physiol. 78: 431-453. Shibley. G. A. 1969. Gastrodermal contractions correlated with rhythmic potentials and pre-locomotor burst in H\dra. Am. Zoo/. 9: 586. Stokes, D. R. 1974. Physiological studies of conducting systems in the colonial hydroid Hydractinia echinala 1. Polyp specialization. J. Exp. Zool. 190: 1-18. Taddei-Ferretti, C., and L. Cordelia. 1976. Modulation of the Hydra attenuate! rhythmic activity: phase response curve. J. Exp. Biol. 65: 737-751. Wagner. A., S. R. Dudgeon, J. R. Vaisnys, and L. W. Buss. 1998. Non-linear oscillations in polyps of the colonial hydroid Podocoryne carnea. Naturwissenschaften 85: 117-120. Winfree, A. T. 1970. An integrated view of the resetting of a circadian clock. / Theor. Biol. 28: 327-374. Zamir, M. 1977. Shear forces and blood vessel radii in the cardiovas- cular system. J. Gen. Physiol. 69: 449-461. Reference: Biol. Bull. 196: 18-25. (February. 1999) High Contents of Trimethylamine Oxide Correlating With Depth in Deep-Sea Teleost Fishes, Skates, and Decapod Crustaceans ROBERT H. KELLY AND PAUL H. YANCEY* Biology Department. Whitman College, Walla Walla. Washington 99362 Abstract. In muscles of shallow-living marine animals, the osmolyte trimethylamine M-oxide (TMAO) is reportedly found (in millimoles of TMAO per kilogram of tissue wet weight) at 30-90 in shrimp, 5-50 in crabs. 61-181 in skates, and 10-70 in most teleost fish. Recently our labo- ratory reported higher levels (83-21 1 mmol/kg), correlating with habitat depth, in deep-sea gadiform teleosts. We now report the same trend in muscles of other animals, collected off the coast of Oregon from bathyal (1800-2000 m) and abyssal plain (2850 m) sites. TMAO contents (mmol/kg SD) were as follows: zoarcid teleosts, 103 9 (bathyal) and 197 2 (abyssal); scorpaenid teleosts, 32 (shallow) and 141 16 (bathyal); rajid skates. 215 13 (bathyal) and 244 23 (abyssal); caridean shrimp, 76 16 (shallow), 203 35 (bathyal), and 299 28 (abyssal); Chionoecetes crabs, 22 2 (shallow) and 164 15 (bathyal). Deep squid, clams, and anemones also had higher contents than shallow species. Osmoconformers showed compensation between TMAO and other osmolytes. Urea contents (typi- cally 300 mmol/kg in shallow elasmobranchs) in skates were 214 5 (bathyal) and 136 9 (abyssal). Glycine contents in shrimp were 188 17 (shallow) and 52 20 (abyssal). High TMAO contents may reflect diet, reduce osmoregulatory costs, increase buoyancy, or counteract de- stabilization of proteins by pressure. Introduction There are two different adaptive strategies that allow marine organisms to regulate cell volume in the face of the Received 1 September 1998: accepted 24 November 1998. * To whom correspondence should be addressed. E-mail: yancey 1 500 m ) Teleost fishes ophidiiform family Zoarcidae (zoarcid eels) Lycenchelys sp. Paehycara sp. Elasmohranch fishes rajiform family Rajidae (skates) Bathyraja spinosissinui Bathyraja unnamed^ (1280-1829 m) O2500 m) Decapod crustaceans suborder Caridea (shrimp) Pandolus danue Pandalopsis umphi Neocrangon abyssorum Decapod crustaceans decapod suborder Brachyura (crabs) Chionoecetes bairdi Chionoecetes angulatus Cephalopod molluscs families Loliginidae. Gonatidae. Onychoteuthidae (squid) Litligo opalescens^ Benyteiithis magister (inner shelf to (30-1500 m near surface) bottom) Gottiitus borealis (epipelagic-abyssal) Moroteuthis robusla (subtidal-bathyal?) Pelecypod molluscs eulamellibranch family Veneridae, septibranchia family Cuspidaridae Saxidomus Cuspidaria glacialis Cuspidaria glacialis giganteusi Antho/oan cnidaria order Actiniaria Urticiiw lofotensis Actinauge abyssorum Actinauge abyssorum * From Eschmeyer et al. (1983); Pearcy et al. (1982); Allen and Smith (1988); Nesis (1987). t Specimens obtained from Pike's Fishmarket, Seattle. t Close relative of Bathyraja trachura (Eschmeyer et al., 1983). provided fresh specimens of Pachygrapsus crassipes (a brachyuran crab) and Squalus acanthias (the dogfish shark). These two species (not listed in Table I) were used to provide additional data for comparison (see Results). Analytical procedures Samples were frozen at 80C on the ships, transported to Whitman College on dry ice. and stored at 70C. Muscle was processed with perchloric acid, then neutralized and filtered as previously described (Wolff et al., 1989). The TMAO content was determined according to the procedure of Wekell and Barnett (1991), by reducing TMAO to tri- methylamine (TMA) with an iron-EDTA reagent. TMA was extracted in toluene and reacted with 0.02iAO on other proteins at high pressure. The stabilization hypothesis is further supported by the finding of reduced glycine in deep-sea shrimp and urea in skates, along with higher TMAO (Fig. 2). Glycine does not generally have as great a protein-stabilizing capability as TMAO (Yancey, 1994). Similarly, urea is a protein desta- bilizer, and its ratio to TMAO is reversed in abyssal skates compared to most shallow species (Fig. 2). Indeed, to our knowledge the urea values for the abyssal skate are the lowest yet reported for a marine cartilaginous fish. Of course, it is possible that substantial urea was lost from deep skates during their 90- to 120-min trips to the surface. As evidence against this urea loss, skates ranged in size from 10-90 cm, but the standard deviations were quite low (6% in abyssal, 2% in bathyal). The stabilization hypothesis could be valid for all species examined here. Wang and Bolen (1997) have shown that unfavorable interactions between TMAO and peptide back- bones stabilize protein structure, supporting earlier hypoth- eses that such osmolyte effects are universal (Yancey et ai, 1982). The effect on proteins under pressure remains spec- ulative, but we know that destabilizing inorganic ions alter water structure in a manner essentially identical to pressure effects (Leberman and Soper, 1995). Perhaps compensatory osmolytes do the opposite (Gillett et , o O is .Q I 50- 25- Q agranular E3 highly granular ' , [ ] ' - 1 ' ' ' PI n . . ' ' ' . ' 7 5 ^ 1 n n 2.0 3.0 4.0 pH 5.0 6.0 Figure 3. Distribution of pH associated with zymosan engulfed by granular and agranular hemocytes after a 30-min exposure to zymosan and an additional 30-min incubation. Hemocytes were classified as granular or agranular on the basis of the overall granularity of the cytoplasm as observed by light microscopy, and the pH of engulfed fluorescent zymosan was calculated from the ratio of the emission intensities of rhodamine and OG 488. The mean pH was 4.9 0.02 for 297 phagosomes within 83 granular hemocytes. but much lower (mean pH 3.1 0.02: P < 0.05) for 364 phagosomes within S3 agranular hemocytes. ganism and cell type. For instance, the lowest phagosomal pH value reported in mouse macrophages after phagocytosis of yeast was 5.0 (Geisow et al.. 1981), and hemocyte phagosomes of Mytilus edulis reached a low pH of 4.5 0.5 (Kroschinski and Renwrantz, 1988). In contrast, diges- tive vacuoles of Paramecium become acidified to pH 3.0 after acidosomes bind with the digestive vacuole (Fok and Allen, 1990). In most cells, acidification of the phagosome lumen may occur in as little as 7-15 min after engulfment of foreign material (Jensen and Bainton. 1973: Geisow et al., 1981: Bassoe and Bjerknes, 1985). and the phagosome in some cell types can remain acidified for 5 h or more after phagocytosis (Horwitz and Maxfield, 1984; Rathman et al., 1996). Like the digestive vacuoles of Paramecium, the hemo- cyte phagosomes of C. virginica become highly acidified. However, the degree of acidification varies, and this varia- tion appears to be related to the final morphology of the hemocyte. Most of the individual hemocytes monitored over time underwent a morphological transition from highly granular to increasingly agranular in appearance after en- gulfment of zymosan, and measurements of phagosomal pH revealed that hemocytes remaining highly granular were not as acidic as those of the agranular form. In fact, preliminary experiments that used fluorescein as the pH-sensitive dye were unsuccessful due to the very low pH of the agranular hemocyte phagosome. The granular hemocytes of bivalves contain many cyto- plasmic granules and morphologically resemble vertebrate immune cells such as macrophages and PMNs (Cheng. 1996: Adema el al., 1991). Like their vertebrate counter- parts, hemocytes may degranulate after engulfing foreign particles (Cheng, 1996). In vertebrate PMNs, degranulation of the cytoplasm only takes place after engulfment of a foreign particle (Hirsch, 1962). and typically occurs within 30 min after phagocytosis (Hirsch and Cohn, 1960). Simi- larly, the transition in oyster hemocyte morphology typi- cally occurred within 0-30 min after engulfment of zymo- san, and granular hemocytes without engulfed particles remained granular throughout the time course. This sug- gests that granular hemocytes containing zymosan became more agranular in appearance over time as the result of increasing phagosome-lysosome fusion and subsequent de- granulation of the hemocyte cytoplasm. This also suggests that the highly acidic pH of the agranular hemocyte phago- some may be the result of increased lysosomal fusion. Although Paramecium phagosomes subsequently become alkaline after lysosomal fusion (Fok and Allen, 1990), de- granulation of mouse macrophages is associated with a gradual reduction in phagosomal pH (Geisow et al.. 1981). Amoeba proteits phagosomes also become further acidified as lysosomal fusion occurs (McNeil et al., 1983). The corresponding decrease in pH from the granular to agranular morphological state suggests the hemocyte is attempting to digest the engulfed zymosan. A similar rapid Table II Typical phagosomal pH values and method of measurement for a variety ofphagocytic cells Minimum pH Organism Cell type Target particle of phagosome Method Source rat PMN* yeast 3.5-4.5 indicator dyes Jensen & Bainton. 1973 mouse macrophage yeast 5.0 fluorescein Geisow el al.. 1981 mouse macrophage latex beads 4.0-5.0 DM-Nerf Rathman et al.. 1996 My/Hut edulis hemocyte yeast 4.5 0.5 indicator dyes Kroschinski & Renwrantz. 1988 Paramecium yeast 3.0 fluorescein Fok & Allen. 1990 PMN polymorphonuclear leukocyte. 32 A. E. BEAVEN AND K. T. PAYNTER decrease in pH occurs in Paramecium, in which the phago- somal pH decreases from 7.0 to 3.0 after the digestive vacuole containing the phagocytized material binds with acidosomes, initi;ing the process of prey killing and pro- tein denaturatin:) (Fok and Allen, 1990). Electron micro- graphs of degi 'iiulated bivalve hemocytes show that diges- tive lamellae form around partially degraded foreign material engulfed within the phagosome, and numerous glycogen granules appear in the cytoplasm (Cheng and Foley. 1975). On the basis of these observations, Cheng and Foley ( 1975) proposed that degranulated hemocytes were in the process of intracellular digestion of engulfed materials. The lowest measured pH of oyster hemocyte phagosomes (pH 2.4) is much more acidic than the lowest pH value estimated from hemocyte food vacuoles of Mytilus edulis (Table II). However, it is important to note that these researchers did not report any change in the granularity of the hemocytes (Kroschinski and Renwrantz, 1988). In ad- dition, the average pH of the agranular form of hemocytes was much lower than the pH optima of digestive enzymes contained within molluscan hemocytes (4.5-5.5). This sug- gests that these enzymes may be inactivated or that perhaps other enzymes with more acidic pH optima become active. Alternatively, it is possible that a subsequent alkalinization of the phagosome occurs, as in Paramecium and Chaos carolinensis (Fok and Allen, 1990; Heiple and Taylor, 1982). However, we have observed that agranular hemocyte phagosomes remained acidified for up to 4 h after phago- cytosis (data not shown). As in vertebrate immune defense cells, phagocytosis of foreign organisms by oyster hemocytes causes a biochem- ical cascade resulting in the production of ROIs such as superoxide anion and hypochlorous acid (HOCL) which contribute to oxidative killing of phagocytized foreign mi- croorganisms (Adema et al, 1991; Anderson et ai, 1992). The production of ROIs by oyster hemocytes reaches a peak 10 to 15 min after introduction of zymosan to hemocytes, and then gradually declines over a period of 120 min (Aus- tin and Paynter, 1995). Myeloperoxidase (MPO). the en- zyme responsible for the production of HOCL, was partially purified from oyster hemocytes and shown to have a pH optimum of 5.5 (Wojcik and Paynter, 1995). The phagoso- mal pH of granular hemocytes ( mean pH 4.9 0.02 ) is very close to the pH optimum of MPO. Thus, the production of ROIs by hemocytes just after stimulation may be the hemo- cyte's first line of defense against invading organisms. Over time, granular hemocytes become increasingly agranular in appearance as lysosomes fuse with the phagosome, result- ing in a concomitant decrease in pH. The production of ROIs may also decline as the pH decreases below the optimal pH of the ROI-producing enzymes such as MPO. In fact, at pH values less than 4, only 40% of MPO would be active (Wojcik and Paynter, 1995). In conclusion, highly and partially granular hemocytes are most often associated with internalized zymosan at the beginning of the time course. Corresponding with this early time point and morphological state is a mean pH similar to optimal pH values of hydrolytic and ROI-producing en- zymes. Over a period of 60 minutes, however, the hemocyte becomes increasingly agranular as lysosomes apparently fuse with the phagosome, further reducing the pH within the phagosome. Although the pH within the phagosome of the highly granular hemocyte is comparable to that of most other phagocytic cells, the phagosomal pH of the agranular hemocyte is more typical of the digestive vacuole in Para- mecium, which reaches a low pH of 3.0 (Fok and Allen, 1990), suggesting that phagosomal acidification is a vital component of the oyster hemocyte' s defense response. Acknowledgments The authors thank Dr. Kenneth Dunn, Indiana University Medical Center, for his help in learning the confocal tech- nique and for his patience in answering all of our questions. Literature Cited Adema, C. M., M. P. W. van der Knapp, and T. Sminia. 1991. Molluscan hemocyte-mediated cytotoxicity: the role of reactive oxygen intermediates. Rev. Aquat. Sci. 4: 201-223. Alvarez. M. R., F. E. Friedl, J. S. Johnson, and G. W. Hinsch. 1989. Factors affecting in vitro phagocytosis by oyster hemocytes. J. Inver- ted. Pathol. 54: 233-241. Anderson, R. S. 1981. Comparative aspects of the structure and function of invertebrate and vertebrate leucocytes. Pp. 629-641 in Invertebrate Blood Cells 2, N. A. Ratcliffe and A. F. Rowley, eds. Academic Press, New York. Anderson, R. S., K. T. Paynter, and E. M. Burreson. 1992. Increased reactive oxygen intermediate production by hemocytes withdrawn from Crassostrea virginica infected with Perkinsus marinus. Biol. Bull. 183: 476-481. Anderson, R. S., E. M. Burreson, and K. T. Paynter. 1995. Defense responses of hemocytes withdrawn from Crassostrea virginica infected with Perkinsus marinus. J. Invertebr. Pathol. 66: 82-89. Antoine, J., E. Prina, C. Jouanne, and P. Bongrand. 1990. Parasito- phorous vacuoles of Leishniania ama:onenesis-ir\fecled macrophages maintain an acidic pH. Infect. Immun. 58: 779-787. Austin, K. A., and K. T. Paynter. 1995. Characterization of the chemi- luminescence measured in hemocytes of the Eastern oyster, Cras- sostrea virginica. J. Exp. Zool. 273: 461-471. Bassoe, C. F., and R. Bjerknes. 1985. Phagocytosis by human leuko- cytes, phagosomal pH and degradation of seven species of bacteria measured by flow cytometry. J. Med. Microbiol. 19: 1 15-125. Black, C. M., M. Paliescheskey, B. L. Beaman, R. M. Donovan, and E. Goldstein. 1986. Acidification of phagosomes in murine macro- phages: Blockage by Nocardia asteroides. J. Inf. Dis. 154: 952-958. Cheng, T. C. 1996. Hemocytes: Forms and functions. Pp. 299-326 in The Eastern Oyster, Crassostrea virginica. V. S. Kennedy and A. F. Newell, eds. Maryland Sea Grant College. College Park. Cheng, T. C., and D. A. Foley. 1975. Hemolymph cells of the bivalve mollusc Mercenaria mercenaria: an electron microscopical study. J. Invertebr. Pathol. 26: 341-351. Cheng, T. C., and G. E. Rodrick. 1975. Lysosomal and other enzymes in the hemolymph of Crassostrea virginica and Mercenaria mercena- ria. Comp. Biochem. Physiol. 52B: 443-447. OYSTER HEMOCYTE PHAGOSOMAL pH 33 Chu, F. E., and J. K. La Peyre. 1993. /Y/Am\in marinn.\ susceptibility and defense-related activities in eastern oysters Crassostrea virginica: temperature effects. O/'.v. Ac/nut. Org. 16: 223-234. Crowle, A. J., R. Dahl. E. Ross, and M. H. May. 1991. Evidence that vesicles containing living, virulent Mycobacterium tuberculosis or My- cohacterium avnim in cultured human macrophages are not acidic. Infect. Imimm. 59: 1823-1831. Dunn, K. W., S. Mayor, J. N. Myers, and F. R. Maxheld. 1994. Applications of ratio fluorescence microscopy in the study ot cell physiology. FASEB J. 8: 573-582. Fisher, W. S. 1985. Structure and functions of oyster hemocytes. Pp. 25-35 in liiiinunitv in Invcnchrates: Cells, Molecules and Defense Reactions. M. Brehelin. ed. Springer-Verlag. New York. Fok, A. K., and R. 1). Allen. 1990. The phagosome-lysosome membrane system and its regulation in Paraincciuin. Int. Rev. Cytol. 123: 61-95. Foley, D. A., and T. C. Cheng. 1972. Interaction of molluscs and foreign substances: the morphology and behavior of hemolymph cells of the American oyster, Crassostrea virginica. in vitro. J. Invertebr. Patlwl. 19: 383-394. Foley. D. A., and T. C. Cheng. 1975. A quantitative study of phagocy- tosis by hemolymph cells of the pelecypods Crassostrea virginica and Mercenaria mercenaria. J. Invertebr. Patlwl. 25: 189-197. Geisow, M. J., P. D. Hart, and M. R. Young. 1981. Temporal changes of lysosome and phagosome pH during phagolysosome formation in macrophages: studies by fluorescence spectroscopy. J. Cell. Bit>l. 89: 645-652. Haugland, R. P. 1996. Handbook of Fluorescent Probes and Research Chemicals. 6th ed. Molecular Probes. Inc.. Eugene. Oregon. Heiple, J. M., and D. L. Taylor. 1982. pH changes in pinosomes and phagosomes in the ameba. Chaos carolinensis. J. Cell Biol. 94: 143- 149. Hirsch. J. G. 1962. Cinemicrophotographic observations on granule lysis in polymorphonuclear leucocytes during phagocytosis. J. Exp. Med. 116: 827-833. Hirsch, J. G., and Z. A. Cohn. 1960. Degranulation of polymorphonu- clear leucocytes following phagocytosis of microorganisms. J. E.\p. Med. 112: 1005-1014. Horwitz, M. 1983. The Legionaires' disease bacterium (Legionella pneumophila) inhibits phagosome-lysosome fusion in human mono- cytes. J. Exp. Med. 158: 2108-2126. Mm uii/. M. A., and F. R. Maxfield. 1984. Legionella pneumophila inhibits acidification of its phagosome in human monocytes. 7. Cell Itiol. 99: 1936-1943. Jensen. M. S., and D. F. Bainton. 1973. Temporal changes in pH within the phagocytic vacuole of the polymorphonuclear neutrophilic leuko- cyte. J. Cell Biol. 56: 379-388. Jones, T. C., and J. G. Hirsch. 1972. The interaction between To.w- pUiama fiiinilii and mammalian cells. II. The absence of lysosomal fusion with phagocytic vacuoles containing living parasites. 7. Exp. Med. 136: 1173-1194. Krosthinski, J., and L. Renwrantz. 1988. Determination of pH values inside the digestive vacuoles of hemocytes from Mytilus edulix. J. In- venebr. Palhol. 51: 73-79. Mauel, J. 1984. Mechanisms of survival of protozoan parasites in mono- nuclear phagocytes. Parasitology 88: 579-592. Maurin, M., A. M. Benoliel. P. Bongrand, and D. Raoult. 1992. Phagolysosomes of Co.\iellu burnetti-'miecled cell lines maintain an acidic pH during persistent infection. Infect. Iinmiin. 60: 5013-5016. McCormick-Ray, M. G., and T. Howard. 1991. Morphology and mo- bility of oyster hemocytes: evidence for seasonal variations. 7. Inver- tebr. Palhol. 58: 219-230. McNeil, P. L., L. Tanasugarn, J. B. Meigs, and I). L. Taylor. 1983. Acidification of phagosomes is initiated before lysosomal enzyme activity is detected. 7. Cell Biol. 97: 692-702. Kathman, M.. M. D. Sjaastad, and S. Falkow. 1996. Acidification of phagosomes containing Salmonella typhimuriitm in murine macro- phages. Infect, linmun. 64: 2765-2773. Rodrick, G. E.. and T. C. Cheng. 1974. Kinetic properties of lysozyme from the hemolymph of Crassostrea virginica. J. Invertebr. Palhol. 24: 41-48. Sibley, L. D., E. Weidner, and J. L. Krahenbuhl. 1985. Phagosome acidification blocked by intracellular Toxoplasma gondii. Nature (Lon- don) 315: 416-419. Vergne, I., P. Constant, and G. Laneelle. 1998. Phagosomal pH deter- mination by dual fluorescence flow cytometry. Anal. Biochem. 255: 127-132. Wojcik, J., and K. T. Paynter. 1995. Putative myeloperoxidase activity from hemocytes of the Eastern oyster, Crassostrea virginica. Am. Zoot. 35: 33A. Reference: Biol. Bull. 196: 34-44. (February. 1999) Differences in the Composition of Adhesive and Non-Adhesive Mucus From the Limpet Lottia limatula ANDREW M. SMITH*, TONYA J. QUICK, AND RACHEL L. ST. PETER Department of Biological Sciences, Butler University, Indianapolis, Indiana 46208 Abstract. The mucus used by the limpet Lottia limatula to form glue-like attachments was compared biochemically to the slippery mucus produced during other activities, such as suction adhesion. Colorimetric assays revealed the pro- tein content of the adhesive mucus to be 2.1 times greater than that of the non-adhesive form, and the carbohydrate content to be 1 .9 times greater. Both forms of mucus con- tained roughly six times as much protein as carbohydrate, and there was no difference in their inorganic elemental compositions. Quantitative analysis of the protein content by SDS-PAGE and a scanning densitometer revealed a similar protein composition in both forms of mucus; but three notable differences emerged. First, the overall differ- ence in protein concentration was confirmed. In addition, there was a 1 1 8 kD protein that was common only in the adhesive mucus, and a 68 kD protein that occurred only in the non-adhesive mucus. Introduction Limpets are known for their ability to attach firmly to rocks in the wave-swept intertidal /one. Recently. Smith (1992) showed that many limpets alternate between two attachment mechanisms. At high tide on the California coast, most limpets use suction adhesion; when the tide goes out, most switch to a glue-like adhesion. Suction adhesion has been studied in limpets and cephalopod molluscs (Smith, 1991a,b; Smith et shown that Ca 2+ release from intra- cellular stores tri.^ered AI discharge in Tubularia mesem- bryanthemum actinulae, followed by the inflow of Ca 2 + ions from the bathing solution into the nematocysts (Kawaii el al, 1997). Furthermore, AI discharge was usually accom- panied by sinuous movement of the aboral tentacle. The involvement of Ca 2+ -mediated signal transduction in ac- tinular settlement and metamorphosis was suspected. To clarify the role of Ca 2 + ions in actinular settlement, we examined AI discharge and larval settlement using intact larvae that were bathed in artificial seawater (ASW) con- taining various concentrations of Ca 2 + and Mg' + ions. The influence of di- and trivalent cations and Ca 2+ -channel blockers (Yang and Sachs, 1989; Santoro and Salleo, 1991; Salleo el al, 1994a, b; Gitter et al. 1994) was also exam- ined. Materials and Methods Reagents Reagent grade artificial seawater (ASW) consisting of NaCl (460 mM), KC1 (10.1 mM). CaCK (9.2 mM), and MgCU (42.6 mM) at pH 7.6 was adjusted by addition of 5 mM imidazole. The concentration of Ca 2+ was reduced, without altering the osmolarity of the ASW, by adjusting Mg 2+ accordingly. Mg 2 + -supplemented ASW with 59, 75, 125, and 206 mM of Mg 2+ was prepared by mixing the regular ASW with 370 mM MgCU aqueous solution (which is isotonic to ASW) in the ratios of 1:19, 1:9. 1:3, and 1:1. respectively, to compensate for the osmolarity. In the case of Ca 2+ -channel blockers. the osmolarity was adjusted by Na + . All Ca 2 + -channel blockers were purchased from Wako Pure Chemical Industries (Osaka, Japan). Biological materials Mature Tubularia mesembryanthemum colonies were collected mainly from submerged fisheries nets and ropes in the vicinity of Nagai Port in Sagami Bay (eastern Japan, 13930'E. 3170'N). Colonies were divided into male and female, and were maintained separately with sand-filtered running seawater at 16 2C as described by Yamashita et al. (1997). After fertilization, branches of female colonies with polyps bearing many actinulae were placed in filtered seawater (FSW; pore size. 0.45 ju,m). and sinking actinulae were collected. A single polyp could produce up to 300 actinulae, which were maintained at 4C prior to use in an experiment. The actinulae were transferred to an ASW- containing beaker and maintained at 21 2C for several hours to recover normal responsiveness. Larval age was defined as time following release from the maternal gono- phore, and 24-h-old actinulae were used in the following experiments unless otherwise stated. Settlement assay Actinular settlement was assayed using six-well polysty- rene plates (Corning Cell Wells. Corning, NY), each well containing 6 ml of ASW and 5 actinulae. The plates were placed on an orbital shaker at 21 2C. The number of settled actinulae was counted under a binocular dissection microscope after 30 h. Each experiment was performed in triplicate. To determine the reversibility of actinular settlement upon replacement of the test ASW with regular ASW. larvae that were largely sinking or attaching to the bottom of the well were washed by several replacements of the surface of the test ASW with regular ASW (roughly half the volume of the test ASW was exchanged with each replacement). Larval behavior and atnchous isorhiza discharge Each actinula was independently examined for the effect of Ca 2+ , Mg 2 + . or Ca 2+ -channel blockers on its behavior when an ATT was contacted with a clean micropipette or on AI discharge. Actinulae in regular ASW were washed and suspended in test ASWs containing various concentrations of Ca 2 + , Mg 2+ , or Ca 2+ -channel blockers. Each actinula was held by suction applied through a micropipette attached to the side of the adhesive protrusion. Behavioral responses to ATT contact with a clean micropipette were examined on suction-held actinula whose ATTs did not contact any sub- strata. For monitoring AI discharge, an ATT of a suction-held actinula was immobilized and attached to the bottom of a petri dish by applying gentle suction through a second micropipette attached to the "wrist," which is the part be- tween an ATT and an aboral tentacle. AI discharge of the immobilized ATT was then triggered by the addition of 100 jul of ASW that contained 200 mM K + ions (K + -ASW) and had been adjusted to the osmolarity of regular ASW by reducing the Na + concentration. AI discharge was observed through the 40 X objective of a microscope, and the number of AI discharged before the K + -ASW application (usually 0-1 ) was subtracted from the final total. To avoid larval disintegration in Ca 2 + -free ASW, a mi- cropipette-held actinula in regular ASW^ was quickly and thoroughly washed and suspended in Ca 2+ -free ASW. Im- mediately afterward, an ATT of the actinula was immobi- li/.ed as described above and AI discharge was triggered. [Ca :+ ], measurement [Ca 2 + ], was measured in whole mounts of living actinu- lae ATT in which AI discharge was induced. The actinulae had been treated with the Ca 2 + -chelating fluorescent indi- CA 2+ DEPENDENCE OF HYDROID SETTLEMENT 47 cator fura-2 (Kawaii el til.. 1997). The fluorescence inten- sities of ATTs from fura-2-loaded actinulae were approxi- mately 100 times stronger than those from non-labeled actinulae, and were strong enough for [Ca 2 ^]; measurement. The value of the fluorescence ratio between excitation at 340 nm and 380 nm (Rwi/w,) was used to indicate [Ca 2+ ] j . The minimum interval was 1 .9 s for ratiometric imaging. Results Ca 2+ dependence ofactinular settlement Reducing the external Ca :+ concentration had a signifi- cant effect on actinular settlement, and the effect was com- petitively antagonized by Mg 2 + (Fig. 1A). Settlement was comparable to normal in 4.6-9.2 mM Ca 2+ , but it was inhibited at lower concentrations (70% 31% at 2.3 mM Ca 2 + , 8% 10% at 1.1 mM Ca 2+ ). Raising the Mg 2 + concentration inhibited actinular set- tlement; 100% settlement was obtained at 75 mM Mg 2 + , 50% at 125 mM Mg 2 + , and 6.7% 1 1.6% at 206 mM. The relative inhibitory effect of Mg 2+ on actinular set- tlement was magnified at lower Ca 2 + . In ASW containing 2.3 mM Ca 2+ , settlement was reduced to 53% 12% at 59 mM Mg 2+ and completely inhibited at 206 mM Mg 2+ . The effects of Ca 2+ and Mg 2+ concentration on settlement were statistically significant as assessed by two-way ANOVA (Table I). Furthermore, there was a significant interaction between Ca 2+ and Mg 2 + . In each case the dose-dependent reduction ofactinular settlement was reversed by rinsing the larvae with regular ASW 9 h after the experiment. Antagonistic effect of Mg~ + Raising the Mg 2+ concentration in an otherwise regular ASW had an inhibitory effect on AI discharge (Fig. IB); the number of discharged AI was 4.0 4.2 at 125 mM Mg 2 + , and AI discharge was completely arrested at 206 mM Mg~ + . The inhibitory effect of Mg 2 ^ increased when the external Ca 2+ ion concentration was lowered. In ASW containing 4.6 mM Ca 2+ . AI discharge was reduced to 5.7 3.8 at 59 mM Mg 2 + and 4.4 3.1 at 75 mM Mg 2+ . In ASW containing 2.3 mM Ca 2+ , discharge was lowered to 4.3 1.4 at 59 mM Mg 2 + , and was completely inhibited at 125 mM Mg 2 + . Two-way ANOVA revealed that the effects of Ca 2+ and Mg 2+ concen- tration on AI discharge were statistically significant, and that there was a significant interaction between Ca 2+ and Mg 2+ (Table I). Again, this dose-dependent reduction of AI dis- charge was reversed by rinsing the larvae with regular ASW 3 h after the experiment. Figure 2 indicates significant enhancement of AI dis- charge in Mg 2 + -reduced ASW (one-way ANOVA; F - 1 1.9759, P < 0.0001 ). In this experiment. AI discharge was triggered by K + -ASW (100 mM). The number of dis- charged AI was 2.1 0.6 at 50 mM Mg 2+ (regular ASW) -''" '.. 1.1 mMCa 2 : 1 80 120 160 external Mg 2 * (mM) 200 40 80 120 160 external Mg 2+ (mM) 200 Figure 1. (A) Actmula settlement in artificial seawater with different concentrations of calcium and magnesium. Five larvae were incubated at specified concentrations of external Ca :+ . Vertical bars indicate standard deviations (n = 5). (B) K + -ASW triggered atrichous isorhiza (AI) dis- charge in artificial seawater containing different concentrations of Ca~ + and Mg :+ ions. Means were obtained by averaging the number of dis- charged AI in each aboral tentacle tip (n = 10). Vertical bars indicate standard deviations. and increased to 7.6 1.6 and 9.2 1.9 at 25 and 10 mM, respectively. Addition of Mg 2 + -reduced ASW (5 mM) also induced sinuous movement of the aboral tentacles. Effects of Ca 2 + -channel blockers on actinular settlement The effects of inorganic Ca 2+ -channel blockers on actinular settlement were examined in various concentrations of chloride salt cations added to regular ASW. In each case, a dose- dependent reduction ofactinular settlement was observed (Fig. 3A). The following ions are listed in order of increasing inhibition efficiency: Co 2+ < Ni 2+ < Cd 2+ < La 3+ < Gd 3 + . 48 S. KAWAII ET AL. Table I Results of two-way ANOVA to assess the effects of Co"" 1 " and Mg : * concentrations on (A I iciinular settlement and (B) atrichous isorhiza (AI) discharge Source* 'If F-ratio P (A) settlement Ca 3 + 3 118.4902 <0.0001 Mg 2+ 4 75.7353 < 0.0001 Ca 2+ -Mg 2+ 12 9.2255 <0.0001 (B) AI discharge Ca 2+ 3 1022.1822 <0.0001 Mg 3+ 4 1012.5841 <0.0001 Ca 2+ -Mg : + 12 534.8241 <0.0001 * Ca 2+ -Mg 2 " 1 " represents the interaction between Ca 2 * and Mg 2+ con- centrations. E o> o> (0 1 10 10" 10 J concentration (uM) The 50% inhibition concentrations (IC 50 ) were approximately 5800, 260, 53, 45, and 7 jjiM, respectively. Observed settle- ment percentages were similar to controls when the blocker was removed 5 h after the experiment. K + -induced AI discharge was also inhibited by these inor- ganic Ca 2+ -channel blockers (Fig. 3B). IC 5I1 values for Gd 3 + , La 3+ , Cd 2+ , Ni 2+ , and Co 2+ , were approximately 5, 41, 78, 500, and 6600 puW, respectively. When 10 4 juMCo 2+ . 10 3 pM Ni 2+ , and 10 2 pM Cd 2+ were added to ASW, AI discharges were lowered to 1.0 0.9, 0.8 1.2, and 0.6 1.1, respec- tively. AI discharges were similar to controls when the blocker o> D) U w 14 12 10 10 20 30 40 50 60 external Mg 2+ (mM) Figure 2. Enhancement of alrichous isorhiza (AI) discharge and settle- ment by low Mg 2 ^ artificial seawater (ASW). Actinulae were bathed in ASW containing various concentrations of Mg 2 f (regular ASW = 42.6 mM Mg 2 * ) for 1 h, and AI discharge was triggered by K ' -ASW (50, 100, or 200 mM) treatment. Means were obtained by averaging the number of discharged AI in each ATT (n = 10). Vertical bars indicate standard deviations. Asterisks (*) indicate significant differences relative to the untreated controls at P < 0.05 by the Tukey-Kramer honestly significant difference test. 10 10 2 10 3 10 4 concentration (|iM) Figure 3. Effects of Ca 2 + -channel blockers on actinular settlement (A) and atrichous isorhiyi (AI) discharge (B). All experiments were performed using sulfate-tree artificial seawater (ASW) at normal pH. Five larvae were incubated for 30 h in ASW containing various concentrations of Ca 2+ - channel blockers. was removed prior to K + -ASW stimulation 1 h after onset of the experiment. ASW containing 25 jjM Gd 3+ or 100 /J.M La 3+ decreased the AI discharge to 1.2 0.8 and 0.8 1.1, respectively, but the discharge was not restored completely when the blockers were removed. Squashing an ATT with a micropipette caused AI discharge even in the presence of 20 jjM Gd 3+ , but sinuous tentacle movement, which was usually observed in regular ASW, was inhibited. [Cir j : transients during atrichous isorhiza discharge Figure 4 compares representative changes of R of ATT during AI discharge in Ca" + -free ASW, Mg J+ -supple- 340/380 CA 2+ DEPENDENCE OF HYDROID SETTLEMENT 49 O CD cc 1.2 2.2 1.6 1.2 B 40 120 160 40 80 120 time (sec) 160 Figure 4. |Ca 2 *], transients during atrichous isorhiza discharge in (A) Ca^-free artificial seawater (ASW), (B) Mg- + -supplemented ASW (200 mM Mg 2+ ), (C) 20 pM Gd' + -ASW. and (D) regular ASW. this is the first demonstration of Ca 2 + -dependent nemato- cyst discharge in larvae. We also confirmed and quantified the Ca 2+ -dependence of actinular settlement in Tubitlaria mesembryanthemum. Reduced-Ca 2 + seawater also inhibits metamorphosis of Hydractinia cchintita plunula larvae (Berking, 1988; Muller, 1985). The EC 50 of external Ca 2 + was 3 mM for T. mesembryanthemum actinular settlement, comparable to that of both H. echimita settlement and actinular AI discharge (Kawaii et ui, 1997). Since actinulae immersed in Ca 2+ -free ASW tended to disintegrate, settle- ment assays were not performed in Ca 2 + -free ASW. The action of K + -ASW was strongly inhibited by in- creasing the Mg 2 + concentration of the bathing solution, and the inhibitory effects of the Mg 2+ ion increased when the external Ca 2+ concentrations were lowered. Similarly, Muller (1985) found that TPA-induced metamorphosis of H. echimita was promoted in Mg 2 + -reduced seawater. Mg 2+ and Ca 2 + ions compete with each other in many biological processes; consequently, we expected that an increase of Mg 2+ concentration in the bathing solution would reduce the Ca 2+ influx, thus inhibiting actinular settlement and AI discharge. However, the rise of R 340 /38o values measured in Mg 2+ -supplemented ASW was equiva- lent to that observed in Ca 2+ -free ASW, and there was no influx of Ca 2+ ions from the Mg 2 + -supplemented ASW to nematocytes. These observations indicate that inhibition of AI discharge by Mg 2+ did not result from the lowering of nematocyte [Ca 2 + ], level by competitive influx of Mg 2 + and Ca 2+ ions. The enhancement of AI discharge and larval settlement by Mg 2+ reduction may be a result of cation- mediated alteration of membrane-associated cell function in signal transduction, and Mg 2+ ions would therefore regulate AI discharge and settlement of the hydroid as an inhibitory element. mented ASW (200 mM Mg 2+ ), 20 yM Gd 3+ -ASW, and regular ASW. Despite the discharge inhibition by Gd 1+ ions, R 340/380 transients were induced by the addition of K + -ASW (200 mM). The rise of RJ^J/MO measured in the presence of Gd 3+ ions was similar to that observed in Ca^-free ASW and considerably lower than that observed in regular ASW [AR, 4()/ 380 = 0.23 0.1 1, 0.18 0.17, and 0.62 0.24, respective- ly]. No significant difference was observed between Ca 2+ -free ASW (A[Ca 2+ ], = 0.31 0.09, n = 6) and Mg 2+ -supple- mented ASW (0.28 0.11, n = 6). Discussion Ca 2 ^ -dependence of settlement and atrichous isorhiza discharge Although external Ca 2+ ions have been reported as being required for nematocyst discharge in several species of cnidaria (Salleo et ai, 1993; Santoro and Salleo, 1991a, b). Involvement of stretch-activated (SA) channels It is natural for us to speculate about involvement of stretch-activated (SA) channels in actinular settlement, be- cause the inhibitory effect of gadolinium ions at low con- centrations in our study is comparable to that of opening SA channels in patch-clamped Xenopus oocytes (Yang and Sachs, 1989). Gd 3+ is the most effective blocker of SA channels found to date, although it also has some effect on Ca 2+ channels (Sadoshima et at.. 1992). We previously demonstrated that both ATT contact with substrata and chemical stimuli are necessary to induce AI discharge (Kawaii et ai, 1997). By itself, a mechanical stimulus, such as immobilization of ATT by suction and vibration applied through a clean micropipette, did not trigger AI discharge. However, contact with a bacterial- film-coated micropipette did trigger discharge and also in- duced sinuous movement of the tentacles. Moreover, K + - ASW treatment did not result in AI discharge from ATTs 50 S. KAWAII ET AL that were not in contact with any substrata. These results suggest that K + -ASW treatment replaced sensory input of chemical stimuli, but did not mimic physical stimuli from ATT contact. Therefore, the observation that the stretching caused by immobilization of ATT did not trigger AI dis- charge does not rule out the involvement of SA channels in the discharge mechanism. Effect of C a 2 + -channel blockers All the Ca 2+ -channel blockers tested for inhibitory ef- fects on K + -ASW triggered AI discharge and larval settle- ment demonstrated similar dose-response curves and similar IC 50 values. This consistency suggests that similar Ca 2 + - channel types were involved in these events. Previously, we showed close relationships between discharge of AIs and sinuous movement of aboral tentacles (Kawaii et al, 1997). Aboral tentacles that had discharged AIs usually initiated sinuous movements, resulting in settlement behavior. Thus AI discharge can be considered as the first step in the settlement process, and inhibition of AI discharge by Ca~ + - channel blockers interrupts the subsequent processes. Squashing of ATTs caused AI discharge, but no sinuous movement was observed when ASW was treated with a Ca 2+ -channel blocker. These results suggest that these cat- ions inhibited not only the AI discharge but also the signal transmission system that follows discharge and leads to the induction of sinuous movement or the aboral tentacle sinu- ous movement itself. Direct monitoring of intracellular Ca 2+ ions also indi- cated that AI discharge required both Ca 2 + -release from intracellular stores and Ca 2+ -influx from bathing solution, and that Ca 2+ -channel blockers inhibited the large Ca 2 ' influx. Taking into consideration that K + -ASW treatment triggered limited AI discharge in Ca 2 + -free ASW (Kawaii et al., 1997) or in ASW containing Ca 2+ -channel blocker, a Ca 2+ release from intracellular stores functioned as the signal transmission system in AI discharge. In the sea anem- one Haliplanella luciae, a precise pharmacological study demonstrated that Ca 2+ acts as a second messenger, and intracellular Ca 2+ stores play an important role in the reg- ulation of nematocyst discharge (Russell and Watson, 1995), via a mechanism of "calcium-induced calcium re- lease" (Endo, 1977). It is conceivable that similar mecha- nisms for regulating nematocyst discharge are present in different types of nematocytes. In Ca 2 + -free ASW or ASW containing 10 /zM Gd 3 + , K + -ASW treatment triggered lim- ited AI discharge and caused slight elevation of R^o/sso- These small [Ca 2+ ], transients could be interpreted as a Ca 2 + release from intracellular stores to nematocyte cy- tosol. In conclusion, AI discharge was the primary action in actinular settlement, and signals from this discharge were spread throughout the larva, initiating actinular metamor- phosis. The inhibitory effect of the Ca 2 + -channel blockers demonstrated that AI discharge and subsequent signal trans- mission require involvement of Ca~ + channels. Acknowledgments The authors thank Dr. E. Hunter (The Centre for Envi- ronment, Fisheries & Aquaculture Science, Suffolk, U. K.) and Dr. C. G. Satuito (JAPAN NUS Co., Ltd., Tokyo, Japan) for their valuable suggestions. Literature Cited Berking, S. 1988. Ammonia, tetraethylammomum, barium, and amilo- ride induce metamorphosis in the marine hydroid Hydractinia. Roiu's Arch. Dev. Biol. 197: 1-9. Endo, M. 1977. Calcium release from the sarcoplasmic reticulum. Physiol. Rev. 57: 71-108. Freeman, G., and E. B. Ridgway. 1987. Endogenous photoprotein. calcium channels and calcium transients during metamorphosis in hydro/.oans. Roiu's Arch. Dev. Biol. 196: 30-50. Freeman, G., and E. B. Ridgway. 1990. Cellular and intracellular pathways mediating the metamorphic stimulus in hydrozoan planulae. Rout's Arch. Dev. Biol. 199: 63-79. Gitter, A. H., D. Oliver, and U. Thurm. 1994. Calcium- and voltage- dependence of nematocyst discharge in Hydra \-ulgaris. J. Comp. Physiol. A 175: 115-122. Kawaii, S., K. Yamashita, M. Nakai, and N. Fusetani. 1997. Intracel- lular calcium transients during nematocyst discharge in actinulae of the hydroid. Tuhularia mesembryanthemum. J. Exp. Zool. 278: 299-307. McKay, M. C., and P. A. V. Anderson. 1988. Preparation and proper- ties of cnidocytes from the sea anemone Anthopleura elegantissima. Biol. Bull. 174: 47-53. Miiller, W. A. 1985. Tumor-promoting phorbol esters induce metamor- phosis and multiple head formation in the hydroid Hydractinia. Dif- ferentiation 29: 216-222. Pantin, C. F. A. 1942. The excitation of nematocysts. J. Exp. Biol. 19: 294-310. Russell. T. J., and G. M. Watson. 1995. Evidence for intracellular stores of calcium ions involved in regulating nematocyst discharge. J. Exp. Zool. 273: 175-185. Sadoshima, J., T. Takahashi, L. Jahan, and S. Izumo. 1992. Roles of mechanosensitive ion channels, cytoskeleton. and contractile activity in stretch-induced immediate-early gene expression and hypertrophy of cardiac myocytes. Proc. Nail. Acad. Sci. USA 89: 9905-9909. Salleo, A., G. Santoro, and P. Barra. 1993. Spread of experimentally induced discharge of the nematocytes in acontia of Calliactis para- silica. Comp. Biochem. Physiol. 104A: 565-574. Salleo, A., G. La Spada, and R. Barbera. 1994a. Gadolinium is a powerful blocker of the activation of nematocytes of Pe/agia nocticuca. J. Exp. Biol. 187: 201-206. Salleo, A., G. La Spada, M. Drago, and G. Curcio. 1994b. Hyposmotic shock-induced discharge in acontia of Calliactis parasitica is blocked by gadolinium. Experientia 50: 148-152. Sanloro, G., and A. Salleo. 1991a. Cell-to-cell transmission in the activation of in situ nematocytes in acontia of Calliactis parasitica. Experientia 47: 701-703. Santoro, G., and A. Salleo. 1991b. The discharge of in situ nematocysts of the acontia of Aiptasia mutabilis is a Ca 2+ -induced response. J. Exp. Biol. 156: 173-185. Tardent, P. 1988. History and current state of knowledge concerning the discharge of cnidae. Pp. 309-332 in The Biology of Nematocysts. D. A. Hessinger and H. M. Lenhoff eds. Academic Press. San Diego. CA 2+ DEPENDENCE OF HYDROID SETTLEMENT 51 Thorington, G. U., and D. A. Hessinger. 1988a. Control of cnida discharge. I. Evidence for two classes of chemoreceptors. Bio/. Bull. 174: 163-171. Thorington. G. U., and D. A. Hessinger. 1988b. Control of discharge: factors affecting discharge of cnidae. Pp. 233-253 in The Biology of Nematocvsts. D. A. Hessinger and H. M. Lenhoff. eds. Academic Press. San Diego. Watson, G. M., and D. A. Hessinger. 1989. Cnidocyte mechanorecep- tors are tuned to the movements of swimming prey by chemoreceptors. Science 243: 1589-1591. Watson, G. M.. and D. A. Hessinger. 1991. Chemoreceptor-mediated elongation of stereocilium bundles tunes vibration-sensitive mechano- receptors on cnidocyte-supporting cell complexes to lower frequencies. J. Cell Sci. 99: 307-316. Yamashita, K., S. Kawaii, M. Nakai, and N. Fusetani. 1997. Behav- iour and settlement of actinula larvae of Tubularia mesembryanthemum Allman. 1871 (Hydrozoa: Tubulariidae). Pp. 512-516 in The Proceed- ings of the 6th 1CCB, J. C. den Hartog, ed. National Museum of Natural History, Leiden. Yanagita, T. M. 1973. The 'cnidoblast' as an excitable system. Publ. Seto Mar. Biol. Lab. 20: 675-693. Yang, X. -C., and F. Sachs. 1989. Block of stretch-activated ion chan- nels in Xenopus oocytes by gadolinium and calcium ions. Science 243: 1068-1071. Reference: Biol. Bull. 196: 52-56. (February. 1999) Free Radicals and Chemiluminescence as Products of the Spontaneous Oxidation of Sulfide in Seawater, and Their Biological Implications DAVID W. TAPLEY'-*. GARRY R. BUETTNER 2 , AND J. MALCOLM SHICK 1 1 Department of Zoology and Center for Marine Studies, University of Maine, Orono, Maine 04469-5751: and : ESR Facility/EMRB 68, College of Medicine, The University of Iowa, Iowa City. Iowa 52242-1101 The discovery ofsymbioses between marine invertebrates and sulfide-oxidizing bacteria at deep-sea hydrothennal vents and in other high-sulfide marine environments has stimulated research into the adaptations of metazoans to potentially toxic concentrations of sulfide. Most of these studies have focused on a particular action of sulfide its disruption of aerobic metabolism hv the inhibition of mito- chondria! respiration ami on the adaptations of sulfide- tolerant animals to avoid this toxic effect (/ ). We propose that sulfidic environments impose another, hitherto over- looked type oftoxicity: exposure to free radicals of oxygen, which may be produced during the spontaneous oxidation of sulfide, thus imposing an oxidative stress. Here we present evidence that oxygen- and sulfur-centered free radicals are produced during the oxidation of sulfide in seawater, and we propose a reaction pathway for sidfide oxidation that is consistent with our obsen'ations. We also show that chemi- huninescence at visible wavelengths occurs during sulfide oxidation, providing a possible mechanism for the unex- plained light emission from hydrothennal vents (2. 3). In the presence of molecular oxygen and trace metal catalysts, hydrogen sulfide spontaneously oxidizes. Oxida- tion-reduction reactions frequently involve free-radical in- termediates, and a metal-catalyzed pathway in which the initial reactions of sulfide oxidation form superoxide and sulfide radicals has been proposed (4). The proposed reac- tion begins with four steps: HS + O, -^ HS' + O,' (1) Received 18 May 1998; accepted 4 December 1998. * Present address: Department of Biology, Salem State College. 352 Lafayette St., Salem. MA 01970-5353. HS' + O, - HO,' + S (2) HS' + O,' ^ S + HO, (3) At near-neutral pH, HO 2 " will immediately protonate: HO ; +H + ^H,O, (4) The proposed metal-catalyzed mechanism underlying reac- tion 1 is (4): M n+ + O, M HS M n+ + HS' (5) (6) Notice that, once either radical is formed, a chain reaction ensues. The superoxide that is formed can undergo a sub- sequent reduction or dismutation to form hydrogen peroxide (H 2 O ; ), a known product of sulfide oxidation (5, 6), which is also produced in reactions 3 and 4. In the presence of transition-metal catalysts, superoxide and H-,0^ will react to form the hydroxyl radical, HO', perhaps the most oxidizing radical that can arise in a biological setting. A similar pathway, in which reaction 1 yields an addition product, was proposed later (7): HS' + O, (7) According to this mechanism, the addition product then reacts with molecular oxygen: HSO,' + O, -^ HSO,' (8) A subsequent series of reactions produces a variety of reactive intermediates, including superoxide, hydrogen per- oxide, and the sulfide radical (7). Although these free-radical mechanisms for sulfide oxi- 52 SULFIDE-GENERATED FREE RADICALS 53 dation have been proposed (4. 7). no direct experimental support tor such a mechanism has been provided to date. We have therefore employed electron paramagnetic reso- nance (EPR) spin trapping to gather direct evidence that free-radical intermediates are produced during sultide oxi- dation. EPR spectrometry is similar to nuclear magnetic resonance (NMR) spectrometry, but relies on magnetic mo- ments resulting from unpaired electrons instead of those from the atomic nucleus. Samples are exposed to micro- wave radiation at a fixed wavelength and amplitude while a magnetic field is swept through an appropriate range of field densities. At appropriate combinations of wavelength and magnetic field strength, the unpaired electrons will resonate, thereby absorbing microwave energy. This absorbance is recorded as the first derivative. In an EPR spectrum, the relative positions of peaks (lines) are more important than their absolute positions, so spectra typically are plotted with no abscissa, only a scale bar indicating the change in mag- netic field strength over a given distance. The ordinate is in arbitrary absorbance units. Spin trapping is a technique for detecting ephemeral radicals by providing a molecule that preferentially reacts with them, forming more stable radical adducts with characteristic spectra. These spectra typically result from a primary peak, due to the radical that was trapped, being split one or more times by adjacent paramag- netic nuclei, typically hydrogen and nitrogen. The splitting of peaks is the result of the magnetic moments of the udduct being oriented either parallel or antiparallel to the magnetic moments of adjacent nuclei, with either orientation being equally likely. The magnetic field of half of the population of adducts will be incrementally increased, while that of the other half will be equally decreased. The values of these splittings (a N and H ) for various adducts of spin-trapping agents are well known and have been tabulated. When sulfide was introduced into artificial seawater (ASW) containing the spin-trapping agent dimethylpyrro- line-N-oxide (DMPO), a prominent EPR spectrum was ob- tained (Fig. 1A, B). The greater intensity of the high-field line compared to the low-field line is consistent with ongo- ing free radical formation by oxidizing sulfide, since the high-field line is detected by the EPR spectrometer several seconds after the low-field line. The appearance of a four- line spectrum with a total width of approximately 45 G suggests that the DMPO-hydroxyl radical adduct (DMPO/ HO"; <7 N = a" = 14.9G) is present (8). But the asymmetry suggests that there are at least two radical species. The second radical should have hyperfine splittings that would produce asymmetries in the two middle lines, yet have a total spectral width near that of DMPO/HO". Appropriate candidates are the sulfide radical adduct, DMPO/S" ( N 16.09G, H = 16.19G) and the sulfite radical adduct, DMPO/*S(V (a N --- 14.5G, a" - 16.0G) (8, 9). The splittings for DMPO/S" are inconsistent with the spectra we obtained; but a computer simulation of the composite of A: Control ] ^f^^^ B: Sulfide C: Simulation D:DMSO Control ^f^^^ E: Sulfide + DMSO 10G Figure 1. Electron paramagnetic resonance (EPR) spectra of DMPO adducts formed during sulfide oxidation in air-saturated artificial seawater (ASW) (27) at pH 7.4 and room temperature (=20C). (A) Control spec- trum of DMPO in ASW with no sulfide added. (B) Spectrum obtained when sulfide (1 mM) is added. (C) Computer simulation of the spectrum in B. a composite of DMPO/HO' (30%l and DMPO/'SOr (70%). (D) Control spectrum of DMPO and dimethyl sulfoxide (DMSO) in ASW with no sulfide added. (E) Spectrum obtained when 1 mM sulfide is added to DMPO and DMSO in ASW. Computer simulation of this spectrum indi- cates these relative abundances: DMPO/HO' (20%), DMPCVSO," (60%). and DMPO/"CH 3 (20%). Horizontal scale = 10 gauss; since in this type of spectrometry the relative positions of the lines are more important than their absolute positions, spectra do not usually include an absolute scale. The vertical axis is in arbitrary units. The vertical ticks in (E) mark the positions of DMPO/TH, peaks. Reaction components, when present, were in the following final concentrations: DMPO, 50 mM; sulfide, I mM: DMSO, 0.7 M. Sulfide was added after all the other reactants, and imme- diately before transferring the sample to the EPR cell. Spectra were obtained with a Bruker ESP 300 EPR spectrometer equipped with a TM, 10 cavity and an aqueous flat cell. Computer simulations of spectra were carried out using SIMEPR software (28). DMPO/HO" and DMPO/"SO 3 ~ (Fig. 1C) reproduces the experimental spectrum well. Chen and Morris (4) postulated superoxide production in their reaction mechanism, but the experiment shown in Figure 1 does not directly support this proposal. However, trapping superoxide is difficult, first because the k obs for its reaction with DMPO is low (=30 M' at pH 7.4; cf. k obs for DMPO/HO* formation is approximately 3.4 X 10 54 D. W. TAPLEY ET AL A/" 1 s~ ' ) ( 10. II). and also because the superoxide adduct (DMPO/'OOH) can undergo further reactions, including reductions (12), to form DMPO/HO' or EPR-silent products (10. 12). We included DMSO in the reaction mixture to distinguish between these potential routes by which DMPO/ HO" can form (Fig. ID, E). The rate constants for the reactions of HO' with DMPO and DMSO are 3.4 X 10 9 M~' s~' and 7 X 10 9 M' ] s" 1 . respectively (11, 13). Under our reaction conditions DMSO should have scav- enged about 96% of the HO' formed; however, the relative abundance of DMPO/HO" in the composite spectrum de- creased by only 33%, from 30% to 20% of the composite area (Fig. IE). This implies that some DMPO/HO' is formed artifactually (10, 14). The DMPO/HO" adduct can also arise from nucleophilic substitution reactions of spin adducts (14 and references therein). For example, DMPO/ "OSO, will hydrolyze to DMPO/HO" in aqueous solution. Although formation of SO 4 " is possible in our experiments, no evidence for DMPO/"OSO,~ (t, /2 = 95s) was seen in the spectra collected. But our data do not rule out the possibility that hydrolysis gave rise to a portion of the DMPO/HO* observed. The presence of sulfide, a strong reductant, sug- gests that direct reduction of DMPO/'OOH to DMPO/HO' (12) may have occurred in these experiments. The conjecture that superoxide is produced during sul- fide oxidation, but is not detected by spin trapping is supported by investigations into the mechanisms of thiol oxidation. Superoxide can be produced during the oxidation of thiols (15-17). but it is difficult to spin trap. In most experiments, superoxide production is demonstrated indi- rectly by including a molecular probe that is indicative of it. We were unable to find a probe for superoxide that would not react with sulfide, and were thus unable by this method to demonstrate superoxide production; but the results when DMSO is included in the reaction mixture, as well as the analogy with thiol oxidation, strongly suggest that superox- ide is produced but not detected in our experiments. The reaction mechanism proposed by Chen and Morris postulates the production of the sulfide radical (reaction 1 ). but no direct evidence for its formation is yet available. The sulfide radical has been spin trapped in anoxic conditions (9). but in our oxic experiments, conversion of the sulfide radical to oxygenated products appears to be efficient. Since the reaction pathways discussed above (reactions 1-8) were first proposed, some of the postulated reactions, as well as other reactions relevant to the mechanism, have been demonstrated and their kinetics quantified. We suggest that reaction 2 is not likely to predominate; we propose instead that the addition reaction (reaction 7) predominates (k 6 = 7.5 X 10 9 A/~' -s~' at pH 7) (18, 19) and its product is then immediately deprotonated at near-neutral pH: Kinetic data for reaction 2 are not available (18). but the rate is not likely to exceed that of reaction 7. Once SO 2 " is formed, it can oxidize to yield SO, and O," (k, ="l X 10 8 AT 1 s~' at pH 6.5) (20). which will subsequently hydrate to form HSO 3 ~: SO,' + O : -> SO, + O,' SO, + + + HSO," (10) (11) If a strongly oxidizing radical is present. HSO, will be oxidized to SO,"". For example, it will react with HO" (k = 5.1 X 10 9 M~' -s~' at pH 11.2) (21): HSO," + HO" - SO,' + H : O (12) HSO,' ^ H + + SO,' (9) Our results (Fig. 1 ) strongly suggest the presence of such an oxidizing radical. Therefore, we propose that the reaction sequence 1, 7, and 9-12 occurs during the oxidation of sulfide. This sequence produces both oxygen- and sulfur- centered radicals, and is consistent with the results of our EPR spin-trapping study (Fig. 1). The oxidation of sulfide is catalyzed by trace metals (4) (see below). Therefore we believe that first-chain initiation is accomplished by reac- tion 6. Sulfide-generated free radicals could impose an oxidative challenge to tissues exposed to them, and could represent a previously unrecognized type of sulfide toxicity: oxidative stress resulting from chronic, subacute exposure to sulfide. Marine animals living in environments where sulfide and molecular oxygen coexist are at risk from exogenously formed free radicals, as well as from radicals resulting from sulfide oxidation within their own tissues. This is true in spite of generally hypoxic conditions in sulfidic environ- ments, since the presence of a strong reductant (sulfide) will enhance the production of free radicals (Fig. 1 ). We have evidence from studies on protobranch bivalves that sulfide exposure does impose an oxidative stress on these animals, and that they possess thermolabile defenses against this (22). Spectra similar to that in Figure IB are obtained when sulfide is added to heat-denatured homogenates of tissues from the protobranch bivalves Soletuya velum and Yoldia linuitiila, but are reduced or absent if undenatured homog- enates are used (22). A free-radical mechanism of sulfide toxicity might ex- plain the symptoms associated with subacute sulfide poi- soning in humans and laboratory animals. The primary symptom of subacute hydrogen sulfide poisoning is local inflammation of moist tissues exposed to the gas (6, 23, 24), especially the conjunctivae of the eye and the respiratory epithelia. In particular, sulfide-induced pulmonary edema is similar to that which appears under pulmonary oxidative stress (6, 25). Since these symptoms are restricted to moist tissues, the mechanism of irritation probably involves the aqueous reactions of hydrogen sulfide, including those pro- ducing radicals. SULFIDE-GENERATED FREE RADICALS 55 600 500 a> CD o o o o o o o - tu 1 s* * \ NxxxXxXxxxxxVfi / \ It^XXXXXXXXXXXX I I xxxxxxxxxvvxx^H ' -.\ K^XXXXXXXXXXXX i iZ i ir> Controls o b b b b b o b c 5-HT [SIN-1] + 10' 4 IV B Percent metamorphosis N> 4^ O) 00 O O O O O O O ;| I -T| T n= * XXXXXXXXXN X ! ! ^ \ I bobbbbbobb'oc Q Q Q Q Q Q Controls [SNAP] + I 5-HT Figure 2. Inhibition of 5-HT-induced metamorphosis by SIN-1 (A) and SNAP (B) at 48 h. (A) Asterisk indicates concentration of unreplaced SIN-1 that significantly inhibited 5-HT-induced metamorphosis at 24 h (41% metamorphosis in 1(T 4 M, ;To.uo5ci> = '5-5) and 48 h (65% meta- morphosis, ^o.oo5cn = 30.77) compared to the 5-HT control. Arrows indicate concentrations that showed significant inhibition compared to 5-HT only at 24 h (e.g.. 59% metamorphosis in 10~ 9 M at 24 h, ^o.oosd) = 13.9). but not at 48 h. All SIN-1 solutions contained 10~ 4 M 5-HT. (B). Asterisk indicates concentration of SNAP that significantly inhibited 5-HT- induced metamorphosis at 24 and 48 h. Arrow indicates concentration that was inhibitory only at 24 h (40% metamorphosis in 10~ 4 M, ^ooosoi = 1 1 .3), but not at 48 h. Solutions of SNAP have a half-life of about 1 h and were changed every 6 h to maintain relatively steady concentrations of NO. D10~ x = degassed solution of 10 x M SNAP plus 10~ 4 M 5-HT; 10~ x = active solution of 10~ x M SNAP plus 10~ 4 M 5-HT. 60 S. J. FROGGETT AND E. M. LEISE en 100 'in -n 80 Q. t_ 60 40 1 20 2 Q) CL Ii n=60 Q Q K H 1 1 I I J T- ~ Controls [L-NAME] B - Arrows indicate concentrations that were significantly effective by 24 h but not at 48 h. n = 30 for 10"" and 10~' M L-NAME. (B) Injections of the inactive isomer D-NAME induced no significant rates of metamorphosis by 48 h. (iS-HT = injected 5-HT). phose by diatoms or associated organisms that occur natu- rally in their littoral habitats (Leise et ai, 1996), but we have no understanding of the time course for metamorpho- sis in that situation. Our results have led us to hypothesize that NO production is necessary for the maintenance of the larval state until an appropriate metamorphic cue is de- tected. Preliminary data from experiments on Pliestilla si- bogae suggest that NO may be active in the metamorphic pathway in this species as well (Meleshkevitch et ai, 1997). The ubiquity of NO in molluscan metamorphosis and its specific actions in this process remain to be determined. Although the tranformations that invertebrate larvae un- dergo in reaction to metamorphic cues are among their most well known activities (Pawlik, 1992), unsuitable habitats can also elicit distinctly negative responses from some larvae, such as those of the polychaetes Nereis vexillosa and Capitella sp. (Woodin, 1991). We recently found a similar effect on llvanassa larvae from one species of benthic diatom. Extracts of cultures of a sheathed pennate diatom species that were isolated from sediments obtained at Myr- tle Grove. North Carolina, inhibit spontaneous metamor- phosis in older (>3 weeks in culture) llyanassa larvae (Leise et ai, 1996; unpubl. data). Such negative metamor- phic actions and the uncertainty of larval encounters with appropriate juvenile habitats suggest that the maintenance of the larval life-history phase is an integral component of the metamorphically competent state. For llyanassa, the production of NO by competent larvae appears to be nec- essary for this purpose. However, maintenance of the larval state is likely to depend upon more than one inhibitory compound. For example, Pires et ai (1996) suggested that norepinephrine might inhibit the circuits controlling meta- morphosis in the slipper limpet Crepidula fornicata. We do not yet know how Il\unassa larvae utilize dopamine or other catecholamines. o) 100 7 ^ 0) S n=60 o ^ 80 . ' Q. o ; 4- # E 60 to | 40 S 20 n n |i i I * noi- y E Controls [L-NMMA] B 100 80 60 40 20 n=60 o g H ~ ^ s Controls [L-NMMA] Figure 4. (A) Injections of all concentrations of L-NMMA induced significant rates of metamorphosis by 24 h in the absence of any inducer. Asterisks indicate concentrations that triggered rates of metamorphosis significantly different from those induced by iFIO (e.g., at ICT 5 M, ATo.oni) = 11.5). Note unusually high levels of metamorphosis in both 5-HT controls. (B) Levels of metamorphosis detected at 48 h. No concentrations remained significantly different from iFIO (^oosm = 6.2). Note the unusually high levels of metamorphosis in both iFIO and FIO. NITRIC OXIDE IN LARVAL ILYANASSA 61 100 r 80 E 60 | 40 S 20 Q CD CL n=60 Q Q O O O O O Controls [Carb-PTIO] Figure 5. No concentrations of injected Carboxy-PTIO (Carb-PTIO), an NO scavenger, induced significant rates of metamorphosis by 48 h when compared to iFIO ( 10~ 5 M ^o.oim = 5.76, 10~ 6 M JCuHim ~ 3.84). n = 30 for 10 " and 10 " M Carboxy-PTIO. Significance levels for 24 and 48 h data were the same. Developing nervous systems in vertebrates and arthro- pods express NO transiently in a variety of areas (Bredt and Snyder. 1994: Truman el ai, 1996: Gibbs and Truman, 1998: Scholz et al.. 1998). NO has been reported to cause growth cone collapse (Renterfa and Constantine-Paton. 1996) and may act in the regulation of neuronal prolifera- tion (Peunova and Enikolopov, 1995; Kuzin et al., 1996), affecting the ability of axons to reach appropriate targets and initiate synaptogenesis (Bredt and Snyder. 1994; Wu et ill.. 1994; Truman et al., 1996; Gibbs and Truman, 1998; Scholz c/ ul.. 1998). Comparable roles for this molecule in molluscs are just beginning to come under investigation. How NO exerts its effects in larval Ilvanassa is still unknown. Typically, NO binds to guanylyl cyclase, stimu- lating the formation of cyclic guanosine 3 '.5' monophos- phate (cGMP) (Murad et al., 1978); the biochemistry of the nitrergic signaling pathway appears to remain applicable to both vertebrate and invertebrate systems (Dawson et al.. 1991: Elphick et al.. 1993; Elofsson et al.. 1993; Huang et al.. 1998). Recent work on the growth and survival of cultured neurons suggests that NO may also affect cGMP- independent intracellular signaling pathways (Gonzalez-Zu- lueta et al., 1997). We cannot yet distinguish between these two mechanisms in our experimental animal. At present, we suggest that NO is produced within the developing mollus- can nervous system and diffuses to its target cells to activate guanylyl cyclase, thereby increasing intracellular levels of cGMP. We hypothesize that high levels of cGMP are nec- essary for the maintenance of larval tissues. We anticipate that neuronal somutu in the apical ganglion the brain re- gion that governs key larval functions will express high levels of NOS. In the presence of a natural metamorphic inducer. nitrergic neurons are probably inhibited, either directly by serotonergic neurons or by feedback from acti- vated NO targets. Activation of serotonergic neurons and the resultant inhibition of NOS activity would decrease levels of cGMP, allowing metamorphosis to proceed. In- vestigations into the downstream actions of NO are just beginning. Given its widespread occurrence in behaviorally signifi- cant neural circuits throughout the animal kingdom, NO would appear to be a relatively ancient neurotransmitter. In adult molluscs, NO functions as an intercellular messenger in behaviorally important circuits. NO appears to be neces- sary for learning in cephalopods (Chichery and Chichery, 1994; Robertson et al.. 1994. 1995. 1996). olfaction in pulmonates (Gelperin, 1994a, b; Gelperin et al., 1996), and feeding in several gastropods (Moroz et al., 1993; Elphick et al.. 1995; Teyke, 1996). Our understanding of the impor- tance of this molecule in developing organisms is still relatively immature, but the growing literature indicates that this molecule can be differentially activated to coordinate specific developmental events occurring throughout a Held of maturing neural tissue (Edelman and Gaily, 1992; Bredt and Snyder, 1994; Wu et al., 1994; Peunova and Enikol- opov, 1995; Kuzin et al.. 1996; Renterfa and Constantine- Paton, 1996: Truman et al., 1996; Gibbs and Truman. 1998: Scholz et al.. 1998). In larvae of marine molluscs, nitrergic pathways may have been exploited to regulate diverse target tissues, much as the ecdysteroids coordinate activity during insect metamorphosis (Riddiford and Truman. 1993). Ec- dysteroid synthesis is inhibited in crustaceans by molt- inhibiting hormone (reviewed in Fingerman, 1997), which may have a molluscan analog in NO. Our comprehension of the mechanisms that drive molluscan metamorphosis will be aided by further explorations of this pathway. Acknowledgments This work was supported in part by NSF grant IBN- 9604516. Literature Cited Barlow, L. A., and J. W. Truman. 1992. Patterns of serotonin and SCP immunoreactivity during metamorphosis of the nervous system of the red abalone, Huliotis rufescens. J. Neurobiol. 23: 829-844. Bland, M. 1995. An Introduction to Medical Statistics. Oxford Univer- sity Press, Oxford. Bonar, D. B. 1978. Ullrastructure of a cephalic sensory organ in larvae of the gastropod Phesti/lii sibogae (Aeolidacea. Nudibranchea). Tissue Cell 10: 153-165. Bonar, D. B., S. L. Coon, M. Walch, R. M. Weiner, and W. Fitt. 1990. Control of oyster settlement and metamorphosis by endogenous and exogenous chemical cues. Bull. Mar. Sci. 46: 484-498. Bredt, D. J., and S. H. Snyder. 1994. Transient nitric oxide synthase neurons in embryonic cerebral cortical plate, sensory ganglia, and olfactory epithelium. Neuron 13: 301-313. Chichery, R., and M.-P. Chichery. 1994. NADPH-diaphorase in a cephalopod brain (Sepia): presence in an analogue of the cerebellum. NeuroReport 5: 1273-1276. Coon, S. L.. and D. B. Bonar. 1986. Norepinephrine and dopaniine content of larvae and spat of the Pacific oyster, Crassostrea gigas. Bio/. Bull. 171: 632-639. Couper, J. M., and E. M. Leise. 1996. Serotonin injections induce 62 S. J. FROGGETT AND E. M. LEISE metamorphosis in larvae of the gastropod mollusc llyanassa obsoleta. Biol. Bull. 191: 178-186. Dawson, T. M., D. S. Bredt, M. Fotuhi, P. M. Hwang, and S. H. Snyder. 1991. Nitric oxide synlhase and neuronal NADPH diaphorase are identical in brain and peripheral tissues. Proc. Natl. Acad. Sci. USA 88: 7797-7801. Edelman, G. M., and J. A. Gaily. 1992. Nitric oxide: linking space and time in the brain. Proc. Null. Acad. Sci. USA 89: 1 1651-1 1652. Elofsson, R., M. Carlberg, L. Moroz. L. Nezlin, and D. Sakharov. 1993. Is nitric oxide (NO) produced by invertebrate neurones? NeuroReport 4: 279-282. Elphick. M. R., I. C. Green, and M. O'Shea. 1993. Nitric oxide synthesis and action in an invertebrate brain. Brain Res. 619: 344-346. Elphick, M. R., G. Kemenes. K. Staras, and M. O'Shea. 1995. Be havioral role tor nitric oxide in chemosensory activation of feeding in a mollusc. J. Neiirosci. 15: 7643-7664. Fingerman, M. 1997. Crustacean endocrinology: a retrospective, pro- spective, and introspective analysis. Physiol. Zool. 70: 257-269. Froggett. S., and E. M. Leise. 1996. Does nitric oxide inhibit metamor- phosis in a larval mollusc? Soc. Neiirosci. Abslr. 22: 364. Froggett, S., and E. M. Leise. 1997. Endogenous nitric oxide inhibits metamorphosis in a larval mollusc. Soc. Neiirosci. Abslr. 23: 1234. Gelperin, A. 1994a. Nitric oxide mediates network oscillations of olfac- tory interneurons in a terrestrial mollusc. Nature 369: 61-63. Gelperin, A. 1994b. Nitric oxide, odour processing and plasticity. Nelh. J. Zool. 44: 159-169. Gelperin, A., D. Kleinfeld, W. Denk, and I. R. C. Cooke. 1996. Os- dilations and gaseous oxides in invertebrate oltaction. J. Neiirobiol. 3(1: 110-122. Gihbs, S. M., and J. W. Truman. 1998. Nitric oxide and cyclic GMP regulate retinal patterning in the optic lobe of Drosophila. Neuron 20: 83-93. Gonzalez-Zulueta, M., H. -Y. Yun, V. L. Dawson, and T. M. Dawson. 1997. Nitric oxide mediates activity-dependent neuronal survival. Soc. Neiirosci. Abslr. 23: 630. Hadfield, M. G., and J. T. Pennington. 1990. Nature of the metamor- phic signal and its internal transduction in larvae of the nudibranch Phestilla sibogae. Bull. Mar. Sci. 46: 455-464. Huang. S.. H. H. Kerschbaum. and A. Hermann. 1998. Nitric oxide- mediated cGMP synthesis in Helix neural ganglia. Brain Res. 780: 329-336. Jacklet, J. W., and M. Gruhn. 1994. Nitric oxide as a putative trans- mitter in Aplysia: neural circuits and membrane effects. Nelh. J. Zool. 44: 524-534. Kempf, S. C., G. V. Chun, and M. G. HadHeld. 1992. An immunocy- tochemical search for potential neurotransmitters in larvae of Pheslilla sibogae (Gastropoda. Opisthobranchia). Comp. Biochem. Physiol. C 101: 299-305. Kempf, S. C., L. R. Page, and A. Pires. 1997. Development of seroto- nin-like immunoreactivity in the embryos and larvae of nudibranch mollusks with emphasis on the structure and possible function of the apical sensory organ. J. Comp. Neurol. 386: 507-528. Ku/iii. B., I. Roberts, N. Peunova, and G. Enikolopov. 1996. Nitric- oxide regulates cell proliferation during Drosophila development. Cell 87: 639-649. Leise, E. M., J. E. Nearhoof, S. J. Eroggett, and L. B. Cahoon. 1996. Benthic diatoms induce metamorphosis in larvae of the caenogastropod mollusc llyanassa obsoleta. Am. Zool. 36: 107 A. Lin, M. -.. and E. M. Leise. 1996a. Gangliogenesis in the prosobranch gastropod Ilyanaxsa obsoleta. J. Comp. Neural. 374: 180-193. Lin, M. -F., and E. M. Leise. 1996b. NADPH-Diaphorase activity changes during gangliogenesis and metamorphosis in the gastropod mollusc Ilyanaxsa obsoleta. J. Comp. Neural. 374: 194-203. Marois, R., and T. J. Carew. 1997. Ontogeny of serotonergic neurons in Aplysia californica. J. Comp. Neurol. 386: 477-490. Meleshkevitch, E. A., D. Y. Budko, S. W. Norbv, L. L. Moroz, and M. G. Hadfield. 1997. Nitric oxide-dependent modulation of the metamorphosis in mollusc Phestilla sibogae (Gastropoda, Nudi- branchia). Soc. Neiirosci. Abslr. 23: 1233. Moroz, L. L., J. H Park, and \V. Winlow. 1993. Nitric oxide activates buccal motor patterns in Lymnaea stagna/is. NeuroReporl 4: 643-646. Morse, D. E. 1994. Molecular mechanisms controlling metamorphosis and recruitment in abalone larvae. Pp. 107-119 in Aba/one of the World. S. A. Shepherd. M. J. Tegner. and S. A. Guzman del Proo, eds. Blackwell. Oxford. Murad. F.. C. Mittal, W. Arnold, S. Katsuki, and H. Kimura. 1978. Guanylate cyclase: activation by a/.ide. nitro compounds, nitric oxide and hydroxyl radical and inhibition by hemoglobin and myoglobin. Adv. Cyclic Nucleotide Res. 9: 145-158. Pawlik, J. R. 1992. Chemical ecology of the settlement of benthic marine invertebrates. Oceanogr. Mar. Biol. Anmi. Rev. 30: 273-335. Peunova, N., and G. Enikolopov. 1995. Nitric oxide triggers a switch to growth arrest during differentiation of neuronal cells. Nature 375: 68-73. Pires, A., and M. G. Hadfield. 1993. Responses of isolated vela of nudibranch larvae to inducers of metamorphosis. J. Exp. Zool. 266: 234-239. Pires, A., J. A Skiendzielewski, and J. V. Mitten. 1996. Depletion of norepinephrine and metamorphosis in a gastropod. Am. Zool. 36: 13A. Pires, A., S. L. Coon, and M. G. Hadfield. 1997. Catecholamines and dihydroxyphenylalanine in metamorphosing larvae of the nudibranch Pheslilla sibogae (Gastropoda: Opisthobranchia). J. Comp. Physiol. A 181: 187-194. Renten'a, R. C., and M. Constantine-Paton. 1996. Exogenous nitric oxide causes collapse of retinal ganglion cell axonal growth cones in vitro. J. Neiirobiol. 29: 415-428. Riddiford, L. M., and J. W. Truman. 1993. Hormone receptors and the regulation of insect metamorphosis. Am. Zool. 33: 340. Robertson, J. D., J. Bonaventura, and A. kohm. 1994. Nitric oxide is required for tactile learning in Octopus vu/garis. Proc. R. Soc. Land. B. 256: 269-273. Robertson, J. D.. J. Bonaventura, and A. Kohm. 1995. Nitric oxide symhase inhibition blocks octopus touch learning without producing sensory or motor dysfunction. Proc. R. Soc. Land. B. 261: 167-172. Robertson, J. D., J. Bonaventura, A. Kohm, and M. Hiscat. 1996. Nitric oxide is necessary for visual learning in Octopus vtilgaris. Proc. R. Soc. Loud. B 263: 1739-1743. Scholz, N. L., E. S. Chang, K. Graubard, and J. W. Truman. 1998. The NO/cGMP pathway and the development of neural networks in postembrvonic lobsters. J. Neiirobiol. 34: 208-226. Sokal. R. R., and F. J. Rolf. 1981. Biometry. W. H. Freeman. New York. Teyke. T. 1996. Nitric oxide, but not serotonin, is involved in acquisition of food-attraction conditioning in the snail Helix pomatia. Neiirosci. Lettr. 206: 29-32. Truman. J. W., J. De Vente, and E. E. Ball. 1996. Nitric oxide- sensitive guanylate cyclase activity is associated with the maturational phase of neuronal development in insects. Development 122: 3949- 3958. Woodin, S. A. 1991. Recruitment of infauna: positive or negative cues? Am. Zool. 31: 797-807. Wu, H. H., C. V. Williams, and S. C. McLoon. 1994. Involvement of nitric oxide in the elimination of a transient retinotectal projection in development. Science 265: 1593-1596. Zar, J. H. 1974. Bioxiulixtical Analysis. Prentice-Hall, Englewood Cliffs. NJ. Zimmer-Faust, R. K., and M. N. Tamburri. 1994. Chemical identity and ecological implications of a waterborne. larval settlement cue. Limnol. Oceanogr. 39: 1075-1087. Reference: Hint. Bull. 196: 63-69. (February. 19991 Developmental Basis of Phenotypic Variation in Egg Production in a Colonial Ascidian: Primary Oocyte Production Versus Oocyte Development J. STEWART-SAVAGE*. BRADLEY J. WAGSTAFF 1 , AND PHILIP O. YUND 2 Department of Biological Sciences, University of New Orleans, New Orleans, Louisiana 70148 Abstract. Colonies of the ascidian Botryllus schlosseri (a cyclical hermaphrodite) exhibit extreme variability in egg production, and there is a large genetic component to this phenotypic variation. Therefore, the developmental bases of variation among different genotypes was investigated. Col- onies differing in egg production (assayed as number of eggs per asexual bud) were cultured in a common garden experiment, and buds were collected and fixed early in the reproductive cycle. The buds were serially sectioned, and the number and size of the oocytes in the developing ovaries were determined for the different genotypes. Because the buds were collected prior to the onset of vitellogenesis. they contained oocytes at the three previtellogenic stages. In reproductive colonies (>0.7 eggs per bud), there were neg- ative relationships between the final number of eggs per bud and ( 1 ) the total number of oocytes present, (2) the number of stage 1 oocytes present, and (3) the number of stage 2 oocytes present. There was no relationship between these parameters in nonreproductive colonies (<0.3 eggs per bud). In contrast, the number of stage 3 oocytes per bud was positively correlated with the final number of eggs per bud in both reproductive and nonreproductive colonies. In re- productive animals there was a negative relationship be- tween the total number of oocytes per bud and the percent- age of oocytes at stage 3 in oogenesis. A principal component analysis revealed that a single vector equally weighted for the number of eggs per bud. the total number of oocytes per bud. and the percentage of oocytes at stage 3 Received 28 May 1998; accepted 9 November 1998. * To whom correspondence should be addressed. E-mail: jssavagecc uno.edu. ' Current Address: Department of Zoology, University of Texas. Austin. TX 78712. 2 Current Address: School of Marine Science, c/o Darling Marine Cen- ter. Walpole. ME 04573. accounted for 84% of the observed variation in reproductive colonies. These data indicate that the phenotypic variation in egg production among the B. schlosseri colonies in the Damariscotta River. Maine, is controlled by genetic varia- tion in both the number of oocytes that populate developing ovaries, and the percentage of oocytes that reach stage 3 in oogenesis. Introduction The production of ova is an exceedingly important aspect of the life-history strategy of any female or hermaphroditic organism. Because egg production is a primary determinant of evolutionary fitness, the mechanisms by which genetic and environmental factors may produce variation in fecun- dity among individuals must be assessed. Nutritional studies in teleosts and lizards have demonstrated that suboptimal diet reduces the number of ovulated eggs, either by reducing the number of oocytes that enter vitellogenesis, or by in- creasing oocyte atresia (Mendez-de la Cruz et ai, 1993; Tyler and Sumpter. 1996). Although a genetic basis for the variation in female fecundity has been demonstrated in many taxa. analysis of the genetic and developmental mech- anisms that produce phenotypic variation in egg production has largely lagged behind (but see Land and Robinson, 1985. for studies in sheep). The genetic and developmental mechanisms controlling egg production in marine inverte- brates are unknown and unstudied. The colonial marine ascidian Botryllus schlosseri has characteristics that make it a desirable candidate for an investigation of the developmental mechanisms controlling intraspecific variation in egg production. First, both ovarian development and oocyte maturation occur in a number of repeated cycles, once an animal attains sexual maturity (Milkman, 1967; Sabbadin and Zaniolo. 1979; Manni et ai. 63 64 J. STEWART-SAVAGE ET AL. 1994; Yund et ai, 1997). Colonies grow by asexual bud- ding, and each developing bud within the colony has the capacity to form a pair of ovaries and testes (though the oogenic potential of buds varies within a colony; Sabbadin and Zaniolo, 1979). Consequently, the number of eggs produced per bud, rather than total colony-wide egg pro- duction, is generally used to assay relative female fecundity. Secondly, phenotypic variation in egg production occurs both within and among populations. The population in the Damariscotta River, Maine, exhibits a continuous range of variation, from to 6 eggs per bud (Yund et /., 1997). In contrast, the population in the Eel Pond at Woods Hole, Massachusetts, has a bimodal distribution in egg produc- tion, with the low mode at 2 eggs per bud and the high mode at 10 eggs per bud (Grosberg, 1988). Finally, in both pop- ulations, phenotypic variation in egg production is known to have both genetic and environmental components (Gros- berg, 1988; Yund et ui, 1997). In this report, we examine both ovarian development and oocyte maturation in the population of B. schlosseri living in the Damariscotta River. We evaluate three mutually compatible hypotheses for the production of a continuous range of phenotypes in the number of eggs produced per bud: ( 1 ) the number of oocytes that populate the developing ovaries within a bud varies by genotype, and a fixed per- centage of these oocytes mature; (2) a fixed number of oocytes populate a developing bud, but different percent- ages of the oocytes mature in different genotypes; (3) a fixed number of oocytes populate a developing bud, and a fixed percentage of them mature, but the number of matur- ing oocytes that become atretic varies by genotype. Materials and Methods Stud\ organism Colonies of Botryllus schlosseri are composed of asexu- ally produced zooids arranged in clusters, or systems, with all zooids in a system sharing a common exhalant siphon. Throughout the life of a colony, all of the zooids undergo a series of synchronous, overlapping sexual and asexual cy- cles. The asexual zooid replacement cycle starts when a new generation of zooids, called buds, form on the side of existing zooids (Berrill, 1941). After about 14 days of development and growth at 16C, these buds swell and their inhalant siphons open. As the new zooids take over the function of the previous generation of zooids, which are quickly resorbed, they enter into their sexual cycle. The reproductive cycle includes the internal fertilization of the mature eggs soon after the inhalant siphons open (Milkman, 1967); the continuous release of sperm starting 16 h later (Stewart-Savage and Yund, 1997); and the brooding of the developing embryos, which are released just before the zooids degenerate (Milkman, 1967). B. schlosseri colonies can be staged according to their 14-day bud development cycle (Berrill, 1941; Izzard, 1973), or according to their 7-day reproductive cycle (Milkman. 1967). Because the next generation of buds is formed halfway through the bud development cycle, there are always three generations of zooids in a colony. During most of the reproductive cycle these three generations are adult zooid, primary bud, and a younger secondary bud. When the zooids of a new gener- ation open their siphons (stage according to Milkman, 1967), the colony contains degenerating zooids, newly opened zooids (which were primary buds), and primary buds (which were secondary buds). Colonv staging and bud collection The B. schlosseri colonies used in this study were col- lected from the Damariscotta River. Maine. Animals were grown on glass microscope slides in the flowing seawater system at the University of Maine's Darling Marine Center in a common garden experiment, and thus experienced identical environmental conditions. Colonies were staged according to Milkman's (1967) six-stage reproductive cy- cle: colonies are at stage 0-1 during the first 24 h after a new generation of zooids open their siphons and at stage 3 I- when the expanding primary buds contain oocytes at the end of vitellogenesis. Over the experimental period, colonies at stage 3-4 were assayed for the number of zooids. the number of primary buds per zooid, and the average number of mature eggs per bud. To assay overall genotypic egg production, the average number of mature eggs per bud was determined by recording the number of eggs per bud in 10 randomly chosen primary buds (Yund et al, 1997). This subsampling technique should minimize the effect of intracolony variation in oogenic potential (Sabba- din and Zaniolo. 1979). When the colonies reached stage 0-1, the area containing the primary buds was collected by making midline cuts through two adjacent zooids and re- moving the entire tunic between them. For each colony, 10 pieces of tunic from at least two areas of the colony were collected and fixed (see below). The colonies were reas- sayed when they reached stage 3 \. The excised buds were fixed in 2% glutaraldehyde in 20 mM TRIS-buffered seawater, pH 8.0. After being fixed for at least 24 h, the buds were rinsed in seawater and then in distilled water, and then stained with Harris hematoxylin (Sigma, St. Louis, MO) for 15 min. After a distilled water rinse, the buds were dehydrated and embedded in methac- rylate plastic (JB-4, Polysciences, Warrington, PA). Blocks, which contained 3-5 buds from a colony, were serially sectioned at 6 ;um. A calibrated reticle at 160X magnifica- tion was used to measure the largest diameter of the oocytes within a bud; the size of elliptical oocytes was determined by averaging the two dimensions. For each animal, the total number of oocytes and the number of oocytes at each stage EGG PRODUCTION IN BOTRYLLUS 65 AS Figure 1. Micrographs of the different stages of oocyte development seen in developing buds of Botryllus schlosseri (stage 7 according to Izzard, 1973). See Results for details of oocyte staging. 1-3, stage 1-3 oocytes; AS. antral sac; P. developing pharynx: T, testis rudiment. (A) Cross section of a bud with a developing testes and stage 2 and 3 oocytes. The lightly stained stage 3 oocytes are surrounded by a darker, cuhoidal follicular layer. 165X, bar = 50 jxm. (B) Higher magnification of stage 2 oocytes shown in A. Small primary follicle cells (arrows) can be seen of oogenesis were determined by averaging the data from 6.4 1.9 (X SD; range: 4-11) buds. Because of the tradeoff between sexual reproduction and asexual growth ( Yund et a!., 1997). the 20 colonies used in this study were all of medium size ( 1 10 44 zooids, X SD) and had equal rates of asexual growth ( 1.6 0.2 buds per zooid, X SD). To limit the effect of food supply on the number of eggs per bud (Grosberg, 1988; Yund et al., 1997), all samples were collected over an 1 1-day period (27 June- 8 July 1995). Six of the 20 colonies were producing few, if any, eggs. Because colony age and the environment can both affect egg production (Grosberg. 1988; Yund et til.. 1997) and we do not know if the lack of egg production in these colonies is the result of genetic factors, environmental factors, or a combination of both, we separated the colonies into "reproductive" ( >0.7 eggs per bud) and "nonreproduc- tive" (<0.3 eggs per bud) groups. Results Oocvte morphology and staging The primary buds collected from B. schlosseri colonies between stages and 1 in the reproductive cycle described by Milkman (1967) were, based on their histological ap- pearance, at stage 7 according to Izzard' s classification ( 1973). As expected, the ovaries of these developing buds contained only previtellogenic oocytes at three stages (Manni et al., 1994) of oogenesis (Fig. 1). The small stage 1 oocytes (Fig. 1C) can be easily distinguished from the other cells in the developing ovary and from the testes rudiment by their larger size (10-15 /im versus 6-8 /nm, respectively) and their large and prominent nucleolus. Stage 2 oocytes are characterized by an increase in both cell and nuclear size, and by an increase in the basophilia of the cytoplasm (Fig. 1A. B). Although Manni et al. (1994) described stage 2 oocytes as being 40-60 /urn, we found definitive stage 2 oocytes that were only 20 jam. Stage 3 oocytes are distinguished from stage 2 oocytes by a further increase in oocyte and nuclear size, by a decrease in cyto- plasmic basophilia, and by the presence of a cuboidal fol- licular layer (Fig. 1A). On the basis of these morphological criteria, stage 3 oocytes ranged in size from 60 to 100 jam. In the 126 buds sectioned, we observed only four atretic stage 3 oocytes (not shown). These oocytes were classified as atretic because test cells had migrated into oocyte cyto- plasm, and the oocyte had no germinal vesicle. These four stage 3 oocytes were excluded from the data set. around each of the six stage 2 oocytes. 650X, bar = 10 jiun. (C) Cross section of a bud containing only stage 1 oocytes. Notice the difference in cell size and nucleolar morphology between the stage 1 oocytes and the cells in the developing testes (area within broken line). 650X. 66 J. STEWART-SAVAGE ET AL. Relationship bet\veen oocyte number and stage, and final egg number As discussed in the Introduction, the average number of eggs produced in each bud (the colony's eggs-per-bud phe- notype) may be controlled by the total number of oocytes within each developing bud. In reproductive colonies (>0.7 eggs per bud), the final eggs-per-bud phenotype is nega- tively related to the total number of oocytes within each bud, whereas in nonreproductive colonies (<0.3 eggs per bud) there is no relationship (Fig. 2A). This relationship indicates that the final number of eggs produced within each bud may be negatively regulated by the number of oocytes that populate the bud, but that the final determination of a colony's eggs-per-bud phenotype must occur during oogen- esis. To determine the stage or stages of oogenesis at which the final eggs-per-bud phenotype is determined, we exam- ined the relationship between the number of oocytes at each stage of oogenesis and the final eggs-per-bud phenotype. In reproductive colonies, there is a negative relationship be- tween the final eggs-per-bud phenotype and both the num- ber of stage 1 and stage 2 oocytes per bud (Fig. 2B and C). In nonreproductive colonies, there is no relationship be- tween the final eggs-per-bud phenotype and the number of stage 1 and 2 oocytes. The number of stage 3 oocytes per bud is positively correlated with the final eggs-per-bud phenotype in both reproductive and nonreproductive colo- nies (Fig. 2D). The correlation of a colony's final eggs-per-bud pheno- type with both the total number of oocytes in each bud (Fig. 2A) and the number of oocytes at stage 3 in oogenesis (Fig. 2D) indicates that the processes determining these two conditions may be coordinated. To determine the relation- ship between these processes, we converted the number of stage 3 oocytes per bud to a percentage to remove the negative relationship between the final eggs-per-bud pheno- type and the total number of oocytes per bud. The number and percentage of oocytes in a bud at stage 3 in oogenesis are equivalent measures of oocyte maturation (R = 0.797). As seen in Figure 3, there is a negative relationship in reproductive colonies between the total number of oocytes per bud and the percentage of those oocytes that have reached stage 3 in oogenesis, whereas there is no relation- ship in nonreproductive colonies. When both variables are plotted against the final eggs-per-bud phenotype, the data points fall in the vicinity of a line (data not shown). A Figure 2. Relationship between the eggs-per-bud phenotype at Milk- man stage 3 and the number of oocytes per bud at Milkman stage 0. See Results for details of oocyte staging. O, nonreproductive colonies (<0.3 eggs per hud): . reproductive colonies (>0.7 eggs per bud). Lines are least squares linear regression of reproductive colonies in A-C and all colonies in D; A: R = 0.756. B: R = 0.774. C: R = 0.568. D: R = 0.780. PQ oo 00 -o m -o ffl a 00 T3 CQ o 00, II L 5 10 15 20 25 Total Oocytes B - ^ 00 4 8 12 Stage 1 Oocytes _ 16 - O 00 , O 12 Stage 2 Oocytes D Tv X O 01234 Stage 3 Oocytes EGG PRODUCTION IN BOTRYLLUS 67 50 40 u oo 2 30 cd 8 o 20 10 5 10 15 20 25 Total Oocytes/Bud Figure 3. Relationship, at Milkman stage 0, between the total number ol oocytes per bud and the percentage of oocytes at stage 3 in oogenesis. I-me is least squares linear regression of reproductive colonies (R = 0.763). O, nonreproductive colonies (<0.3 eggs per bud); . reproductive colonies (>0.7 eggs per bud). principal component analysis (Table I) reveals that a single vector equally weighted for the three variables accounts for 84% of the variation. Discussion The strong relationship between a B. schloaseri colony's final eggs-per-bud phenotype and both the total number of oocNtes that are in the developing ovary and the number of oocytes at stage 3 of oogenesis indicates that two separate mechanisms operate to determine the final number of eggs that a colony produces. In reproductive colonies, a variable number of oocytes populate a developing ovary, a variable percentage of those oocytes reach stage 3 in oogenesis, and the final eggs-per-bud phenotype is determined by the neg- ative relationship between the two. Although the strong relationship between these two variables in reproductive colonies suggests that they may be genetically linked, the data from nonreproductive colonies demonstrate that the control of oogenesis can be uncoupled from the number of oocytes that populate a developing ovary. Control of the number of oocytes within u developing hud The development of the ovaries in B. schloxseri is not a one-time event, but occurs during each asexual cycle. Studies on the development of buds in Botryllus have demonstrated that germ cells are not seen in the devel- oping bud until after it is vascularized (Izzard, 1973; Mukai and Watanabe, 1976; Sabbadin and Zaniolo, 1979; Manni et 10 4 larvae per bowl), and an even layer of zooxanthellae was pipetted along the bottom of the bowls. Several drops of homogenized Anemia were added to the appropriate treatments. Zooxanthellae and Anemia slurry were removed either 4 or 24 h later (Table I) by concen- trating larvae on a filter and placing them into clean filtered seawater. Some larvae from each treatment were observed under a compound microscope, either immediately after zooxanthellae were removed or 24 h later, to determine if they had become infected with zooxanthellae. For treatments Cl and C2 (Table I), we used the follow- ing method to determine the fraction of larvae that became infected. Twenty-four h after the larvae were exposed to zooxanthellae, the water in the larval cultures was swirled and one aliquot was removed from each replicate. Between 25 and 56 larvae per aliquot were examined under a com- pound microscope to count how many contained zooxan- thellae. Lan'al development To observe and quantify the developmental progression of both azooxanthellate and zooxanthellate larvae, six rep- licate cultures of each were maintained in plastic 6-well culture dishes (300-500 larvae per well in 5 ml of filtered seawater). Water was changed roughly once a day. Larval development was monitored for about 2 weeks. Each rep- licate well was placed haphazardly under a dissecting mi- croscope, and within the field of view, the number of larvae at each developmental stage was counted. Electron microscopy To follow the process of zooxanthella incorporation into host tissue, larvae from treatment C2 were sampled and fixed for electron microscopy 1 and 24 h after zooxanthellae were added to larval cultures. The larvae were placed in sampling cups, which were prepared by cutting off the bottoms of microfuge tubes and affixing 50-/j,m mesh across the bottom. The cups were placed in 1 % glutaraldehyde in phosphate-buffered saline (PBS. 0.1 M sodium phosphate. 0.45 M sodium chloride, pH 7.2) for 1 h; rinsed 3x10 min in PBS: postfixed for I h in 1% osmium tetroxide in PBS: rinsed 3 X 10 min in PBS; and dehydrated for 15 min each in 30%, 50%, and 70% ethanol, and then 1 h each in 80%, 95%, and 3 X 100% ethanol. Samples for scanning electron microscopy were dried for 15 min in hexamethyldisilane. Table I Experimental treatments of Fungia scutaria lan-ae Treatment: Developmental stage Source of Algae Anemia added? Exposure duration Infection determined A: embryo-early planu'.a (0-12 h old) F. sciituna no overnight immediately B: early planula (1-2 days old) F. scutaria no overnight immediately Cl: fully developed planula (3 days old)* F. scutaria no 4 h after 24 h C2: fully developed planula (3 days old)* yes 4 h after 24 h Dl: fully developed planula (3 days old)* A. pallttla no 4 h after 24 h D2: fully developed planula (3 days old)* yes 4 h after 24 h El: fully developed planula (3 days old)* C. .xamachana no 4 h after 24 h E2: fully developed planula (3 days old)* yes 4 h after 24 h F: polyp (after metamorphosis) F. scutaria no overnight after 24 h * Planulae were considered fully developed once they had acquired the ability to feed. ONSET OF SYMBIOSIS IN FUNGIA SCUTAR1A 73 mounted on stubs, coated with 60:40 Au:Pd, and viewed on an Amray 3300FE scanning electron microscope. Samples for transmission electron microscopy were infiltrated with Spurr's resin in 1:1 ethanol:resin for 2.5 h, 1 :3 mix for 2.5 h, 2 X 100% resin for 1 h. and 100% resin overnight at 60C. Thin sections were prepared on an ultramicrotome, stained with uranyl acetate and lead citrate, and viewed on a Philips CM 12 transmission electron microscope. Polyacrylamide gel electrophoresis We prepared one-dimensional SDS-PAGE protein pro- files of both azooxanthellate and zooxanthellate larvae through development (eggs through 6-day-old larvae). For each sample, about 1000 larvae were counted, collected by centrifugation, and frozen at 80C. Protein extracts were prepared by homogenizing frozen larvae over ice in a ground-glass grinder in KM) p.\ of homogenization buffer (40 mM Tris-HCl, 10 mM EDTA. protease inhibitor cock- tail (Sigma). pH 7.4). Homogenates were centrifuged for 10 min at 14.000 X g to pellet zooxanthellae and animal debris. The protein concentration of the supernatant was deter- mined spectrophotometrically (Bradford. 1976); larvae con- tained approximately 50-100 ng protein/larva. Larval pro- teins were resolved on 12.5% SDS-PAGE gels under reducing conditions (methods modified from Laemmli. 1970). Gels were silver stained (methods modified from Heukeshoven and Dernick. 1986) and scanned on an Im- agernaster desktop scanner (Pharmacia) and analyzed using Irnagemaster software (Pharmacia). Results Larval development Larval development was observed over three summers (1994, 1996. 1997). Larvae from all years followed the same progression of developmental stages, as illustrated in Figure 1 A and detailed below in Figure 2, progressing from swimming to creeping to settled. The duration of each developmental stage, however, was variable; for the later stages it differed by up to several days both within and among replicates. Figure IB shows the time course of developmental events for zooxanthellate larvae from 1996. All larvae progressed through the following series of stages. Within 12 h after fertilization, slowly moving, cili- ated spherical planulae developed: within 24 h, barrel- shaped planulae, roughly 100 /nm in length (shown in Fig. 2 A), had developed and were actively swimming at all depths in the culture dishes. By day 3, larvae had fully formed mouths and functional gastric cavities, and were capable of feeding. Upon addition of food (homogenized Anemia), larvae ceased swimming and dropped to the bot- tom of the dish. They extruded mucus, their oral ends expanded, and they ingested whatever they landed on. in- cluding experimentally added zooxanthellae. As they fed, their gastric cavities became filled with participate matter (Fig. 2B). Some larvae resumed swimming while trailing a strand of mucus; the mucus trapped particulate matter that slowly entered the mouth. Larvae continued to feed for several hours and then resumed swimming. Except for Figure 1. Progression of developmental events in Fttn^ut \citttirui larvae. (A) Schematic representation of developmental stages from the early planula through metamorphosed polyp. (B) Example of the time course of developmental events. Data shown are from zooxanthellate larvae in 1996. Larvae were infected with zooxan- thellae on day 3 and then divided into six replicate dishes, which were monitored daily. Each point represents data pooled from the six replicates. Larvae progressed from swimming to creeping to settling. 74 J. A. SCHWARZ ET AL. figure 2. Light micrographs of stages in the development of Fiingia scularia larvae. (A) Two-day-old planula larva, prior to development of a mouth. (B) Three-day-old feeding planula (m = mouth, mf = mucous strand with food particles attached, z = zooxanthella). (C) Polyp with tentacles, 6 days after settling. Zooxanthellae are visible as golden spheres. Planula length and polyp diameter, approximately 100 /xm. zooxanthellae, all ingested paniculate matter was digested or expelled by the following day. When larvae were about 4 days old, they assumed a ball shape, ceased active swim- ming, and began creeping slowly over the substrate. Starting on day 5, the ball-shaped larvae began to settle. They spread out over the substrate and metamorphosed into volcano- shaped polyps, which began to develop tentacle buds sev- eral days after metamorphosis (Fig. 2C). Acquisition of zooxanthellae and onset of symbiosis Prior to the development of a functional mouth on day 3, planulae of F. sciituria did not become infected by experi- mentally added zooxanthellae. Once the mouth was func- tional, however, the planulae were able to acquire zooxan- thellae. When stimulated to feed, larvae indiscriminately ingested any particulate matter, including experimentally added zooxanthellae. Zooxanthellae either were ingested as part of a larger mass that was fully engulfed by the mouth, or they adhered to mucous strands that were ingested by the larvae. Figure 3A shows a zooxanthella adhered to a larval mucous strand, and Figure 3B shows several zooxanthellae surrounding and contained within the oral cavity of a larva. One hour after zooxanthellae were added, larvae were sam- pled and fixed for transmission electron microscopy. Figure 4 shows a representative planula 1 h postfeeding. in longi- tudinal section, with several algae resident in endodermal Figure 3. Scanning electron micrographs detailing zooxanthella acqui- sition by 3-day-old Fiingia scuraria planulae. (A) Feeding planula with zooxanthella adhered to mucous strand (m = mouth, z = zooxanthella). (B) Feeding planula. with multiple zooxanthellae entering the mouth. Larvae were fixed for electron microscopy 1 h after exposure to zooxan- thella isolates and homogenized Anemia (see Methods). Bars = 10 ij,m. ONSET OF SYMBIOSIS IN FUNGIA SCUTAKIA 75 ec Figure 4. Transmission electron micrograph of a longitudinal section through a Fungia scutaria larva infected with zooxanthellae. Thickened oral end at lower left. Zooxanthellae appear in the endoderm as dark spheres. Light ellipses, mostly in the ectoderm, are poorly preserved nematocysts. ec = ectoderm, en = endoderm, z = zooxanthella. Bar = 20 /im. cells. Micrographs suggest that zooxanthellae are phagocy- tosed by endodermal cells lining the coelenteron (Fig. 5 A, B) and appear in both endodermal (Fig. 5C) and ectodermal tissue (Fig. 5D). Although zooxanthellae were still present in ectoderm 24 h later, we did not determine how long zooxanthellae remained within the ectoderm or whether they eventually migrated into the endoderm or were di- gested or expelled from the host. Larvae were not limited to forming an association with a specific strain of zooxanthellae; planulae were capable of becoming infected by zooxanthellae isolated from F. scu- taria (Treatment C2) and Aiptasia pallida (Treatment D2), as well as by cultured zooxanthellae from Cassiopeia xani- acliana (Treatment E2) (see Table I). To determine whether the host had retained zooxanthellae, larvae from Treatments C2 and D2 were observed over a period of 10 to 14 days. Larvae that had acquired zooxanthellae on day 3 remained infected as they progressed through development and meta- morphosis into polyps. Infection by zooxanthellae was not required for metamor- phosis: both zooxanthellate and azooxanthellate larvae suc- cessfully settled and metamorphosed into polyps (Fig. 6). Larvae infected with zooxanthellae from F. scutaria (Treat- ment C2) and A. pallida (Treatment D2) both underwent metamorphosis (we did not monitor settlement for larvae infected with zooxanthellae cultured from C. xamachana). Aposymbiotic polyps were able to ingest experimentally added zooxanthellae via ciliary currents produced by the polyps that swept paniculate matter, including zooxanthel- lae, over and into their mouths. Observations over the 6 days following showed that the zooxanthellae were retained within the polyps throughout this period. The proportion of larvae that became infected by zoo- xanthellae isolated from adult F. scutaria (Treatment C) depended on the strength of the feeding response. Feeding was observed to be strongly stimulated (i.e.. virtually all larvae began to feed) by the addition of homogenized Ar- temia, but was also stimulated to a lesser extent (i.e., some larvae began to feed) simply by the addition of zooxan- thella-isolates, which contained residual animal host tissue. We quantified the effect of larval feeding strength on zoo- xanthella acquisition for treatments Cl (zooxanthellae alone) and C2 (zooxanthellae and Artemia). In the zooxan- thellae alone treatment, 25.0% 0.02% ( = 2) of larvae acquired zooxanthellae, whereas 96.8% 0.01% ( = 2) became infected when exposed to both zooxanthellae and Anemia. It was clear that larvae in Treatments D and E also became infected at a higher rate when exposed to both zooxanthellae and homogenized Anemia than to zooxan- thellae alone, although the results were not quantified. An experiment in 1996 provided preliminary evidence that symbiotic state may influence developmental events in F. scutaria. Zooxanthellate larvae settled and metamor- phosed earlier than azooxanthellate larvae, most of which became arrested in the "ball stage" and then eventually died (Fig. 7). However, the same experiment repeated in 1997 showed low rates of metamorphosis for both zooxanthellate and azooxanthellate larvae and no difference in the timing of metamorphosis. Lan'al protein profiles Protein profiles of larvae through development showed changes with the age of the larvae. Two bands, at 84 and 79 kilodaltons (kDa), were abundant in eggs and 1 -day-old larvae (Fig. 8A). As shown in Figure 8B. this protein doublet comprised a significant proportion (36%) of total protein in 1 -day-old larvae, but was almost absent by day 6. The apparent depletion of this protein corresponds to the onset of settlement and metamorphosis. The abundance of the putative yolk protein did not differ between 6-day-old azooxanthellate and zooxanthellate larvae. Discussion Lan'al development and acquisition of zooxanthellae Development in Fungia scutaria was similar to that re- ported in other broadcast-spawning species of coral (Bab- cock and Hey ward. 1986; review in Harrison and Wallace, 1990). Planula larvae had fully developed within 24 h after fertilization, which is within the range of one to several days reported for other species. Larvae of F. scutaria were about 100 /urn long, ciliated, and barrel-shaped; they exhibited active swimming behavior until they settled at an age of 5 76 J. A. SCHWARZ ET AL. \ ** ec ^,$Sr* *** * ec * ^M ,;,,'l Figure 5. Transmission electron micrographs of onset of symbiosis between Fiing/n xcutariti planulae and zooxanthellae. (A) Section through endoderm and gastric cavity of a planula showing initial contact between an endodermal cell and a zooxanthella. Host endodermal membranes are very closely associated with the alga. (B) Endodermal cell partially surrounding a zooxanthella, suggesting that the alga is being ingested by the host cell. (C) Zooxanthella resident within a vacuole in an endodermal cell. (D) Two zooxanthellae in gastric cavity (one is being phagocytosed) and one resident within a vacuole in an ectodermal cell, gc = gastric cavity ec = ectoderm, en = endoderm, z = zooxanthella. Bars = S jum. days to approximately 2 weeks. This appearance and be- havior is typical for externally developed planula larvae. Very little is known about the feeding ability or behavior of coral planulae. Although it appears that many species, particularly brooding species, produce a nonfeeding larva, the ability to feed has probably gone unrecognized in some species because rearing techniques generally do not expose larvae to a source of paniculate food. We found that the feeding behavior of F. scuturui was very similar to that reported for the temperate coral Caryophvllia smithi (Tranter et al., 1982) and for the temperate sea anemones Anthopleura elegantissima and A. xanthogrammica (Siebert. 1974; Schwarz, 1996). Feeding consisted of a mouth-opening response to the addition of ground animal tissue, as well as secretion of mucous strands that trapped participate matter for ingestion. Although most scleractinian coral species spawn azoo- xanthellate gametes that develop into azooxanthellate plan- ulae (review in Richmond, 1997), little is known about how planulae might acquire zooxanthellae from the environ- ment. The results of this study support the idea that for corals, competency for infection by zooxanthellae may gen- erally depend on the development of a functional mouth. We found that F. scitraria did not become infected by experimentally added zooxanthellae until after a mouth developed. Once the mouth was functional, all developmen- tal stages were competent to become infected. Reports of infection events in other species support this hypothesis species that are infected at the polyp stage appear to have a nonfeeding planula that does not develop a mouth until the polyp stage (Kinzie, 1974: Babcock and Hey ward, 1986; Benayahu et ai, 1989). Studies of the feeding behavior of planulae also support this hypothesis: planulae of both F. scuttiria (this study) and A. elegantissima (Schwarz. 1996) ONSET OF SYMBIOSIS IN FUNGIA SCVTARIA 77 N V^?^fes*^ i-i i-i P 2345 Figure 6. Light micrographs of newly settled polyps of Fungia scu- iiirin. (A) A/ooxanthellate polyp (m = mouth). (B) Zooxanthellate polyp. Zooxanthellae appear as brown spheres in the polyp. The two polyps shown in this figure were settled in the same dish, adjacent to one another. Contaminating diatoms appear as small ellipses around the polyps. Polyp diameter = 100 jxm. exhibit feeding behavior that leads to the ingestion of zoo- xanthellae. It will be interesting to determine whether other species that produce a feeding planula larva acquire zoo- xanthellae in the same manner as shown for F. scutaria and A. elegantissima. Both endodermal and ectodermal cells incorporated zooxanthellae within 1 h after larvae were exposed to zoo- xanthellae. The appearance of zooxanthellae in ectodermal tissue was surprising because zooxanthellae phagocytosed by endodermal cells would not be expected to be trans- ported into tissues where they do not ultimately reside. We did not determine how the zooxanthellae entered the ecto- derm. Future work will include long-term sampling of newly infected larvae to investigate the fate of the ectoder- mal zooxanthellae. Horizontal transmission of symbionts would appear to be disadvantageous for obligately symbiotic species because of 100 789 Larval age (days) 10 11 Figure 7. Effect of symbiotic state on larval settlement. Results from 1996 experiment. Nearly 100% of zooxanthellate planulae underwent settlement and metamorphosis by day 10. whereas most azooxanthellate planulae failed to settle. Each point represents data pooled from six replicate dishes for each treatment. 43 30- 20- kDa 34567 Larval age (days) Figure 8. Protein profiles and abundance of putative yolk protein in Fungia scutaria larvae. (A) Silver-stained ID SDS-polyacrylamide gel of total protein extracted from eggs (lane 2), 1-day-old larvae (lane 3), 6-day-old azooxanthellate larvae (lane 4), and 6-day-old zooxanthellate larvae (lane 5). Each lane contained 1.2 fig protein. Molecular weight standards in lane 1. Arrow highlights a putative yolk protein doublet (84 and 79 kDa) that is abundant in eggs and 1-day larvae but absent by day 6. ( B I Decline in abundance of putative yolk protein through larval develop- ment. The depletion of putative yolk protein corresponded with the onset of settlement on day 7. There was no difference in putative yolk protein abundance in azooxanthellate and zooxanthellate larvae (days 5 and 6): data shown represent the average of the two treatments. the possibility that infection may not occur. However, for planulae dispersed to areas with different environmental conditions, the ability to acquire zooxanthellae from the environment might confer a greater advantage to the host than directly inheriting maternal zooxanthellae. This study found that planulae of F. scutaria were capable of forming an association with members from three clades of zooxan- thellae classified by Rowan and Powers (199 la, b); zoo- xanthellae from C. xamachana are in group A, those from A. pallida are in group B, and those from F. scutaria are in group C. The degree to which zooxanthellae from different clades persist in F. scutaria remains to be investigated, but our results suggest considerable flexibility in host-symbiont specificity in this species. In contrast, planulae of A. elegan- tissima, although able to form an association with zooxan- 78 J. A. SCHWARZ ET AL thellac recently isolated from a conspecific adult, were unable to do so with cultured S. californium, which is the species reported to occur in A. elegantissima (Banaszak et til., 1993: Schwar/,. 1996). The finding that a stronger larval feeding response re- sulted in higher rates of infection indicates that larval feed- ing behavior may play an important role in acquiring zoo- xanthellae from the ambient environment. Because so little is known about the distribution and abundance of zooxan- thellae in the natural environment, it is difficult to speculate on potential sources of these symbionts. However, one source that is likely to occur in abundance is mucus expelled by corals. Cnidarian hosts regularly expel mucus containing high concentrations of zooxanthellae (Steele, 1975; McClo- skey et al., 1996; Schwarz. pers. obs.), and increased rates of expulsion have been reported to accompany spawning (Montgomery and Kremer, 1995; D. Krupp, pers. obs.). Although this study did not examine whether planulae of F. saitaria will feed on coral mucus, planulae of the sea anemone Anthopleura elegantissima did feed on mucus expelled by adults and became infected by the zooxanthel- lae within it (Schwarz, 1996). These results suggest that ingestion of zooxanthellae could occur either at the spawn- ing site or at the sites in which the larvae ultimately settle, allowing them to acquire symbionts adapted to different environments. Effect of symbiont acquisition on lan-al development Zooxanthellae are known to affect the physiology of their adult hosts, and the acquisition of zooxanthellae by larval hosts probably influences larval development. For example, the acquisition of symbionts may act as a settlement cue. An experiment in 1996 demonstrated that zooxanthellate larvae settled earlier than azooxanthellate larvae (Fig. 6) indeed, most azooxanthellate larvae failed to settle. However, the same experiment repeated the following year showed no differences in settlement (data not shown). It is possible that symbiotic state does influence developmental events but either acts in concert with, or is overridden by, environmen- tal variables such as temperature. The 1996 experiment was conducted during a period of anomalously warm water temperatures that induced a bleaching event on the reef flat in Kaneohe Bay. whereas the 1997 experiment was charac- terized by normal temperatures. Thus larval development may have been influenced more strongly by water temper- ature than by symbiotic state. Experimental manipulation of environmental parameters will allow us to examine this question in more detail. Potential effect of symbiont acquisition tin larval energetic strategies and dispersal The larval stage serves as a means for dispersal in the life histories of sessile marine invertebrates. The lensth of the larval stage depends in part on the amount of energy avail- able for metabolism (Boidron-Metairon, 1995; Levin and Bridges, 1995). Larvae of F. scutaria have several potential sources of energy that may allow them to extend the larval stage sufficiently to explain their widespread occurrence throughout the Pacific. First, larvae may initially obtain nutrition from yolk protein supplied through the egg. The presence and the pattern of decline of two abundant 84 kDa and 79 kDa proteins and the correlation between their depletion and the onset of settlement suggest that larvae may metabolize this protein over the course of their devel- opment. Second, once the mouth has developed, larvae may obtain energy through feeding. Third, larvae that have ac- quired zooxanthellae may receive nutrition in the form of organic carbon translocated by zooxanthellae. Richmond (1981, 1987) demonstrated that symbiotic planulae of the coral Pocillopora damicornis received about 13%-27% of the carbon fixed by zooxanthellae. Each of these modes of nutrition may operate at different times in development, and each may function to extend the length of the dispersal stage. Acknowledgments This work was supported by grants from the Office of Naval Research (NOOU149710101) to V. M. W.. and from Sigma Xi and Oregon State University Zoology Department to J. A. S. We thank P. Jokiel, B. Kinzie, F. Cox, and the staff of the Hawaii Institute of Marine Biology for facilities, zooxanthella cultures (B. Kinzie), and support. We thank the OSU Department of Botany and Plant Pathology Elec- tron Microscope facility staff for their assistance. Rick Jones prepared the illustration for Figure 1. This is Contri- bution #1043 from the Hawaii Institute of Marine Biology. Literature Cited Babcock, R. C., and A. J. Heyward. 1986. Larval development of certain gamete-spawning scleractinian corals. Coral Reefs 5: 1 1 1-1 16. Banaszak, A. T., R. Iglesias-Prieto, and R. K. Trench. 1993. Scrip- sie/la velellae sp. nov. (Peridiniales) and Gloeoedinium viscitm sp. nov. (Phytodiniales). dinoliagellate symbionts of two hydrozoans (Cni- daria). J. Phycul. 29: 517-528. Benayahu, Y., Y. Achituv, and T. Berner. 1989. Metamorphosis of an octocoral primary polyp and its infection by algal symbiosis. Symbiosis 7: 159-169. Boidron-Metairon, I. F. 1995. Larval nutrition. Pp. 223-238 in Ecology of Marine Invertebrate Larvae. Larry R. McEdward. ed. CRC Press. Boca Raton, FL. Bradford, M. B. 1976. A rapid and sensitive method for the quuntitation of mierogram quantities of protein utili/ing the principle of protein-dye binding. Anal. Biochem. 72: 248-254. Colley, N. J.. and R. K. Trench. 1983. Selectivity in phagocytosis and persistence of symbiotic algae by the scyphistoma stage of the jellyfish Cassiopeia xmiiuchana. Proc. R. Six. Lund. K 219: 61-82. Davy, S. K., I. A. N. Lucas, and J. R. Turner. 1997. Uptake and persistence of homologous and heterologous zooxanthellae in the tem- ONSET OF SYMBIOSIS IN FUNG1A SCUTAKIA 79 perate sea anemone Cereu\ pedunculate (Pennant). Biol. Hull. 192: 208-216. M.m-l.is. A. K. 1994. Symbiotic Interaction!,. Oxford University Press, Oxford. US pp. Fadlallah. Y. H. 1983. Sexual reproduction, development and larval biology in sclcractmian corals. Coml Reefs 2: 129-150. Kilt. \V. k.. and K. K. Trench, I983a. Infection of invertebrate hosts with the symbiotic dinoflagellate Symbiodinium microadriaticum. En- doc\tobii>. 2: 675-681. Fitt. \V. K., and R. K. Trench. 1983h. Endocytosis of the symbiotic dinoflagcllatc Symbiodinium microadriaticum Freudenthal by endoder- mal cells of the scyphistomae of Cassiopeia \amai liana and resistance of the algae to host digestion. ./. Cell Sci. 64: 195-212. Harrison. P. I,., and C. C. Wallace. 1990. Reproduction, dispersal and recruitment of scleractinian corals. Pp. 133-207 in Ecosystems of the World: Coral Reefs. Vol. 25. Z. Dubinsky. ed. Elsevier. New York. Heukeshoven, J.. and R. Dernick. 1986. Neue Ergebnisse zum Mecha- nismus der Silberfarbung. Electron/lores Forum 86: 22-27. kinzie. R. A. 1974. Experimental infection of aposymbiotic gorgonian polyps with /ooxanthellae. J. Exp. Mar. Bio/. Ecol. 15: 335-345. Krupp. I). A. 1983. Sexual reproduction and early development of the solitary coral Fungiii scularia (Anthozoa: Scleractima). Coral Reefs 2: 159-164. Laemmli. I'. K. 1970. Cleavage of structural proteins during the assem- bly of the head of bactenophage T4. Nature 227: 680-685. Levin, L. A., and T. S. Bridges. 1995. Pattern and diversity in repro- duction and development Pp. 1 18 in Ecology of Marine Invertebrate Lun-ae. Larry R. McEdward. ed. CRC Press. Boca Raton. FL. McCloskey, L. R., T. G. Cove, and E. A. Verde. 1996. Symbiont expulsion from the anemone Anthopleura elegantissima (Brandt) (Cni- daria: Anthozoa). J. Exp. Mar Biol. Ecol. 195: 173-186. Montgomery, M. k., and P. M. Kremer. 1995. Transmission of sym- biotic dinoflagellates through the sexual cycle of the host scyphozoan Liintche unxtticulatu. Mar. Biol. 124: 147-155. Richmond. R. H. 1981. Energetic considerations in the dispersal of Pocillopora damicornix (Linnaeus) planulae. Proc. 4th Int. Coral Reef Symp.. Manila 2: 153-156. Richmond. R. H. 1987. Energetics, competency, and long-distance dis- persal ol planula larvae of the coral Pocillopora damicomis. Mar. Biol. 93: 527-533. Richmond, R. H. 1997. Reproduction and recruitment in corals: critical links in the persistence of reefs. Pp. 1 75-197 in Life and Death of Coral Reefs. C. Birkeland. ed. Chapman and Hall. New York. Richmond, R. H., and C'. L. Hunter. 1990. Reproduction and recruit- ment of corals: comparisons among the Caribbean, the tropical Pacific, and the Red Sea. Mar. Ecol. Prog. Ser. 60: 185-203. Ro\van, R.. and N. kiumltmi. 1995. Intraspecific diversity and ecolog- ical zonation in coral-algal symbiosis. Proc. Natl. Acad. Sci. USA 92: 2850-2853. Rowan. R., and D. A. Powers. 1991a. Molecular genetic identification of symbiotic dinoflagellates (/ooxanthellae). Mar. Ecol. Prog. Ser. 71: 65-73. Rowan, R., and I). A. Powers. 1991h. A molecular genetic classification of zooxanthellae and the evolution of animal-algal symbioses. Science 251: 1348-1351. Schwarz, J. A. 1996. Feeding behavior and acquisition of zooxunthellae by the planula larvae of the temperate sea anemone Anthopleura elegantissima. Masters Thesis, University of California Santa Cruz. Seihert, A. E. 1974. A description of the embryology, larval develop- ment, and feeding of the sea anemones Anthopleura e/egantissima and .4. xanthogrammica. Can. J. Zoo/. 52: 1383-1388. Steele, R. D. 1975. Stages in the life history of symbiotic zooxanthellae in pellets extruded by its host Aiplasia tagetes (Duch. and Mich.) (Coelenterata. Anthozoa). Biol. Bull. 149: 590-600. Sugiura. Y. 1964. On the life history of rhizostome medusae. II. Indis- pensability of zooxanthellae. Embryologia 8: 223-233. Tranter. P. R. G., D. N. Nicholson, and D. Kinchington. 1982. A de- scription of spawning and post-gastrula development of the cool tem- perate coral, Caryophyllia smithi. J. Mar. Biol. Assoc. UK. 62: 845-854. Trench, R. K. 1980. Integrative mechanisms in mutualistic symbioses. Pp. 275-279 in Cellular Interactions in Symbiosis and Parasitism. C. B. Cook, P. W. Pappas, and E. D. Rudolph, eds. Ohio State University Press, Columbus, OH. Trench, R. K. 1987. Dinoflagellates in non-parasitic symbioses. Pp. 530-570 in The Biology of Dinoflagellates. F. J. R. Taylor, ed. Black- well Scientific, Oxford. Reference: Bio/. Bull. 196: 80-93. (February, 1999) Molecular Determination of Species Boundaries in Corals: Genetic Analysis of the Montastraea annularis Complex Using Amplified Fragment Length Polymorphisms and a Microsatellite Marker JOSE V. LOPEZ', RALF KERSANACH, STEPHEN A. REHNER 2 , AND NANCY KNOWLTON 3 Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Republic of Panama Abstract. Analyses of DNA have not been widely used to distinguish coral sibling species. The three members of the Montastraea annularis complex represent an important test case: they are widely studied and dominate Caribbean reefs. yet their taxonomic status remains unclear. Analysis of amplified fragment length polymorphisms (AFLPs) and a microsatellite locus, using DNA from sperm, showed that Montastraea faveolata is genetically distinct. One AFLP primer yielded a diagnostic product (880 bp in M. faveolata, 920 bp in M. franksi and M. annularis) whose homology was established by DNA sequencing. A second primer revealed a 630 bp band that was fixed in M. faveolata. and rare in M. franksi and M. annularis: in this case homologies were confirmed by Southern hybridizations. A tetranucle- otide microsatellite locus with several alleles exhibited strong frequency differences between M. faveolata and the other two taxa. We did not detect comparable differences between M. annularis and M. franksi with either AFLPs (12 primers screened) or the microsatellite locus. Comparisons of AFLP patterns obtained from DNA from sperm, somatic tissues, and zooxanthellae suggest that the technique rou- tinely amplifies coral (animal) DNA. Thus analyses based Received 5 June 1998; accepted 1 December 1998. 1 Current address: Division of Biomedical Research, Harbor Branch Oceanographic Institution. 5600 U.S. 1 North. Ft. Pierce. FL 34946; E-mail: Lopez@hboi.edu 2 Current address: Department of Biology, P.O. Box 23360, University of Puerto Rico, Rio Piedras, San Juan, Puerto Rico, 00931. 1 Also at Marine Biology Research Division 0202. Scripps Institution of Oceanography, University of California San Diego, La Jolla, CA 92093- 0202; E-mail: nknowlton@uscd.edu Abbreviations: AFLP - amplified fragment length polymorphism, a registered trademark of Keygene. on somatic tissues may be feasible, particularly after diag- nostic differences have been established using sperm DNA. Introduction The recognition of species boundaries in sympatry is straightforward in principle, because the absence of inter- breeding implies the existence of at least some fixed genetic differences between taxa (Avise and Ball, 1990). However, the number of such differences may be very small if the isolation of taxa is recent or the rate of evolution is slow. If in addition sporadic hybridization occurs, the problem of defining species becomes particularly difficult (e.g., Howard et al., 1997). Closely related coral species appear to be especially challenging in this regard (Veron. 1995; Knowlton and Weigt, 1997). Species boundaries are in flux for a number of well-studied groups (e.g., Miller and Babcock, 1997; Miller and Benzie, 1997; Odorico and Miller, 1997; Willis et al.. 1997; Knowlton and Budd, unpubl.), and it is unclear whether these controversies are due to the technical chal- lenge of finding diagnostic characters between generally similar but reproductively isolated taxa, or alternatively, to the blurring of species boundaries by hybridization (Veron, 1995; Knowlton and Weigt, 1997; Willis et al., 1997). Molecular methods have great potential to resolve the na- ture of species boundaries because of the large number of unambiguous characters they provide (Avise. 1994). A clear example of these issues is presented by the proposed members of the Montastraea annularis species complex: M. annularis (formerly morphotype I or columnar morph). M. faveolata (formerly morphotype II or massive morph). and M. franksi (formerly morphotype III or bumpy morph) (Knowlton et al., 1992; Van Veghel and Bak, 1993: 80 GENETIC ANALYSIS OF MONTASTRAEA 81 Weil and Knovvlton, 1994). In sympatry, these taxa differ in colony morphology, growth rate, stable isotope chemistry, aggressive behavior, allozymes. corallite structure, and life history (Tomascik. 1990: Van Veghel and Bak, 1993. 1994: Van Veghel, 1994; Van Veghel and Kahmami. 1994; Weil and Knovvlton, 1994; Van Veghel and Bosscher, 1995; Van Veghel c/ al.. 1996; Szmant et al.. 1997; Knowlton and Budd. unpubl.). Such concordance of suites of independent characters in sympatric taxa strongly suggests reproductive isolation (Avise and Ball. 1990). and differences in the timing of spawning and apparent barriers to interspecific fertilization (Knowlton et al.. 1997) also support this inter- pretation (but see Szmant et al.. 1997). Overall, these data support separate species status regardless of the species concept used (Templeton, 1989: Cracraft. 1989; Mallet, 199?; Knowlton and Weigt, 1997). Nevertheless, a preliminary molecular survey revealed no fixed DNA sequence differences among these taxa in two regions that might, a priori, be expected to have them: the ITS regions of rDNA and an intron in a /3-tubulin gene (Lopez and Knowlton. 1997). Sequence-based methods can only be used to examine a limited stretch of DNA, however, and methods that screen a larger proportion of the genome appear to offer greater promise (Lopez and Knowlton, 1997). One such approach is analysis of amplified fragment length polymorphisms (AFLPs), which screens for poly- morphisms at. or adjacent to. restriction endonudease sites (Zabeau and Vos, 1995). In a preliminary survey, we found evidence for potentially diagnostic differences between M. faveolata and M. franksi using two AFLP primers (Lopez and Knowlton. 1997). In the present study we wished to ( 1 ) determine whether these apparently diagnostic AFLP differences hold up when sample sizes are increased. (2) screen additional AFLP primers to see if any show promise for distinguishing M. anniilurix and M. franksi, (3) determine, using Southern hybridization and DNA sequencing, whether apparently similar AFLP bands are indeed homologous, and (4) assess whether diagnostic polymorphisms detected using high- quality DNA derived from sperm could also be seen in more readily collected, but potentially less pure, somatic tissue samples. During earlier work on tubulin introns (Lope/ and Knowlton, 1997), we also uncovered a tetranucleotide mi- crosatellite locus (here called Mfra-gtttl) in a genomic clone derived from M. franksi. Microsatellite or simple repeat loci have become increasingly important tools in evolutionary and population studies because of their high levels of polymorphism and codominant inheritance (Jarne and Lagoda, 1996). Here we report on evidence for allelic frequency differences at this locus among the Montastraea taxa. Materials and Methods Sample acquisition and DNA preparation All corals were collected from the San Bias Islands, Panama. Colonies were identified to species in the field, based on colony morphology, and brought to waters near the laboratory shortly before the anticipated date of spawning (Knowlton et ai. 1997). At dusk, each colony was placed in a separate container: spawning generally occurred 2-4 h after sunset, and the gamete bundles were collected imme- diately after release. The gamete bundles from each con- tainer were washed separately over plankton netting. The eggs were retained on the netting, while the sperm passed through with the wash water, which was collected and centrifuged. The pelleted sperm were quick frozen (details in Lopez and Knowlton, 1997). Abundant DNA (hundreds of micrograms) was extracted from 1-2 ml of highly con- centrated sperm solution using standard techniques (Sam- brooks et al.. 1989), as previously described (Lopez and Knowlton. 1997). Sperm provide an ideal source of coral DNA (McMillan et al.. 1988), but they cannot be collected routinely. How- ever, high molecular weight DNA is difficult to extract from somatic tissues (McMillan et al., 1988) and may be con- taminated by DNA from symbiotic dinoflagellates (zooxan- thellae), which the gametes of Montastraea lack (Szmant, 1991). To determine whether DNA extracted from somatic tissues is of sufficient quality for AFLP analyses, we com- pared the analyses of DNA from sperm with those of DNA from somatic tissues from the same colonies. DNA from somatic samples was extracted according to the protocol of Rowan and Powers ( 1991 ), except that tissue was removed from 25-50 cm 2 of coral with an airbrush at 75-100 psi and suspended in 5-20 ml of L buffer ( 100 mM EDTA, 10 mM Tris-Cl. pH 7.6). To enrich for coral (animal) DNA within somatic tissues, frozen samples were ground in a glass homogenizer 5-10 times and centrifuged in an RT6000B Sorvall centrifuge at 50-100 X g for 10 min at room temperature. This spin was repeated one or two times for samples especially rich in zooxanthellae. The animal-en- riched DNA was then incubated for 3 h in 20-50 /j.g/ml proteinase K with \7c SDS (final concentration), followed by successive phenol :chloroform extractions (Sambrook et al.. 1989). DNA that remained resistant to restriction diges- tions was further purified by the GeneClean (Bio 101) protocol. To clarify further the potentially confounding con- tribution of zooxanthellae in coral somatic samples, we also analyzed zooxanthella DNA provided by Rob Rowan. This DNA came from other colonies of Montastraea (primarily M. faveolata) from the same region, and was not necessarily entirely free of coral (animal) DNA. 82 AFLP-PCR J. V. LOPEZ ET AL. The AFLP method and preparation of templates using the Pst I adapter system have been described in detail (Mueller et al. 1996). Genomic DNA was cut at specific 6-base recognition sequences by the Pst I restriction enzyme, and then^a synthetic, 21 bp adapter was ligated to the ends of the fragments. The polymerase chain reaction (PCR) was then used to amplify these restriction fragments, using primers matching the adapter sequence. To limit the number of different" fragments that are amplified (and hence improve the clarity of the resulting products), several additional, arbitrarily chosen bases were added to the PCR primers at their 3' ends. These additional bases (by which primers are identified, e.g., ATG or GGAG) overlap with genomic DNA beyond the restriction site, and amplify the subset of frag- ments that contain the additional nucleotides. We used the same methods and extension primers (ATG. GGAG) as previously reported (Lopez and Knowlton, 1997; note, however, that in our earlier report, the GGAG primer was incorrectly listed as GAG). We also used primers with the following 3' extensions: ATT, GAC, GTG, ATC. TGT, ACT. TTG, AGC, TAG. and ACGC. The PCR profile for all AFLP extension reactions was 94C/45 s, 60C/60 s, and 72C/90 s for 30 cycles, using an MJ Research PTC- 100 or PTC-200 thermocycler. Typically, a "preamplification" PCR was performed with an extension primer possessing one additional base (A, C, G or T) (Vos el at.. 1995). This reaction enriches for the subset of amplifiable templates possessing the extra nucleotide, improves the targeted band signal, and reduces background. The preamplification PCR was run with the same AFLP profile as above, but with fewer (20) cycles. However, this preamplification protocol did not improve the clarity of the patterns for the GGAG primer. The best electrophoretic resolution of PCR products was obtained with 1.2%-1.4% agarose/TBE gels (contain- ing at least 50% Metaphor agarose, FMC) run at 5.4 V/cm. The agarose-based technique used here and in our previous study does not require radioactive nucleotides, is less toxic, and is relatively easy to perform (Mueller et al.. 1996; Lopez and Knowlton. 1997). although it yields fewer dis- crete bands per lane (6-12) than the original poly aery 1- amide gel electrophoresis (PAGE) method (Vos et al.. 1 995 ) due to poorer resolution of fragments of less than 400 bp. All AFLP analyses shown here were performed more than once to ensure reproducibility, and AFLP-PCRs with no DNA added served as controls for contamination. AFLP banding patterns and DNA sequences (see below) were analyzed with RFLPscan (Scanalytics), BLAST (Alt- schul et al.. 1990). and MAPD (Yuhki and O'Brien. 1990; Stephens et al.. 1992). Only bands in the 0.3-1.6 kb size range were considered, since variation in higher molecular weight bands is more difficult to interpret due to potentially inconsistent amplification of large DNA fragments and the possibility of incomplete restriction enzyme digestion. RFLPscan (Scanalytics) converted band patterns into binary presence/absence characters for each sample and computed distance estimates for pairwise comparisons based on band sharing (Nei and Li. 1979). Cloning, Southern hybridisation, ami sequencing Both gel-purified and cloned AFLP fragments were used as hybridization probes and as sequencing templates. Spe- cific AFLP bands were dissected from low melting temper- ature aearose (NuSieve/Metaphor, FMC) gels, added to 200 ,11! of distilled H : O. and melted at 65C. When this DNA was used as a template in a "re-amplification" PCR with the original primer to obtain more material, PCR products were visualized on an agarose minigel to verify that a single band of the correct size had been amplified. Alternatively, AFLP fragments were cloned directly into pGEM-T vectors (Pro- mega). Probes for Southern hybridizations were labeled with a[ 32 P]-dCTP via nick-translation to a specific activity of 10"-10 7 cpm//ag (Sambrook et al., 1989). After separation on agarose gels. AFLP products were blotted onto nylon membranes (Duralon, Stratagene) ac- cording to standard procedures (Southern, 1975; Sambrook et al.. 1989). Higher molecular weight fragments (> 1 kb) appeared to be transferred to membranes less efficiently than smaller fragments, probably because of the relatively high gel concentrations of agarose (1.2%-1.4%) that were used. After hybridization and stringent washing (in 0.1 X SSC and 0.5% SDS at 50C) (Sambrook et al.. 1989), filters were exposed to Kodak XAR-2 X-ray film, generally for 2-3 days. Informative fragments that were generated using the GGAG primer were either directly sequenced after cleaning the PCR product with QIAquick PCR purification kits (Quiagen), or sequenced from plasmid-cloned fragments purified with Wizard miniprep kits (Promega). Sequencing reactions were run on automated DNA sequencers (ABI 373A or 377. Perkin Elmer), initially using primers com- plementary to T7 or SP6 promoter regions, following the standard cycle sequencing protocols (ABI, Perkin Elmer) used previously (Lopez and Knowlton. 1997). The follow- ing primers were then designed and used to obtain complete sequences for the GGAG 880 and 920 bp fragments; 5' CCCTGATCAGTATTTTGGG 3' (880i), 5' TTGGAATA- TTTGCCTTACCG 3' (880f), and 5' GGAGGGCTCTGT- TATTCTATC 3' (880r). The 880f primer (slightly internal to 880i) matches available Montastraea sequences, and when used with primer 880r yielded products of 837 or 804 bp. Microsatellite analyses The microsatellite locus Mtra-gtttl was initially detected in a clone (tub29A) derived from M. franksi (no. 426) that was recovered while we were screening for taxonomically GENETIC ANALYSIS OF MONTASTRAEA 83 informative /3-tubulin introns (Lope/, and Knowlton, 1997). Its occurrence and polymorphism in other Montastraea species were determined by designing the following 2 oli- gonucleotide primers, which are complementary to the genomic sequences flanking Mt'ra-gtttl: Sputlf-5' AAACA TACGG CCAGT GCTGG 3' and Sput2rc - 5' GAAAA GAGCA ATCTT TTGTA TGGTG 3'. The PCR profile used for Mfra-gtttl amplification from genomic DNA was 94C/40 s. 60C/45 s, and 72C/60 s for 30 cycles. All PCRs shown here were reproducible and included negative controls. The resolution of PCR products was better when 4.0% agarose (Metaphor. FMC) TBE gel electrophoresis was used, and banding patterns were confirmed by poly- acrylamide gel (10%) electrophoresis using the entire PCR product (approximately 1 /u.g DNA). Results AFLP hand patterns Band patterns produced using the GGAG primer showed a clear diagnostic difference between M. faveolata and M. franksi (Fig. 1A). which confirms our previous results ob- tained with smaller sample sizes (Lopez and Knowlton, 1997). The 920 bp GGAG band was absent from, and the 880 bp band was present in. all 16 M. faveolata tested (including 6 previously analyzed), while the reciprocal pat- tern occurred in 15 M. franksi (including 7 previously analyzed). A third band, migrating at around 850 bp, may also occur at significantly different frequencies in the two taxa, but our sample sizes are too limited to test for this. The ATG primer also provided evidence of genetic dif- ference between M. faveolata and M. franksi: as previously reported (Lopez and Knowlton, 1997), the 630 band was characteristically present in the former and absent from the latter (Fig. IB). In this case, however, increasing the sample size indicated that this difference between the species is not fixed: the 630 bp band was present in all 16 individuals of M. faveolata tested ( including 7 from the previous study ). but it also appeared in one individual of M. franksi (no. 19, lane 9). The remaining 14 M. franksi (including 6 previ- ously studied) lacked this band. In contrast to the clear differences separating M. faveo- lata from M. franksi, no diagnostic bands separated M. franksi and M. anmdaris. This was true, not only for the ATG (Fig. IB) and GGAG primers (five M. annul aris analyzed, data not shown), but also for 10 additional prim- ers that were screened (data not shown). The ATT primer yielded band patterns with the strongest quantitative differ- ences (Fig. 2). but the differences are not statistically sig- nificant by a chi-square test, once a Bonferroni correction (Rice. 1989) for the total number of bands examined is applied (see legend Fig. 2). Moreover, mean average per- cent difference (MAPD; Yuhki and O'Brien, 1990) in ATT band-sharing values among samples of M. anmdaris (28%) and among samples of M. franksi (22%) were very similar to the value calculated for interspecific comparisons be- tween these taxa (27%). Homology of bands Southern hybridization with DNA probes derived from the 630 bp ATG bands provided further insights into the nature of genetic differences between M. faveolata and M. franksi. In general, results were better when the hybridiza- tion probes were derived from DNA clones. Multiple bands were labeled when probes were derived from gel-isolated ATG bands, suggesting that the 630 bp bands were contam- inated with fragments of different molecular weight. Southern hybridizations showed that the 630 bp ATG bands found in all M. faveolata and one M. franksi (no. 19) (Fig. 3A) are homologous (Fig. 3B). Moreover, another M. franksi (no. 20) possessed a higher molecular weight frag- ment that also hybridized to the 630 bp ATG probe (Fig. 3B). This larger fragment may be the same one visible in Figure 3A, and probably arose by one or more DNA inser- tions at the 630 bp locus. Unfortunately, similar Southern hybridizations using the diagnostic 880 GGAG fragment as a probe were unsuccessful due to our inability to use the preamplification protocol. We therefore used DNA sequences to evaluate the ho- mologies of the 880 and 920 bp GGAG bands. Partial DNA sequences were initially obtained from both 5' and 3' ter- mini of the GGAG 880 (M. faveolata) and GGAG 920 (M. franksi) fragments. These preliminary sequences permitted the design of PCR primers by which the corresponding locus was amplified from genomic DNA. The GGAG 880 locus was amplified consistently from samples of M. faveo- lata (Fig. 4A), but a larger band appeared for M. franksi (Fig. 4A) and M. annularis (data not shown) when the same primers were used. DNA sequencing confirmed the homol- ogy of PCR products for 9 M. franksi, 6 M. anmdaris, and 7 M. faveolata (sequences deposited in Genbank AF1 101 14-AF1 10129, API 12346- API 12351). Three inser- tions or deletions (22, 8, and 3 bp) constituted the primary differences between M. faveolata and the other two taxa (Fig. 4B), as would be expected from the estimated size differences in the 880 and 920 bp bands. There were also 5 nucleotide substitutions [4 transitions, 1 transversion in 837 bp within the GGAG 920 fragment (see methods); data not shown] that distinguished M. faveolata from M. annularis and M. franksi. The sequences exhibited d(AT) contents of 58%-64%, and did not resemble sequences in current da- tabases (GenBank and EMBL; April 1998). The lack of significant open reading frames in the sequences suggests that they do not represent protein-encoding regions. 84 J. V. LOPEZ ET AL. A. -GGAG M. faveolata M. franksi bp 880 920 B. A/, faveolata -ATG M. franksi M. annularis bp 630 Figure 1. APLP hand patterns. Samples are grouped by species, and individual sample numbers are indicated above each lane. A 1.0 kb ladder (Gibco/BRL, Bethesda) was used as the molecular weight standard. (A) AFLP patterns derived with the GGAG primer. Species-specific bands at 880 and 920 bp are indicated by arrows. (B) AFLP patterns derived with the ATG primer. The 630 bp band is indicated by an arrow. GENETIC ANALYSIS OF MONTASTRAEA 85 -ATT M. annularis M. franksi Figure 2. Comparison ot Montastraea franksi and M. annularis AFLP patterns obtained with the ATT primer. Polymorphic bands showing the greatest frequency differences between the species are marked by asterisks. The most extreme frequency difference (band indicated by upper asterisk, present in 5 of 13 versus 12 of 13 individuals of M. annularis and M. franksi, respectively; not all samples shown) was individually significant (chi-square = 6.1. P < 0.02). However, this difference is not significant when a Bonferroni correction (Rice, 1989) for the total number of bands (18) is applied (P must be less than 0.003). Comparisons of band patterns from gametes, somatic tissues, and zooxanthellae The ATG patterns for DNA derived from sperm and from somatic tissue were generally consistent, and the taxonom- ically informative 630 bp band was conspicuous in analyses of somatic tissues from M. faveolata (Fig. 5A). Reproduc- ibility for the GGAG primer was poorer due to our inability to use the preamplification protocol, but diagnostic GGAG bands at 920 and 880 bp were visible in analyzed samples of DNA from somatic tissues from M. franksi and M. faveo- lata, respectively (Fig. 5B). This suggests that AFLP anal- yses can be informative with somatic tissues, especially once diagnostic patterns have been established with sperm samples. The general lack of higher molecular weight AFLP bands from analyses of somatic tissue compared to those of gamete samples may be due to degradation during DNA purification of somatic samples (e.g., McMillan el ai, 1998) or to the presence of contaminants that interfered with the reactions, but these bands were typically not scored. Some differences between somatic and gamete samples (e.g., for M. franksi no. 467 in Fig. 5 A) cannot currently be ex- plained. To determine how zooxanthella-derived bands in par- ticular might confuse the interpretation of analyses of somatic tissues, we obtained AFLP bands from DNA purified from zooxanthella types A, B, and C from Mon- tastraea (see Rowan and Knowlton. 1995) (Fig. 6). There may be some potential for confusion between the diag- nostic 630 bp band in M. faveolata and similarly-sized bands from zooxanthella types A and B, although these zooxanthella bands may in fact be due to coral (animal) contamination. In general, the similarity of the gamete and somatic tissue samples (Fig. 5) and the difference between zooxanthella-enriched and zooxanthella-absent (sperm) samples (Fig. 6), suggest that for these corals, the AFLP technique primarily amplifies coral (animal) DNA from somatic tissue samples. Microsatellite locus Analysis of a clone from M. franksi derived from a PCR amplification product using primers for (3-tubulin revealed a microsatellite locus (Mfra-gtttl) whose core repeat sequence (GTTT) was perfectly repeated 9 times (EMBL accession number AJ223626). It is similar (but not identical) to simple repeats in other scleractinian corals (McMillan et ai, 1991). Analysis of additional samples using the same primers revealed that a smaller allele (approximately 160 bp, its size presumably due to 86 J. V. LOPEZ ET AL A. M. faveolata M. franksi B. M. faveolata M. franksi Figure 3. Southern hybridi/ution experiments with the ATG 630 bp fragment to determine band homolo- gies. (A) Ethidium bromide-stained agarose gel used for Southern blotting, showing typical AFLP patterns obtained using the ATG primer. (B) Autoradiograph produced with probe for the ATG 630 bp fragment. A cloned 630 bp ATG fragment was radiolabeled and used for probing the filter of the gel in (A). This fragment also hybridized to the 1.6 kb fragment in the marker lane (M). a loss of 2-3 repeat units) is the most common allele in both M. franksi and M. annultiris (Fig. 7; Table I). Two individuals appeared to be heterozygous for the 160 bp and 169 bp alleles (i.e., two bands amplified; data not shown). One sample (from M. anniiluris no. 27, Fig. 7) yielded three bands ( 160 bp, 169 bp, and an intermediate band migrating between them); this pattern suggests the presence of an additional locus, although it could be a GENETIC ANALYSIS OF MONTASTRAEA 87 A. M. faveolata M. franksi kh l.o B. 131 141 I M. franksi ACCTATTTCCCTAAA ATTTCTCGC M. franksi M. franksi M. annularis M. annularis M. faveolata . . M. faveolata . . M. faveolata . . M. faveolata . . M. faveolata . . 361 371 501 511 I II I M. franksi GAGTTTAACAACT AACCTTCTGCGTT M. franksi A M. franksi A M. annularis A M. annularis A M. faveolata . .... .... M. faveolata . .... .... M. faveolata . .... .... Af. faveolata . .... .... M. faveolata . .... .... Figure 4. Agarose gel and sequence alignments showing the differences between the 880 and 920 bp GGAG fragments. (A) Fragments amplified from genomic DNA of Monlustraea faveolata and M. franksi using the Montastraea-biased primers. (B) Sequence alignment showing the regions that generate the difference in size of the 880 and 920 bp fragments. PCR artifact. In contrast, most samples of M. ftnenlutu yielded a higher molecular weight smear above 220 bp, rather than discrete 160 or 169 bp bands, when using the same primers and PCR conditions ("null" alleles. Fig. 7). Overall, this microsatellite locus suggests that genetic differences exist between M. faveolata and the other two taxa. but determining the precise nature of these differ- ences would require further analyses. J. V. LOPEZ ET AL A. -ATG M.faveolata M. franksi bp 630 G - Gamete DNA S - Somatic Tissue DNA B. -GGAG M. faveolata M. franksi Figure 5. Comparison of AFLP patterns from gametic (G) and somatic (S) tissue samples. DNA derived from sperm and from somatic tissue of the same Montaxtraea colony were analyzed in parallel AFLP-PCRs. using identical conditions. (A) Results from ATG primer. (B) Results from GGAG primer. Diagnostic bands are identified by arrows. GENETIC ANALYSIS OF MONTASTKAEA 89 Table I Zooxanthellae Mfra-gtttl alli'lc distributions in members of the Montastraea annularis ABC Fav complex bp 630 Figure 6. AFLP assay of zooxanthella samples. The ATG primer was used after PCR preamplification of zooxanthella templates (see methods). Identical conditions for AFLP analyses were used on both Zooxanthellae and coral (Montaxtraea faveolala) DNA samples. Three faint bands (indi- cated by arrows) obtained from Type A Zooxanthellae appear similar to the three dominant AFLP bands obtained from M. faveolata (550, 630. 750 bp) shown in Fig. IB; these may be due to coral (animal) contamination of the zooxanthella DNA. Allele size*/pattern Species 160 \W 160/169 Null Total M. franksi X 1 1 1 11 M. annularis 7 2 1 1 II M. faveolata 1 11 12 * Sizes indicated are approximate (see legend Fig. 7). Discussion Status of the three members of the Montastraea annularis complex When the specific status of taxa in sympatry is question- able, multiple, independent, fixed differences provide com- pelling evidence for the lack of effective interbreeding (Avise and Ball, 1990). The reciprocal presence/absence pattern for the GGAG 880 and 920 bp bands appears to represent one such fixed difference between M. faveolata and the other two taxa. In addition, strong frequency differ- ences at the ATG 630 locus and failure to amplify the 160 or 169 bp alleles at the microsatellite locus in most M. faveolata also point to the distinctiveness of this species. The significance of these genetic differences is further sup- ported by other biological differences that distinguish the taxa (Tomascik, 1990: Hayes, 1990; Knowlton et ai, 1992; Van Veghel and Bak, 1993, 1994; Van Veghel. 1994; Van Veghel and Kahmunn, 1994; Van Veghel and Bosscher, 1995; Van Veghel et ai. 1996; Weil and Knowlton, 1994; Szmant et ai, 1997; Knowlton et nl.. 1997; Knowlton and Budd, unpubl.). M. faveolata M. franksi M. annularis bp 200 154 Figure 7. A subset of the Monlastraea samples assayed for the Mfra-gtttl microsatellite locus. Sample number and species identity for each coral colony are shown above gel. Sizes refer to two bands in the molecular weight markers (M). Representative "null" Mfra-gtttl patterns are shown in the first five lanes. Samples from M. annularis were run on a separate gel. Samples from two individuals of M. franksi (nos. 312, 408) yielded bands that appear to be slightly larger than 160 bp; confirmation of their distinctiveness would require additional analysis. 90 J. V. LOPEZ ET AL. The nature of a species boundary between M. annularis and M. franksi remains much more problematic. No tech- nique used to date has revealed fixed genetic differences between them, despite marked differences in both aggres- sive behavior (Weil and Knowlton, 1994; Van Veghel and Bak, 1493) and the timing of spawning (Knowlton et ai, 1997; Szmant et ai., 1997). More than rare hybridization would presumably erode the predictable association be- tween colony morphology and these other biological char- acteristics, but genetic evidence supporting this otherwise reasonable argument is lacking. Nevertheless, negative re- sults for any single gene is weak evidence to support syn- onymizing species, particularly when, as is the case here, other types of data point to the existence of reproductive barriers. A sobering example of the limitations of negative genetic evidence is provided by Howard et al. ( 1997), who found only six species-specific markers distinguishing two species of oaks, despite having screened 700 10-bp primers. Molecular characters also provide an ideal means for statistically analyzing the probability of encountering par- ticular combinations of characters, including those that would be expected in an Fl hybrid (Lessios and Pearse, 1996; Boeklen and Howard, 1997; Suchanek et al.. 1997; Foltz, 1997). The only individual with an atypical allele for its species (M. franksi no. 19, with the ATG 630 bp band characteristic of M. faveolata; Fig. IB) had the typical M. franksi band size for GGAG (fig. 2A in Lopez and Knowl- ton, 1997). This suggests that this individual is not an FI hybrid, although this pattern could reflect introgression. Using these and additional loci to screen for hybrids in natural populations will allow us to determine whether Veron's ( 1995) proposal of frequent hybridization applies to this species complex. If Fl hybrids are not detected in large surveys, then the rare occurrence of atypical alleles at some loci probably reflects the fact that ancestral polymorphisms have not yet been completely sorted with respect to current species boundaries (Pamilo and Nei, 1988; Moore, 1995). The finding of genetic differences between M. faveolata and the other two taxa in Panama should allow us to determine whether the same patterns occur at other loca- tions within the range of these species. Of particular interest will be sites in the northern Caribbean. Fertilization studies in the Florida Keys do not reveal clear barriers between M. faveolata and the other taxa (Szmant et al.. 1997), in con- trast to results from similar studies in Panama (Knowlton et al.. 1997; Levitan and Knowlton. pers. obs.). Occasional colonies that exhibit mosaic growth forms between M. faveolata and M. annularis have also been observed in both the Bahamas (Knowlton, pers. obs.) and the Dry Tortugas (E. Weil, pers. comm.). The same primers that amplified the Mfra-gtttl microsatellite locus in Panama amplified a sim- ilar 169 bp band in two Montastraea colonies of uncertain taxonomic status from the Florida Keys, suggesting that at least some of the markers we have developed for corals from Panama will have broad geographic utility (Cook et al.. 1991). Molecular genetic analyses of scleractinian corals Until recently, protein electrophoresis was the primary tool for genetic studies of corals, primarily at the level of species (Ohlhorst, 1984; Ayre et al.. 1991; Weil, 1993; Potts et al., 1993; Stobart and Benzie, 1994; Weil and Knowlton, 1994; Garthwaite et ai. 1994; Miller and Benzie, 1997) and population (Stoddart, 1984a, 1984b; Hey ward and Stoddart, 1985; Willis and Ayre, 1985; Ayre and Willis, 1988; Hunter, 1993; Hellberg, 1994). More recently, DNA-based techniques have been used to determine higher level phy- logenies (McMillan and Miller, 1990; McMillan et al.. 1991; Chen et al.. 1995; Veron et al.. 1996; Romano and Palumbi, 1996, 1997), and to analyze or recognize species and populations (McMillan and Miller, 1989; Beauchamp and Powers, 1996; Odorico and Miller, 1997; Lopez and Knowlton, 1997; Hunter et ai. 1997; Takabayashi et ai. 1998). This is, in principle, straightforward given the wide applicability of the methods; but in practice, the scleractin- ian coral genome has provided several surprises that remain poorly understood. For example, Romano and Palumbi (1996) used mitochondrial 16S rDNA sequences to define two distantly related clades, whose 29% sequence diver- gence implied a split predating the origin of coral skeletons 240 million years ago. Nevertheless, three individuals con- tained sequences from both of these highly divergent clades. Odorico and Miller (1997) also found highly divergent ITS and 5.8S nuclear rDNA sequences within single individuals of several Acropora species. These patterns could be inter- preted as evidence for evolutionary reticulation. However, extensive inter-individual variation without intra-colony variation has also been reported (e.g., 31% sequence vari- ation among 12 individuals of Stylophora pistillata: Tak- abayashi et al., 1998). Individual genes can also show quite different evolutionary patterns in different coral taxa: ITS sequences exhibit modest to considerable variability be- tween congeneric species in Acropora (Odorico and Miller, 1997), Porites (Hunter et a I.. 1997). and Balanophyllia (Beauchamp and Powers, 1996), but very little between members of the Montastraea annularis complex (Lopez and Knowlton, 1997; Szmant et al., 1997). Identical 16S rDNA sequences for corals in different genera (Romano and Palumbi, 1996) are also surprising. When genetic variation is low, sequencing individual genes may be less efficient than the use of approaches that screen broadly across the genome. Of these, analysis of AFLPs has considerable promise because it is straightfor- ward, relatively inexpensive, and accessible. It is also prob- ably more reproducible than RAPDs, and therefore more suitable for analyses of field samples (e.g., Janssen et al., 1996; Huys et al., 1996; Majer et al.. 1996; Folkertsma et GENETIC ANALYSIS OF MONTASTRAEA 91 al.. 1996; this study). Many AFLP loci have already been shown to be inherited in a Mendelian fashion (Vos et at.. 1995), although like RAPDs they are dominant markers. Although allozymes remain a valuable tool because of their codominant inheritance and accessibility, the relatively small number of potential loci that can be reliably scored in scleractinians limits their usefulness for discriminating very similar species. AFLP loci can also be further explored using the standard techniques of molecular biology. These more time-consum- ing and expensive steps are recommended whenever poten- tial inherent biases in PCR-based methods have not been explored (Vos et al., 1995). More detailed analysis is also essential for understanding the genetic basis of different band patterns and confirming which bands are homologous. The results of our studies of Montastraea support the im- portance of such additional analyses. For example, the GGAG band pattern differences be- tween M. faveolata and the other taxa could in principle have been due to a difference at one locus (resulting in change in fragment size), or differences at two loci (each with a visible band and a null allele). The ability of primers based on the 880 bp band to amplify what appears to be the 920 bp band and the homology of sequences from these amplifications support the former interpretation. When there are many differences between taxa, and distinguishing taxa is the only goal, then knowing the exact number of inde- pendent loci is perhaps not a serious issue. However, when there are relatively few loci that distinguish taxa (as is the case here), or when one wishes to recognize hybrids, un- derstanding the basis of observed differences is particularly important. Interpreting similarity between bands can be likewise complex due to the possibility of comigration of non-ho- mologous fragments (Rieseburg, 1996; Grosberg et al., 1996). Thus we cannot be sure that the AFLP bands shared between M. annularis and M. franksi are homologous, al- though this seems likely based on the overall genetic sim- ilarity of these two taxa (Van Veghel and Bak, 1993; Weil and Knowlton, 1994; this study). Assessing homology is particularly important in the interpretation of unusual band- ing patterns for example, the ATG 630 bp band in a single individual of M. franksi, which was found to be homologous to the ATG 630 bp band characteristic of M. faveolata. DNA-based methods for analysis of intraspecific gene flow will be especially difficult when species themselves are poorly defined. For Montastraea, there may be a narrow technical window between methods that can detect differ- ences among species, and methods that can detect differ- ences among populations or clones within species (e.g.. Coffroth, 1997; Sites and Crandall, 1997). This should be a high priority for future work, as effective conservation biology depends on determining whether regions are genet- ically interconnected to the extent predicted by current patterns (Roberts, 1997). Acknowledgments Early versions of this manuscript were improved by thoughtful discussions with J.C. Stephens and Tom Laugh- lin. Javier Jara, Juan Mate, and Rob Rowan helped collect the corals and sperm samples. The manuscript was im- proved by comments from Rob Rowan and anonymous reviewers. The Smithsonian Institution provided financial support. The Government of Panama (Institute Nacional de Recursos Naturales Renovables and Recourses Marines) and the Kuna Nation granted permission for field work and collecting. Literature Cited Altschul, S. F., W. Gish, W. Miller. E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. J. Mol. Biol. 215: 403-410. Avise, J. C. 1994. Molecular Markers. Natural History and Evolution. Chapman and Hall, New York. Avise, J. C., and R. M. Ball Jr. 1990. Principles of genealogical concordance in species concepts and biological taxonomy. Oxford Sun: Evol. Biol. 7: 45-67. Ayre, D. J., and B. L. Willis. 1988. Population structure in the coral Pavona cactus: clonal genotypes show little phenotypic plasticity. Mar. Biol. 99: 495-505. Ayre, D. J., J. E. N. Veron, and S. L. Dufty. 1991. The corals Acropora palifera and Acropora cuneata are genetically and ecologically distinct. Coral Reefs 10: 13-18. Beauchamp, K. A., and D. A. Powers. 1996. Sequence variation of the first internal spacer (ITS-1) of ribosomal DNA in ahermatypic corals from California. Molec. Mar. Biol. Biotechnol. 5: 357-362. Boecklen, W. J., and D. J. Howard. 1997. Genetic analysis of hybrid zones: numbers of markers and power of resolution. Ecology 78: 2611-2616. Chen, C. A., D. M. Odorico, M. Ten Lohuis, J. E. N. Veron, and D. J. Miller. 1995. Systematic relationships within the Anthozoa (Cni- daria: Anthozoa) using the 5'-end of the 28S rDNA. Mol. Phylogenet. Evol. 4: 175-183. Coffroth, M. A. 1997. Molecular approaches to the study of clonal organisms: deciphering the alphabet soup. Proc. 8th Int. Coral Reef Symp. 2: 1603-1608. Cook, C. B., J. V. Lopez, P. Eng, and E. M. Mueller. 1998. Geographic variation of reef corals in the Montastraea annularis complex: Florida vs. Panama. Abstr. 26th Benthic Ecology Meet. p. 27. Cracraft, J. 1989. Speciation and its ontology: the empirical conse- quences of alternative species concepts for understanding patterns and processes of differentiation. Pp. 28-59 in Speciation and Its Conse- quences. D. Otte and J. A. Endler. eds. Sinauer, Sunderland. MA. Folkertsma, R. T., J. N. A. M. R. van der Voort, K. E. de Groot, P. M. van Zandvoort, A. Schots, F. J. Gommers, J. Helder, and J. Bak- ker. 1996. Gene pool similarities of potato cyst nematode popula- tions assessed by AFLP analysis. Mol. Plant-Microbe Interact. 9: 47-54. Foltz, D. W. 1997. Hybridization frequency is negatively correlated with divergence time of mitochondrial DNA haplotypes in a sea star (Lep- tasterias spp.) species complex. Evolution 51: 283-288. Garthwaite, R. L., D. C. Potts, J. E. N. Veron, and T. J. Done. 1994. Electrophoretic identification of poritid species (Anthozoa: Sclerac- tinia). Coral Reefs 13: 49-56. 92 J. V. LOPEZ ET AL. Grosberg, R. K., D. R. Levitan, and B. B. Cameron. 1996. Charac- terization of genetic structure and genealogies using RAPD-PCR mark- ers: A random primer for the novice and the nervous. Pp. 67-100 in Molecular Zoology: Advances, Strategies ana' Protocols, J. D. Ferraris and S. R. Palumbi, eds. Wiley-Liss, New York. Hayes, J. A. 1990. Distribution, movement and impact of the corallivo- rous gastropod Coralliophila abbreviate (Lamarck) on a Panamanian patch reef. J. Exp. Mar. Biol. Ecol. 142: 25-42. Hellberg, M. E. 1994. Relationships between inferred levels of gene flow and geographic distance in a philopatric coral, Balanophy/lia elegans. Evolution 48: 1829-1854. Heyward, A. J., and J. A. Stoddart. 1985. Genetic structure of two species of Montipora on a patch reef: conflicting results from electro- phoresis and histocompatibility. Mar. Biol. 85: 117-121. Howard, D. J., R. W. Preszler, J. Williams, S. Fenchel, and W. J. Boecklen. 1997. How discrete are oak species? Insights from a hybrid zone between Quercus grisea and Quercus gambelii. Evolution 51: 747-755. Hunter, C. L. 1993. Genotypic variation and clonal structure in coral populations with different disturbance histories. Evolution 47: 1213- 1228. Hunter, C., C. W. Morden, and C. M. Smith. 1997. The utility of ITS sequences in assessing relationships among zooxanthellae and corals. Proc. 8th Int. Coral Reef Symp. 2: 1599-1602. Huys, G., I. Kersters, R. Coopman, P. Janssen, and K. Kersters. 1996. Genotyping diversity among aeromonas isolates recovered from drink- ing water production plants as revealed by AFLP analysis. Syst. Appl. Micro. 19: 428-435. Janssen, P., R. Coopman, G. Huys, J. Swings, M. Bleeker, P. Vos, M. Zabeau, and K. Kersters. 1996. Evaluation of the DNA fingerprint- ing method AFLP as a new tool in bacterial taxonomy. Microbiology 142: 1881-1893. Jarne, P., and P. J. L. Lagoda. 1996. Microsatellites, from molecules to populations and back. Trends Ecol. Evol. 11: 424-429. Knowlton, N., and L. A. Weigt. 1997. Species of marine invertebrates: a comparison of the biological and phylogenetic species concepts. Pp. 199-219 in Species: the Units of Biodiversity, M. F. Claridge, H. A. Dawah, and M. R. Wilson, eds. Systematics Assoc. Spec. Vol. Ser. 54, Chapman and Hall. London. Knowlton, N., E. Weil, L. A. Weigt, and H. M. Guzman. 1992. Sibling species in Montastraea annu/aris, coral bleaching, and the coral cli- mate record. Science 255: 330-333. Knowlton, N., J. L. Mate, H. M. Guzman, R. Rowan, and J. Jara. 1997. Direct evidence for reproductive isolation among the three species of the Montastraea annularis complex in Central America (Panama, Hon- duras). Mar. Biol. 127: 705-7 1 1 . Lessios, H. A., and J. S. Pearse. 1996. Hybridization and introgression between Indo-Pacific species of Diadema. Mar. Biol. 126: 715-723. Lopez, J. V., and N. Knowlton. 1997. Discrimination of species in the Montastraea annu/aris complex using multiple genetic loci. Proc. 8th Int. Coral Reef Symp. 2: 1613-1618. Majer, D., R. Mithen, B. G. Lewis, P. Vos, and R. P. Oliver. 1996. The use of AFLP fingerprinting for the detection of genetic variation in fungi. Mycol. Res. 100: 1 107-1 111. Mallet, J. 1995. A species definition for the Modern Synthesis. Trends Ecol. Evol. 10: 294-299. McMillan, J., and D. J. Miller. 1989. Restriction analysis and DNA hybridization applied to the resolution of Acropora nobilis from Acro- pora formosa. Proc. 6th Intl. Coral Reef Symp. 2: 775-777. McMillan, J., and D. J. Miller. 1990. Highly repeated DNA sequences in the scleractinian coral genus Acropora: evaluation of cloned repeats as taxonomic probes. Mar. Biol. 104: 483-487. McMillan, J., D. Yellowlees, A. Heyward, P. Harrison, and D. J. Miller. 1988. Preparation of high molecular weight DNA from her- matypic corals and its use for DNA hybridization and cloning. Mar. Biol. 98: 271-276. McMillan, J., T. Mahoney, J. E. N. Veron, and D. J. Miller. 1991. Nucleotide sequencing of highly repetitive DNA from seven species in the coral genus Acropora (Cnidaria: Scleractinia) implies a division contrary to morphological criteria. Mar. Biol. 110: 323-327. Miller, K., and R. Babcock. 1997. Conflicting morphological and re- productive species boundaries in the coral genus Platygvra. Biol. Bull. 192: 98-110. Miller, K. J., and J. A. H. Benzie. 1997. No clear genetic distinction between morphological species within the coral genus Platygvra. Bull. Mar. Sci. 61: 907-917. Moore, W. S. 1995. Inferring phylogenies from mtDNA variation: mi- tochondrial-gene trees versus nuclear-gene trees. Evolution 49: 718- 726. Mueller, U. G., S. E. Lipari, and M. G. Milgroom. 1996. Amplified fragment length polymorphism (AFLP) fingerprinting of symbiotic fungi cultured by the fungus-growing ant Cvphomvrmex minutus. Mol. Ecol. 5: 119-122. Nei, M., and W. H. Li. 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76: 5269-5273. Odorico, D. M., and D. J. Miller. 1997. Variation in the ribosomal internal transcribed spacers and 5.8S rDNA among five species of Acropora (Cnidaria; Scleractinia): Patterns of variation consistent with reticulate evolution. Mol. Biol. Evol. 14: 465-473. Ohlhorst, S. L. 1984. The use of polyacrylamide gel electrophoresis in coral taxonomy. Palaeont. Am. 54: 45-48. Pamilo, P., and M. Nei. 1988. Relationships between gene trees and species trees. Mol. Biol. Evol. 5: 568-583. Potts, D. C., A. F. Budd, and R. L. Garthwaite. 1993. Soft tissue vs. skeletal approaches to species recognition and phylogeny reconstruc- tion in corals. Cour. Forsch.-Inst. Senckenberg 164: 221-231. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. Rieseberg, L. H. 1996. Homology among RAPD fragments in interspe- cific comparisons. Mol. Ecol. 5: 99-105. Roberts, C. M. 1997. Connectivity and management of Caribbean coral reefs. Science 278: 1454-1457. Romano, S. L., and S. R. Palumbi. 1996. Evolution of scleractinian corals inferred from molecular systematics. Science 271: 640-642. Romano, S. L., and S. R. Palumbi. 1997. Molecular evolution of a portion of the mitochondria! 16s ribosomal gene region in scleractinian corals. J. Mol. Evol. 45: 397-41 1. Rowan, R., and N. Knowlton. 1995. Intraspecific diversity and ecolog- ical zonation in coral algal symbiosis. Proc. Natl. Acad. Sci. USA 92: 2850-2853. Rowan, R., and D. A. Powers. 1991. Molecular genetic identification of symbiotic dinoflagellates (zooxanthellae). Mar. Ecol. Prog. Ser. 71: 65-73. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular Clon- ing: A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Sites, J. W. Jr., and K. A. Crandall. 1997. Testing species boundaries in biodiversity studies. Conserv. Biol 11: 1289-1297. Southern, E. M. 1975. Detection of specific sequences among DNA fragments separated by gel electrophoresis. / Mol. Biol. 98: 503-517. Stephens, J. C., D. A. Gilbert, N. Yuhki, and S. J. O'Brien. 1992. Estimation of heterozygosity for single-probe multilocus DNA finger- prints. Mol. Biol. Evol. 9: 729-743. Stobart, B., and J. A. H. Benzie. 1994. Allozyme electrophoresis dem- onstrates that the scleractinian coral Montipora digitata is two species. Mar. Biol. 118: 183-190. Stoddart, J. A. 1984a. Genetic differentiation amongst populations of GENETIC ANALYSIS OF MONTASTRAEA 93 the coral Pocillopora damicornis off southwestern Australia. Coral Reefs 3: 149-156. siuilil.M i. J. A. 1984b. Genetical structure within populations of the coral Pocillopora damicornis. Mar. Biol. 81: 19-30. Suchanek, T. H., J. B. Geller, B. R. Kreiser, and J. B. Mitton. 1997. Zoogeographic distnbutions of the sibling species Mvtilus galloprovin- cialis and M. trossulus (Bivalvia: Mytilidae) and their hybrids in the north Pacific. Biol. Bull. 193: 187-194. Szmant, A. 1991. Sexual reproduction by the Caribbean corals Montas- traea annularis and M. cavernosa. Mar. Ecol. Prog. Ser. 74: 13-25. Szmant, A. M., E. Weil, M. W. Miller, and D. E. Colon. 1997. Hy- bndization within the species complex of the scleractinian coral Mon- taslraea annularis. Mar. Biol. 129: 561-572. Takabayashi, M., D. Carter, S. Ward, and O. Hoegh-Guldberg. 1998. Inter- and intra-specific variability in rDNA sequence in the ITS region of corals. Proc. Austral. Coral Reef Soc. 75th Ann. Conf. 237-244. Templeton, A. R. 1989. The meaning of species and speciation: a genetic perspective. Pp. 3-27 in Speciation and its Consequences, D. Otte and J. A. Endler, eds. Sinauer. Sunderland, MA. Tomascik, T. 1990. Growth rates of two morphotypes of Montastraea annularis along a eutrophication gradient. Barbados. W. I. Mar. Poll. Bull. 21: 376-380. Van Veghel, M. L. J. 1994. Reproductive characteristics of the poly- morphic Caribbean reef building coral Montastraea annularis. I. Ga- metogenesis and spawning behavior. Mar. Ecol. Prog. Ser. 109: 209- 219. Van Veghel, M. L. J., and R. P. M. Bak. 1993. Intraspecific variation of a dominant Caribbean reef-building coral, Montastraea annularis: genetic, behavioral and morphometric aspects. Mar. Ecol. Prog. Ser. 92: 255-265. Van Veghel, M. L. J., and R. P. M. Bak. 1994. Reproductive charac- teristics of the polymorphic Caribbean reef-building coral Montastraea annularis. III. Reproduction in damaged and regenerating colonies. Mar. Ecol. Prog. Ser. 109: 229-233. Van Veghel, M. L. J., and H. Bosscher. 1995. Variation in linear growth and skeletal density within the polymorphic reef building coral Montastraea annularis. Bull. Mar. Sci. 56: 902-908. Van Veghel, M. L. J., and M. E. H. Kahmann. 1994. Reproductive characteristics of the polymorphic Caribbean reef building coral Mon- tastraea annularis. II. Fecundity and colony structure. Mar. Ecol. Prog. Ser. 109: 221-227. Van Veghel, M. L. J., D. F. R. Cleary, and R. P. M. Bak. 1996. Interspecific interactions and competitive ability of the polymorphic reef-building coral Montastraea annularis. Bull. Mar. Sci. 58: 792- 803. Veron, J. E. N. 1995. Corals in Space and Time: The Biogeography and Evolution of the Scleractinia. UNSW Press, Sydney, Australia. Veron, J. E. N., D. M. Odorico, C. A. Chen, and D. J. Miller. 1996. Reassessing evolutionary relationships of scleractinian corals. Coral Reefs 15: 1-9. Vos, P., R. Hogers, M. Bleeker, M. Reijans, T. van de Lee, M. Homes, A. Frijters, J. Pot, J. Peleman, M. Kuiper, and M. Zabeau. 1995. AFLP: a new technique for DNA fingerprinting. Nucl. Acids Res. 23: 4407-4414. Weil, E. 1993. Genetic and morphological variation in Caribbean and eastern Pacific Porites (Anthozoa, Scleractinia). Preliminary results. Proc. 7th Inter. Coral Reef Symp. 2: 643-656. Weil, E., and N. Knowlton. 1994. A multi-character analysis of the Caribbean coral Montastraea annularis (Ellis and Solander. 17861 and its two sibling species, M. fave olata (Ellis and Solander. 1786) and M. franksi (Gregory, 1895). Bull. Mar. Sci. 55: 151-175. Willis, B. L., and D. J. Ayre. 1985. Asexual reproduction and genetic determination of growth form in the coral Pavona cactus: biochemical genetic and immunologic evidence. Oecologia 65: 516-525. Willis, B. L., R. C. Babcock, P. L. Harrison, and C. C. Wallace. 1997. Experimental hybridization and breeding incompatibilities within the mating systems of mass spawning reef corals. Coral Reefs 16, Suppl.: S53-S65. Yuhki. N., and S. J. O'Brien. 1990. DNA variation of the mammalian major histocompatibility complex reflects genomic diversity and pop- ulation history. Proc. Natl. Acad. Sci. USA 87: 836-840. Zabeau, M., and P. Vos. 1995. Selective restriction fragment amplifi- cation: a general method for DNA fingerprinting. European Patent Application. Publ # 0534858-A1. Office Europeen des Brevets. Paris. Reference: Biol. Bull. 196: 94-104. (February, 1999) Ultraviolet Radiation and Distribution of the Solitary Ascidian Corella inflata (Huntsman) BRIAN L. BINGHAM 1 '* AND NATHALIE B. REYNS 2 1 Huxley College of Environmental Studies, Western Washington University, Belling/tain, Washington 98225; and 2 Marine Sciences Research Center, State University of New York, Stony Brook, New York 11794 Abstract. The solitary ascidian Corella inflata is a com- mon fouling organism in many areas of Puget Sound and the San Juan Archipelago, Washington, USA. Despite its abun- dance, it is conspicuously absent from areas that receive direct sunlight. Previous work suggests that ascidians in unshaded habitats can be overgrown and killed by algal overgrowth. In this study, we tested the hypothesis that UV irradiation contributes to C. inflata distribution by killing individuals exposed to direct sunlight. To test this, we exposed C. inflata embryos, larvae, juveniles, and adults to UV irradiation and measured the responses. We also tested for UV-absorbing compounds in larvae, juveniles, and adults. In the laboratory, UV significantly damaged all life stages; the earliest stages were most vulnerable. A 3-week UV exposure significantly shortened adult life span. Juve- niles suffered 100% mortality after only 3 days. Tadpole larvae decreased settlement and metamorphosis after 1 day of UV exposure, and embryos exhibited developmental abnormalities after only 30 minutes of exposure. None of the life-history stages had apparent UV-absorbing com- pounds. Given the vulnerability of this species to UV, we suggest that its unique life-history traits (i.e., time of spawn- ing, brooding behavior, length of larval life) help it persist in its preferred habitat and avoid dispersal into inappropri- ate, UV-exposed areas. Introduction Corella inflata (Huntsman) is a solitary ascidian common throughout Puget Sound, Washington, and in waters off the west coast of British Columbia. It occurs from the intertidal zone to 45 m (Van Name. 1945) but is most abundant on Received 26 November 1997; accepted 6 November 1998. * To whom correspondence should be addressed. E-mail: binghamls 1 cc.wwu.edu docks and pilings (Young, 1982). Adults, which may reach 5 cm in length, have a thin, transparent outer tunic. This is in marked contrast to the tough, opaque tunic that protects most other solitary ascidians. Lambert (1968) studied a population of C. inflata in a Puget Sound marina for 12 months and observed mass mortality in the early spring. The mortality coincided with a period of heavy diatom growth, and Lambert suggested that smothering diatom mats were responsible for the ascidian deaths. This conclusion was supported by observations that mass mortality occurred only in areas exposed to full solar radiation; C. inflata in shaded habitats survived. Recent research suggests that another factor may contrib- ute to mortality of C. inflata in exposed areas. Increasing interest in the status of the stratospheric ozone layer has led to intense study of the deleterious effects of ultraviolet radiation (UV). Jokiel ( 1980) first demonstrated the damag- ing effects of UV radiation on tropical marine invertebrates (including ascidians). More recent work has shown that the UV-B portion of the spectrum (280-320 nm) is particularly lethal to marine bacteria, plankton, invertebrates, and fish (reviewed by Worrest, 1982; Hardy and Gucinski, 1989; see also Snick et al., 1991 ; Karentz et a!.. 1991 ; Karentz. 1994a, b). There are indications that marine invertebrate embryos and larvae may be particularly sensitive to solar radiation (Damkaer et al, 1980; Pennington and Emlet, 1986; Bier- mann et al., 1992; Adams and Shick, 1996). The habitat (primarily shallow water) and anatomy (thin, transparent tunic) of C. inflata may make it particularly vulnerable to UV damage, and the effects may not be limited to adult animals. Unlike most solitary ascidians, C. inflata holds its embryos in a spacious brood chamber. The eggs are buoyant and float to the top of the chamber where there is potential for UV-induced damage during develop- ment. 94 UV AND CORELLA INFLATA DISTRIBUTION 95 The purpose of this study was to determine how UV radiation affects Corella inflata. We examined field distri- butions of ascidiuns to determine whether population den- sity was correlated with UV exposure. We measured the vulnerability of embryos, larvae, juveniles, and adults to UV damage. Finally, we examined larvae, juveniles, and adults for UV-absorbing compounds. Materials and Methods Field sampling We selected docks in two marinas (Anacortes Marina and Skyline Marina) in Anacortes, Washington, for the field portion of this study. These marinas have similar construc- tion designs, with docks that project out into a sheltered embayment. A roof shades inner dock slips, but outer slips are not covered and receive full sunlight. Because the docks rise and fall with the tides, they harbor diverse invertebrate communities including sponges, hydrozoans, polychaetes, barnacles, bivalves, bryozoans, and ascidians (particularly Corella inflata). Macroalgae are absent from most docks at both locations. At each marina, we chose four nonadjacent slips (two shaded and two exposed) for detailed study. Five permanent sampling locations were marked at 2.5-m intervals along each slip. We measured C. inflata densities at each location by placing a 25 X 25 cm quadrat on the side of the slip just below the water's surface and counting all C. inflata indi- viduals within the quadrat. We made counts in June and again in August, 1994. To determine how the density of C. inflata related to UV exposure, we measured spectral irradiance 5 cm below the water's surface at each quadrat location. Measurements (from 300-700 nm at 2-nm intervals) were made with a LI-COR 1800UW spectroradiometer (see Kirk et /., 1994, for a discussion of the measurement characteristics of this instrument) on a clear sunny day ( 15 July 1994) at 1300 h. We also measured light attenuation in the marinas by taking spectral scans at 50 cm and 1 m at several exposed quadrat locations. Finally, we went just outside the entrance to the Skyline Marina and measured light at 1-m intervals from the surface to 20 m. This allowed us to determine penetration of UV-A and UV-B in the local waters. Because we could not measure the complete UV-B spec- trum (from 280-320 nm) due to limitations of the instru- ment, we used a 2nd order regression of the data from 300-320 nm to extrapolate the curves from 280 to 300 nm. We then integrated the data files to get separate measure- ments of total UV-B (280-320 nm). UV-A (320-400 nm), and PAR (photosynthetically active radiation, 400-700 nm). Because algal overgrowth could be a confounding factor in our sampling, we monitored algal growth at the study sites. We hung transparent acrylic strips (2-cm wide X 15-cm long) at all quadrat locations. After 6 weeks (1 July-15 August. 1994), they were collected and two 4-cm 2 areas, one at 5 cm and the other at 15 cm, were wiped with a OFF filter. We extracted the filters in 90% acetone for 24 h and made fluorornetric readings of chlorophyll a (Parsons et ai, 1984). The two values were averaged to give a mea- surement of algal growth at each quadrat location. Laboratory experiments To measure the sensitivity of C inflata to UV, we used an enclosed, flow-through seawater tank equipped with two UV bulbs (Q-Panel Company, UVB-313) and two cool- white fluorescent bulbs. A cellulose acetate bulb sheath filtered out wavelengths less than 290 nm. The tank was divided into two sections with nylon netting. The control section was covered by a UV-filtering shield (Atohaas North America, Plexiglas UF3, 3-mm thickness) that re- flected wavelengths below 400 nm. To compare light in the two treatments, we measured the irradiance spectra in the shielded and unshielded portions of the tank (at 5-cm depth) with the spectroradiometer. Adult sensitivity. To determine if UV exposure affects adult C. inflata, we placed 20 individuals in 240-ml polypropylene cups from which the bottoms had been re- moved and replaced with Nitex screen. A slit in a piece of open-cell foam attached to the inside wall of the cups held the ascidians in a normal position (horizontal with the excurrent siphon and brood chamber up; Young, 1988) but was flexible enough to allow normal feeding. The cups were randomly assigned to either UV-exposed (/; = 10) or shielded (n = 10) treatments. Foam collars kept the cups floating with the adults about 3 cm below the surface of the water. The tank was placed on a 15:9 h light/dark cycle and survival was monitored for 22 days. Life spans (up to 22 days) in the two treatments were compared by one-way ANOVA. We set a at 0.05 for all statistical analyses and used Hartley's F max test (Sokal and Rohlf, 1995) to test for homogenous variances before each analysis. Juvenile sensitivity. We collected larvae by exposing adult C. inflata to bright light shortly after collection. Thirty larvae were placed in each of 12 roughened polystyrene petri dishes (100-mm diameter, 15-mm deep) and allowed to settle. Positions of the juveniles were marked with a permanent ink marker on the back of the dishes. We filled the dishes with fresh seawater and randomly assigned them to the UV-exposed (n = 6) or shielded (n = 6) treatments. Mortality was determined after 4 days of treatment. Because of unequal variances, we used a Kruskal-Wallis nonpara- metric ANOVA to compare treatments. Lan'al sensitivity. Tadpole larvae were exposed to UV radiation to measure effects on settlement. Six petri dishes containing 30 newly released, swimming tadpole larvae were randomly assigned to the UV-exposed (n = 3) or 96 B. L. BINGHAM AND N. B. REYNS shielded (/; = 3) sections of the tank. Percent settlement in each treatment was determined after one light cycle (15 h UV exposure). Treatments were compared by one-way ANOVA. We also subjected tadpole larvae to different periods of UV exposure to determine a damage threshold. Twenty-one replicate petri dishes (with 30 newly released tadpole larvae per dish) were prepared. Three dishes (the 0-exposure con- trol) were placed in the shielded portion of the tank. The other 18 dishes were exposed to the UV light. At 30-min intervals, we moved three randomly chosen plates into the shielded portion of the tank until all plates had been moved over. Twenty-four hours later, percent settlement was de- termined for all plates. We analyzed the relationship be- tween UV exposure time and percent settlement by simple linear regression. Sensitivity of developing embryos. Approximately 75 adults were light shocked to obtain fertilized eggs. Eggs from all the adults were mixed and 30 were arbitrarily assigned to each of 21 petri dishes. The developing embryos were exposed to UV radiation in the experimental tank (in groups of three replicate dishes). Exposure intervals were staggered as described above to determine the threshold at which development becomes abnormal. After 24 h, all dishes were examined and the proportion of eggs that had (1) reached the tadpole stage, (2) arrested development at the morula stage, or (3) become abnormal at, or before, the 16-cell stage was determined. We used a chi-square test for independence to determine whether length of UV exposure affected the developmental stage the embryos reached. To avoid sacrificial pseudoreplication (Hurlburt, 1984), we an- alyzed only one randomly chosen replicate from each ex- posure period. Outplants To determine if recruiting C. inflata could survive in exposed areas of the marina where they were not usually found, newly settled ascidians were transplanted into the field. In the laboratory, 30 tadpole larvae were placed in 100-mm diameter petri dishes (n = 20) and allowed to settle. Settlement locations were marked on the back side of each dish with a permanent marker and five marked plates were hung vertically, front side out, on four slips in Skyline Marina (two shaded and two unshaded). After 6 days, we compared survival of the juveniles in the shaded and ex- posed treatments with a Kruskal-Wallis nonparametric ANOVA. Recruitment We monitored C. inflata recruitment to determine whether larvae ever colonize exposed docks. Six 100-mm diameter cement plates were hung from two shaded and two unshaded slips in Anacortes Marina and left in place from 8 July to 15 August, 1994. We then collected the plates, counted the recruits, and used correlation analysis to test for relationships between irradiance, algal growth, and adult density and number of recruits. If algal overgrowth causes mortality of C. inflata at our site, larvae should avoid settling on algal-covered surfaces (such as those in exposed sites). We tested this by placing 15 roughened petri dishes (100-mm diameter) in Skyline Marina. Three of the dishes hung from a covered dock where darkness prevented algal growth. The remaining 12 dishes hung from a dock exposed to full sunlight. After 2 weeks, the exposed dishes had developed a layer of fila- mentous algae. We collected all the dishes and divided them among three treatments: no algal cover (3 dishes from the covered dock), 100% algal cover (3 dishes from the exposed dock), and 50% algal cover (9 exposed dishes that we carefully scraped to remove the algae from half of the bottom surface). The dishes were filled with fresh seawater and 30 C. inflata tadpole larvae were added to each. After 24 h, settlement in each dish was recorded. Settlement in 100% algal-covered dishes (n = 3) was compared to dishes with no algae (n = 3) with one-way ANOVA. We used a paired Student's / test to compare larval settlement in clean or algal-fouled halves of the scraped dishes (n = 9). UV-absorbing compounds To determine whether C. inflata has UV-absorbing sub- stances, we extracted compounds from three large adults (between 1.5 and 2.5 cm long), five juveniles (less than 1.0 cm long), and the pooled tadpole larvae from 50-75 adults. The tunics were removed from the adults and analyzed separately from the bodies (we hypothesized that UV-ab- sorbing compounds, if present, would be concentrated in the tunic). The bodies were carefully cleaned to remove all feces and gut material and ensure that only C. inflata com- pounds were measured. Tadpole larvae were collected by light shocking adults. We lyopholized and extracted the samples in 80% methanol as described by Karentz et al. (1991). Absorbance of the extracts was measured spectro- photometrically at 2-nm intervals from 280 to 400 nm. The dry weights of the C. inflata samples were unequal (large tunics = 120 mg, large bodies = 100 mg, small tunics = 170 mg, small bodies = 100 mg, tadpole larvae = 120 mg). To permit direct comparison among the samples, we calculated relative absorptivity by the following equa- tion. relative absorptivity Measured absorbance /tissue dry weight (mg)\ path length (cm) X - \ solvent volume (ml) / This equation is derived from Beer's law (Skoog and UV AND CORELLA INFLATA DISTRIBUTION 97 West, 1974) except that absorptivity in Beer's law is the molar absorptivity of a single absorbing substance. Since we do not know the composition of our extracts, we consider absorptivity in this case to be a relative measure. To test the effectiveness of the tunic as a physical barrier to UV damage, we dissected the tunics from one juvenile and one adult C. inflata, placed them directly in the spec- trophotometer light path and measured absorbance from 280 to 400 nm. By substituting tunic thickness (=900 /mm) for path length and tunic density (g cm 3 ) for the concentra- tion measure in the above equation, we calculated relative absorptivities that were directly comparable to those mea- sured with the extracts. Results Field sampling Corella inflata was entirely restricted to shaded slips in both marinas, and densities decreased toward the slip ends where shading was less complete. At both marinas, there was a strong negative relationship between algal growth (as measured by Chi a concentration on acrylic strips) and C. inflata density and between UV irradiance and C. inflata density (Fig. 1). The relationship between C. inflata density and UV-B irradiance alone followed the same pattern and had similarly high correlations (r = -0.88, P < 0.001 for Anacortes Marina and r = -0.68, P < 0.001 for Skyline Marina). There was an apparent threshold in both irradiance and chlorophyll concentration above which few C. inflata individuals occurred. This threshold corresponded to the point at which slips were no longer shaded. There was essentially no change in densities of ascidians between the two sampling dates (June and August). A vertical profile showed a logarithmic decrease in UV irradiance with depth (Fig. 2). UV-B was detected to 5 m. Levels of UV-B corresponding to threshold intensities for adult distributions in the marinas (Fig. 1) occurred at depths between 1.5 and 2 m. UV-A was measurable to 14 m. Laboratory experiments Light levels in our experimental tank differed signifi- cantly from those measured in the field. Although total UV-B irradiances were remarkably similar, total UV-A was 30 times lower in our tank than in the field; total PAR was 32 times lower (Table I). Closer examination of individual spectra reveals that tank and field light quality differed within these wavelength ranges. UV in the experimental tank was skewed to the more biologically damaging, low wavelengths (Fig. 3). Without a weighting function for this species, we cannot determine how this affected the experi- mental animals. It is likely, however, that the tank overem- phasized low-wavelength UV-B effects and under-empha- sized UV-A and PAR effects. The Plexiglas shield effectively removed all UV-B and UV-A from the shielded portion of the tank. The shield also caused an 8.5% drop in total PAR, with the greatest effect at wavelengths between 400 and 415 nm (Table I, Fig. 4). Adult sensitivity. At irradiances present in the tank, UV damaged all C. inflata life-history stages. After 8 days, exposed adults became opaque and many were dead after 10 days; after 21 days, all exposed animals were dead. Control animals also experienced some mortality, probably due to handling. However, mortality rates were lower and leveled off after 17 days. The average life span of the UV-exposed adults (mean = 13.9 1.2 d SD) was significantly lower than that of the shielded controls (18.0 1.1 d, F = 5.8, P = 0.02) within the 21 days of the experiment. This result underestimates the true effect because the experiment was terminated at 22 days and many of the control animals were still living. Juvenile sensitivity. UV exposure affected juveniles even more strongly than it did adults. In the laboratory, juveniles in the exposed treatment showed 100% mortality after only 4 days (controls had 57.0% 11% mortality, H = 9.46, P = 0.002). Similar results were seen in the field outplants. All juveniles had disappeared from dishes in the unshaded habitat after only 3-6 days, whereas 36.6% 8.3% of the shaded juveniles were still alive at the end of the experiment (H = 7.27, P = 0.007). Examination of the dishes in which the juveniles were outplanted revealed minimal algal growth, suggesting that overgrowth was not responsible for the mortality. Larval sensitivity. Very few C. inflata larvae exposed to UV light settled successfully in the laboratory (2.3% 0.4% compared to 56.8% 0.8% for the shielded individ- uals, F = 56.4, P = 0.001). This underestimates the true effect because many larvae that had not settled in the shielded treatment were still actively swimming; all ex- posed larvae were dead. The effects of exposure were cu- mulative; as exposure time increased, settlement success decreased (Fig. 5). Sensitivity of developing embryos. UV interfered with the normal development of C. inflata embryos (Fig. 6, x* for one set of replicates = 303.48, P < 0.001). Even limited exposure had significant developmental consequences. Af- ter only 30 min of exposure, less than 20% of the embryos developed to tadpole stage. As exposure time increased, early cleavages became more abnormal, with many em- bryos failing to pass the 4- or 8-cell stages. No normal development occurred after 1.5 h of UV exposure. Settlement and recruitment Of 199 C. inflata individuals recruiting to plates in Anacortes Marina, 181 (90.9%) were on plates on shaded slips. The remaining 18 recruits on exposed plates were 98 100 -, 80 - 60 - 40 - 20 - B. L. BINGHAM AND N. B. REYNS Anacortes Marina 100 -i r = -0.65 p = 0.001 80 - 60 - 40 - 20 - - r = -0.89 p< 0.001 0.01 0.1 10 0.1 10 "c 10 80 - 60 - 40 - 20 - - Skyline Marina r = -0.48 p = 0.029 100 -i 80 - 60 - 40 - 20 - 0.01 0.1 Chi a 10 0.1 r = -0.73 p< 0.001 10 cm"") -2x Irradiance (W m" ) Figure 1. Density of Corella inflata in Anacortes and Skyline Marinas. Densities are plotted as a function of Chi a concentration (based on overgrowth of clear acrylic strips) and UV irradiance (300-400 nm). Correlation coefficients (r) and P values are shown. on the lower edges where they were shaded by the plate itself. In laboratory assays, a developed algal mat did not affect C. inflata settlement. Similar numbers of larvae settled successfully and metamorphosed on clean and algal-fouled surfaces, and larvae showed no preference for clean rather than algal-fouled surfaces when given a choice (Fig. 7). However, because of low replication, the power of these tests was low (/3 = 0.65 for the single choice and 0.24 for the preference experiment; Cohen, 1988). Levels of algal fouling in the dishes reached 0.08 ju-g Chi a cm" 2 . UV-absorbing compounds All methanol extracts showed minor absorptivity peaks in the UV range (Fig. 8). Tunic extracts absorbed much less UV than extracts of the bodies. Absorbance was greatest for the bodies of small specimens of C. inflata. Tadpole extracts showed a small peak at 290 nm. The peaks (A max ) differed slightly among the samples, but all were between 294 and 300 nm. None fell within the wavelength range in which sunscreening mycosporine-like amino acids normally occur. The tunic provided little physical barrier to UV; absorptivi- UV AND CORELLA INFLATA DISTRIBUTION 99 0001 I o -\ Irradiance (W m" ) 0.01 0.1 I 10 UVB = 10" 675 ( De P th > + 0.153 r = 0.98 374 6 - 3 8 H Q. 10 - 12 - 14 - 16 - Figure 2. Penetration of UV-B (300-320 nm) and UV-A (320-400 nm) in the water column at the entrance to Skyline Marina. Simple regression lines and equations are shown. Measurements were made at 1300 h on 15 July 1994. ties were only slightly above those seen in the chemical extracts (Fig. 8). For comparison, we calculated the relative absorptivity of mylar (a synthetic material that filters wave- lengths <315 nm). Mylar relative absorptivity at 294 nm was 0.34. Discussion Unlike many Puget Sound ascidians, which occur primar- ily in subtidal benthic habitats, Corella inflata is found in greatest abundance on floating docks (Young, 1982). It is not normally found on fixed intertidal structures because it does not tolerate desiccation. The dock habitat may be a refuge from benthic predators that prefer this species to other ascidians that have thicker, heavier tunics (Young, 1985). However, although these artificial substrates may provide protection from predators, they expose C. inflata to hazards associated with high levels of solar radiation. Young and Chia ( 1984) outplanted juveniles of C. inflata to subtidal locations at 4.5-m depths in shaded and unshaded dishes. Mortality was significantly higher in the unshaded dishes, presumably due to algal overgrowth (though this is still within the range to which UV penetrates; Fig. 2). Since algal overgrowth apparently can kill juveniles, we predicted that larvae should detect and avoid algal-filmed surfaces, particularly since larvae of many ascidian species show strong settlement specificity (reviewed by Svane and Young, 1989). In laboratory testing, however, larvae settled readily on surfaces that were covered with filamentous algae, sometimes attaching directly to the algae themselves. Goodbody (1963) and Goodbody and Gibson (1974) out- planted juvenile Ascidia nigra (another solitary ascidian) to the field at depths of 1.2 and 2.1 m and measured cohort survival. They found very high mortality, particularly among individuals at the shallower depth. Mortality was significantly lower in shaded treatments, particularly at the 1.2-m depth. The authors hypothesized that one source of the mortality was smothering from accumulation of benthic diatoms or filamentous algae in the unshaded treatments. The mortality of the outplanted cohorts was highest within the first 15 days. Interestingly, the deep black pigmentation characteristic of adult Ascidia nigra does not develop until about the 15th day after larval settlement, and the authors noted that mortality decreased sharply after the pigment appeared (Goodbody and Gibson, 1974). Our results with C. inflata suggest that UV could be an alternative source of mortality for young ascidians in exposed habitats. If the habitat of C. inflata makes this species vulnerable to UV damage as suggested by our results we expect it to show adaptations to avoid UV exposure in shallow-water habitats. Many organisms possess UV-absorbing, mycospo- rine-like amino acids (MAAs) that may prevent UV damage (reviewed by Karentz, 1994a). The Antarctic ascidian (Mol- gula enodis) contains seven MAAs (Karentz et ai, 1991) with distinct absorbance peaks around 330 nm; the Western Pacific Halocynthia roretzi contains a UV-absorbing sun- screen that absorbs maximally at 337 nm (Kobayashi et ai, 1981). C. inflata absorptivities peaked between 290 and 300 nm, suggesting that MAAs are not responsible for the UV ab- sorbance we measured. The body of this ascidian contains uric acid crystals that are deposited there as metabolic waste products (Lambert et al., 1998). It is possible that uric acid provides some limited protection from UV damage (a uric acid absorbance peak occurs at 292 nm). Another possibility Table I Comparison of UV-B (300-320 nm). UV-A (320-400 nm), and PAR (400-700 nm) in Skyline Marina. Anacortes, WA. where Corella inflata populations were monitored (15 July 1994) and in the tank where experiments were done (compare Fig. 3) Intensity (W m" "-)* Depth Location (cm) UV-B UV-A PAR Tank UV-exposed 5 1.10 1.12 9.99 Shielded 5 0.05 0.06 9.14 Marina 5 1.80 34.59 328.70 50 0.56 23.02 248.00 100 0.37 15.68 226.60 * Measured with a LI-COR 1 800UW spectroradiometer. 100 c 2.0 -| 1.5 - 1.0 - Anacortes Marina: 5-cm depth B. L. BINGHAM AND N. B. REYNS Anacortes Marina: 50-cm depth ^ 2 - 1 i H 1.5 H b ^H 1.0 - 0.5 - 0.0 Anacortes Marina: 100-cm depth r~ 300 Experimental tank: 5-cm depth 400 500 600 700 300 400 500 600 700 Wavelength (nm) Figure 3. Irradiances at 5, 50, and 100 cm in Skyline Marina and in the flow-through seawater tank used for laboratory experiments (5-cm depth). Due to instrument constraints, we were unable to measure the UV-B below 300 nm. Lines on the figures between 290 and 300 nm are extrapolated second-order regressions from the data for 300 to 320 nm. is that the absorbance we saw came from cell secretions known as "ornaments" (Cloney, 1990). Embryonic test cells produce ornaments that are deposited on the larval tunic of C. inflata and other ascidians (Cloney and Cavey, 1982; Cloney, 1994). The ornaments are composed largely of opal, a form of silicon dioxide (Monniot et al, 1992). It is unlikely that the opal alone absorbs UV, but if the orna- ments contain other UV-absorbing substances, they might provide larvae with some protection (R. A. Cloney, pers. comm.). With little structural or chemical protection, C. inflata appears vulnerable to UV damage in all life-history stages. Embryos were the most vulnerable. Even short exposures caused developmental abnormalities; after only 30 min of laboratory UV exposure, many embryos arrested at the morula stage. This agrees with Jeffrey (1990), who also found that UV-irradiated ascidian embryos failed to gastru- late. Although embryos were most sensitive to UV, larvae, juveniles, and adults were also affected. We suggest that the unique life-history characteristics of C. inflata allow the populations to persist in shallow habitats despite this vul- nerability. 0.15 n Without UV-absorbing Plexiglas With UV absorbing Plexiglas 0.10 - & 0.05 - 0.00 700 Wavelength (nm) Figure 4. Irradiance spectra in the shielded and unshielded treatment portions of the experimental tank. Wavelengths from 290-300 nm are extrapolated second-order regressions from the data for 300 to 320 nm. UV AND CORELH 1NFLATA DISTRIBUTION 101 100 c 1.5 cm long; small individuals were <1.0 cm long. All gut contents were cleaned out during preparation to ensure that only C. inflata compounds were measured in the body samples. The area of normal mycosporine-like amino acid (MAA) absorbance peaks is indicated. Inset graphs are absorptivities for intact whole tunic. A majl are maximum absorptivity peaks. See the text for absorptivity calculations. UV AND CORELLA INFLATA DISTRIBUTION 103 higher, and it seems likely that those wavelengths also have a significant impact. The relative importance of UV-A, UV-B. and PAR to C. inflata populations merits further study. In our laboratory work, the UV/PAR ratios were higher than they would be under natural conditions. Many organisms possess DNA repair systems that are activated by short-wave visible light (reviewed by Mitchell and Karentz, 1993). If C. inflata has such systems, they may have been inoperable given the low PAR they received in our labora- tory trials. The ideal way to test this possibility is to repeat the experiments under natural light in the field. Whether UV effects are important for other ascidian species is unknown. Nearly 40 years ago, Endean (1961) suggested that pigmentation in the tunic of Phallusia nni- millata protects that species from ultraviolet damage. To our knowledge, this has never been tested for P. mamillata or any other ascidian. Using a bacterial dosimeter, Karentz and Lutze (1990) detected significant UV-B radiation to 10 m and documented effects of irradiation to 20 and 30 m in Antarctic waters. Depth of UV penetration will depend on local conditions. However, in moderately clear, shallow waters, it is highly likely that UV will have significant ecological effects that may be detectable in life histories of the species that persist there. Acknowledgments We thank Steve Hey wood and Gene Me Keen for logis- tical assistance. Charley Lambert, Gretchen Lambert, Rich- ard Cloney, and Jody Steelier provided unpublished infor- mation and/or valuable discussion and insight. We are also indebted to Deneb Karentz for patient guidance and sug- gestion. We appreciate the very helpful comments of two anonymous reviewers. The owners, staff, and residents of Anacortes, Cap Sante. Anchor Cove, and Skyline Marinas provided access to and use of their facilities. This work was supported by NSF grants OCE-9340328, USE-9650124, and USE-9151453. Literature Cited Adams, N. L., and J. M. Shick. 1996. Mycosponne-Iike amino acids provide protection against ultraviolet radiation in eggs of the green sea urchin Strongylocentrotus droebachiensis. Photochem. Photohiot. 64: 149-158. Biermann, C. H., G. O. Schinner, and R. R. Strathmann. 1992. In- fluence of solar radiation, microalgal fouling, and current on deposition site and survival of embryos of a dorid nudibranch gastropod. Mar. Ecol. Prog. Ser. 86: 205-215. Child. C. M. 1927. Developmental modification of the larval stage in the ascidian Corella willmeriana. J. Morphol. 44: 467-517. Cloney, R. A. 1990. Larval tunic and the function of the test cells in ascidians. Acta Zoo/. 71: 151-159. Cloney, R. A. 1994. Test cell secretions and their functions in ascidian development. Pp. 77-95 in Development of Marine Invertebrates. W. H. Wilson. S. A. Strickler. and G. L. Shinn. eds. Johns Hopkins University Press. Baltimore. Cloney, R. A., and M. J. Cavey. 1982. Ascidian larval tunic: extraem- bryonic structures influence morphogenesis. Cell Tissue Res. 222: 547-562. Cohen, J. 1988. Sttitixtical Power Analysis for the Behavioral Sciences. Earlbaum. Hillsdule, New Jersey. 567 pp. Cohen. S. 1990. Outcrossing in field populations of two species of self-fertile ascidians. J. Exp. Mar. Biol. Ecol. 140: 147-158. Damkaer. D. M.. D. B. Dey, G. A. Heron, and E. F. Prentice. 1980. Effects of UV-B radiation on near-surface zooplankton of Puget Sound. Oecologia 44: 149-158. Kndean, R. 1961. The test of the ascidian, Phallusia mammillata. Quart. J. Microsc. Soc. 102: 107-117. Goodbody, I. 1963. The biology of Ascidia nigra. II. The development and survival of young ascidians. Biol. Bull. 124: 31-44. Goodbody, I., and J. Gibson. 1974. The biology of Ascidia nigra (Savigny) V. Survival in populations settled at different times of the year. Biol. Bull. 146: 217-237. Hardy, J.. and H. Gucinski. 1989. Stratospheric ozone depletion: im- plications for marine ecosystems. Oceanography. November 1989: 18-21. Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54: 187-211. Hiius, J. 1939. The effect of light on the spawning in ascidians. Avh. Nor. Vidensk. Akad. Oslo. Mat. Naturvidensk. Kl. 4: 5-49. Jeffrey, W. R. 1990. Ultraviolet irradiation during ooplasmic segrega- tion prevents gastrulation, sensory cell induction, and axis formation in the ascidian embryo. Dev. Biol. 140: 388-400. Jokiel, P. L. 1980. Solar untraviolet radiation and coral reef epifauna. Science 207: 1069-1071. Karentz, D. 1994a. Ultraviolet tolerance mechanisms in Antarctic ma- rine organisms. Antarct. Res. Ser. 62: 93-1 10. Karentz, D. 1994b. Prevention of ultraviolet radiation damage in Ant- arctic marine invertebrates. NATO ASl Series 1 18: 176-179. Karentz, D., and L. H. Lutze. 1990. Evaluation of biologically harmful ultraviolet radiation in Antarctica with a biological dosimeter designed for aquatic environments. Limnol. Oceanogr. 35: 549-561. Karentz, D., F. S. McEuen, M. C. Land, and W. C. Dunlap. 1991. Survey of mycosporine-like amino acid compounds in Antarctic marine organisms: potential protection from ultraviolet exposure. Mar. Biol. 108: 157-166. Kirk. J. T. O., B. R. Hargreaves, D. P. Morris, R. B. Coffin, B. David. D. Grederickson, D. Karentz, D. R. S. Lean, M. P. Lesser, S. Madronich, J. H. Morrow, N. B. Nelson, and N. M. Scully. 1994. Measurements of UV-B radiation in two freshwater lakes: an instru- ment intercomparison. Arch. Hydmbiol. Beih. 43: 71-99. Kobayashi, J., H. Nakamura, and V. Hirata. 1981. Isolation and structure of a UV-absorbing substance 337 from the ascidian Halocyn- thui roret-i. Tetrahedron Lett. 22: 3001-3002. Lambert, C. C., and G. Lambert. 1978. Tunicate eggs utilize ammo- nium ions for flotation. Science 200: 64-65. Lambert, C. C., I. M. Lambert, and G. Lambert. 1995. Brooding strategies in solitary ascidians: Corella species from north and south temperate waters. Can. J. Zoo/. 73: 1666-1671. Lambert, C. C., G. Lambert, G. Crundwell, and K. Kantardjieff. 1998. Uric acid accumulation in the solitary ascidian Corella inflata. J. Exp. Zoo/. 282:323-331. Lambert, G. 1968. The general ecology and growth of a solitary ascid- ian. Corella willmeriana. Biol. Bull. 135: 296-307. Lambert, G., C. C. Lambert, and D. P. Abbott. 1981. Corella species in the American Pacific Northwest: distinction of C. inflata Huntsman. 1912 from C. willmeriana Herdman. 1898 (Ascidiacea. Phlebo- branchia). Can. J. Zoo/. 59: 1493-1504. Mitchell, D. L., and D. Karentz. 1993. The induction and repair of DNA photodamage in the environment. Pp. 345-377 in Environmental 104 B. L. BINGHAM AND N. B. REYNS UV Photobiology. A. R. Young. J. Moan, L. O. Bjorn, and W. Nultsch, eds. Plenum Press, New York. Monniot, K. R. Martoja, M. Truchet, and F. Frohlich. 1992. Opal in ascidians: a curious bioaccumulation in the ovary. Mar. Bio!. 112: 283-292. Olson, R. R. 1983. Ascidian-Pror/i/oron symbiosis: the role of larval photoadaptations in midday larval release and settlement. Biol Bull 165: 221-240. Parsons, T. R., Y. Maita, and C. M. Lalli. 1984. A Manual of Chemical and Biological Methods for Seawater Analysis. Pergamon Press, New York, New York. 173 pp. Pennington, J. T., and R. B. Emlet. 1986. Ontogenetic and diel vertical migration of a planktonic echinoid larva. Dendraster excentricus (Eschscholtz): occurrence, causes, and probable consequences. J. Exp. Mar. Biol. Ecol. 104: 69-95. Shick, J. M., M. P. Lesser, and W. R. Stochaj. 1991. Ultraviolet radiation and photooxidative stress in zooxanthellate Anthozoa: the sea anemone Phyllodiscus semoni and the octocoral Clavularia sp. Symbi- osis 10: 145-173. Skoog, D. A., and D. M. West. 1974. Analytical Chemistry. Holt, Rinehart, and Winston, New York. 598 pp. Sokal, R. R., and F. J. Rohlf. 1995. Biometry. W. H. Freeman, New York. 887 pp. Svane, I., and C. M. Young. 1989. The ecology and behaviour of ascidian larvae. Oceanogr. Mar. Biol. Annu. Rev. 27: 45-90. Van Name, W. 1945. North and South American ascidians. Bull. Amer. Mus. Nat. Hist. 84: 1-476. Worrest, R. C. 1982. Review of the literature concerning the impact of UV-B radiation upon marine organisms. Pp. 429-457 in The Role of Solar Ultraviolet Radiation in Marine Ecosystems. J. Calkins, ed. Plenum Press, NY. Young, C. M. 1982. Larval behavior, predation, and early post-settle- ment mortality as determinants of spatial distribution in subtidal soli- tary ascidians of the San Juan Islands, Washington. Ph.D. dissertation. University of Alberta, Edmonton, Alberta, Canada, 260 pp. Young, C. M. 1985. Abundance of subtidal solitary ascidians in the San Juan Islands, Washington, as influenced by food preferences of the predatory snail Fusitriton oregonensis. Mar. Biol. 84: 309- 321. Young, C. M. 1988. Ascidian cannibalism correlates with larval behav- ior and adult distribution. J. Exp. Mar. Biol. Ecol. 117: 9-26. Young, C. M., and F. S. Chia. 1985. An experimental test of shadow response function in ascidian tadpoles. J. Exp. Mar. Biol. Ecol. 85: 165-175. Reference: Biol. Bull. 196: 105-112. (February. 1999) Time in Residence Affects Escape and Agonistic Behavior in Adult Male American Lobsters S. I. CROMARTY*, J. MELLO, AND G. KASS-SIMONf Biological Sciences Department, University of Rhode Island, 100 Flagg Road, Kingston, Rhode Island 02881-0816 Abstract. Acquisition and retention of a shelter by a lob- ster are two of the variables that play a role in lobster agonistic interactions. Since shelter procurement and reten- tion are important for lobster survival, behaviors related to this activity frequently outrank other daily behaviors (e.g., searching for food). Here, we examine the effects of time in residence on the parameters of the escape response of the American lobster, Homarus americanus. Adult male intermolt lobsters (Stage C 4 ) were placed in an experimental tank for three different time periods (one hour, 24 hours, and 48 hours). The probability of eliciting an escape response was inversely related to the time spent in the tank. Eighty percent of the animals in residence for 1 h tailflipped in response to a threat, whereas only 14% of the animals in residence for 48 h tailflipped. There were also significant changes in some of the pa- rameters of the escape response among animals in residence for 24 h compared to those in residence for 1 h. The number of tailflips and the distance traveled were reduced, although frequency, velocity, acceleration, force, and work factors were not significantly different. Furthermore, with increased time in residence, lobsters switched from an avoidance or escape-prone behavior to an aggressive-prone behavior. Most of the animals in residence for 48 h approached and attacked a threat-stimulus rather than fleeing from it. On an empirically defined "index of aggressiveness," in which various behaviors were numeri- Received 22 July 1998; accepted 14 October 1998. * Present address: Department of Neurobiology, Harvard Medi- cal School, 220 Longwood Avenue, Boston, MA 02115. E-mail: Stuart_Cromarty@hms. harvard.edu t To whom correspondence should be addressed. E-mail: avflOl @uriacc. uri.edu Abbreviations: FEP. Fisher exact probability test; SSI, subsequent swims of first half; SS2, subsequent swims of second half. cally ranked from least aggressive (0) to most aggressive (6), animals residing in the tank for 1 h had an average index value of 0.1 compared to a value of 5.0 for animals in residence for 48 h. These findings are consonant with the suggestion that lobsters that have occupied a given space for an extended period of time take possession of the site and defend it instead of fleeing when threatened with a threat-inducing stimulus; it supports the idea that shelter retention increases aggressiveness and diminishes avoidance behaviors. Introduction Shelter selection by American lobsters, as they make the transition from their pelagic phase to a benthic existence, has been studied in detail (Cobb and Wahle, 1994; a re- view). Postlarval and early juvenile lobsters actively seek out cobble and boulder habitats for shelter a habitat that offers the most protection from predators and conspecifics as well as protection against storm surges and rapid currents (Hudon, 1987; Abe et ai, 1988; Incze and Wahle, 1991; Wahle and Steneck, 1991, 1992). Such shelters may be limited in number since cobble and boulder habitats com- prise only 10% of the sea-floor (Kelley, 1987) the habitat that is most favored by lobsters. In addition, as lobsters grow (up to five orders of magnitude increase in body mass), their aggressiveness increases and their behavior becomes less cryptic, so crowding may become more of a problem (Cobb and Wahle, 1994). Shelter procurement and retention may therefore be an important determinant of overall behavior in lobsters, with dominant individuals procuring and retaining the most fa- vorable shelters, which in turn may be important for an individual's survival (Cobb, 1971; Stewart, 1972). Re- cently, Spanier and others (1998) assessed the behavior of juvenile American lobsters under predation risk in labora- 105 106 S. I. CROMARTY ET AL. lory settings. In the presence of a predator fish, the tautog, Tautoga onitis, the dominant lobster appeared to guard the available shelters; subdominant lobsters, which did not guard shelters, had a mortality rate seven times higher than that of dominant lobsters. Moreover, field observations by O'Neill and Cobb (1979) showed that intruders were less able to procure a shelter if the current resident had occupied it for a certain period of time. In laboratory settings, factors that establish the domi- nance of an individual over other conspecifics and allow dominant individuals to have an advantage in procuring, capturing, and more importantly holding on to a suit- able shelter include a greater carapace length and claw size (Scrivener, 1971), the sex of the individual (male), and molt stage (O'Neill and Cobb, 1979). O'Neill and Cobb (1979) found that shelter familiarity did not affect shelter procurement or retention in the labo- ratory, but that in the field, lobsters already in a shelter were more likely to retain that shelter. On the other hand, Amer- ican lobsters have been observed to display increased ag- gressiveness after being isolated in individual tanks (E. Kravitz, pers. comm. to S.C.), and during our earlier studies, we observed that it was harder to induce animals to tailflip after they had been kept for some time in isolated tanks. Therefore, the question arises as to the extent to which an animal's time in residence is a significant determinant of its aggressive and avoidance behavior. To explore this ques- tion, we placed animals in an isolated experimental tank for periods of 1, 24, or 48 h and videotaped their responses to a nontactile threatening stimulus introduced into the tank. We now present evidence that with increased time in residence, lobsters not only are less likely to flee from a threat, but also will confront it with increasing aggressive- ness. Materials and Methods Procedures and experimental protocols are essentially the same as those described in Cromarty et al., 1991, 1998), but are summarized here with relevant differences included. Animals Adult American lobsters (carapace length 78 to 84 mm) were obtained directly from an offshore lobster vessel fish- ing in Narragansett Bay, Rhode Island. Animals were housed at the Narragansett Bay Campus of the University of Rhode Island in separate, but connecting, tanks in a free- flow seawater system at ambient temperatures ranging from 16 to 22C under a 14-h-light/10-h-dark illumination cycle. The animals were fed three times per week on a mixed diet of squid, crab, and fish, but were not fed for 48 h prior to an experiment. Six hours before an experiment, an animal was moved to the Kingston campus of the university, where it was placed in a holding tank (30 cm 3 ). The tank was supplied with its own air supply, and water was obtained from the same source used to supply the tanks at the URI Bay Campus. Experiments Experiments were conducted from May to October, to avoid possible seasonal differences in behavior. Lnenicka and Zhao (1991) documented seasonal differences in the physiology and morphology of crayfish neuromuscular terminals which suggested that lobster escape parameters might differ seasonally. Experiments were carried out between 1200 and 1700 h in an aquarium filled with filtered recirculated seawater from Narragansett Bay. Salinity was measured before each experiment, and ranged between 29%c and 33%o. Water was replaced or added as necessary to maintain salinity within this range. The experimental tank was kept between 18C and 20C by a Frigid Units AE-234 AG-602 chiller. The chiller was turned off before the start of the experiment. The experimental area consisted of an open-ended tank (1.0 m L X 0.3 m W X 0.3 m H) immersed in a larger main tank (2.2 m L X 0.75 m W X 0.91 m H). The layout was designed so that a threatening stimulus could be introduced at the open end of the experimental tank. A weighted wooden partition with a pulley acted as a blind and a separation from the main tank at the open end. To ensure that lobsters were initially at the closed, non-stimulus end, a light was placed at the open end, causing the lobster to move towards the darkened closed end. The partition was raised once the lobster had reached the closed end, while the light was moved to the closed non-stimulus end. This served to "push" the animal back towards the open (stimulus), now darkened, end. A piece of PVC tubing (0.15 m L X 0.10 m W) weighted to 1 .45 kg with pebbles served as the threat- ening visual stimulus. The stimulus was raised above the open end and was released into the water at a preset distance of 10 cm (measured from the open edge of the tank to the lobster) whenever a lobster approached the open end after the designated residency period. One hour before the ex- periment, the physical condition of each animal was checked. Animals were used only if they moved around the tank or exhibited antennule flicking. Cameras were placed in two positions (a Sony camcorder above the tank and a Panasonic WV-CD20 camera to the side), and experiments were recorded (Panasonic AG-6010 and Panasonic NV-8950) simultaneously from the vertical and horizontal perspectives. Video recordings of each lob- ster were analyzed frame-by-frame. For measurements of distance traveled, a metric grid divided into 0.5-cm units was painted onto the side of the experimental tank. Trans- parent overlays on the video monitor were later used to record the escape swimming distance of each animal. Dis- tance traveled along the length of the tank was measured by DURATION OF SITE RESIDENCY INFLUENCES LOBSTER BEHAVIOR 107 using the position of the tip of the lobster's rostrum as a guide, and the number of tailflips was counted; time was automatically recorded on the videotape. An independent observer inspected all recordings and rejected runs in which the experimental parameters were not strictly adhered to (e.g., cases in which the stimulus was released closer than 10 cm to the experimental animal). After each experiment, the animal's molt stage was de- termined by examining cuticular changes and setal devel- opment in the pleopods (Aiken, 1973, 1980). Only stage C 4 (intermolt) animals were used, since probability of escape depends on the molt stage of each lobster (Cromarty et al, 1991, 1995). Measurements of carapace length, cutter length, lobster weight and volume, temperature, and salinity were recorded at the end of each experimental trial. Analysis of the escape response follows our previous protocol (Cromarty et al, 1991, 1998). To analyze the escape parameters, the response was broken into two ele- ments the initial tailflip, henceforth called the "power swim," followed by the numerous subsequent tailflips, called "subsequent swims." (The number of subsequent swims in this study ranged from one to six.) A tailflip, or swim, is defined as beginning immediately after the start of abdominal flexion and ending at abdominal extension. The following characteristics of the escape response were analyzed for each lobster: distance traveled (centimeters), number of tailflips, duration (seconds), frequency of tailflips (tailflips per second), velocity (meters per second), acceler- ation (meters per second squared), force (mass X acceler- ation), work (force X distance), distance swum per weight (meters per kilogram), and distance swum per lobster body length. The last two parameters were measured to determine whether individual lobster variability in weight and size could alter the significance of a parameter, even though weight and size were not significantly different (using an ANOVA) among the animals in the three resident periods. As in previous analyses, in evaluating acceleration, the added-mass forces (Batchelor, 1967) that act on accelerating bodies in fluids were ignored since these are a multiple of mass and would act equally on all animals of the same weight (see Cromarty et al., 1991). The analysis of the escape response is designed to reflect relative changes in lobster escape behavior and not the kinematic relationships investigated by other researchers (Batchelor, 1967; Daniel and Meyhofer, 1989; Nauen and Shadwick, 1993). Each of the escape parameters was analyzed for ( 1 ) the entire escape response; (2) the initial power swim; (3) the subsequent swims over the entire subsequent swimming distance; and (4) the subsequent swims in each half of that distance, since earlier experiments showed that there were differences in the total subsequent swimming distance trav- eled by lobsters. We therefore divided the distance traveled in the subsequent swims by half and analyzed each half (Cromarty et al., 1991, 1998). Because the distances dif- fered and because each distance was divided equally in half for each escape sequence for each animal, no data are available to compare distance traveled between the two halves of the subsequent swims for each residency group. To quantify the degree of "aggression" in the post-stim- ulus behavior of each animal, we ranked this behavior on a scale of to 6 and subjectively ordered behavior towards the stimulus as follows: = back away, never approach 1 = approach but remain more than one bodylength away 2 = approach within one bodylength 3 = approach, touch 4 = approach, touch, grasp 5 = approach, touch, grasp, and tug/pull 6 = approach, touch and grasp, tug/pull, and an offensive tailflip Statistical analysis Differences in weight, carapace length, and cutter size among the three residency periods were determined by parametric analysis of variance (ANOVA). The Fisher exact probability test (FEP) was used to determine differences in the probabilities of escape among the three resident periods. Due to a non-normal distribution of data, Kruskal-Wallis (KW) tests were run for all the escape parameters except the subsequent swims and the "aggression intensity index." The first and second halves of the subsequent swims were com- pared with a multiple analysis of variance (MANOVA) and a repeated measures follow-up test. A trend was considered to exist if the P value for a parameter ranged between 0.05 < P < 0. 1 . Values were considered significant at P < 0.05 for all the statistical tests. Both ANOVAs and MANO- VAs were run on SPSS software ver. 6.6.1 (SPSS Inc., Chicago). Results Weight (in grams)/carapace length (in millimeters )/cutter length (in millimeters) There were no significant differences in the weights (mean SEM for all) of 1-h (415 29), 24-h (414 32), and 48-h (427 31) resident lobsters (ANOVA, F(2, 27) = 0.35, P = 0.71). No significant differences were found in the carapace lengths of 1-h (78 4.9), 24-h (79 2), and 48-h (79 3) resident lobsters (ANOVA, F(2, 27) = 0.17, P = 0.84); similarly there were no significant differences in the cutter lengths of 1-h (107 3), 24-h (109 3), and 48-h ( 108 2) resident lobsters (ANOVA, F(2, 27) = 0.69, P = 0.61). Effects of residence time on the probability of escape Probability of escape. The probability of escape for ani- mals in the three residence time periods is summarized in 108 S. I. CROMARTY ET AL. Table I Escape and post-threat behavior of animals in each residency period: some lobsters initially tailftipped and then re-approached the stimulus or re-approached and attacked it Immediate response Secondary responses Back up Approach only Approach and Attack Residence period (h) n Tailflip 1 24 48 10 13 7 8 8 1 9 6 1 7 7 5 7 Table I. Of the 1-h, 24-h, and 48-h resident lobsters, 8 out of 10, 8 out of 13, and 1 out of 7, respectively, escaped when the stimulus was introduced. These probabilities were significantly different (FEP, P = 0.001). Escape parameters for one- and 24-hour resident male adult lobsters. Because only one lobster among the 48-h resident group could be induced to tailflip, only 1-h and 24-h residents were compared. Three lobsters in the 1-h resident period were not analyzed, because their gross swimming pattern deviated from a rectilinear motion; thus five 1-h lobsters and eight 24-h lobsters were used in the analyses. Total escape response (initial power swim plus subse- quent swims). One-hour resident lobsters swam farther (KW, x 2 = 6.19, P = 0.012) and took more tailflips (KW, X 2 = 5.67, P = 0.017) than did 24-hour resident animals (Fig. 1A, B, respectively). Distance-dependent parameters, such as distance traveled per body length and distance traveled per weight, were significantly shorter among the 24-h lobsters (KW, P = 0.02; Fig. 1C, D). Although time spent escaping, frequency of tailflips, velocity, acceleration, force, and work were not significantly different at the P ^ 0.05 level between the two groups (KW, P > 0.05), a trend at 0.05 < P < 0.10 towards a decrease in time, frequency, and velocity was observed for the 24-h resident lobsters (Fig. 2A, B, and C). Initial power swim. There was a significant decrease in the acceleration, force, and work of the power stroke in lobsters that were in residence for 24 h (KW, P < 0.05; Fig. 3A, B, and C) and a significant increase in the duration of the initial power stroke (KW, P = 0.014; Fig. 3D). A trend towards decreased distance traveled and lower velocity in the 24-h resident lobsters was observed (KW, 0.05 < P < 0.10). Total subsequent swims. Five males in the 1-h category and six males in the 24-h category were compared, since two lobsters in the 24-h resident period took only one tailflip with no subsequent swims. Significant differences in sub- sequent swim duration were found between animals in the 1-h to 24-h resident periods. Significant differences in sub- sequent swim duration (MANOVA, F(l, 9) = 3.92, P = 0.04), and number of tailflips (MANOVA, F(l, 9) = 7.42, P = 0.02) were found, with 24-h resident lobsters spending less time tailflipping and taking fewer tailflips (Fig. 4A, B). Post-threat behavior In the analysis of post-threat behaviors, the percentage of lobsters that re-approached the stimulus was 0% (0 out of 10), 54% (7 out of 13), and 100% (7 out of 7), in the residency periods of 1 h, 24 h, and 48 h, respectively (Table I). Of these, 0% (0 of 10) of the 1-h, 38% (5 of 13) of the 24-h, and 100% (7 of 7) of the 48-h lobsters approached and attacked the stimulus with high intensity after it had been o.o 0.2 D.s TOTAL DISTANCE (m) 0.8 1.0 o- 246 TOTAL NUMBER OF TAILFLIPS 1 BODYLENGTHS TRAVELED o- 01234 TOTAL DISTANCE PER WEIGHT (m/Kg) Figure 1. Parameters of the entire escape response exhibited by lob- sters in residence for 1 h ( = 5) and 24 h (n = 8). An asterisk (*) indicates results that are significantly different at P & 0.05. (A) Mean distance traveled in meters (m). (B) Mean number of taiiflips. (C) Mean distance traveled divided by lobster body length; represented as number of body lengths traveled. (D) Mean distance traveled in meters (m) divided by lobster weight in kilograms (kg). DURATION OF SITE RESIDENCY INFLUENCES LOBSTER BEHAVIOR 109 E i _ e X c - e 0. X = o c 50 i a.s i.o i.s 2.0 TOTAL TIME ' ' 0123456 TOTAL FREQUENCY (TF/S) Figure 2. Parameters of the entire escape response exhibited by lob- sters in residence for 1 h (n = 5) and 24 h (n = 8). (A) Mean time in seconds (s). (B) Mean frequency of tailflips in tailflips per second (TF s~'). (C) Mean velocity in meters per second (m s" 1 ), acceleration in meters per second squared (m s" 2 ), force in newtons (kg m s 2 ), and work in joules (J) of tailflips. presented. On the aggression intensity index, (with prede- termined levels of aggressive behavior, ranked from to 6, see Methods), behavioral responses after the stimulus was introduced were statistically different (Fig. 5A-C), with lobsters in residence for 48 h displaying more intensively aggressive behaviors (average index of 5.0 0.8) than those in residence for 24 h (average index of 2.1 2.2; KW, P = 0.03) and 1 h (average index of 0.1 0.3; P = 0.006). Although it is possible that stress influenced these behav- iors, no physical evidence of stress-related behavior was observed. Animals in all three residency periods exhibited typical active searching and antenna whipping behavior. (Lobsters that exhibit stress show no movement in the experimental tank and do not antenna whip [pers. obs.].) In addition, lobsters in the 48-h residency period showed a distinct change in behavior compared to the 24-h resident animal, exhibiting more grasping, pulling, and offensive tailflips. After the experiment, all animals (while still in the experimental tank) ate the food presented to them, indicat- ing that they were not under stress, since sick or stressed animals avoid food altogether (pers. obs.). Discussion In this study, we have shown that lobsters residing in an experimental tank for 24 or 48 hours exhibit a reduced tendency to escape in response to a threat and an increase in post-threat aggressiveness. o 93 o- 10 POWER STROKE ACCELERATION (m/s/s) I 99 o- - 1 2 POWER STROKE FORCE (Kgm/s/s) o- 50 0.00 O.OS 0.10 0.1S 0.20 POWER STROKE WORK (J) 0.25 D 0.0 0.1 0.2 0.3 0.4 0.5 POWER STROKE TIME (s) 0.6 Figure 3. Parameters of the power stroke exhibited by lobsters in residence for 1 h (n = 5) and 24 h (n = 8). An asterisk (*) indicates results that are significantly different at P s 0.05. (A) Mean acceleration in meters per second squared (m s~"). (B) Mean force in newtons (kg m s~"). (C) Mean work in joules (J). (D) Mean duration in seconds (s). 10 S. I. CROMARTY ET AL. O so o SSTIMR.l SSTIME2 0.0 0.2 0.5 0.8 TIME (s) 1.0 1.2 X O 93 O C3 SSTFl D SSTF2 NUMBER OF TAILFLIPS Figure 4. Parameters of subsequent swims of the escape response exhibited by lobsters in residence for 1 h (;i = 5) and 24 h (11 = 8). An asterisk (*) indicates results that are significantly different at P s 0.05. (A) Mean duration in seconds (s) for the first (SSTIME1) and second (SS- TIME2) halves of the subsequent swimming distance. (B) Mean number of subsequent swims in the first (SSTFl) and second (SSTF2) halves of the subsequent swimming distance. B L. 01 0.1 0.3 (* ##) jg *" O J 6- 1 " ft- m 6- 4- 2- The measured parameters of the escape response (dis- tance traveled, number of tailflips, acceleration, force, etc.) were all reduced in the 24-h residency group, while time spent escaping was increased. The efficacy of the initial power stroke was reduced, resulting in animals that ap- peared reluctant to escape. The power swim took longer to complete and the distance traveled was reduced, resulting in a lower initial velocity and acceleration (Figs. 1 and 2). Lobsters residing in the tank for 48 h simply did not tailflip, and therefore their escape response sequences could not be compared to those of 1 -h and 24-h residency animals. The post-stimulus behavior of the 24-h and 48-h resident lobsters was different from that of the 1-h residents. None of the latter attacked the stimulus, but 38% of the 24-h and 100% of the 48-h lobsters attacked the stimulus with high intensity, as reflected in the aggression intensity index. In particular, the 48-h resident lobsters exhibited the highest intensity of aggressive behaviors towards the stimulus (Fig. 5C). Therefore, as residency period was increased from 1 h to 48 h (in an experimental tank), escape behavior decreased and directed aggression towards the stimulus increased. Our experiments were designed to test whether time in residence affects the probability that a lobster will respond with an escape response when threatened. Since animals tailflipped when initially caught and again when placed in the experimental tank, we tried to minimize handling by capturing them with a net (in their holding tanks) and then moving them directly to the experimental tank. To avoid handling-induced arousal, we did not re-handle 24-h and 48-h lobsters once they were placed in the experimental tank. In preliminary experiments we found that any attempt to recapture an animal caused it to tailflip wildly around the tank. In any event, although the interval between handling and presentation of a threat may have been different in the three groups, it is unlikely that a single prior handling would 2.1 2.2 (* 5.0 0.8 (' ## ) 3456 0123456 0123456 Aggression Intensity Index Figure 5. Average aggression intensity index of post-threat behaviors directed towards the stimulus by lobsters in residence for 1 h (A). 24 h (B). and 48 h (C). Index values range from (back away, never approach) to 6 (grasp, pull, tug, and aggressive tailflip). The asterisks and number symbols indicate the level of significance: * 1 h < 24 h; P = 0.03; ## 1 h < 48 h; P = 0.0006; ** 24 h < 48 h; P = 0.009. DURATION OF SITE RESIDENCY INFLUENCES LOBSTKR BEHAVIOR produce significant differences in the level of arousal in 24-h and 48-h residents. Nonetheless, if prior handling contributed to the observed time-dependent reluctance to escape and increase in aggres- siveness, then our findings suggest that an animal becomes more aggressive the longer it is left undisturbed in a site. Many other factors appear to affect an animal's ability to procure and retain a shelter (O'Neill and Cobb, 1979). These include differences in dominance due to body length, weight, and claw size (Scrivener. 1971). Reproductive sta- tus is also important. In a related decapod, maternal female crayfish residents won 92% of their encounters with male intruders (Figler el ul.. 1995). Gravid lobsters exhibit a distinct reduction in escape behavior and are more likely than males to attack an approaching stimulus (Cromarty et al.. 1998). Odor cues for discrimination of familiar and dominant lobsters (Karavanich and Atema, 1991. 1993, 1998a, b) and sex-identifying urine and molt signals in lobsters (Atema and Cowan, 1986), plus visual cues (Bruski and Dunham, 1987), are most likely important sensory cues for procuring and residing in a shelter. Crayfish rapidly learn to discriminate between changing spatial configura- tions (Sandeman and Varju, 1988; Varju and Sandeman, 1989; Basil and Sandeman, 1997) and. as shown by direct measurements of electrical heart activity (Shuranova and Burmistrov, 1995), are constantly sampling their surround- ings (i.e., predators or conspecific intruders, currents, shad- ows, food availability, etc.). An important determinant in shelter retention may be social contact with conspecifics. Hoffman et al. (1975) showed that lobsters who were in visual contact with other animals or were communally housed were less aggressive. In this regard, Yeh et ul. (1996, 1997) found that crayfish social status and social experience determine the effect of serotonergic modulation on the lateral giant motor neuron that mediates one form of escape behavior. Social isolation has also been found to cause dramatic increases in intraspe- cific aggression in mice (Valzelli, 1973). In an early study, O'Neill and Cobb (1979) found that in the laboratory, shelter familiarity did not affect shelter pro- curement or retention, whereas in the field, resident lobsters were more likely to retain their shelters. These observed differences might have been due to the above-mentioned experimental conditions, i.e.. type and duration of the ex- perimental housing of animals. In our experimental conditions there were no other ani- mals, no places to seek shelter, and only one avenue of escape. Therefore, an animal's immediate response under these circumstances may be considered to reflect and be driven by the differential effects of having been undisturbed in a familiar environment for different lengths of time. That is, our experiments seem to reveal that the patterns associ- ated with either aggression and dominance, on the one hand, or avoidance and submission, on the other, become mani- fested as a function of how the animal assesses its own place in its immediate environment. This assessment then be- comes a determinant of how an animal responds to other conspecifics and might be considered operationally as a "motivational state." A change in "motivational state" has recently been tested in the European hermit crab Pagurus bernhardus during agonistic encounters (Elwood et al.. 1998). This study showed that the duration and severity of the startle/threat response are inversely related to the "mo- tivational state" of the animal to continue the previous activity, namely of fighting for a more suitable shelter inhabited by a conspecific. In summary, our experiments indicate that time in resi- dence and isolation are important physiological determi- nants of a lobster's behavior and can cause it to switch from an avoidance-prone to an aggression-prone state. Acknowledgments We thank Bill Mac Elroy for allowing us to collect nongravid and male animals while he was fishing offshore and Tom Angell of the Rhode Island Department of Envi- ronmental Management for collecting gravid animals off- shore. (All lobsters were returned to the Bay after experi- mentation was completed.) Thanks to Dr. Mike Clancy and Ms. Kathy Castro for help in lobster collection and main- tenance. Drs. Stanley Cobb and Frank Heppner kindly pro- vided laboratory space and equipment. We also thank Ms. Malia Schwartz for critiquing an earlier draft of the manu- script. This research was supported by a Whitehall Founda- tion grant to G.K-S and Grant-in Aid of Research from Sigma Xi and Lerner Gray Grants for Marine Research to S.I.C. Literature Cited Able, K. W., K. L. Heck, M. P. Fahay, and C. T. Roman. 1988. Use of salt-marsh peat reefs by small juvenile lobsters on Cape Cod, Massachusetts. Estuaries 11:83-86. Aiken, D. E. 1973. Proecdysis. setal development, and molt prediction in the American lobster (Homarus americanus). J. Fish. Res. Board Can. 30:1334-1337. Aiken, D. E. 1980. Molting and growth. Pp. 91-163 in The Biology and Management of Lobsters, vol. 1, J. S. Cobb and B. F. Phillips, eds. Academic Press. New York. Atema, J., and D. Cowan. 1986. Sex-identifying urine and molt signals in lobster (Hinuru\ umericanus). J. Chem. Ecol. 12:2065-2080. Basil, J., and D. Sandeman. 1997. Australian freshwater crayfish detect small spatial changes in their habitat using only tactile cues. Sue. Neurosc. Abslr. 23:527.17. Batchelor, G. K. 1967. An Introduction to Fluid Dvnamics. Cambridge University Press, London. Bruski, C. A., and 1). W. Dunham. 1987. The importance of vision in agonistic communication of the crayfish Orconecles rusticus. I. An analysis of bout dynamics. Behaviour 63:83-107. Cobb, J. S. 1971. The shelter behavior of the American lobster. Homarus americanus. Ecology 52:108-115. 112 S. I. CROMARTY ET AL. Cobb, J. S., and R. A. Wahle. 1994. Early life history and recruitment processes of clawed lobsters. Cruslaceana 67:1-25. C'romarty, S. I., J. S. Cobb, and G. Kass-Simon. 1991. Behavioral analysis of the escape response in juvenile American lobsters, Homarus americanus over the molt cycle. J. Exp. Biol. 158:565-58 1 . C'romarty, S. I., J. S. Cobb, and G. Kass-Simon. 1995. Adult and juvenile lobsters (Homarus americanus). differ markedly in the neuro- muscular physiology and behavior of the escape response over the molt cycle. Soc. Neurosci. Abstr. 25:201. Cromarty, S. I., J. Mello, and G. Kass-Simon. 1998. Behavioral anal- ysis of escape behavior in male, and gravid and non-gravid, female lobsters. Biol. Bull. 194:63-7 1 . Daniel, T. L., and E. Meyhofer. 1989. Size limits in escape locomotion of caridean shrimp. J. E-\p. Biol. 143:245-265. Elwood, R. W., K. E. Wood, M. B. Gallagher, and J. T. A. Dick. 1998. Probing motivational state during agonistic encounters in animals. Nature 393:66-68. Figler, M. H., M. Twum, J. E. Finkelstein, and H. V. S. Peeke. 1995. Maternal aggression in red swamp crayfish (Procambarus clarkii. Girard): The relation between reproductive status and outcome of aggressive encounters with male and female conspecifics. Behaviour 132:107-125. Hoffman, R. S., P. J. Dunham, and P. V. Kelly. 1975. Effect of water temperature and housing conditions upon the aggressive behavior of the lobster Homarus americanus. J. Fish. Res. Board Can. 32:713-717. Hudon, C. 1987. Ecology and growth of postlarval and juvenile lobster Homarus americanus. off lies de la Madeleine (Quebec). Can. J. Fish. Auuat. Sci. 50:1422-1434. Incze, L., and R. A. Wahle. 1991. Recruitment from pelagic to early benthic phase in lobsters (Homarus americanus). Mar. Ecol. Progr. Ser. 79:77-87. Karavanich, C., and J. Atema. 1991. Role of olfaction in recognition of dominance in the American lobster (Homarus americanus). Biol. Bull. 181:359-360. Karavanich, C., and J. Atema. 1993. Agonistic encounters in the American lobster, Homarus americanus: Do they remember their op- ponents? Biol. Bull. 185:321-322. Karavanich, C., and J. Atema. 1998a. Olfactory recognition of urine signals in dominance fights between male lobsters. Behavior 135:719- 730. Karavanich, C., and J. Atema. 1998b. Individual recognition and mem- ory in lobster dominance. Anim. Behav. 56. (in press). Kelley, J. 1987. A treatise on glaciated coast. Pp. 151-176 in Sedimen- tary Environments Along Maine's Estuarine Coastline, D. M. Fitzger- ald and P. S. Rosen, eds. Academic Press, New York. Lnenicka, G. A., and Y. Zhao. 1991. Seasonal differences in the phys- iology and morphology of crayfish motor terminals. J. Neurobiol. 22:561-569. Nauen, J. C., and R. E. Shadwick. 1993. Biomechanics of tail-flipping by the spiny lobster Panuliriis interruptus. Am. Zoo/. 32(5): 127A. O'Neill, D. J., and J. S. Cobb. 1979. Some factors influencing the outcome of shelter competition in lobsters (Homarus americanus). Mar. Behav. Physiol. 6:33-45. Sandeman, D., and D. Varju. 1988. A behavioral study of tactile localisation in the crayfish Cherax destructor. J. Comp. Physiol. 163: 525-536. Scrivener, J. C. E. 1971. Agonistic behavior of the American lobster, Homants americanus (Milne-Edwards). Fish. Res. Bd. Can. Tech. Rep. 235:1-128. Shuranova, Z. H., and I. U. Burmistrov. 1995. The orienting reaction in invertebrates. Zh. Vyssh. Nen: Deal. Im. IP. Pavola. 45(31:510-520. Spanier, E., T. P. McKenzie. J. S. Cobb, and M. Clancy. 1998. Be havior of juvenile American lobsters, Homarus americanus. under predation risk. Mar. Biol. 130:397-406. Stewart, L. L. 1972. The seasonal movements, population dynamics and ecology of the lobster, Homarus americanus (Milne Edwards) off Ram Island, Connecticut, Ph.D. Thesis. University of Connecticut, 1 52 pp. Valzelli, L. 1973. The "isolation syndrome" in mice. Psychopharma- cologia 31:305-320. Varju, D., and D. Sandeman. 1989. Tactile learning in a new habitat and spatial memory in the crayfish Chera.\ destructor. Page 2 la in Dynamics and Plasticity in Neuronal Systems. N. Eisner and W. Singer, eds. Thieme, Stuttgart. Wahle, R. A., and R. S. Steneck. 1991. Recruitment habitats and nursery grounds of the American lobster Homarus americanus: a demographic bottleneck. Mar. Ecol. Progr. Ser. 69:231-243. Wahle, R. A., and R. S. Steneck. 1992. Habitat restrictions in early benthic life: experiments on habitat selection and in situ predation with the American lobster. J. E.\;i. Mar. Biol. Ecol. 157:91-114. Yeh, S. R., R. A. Fricke, and D. H. Edwards. 1996. The effect of social experience on serotonergic modulation of the escape circuit of crayfish. Science 271:366-369. Yeh, S. R., B. E. Musolf, and D. H. Edwards. 1997. Neuronal adapta- tions to changes in the social dominance status of crayfish. J. Neurosci. 17:697-708. Reference: Bio/. Bull 196: 113-120. (February, 1999) Tunic Morphology and Cellulosic Components of Pyrosomas, Doliolids, and Salps (Thaliacea, Urochordata) EUICHI HIROSE 1 *, SATOSHI KIMURA 2 . TAKAO ITOH 2 . AND JUN NISHIKAWA 3 1 Department of Chemistry, Biology and Marine Science, Faculty of Science, University of the Ryukyus, Nishihara, Okinawa 903-0213, Japan: 2 Wood Research Institute, Kyoto University, Uji, Kyoto 611. Japan; and 3 Ocean Research Institute, University of Tokyo, Nakano, Tokyo 164-8639, Japan Abstract. The morphology and cellulosic composition of the tunic was studied in pelagic tunicates (3 pyrosomas, 2 doliolids, and 13 salps). The tunic is transparent and gelat- inous, consisting of an electron-dense cuticular layer with a fibrous tunic matrix. The thickness and density of the cu- ticular layer and of the tunic matrix differ from species to species. In some salps, the cuticular layer has numerous minute protrusions that are structurally identical to those found in several ascidians. Free mesenchymal cells (tunic cells) are distributed in the tunic. Whereas the number of tunic cells in the pyrosomas is similar to that in ascidians. there are many fewer tunic cells in doliolids and salps. These differences may be caused by the different functions of the tunic in each group. The existence of cellulose in the tunic was confirmed using electron diffraction in all of the species studied thus far. Their diffractograms indicate that the cellulose microfibrils consist of nearly pure 1/3 of the allomorph. These results show that tunic morphology and cellulosic composition are similar in ascidians and thali- aceans (pyrosomas, doliolids, and salps). The tunic is con- sidered to be a homologous tissue in these animals, and their most recent common ancestor would have possessed this tissue. Introduction Members of the phylum Chordata are characterized by having a notochord during some stage of development. Urochordata (also called Tunicata) is one of three subphyla Received 8 June 1998; accepted 13 October 1998. * To whom correspondence should be addressed. E-mail: euichi @sci.u-ryukyu. ac.jp in the phylum Chordata. The name Tunicata is derived from the unique integumentary tissue, called the tunic, that en- tirely covers the epidermis. The Urochordata includes three classes; all of the species possess tunic in the classes As- cidiacea and Thaliacea, whereas the presence of tunic is not well documented in the class Appendicularia. The tunic is a peculiar tissue among metazoans because of its cellulosic components (De Leo et al., 1977) and the presence of free-living cells (tunic cells) in the tunic, that is, outside the epidermis. To date, the biology and biochemistry of the tunic have been studied mainly in ascidians, sessile forms of tunicates, but they have not been well investigated in pe- lagic tunicates. In ascidians, many types of tunic cells have been de- scribed, and they are involved in various biological func- tions, such as phagocytosis (De Leo et al., 1981; Hirose et al., 1994), conduction of impulses (Mackie and Singla. 1987), contractility of the tunic (Hirose and Ishii, 1995). bioluminescence (Aoki et al., 1989; Chiba et al., 1998), photosynthetic symbiosis (Hirose et al.. 1996b), and al- lorecognition (Hirose et al.. 1997c). The tunic is overlaid by a cuticular layer that sometimes has a subcuticular layer beneath it. In several ascidian species, the cuticular surface has numerous minute protrusions that are 100 nm high or less. Descriptions of the cuticular fine structures in 116 ascidian species indicate that the presence or absence of cuticular protrusions has phylogenetic significance (cf. Hi- rose et al, 1997b). Little information has been accumulated on the tunic of pelagic tunicates, such as pyrosomas, do- liolids. and salps (reviewed in Welsh, 1984, and Bone, 1998). In this study, we investigated the tunic morphology of thaliaceans, with special attention to the distribution of the tunic cells and the fine structure of the cuticule. 113 114 E. HIROSE ET AL Almost all ascidians studied to date have been found to contain cellulose I microfibrils in the tunic (Yamamoto el al, 1989; Van Daele el al., 1992; Kimura and Itoh, 1996; Okamoto el al., 1996). In pelagic tunicates, however, re- search has thus far shown cellulose I microfibrils with high crystallinity in the tunic of only one species of Salpidae, Salpa fiisifonnis (Belton er al., 1989). It is not yet known whether other pelagic tunicates can make cellulose. Our study also focuses on the existence and characterization of cellulose in pelagic tunicates. This report deals with the tunic morphology and cellulo- sic components of 1 8 thaliacean tunicates from all orders of Thaliacea. The results provide information valuable for better understanding the diversity and evolution of the tunic in conjunction with the phylogeny of tunicates. Materials and Methods Sample collection and fixation We examined 3 species of pyrosomas, 2 species of doliolids, and 13 species of salps (see Table I), which were collected in several net tows taken southeast of Tokyo, Japan (34 11'- 3504' N, 13906'-14332' E). Apparently intact animals were sorted from the samples on board ship and prefixed immediately in 2.5% glutaraldehyde-0.45 M sucrose-0.1 M cacodylate (pH 7.4) at room temperature. Large species, such as Thefts vagina. Salpa fiisifonnis, and lasis -onaria. were dissected in the fixation medium, and the tunic near a gill bar was used for the following examinations. Microscopy for the ohsen-ation of tunic morphology After a brief rinse in 0.45 M sucrose-0.1 M cacodylate (pH 7.4), the specimens were postfixed in 1% osmium tetroxide-0.1 M cacodylate for 1.5 h. dehydrated through an ethanol series, cleared with H-butyl glycidyl ether, and em- bedded in low-viscosity epoxy resin. Thick sections were stained with 1% toluidine blue for light microscopy. Thin sections were double stained and examined in a Hitachi HS-9 transmission electron microscope at an accelerating voltage of 75 kV. In some of the prefixed specimens, the tunic was isolated from the other tissues and observed using a light microscope equipped with Nomarski differential interference contrast (DIG) and phase contrast optics. Microscop\ for the analvsis of cellulose fibers To examine replicas of the cellulose fibers, the samples were treated with 5% KOH overnight at room temperature, followed by 2 h of bleaching in 0.34% NaCIO-,. buffered at pH 4.9 in 50 mM acetate buffer, at 80C. These treatments were repeated three times, at which point the tissue became completely white. The purified tunic samples were trans- ferred to an acetyl cellulose film, air dried, then unidirec- Table I Tunic Cuticle Structures in Thaliacea Species a Cuticular protrusions' 1 Subclass Pyrosomata Order Pyrosomatida Family Pyrosomatidae Pvroslremma agassizi Pyrosoma ahemiosum Pyrosoma atlanticum Subclass Myosomata Order Doliolida ( = Cyclomyaria) Family Doliolidae Dolioletta gegenbanri (gono. I Doliohtm nationalis (gono.) Order Salpida (=Desmomyaria) Family Salpidae Cvclosalpa affiiiis (agg.) N Cyclosalpa polae (agg.) Cvclosa/pa quadriluminis forma parallela (agg.) N lasis zonaria (agg.) Metcalfina hexagona (sol.) Pegea confoederata (agg.) N Salpa fiisifonnis (sol.) Salpa fiisifonnis (agg.) Thalia dear (sol.) Thalia democratica (sol.) + Tluilia urientalis (sol.) Thetys vagina (agg.) Traustedtia multitentaculata (agg.) Reneriella retracta (sol.) N a gono. = gonozooid; agg. = aggregate zooid; sol. = solitary zooid. h = cuticular layer was not observed clearly; N = cuticular layer was not observed (too lucent or bad preservation). A plus sign indicates the presence of minute protrusions. A minus sign indicates the absence of minute protrusion (the surface of the cuticle is flat). tionally shadowed at 45 with platinum-carbon and coated with carbon at 2 X 10~ 4 Pa by a BAF 400D freeze-etch apparatus (Balzers, Liechtenstein). Replicas were cleaned in a 5% sodium dichromate-50% sulfuric acid mixture (w/v) and mounted on Formvar-coated copper grids for observa- tion with a transmission electron microscope (JEOL, JEM- 2000EXII) operating at an accelerating voltage of 100 kV. For observation of the selected area electron diffraction, purified tunic samples were mounted on carbon-coated grids after homogenization with liquid nitrogen using mortar and pestle. The electron diffraction patterns were obtained with a JEM-2000EX1I transmission electron microscope operat- ing at an accelerating voltage of 100 kV. Results Tunic morphology All of the species examined in this study have a trans- parent, gelatinous tunic that covers the epidermis (Fig. 1 ). TUNIC OF THAL1ACEANS 115 The hardness of the tunic varies among species. The tunic is very soft and fragile in some species, such as Dolioluin nationalis, Cyclosalpa /wlae. Pegea confoederata, and Ret- tcricllci retracta. The soft tunic was usually only lightly stained in both light and electron microscopic preparations, suggesting that the structural components of the tunic are present in low density in these species. The thickness of the tunic also varies from one species to another: in some species it has a thickness of several millimeters or more (e.g., Pyrosonia aherniosum, Metculfina hexugona, and Thetys vagina), whereas in others (Dolioletta gegenbauri and Dolioluin national is) it is a thin sheath of 1-2 ;u,m. However, the thickness we measured includes substantial error attributable to shrinkage of the specimens during fix- ation, embedding, and sectioning. In all of the species examined, bacteria were rarely found within the tunic. In the salps and doliolids (species of the subclass Myo- somata), we found many fewer tunic cells than in the ascidian species (Fig. 2A). The sparsely distributed tunic cells are amoeboid-shaped, with extending pseudopodia (Fig. 2, B and C). In contrast, the pyrosomas have many tunic cells of several types (Fig. 3), one of which forms a cellular network in the tunic (Fig. 4A). In histological sections, elongated forms of tunic cells that appear to cor- respond to cells of this network form a line (Fig. 4. B and C). The network probably occurs in a specific layer in the tunic. Salps have two forms of zooids in their life cycle: solitary (asexual) and aggregate (sexual) zooids. We examined the tunic of both forms in Salpa fusiformis. Although the tunic shape differs between the two forms and the tunic is usually thicker in the solitary zooids, there are no prominent differ- ences in the morphology of tunic cells or in the fine struc- ture of the tunic cuticle. Tunic cuticle The tunic cuticle is an electron-dense layer covering the tunic matrix. In some of the species that have a very soft tunic, we could not clearly distinguish the cuticular layer or the tunic matrix, or both, in thick and thin sections (Table I). Perhaps the stainability and electron density of the cuticular layer were too low to be detected or the tunic and cuticle were so fragile that they were poorly fixed or broken in these specimens. The thickness of the cuticular layer is about 10-20 nm in 8 of 10 species in which we clearly observed the fine structure of the tunic cuticle (Fig. 5, A-E). The other two species, lasis zonaria and Thetys vagina, have a cuticular layer of 0.5-1.0 ;um thick, including a subcuticular layer (Fig. 5F). The minute cuticular protrusions (about 50 nm in height or less) were seen only in Thalia democratica. Thalia orientalis, and Thetys vagina (Fig. 5, E and F: Table I). In general, when the tunic is harder, the cuticular layer is thicker and the fibrous components of the tunic matrix are stained more densely. Cellulose fihers Figure 6 shows replica images (A, C, E, and G) and electron diffractograms (B, D, F, and H) of purified tunic- fibers in Pyrosonni atlanticuni (A and B), Dolioluin nn- tionalis (C and D). lasis zonaria (E and F), and Pegea confoederata (G and H). The electron diffractograms with three to five spots (1 10, HO, 200, 002, and 004) in each figure indicate that the tunic of pyrosomas, doliolids, and salps contains cellulose I microfibrils with high crystallin- ity. No super-lattice reflections originating from triclinic crystalline cellulose (I) were observed. In some speci- mens, a 002 reflection spot originating from monoclinic crystalline cellulose (1)3) was clearly observed (Fig. 6D). The diffractograms obtained from Pyrostremma agassizi. Dolioletta gegenbauri, Salpa fusiformis, and Retteriella re- tracta are essentially identical to those shown in Figure 6. The 1 10 diffraction spot of Pyrosonia atlanticuni was often stronger than llO (Fig. 6B). Pyrosonia atlanticuni and Do- lioluin nationalis have cellulose microfibrils of 20-nm mean width (Fig. 6, A and C). whereas the microfibrils of salps are 18-nin mean width (Fig. 6, E and G). Bundles with two to six cellulose microfibrils were often observed in specimens of /. zonaria and P. confoederata (Fig. 6, E and G. arrow- heads). Discussion The functions of the tunic would be very different be- tween benthic and pelagic forms of tunicates. In benthic environments, many organisms are concentrated at high densities and diversities, so the primary functions of the tunic of benthic tunicates would be protection against pred- ators, including bacterial infections, and attachment to the substratum. The tunic might also assist in competition for space. Therefore, benthic tunicates would benefit from hav- ing a thick, hard tunic containing many tunic cells with a variety of functions that contribute to body protection. In contrast, pelagic tunicates do not need a tunic for settlement or for occupying space. Hard, heavy tunics like those pos- sessed by ascidians would be unsuitable for maintaining the neutral buoyancy of pelagic tunicates, even though they might be protective. Moreover, since pelagic tunicates are heavily preyed upon by sight predators such as fish, a thin transparent tunic that transmits light would be a distinct advantage to the pelagic forms. Thus, it is reasonable for pelagic tunicates to posses relatively soft, fragile tunics. Whereas the salps and doliolids examined in this study contained few tunic cells, pyrosomas had relative!) high numbers of tunic cells, comparable to the numbers found in ascidians. If one assumes that tunic cells have evolved primarily for protection against predators and tend to be E. HIROSE ET AL. 1A TUNIC OF THALIACEANS 17 Figure 1. Histological sections of the tunic stained with toluidine blue. The tunic matrix fills the space between the lunic cuticle (c) and the epidermis (e). (Al Pyrosoma atlanlicwn. Arrowheads indicate some tunic cells. (B) Dulitiletia gegenhauri (gonozooid) has a very thin tunic layer (indicated by two arrowheads). (C) Cyclosalpa polae (aggregate zooid). (D) Thalia tmenlalis (solitary zooid). Scale bars = 50 fim (A), 10 ^m (B-D). Figure 2. Tunic cells of Salpa fusiformis (aggregate zooid). Tunic cells (arrowheads) are sparsely distributed in the tunic (A. phase contrast). Tunic cells are amoeboid-shaped with pseudopodia (B. Nomarski differential interference contrast; C, histological section). Scale bars = 50 fim (A). 10 p,m (B and C). Figure 3. Tunic cells of Pyroslremma agassizi. Several types of tunic cells are distributed in the tunic (A, Nomarski differential interference contrast; B and C, histological sections), c. tunic cuticle. Scale bars = 50 jj.m (A). 10 /xm (B and C). Figure 4. Multipolar tunic cells form a cellular network in the tunic of Pyrostremma agassizi. (A. Nomarski differential interference contrast). Tunic cells (arrowheads) forming a line are occasionally found in the tunic and probably correspond to the cellular network (B. histological section; C enlargement of B). c, tunic cuticle; e, epidermis. Scale bars = 50 /xm (A and B). 10 jam (C). well developed in animals that lack the ability to escape, these cells should be more important to pyrosomas than to doliolids and salps. Although all three groups use jet pro- pulsion, pyrosomas swim slowly, using the water streams from the feeding currents of each zooid. whereas doliolids and salps swim quickly, using water pulses produced by the Figure 5. Transmission electron microscopy of the tunic cuticle of some thaliaceans. (A) Pyrosoma atlanticum. (B) Doliolerta gegenbauri (gonozooid). (C) lasis zonaria (aggregate zooid). (D) Metcalfina hex- agona (solitary zooid). (E) Thalia democratica (solitary zooid). (F) Thetys vagina (aggregate zooid). Arrowheads in E and F indicate cuticular protrusions. Magnifications of A, B, and D-F are identical. Scale bars = 0.2 /urn (A), 1 /urn (C). 118 E. HIROSE ET AL 004 H Figure 6. Replication images of purified cellulose microtibril (A, C, E, and G) and its electron diffractogram (B, D, F, and H) of P\r<>Mna uilanlicum (A and B). Do/in/uin nalionalis (C and D), la\is zonaria (E and F), and Pegeu canfoctlerata (G and H). All figures are of the same magnification. The diffractograms show that all specimens are composed of cellulose I microfibrils with high crystallinity. Bundles with two to six cellulose mierofibrils were often observed in /. zonaria and Pcgea confoederata (E and G. arrowheads). rapid contraction of circular muscle bands responding to stimulation (e.g.. Bone and Trueman, 1983, 1984; Nish- ikawa and Terazaki, 1994; Bone, 1998). The tunic cells in salps and doliolids have an amoeboid shape and many pseudopodia, characteristics that are sug- gestive of their motility within the tunic. Because they are very similar in morphology to the tunic phagocytes of some ascidians (cf. Hirose et /., 1994, 1996a, b), these cells might be phagocytes. Among several types of tunic cells in pyrosomas, one type forms a cellular network. This network TUNIC OF THALIACEANS 119 might facilitate coordination among the zooids in a colony. A cellular network has been described in the tunic of some colonial ascidians; it seems to be involved in impulse con- duction (Mackie and Singla, 1987) and tunic contractility in response to wounding (Hirose and Ishii, 1995; Hirose el al., 1997a). Although bacteria are often found within the tunic of some ascidians (cf. Hirose and Saito. 1992; Hirose et al., 1996a), they are rarely found within the tunic of pelagic tunicates. Perhaps the tunic of the pelagic species contains some antibiotic substances, or maybe bacterial infections are less common in the habitat of the pelagic tunicates than in that of the sessile forms. If some tunic cells of pelagic tunicates are phagocytic, they may help to keep the tunic sterile. To date, the fine structure of the tunic cuticle has been described in 1 16 ascidian species covering all of the fami- lies and subfamilies of the class Ascidiacea except for the phlebobranch families Octacnemidae and Plurellidae (Hi- rose et al., 1990, 1992, 1997b). The presence of cuticular protrusions appears to have a phylogenetic significance in Ascidiacea, because there is a general stability of the char- acter-state distribution (presence or absence) within the families or subfamilies in the traditional classification. The authors concluded that the common ancestor of ascidians lacked cuticular protrusions and that the protrusions possi- bly emerged independently in several lineages (Hirose et al.. 1997b). With respect to ascidian phylogeny, pelagic tunicates can be considered the "out-group." Our results tend to support this concept, because many of the examined species do not have cuticular protrusions. The common ancestor of tunicates probably did not have cuticular pro- trusions and this character may have been independently acquired in some lineages of ascidians and thaliaceans. If cuticular protrusions emerged repetitively and became fixed in certain tunicate lineages, these protrusions should have some adaptive purpose that is as yet unknown. Using electron diffraction, the existence of cellulose was confirmed in the tunic of three groups of pelagic tunicates: pyrosomas, doliolids. and salps. The tunic of these pelagic tunicates consisted of cellulose I microfibrils with high crystallinity and large dimensions and belonging almost entirely to the cellulose Ij3 allomorph. These features are similar to those of the cellulose found in ascidians (Yamamoto et al., 1989: Van Daele et al.. 1992; Kimura and Itoh. 1996; Okamoto et al., 1996), suggesting that cellulose synthetic ability is an inherited characteristic com- mon to ascidians and thaliaceans. The tunic cellulose in thaliaceans may be synthesized by terminal complexes (TCs) in the plasma membrane of epidermal cells, as occurs in ascidians (Kimura and Itoh. 1996). The visualization of TCs in these thaliaceans is required for further discussion of the evolution of cellulose synthesis in tunicates. In this study, we did not examine appendicularians, the other member of the pelagic tunicates. Adult appendicular- ians do not have integumentary tissue outside the epidermis, but they secrete a mucous substance that forms their "house." which is a feeding apparatus. In embryonic and juvenile stages, the entire animal is covered by an acellular membrane, but it is not certain whether the membrane is equivalent to the tunic (Fenaux, 1998). The tunic is a synapomorphic character in tunicates, and it is generally believed to have evolved monophyletically. Our present results demonstrating that the tunic morphol- ogy and cellulosic components are fundamentally the same in ascidians, pyrosomas, doliolids, and salps are consis- tent with that view. The numbers of tunic cells in doliolids and salps are much smaller than those in ascidians and pyrosomas; this difference could reflect different functions of the tunic as an integumentary tissue. Thus, perhaps the tunic has become diversified in its properties and functions, thereby making it an attractive model for studying the evolution of adaptive tissue functions. Acknowledgments We thank the captain and the crews of R/V Tansei Mam of the Ocean Research Institute, University of Tokyo, for material collection at sea. We also wish to thank anonymous referees for their valuable comments and discussions. This study was supported by Grants-in-Aid for the Encourage- ment of Young Scientists to E. H. and S. K.. by Grants-in- Aid for Research for the Future Program (JSPS-RFTF 96L00605) to T. I. from the Ministry of Education, Culture and Science of Japan, and by a predoctoral fellowship to S. K. from the Japanese Society for the Promotion of Science. The study was partly carried out at Shimoda Ma- rine Research Center, University of Tsukuba (contribution #625). Literature Cited Aoki. M., K. Hashimoto, and H. Watanabe. 1989. The intrinsic origin of bioluminescence in the ascidian. Clavelina minima. Biol. Bull. 176: 57-62. Belton P. S., S. F. Tanner, N. Cartier, and H. Chanzy. 1989. High- resolution solid-state "C nuclear magnetic resonance spectroscopy of tunicin. an animal cellulose. Macromolecules 22: 1615-1617. Bone, Q. 1998. The Biology of the Pelagic Tunicate. Oxford University Press. Oxford. Bone, Q., and E. R. Trueman. 1983. Jet propulsion in salps (Tunicata: Thaliacea). J. Zool. Lnd. 201: 481-506. Bone. Q., and E. R. Trueman. 1984. Jet propulsion in Doliolum (Tu- nicata: Thaliacea). J. Exp. Mar. Biol. Ecol. 76: 105-118. Chiba, K., M. Hoshi, M. Isobe, and E. Hirose. 1998. Bioluminescence in the tunic of the colonial ascidian, Clavelina minima: Identification of luminous cells in vitro. J. Exp. Zool. 281: 546-553. De Leo, G., E. Patricolo. and G. D'Ancona Lunetta. 1977. Studies on the fibrous components of the test of dona intestinalis Linnaeus. I. Cellulose-like polysaccharide. Ada Zool. (Stockh.i 58: 135-141. De Leo, G., E. Patricolo, and G. Frittita. 1981. Fine structure of the tunic of dona intestinalis L. II. Tunic morphology, cell distribution and their functional importance. Acta Zool. (Stockh.) 62: 259-271. 120 E. HIROSE ET AL Fenaux, R. 1998. Life history of the Appendicularia. Pp. 151-159 in The Biology of the Pelagic Tunicate, Q. Bone, ed. Oxford University Press, Oxford. Hirose, E., and T. Ishii. 1995. Microfilament contraction promotes rounding of tunic slices: an integumentary defense system in the colonial ascidian Aplidium yamazii. Biol. Bull. 189: 29-35. Hirose, E., and Y. Saito. 1992. Threadlike bacteria in the tunic of a botryllid ascidian. (Note). Hiyoshi Rev. Natur. Sci. Keio Univ. 12: 108-110. Hirose, E., Y. Saito, K. Hashimoto, and H. Watanabe. 1990. Minute protrusions of the cuticle: fine surface structures of the tunic in ascid- ians. J. Morphol. 204: 67-73. Hirose, E., T. Nishikawa, Y. Saito, and H. Watanabe. 1992. Minute protrusions of ascidian tunic cuticle: some implications for ascidian phylogeny. Zool. Sci. (Tokyo) 9: 397-404. Hirose, E., T. Ishii, Y. Saito, and Y. Taneda. 1994. Phagocytic activity of tunic cells in the compound ascidian Aplidium yamazii (Polyclinidae, Aplousobranchia). Zoo/. Sci. (Tokyo) 11: 203-208. Hirose, E., M. Aoki, and K. Chiba. 1996a. Fine structures of tunic cells and distribution of bacteria in the tunic of the luminescent ascidian Clavelina miniata (Ascidiacea. Urochordata). Zoo/. Sci. (Tokyo) 13: 519-523. Hirose, E., T. Maruyama, L. Cheng, and R. A. Lewin. 1996b. Intra- cellular symbiosis of a photosynthetic prokaryote, Prochloron sp., in a colonial ascidian. Invertebr. Biol. 115: 343-348. Hirose, E., T. Ishii, and Y. Taneda. 1997a. Two modes of tunic cuticle formation in a colonial ascidian Aplidium \amazii. responding to wounding. Dev. Comp. Immunol. 21: 25-34. Hirose, E., G. Lambert, T. Kusakabe, and T. Nishikawa. 1997b. Tu nic cuticular protrusions in ascidians (Chordata, Tunicata): a perspec- tive of their character-state distribution. Zool. Sci. (Tokyo) 14: 683- 689. Hirose, E., Y. Saito, and H. Watanabe. 1997c. Subcuticular rejection: an advanced mode of the allogeneic rejection reaction in the compound ascidians. Botrylloides simodensis and B. fuscus. Biol. Bull. 192: 53- 61. Kiiinii ;i, S., and T. Itoh. 1996. New cellulose synthesizing complexes (terminal complexes) involved in animal cellulose biosynthesis in the tunicate Metandrocarpa uedai. Protoplasms 194: 151-163. Mackie, G. O., and C. L. Singla. 1987. Impulse propagation and con- traction in the tunic of a compound ascidian. Biol. Bull. 173: 188-204. Nishikawa, J., and Terazaki, M. 1994. Measurement of swimming speed and pulse rate of salps using a video equipment. Bull. Plankton Soc. Jpn. 41: 170-173. Okamoto, T., J. Sugiyama, and T. Itoh. 1996. Structural diversity of ascidian cellulose. Wood Res. 83: 27-29. Van Daele, Y., J.-F. Revol, F. Gaill, and G. Gofflnet. 1992. Charac- terization and supramolecular architecture of the cellulose-protein fibrils in the tunic of the sea peach (Halocynthia papillosa, Ascidiacea, Urochordata). Biol. Cell 76: 87-96. Welsh, U. 1984. Urochordata. Pp. 800-8 16 in Biology of the Integument I. Invertebrates. J. Bereiter-Hahn. A. G. Matoltsy, and K. S. Richards, eds. Springer-Verlag, Berlin. Yamamoto, H., F. Horii, and A. Hirai. 1989. Structural changes of native cellulose crystals induced by annealing in aqueous alkaline and acid solutions at high temperatures. Macromolecules 22: 4130-4132. THE BIOLOGICAL BULLETIN Editor Associate Editors Section Editor Online Editors Editorial Board Editorial Office APRIL 1999 MICHAEL J. GREENBERG Louis E. BURNETT CHARLES D. DERBY MICHAEL LABARBERA RUDOLF A. RAFF SHINYA INOUE, Imaging and Microscopy JAMES A. BLAKE. Keys to Marine Invertebrates of the Woods Hole Region WILLIAM D. COHEN. Marine Models Electronic Record and Compendia PETER B. ARMSTRONG ANDREW R. CAMERON ERNEST S. CHANG THOMAS H. DIETZ RICHARD B. EMLET DAVID EPEL GREGORY HINKLE MAKOTO KOBAYASHI DONAL T. MANAHAN MARGARET MCFALL-NGAI MARK W. MILLER TATSUO MOTOKAWA YOSHITAK.A NAGAHAMA SHERRY D. PAINTER J. HERBERT WAITE RICHARD K. ZIMMER-FAUST PAMELA CLAPP HINKLE VICTORIA R. GIBSON CAROL SCHACHINGER PATRICIA BURNS The Whitney Laboratory, University of Florida Grice Marine Biological Laboratory, College of Charleston Georgia State University University of Chicago Indiana University Marine Biological Laboratory ENSR Marine & Coastal Center. Woods Hole Hunter College, City University of New York University of California, Davis California Institute of Technology University of California, Davis Louisiana State University Oregon Institute of Marine Biology. Univ. of Oregon Hopkins Marine Station, Stanford University Cereon Genomics, Cambridge, Massachusetts Hiroshima University of Economics, Japan University of Southern California Kewalo Marine Laboratory. University of Hawaii Institute of Neurobiology, University of Puerto Rico Tokyo Institute of Technology, Japan National Institute for Basic Biology. Japan Marine Biomed. Inst., Univ. of Texas Medical Branch University of California. Santa Barbara University of California, Los Angeles Managing Editor Staff Editor Editorial Assistant Subscription & Advertising Secretary Published by MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASSACHUSETTS Cover Cyanca cci/ullata, the lion's mane jellyfish, is fea- tured on our cover this month as an artistic memo- randum to our readers that the Keys to Marine Invertebrates of the Woods Hole Region, edited by Ralph A. Smith (1964), are being revised online (see the Announcement on page 121). Moreover, the first segment of the revision has been com- pleted: it is the section on Scyphozoa, authored by Dale Calder, who is at the Centre for Biodiversity and Conservation Biology of the Royal Ontario Museum in Toronto. The drawing on our cov- er by Patrice Stephens-Bourgeault. also of the Royal Ontario Museum is one of the plates from the new section. It is adapted from several sources, including a lithograph by Auguste Sonrel, drawn for Louis Agassiz, who published it in 1862. C\anea is a frequent visitor to the Woods Hole region during winter and spring, and occasionally in the summer. There may be two varieties that differ in size, color, and some minor morphological features. The larger version, found primarily in the North At- lantic, can attain a size of up to 2.5 m across the saucer-shaped bell with tentacles up to 40 m long. The tentacles of the medusa on our cover are contracted; if they were relaxed they would extend about six feet below the bottom of the journal. When it is hunting, Cyanea sinks slowly, with its tentacles relaxed and spread out around it in a circular net. Its weapons are thousands of nemato- cysts. comprising euryteles and isorhizas, and its prey includes fishes and polychaete worms. Large arctic animals have a sting that can raise painful weals on humans, but are not known to be fatal notwithstanding the assertion of Sherlock Holmes in "The Adventure of the Lion's Mane." Still, the nematocysts continue to be of interest to students of marine venoms. Among Metazoa. the Cnidaria have the most prim- itive nervous systems, and these have been studied as an approach to the fundamental properties of neuronal mechanisms. Cyaneci capillata, in particular, provides useful preparations with which to investigate synaptic transmission and ion channels. Most recently, these primitive animals have been found to have very con- ventional voltage-gated calcium channels, and contin- ued investigation of these proteins will help us under- stand the relationship between structure and function in calcium channel subunits ( see Jeziorsky et ul. 1 998. J. Biol. Cliem. 273: 22792-99). CONTENTS VOLUME 196. No. 2: APRIL 1999 Announcement: A.'cv> to Marine Invertebrates of tltc Woods Hole Region . . 121 RESEARCH NOTE O Foighil, IM. ii ni. ml. Bruce A. Marshall, Thomas J. Hilbish, and Mario A. Pino Trans-Pacific range extension by rafting is inferred for the flat oyster Ostrea chi/ensis 122 NEUROBIOLOGY Schmidt, Manfred, and Steffan Harzsch Comparative analysis of neurogenesis in the central olfactory pathway of adult decapod crustaceans by in vivo BrdU-labeling 127 PHYSIOLOGY Frank, Tamara M. Comparative study of temporal resolution in the vi- sual systems of mesopelagic crustaceans 137 Coughlin, David J., and Lawrence C. Rome Muscle activity in steady swimming scup, Stenotomus rhrysopi, varies with fiber type and body position ... 145 Khan, Hamid R., Da\id A. Price, Karen E. Doble, Mi- chael J. Greenberg, and A.S.M. Saleuddin Osmoregulation and FMRFamide-related peptides in the salt marsh snail Melampus bidentatus (Say) (Mol- lusca: Pulmonata) 153 BEHAVIORAL PHYSIOLOGY McGaw. I.J., C.L. Reiber, andJ.A. Guadagnoli Behavioral physiology of four crab species in low- salinity 163 DEVELOPMENT AND REPRODUCTION Vavra, Jay, and Donal T. Manahan Protein metabolism in lecithotrophic larvae (Gas- tropoda: Haliotii rufescena) 177 Biggers, William J., and Hans Laufer Settlement and metamorphosis of Capitella laiTae in- duced by juvenile hormone-active compounds is me- diated bv protein kinase C and ion channels 187 Unuma, Tatsuya, Takeshi Yamamoto, and Toshio Akiyama Effect of steroids on gonadal growth and gametogen- esis in the juvenile Red Sea urchin Pseudocentrotus depressus 199 ECOLOGY AND EVOLUTION Okamura. Beth, and Julian C. Partridge Suspension feeding adaptations to extreme flow en- vironments in a marine bryozoan 205 Schmidt, Hartmut, and Wilfried Wesdieide Genetic relationships (RAPD-PCR) between geo- graphically separated populations of the "cosmopol- itan" interstitial polvchaete Hesionides gohari (Hesion- idae) and the evolutionary origin of the freshwater species Hesionides riegeronim 216 THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is published six times a year by the Marine Biological Laboratory. 7 MBL Street. Woods Hole, Massachusetts 02543. Subscriptions and similar matter should be addressed to Subscription Manager. THE BIOLOGICAL BULLETIN, Marine Biological Laboratory. 7 MBL Street, Woods Hole, Massachusetts 02543. Subscription per year (six issues, two volumes): $205 for libraries; $95 for individuals. Subscription per volume (three issues): $102.50 for libraries; $47.50 for individuals. Back and single issues (subject to availability): $40 for libraries; $20 for individuals. Communications relative to manuscripts should be sent to Michael J. Greenberg, Editor-in-Chief, or Pamela Clapp Hinkle, Managing Editor, at the Marine Biological Laboratory, 7 MBL Street, Woods Hole. Massachusetts 02543. Telephone: (508) 289-7428. FAX: 508-457-1924. E-mail: pclapp@mbl.edu. http://www.mbl.edu/BiologicalBulletin/ The home page for the electronic companion to THE BIOLOGICAL BULLETIN the Marine Models Electronic Record and other BIOLOGICAL BULLETIN publications is available on the World Wide Web at the address shown above. THE BIOLOGICAL BULLETIN is indexed in bibliographic services including Index Medicus and MEDLINE. Chemical Abstracts. Current Contents, Elsevier BIOBASE/Citrrent Awareness in Biological Sciences, and Ge o Abstracts. Printed on acid free paper, effective with Volume 180, Issue 1. 1991. POSTMASTER: Send address changes to THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, 7 MBL Street, Woods Hole, MA 02543. Copyright 1999. by the Marine Biological Laboratory Periodicals postage paid at Woods Hole, MA. and additional mailing offices. ISSN 0006-3 185 INSTRUCTIONS TO AUTHORS The Biological Bulletin accepts outstanding original research reports of general interest to biologists throughout the world. Papers are usually of intermediate length (10-40 manuscript pages). A limited number of solicited review papers may be accepted after formal review. A paper will usually appear within four months after its acceptance. Very short, especially topical papers (less than 9 manuscript pages including tables, figures, and bibliography) will be published in a separate section entitled "Research Notes." A Research Note in The Biological Bulletin follows the format of similar notes in Nature. It should open with a summary paragraph of 150 to 200 words comprising the introduction and the conclusions. The rest of the text should continue on without subheadings, and there should be no more than 30 references. References should be referred to in the text by number, and listed in the Literature Cited section in the order that they appear in the text. Unlike references in Nature, references in the Research Notes section should conform in punctuation and arrangement to the style of recent issues of The Biological Bulletin. Materials and Methods should be incorpo- rated into appropriate figure legends. See the article by Loh- mann et til. (October 1990. Vol. 179: 214-218) for sample style. A Research Note will usually appear within two months after its acceptance. The Editorial Board requests that regular manuscripts con- form to the requirements set below: those manuscripts that do not conform will be returned to authors for correction before review. 1. Manuscripts. Manuscripts, including figures, should be submitted in quadruplicate, with the originals clearly marked. (Xerox copies of photographs are not acceptable for review pur- poses.) The submission letter accompanying the manuscript should include a telephone number, a FAX number, and (if possible) an E-mail address for the corresponding author. The original manu- script must be typed in no smaller than 12 pitch or 10 point, using double spacing (including figure legends, footnotes, bibliography, etc.) on one side of 16- or 20-lb. bond paper, 8 by 1 1 inches. Please, no right justification. 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 r inch (2.5 cm). Manuscripts should conform to the Council of Biologv Editors Style Manual, 5th Edition (Council of Biology Editors, 1983) and to American spelling. Unusual abbreviations should be kept to a minimum and should be spelled out on first reference as well as defined in a footnote on the title page. Manuscripts should be divided into the following components: Title page. Abstract (of no more than 200 words). Introduction, Materials and Methods, Results, Discussion, Acknowledgments, Literature Cited, Tables, and Figure Legends. In addition, authors should supply a list of words and phrases under which the article should be indexed. 2. Title page. The title page consists of a condensed title or running head of no more than 35 letters and spaces, the manuscript title, authors' names and appropriate addresses, and footnotes listing present addresses, acknowledgments or contribution num- bers, and explanation of unusual abbreviations. 3. Figures. The dimensions of the printed page. 7 by 9 inches, should be kept in mind in preparing figures for publication. We recommend that figures be about 1 times the linear dimensions of the final printing desired, and that the ratio of the largest to the smallest letter or number and of the thickest to the thinnest line not exceed 1:1.5. Explanatory matter generally should be included in legends, although axes should always be identified on the illustra- tion 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 and cited, in order, in the text. The author's name and an arrow indicating orientation should appear on the reverse side of all figures. Color: The Biological Bulletin will publish color figures and plates, but must bill authors for the actual additional cost of printing in color. The process is expensive, so authors with more than one color image should consistent with editorial concerns, especially citation of figures in order combine them into a single plate to reduce the expense. On request, when supplied with a copy of a color illustration, the editorial staff will provide a pre-publi- cation estimate of the printing cost. 4. Tables, footnotes, figure legends, etc. Authors should follow the style in a recent issue of The Biological Bulletin in preparing table headings, figure legends, and the like. Because of the high cost of setting tabular material in type, authors are asked to limit such material as much as possible. Tables, with their headings and footnotes, should be typed on separate sheets, num- bered with consecutive Roman numerals, and placed after the Literature Cited. Figure legends should contain enough informa- tion 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 contri- bution numbers, and explanation of unusual abbreviations. All such footnotes should appear on the title page. Footnotes are not normally permitted in the body of the text. 5. Literature cited. In the text, literature should be cited by the Harvard system, with papers by more than two authors cited as Jones et al.. 1980. Personal communications and material in prep- aration 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 ABSTRACTS, with the minor differences set out below. The most generally useful list of bio- logical journal titles is that published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of Serials; the most recent issue). Foreign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PERIODICALS, may find a booklet published by the Biological Council of the U.K. (obtainable from the Institute of Biology. 41 Queen's Gate, London, S.W.7, England, U.K.) useful, since it sets out the WORLD LIST abbreviations for most biological journals with notes of the USASI abbreviations where these differ. CHEMICAL ABSTRACTS publishes quarterly supplements of addi- tional 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.Ph\siol.) E. Unusual words in journal titles should be spelled out in full, rather than employing new abbreviations invented by the author. For example, use Rit Visindafjelags Islendinga without abbreviation. F. All single word journal titles in full (e.g.. Veliger, Ecol- ogy. Brain). G. The order of abbreviated components should be the same as the word order of the complete title (i.e., Proc. and Trans. placed where they appear, not transposed as in some BIOLOGICAL ABSTRACTS listings). H. A few well-known international journals in their 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, page proofs, and charges. Authors of articles in black and white (no color figures) receive their first 50 reprints (without covers) free of charge. Color reprints and additional black-and-white reprints may be purchased: authors will receive order forms. Reprints normally will be delivered about 2 to 3 months after the issue date. Authors (or delegates for foreign authors) will receive page proofs of articles shortly before publi- cation. They will be charged the current cost of printers' time for corrections to these (other than corrections of printers' or editors' errors). Other than these charges for authors' alterations. The Biological Bulletin does not have page charges. Reference: Biol. Bull. 196: 121. (April 1999) ANNOUNCEMENT Keys to Marine Invertebrates of the Woods Hole Region The Biological Bulletin is undertaking a complete revision of Ralph Smith's classic handbook: Keys to Marine Invertebrates of the Woods Hole Region. This first revision will be carried out electronically, and will be directed and edited by James A. Blake (ENSR Marine and Coastal Center, Woods Hole). To date, the original edition of the Keys has been republished at for ready access. More- over, a revised section on Cnidaria: Scyphozoa, by Dale Calder of the Centre for Biodiversity and Conservation Biology at the Royal Ontario Museum in Toronto, has now been completed, reviewed, and posted on the web- site, replacing the original section. Keys to Marine Invertebrates of the Woods Hole Region grew out of the Systematics-Ecology Program, which oper- ated at the Marine Biological Laboratory (MBL) during the 1960s. Melbourne R. Carriker, who directed this program, noted that illustrated, referenced, and indexed checklists and keys to the common organisms of the region were urgently needed by non-systematists in the scientific communities in and around Woods Hole. The region was defined to include collecting sites on both sides of Cape Cod, the islands of Martha's Vineyard, Nantucket, and the Elizabeth Islands, and the deeper waters of Cape Cod Bay, Nantucket Sound, Vineyard Sound, and Buzzards Bay essentially all of southeastern New England. The first response, in the 1960s, to the need for an invertebrate guide was a collection of looseleaf keys pro- duced by the staff of the Invertebrate Zoology course at the MBL. These documents were added to or rewritten by a group of contributing experts, then compiled and edited by the late Ralph I. Smith, and published by the MBL in 1964. But the handbook that Smith produced was not intended to be an exhaustive survey of the rich and varied fauna of the Woods Hole Region: rather it was meant to provide the keys, illustrations, and glossaries that would facilitate a routine identification of those organisms commonly used in research and in classroom activities. Annotated species lists and bibliographies were also included as a guide to more comprehensive resources or to specialists who could help with organisms that resisted identification. The Keys quickly became an important resource, and they are still used as a standard faunal reference. The 1964 edition of the Keys was intended by Smith to be only the first of several editions, as the coverage was un- even, some especially microscopic or planktonic taxa were excluded, and some groups were represented only at higher taxonomic levels. Inaccuracies and lacunae were therefore expected, and revisions were invited. But no revisions were ever made, and from the standpoint of taxonomy, this man- ual is now hopelessly out of date. Though the Kevs are dated, their utility remains undimin- ished, and these considerations led the editors of The Bio- logical Bulletin to essay the first revision in 35 years. Because the Ke\s (like any catalog) are composed of many independent sections, the unique advantages of online edit- ing and publication are especially applicable to the project. For example, as a section, or even a subsection, is revised, the new key can be added at once to the online version of the manual. The updating can therefore proceed continu- ously and smoothly, even as the manual is being used. Moreover, because the Keys can be downloaded and printed in whole or in part, as necessary, the requirement for stan- dard, printed, bound copies of a constantly changing prod- uct may be precluded. Because few of the original contributors to the Keys are available to update their chapters, we have begun to search for scientists who are qualified and interested in revisiting existing chapters, or contributing new chapters for taxa not included in the 1964 edition. Unfortunately, we have not yet identified qualified systematists for some groups, reflecting the current lack of taxonomic expertise. Nevertheless, we hope to have the Keys substantially updated by the end of the year 2000, a fitting tribute, we think, to Ralph Smith and his contributing authors. Anyone who would like to contribute to this project at any taxonomic level should contact James Blake directly. He can be reached at ENSR Marine and Coastal Center, 89 Water Street, Woods Hole, MA 02543; telephone 508-457- 7900: e-mail jablake@ix.netcom.com. Michael J. Greenberg Editor-in-Chief 121 Reference: Biol. Bull. 196: 122-126. (April Trans-Pacific Range Extension by Rafting Is Inferred for the Flat Oyster Ostrea chilemis DIARMAID 6 FOIGHIL 1 *, BRUCE A. MARSHALL 2 , THOMAS J. HILBISH 3 , AND MARIO A. PINO 4 1 Museum of Zooloifv and Department of Biology, University of Michigan, Ann Arbor, Michigan 48109-1079; 2 Museum of New Zealand, P.O. Box 467, Wellington, New Zealand; 3 Belle W. Baruch Institute, University of South Carolina, Columbia, South Carolina 29208; and 4 Instituto de Geociencias, Universidad Austral de Chile, Casilla 567, Valclivia, Chile Stretches of deep ocean are potent burners to the disper- sion of nearshore. benthic marine taxa. Such obstacles can be overcome, however, by species that have either a pro- tracted pelagic larval development or a benthic life-history stage that can be transported by rafting (1, 2). The oyster Ostrea chilensis lacks an extended pelagic larval phase and has discrete populations in New Zealand and Chile that are separated by one of the largest extents (>7000 km) of open ocean on the planet. We tested competing dispersal hypoth- eses for this species by using ontogenetically informative, dated fossil and sub-fossil shell material, as well as molec- ular phylogenetic analyses. Our data show that dispersal by rafting is by far the most likely explanation for trans-Pacific range extension by this oyster, and we reject competing hypotheses of vicariance, anthropogenic introduction, and dispersal by ancestral lineages with extended larval devel- opment. The presence of O. chilensis in Chile is important because it clearly demonstrates that transoceanic range ex- tension by rafting is potentially available to a significant traction of nearshore marine biotas. A prolonged pelagic larval phase facilitates the crossing of open-ocean barriers by nearshore benthic taxa (3. 4). and an evolutionary important linkage has been proposed be- tween this mode of development and a suite of species-level traits, including enhanced geographic range (5, 6). How- ever, there are als<> many nearshore examples of individual species (7), clonal taxa (8 V). and significant fractions of entire faunas (10) that lack significant pelagic larval devel- opment and have biogeographic ranges that span oceanic Received 28 August 1998; accepted 14 January 1999. * To whom correspondence should be addressed. E-mail: diarm,nd<" umich.edu barriers to dispersal. Spontaneous rafting events, in which sessile life-history stages are passively transported on drift- ing objects, are common in the marine environment and have probably acted as an alternative long-distance dis- persal mechanism for major subsections of shallow-water hard-bottom faunas over ecologically significant timescales (7-1 1 ). In the case of individual species, however, claims of long-distance range extension by rafting are usually con- founded by the alternative possibilities of undocumented anthropogenic introductions that predate biotic surveys (12), or of transoceanic colonization by ancestral lineages having extended larval development, with the subsequent evolutionary loss of such larvae (13). In this study, we present a specific case of transoceanic range extension by rafting, possibly the first from which the alternative dis- persal hypotheses can be excluded with confidence. Ostrea chilensis is unique among oysters in lacking an obligate feeding pelagic (planktotrophic) larval develop- ment; larvae typically metamorphose less than 2 hours after their release from parental brood chambers (14). This greatly abbreviated pelagic phase is associated with a highly distinctive larval shell morphology (15) and is a derived condition among flat oysters ( 16). O. chilensis occurs in two regional populations separated by 7400 km of uninterrupted south Pacific Ocean (Fig. 1). It is found throughout the nearshore waters of New Zealand (including the Chatham Islands) to depths of 550 m (17) and also in the shallow subtidal zone of southern Chile (18). New Zealand and Chile share some relict Gondwanaland taxa (19). but this oyster does not qualify as such because it is unknown from pre-Holocene Chilean fossil strata (20), its appearance in the New Zealand Pliocene fossil record (21) greatly post-dates 122 OSTREA CHILENSIS PHYLOGEOGRAPHY 123 Figure 1. Polar View of the South Pacific showing Osticu chilciiM-, sampling locations (large arrows) for molecular analysis in New Zealand: Moturekareka Island, Hauraki Gulf (North Island), Foveaux Strait (South Island, 4643' S. 1683' E); and in Chile: Quempillen. Isla Chiloe. Small arrows show predominant surface circulation patterns. the separation of New Zealand from Gondwanaland (22), and its New Zealand and Chilean populations are almost indistinguishable allozymically [genetic similarity = 0.991; a value of 1.0 indicates genetic identity (18)]. The fossil record suggests that the New Zealand popula- tion is ancestral, and it is speculated that the Chilean pop- ulation was founded by rafted, post-metamorphic New Zea- land oysters transported via the Antarctic circumpolar and Humboldt currents (18). Alternatively, Chilean populations might have been established by the transoceanic dispersal of ancestral New Zealand larvae with extended planktotrophic pelagic development, followed by independent losses of this developmental mode in both source and founder popula- tions. The simplest hypothesis is that the Chilean population resulted from an undocumented human introduction event, a frequent occurrence for commercially important oysters (23. 24). Samples of sub-fossil O. chilensis were collected for radiocarbon dating from the Raqi-Tubul estuary, a southern Chilean location where this species has been regarded as the sole native oyster (25). Radiocarbon age estimates (95% confidence intervals) of 953-1238 and 2998-3383 years before present (y BP) were obtained for specimens taken respectively from salt marsh sediments and from an oyster midden site (26). Some of the midden specimens had at- tached juvenile oysters (spat) that retained well-preserved larval shells (prodissoconchs) displaying the diagnostic morphological characteristics of this species of oyster (Fig. 2a). Unequivocal evidence for human settlement in New Zealand dates from about 850 y BP (27). The presence of significant populations of O. chilensis in South America 2000 years prior to this date precludes anthropogenic intro- duction as a credible explanation for transoceanic range extension in this species. Although O. chilensis transversed the South Pacific with- out human intervention, the Chilean midden data do not reveal whether this occurred via planktotrophic larval dis- Figure 2. Scanning electron micrographs of prodissoconchs preserved on juvenile oyster shells from (2a) Chilean oyster midden sample radio- carbon-dated to 2998-3383 years y BP (95% probability), prodissoconch length = 469 jam (UMMZ 255348); (2b) New Zealand North Island upper Nukumaruan Stage ( 1.6-2.0 mya) strata, prodissoconch length = 470 /urn (NMNZ M. 42778). Both specimens display the non-umbonate. D-shaped, >425 /nm in length, flattened prodissoconch valves diagnostic for Ostrea chilensis (15). Scale bars = 150 jum. Sub-fossil Chilean samples of O. chilensis were obtained by M. Pino at the Raqi-Tubul estuary (3713'30" S, 7326' W) in December 1996. Shell samples were recovered from a midden, predominantly (70%-80%) composed of oysters (26). situated 5 m above present sea level on the west slope of "La Isla" hill. Additional sub-fossil oyster shells were recovered from adjacent salt marsh sediments (1.5 m depth). Sub-fossil Chilean samples were subjected to radiocarbon dating analyses by Beta Analytic Inc. 4985 S.W. 74 Court, Miami, Florida 33155. New Zealand fossil flat oysters with characteristic O, chilensis prodissoconchs were sampled by B. A. Marshall and M. B. Willoughby from late Pliocene horizons on the north bank of Mangahao River, 1.5 km south of Mangahao, Wairarapa. North Island (map reference NZMS 260/ 462805). Prodissoconch samples were gold-coated and examined with a Hitachi S-3200N scanning electron microscope. 124 D. O FOIGHIL ET AL. persal. Previous studies of fossil O. chilensis in New Zea- land were based on adult shell morphology (21) and do not provide insights into when planktotrophic larval develop- ment was lost in this oyster lineage. We examined museum fossil holdings of O. chilensis and found individuals from North Island strata [late Pliocene, upper Nukumaruan Stage (1.6-2.0 mya)], with a prodissoconch structure identical to that of modern specimens (Fig. 2b). Earlier Pliocene spec- imens did not yield interpretable prodissoconchs, but the North Island populations of this species had fully evolved the present day nonplanktotrophic larval development prior to the Pleistocene. This is pertinent because molecular phy- logenetic analysis of O. chilensis populations (Fig. 3) indi- cates that ( 1 ) a pronounced mitochondrial dichotomy exists among North and South Island samples, differing by 19 nucleotide substitutions (3.1% sequence divergence for a 609 nucleotide fragment of cytochrome oxidase I); (2) North Island and Chilean samples form sister lineages, differing by four nucleotide substitutions (0.6% sequence divergence); (3) ancestral North Island lineages were the probable source populations for trans-Pacific colonists. Joint consideration of the fossil and molecular data al- lows a comparative evaluation of the two remaining dis- persal hypotheses. Rafting is compatible with a single loss of planktotrophic larval development in a common ancestral lineage of O. chilensis and allows for a Pleistocene/Holo- cene trans-Pacific colonization event. Conversely, dispersal by planktotrophic larvae requires that trans-Pacific coloni- zation be pre-Pleistocene. and that loss of such larvae (a condition unique to O. chilensis among the Ostreacea) oc- curred on three independent occasions after the respective branching of South Island, North Island, and (pre-Pleisto- cene and post-trans-Pacific colonization) Chilean lineages. This mechanism of dispersal also requires the assumption that putative ancestral larval cohorts were sufficiently long- lived and numerous to overcome mortality and (enormous) diffusion effects over a >7000-km linear dispersive path- way, and to recruit in viable breeding densities in Chile. Extant flat oyster species with planktotrophic larvae have pelagic periods ranging from 6 to 33 days (29). The effec- tive dispersal distances [10 s-100 s of km (30)] calculated from these periods are insufficient for trans-Pacific coloni- 43 O. angasi O. aupouria North Island Chiloe South Island o- Figure 3. Most parsimonious tree ( 1 74 steps. CI = 0.948; RI = 0.775 ), obtained by an exhaustive search for optimal tret based on an mt COI gene fragment [609 nucleotides. homologous to positions 52-669 of the chiton nit gen '- i| sequenced for New Zealand South Island (n = 12). New Zealand North Island < = 13). and Chilean (Isla Chiloe, n = 16) samples of Ostrea chilensis. No intra-populational variation was observed. Regional flat oyster taxa, O. aupouria (New Zealand) and O. angasi (Australia), were employed as outgroups. The number of steps are indicated above each branch and brackets show, respectively, the decay index and bootstrap values supporting each node. Molecular techniques and phylogenetic analyses were as previously described (24), and GenBank accession numbers for the five haplotypes are API 12285-9. OSTREA CHILENSIS PHYLOGEOGRAPHY 125 zation. even if calculated for the most favorable oceanic currents: transoceanic, non-meandering flow rates of 125 cm s~ : [<3600 km (31)]. On grounds of phylogenetic parsimony (Fig. 3). apparent absence of Pliocene/Pleisto- cene fossil specimens in Chile (20), and absence of excep- tionally long-lived/teleplanic (2) larvae in flat oysters (29), the ancestral planktotrophic larval dispersal hypothesis is rejected as a plausible explanation for trans-Pacific range extension in O. chilensis. We conclude that dispersal by rafting is by far the most likely explanation for the trans-Pacific range extension of O. chilensis. Pumice, an effective long-distance rafting vector for hard-bottom, shallow-water, suspension-feeding epi- benthos (11), may have served as the transport platform, because very significant quantities of this buoyant material have been released by repeated eruptions of the North Island Taupo Volcanic Zone since the Pliocene (32). Further ge- netic characterization of the North Island and Chilean pop- ulations may uncover haplotypes common to both popula- tions, a finding that would be necessary to entertain the possibility of continued trans-Pacific gene flow. In principle, there may be no maximum dispersal distance for rafting organisms (7. 11), and the presence of O. chil- ensis in Chile highlights the existence of an alternative dispersal mechanism, of truly transoceanic scope, which is potentially available to many members of nearshore faunas (7-11). A broad cross-section of the North Island hard- bottom epibenthos was probably also rafted across the Pa- cific, including indirect developing species with an obligate planktotrophic larval stage. Rafted innocula of such taxa are less likely to have become established in Chile because an obligate extended pelagic larval phase would dilute popu- lation densities below the concentrations needed for repro- ductive success in the crucial initial generations of the new colonies (7-9. 11). However, careful analysis of Chilean taxa with reduced or absent pelagic larval development including sponge, ascidian, cheilostome bryozoan and hy- droid faunas (8) may reveal additional cryptic rafted im- migrants from New Zealand. Such taxa would have New Zealand sister lineages and might be identified by morpho- logical or molecular characterization of epifaunal residues from Chilean midden shells that predate human settlement of New Zealand. Our molecular results are in striking contrast to those from another oyster, Crassostrea virginica; this species undergoes a planktotrophic larval development, but it ex- hibits greater intraspecific genetic divergence values over a few hundred kilometers of contiguous southeastern Florid- ian coastline (33) than do North Island and Chilean popu- lations of O. chilensis. The contrasting phylogeographic patterns of these developmentally heterogeneous oyster taxa indicate that assumptions of marine invertebrate genetic structuring based solely on extrapolations from pelagic lar- val periods may lead to spectacular error. Acknowledgments We are indebted to our colleagues who graciously pro- vided ethanol-fixed oyster tissue samples: A. Jeffs, M. Barker, A. Frazer, O. Chaparro, J. Toro, R. Ward, and J. Nell. Manuscript drafts were improved by the comments of G. Paulay. T. Baumiller. W. Brown, S. Palumbi, and two anonymous reviewers. This study was supported by NSF award 9617689 to D. 6 Foighil. Literature Cited 1. Vermeij, G. J. 1987. The dispersal barrier in the tropical Pacific: implications for molluscan speciation and extinction. Evolution 41: 1046-1058. 2. Scheltema, R. S. 1988. Initial evidence for the transport of teleplanic larvae of benthic invertebrates across the east Pacific barrier. Biol. Bull. 174: 145-156. 3. Kohn, A. J., and F. E. Perron. 1994. Life history and biogeography patterns in Conns. Oxford Biogeography Series. 9. Clarendon Press. Oxford. 1 10 pp. 4 Lessios, H. A., B. D. Kessing, and D. R. Robertson. 1998. Massive gene flow across the world's most potent marine biogeographic bar- rier. Proc. R. Soc. Land. B 265: 583-588. 5. Jablonski, D. 1986. Larval ecology and macroevolution in marine invertebrates. Bull. Mar. Sci. 39: 565-587. 6. Jablonski, D. 1987. Heritability at the species level: analysis of geographic ranges of Cretaceous mollusks. Science 238: 360-363. 7. Johannesson, K. 1988. The paradox of Rockall: why is a brooding gastropod (Littorina saxatilis) more widespread than one having a planktonic larval stage (L. littorea)! Mar. Biol. 99: 507-513. 8. Jackson, J. B. C. 1986. Modes of dispersal of clonal benthic inver- tebrates: consequences for species' distributions and genetic structure of local populations. Bull. Mar. Sci. 39: 588-606. 9. 6 Foighil, D. 1989. Planktotrophic larval development is associated with a restricted geographic range in Lasaea, a genus of brooding, hermaphroditic bivalves. Mar. Biol. 103: 349-358. 10. Davenport, J., and H. Macalister. 1996. Environmental conditions and physiological tolerances of intertidal fauna in relation to shore zonation at Husvik. South Georgia. J. Mar. Biol. Assoc. UK 76: 985-1002. 1 1 . Jokiel, P. L. 1990. Transport of reef corals into the Great-Barrier Reef. Nature 347: 665-667. 12. Carlton, J. T., and J. Hodder. 1995. Biogeography and dispersal of coastal marine organisms: experimental studies on a replica of a 16th-century sailing vessel. Mar. Biol. 121: 721-730. 13 Jablonski, D., and R. A. Lutz. 1983. Larval ecology of marine benthic invertebrates: paleobiological implications. Biol. Rev. Camb. Phi/os. Soc. 58: 21-89. 14. Millar, R. H., and P. J. Hollis. 1963. Abbreviated pelagic life of Chilean and New Zealand oysters. Nature 197: 512-513. 15. Chanley, P., and P. Dinamani. 1980. Comparative descriptions of some oyster larvae from New Zealand and Chile, and a description of a new genus of oyster, Tiostrea. N. Z. J. Mar. Freshwater Res. 14: 103-120. 16. Jozefowitz, C. J., and D. 6 Foighil. 1998. Molecular phylogenetic analysis of Southern Hemisphere flat oysters based on partial mito- chondrial 16S rDNA gene sequences. Mol. Phylogenet. Evol 10:426- 435. 1 7. Jeffs, A. G., and R. G. Creese. 1996. Overview and bibliography of research on the Chilean oyster Tiostrea chilensis (Phillippi, 1845) from New Zealand waters. J. Shellfish Res. 15: 305-3 1 1 . 18 Buroker, N. E., P. Chanley, and P. Dinamani. 1983. Systematic 126 D. O FOIGHIL ET AL. status of two oyster populations of the genus Tiostrea from New Zealand and Chile. Mar. Biol. 77: 191-200. 19. Cooper, R. A., and P. R. Millener. 1993. The New Zealand biota: historical background and new research. TREE 8: 429 133. 20. Herm. D. 1969. Marines Plioziin und Pleistozan in Nord und Mittel- Chile unter besonderer Berticksichtigung der Entwicklung der Mol- lusken-faunen. Zitte/iaiui 2: 159 pp. 21. Beu, A. G., and P. A. Maxwell. 1990. Cenozoic Mollusca of New Zealand. N. Z. Geol. Sun: Paleontol. Bull. 58: 336-341. 22. Weissel, J. K., and D. E. Hayes. 1977. Evolution of the Tasman Sea reappraised. Earth Planet. Sci. Lett. 36: 77-84. 23. Carlton, J. T., and R. Mann. 1996. Transfers and world-wide introductions. Pp. 691-706 in The Eastern Oyster Crassostrea vir- ginica. V. S. Kennedy, R. I. E. Newell, and A. F. Eble, eds. Maryland Sea Grant. College Park. Maryland. 24. 6 Foighil, D., P. M. Gaffney, A. E. Wilbur, and T. J. Hilhish. 1998. Mitochondria! cytochrome oxidase I gene sequences support an Asian origin for the Portuguese oyster, Crassostrea angulatti. Mar. Biol. 131: 497-503. 25. Solis, I. F. 1967. Observaciones biologicas en Ostras (Ostrea chil- ensis Philippi) de Pullinque. Biol. Pesq. (Chile) 2: 51-82. 26. Bustos, V., N. Vergara, and Z. Seguel. 1997. Los conchales an- tropicos de ostras en la micro-area de Raqui-Tubul del golfo de Arauco, Vii Region. Abstract, XIV Congreso Nacional de Arqueologia Chilena, Copiapo, p. 41. 27 McGlone, M. S., A. J. Anderson, and R. N. Holdaway. 1994. An ecological approach to the Polynesian settlement of New Zealand. Pp. 136-163 in The Origins of the First New Zealanders, D. G. Sutton, ed. Auckland University Press. Auckland, NZ. 28. Boore, J. L., and W. M. Brown. 1994. Complete DNA sequence of the mitochondria! genome of the black chiton, Kalharina innicata. Genetics 138: 423-443. 29. Buroker, N. E. 1985. Evolutionary patterns in the family Ostreidae: larviparous vs. oviparous. J. E.\p. Mar. Biol. Ecol. 90: 233-247. 30. Okubo, A. 1971. Oceanic diffusion diagrams. Deep-Sea Res. 18: 789-802. 31. Scheltema, R. S. 1992. Passive dispersal of planklonic larvae and the biogeography of tropical sublittoral invertebrate species. Pp. 195- 202 in Marine Eutrophication anil Population Dynamics. Proceedings of the 25th EMBS, G. Colombo et a/., eds. Olsen and Olsen, Fredens- borg. Denmark. 32. Wilson, C. J. N., B. F. Houghton, and E. F. Lloyd. 1986. Volcanic history and evolution of the Maroa-Taupo area, central North Island. Pp. 194-223 in Late Cenozoic Volcanism in Ne\v Zealand. I. E. M. Smith, ed. The Royal Society of New Zealand, Bulletin 23. 33. Reeb, C. A., and J. C. Avise. 1990. A genetic discontinuity in a continuously distributed species: mitochondrial DNA in the American oyster, Crassostrea virginica. Genetics 124: 3974. Reference: Bio/. Bull. 196: 127-136. (April 1999) Comparative Analysis of Neurogenesis in the Central Olfactory Pathway of Adult Decapod Crustaceans by In Vivo BrdU Labeling MANFRED SCHMIDT 1 * AND STEFFEN HARZSCH 2 Zoologisches Institnt und Zoologisches Museum, Neurophysiologie, Universitdt Hamburg, and Institut fur Biologie, Technische Universitdt Berlin, Germany: and 2 Fakultat fur Biologic, Neuroanatomie, Universitdt Bielefeld, Germany Abstract. To examine the distribution of neurogenesis in the central olfactory pathway of adult decapod crustaceans, we labeled, in vivo, six species of decapod crustaceans representing most infraorders (shrimps, spiny lobsters, clawed lobsters, crayfish, hermit crabs, true crabs) with the proliferation marker 5-bromo-2'-deoxyuridine (BrdU). In all tested species a group of small, neuron-like nuclei in the lateral soma clusters of the brain was labeled; the lateral soma clusters are composed of the cell bodies of ascending olfactory projection neurons. In only a few instances did groups of BrdU-positive nuclei also occur in the other soma clusters of the central olfactory pathway. In the spiny lobster (Panulirus argus), a group of small neuron-like nuclei was labeled in the medial soma clusters containing the cell bodies of local interneurons of the olfactory deutocerebrum. In the hermit crab (Pagurus bernhardus), and the true crab (Cancer pagurus), a group of small neuron-like nuclei was labeled in soma clusters located in the eyestalks. These soma clusters probably contain the cell bodies of local interneurons of the hemiellipsoid bodies, to which the ol- factory projection neurons ascend. These results indicate that neurogenesis occurs among olfactory projection neu- rons in the adult brain. Among the other neuronal types of the central olfactory pathway, however, neurogenesis is restricted to specific taxa. The persistence of neurogenesis among the different neuronal types of the central olfactory pathway throughout adult life suggests an enormous struc- tural plasticity of brain circuitry that may enable the long- Received 4 November 1997: accepted 6 January 1999. * To whom correspondence should be addressed. Universitat Hamburg. Zoologisches Institut und Zoologisches Museum. Neurophysiologie. Martin-Luther-King-Platz 3. 20146 Hamburg. Germany. E-mail: mschmidt@zoologie.uni-hamburg.de lived decapod crustaceans to adapt to changing olfactory environments. Introduction It was commonly believed that neurogenesis is rare, if not totally absent, in the central nervous system of most adult animals (e.g., Purves, 1988), although de novo generation of neurons occurs in the brain of adult birds (e.g., Paton and Nottebohm, 1984), fish (e.g., Raymond and Easter, 1983), and mammals (Kaplan and Hinds, 1977; Bayer, 1982; Kaplan et ai, 1985; Crespo el ai, 1986). During the last decade, the development of new techniques to label and culture proliferating cells fostered a surge of research that conclusively demonstrated neurogenesis in the adult mam- malian brain (Corotto et ai, 1993; Luskin, 1993; Okano et ai, 1993; Morshead et ai, 1994; Kuhn et ai. 1996). Adult neurogenesis also occurs in the arthropod brain (Cayre et ai, 1994; Cayre et ai, 1996: Schmidt, 1997; Sandeman et ai, 1998). Thus, neurogenesis in the adult brain is now known to be a continuous process that contributes signifi- cantly to the neuronal population of specific brain areas. In the mammalian brain, two regions in particular acquire new neurons throughout adult life: the dentate gyrus of the hippocampus (Bayer. 1982; Crespo et ai, 1986; Okano et ai, 1993: Kuhn et ai, 1996) and the olfactory bulb (Kaplan etai, 1 985; Corotto etal, 1993; Luskin. 1993; Okano et ai, 1993). In the dentate gyrus, new granule cells develop from progenitor cells in the subgranular region (Crespo et ai, 1986; Okano et ai, 1993; Kuhn et ai, 1996). Although the rate of neurogenesis decreases significantly with age (Kuhn et ill.. 1996), the number of granule cells increases contin- uously (Crespo et ai, 1986), indicating that cell death is rare. The olfactory bulb acquires local interneurons of dif- 127 128 M. SCHMIDT AND S. HARZSCH ferent types (granule cells, periglomerular cells) almost linearly with age, although many of the new neurons do not survive, indicating a substantial turnover (Kaplan el a!., 1985; Corotto el al. 1993; Luskin, 1993; Okano el ai, 1993). The new neurons in the olfactory bulb originate from constitutively proliferating cells in the subependymal zone of the lateral ventricles (Moreshead et ul.. 1994), some of which migrate rostrally along a highly restricted pathway (the rostral migratory stream) and differentiate into neurons upon reaching the olfactory bulb (Luskin, 1993). Evidence for neurogenesis in the adult arthropod brain was obtained in diverse insect species and in two species of decapod crustaceans, Carcinus maenas (shore crab) and Cherax destructor (Australian crayfish) (Bieber and Fuld- ner, 1979; Technau, 1984; Cayre et al., 1994. 1996; Schmidt. 1997; Sandeman et al,, 1998). As in mammals, in arthropods adult neurogenesis is restricted to or prominent in specific brain areas. In adult insects, only the mushroom bodies, which represent the second stage of the central olfactory pathway and mediate olfactory learning and mem- ory (see reviews by Davis, 1993; Menzel et ul., 1994), acquire new local interneurons. the Kenyon cells (Bieber and Fuldner, 1979; Technau. 1984; Cayre et al., 1994, 1996). Surprisingly, only in some insect species do adults show neurogenesis of Kenyon cells (fly: Drosophila mela- nogaster; crickets: Acheta domesticus, Gryllus bimaculatus, Gryllomorpha dalmatina; beetles: Aleochara cnrtiila, Tenebrio molitor, Zophobas spec., Harmonia axyridis), whereas other species (bee: Apis mellifera; locust: Locusta migratoria: cockroach: Periplaneta americana) do not (Bieber and Fuldner. 1979; Technau, 1984; Cayre et ai, 1994, 1996; Fahrbach et ai, 1995). In adult insects, neuro- genesis appears to be driven by typically large neuroblasts or ganglion mother cells that continue to be mitotically active into the imaginal stage (Technau, 1984; Cayre et ai, 1994, 1996). Evidence for adult neurogenesis in decapod crustaceans occurs in the central olfactory pathway (Schmidt, 1997; Sandeman et al., 1998). In decapods, as in insects, this pathway is composed of two elements: bilateral glomerular neuropils of the deutocerebrum, where olfactory afferents transmit information onto local interneurons and projection neurons (antennal lobes in insects; olfactory lobes in deca- pods); and bilateral second-order neuropils of the protoce- rebrum, to which the projection neurons ascend (mushroom bodies in insects; ii, H.-llipsoid bodies in decapods) (e.g., Boeckh et al.. IMX^I tem et al., 1988; Mellon et al., 1992). In adult shore crabs (Carcinus maenas) and Austra- lian crayfish (Chera.\ desti i >). in vivo labeling with the thymidine analog 5-broiiio-. -oxyuridine (BrdU) demon- strated neuronal proliferation i : he lateral soma clusters (LC) containing the somata of ill. -tory projection neurons (Schmidt, 1997; Sandeman et a , ,998). In Carcinus mae- nas, neuronal proliferation also occurred in clusters of cell bodies associated with the hemiellipsoid bodies (HBC) but not in the medial soma clusters (MC) containing the somata of local interneurons restricted to the olfactory deutocere- brum (Schmidt, 1997). In late juveniles of Cherax destruc- tor, however, proliferation was detected in the MC (Sande- man et al., 1998). In both analyzed species, proliferation in the LC is associated with a continuous linear increase in the number of mature olfactory projection neurons throughout the life span of postlarval animals (Schmidt, 1997; Sande- man et al.. 1998). In Carcinus maenas. the proliferating cells in the LC and the HBC are tightly packed in a very restricted area close to the respective neuropil, thus delin- eating these areas as proliferative zones. In location and morphology, the proliferating cells deviate considerably from typical neuroblasts or ganglion mother cells that are still mitotically active in the LCs of early crab juveniles (Harzsch and Dawirs, 1996). This indicates that neurogen- esis in the central olfactory pathway of adult shore crabs occurs by a different mechanism than during embryonic and larval development. In this study we ask whether the neurogenesis that has been observed in the central olfactory pathway of adult shore crabs (Carcimts maenas) and Australian crayfish (Cherax destructor) can be generalized to other decapod crustaceans. In vivo labeling with BrdLI in six species rep- resenting most of the important decapod taxa (Caridea = shrimps, Palinura = spiny lobsters, Astacidea = clawed lobsters and crayfish. Anomura = hermit crabs, Brachy- ura = true crabs) revealed that proliferation among olfac- tory projection neurons is a common phenomenon in adult decapods. Significant differences between species exist in the occurrence of neurogenesis among local interneurons of the olfactory deutocerebrum and among neurons that prob- ably represent local interneurons of the hemiellipsoid bod- ies. We conclude that the central olfactory pathway of decapod crustaceans retains a lifelong structural plasticity, whose regulation and physiological significance remains to be elucidated. Materials and Methods Experiments were performed on adult, sexually mature specimens (and in case of Paniilirus argus, also on late juveniles) of six species of decapod crustaceans (Table I). All animals were kept in tanks with running seawater or fresh water without feeding for at least 2 days prior to the experiments. For in vivo labeling. 5-bromo-2'-deoxyuridine (BrdU) was injected into the hemolymph of experimental speci- mens (5 milligrams of BrdU per 100 gram of body weight in a 0.5% BrdU solution in the appropriate saline). After survival times of 6 to 24 h, the brains of the animals were fixed with Bodian # 2 fixative (90 ml 80% ethanol, 5 ml formol, 5 ml glacial acetic acid), for 1 t h at room tem- perature. After rinsing in 0.1 M Sorensen phosphate buffer (SPB), the brains were embedded in gelatin and cut on a NEUROGENESIS IN OLFACTORY PATHWAY OF ADULT DECAPODS Table I Suininury of decapod cru.\tM-ftin ,*/><( 10 nn-cMii>tilfil 129 Scientific name Systematic status Carapace length (cm) Weight (g) Source of material Trivial name Infraorder* Family /it Sicyonia Rock shrimp Penaeidea Sicyonidae 3 15-17 Dr. J. Nunez, Whitney Laboratory. brevirostris [Caridea] University of Florida Piiniilini\ Caribbean spiny Palinura Palinuridae 7 2.8-6.0 15-208 Dr. B. W. Ache, Whitney Laboratory, argus lobster [Achelata] University of Florida Hoiiiiirus American lobster Astacidea Homaridae 5 11.0-13.5 500-750 Commercial supplier amerifuiiu.'i [Homurida] Cherax Australian crayfish Astacidea Parastacidae 3 5.2-5.4 50-61 Commercial supplier destructor [Astacida] Pagurus Hermit crab Anomura Paguridae 5 1.6-2.8 3.5-29.7 Biologische Anstalt Helgoland bernhardus [Anomalal Cancer Common crab Brachyura Cancridae 5 10.2-16.0 167-688 Biologische Anstalt Helgoland pagitnts [Brachyura] * Name according to Gruner (1993), followed [in brackets] by name according to terminology proposed in most recent investigations of decapod phylogeny by Sandeman ct ui. ( 1993) and Scholtz and Richter (1995). f n = number of specimens examined. vibratome into serial sections with a thickness of 70 or 100 /Am. After degelatinization in a warm water bath, incubation in 2 N HC1 for 20 min. and rinsing (4 X 15 min) in SPB. the free-floating sections were incubated overnight in anti-BrdU ("Cell Proliferation Kit" from Boehringer or anti-BrdU con- taining nucleases from Amersham). Subsequently, sections were rinsed in SPB (4 X 30 min) and incubated for 4 h in goat anti-mouse CY3-labeled secondary antibody (Jackson Immunoresearch) diluted 1:100 in SPB. After final rinsing in SPB (4 x 30 min), the labeled sections were coverslipped in glycerol/SPB. Sections were viewed and photographed using a micro- scope with bright-field optics and epifluorescence (Olympus BH-2 or IX-50). Despite increased section thickness (70 jam compared to the 20-/zm-thick cryostat sections normally used in BrdU immunostainings of mammalian tissue), no problems with antibody penetration were encountered. In- tensely labeled nuclei were detectable throughout the entire thickness of the vibratome sections. The method we used here for the in vivo BrdU-labeling was the same as in a previous study (Schmidt. 1997). where it was shown to be highly reliable since in all 16 shore crabs (Carcinus nuie- nas) treated by this method, somata in the LC were dis- tinctly labeled. Results In vivo labeling with BrdU resulted in the clear visual- ization of nuclei in soma clusters of the central olfactory pathway in most of the specimens. All tested species showed labeling of nuclei in the paired lateral soma clusters (LC; soma cluster # 10 according to Sandeman ct al., 1992) of the olfactory deutocerebrum (Figs. 1; 2a, b; 3a; 4; 5a, b; 6a). These clusters exclusively contain somata of ascending olfactory projection neurons (OPN). The BrdU-positive nu- clei in the LC differed from other labeled nuclei that were present throughout almost all brain areas by their specific location, arrangement, and shape. The labeled nuclei in the LC always occurred as a tightly packed, string- or ball-like group that was located in the ventral part of the cluster directly adjacent to the neuropil of the olfactory lobe (OL). They were spherical and slightly smaller than were the Figure 1. Proliferating cells in the central olfactory pathway of adult rock shrimp (Sicyuniu hrevirostris: Penaeidea (Caridea]; Sicyonidae) iden- tified by in vivo BrdU-labeling. Light micrographs of vibratome sections. In the lateral soma cluster (LC) a compact, string-like group of small BrdU-positive nuclei (arrow) is located close to the neuropil of the olfac- tory lobe (OL). Inset: higher magnification shows spherical, and hence neurnnal shape of labeled somata. 130 M. SCHMIDT AND S. HARZSCH Figure 2. Proliferating cells in the central olfactory pathway of late juvenile (a-c; 49-57 g) and adult (d. e; 140 g) spiny lobster (Punulirus argus: Palinura [Achelata]; Palinuridae) identified by in rnv> BrdU-labeling. Light micrographs of vibratome sections, (a) Overview of olfactory deutocerebrum (hemibrain). In the lateral sonia cluster (LC) and the medial soma cluster (MC), a compact group of small BrdU-positive nuclei (arrows) is located close to the neuropil of the olfactory lobe (OL). Note presence of the large accessory lobe (AL). (b. e ) BrdU-positive nuclei in the lateral soma cluster at higher magnification. Note spherical and hence neuronal shape of labeled nuclei in e. (c, d) BrdU-positive nuclei in the medial soma cluster at higher magnification. Note spherical and hence neuronal shape of labeled nuclei in d. unlabeled nuclei of OPN somata located in the periphery of the LC (Fig. 2b, e). For the species in which specimens of different size (and hence age) were tested (Ptimilinis argus. Cancer pagurus), only in Canca t "u gurus did we find obvious differences in the number of BrdU-positive nuclei per LC between spec- imens. In Cancer pagurus, the LC of three large specimens (650-700 g and ca. 16 cm carapace width) showed no labeled somata or only - u uit were very weakly labeled (some nuclei belonging to < icr cell types were intensely labeled, showing that BrdU 1 .id been available for incorpo- ration). However, typical groups ; \ strongly BrdU-positive nuclei occurred in all LCs in two smaller, and hence younger, adults ( 160-200 g and 10-1 1 cm carapace width). In Panulirus argus ( 15-208 g and carapace length of 2.8-6 cm), we detected no obvious difference in the number of BrdU-positive nuclei per LC. The number of BrdU-positive nuclei per LC differed considerably among species. In Sicyimia brevirostris. only a few nuclei (less then 15) per LC were BrdU-positive (Fig. 1). In Cherax destructor (Fig. 4), as well as in Pagurus bernhardiis (Fig. 5) and Cancer pagurus (Fig. 6), the num- ber of BrdU-positive nuclei per LC was somewhat higher (20 to 40). In Homants americanus (Fig. 3a), about 50 nuclei per LC, and in Panulirus argus (Fig. 2a, b, e), more than 80 nuclei per LC were BrdU-positive. Of the species included in this study, two (Panulirus argus and Homarus americanus) showed proliferation not only in the LC but also in the paired medial soma cluster (MC; soma clusters # 9 and 11 according to Sandeman el NKUROGENESIS IN OLFACTORY PATHWAY OH ADULT DECAPODS 131 Figure 3. Proliferating cells in the central olfactory pathway of adult American lobster (Homurus nmeri- ctinus: Astacidea [Homarida]: Homaridae) identified by in vivo BrdU-labeling. Light micrographs of vibratome sections, (a) In the lateral soma cluster (LC). a compact group of small BrdU-positive nuclei (arrow) is located close to the neuropil of the olfactory lobe (OL). (b) In the medial soma cluster (MC) of this specimen, a string of slightly elongated nuclei is BrdU positive (arrows). /., 1992) containing the cell bodies of local interneurons of the olfactory deutocerebrum (Figs. 2a, c, d, 3b). Only Pann- lirits argus showed a tightly packed group of BrdU-positive nuclei in the MCs in all tested specimens, which ranged from late juveniles to mid-sized sexually mature adults. In each case this group comprised 15-25 strongly labeled spherical nuclei that were slightly smaller than the nuclei of somata located in the periphery of the MC (Fig. 2c, d). The group of BrdU-positive nuclei was located in the ventral- most area of the MC adjacent to the neuropils of the olfac- Figure 4. Proliferating cells in the central olfactory pathway of adult Australian crayfish (Clierax destructor: Astacidea [Astacida]; Parastaci- dae) identified by in vivo BrdU-labeling. Light micrograph of a vibratome section. A compact group of small Brdl'-positive nuclei (arrow) is located in the LC close to the neuropils of the olfactory lobe and the accessory lobe. tory and accessory lobes (Fig. 2a, c). In Hoimirns america- nus. evidence for proliferation in the MC occurred in only two of five tested specimens. The MCs of these two animals contained a group of about 10 strongly BrdU-positive nuclei that were slightly elongated and arranged in a string trans- versing the cluster from its innermost aspect to the periph- ery (Fig. 3b). Two of the species included in this study, the hermit crab (Pagurus bemhardus) and the true crab (Cancer pagurus), possessed BrdU-positive nuclei arranged in an obvious group in the LC and in a paired soma cluster adjacent to the hemiellipsoid body (HBC). The morphology of the neurons whose somata constitute the HBCs has not yet been eluci- dated, and it remains unclear whether they represent local interneurons of the hemiellipsoid bodies. In Pagurus bem- hardus, the eyestalks (n = 10) of all specimens had a group of 6-15 BrdU-positive nuclei in the HBC (Fig. 5c). In only two of the eyestalks of the three large specimens of Cancer pagurus did we find a cluster of BrdU-positive nuclei in the HBC. In contrast, the eyestalks (;; = 4) of the two smaller specimens contained a cluster of BrdU-positive nuclei in the HBC. In each case, the number of BrdU-positive nuclei ranged from 20 to 30 per HBC (Fig. 6b). The BrdU-positive nuclei in the HBC of Pagurus bemhardus and Cancer pagurus were spherical and slightly smaller than the nuclei of unlabeled somata in the periphery of the HBC. In each case, the group of BrdU-positive nuclei was located close to the HB neuropil. None of the other species had a group of BrdU-positive nuclei present in the HBC. This negative result is supported best by the spiny lobster (Panulinia argus). where 13 eyestalks were analyzed, and the rock 132 M. SCHMIDT AND S. HARZSCH Figure 5. Proliferating cells in the central olfactory pathway of adult hermit crab (Pagurus bernlninhis: Anomura [Anomala]; Paguridae) identified by in vim BrdU-labeling. Light micrographs of vibratome sections. (a. b) Central brain. In the lateral soma clusters (LC). a compact group of small BrdU-positive nuclei (arrows) is located close to the neuropil of the olfactory lobe (OL). (c) Eyestalk ganglion. In the soma cluster (HBC) of the hemiellipsoid body (HB). a compact group ot small BrdLI-posilive nuclei is located close to the HB neuropil. Note scattered BrdU-positive nuclei (arrowheads) in the soma clusters of visual neuropils (medulla interna = MI; medulla externa = ME). Inset: Higher magnification shows shows spherical and hence neuronal shape of BrdU-positive nuclei. shrimp (Sicyonia brevirostris), where 5 eyestalks were an- alyzed. However, in all tested species, very few nuclei located in the HBC were labeled in most specimens; the same was true for other soma clusters of the eyestalk gan- glia, especially the soma clusters of the three visual neuro- pils. Most of these BrdU-positive nuclei were either flat and elongated, and thus probably represent glial cells; or they were more spherical but clearly associated with arterioles passing through the soma clusters, and thus probably rep- resent connecthr tissue cells constituting the wall of these hemolymph vesM. Hut some nuclei in soma clusters were spherical and not ob iously associated with arterioles; thus they might represent lu-uronal precursor cells. Discussion We demonstrate by in vm> labeling with BrdU that mi- totic activity in the neuronal soma clusters of the central olfactory pathway appears to be a common feature of adult decapod crustaceans; this result corroborates and substan- tially extends previous findings for juvenile and adult shore crabs (Ccircintis niaenas) and for the Australian crayfish (Che rax destructor) (Schmidt, 1997; Sandeman et ui, 1998). In all three studies the BrdU-positive nuclei in the soma clusters of the central olfactory pathway were small and almost spherical, and clearly resemble neurons. How- ever, positive BrdU-labeling of neuron-like nuclei in soma clusters that are predominantly, but not exclusively, com- posed of neuronal cell bodies (the clusters also contain the somata of glial cells and of cells forming the walls of arterioles) is not conclusive evidence for the generation of new neurons, i.e., neurogenesis. BrdU-labeling could be explained in several alternative ways: differentiation of pro- liferating cells into non-neuronal cells, programmed cell death shortly after birth of new cells, or DNA synthesis without subsequent mitosis. In the case of Carcinus maenas, however, various lines of evidence rule out these alterna- tives and thus strongly support the notion that proliferation in the soma clusters of the central olfactory pathway indeed NEUROGENESIS IN OLFACTORY PATHWAY OF ADULT DECAPODS 133 Figure 6. Proliferating cells in the central olfactory pathway of adult common crab (Cwu-cr pagurus: Brachyura [Brachyura]; Cancridae) identified by in rnvi BrdU-labeling. Light micrographs of vibratome sections, (a) Central brain. In the lateral soma cluster (LC). a compact group of small BrdU-positive nuclei (arrows) is located close to the neuropil of the olfactory lobe (OL). (b) Eyestalk ganglion. In the soma cluster (HBC) of the hemiellipsoid body, a compact group of small BrdU-positive nuclei is present (arrow). reflects neurogenesis (Schmidt, 1997). This conclusion is substantially supported by the finding that in Ciircinus niae- nas and in Che rax destructor, the olfactory projection neu- rons whose somata constitute the LCs increase continuously in number throughout juvenile and adult growth (Schmidt. 1997; Sandeman et cii, 1998). Throughout juvenile and adult life the number of olfactory projection neurons dou- bles in Carcinus maenas (from ca. 24.000 to cci. 48.000 per brain: Schmidt. 1997). and it approximately triples in Che rax destructor (from ca. 60.000 to ca. 180.000 per brain: Sandeman et /.. 1998). Thus, the generation of new olfactory projection neurons throughout juvenile and adult life appears to be well established in these species. The groups of proliferating cells in the respective soma clusters most likely reflect the cellular basis of this neuronal prolif- eration. We conclude that, at least for the LC, the occur- rence of groups of BrdU-positive, neuron-like somata truly indicates neurogenesis. although a direct confirmation of this conclusion (for instance by double-labeling BrdU-pos- itive cells with neuronal markers) is still lacking. Another important issue is whether the occurrence of groups of BrdU-positive nuclei in the MC and the HBC can also be regarded as an indication of neurogenesis. At present the answer is uncertain, since counts of the respective neu- ron types have not yet been performed in the species we studied. Two lines of evidence support this assumption for the HBC. First, the BrdU-positive nuclei in the HBC are very similar in size, shape, and location (at the inside of the cluster, directly adjacent to the neuropil) to the BrdU- positive nuclei in the LC, in all three species in which they were found (this study; Schmidt, 1997). Second, in Carci- nus maenas. groups of BrdU-positive neuron-like nuclei were present in the HBC after a long post-injection survival time (1 month), ruling out the possibilities that newly born cells collectively either undergo programmed cell death or differentiate into non-neuronal cells (Schmidt. 1997). For the MC. only the very similar size, shape, and location (also at the inside of the cluster, directly adjacent to the neuropil) of the BrdU-positive nuclei to the BrdU-positive nuclei in the LC supports the assumption that mitotic activity as shown by the BrdU-labeling in this study reflects neurogen- esis. In a strict sense this argument applies only to the spiny lobster (Panulirus argus). in which BrdU-positive nuclei were consistently present in the MC and were indeed indis- tinguishable from their counterparts in the LC in terms of the above mentioned criteria. However, in the American lobster (Homants americamis). where we also observed BrdU-positive nuclei in the MC, the situation is not as clear. BrdU-positive nuclei forming a group were present in the MCs in only two of the five tested specimens. Furthermore, the labeled nuclei were slightly elongated compared to the BrdU-positive nuclei of the LC, and the group they formed was neither compact nor located close to the OL/AL neu- ropil (as in the spiny lobster). Therefore it remains ques- tionable whether the proliferating cells in the MC of Ho- manis americamis give rise to new neurons or to other cell types. The shape and arrangement of these BrdU-positive nuclei suggest that they might represent cells forming the wall of arterioles. Double labeling experiments with neuro- nal markers are necessary to clarify this point. Recently, the occurrence of BrdU-positive nuclei in the MC was reported lor juveniles and adults of the Australian crayfish (Cherax destructor) (Sandeman et a/.. 1998). Here we could not reproduce this result in the three adult specimens tested, and 134 M. SCHMIDT AND S. HARZSCH Sicyonia brevimstns Panulirus argus Homarus amencanus Cherax destructor Pagurus bernhardus Cancer pagurus Carcmus maenas "SHR MPS" "SPINY LOBSTERS' CLAWED OBSTERS CRAYFISH "HERMIT CRABS" "TRUE CRABS Caridea Achelata Homanda Astacida Anomala Brachyura Figure 7. Occurrence of neurogenesis in the central olfactory pathway of juvenile and adult decapod crustaceans. Decapod phylogeny and terminology of taxa are according to Sandeman et al., (1993) and Scholtz and Richter (1995). Common names of taxa included in this study are given in parenthesis. Names of species studied to date (Carcinus maenas from Schmidt. 1997; Cherax destructor from Sandeman et al., 1998, and from this study, all others from this study) are given in italics. Brain morphology highlights the central olfactory pathway with olfactory lobe, accessory lobe, hemiellipsoid body, the connecting olfactory globular tracts and associated soma clusters: lateral cluster (LC). medial cluster (MC), and hemiellipsoid body cluster (HBC). Soma clusters with typical groups of BrdU-positive nuclei are in black; in questionable cases they are dotted. Note that neuronal proliferation in the LC indicating neurogenesis among olfactory projection neurons is a common feature of adult decapod crustaceans. Proliferation in the other soma clusters of the central olfactory pathway (in the MC indicating neurogenesis among local interaeurons of the olfactory deutocerebrum; in the HBC indicating neurogenesis among local interneurons of the hemiellipsoid body) appears to be linked to the occurrence (in spiny lobsters, clawed lobsters, and crayfish) and later reduction (in hermit crabs and true crabs) of the accessory lobe in the evolutionary history of decapods. the reasons for this difference remain unclear. One possible explanation is that the adult animals tested by Sandeman et al., (1998) were significantly younger than our specimens, since the position of the BrdU-positive nuclei of the adult illustrated in their repo.t (in the center of the MC and at the outer edge of the i ^embles the situation in early juveniles. Our finding that no Bi , 'ive somata were present in the LCs of three large adults of Cancer pagurus after an appropriate post-injection survival lime (5.5 h) is the only instance so far of a negative result of this kind. In the previous study (Schmidt, 1997). all specimens of Carcinus maenas (n = 16), and in this study, all other specimens of Cancer pagurus (n = 2) as well as all specimens of the other species (/; = 20) showed a typical group of BrdU- positive nuclei in each LC. This indicates that proliferation in the LC is a continuous process that apparently is not related to the molt cycle. That the negative result in Cancer pagurus occurred in very large and hence old adults sug- gests that, at least in some decapod species, the generation of new olfactory projection neurons may not continue for the entire life-span, but ceases in senescent animals. For some species of decapod crustaceans, enough growth data are available to allow absolute age to be estimated from body size. According to such estimates, the large specimens of Cancer pagurus, in which we did not detect BrdU- NEUROGENESIS IN OLFACTORY PATHWAY OF ADULT DECAPODS 135 positive nuclei in the LC, were at least 7 years old; whereas the smaller adults, in which we found typical groups of BrdU-positive nuclei in each LC. were about 5 years old (Gruner, 1993). In comparison, the American lobsters that showed typical groups of BrdU-positive nuclei in each LC in our experiments were at least 8 and probably more than 10 years old (Hartnoll. 1982) and the largest specimen of Panulinis argus in our study that also showed normal proliferation in the LC was at least 5 years of age (Travis. 19541. Common features of the proliferation in the soma clusters of the central olfactory pathway in adult decapod crusta- ceans include the very uniform relative size, the shape, and the location of the BrdU-positive nuclei (Schmidt, 1997: this study). In each case these nuclei are among the smallest of the respective soma clusters; they are mostly spherical and thus resemble neurons; and they are located in very restricted regions of the respective soma clusters close to the neuropil. These cells apparently represent a unique type of presumptive neuronal precursor cell differing in several aspects from the neuronal precursor cells that are active during embryonic and larval development of crustaceans and insects. The latter precursor cells are neuroblasts that differentiate from the ectoderm and ganglion mother cells that arise from the neuroblasts through a series of unequal mitoses; each produces two neurons by an equal mitosis (e.g., Thomas el al.. 1984; Harzsch and Dawirs, 1994). Typically, neuroblasts and ganglion mother cells are con- siderably larger than terminally differentiated neurons, and they are located at the outer edge of the soma clusters formed by their mitotic activity (e.g., Nordlander and Ed- wards. 1970; Ito and Hotta. 1992; Harzsch and Dawirs, 1994). The mitotically active precursor cells that were re- cently demonstrated in the LC of early postlarval crabs share these criteria and therefore are regarded as delayed ganglion mother cells (Harzsch and Dawirs, 1996). We did not find BrdU-positive nuclei resembling typical large neu- roblasts or ganglion mother cells in any specimen, indicat- ing that the mitotic activity of these cells ceases during juvenile development. Since BrdU-positive groups of nuclei were detected in the LC of all tested species representing the major infraorders of decapod crustaceans (Fig. 7). we conclude that prolifer- ation in the LC and hence neurogenesis of olfactory pro- jection neurons most likely represents a general principle common to the brain of late juvenile and adult decapod crustaceans. In contrast, neurogenesis of local interneurons of the olfactory deutocerebrum and of neurons whose so- mata constitute the hemiellipsoid body clusters clearly is not a common feature among adult decapod crustaceans, since evidence for neuronal proliferation in the MC and in the HBC. respectively, occurs only in some of the tested spe- cies. In the case of the MC, we found clear evidence for neuronal proliferation only in the spiny lobster (Panulinis argus). Although we detected BrdU-positive nuclei in the MCs of two American lobsters (Homanis aineiicaims), it remains questionable whether they represent neuronal pre- cursor cells. Proliferation has recently been reported (San- deman et al., 1998) in the MCs of the Australian crayfish (Cherax destructor), but we were unable to reproduce this result. In the case of the HBC, on the other hand, evidence for neurogenesis (i.e., clustered BrdU-positive, neuron-like nuclei) occurs only in the brachyuran crabs Ccircintis nnic- nas (Schmidt. 1997) and Cancer pagurus and in the hermit crab Pagurus henihardus. Clearly, not enough species have been tested to exclude simple species-specific differences as the basis of this variability. However, it is striking that presumptive neurogenesis of local interneurons of the ol- factory deutocerebrum appears to be restricted to species with accessory lobes (Fig. 7) that is, glomerular neuropils of the deutocerebrum connected to the olfactory lobes (Schmidt and Ache, 1997) whereas neurogenesis of neu- rons residing in the HBC occurs only in species in which the accessory lobes are thought to be reduced secondarily (Scholtz and Richter, 1995; Sandeman et al., 1993). Clearly, the function of adult neurogenesis in decapod crustaceans and other animals remains to be determined. From the studies in decapod crustaceans, as well as from those in vertebrates and insects, it emerges that adult neu- rogenesis is either limited to or is at least very prominent in the central olfactory pathway (e.g., Kaplan et al., 1985; Corotto et al., 1993; Luskin, 1993; Okano et al., 1993; Cayre et al., 1994, 1996; Schmidt, 1997). The olfactory pathway also is the only sensory pathway in which primary receptor neurons show a lifelong, continuous turnover. This has been known for vertebrates including mammals for many years (e.g.. Graziadei, 1973) and recently was also found for a decapod crustacean. Cherax destructor (Sande- man and Sandeman, 1996). It seems reasonable to speculate that neurogenesis in the olfactory pathway of the adult brain may be linked to the continuous turnover of olfactory re- ceptor neurons in the periphery. This could allow a lifelong "adaptation" of the olfactory system to ever-changing ol- factory environments and would be advantageous for long- lived animals such as most vertebrates and decapod crusta- ceans. Acknowledgments This study was supported in part by grants from the Deutsche Forschungsgemeinschaft (Schm 738/4-1; Ha 2540/1-1). We wish to thank Dr. B. W. Ache (Whitney Laboratory, University of Florida) for providing lab space and spiny lobsters, and Dr. J. Nunez (Whitney Laboratory, University of Florida) for providing rock shrimp. Literature Cited Bayer, S. A. 1982. Changes in the total number of dentate granule cells in juvenile and adult rats: a correlated volumetric and 3H-thymidine autoradiographic study. Exp. Brain Res. 46: 315-323. 136 M. SCHMIDT AND S. HARZSCH Bieber, M., and P .'drier. 1979. Brain growth during the adult stage of a holome! . ious insect. Naturwissenschaften 66: 426. Blaustein. I) ' , <'. D. Derby, R. B. Simmons, and A. C. Beall. 1988. Structup . : the brain and medulla terminalis of the spiny lobster Piiniil ../yn.s and the crayfish Procambarus clnrkii, with an empha- sis on ..itactory centers. J. Crust. Bin/. 8: 493-519. Boeckh, J., K. D. Ernst, H. Sass, and U. VValdow. 1984. Anatomical and physiological characteristics of individual neurons in the central antennul pathway of insects. J. Insect Physio/. 30: 15-26. Cayre, M., C. Strambi, and A. Strambi. 1994. Neurogenesis in an adult insect brain and its hormonal control. Nature 368: 57-59. C'avre, M., C. Strambi. P. Charpin, R. Augier, M. R. Meyer, J. S. Edwards, and A. Strambi. 1996. Neurogenesis in adult insect mush- room bodies. J. Comp. Neural. 371: 300-310. Corotto, F. S., J. A. Henegar, and J. A. Maruniak. 1993. Neurogenesis persists in the subependy mal layer of the adult mouse brain. Neurosci. Lett. 149: 111-114. Crespo, D., B. B. Stanfield, and W. M. Cowan. 1986. Evidence that late-generated granule cells do not simply replace earlier formed neu- rons in the rat dentate gyrus. Exp. Brain Res. 62: 541-548. Davis, R. L. 1993. Mushroom bodies and Drosophila learning. Neuron 11: 1-14. Fahrbach, S. E., J. L. Strande. and G. E. Robinson. 1995. Neurogen- esis is absent in the brains of adult honey bees and does not explain behavioral neuroplasticity. Neurosci. Lett. 197: 145-148. Graziadei, P. P. C. 1973. Cell dynamics in the olfactory mucosa. Tissue Cell 5: 113-131. Gruner, H.-E., ed. 1993. Lehrbuch der spe;iellen Zoologie. Band I: Wirbe/lose Tiere. 4. Teil: Arthropoda (ohne Insecta). Gustav Fischer Verlag. Stuttgart. Hartnoll, R. G. 1982. Growth. Pp. 1 1 1-196 in The Biology of Crustacea. Vol. 2. Emhryo/ogy, Morphology, and Genetics. L. G. Abele. ed. Academic Press, New York. Harzsch, S., and R. R. Dawirs. 1994. Neurogenesis in larval stages of the spider crab Hyas araneus (Decapoda, Brachyura): proliferation of neuroblasts in the ventral nerve cord. Rota's Arch. Dev. Bio/. 204: 93-100. Harzsch, S., and R. R. Dawirs. 1996. Neurogenesis in the developing crab brain: postembryonic generation of neurons persists beyond meta- morphosis. J. Neurobiol. 29: 384-398. Ito, K., and V. Hotta. 1992. Proliferation pattern of postembryonic neuroblasts in the brain of Drosophila melanogaster. Dev. Biol. 149: 134-148. Kaplan. M. S.. and J. \V. Hinds. 1977. Neurogenesis in the adult rat: electron microscopic analysis of light radioautographs. Science 197: 1092-1094. Kaplan, M. S., N. A. McNelly. and J. W. Hinds. 1985. Population dynamics of adult-formed granule neurons of the rat olfactory bulb. J. Comp. Neural. 239: 117-125. Kuhn, H. G., H. Dickinson-Anson. and F. H. Gage. 1996. Neurogen- esis in the dentate gyrus of the adult rat: age-related decrease of neuronal progenitor proliferation. J. Neurosci. 16: 2027-2033. Luskin, M. B. 1993. Restricted proliferation and migration of postnatally generated IK d rived from the subventricular zone. Neuron 11: 173-189. Mellon, D., Jr., V. id M. D. Lawrence. 1992. Anatomv and fine structure of neurons in the deutocerebral projection pathway of the crayfish olfactory system. J. Comp. Neural. 321: 93-111. Menzel, R.. C. Durst, J. Erber, S. Eichmiiller, M. Hammer, H. Hilde- brandt, J. Mauelshagen, U. Miiller, H. Rosenboom, J. Rybak, S. Schafer, and A. Scheidler. 1994. The mushroom bodies in the honeybee: from molecules to behavior. Pp. 81-102 in Fonschntte der Zoologie, Vol. 39. Neural Basis of Behavioural Adaptations. K. Schild- berger and N. Eisner, eds. Gustav Fischer Verlag. New York. Morshead, C. M., B. A. Reynolds, C. G. Craig, M. W. McBurney. W. A. Staines, D. Morassutti, S. Weiss, and D. Van der Kooy. 1994. Neural stem cells in the adult mammalian forebrain: a relatively qui- escent subpopulation of subependymal cells. Neuron 13: 1071-1082. Nordlander, R. H., and J. S. Edwards. 1970. Postembryonic brain development in the monarch butterfly. Danaus p/e.\ippus plexippus L. III. Morphogenesis of centers other than the optic lobes. Wilhe/m Roux' Arch. 164: 247-260. Okano, H. J., D. \V. Pfaff, and R. B. Gibbs. 1993. RB and Cdc2 expression in brain: correlations with 3 H-thymidine incorporation and neurogenesis. J. Neurosci. 13: 2930-2938. Paton, J. A., and F. N. Nottebohm. 1984. Neurons generated in the adult brain are recruited into functional circuits. Science 225: 1046- 1048. Purves, D. 1988. Body and Brain. A Trophic Theory of Neuronal Con- nections. Harvard University Press, Cambridge. Raymond, P. A., and S. S. Easter, Jr. 1983. Postembryonic growth of the optic tectum in goldfish. I. Location of germinal cells and numbers of neurons produced. J. Neurosci. 3: 1077-1091. Sandeman, R. E.. and D. C. Sandeman. 1996. Pre- and postembryonic development, growth and turnover of olfactory receptor neurons in crayfish antennules. J. Exp. Biol. 199: 2409-2418. Sandeman, D., R. Sandeman, C. Derby, and M. Schmidt. 1992. Mor- phology of the brain of crayfish, crabs, and spiny lobsters: a common nomenclature for homologous structures. Biol. Bull. 183: 304-326. Sandeman, D. C.. G. Scholtz, and R. E. Sandeman. 1993. Brain evolution in decapod Crustacea. J. Exp. Zool. 265: 1 12-133. Sandeman, R., D. Clarke, D. Sandeman. and M. Manly. 1998. Growth-related and antennular amputation-induced changes in the ol- factory centers of crayfish brain. /. Neurosci. 18: 6195-6206. Schmidt, M. 1997. Continuous neurogenesis in the olfactory brain of adult shore crabs, Carcinus maenas. Brain Res. 762: 131-143. Schmidt, M., and B. W. Ache. 1997. Immunocytochemical analysis of glomerular regionalization and neuronal diversity in the olfactory deu- tocerebrum of the spiny lobster. Cell Tissue Res. 287: 541-563. Scholtz, G., and S. Richter. 1995. Phylogenetic systematic!, of the reptantian Decapoda (Crustacea, Malacostraca). Zool. J. Linn. Soc. 113: 289-328. Technau, G. M. 1984. Fiber number in the mushroom bodies of adult Drosophila melanogaster depends on age. sex and experience. J. Neu- rogenet. 1: 1 13-126. Thomas, J. B., M. J. Bastiani, M. Bate, and C. S. Goodman. 1984. From grasshopper to Drosophiht: a common plan tor neuronal devel- opment. Nature 310: 203-207. Travis, D. F. 1954. The molting cycle of the spiny lobster Panulirus argus Latreille. I. Molting and growth in laboratory-maintained indi- viduals. Biol. Bull. 107: 433-450. Reference: Biol. Bull. 196: 137-144. (April 1999) Comparative Study of Temporal Resolution in the Visual Systems of Mesopelagic Crustaceans TAMARA M. FRANK Harbor Branch Oceanographic Institution, 5600 U.S. 1 N. Ft. Pierce. Florida 34946 Abstract. The temporal characteristics of the visual sys- tems of eight species of mesopelagic crustaceans were stud- ied using the electroretinogram (ERG). Experiments were conducted on shipboard, using dark-captured specimens collected off the south coast of Cuba. As one would expect based on the relative intensity differences in their light environments, the deepest living species. Systellaspis debi- lis and Sergio filictum, have low maximum critical flicker fusion frequencies (CFFs) of 21-25 Hz. whereas the shal- lower living species Oploplwrus gracilirostris and Janicella spinacauda have higher maximum CFFs (31-32 Hz). One of the shallowest living species, Funchalia villosa, has an unusually low maximum CFF (24 Hz), which may be a function of working with a dark-adapted eye. Two of the bilobed euphausiid species, Nematobrachionflexipes and N. sexspinosus, have very high maximum CFFs (44-57 Hz), comparable to those of surface-dwelling crabs, even though they live between 400 and 600 m. The maximum CFF of Sr\-Iocheiron imixininni. a shallower living bilobed eu- phausiid, is only 36 Hz, indicating that maximum CFF among the euphausiids cannot be correlated with depth of occurrence. The unusually high flicker fusion frequency of the deeper living euphausiids may be correlated to their preference for bioluminescent prey. Introduction Autrum's studies of insect photoreceptors in the 1950s gave rise to the idea that the response dynamics of the retina match the habitat and lifestyle of the organism. In these classic studies, he established that the eyes of rapidly mov- ing day-active species have better temporal resolution, as indicated by flicker fusion frequencies of 200-300 Hz. than the eyes of slower moving night-active forms, with flicker fusion frequencies of 10-20 Hz (Autrum, 1950, 1958: Au- Received 4 June 1998; accepted 9 December 1998. trum and Stocker, 1952). These extracellular studies, using the electroretinogram (ERG), which is the summed mass response from a large number of receptor cells, were later supported by studies on the intracellular responses of single cells (Howard et ai, 1984; de Souza and Ventura, 1989). Since the light environment of mesopelagic crustaceans is similar to that of nocturnal insects, one might predict that they would also have fairly low temporal resolution, and that temporal resolution would be correlated with daytime depth of occurrence. Although the spatial resolution, which is a function of the structure and optics of photoreceptors, has been studied in a number of mesopelagic species (see Cronin, 1986; Land, 1990, for review), the temporal reso- lution, which is a function of the membrane properties of the receptor cells themselves (see Weckstrom and Laughlin. 1995, for review), has received little attention. As shown by the theoretical analysis of Srinivasan and Bernard (1975), visual acuity is dependent on both the spatial resolution and the temporal resolution of the eye, because for most organ- isms, visual targets are rarely stationary. Either the organ- isms themselves are actively moving, so the environment is in motion with respect to their photoreceptors, or their photoreceptors are moving because of muscle tremor or nystagmus. Srinivasan and Bernard (1975) determined that at angular velocities (of the object with respect to the organism viewing it) above a critical value, spatial resolu- tion is more dependent on the temporal properties of the photoreceptor cells than on the structural optics of the eye. Therefore, for any comprehensive analysis of the visual system of an organism, both the spatial and temporal char- acteristics of the photoreceptor need to be studied. Studies on temporal resolution in mesopelagic organisms are rare, due to the difficulties in collecting visually com- petent organisms and keeping them alive during transport back to shore-based labs. The only published study on the temporal characteristics of the visual systems of mesope- 137 138 T. M. FRANK lagic organising is by Moeller and Case (1995), in which they demon slrated that two species of deep-sea crustaceans have very low flicker fusion frequencies (8-12 Hz), com- pared to much higher maximum flicker fusion frequencies (50-60 Hz) for shallow-water crabs (Brocker, 1935; Cro- zier and Wolf, 1939). However, Moeller and Case (1995) measured the critical flicker fusion frequency at threshold light intensities, which is difficult to compare between spe- cies because ( 1 ) the threshold sensitivity of a crustacean eye measured using the ERG technique varies considerably from preparation to preparation (pers. obs.), and (2) critical flicker fusion frequency is dependent upon the intensity of the stimulus light (Brocker. 1935; Crozier and Wolf, 1939; Croziere//., 1939). A less problematic characteristic to use for comparative studies of temporal resolution is the nia.\i- nnini critical flicker fusion frequency. This is the maximum flicker rate that the eye is capable of following at any light intensity. In the current study, the maximum critical flicker fusion frequencies of the photoreceptors of eight species of mesopelagic crustaceans from a variety of depths (250-900 m) were examined using ERG recordings. The results of these experiments indicate that several species from this dim light environment have surprisingly high maximum flicker fusion rates. These unusually high rates do not ap- pear to be a function of depth of occurrence, but rather, appear to be more closely correlated with the biolumines- cence of the preferred prey. Materials and Methods Animal collections The crustacean species used in this study (Table I) were collected off the south coast of Cuba on a research cruise aboard the RV Seward Johnson, with a 2.4 X 1.8 m Tucker Table I Daytime ilcprh distribution of crustaceans in thi.\ Species Depth (ml* Family Euphausiacidae Stylocheiron ^^n Nematobrachi* Nematobrachion 650 nm; Wratten Filter 79B). Signals were ampli- fied with a Haer Microelectrode Amplifier (Model X Cell-3) used in conjunction with a high-impedance probe to eliminate electrode polarization artifacts (Kugel, 1977). Low-frequency filters were set to minimal filtering (0.01- 0.1 Hz) to minimize distortion of the AC-amplified signal. Data were digitized using a program written in LabView (National Instruments, Inc.), and stored to disk for later analysis. This study was conducted on shipboard because the or- ganisms used will not survive transport to a shore-based laboratory. Due to the difficulties inherent in working on a moving and rolling vessel, only extracellular electrophysi- ology was possible. Light stimuli Test flashes of 490 nm light from an American Instru- ments SA monochromator (Model H-20) were delivered to the eye via a fused silica light guide, positioned so the circle TEMPORAL RESOLUTION IN CRUSTACEAN VISION 139 of light at the output was larger than the eye. A piece of lens tissue between the light guide and the eye served as a diffuser. Flash duration was controlled by a Uniblitz shutter (Model T132) under computer control, such that a 50% duty cycle (50:50 light:dark ratio) was maintained. Irradiance was controlled with a neutral-density wheel driven by a stepper motor under computer control, and calibrated with a UDT optometer (United Detector Technology Model S370) and radiometric probe with point calibrations provided by UDT. Procedure Although mesopelagic crustaceans are relatively insensi- tive to red light (Frank and Case. l"88a), the dim red preparation light did produce a small degree of light adap- tation. Therefore, after electrode placement, the specimen was dark-adapted until the response to a test flash, given every 5 min, had not changed for 1 h, indicating that the eye was in its fully dark-adapted condition. A flickering light stimulus of 1.5-s duration was then presented to the dark- adapted eye. To ensure that every flicker stimulus was presented to a fully dark-adapted eye, a dim 100-ms test flash that elicited a 50-juV response (the smallest response reliably discernible from background noise) in the fully dark-adapted eye was presented to the eye after every flicker stimulus, and the eye was allowed to re-dark-adapt until the test flash response had recovered to 50 /J.V, before the next flicker stimulus was presented. The flicker rate for subse- quent stimuli was increased until critical flicker fusion was achieved; this is defined as the point at which the eye can no longer produce a modulated electrical signal that remains in phase with the flickering light. Irradiance was then in- creased by one log unit, and the flicker rate of the stimulus light was increased until fusion was again achieved. Maxi- mum critical flicker fusion frequency (CFF) is defined as the point at which further increases in irradiance do not result in a faster flicker fusion frequency. All maximum CFFs are in reference to dark-adapted eyes. The responses to single 100-ms flashes of 490-nm light of varying irradiances were also measured, starting from irra- diances generating a threshold response, and continuing to the point at which further increases in irradiance produced no further increases in response amplitude. In some prepa- rations, the response was not saturated at the maximum stimulus light irradiance. For those preparations in which response saturation was reached, the log 1 10 was determined; this is the log of the stimulus irradiance eliciting a response that is 10% of the maximum amplitude. The log l l() was used as an estimate of the relative sensitivity of the photo- receptors (after de Souza and Ventura, 1989). For all species, latencies were measured using a response amplitude that was 10% of the maximum response. Re- sponse latency is defined as the time from the start of the light stimulus to the start of the ERG. Results The critical flicker fusion frequency measured via the extracellular electroretinogram is dependent on a variety of factors, namely adaptational state, background intensity, stimulus intensity, and the subtended visual angle of the source. In this study, all factors, with the exception of the stimulus intensity, were equal for all the species: they were all completely dark-adapted before being presented with a flickering light stimulus, the background intensity was 0. and the eye was always bathed with a circle of light that was larger than the eye itself. The only variable factor was the stimulus intensity. Although the irradiance of the light source was calibrated, the response to the light stimulus depended on the position of the electrode in the eye, so that a 100-|U,V response might be generated by a dim stimulus in one specimen and a brighter stimulus in another specimen of the same species. As shown in Figure 1, the CFF de- pended on irradiance level, with lower irradiances evoking lower CFFs and higher irradiances evoking higher CFFs. Therefore, to ensure that the same parameter was measured for all species in this comparative study, the maximum CFF. which is the highest flicker rate that the eye is capable of following at any intensity, was used. Maximum critical fusion frequency The maximum CFFs were measured for eight species of mesopelagic crustaceans from a variety of depths (Table I). Oplophorus qracilirostris ^ 35-r o c CD 30- A ** * T CT 2? 25- LL 20 1 C 9 o 'w 15- LL 10 * CD g 5 - LTJ - 7 8 9 10 11 12 13 Log Irradiance Figure 1. Flicker fusion frequency (Hz) as a function of irradiance (photons cm~- s~ ' ) for Oplophorus graciliroslris. Flicker fusion frequency increases as the irradiance increases, up to the maximum critical flicker fusion frequency (arrow), the point at which further increases in irradiance do not result in a more rapid flicker fusion frequency. 140 T. M. FRANK Three species, Sergia filictum, Systellaspis debilis, and Fun- chciliu rillosa. had relatively low maximum CFFs, between 20 and 25 Hz (Table II), as one would expect from species coming from a very low light environment. A representative example of an ERG is shown for Systellaspis debilis in Figure 2A. Three species, Oplophoms gracilirostris, Juni- cella spinacauda, and Stylocheiron maximum, had some- what higher rates, between 31 and 36 Hz; and two species, Nematobrachion flexipes and Nematobrachion sexspinosus, had extremely high CFFs. considering their dim light envi- ronment, of 44 and 57 respectively (Table II). A represen- tative example of an ERG for N. sexspinosus is shown in Figure 2B. Sensitivity The overall sensitivity of the eye was estimated by de- termining the log of the irradiance (log I,,,) required to produce a response that was 10% of the amplitude of the maximum response the eye was capable of generating. In several preparations, the maximum response was not seen, and the log 1,,, could not be determined. As shown in Table II, there is a trend towards lower sensitivity to light as the response dynamics of the eye speeds up. Systellaspis debi- lis, the species with the lowest maximum CFF. had, accord- ing to the log l lo , the most sensitive eye, while N. sexspi- nosus, the species with the highest maximum CFF, has the lowest sensitivity to light. Response latency As an indicator of the speed of transduction in the pho- toreceptors of the various species, latency from the start of the light stimulus to the start of the photoreceptor response was measured, using a response amplitude that was 10% of Table II Mean values for temporal resolution (max CFF). sensitivitv (Log l lu ) ami response latencies obtained from ERG data Log I 10 Latency Species Max CFF IH/J ( photons cm " - s ~ ' I (ms) Systellaspis debilix 2 1 ( 0.6; n = 4) 8.9 ( 0.06) 58 ( 6.9) Funchalia w7/,.w. 24 (0.5; n = 2) 9. 3 (0.0) 75 (2.0) Sergia filictum 15 (n = 1) NA NA Janicella spinacat,. (0.3; n = 3) 9.4 (0.06) 49 (3. 5) Oplophoms gracilirostris '. n = 2) 9.4 (0.09) 42 (5. 5) Stylocheiron maximum 36 (n 10.0 42 Nematobrachion flexipes 44 ( I (). /, >) 10.2 ( 0.38) 22 ( 0.5) Nematobrachion sexspinosus 56 (2.0; n = 4) 10.8 (0.18; n = 2) 16(2.5) Species are ranked by max CFF. from lowest to highest. Standard errors and number of specimens tested are in parenthesis. Systellaspis debilis 20 22 H Z B ERG S ERG S ERG S Nematoscelis sexspinosus AAMlWimilMAJMlMIWlJl 63 Hz 65 Hz Figure 2. Representative examples of species with low and high flicker fusion frequencies. ERG designates the response recorded from the eye; S designates the flickering light stimulus. The data shown are from the last 0.4 s of the 2-s stimulus pulse. (A) The ERG from Systel/aspis debilis was able to follow the stimulus light at 20 Hz cycle for cycle; at 22 Hz. the ERG response was lagging behind the light stimulus and "missing" cycles. ( B ) The ERG from Nematoscelis sexspinosus was able to follow the stimulus light at 60 Hz. At 63 Hz. the ERG appears to be in phase with the stimulus light, but careful examination of the data shows that 25 flashes of light were given, but only 23 responses were produced. By 65 Hz, the lag is even greater, and the ERG is clearly "missing" cycles. The CFF of this specimen is 60 Hz. since this is the last recorded frequency at which the ERG was able to match, cycle for cycle, the phase of the stimulus light. the maximum amplitude. Species with lower flicker fusion frequencies also have longer latencies, indicative of slower eyes, and species with higher flicker fusion frequencies have eyes with much faster response dynamics (Table II). Discussion The vertical distributions of the species examined in this study have not been determined for the south coast of Cuba. The vertical distribution data in Table I are for the Gulf of Mexico, except for the euphausiids. Since Gulf of Mexico water originates in the Caribbean Sea (Nowlin, 1971), and both areas have Jerlov's Type 1 or 1 A water (Jerlov. 1976). it is likely that these two areas would have similar species assemblages and distribution patterns. The abundance of the TEMPORAL RESOLUTION IN CRUSTACEAN VISION 141 three species ofeuphausiids in this study was extremely low in the Gulf of Mexico (Kinsey and Hopkins, 1994), as it was off the coast of Cuba (pers. obs.). and comprehensive data on depth distribution are not available for these areas. The data presented in the table are for the tropical Pacific, where Jerlov's Type 1 or 1A water is also present. The depth distributions of nine relatively abundant euphausiid species in the Gulf of Mexico (Kinsey and Hopkins. 1994) were compared with those provided by Roger (1978) for the same species in the tropical Pacific, and the depth ranges proved to be the same in the two areas. Therefore, it is likely that the data presented for the vertical distribution of Nemato- brachion flexipes, N. se.\spinosus, and Stylocheiron maxi- mum in the tropical Pacific would also apply to the Gulf of Mexico. Depth vs. critical flicker fusion Shallower depth ranges do not necessarily mean a brighter light environment: an organism found at 300 m in murks water might see significantly less light than an or- ganism found at 500 m in very transparent water. However, in this study, the depth of occurrence is an indication of relative light intensity, as the water is all Jerlov's Type ! or 1A (Jerlov. 1976). Most of the species in this study live below 400 m. At 400 m during the day. in Jerlov's type 1 water, downwelling ambient light has been reduced to less than 0.0001% of the surface irradiance (Jerlov, 1976). This is about as bright as the moonlight seen by nocturnal insects, in that light from a full moon is 0.0001% of daytime illumination (Munz and McFarland, 1973; Land. 1981). Therefore, one would expect mesopelagic crustaceans to have relatively low maximum CFFs. indicative of low tem- poral resolution, as has been found in nocturnal insects (Autrum. 1950, 1958, 1984; Howard et al.. 1984: de Souza and Ventura. 1989; Laughlin and Weckstrom, 1993). In addition, one might expect the deepest dwelling species, which live in the dimmest light, to have the lowest maxi- mum CFFs. This is certainly the case for the two deepest living species in this study an oplophorid, Systellaspis debilis, and a sergestid. Sergio filictum which have max- imum CFFs of 20-25 Hz (Fig. 3). equivalent to those of the nocturnal slow moving insects studied by Autrum (1950. 1958). Oplophorus gracilirostris and Janicella spinacauda are in the same family as S. debilis. but have higher maxi- mum CFFs between 31-32 Hz. Their daytime depth range is about 100 m shallower than that of S. debilis (Fig. 3). so a higher maximum CFF is not unexpected. A previous study by Moeller and Case (1995) reports a critical flicker fusion frequency of 12 Hz for Oplophorus spinosus, which has a depth distribution similar to that of O. gracilirostris. How- ever, those authors were using a light that was only 1 log unit above the irradiance that produced a threshold re- sponse, and as shown by Figure 1. a much lower critical Maximum CFF vs. Depth 60 50 - 40 - 30 - 20 10 - N. sexspinosus ' N. flexipes S maximum O. gracilirostris J spinacauda F. villosa S. filictum S. debilis 200 300 400 500 600 700 Daytime Depth (m) 800 900 1000 Figure 3. Maximum critical flicker fusion frequency (CFF) as a func- tion of daytime depth distribution for the eight species in this study. flicker fusion frequency would result from such a dim light stimulus. At first glance, the very low maximum CFF of the penaeid. Funclmlia villosa, is somewhat puzzling. It is one of the shallowest living species, and also possesses an eye with a fairly high spatial resolution (Herring and Roe. 1988). High spatial resolution is usually correlated with a higher light environment, and is also associated with a comparatively higher temporal resolution (Srinivasan and Bernard, 1975). However, these measurements of maximum CFF were made in a completely dark-adapted eye; the eye of F. villosa, which is of the superposition type, possesses migrating screening pigments, which changes the eye from a spatially acute, apposition-like eye during the day to an eye with less spatial resolution, but greater sensitivity, at night (Herring and Roe. 1988). Maximum CFF is known to be higher in the light-adapted vs. the dark-adapted eyes of shallow-water crustaceans that possess mobile screening pigments (Crozier and Wolf. 1939; Crozier et al.. 1939: B rocker, 1935). but mobile screening pigments are usually not found in mesopelagic species (see Hallberg and Elofs- son. 1989. for review). F. villosa appears to be an exception to this rule. Looking at the other species in this study, the 142 T. M. FRANK screening pigments in the photoreceptors of Oploplwrus spinosus (Welsh and Chace. 1937; Land, 1976; Gaten et ai, 1992). a close relative of O. gracilirostris with the same depth distribution, and Systellaspis debilis (Gaten el ai, 1992) do not appear to be mobile. Although some species of sergestids may possess screening pigments (Welsh and Chace, 1938), it is not known whether these are mobile. Chun (1896). Zimmer (1956), and Meyer-Rochow and Walsh (1978) found no evidence of the migration of screen- ing pigments in euphausiid eyes, and while Kampa (1965) indicates that some migration does occur. Land et al. ( 1979) conclude that this migration is not of sufficient magnitude to affect the spatial resolution of the eye. It remains to be seen whether F. villosa, whose spatial resolution is clearly higher when its screening pigments are in the light-adapted posi- tion, will also demonstrate a higher temporal resolution under light adaptation that is more consistent with its day- time depth distribution. Future studies will include deter- mining the temporal resolution of F. villosa, as well as that of other species without migrating screening pigments, un- der light-adapted conditions. The three euphausiid species in this study (Nematobra- chion sexspinosus, N.flexipes, and Stylocheiron maximum) all have bilobed eyes, and Chun (1896) determined (mor- phologically) that the upper lobe, which is oriented upwards toward the brighter downwelling light, has higher spatial resolution than the lower lobe, which is oriented downwards towards dimmer upwelling light. Due to limitations set by working on shipboard, the responses from receptor cells in the upper lobe could not be isolated from responses from receptor cells in the lower lobe. Therefore, it remains to be determined whether the temporal resolution differs between the two lobes. However, it is clear that two of these species. N. sexspinosus and N. flexipes, have the highest maximum CFFs of all the species in this study, comparable to CFFs reported for shallow-water crabs (Brocker, 1935; Crozier and Wolf. 1939). This is unexpected, because their daytime depths of occurrence are in the middle of the depth distri- bution for the eight species in the study (Fig. 3). These two species live at a depth at which the downwelling light intensity is roughly the same as that experienced by noc- turnal insects under a full moon (see above), yet they have a substantially higher maximum CFF (40-60 Hz) than most nocturnal insects (10-20 Hz; Autrum. 1950, 1958). Laugh- lin and Weckstrom ( 1993) demonstrated that the benefits of high temporal res; ion are limited for nocturnal, generally slow moving insects, and that the metabolic price for im- proving the temporii! hh of vision is substantial. In addition to the metabolic < ., ''ise, fast photoreceptors have a lower sensitivity than slow photoreceptors (Laughlin, 1990). which would be a distuu Disadvantage to organisms living in a light-limited environment. However, these con- clusions were drawn for terrestrial organisms, and other factors must be taken into account in the oceanic realm. Critical flicker fusion and bioluminescence Although the crustaceans in this study share the same light regime as nocturnal insects with respect to background illumination, many of their prey are bioluminescent. Biolu- minescence, which is very rare in the terrestrial environ- ment, is an extremely common phenomenon in the oceanic realm. In the mesopelagic zone (200-900 m), luminescence has been found in up to 75% of the fish species (Herring and Morin, 1978) and 79% of the shrimp species (Herring, 1976). Nocturnal terrestrial insects must image dark objects against a dimly lit background, so a higher temporal reso- lution, with the resulting decrease in contrast sensitivity, would be a considerable disadvantage. In the ocean, it might be advantageous for predatory carnivorous species to sac- rifice sensitivity for the ability to more accurately track a glowing or flashing prey item. All euphausiid species with bilobed eyes possess an ex- tremely elongated second or third thoracic leg, some with clawlike chelae at the end, which is hypothesized to be an adaptation for active carnivorous feeding (Mauchline and Fisher, 1969). If some of these species specialized in cap- turing bioluminescent prey, the greater contrast between a bioluminescent prey item against a dim background vs. a dark prey item against a dim background would make it advantageous for these species to sacrifice sensitivity (and hence contrast detection) in return for greater temporal resolution (and hence tracking ability). This rationale would likewise explain the puzzling result that Stylocheiron max- imum, which is also a bilobed euphausiid with an elongated thoracic appendage, has the shallowest depth distribution of the three euphausiid species but also the lowest maximum CFF (Fig. 3). S. maximum eats primarily copepods in the genera Oithona. which is nonluminescent, and Oncaea, of which only one species is known to be bioluminescent (see Herring, 1985, for review), and this bioluminescent species is not present in the Gulf of Mexico (Kinsey and Hopkins, 1994). Since the biomass and species distribution off Cuba is similar to that of the Gulf of Mexico, it is likely that 5. maximum is eating primarily nonbioluminescent prey in Cuban waters as well. On the other hand, the primary prey item of N. sexspinosus and N. flexipes. the two species with the highest critical flicker fusion frequencies, is an active bioluminescent copepod called Pleuromamma (Hu. 1978; Kinsey and Hopkins, 1994). all species of which emit some form of bioluminescent spew (see Herring, 1985. for re- view). To further support the argument that this visual adaptation is driven by bioluminescence, the Nematobra- chion species, as mentioned above, have a deeper depth range than S. maximum, but possess a less sensitive eye, according to the log l l() values (Table II). Following Au- trum' s hypothesis, one would predict that the organism from the dimmer light regime would have the more sensi- tive eye with slower response dynamics (assuming similar TEMPORAL RESOLUTION IN CRUSTACEAN VISION 143 activity levels). However, because the Nematobrachion spe- cies are specializing in bioluminescent prey, the advantages of a higher temporal resolution might outweigh the advan- tages of a more sensitive eye. Other species, however, might benefit from having an eye with lower temporal resolution. If the preferred biolumines- cent prey were a slow moving item that glowed, such as some species of gelatinous zooplankton, or marine snow colonized by bioluminescent bacteria, an eye with a lower temporal resolution assuming this meant a longer integra- tion time (see below) would actually be advantageous. This benefit would only apply to a dim "slow" signal, such as a glow or a flash with a slow rise time; a brief dim flash with a rapid rise time would be equally difficult to detect by either slow or fast photoreceptors (for a comprehensive discussion of frequency coding, see Laughlin, 1981; Laugh- lin and Weckstrom, 1993). As stated above, the advantage of possessing an eye with a lower flicker fusion frequency depends on the assumption that a lower CFF is correlated with a longer integration time. The integration times of the eyes of the species in this study have not been measured, but de Souza and Ventura (1989) found that critical duration, another temporal characteristic of a photoreceptor that is determined electrophysiologically, is directly related to integration time, so that a long critical duration indicates a long integration time. Since the maxi- mum CFF can be equated to the reciprocal of the critical duration (Matin. 1968), the low CFFs of most of the crus- taceans in this study (the Nematobrachion species being the exception) indicate that they possess photoreceptors with fairly long integration times, and therefore might be well adapted for detecting dim, glowing bioluminescence. The conclusion that eyes with lower CFFs have slower response dynamics is supported by the latency data, in that the eyes with lower maximum CFFs also have longer response la- tencies (Table II). In conclusion, it appears that while the mesopelagic light environment is similar to that of nocturnal insects with respect to background light, the temporal resolutions of several species found in this environment are substantially higher than would have been predicted on the basis of background light alone. In addition, the hypothesis that temporal resolution would be correlated with daytime depth distribution is not supported by these data. However, this preliminary study indicates that when bioluminescence is taken into account, Autrum's hypothesis that the response dynamics of the retina match the habitat and lifestyle of the organism appears to be valid in the oceanic realm as well as in the terrestrial environment. Acknowledgments I thank Discovery Channel for allowing me to participate on their expedition off the coast of Cuba, and the captain and crew of the RV Seward Johnson for their great assis- tance with animal collection. I also thank Dr. Simon Laugh- lin for his valuable advice, and Dr. Edith Widder, Dr. Sonke Johnsen. and Ms. Erin Fisher for their helpful comments on the manuscript. Additional funding for this work was pro- vided by NSF Grant No. #OCE-93 13972 to T. M. Frank and E. A. Widder. Harbor Branch Oceanographic Institution Contribution #1270. Literature Cited Autrum, H. 1950. Die Beliditungspotentiale und das Sehen der Insekten (Untersuchungen an Calliphora und Dixippus). Z. Vergleich. Physiol. 32: 176-227. Autrum, H. 1958. Electrophysiological analysis of the visual systems in insects. Exp. Cell Res. 5: 426-439. Aulrum, H. 1984. Comparative physiology of invertebrates: Hearing and vision. Pages 1-19 in Foundations of Sensory Science. Vol 1, W. W. Dawson and J. M. Enoch, eds. Springer- Verlag, Berlin. Autrum, H., and M. Stocker. 1952. Uber optische Verschmelzungsfre- quenzen und stroboskopisches Sehen bei Insekten. Biol. Zentr. 71: 129-152. Brocker, H. 1935. Untersuchungen liber das Sehvermogen der Ein- siedlerkrebse. Zoo/. Jahrb. Abt. Allgem. Zool. Physiol. Tiere 55: 399- 430. Chun, C. 1896. Atlantis: Biologische Studien iiber pelagische Organis- men. Zoologica 7: 1-260. Cronin, T. W. 1986. Photoreception in marine invertebrates. Am. Zool. 26: 403-415. Crozier. W. J., and E. Wolf. 1939. The flicker response contour for the crayfish. J. Gen. Physiol 23: 1-10. Crozier. W. J., E. Wolf, and G. Zerrahn-Wolf. 1939. The flicker response contour for the isopod Asellus. J. Gen. Physiol. 22: 451-462. de Souza, J. M., and D. F. Ventura. 1989. Comparative study of temporal summation and response form in hymenopteran photorecep- tors. J. Comp. Physiol. 165: 237-245. Flock, M. E., and T. L. Hopkins. 1992. Species composition, vertical distribution, and food habits of the sergestid shrimp assemblage in the eastern Gulf of Mexico. J. Crust. Biol 12: 210-223. Frank, T. M., and J. F. Case. 1988a. Visual spectral sensitivities of bioluminescent deep-sea crustaceans. Biol. Bull. 175: 261-273. Frank, T. M., and J. F. Case. 1988b. Visual spectral sensitivities of the bioluminescent deep-sea mysid. Gnathophausia ingens. Biol. Bull. 175: 274-284. Gaten, E., P. M. J. Shelton, and P. J. Herring. 1992. Regional mor- phological variations in the compound eyes of certain mesopelagic shrimps in relation to their habitat. J. Mar. Biol. Ass. UK 72: 61-75. Hallherg, E., and R. Elofsson. 1989. Construction of the pigment shelf of the crustacean compound eye: A review. J. Crust. Biol. 9: 359-372. Herring, P. J. 1976. Bioluminescence in decapod Crustacea. J. Mar. Biol. As.v. t/A'56: 1029-1047. Herring, P. J. 1985. Bioluminescence in the Crustacea. J. Crust. Biol. 5: 557-573. Herring, P. J., and J. G. Morin. 1978. Bioluminescence in fishes. Pp. 273-239 in Bioluminescence in Action, P. J. Herring, ed. Academic Press, New York. Herring, P. J., and H. S. J. Roe. 1988. The photoecology of pelagic oceanic decapods. Symp. Zool Soc. Land. 59: 263-290. Hopkins, T. L., J. V. Gartner, and M. E. Flock. 1989. The caridean shrimp (Decapoda: Natantia) assemblage in the mesopelagic zone of the eastern Gulf of Mexico. Bull Mar. Sci. 45: 1-14. 144 T. M. FRANK Hopkins, T. L., M. E. Flock, J. V. Gartner, Jr.. and J. J. Torres. 1994. Structure and trophic ecology of a low latitude midwater decapod and mysid assemblage. Mar. Ecol. Prog. Ser. 109: 143-156. Howard, J., A. Dubs, and R. Payne. 1984. The dynamics of photo- transduction in insects. A comparative study. J. Comp. Ph\sil. 154A: 707-718. Hu, V. J. H. 1978. Relationships between vertical migration and diet in four species of euphausiids. Limnol. Oceanog. 23: 296-306. Jerlov, N. G. 1976. Marine Optics. Elsevier. Amsterdam. Kampa, E. M. 1965. The euphausiid eye a re-evaluation. Vision Res. 5: 475-481. Kinsey, S. T., and T. L. Hopkins. 1994. Trophic strategies of euphau- siids in a low-latitude system. Mar. Biol. 118: 651-661. Kugel. M. 1977. The time course of the electroretinogram of compound eyes in insects and its dependence on special recording conditions. J. Exp. Biol. 71: 1-6. Land, M. F. 1976. Superposition images are formed by reflection in the eyes of some oceanic decapod Crustacea. Nature 263: 764-765. Land. M. F. 1981. Optics and vision in invertebrates. Pp. 471-592 in Handbook of Sensory Physiology. Vol VII/6B. H. Autrum. ed. Springer- Verlag. Berlin. Land, M. F. 1990. Optics of the eyes of marine animals. Pp. 149-166 in Light and Life in the Sea, P. J. Herring. A. K. Campbell. M. Whittield. and M. Maddock, eds. Cambridge Univ. Press. Cambridge. Land, M. F., F. A. Burton, and V. B. Meyer-Rochow. 1979. The optical geometry ot euphausiid eyes. J. Comp. Physiol. 130: 49-62. Laughlin. S. B. 1981. Neural principles in the peripheral visual systems of invertebrates. Pages 133-280 in Handbook of Sensory Physiology. Vol. VI1/6B. H. Autrum. ed. Springer- Verlag. Berlin. Laughlin, S. B. 1990. Invertebrate vision at low luminances. Pp. 223- 250 in Night Vision. R. F. Hess. L. T. Sharpe. and K. Nordby. eds. Cambridge Univ. Press, Cambridge. Laughlin, S. B.. and M. Weckstrom. 1993. Fast and slow photorecep- tors a comparative study of the functional diversity of coding and conductances in the Diptera. J. Comp. Physiol. 172: 593-609. Loevt, E. R. 1976. Light and photoreceptor degeneration in the Norway lobster. Nephrops non'egiciis (L). Proc. Roy. Soc. Loud. 193B: 31-44. Matin, L. 1968. Critical duration, the differential luminance threshold. critical flicker frequency and visual adaptation: a theoretical treatment. J. Opt. Soc. Am. 58: 404-415. Mauchline, J.. and L. R. Fisher. 1969. The biology of euphausiids. Pages 1-454 in Advances in Marine Biology. Vol. 7. R. S. Russel and M. Yonge. eds. Academic Press. London. Meyer-Rochow, V. B., and S. Walsh. 1978. The eyes of mesopelagic crustaceans. III. Thysanopoda tricuspidata (Euphausiacea). Cell Tiss. Res. 195: 59-79. Moeller, J. F., and J. F. Case. 1995. Temporal adaptation in visual systems of deep-sea crustaceans. Mar. Biol. 123: 47-54. Munz. F. W., and W. N. McFarland. 1973. The significance of spectral position in the rhodopsins of tropical marine fishes. Vision Res. 13: 1829-1874. Nilsson, H. L., and M. Lindstrom. 1983. Retinal damage and sensitivity loss of a light-sensitive crustacean compound eye (Cirolanu borealis); electron microscopy and electrophysiology. J. E\p. Biol. 107: 277-292. Nowlin, \V. D. 1971. Water masses and general circulation of the Gulf of Mexico. Oceanol. Int. 6: 28-33. Roger, C. 1978. Bioecological Sheets on Tropical Pacific Eiiphansiiils. Orstom. Paris. Srinivasan, M. V., and G. D. Bernard. 1975. The effect of motion on visual acuity of the compound eye: A theoretical analysis. Vision Res. 15: 515-525. Weckstrom. M., and S. B. Laughlin. 1995. Visual ecology and voltage- gated ion channels in insect photoreceptors. Trends Neiimsci. 18: 17-21. Welsh. J. H., and F. A. Chace, Jr. 1937. Eyes of deep sea crustaceans. I. Acanthephyridae. Biol. Bull. 72: 57-74. Welsh, J. H., and F. A. Chace, Jr. 1938. Eyes of deep sea crustaceans. II. Sergestidae. Biol. Bull. 74: 364-375. Zimmer. C. 1956. Euphausiacea. Pages 1-286 in Bronns Klassen des Tierreichs. Bd. 5, H. E. Gruner, ed. Geest and Portig, Leipzig. Reference: Biol. Bull. 196: 145-152. (April 1999) Muscle Activity in Steady Swimming Scup, Stenotomus chrysops, Varies With Fiber Type and Body Position DAVID J. COUGHLIN 1 AND LAWRENCE C. ROME 2 ^Science Division, Widener University. One University Place, Chester, Pennsylvania 19013: and 2 University- of Pennsylvania, Department of Biology, Leidy Laboratory. Philadelphia, Pennsylvania, 19104. and Coastal Research Laboratory. Woods Hole Oceanographic Institution. Woods Hole. Massachusetts 02543 Abstract. The red and pink aerobic muscle fibers are used to power steady swimming in fishes. We examined red and pink muscle recruitment and function during swimming in scup, Stenotomus chrysops, through electromyography and high-speed cine. Computer analysis of electromyograms (EMGs) allowed determination of initial speed of muscle recruitment and duty cycle and phase of muscle electromyo- graphic activity for both fiber types. This analysis was carried out for three longitudinal positions over a range of swimming speeds. Fiber type and longitudinal position both affected swimming speed of initial recruitment. Posterior muscle is recruited at the lowest swimming speed, whereas more anterior muscle is not initially recruited until higher speeds. At more anterior positions, the initial recruitment of pink muscle occurs at a higher swimming speed than the recruitment of red muscle. The duty cycle of pink muscle EMG activity is significantly shorter than that of red muscle, reflecting a difference in the onset time of activation during each cycle of length change: pink muscle onset time follows that of red. The different patterns of usage of red and pink muscle reflect differences in their contraction kinetics. Be- cause pink muscle generates force more rapidly than red muscle, it can be activated later in each tailbeat cycle. Pink muscle is used to augment red muscle power production at higher swimming speeds, allowing a higher aerobically based steady swimming speed than that possible by red muscle alone. Received 26 August 1998; accepted 22 January 1999. Email: coughlin(s> pop I. science. widener.edu Introduction Steady swimming in teleosts is powered by aerobic mus- cle fibers (Rome et ai. 1984; Bone. 1989), including the slow-twitch red muscle and the intermediate-twitch pink muscle (Coughlin and Rome. 1996). These muscle fibers are arranged in relatively thin longitudinal bands, with the pink muscle medial to the red muscle. During steady swimming, red muscle is used at the lowest swimming speeds (Rome et ai. 1984). Pink muscle reportedly is not recruited until intermediate swimming speeds (Johnston et ai, 1977). At the maximum steady swimming speed, both red and pink muscle are recruited (Coughlin and Rome, 1996). Muscle function and power production in a steadily swimming fish have recently been detailed for scup (Rome et ai, 1993; Coughlin and Rome, 1996: Coughlin et ai. 1996). These studies have employed the workloop tech- nique (Josephson, 1985) to examine power production by isolated muscle bundles activated using patterns of muscle length change and muscle stimulation that had been re- corded in vivo from swimming fish. At the maximum steady swimming speed (the highest speed before white muscle recruitment), scup generate most of the power for swim- ming by using the aerobic muscle fibers of the posterior myotomes (Rome et ai, 1993; Coughlin and Rome, 1996). Coughlin el ai ( 1996) found several differences between the contraction kinetics of red and pink muscle. First, pink muscle has faster kinetics than red muscle. In isometric contractions, the rates of activation and relaxation of pink muscle are about twice those of red muscle. These rates will influence power production by each muscle type, and there- fore, they may affect how the muscle is recruited and activated in a swimming fish. During oscillatory activity such as swimming, muscle must alternately be activated and 145 146 D. J. COUGHLIN AND L. C. ROME relax during each tailbeat cycle. The muscle is activated to generate force during shortening and must relax at the end of shortening to minimize the "negative" work done on the muscle to lengthen it. For most fish muscle, activation occurs during the lengthening phase, giving the muscle time to generate force so that force peaks as shortening begins. Because pink muscle activates more quickly than red mus- cle, the onset of pink muscle stimulation can lag behind that of red muscle during each cycle of length change, as we have previously found at maximal swimming speeds (Coughlin and Rome, 1996). This appeared to help mini- mize the negative work that is due to a high force level at the end of lengthening. Finally, because the offset of red and pink muscle electromyographic activity during each cycle of length change occurs at the same time, this also resulted in the EMG duty cycle of pink being less than that of red muscle at the highest aerobic swimming speed. In this study, we analyze whether these patterns of recruitment occur at submaximal steady swimming speeds as well. Under optimized conditions of oscillatory activity and at oscillation frequencies greater than 4 Hz, pink muscle pro- duces more mass-specific power than red muscle. However, at low frequencies of oscillation, isolated scup pink muscle shows a marked reduction in power production compared to red muscle. In workloop experiments at oscillation frequen- cies below 4 Hz. force levels drop rapidly in pink muscle during the shortening phase of each length change. The mechanistic reasons for this drop in force during shortening at low frequencies are not known (Coughlin et til., 1996), but this observation leads to a prediction that pink muscle should not be recruited at low swimming speeds that cor- respond to relatively low tailbeat frequencies. In this study, we used eleetromyography to record the activity of red, pink, and white muscle in swimming scup over a range of steady swimming speeds. We examined the effects of swimming speed, longitudinal position on the fish, and fiber type on muscle electromyographic activity param- eters such as swimming speed of initial recruitment, relative phase of red and pink muscle activation, and duration of electromyographic activity. Also, for a limited number of swimming speeds, the phase of EMG relative to the cycle of length change could be determined. Data collected permit an analysis of how pink and red muscle are used during steady swimming in scup over a range of speeds. ; ials and Methods Scup, collected in . , i 994 at Cape Cod, Massachusetts (;i = : 12. length SI. 21.8 1.4 cm), swam in a recirculating water treadmill (I >me ct til., 1990). Electro- myograms (EMGs) were recorded from fish swimming at speeds ranging from 1.5 to 3 body lengths (BL) per second at 10C (-30 to 60 cm s~') and from 2.? to 4.25 BL per second at 20C (-50-85 cm s ' ). These are the ranges of steady, aerobic swimming speeds exhibited by these fish. For each temperature, the fish would not reliably swim at lower speeds, and swimming at higher speeds was "un- steady." indicating anaerobic activity due to the recruitment of white muscle. For each fish, EMGs were recorded simultaneously from red and pink muscle at one of three longitudinal positions along the fish: the ANT-1, ANT-2, and MID positions, which are defined as 28%, 40%, and 55% of the total length from the anterior tip or snout (Rome et til.. 1993). Record- ings were not made in the POST position (70%), because this position had no distinct layer of pink muscle (Zhang et til., 1996). When possible, recordings were made from two positions simultaneously (Table I). In addition, white mus- cle recordings were made at the ANT-2 position in all fish. The recording technique has been previously detailed (Rome et /., 1990, 1992). Fish were anesthetized with tricaine methanosulfate ( MS-222 ) at a dosage of 50 mg 1 ~ ' . They were maintained during the surgical procedure with regulated respiratory current pumped across their gills. A hypodermic needle was used to insert twisted wire (Teflon- coated Medwire) bipolar electrodes into the muscle. All wires were sutured at their point of entry and collectively near the back of the dorsal fin. Fish recovered quickly from anesthesia when returned to the holding tank. Fish were swum 24 h after surgery at either 10 or 20C. Fish swum at two temperatures were allowed to adjust to a change of temperature for 48 h. Grass amplifiers filtered the electro- myographic signal with a bandwidth of 10 to 3000 Hz and a 60-Hz notch filter. The placement of electrodes into either Table I S/>('fi/c;i.v and conditions used for electromyographic recordings of the wimming musculature of scup Fish nuiiituM Total length (cm) Aerobic tiher position Temperature (C) 2 23.1 ANT-1 20 4 21.0 ANT-2, MID 20 7 22.2 ANT-2. MID 20 8 23.2 ANT-2 20 9 20.3 ANT-1 20 10 22.0 ANT-1, ANT-2 20 1 1 21.1 ANT-1, MID 20 19 23.7 ANT-2. MID 10,20 20 23.6 ANT-2. MID 10,20 22 19.7 ANT-1, ANT-2 10.20 24 20.5 ANT-2. MID Id 25 21.0 ANT-1. ANT-2 II) For all fish, white muscle activity was recorded at ANT-2. For each fish, the aerobic muscle fibers (red and pink) were recorded at one or two positions tor one or two temperatures. For each position and fiber type, electromyographic activity was analyzed for a minimum of lour tailbeat cycles. PINK AND RED MUSCLE ACTIVITY IN SCUP 147 the thin pink layer or the overlying red muscle was verified through dissection after each experiment. Also, the nature of the signal indicated that cross-talk between red and pink muscle recordings was not occurring. The pink and red muscle EMG waveforms at one position were not the same. The EMGs from swimming fish were analyzed using custom macros in DATAPAC software. The semi-auto- mated analysis of the EMG computer files permitted char- acterization of bursts of electromyographic activity. For each muscle fiber type at each longitudinal position, several variables relating to muscle activity were measured at each swimming speed. First, swimming speed of initial recruit- ment, or the minimum speed at which bursts of activity could be detected, was determined for each fiber type at each position. Bursts were identified by an algorithm that examined the first derivative at each point in a rectified EMG trace. For each rectified trace, a threshold was set to distinguish the slowly varying background noise from the spikes of electromyographic activity. Spikes were identified as points in the trace that exceeded this threshold, and each burst was then identified as a string of consecutive spikes. Burst duration, or the length of each electromyographic burst, was expressed as duty cycle, the duration of muscle activity as a proportion of the period of tailbeat oscillation. The relative timing of the onset of activity in red and pink muscle in each tailbeat cycle was also analyzed. The phase difference between the times of activity onset in red and pink muscle was determined for each fish at each longitu- dinal position for each swimming speed. Phase difference was expressed as a proportion of the oscillation period. Positive values indicate that the activity of red muscle occurred before that of pink muscle; negative values indi- cate that pink muscle activity had an earlier onset. For swimming at 2.5 and 4.0 BL s~' at 20C, patterns of muscle length change were also determined. The fish were filmed from above with high-speed cine. The films were synchronized with the EMG traces (Rome. 1995). Body curvature and the associated muscle length change charac- teristics, including the amplitude and frequency of muscle oscillation, were then determined from the films (Rome et ai. 1992. 1993; Coughlin and Rome, 1996). For these swimming bouts, the relative phase of the electromyo- graphic activity with respect to the muscle length change was calculated for both fiber types. Phase was defined as the timing of the electrical activity with respect to the beginning of muscle shortening (maximum length) and was deter- mined separately for red and pink muscle as previously reported (Coughlin and Rome, 1996). The phase was ex- pressed as a proportion of the period of tailbeat oscillation and is similar to the On-/3 shift reported by Jayne and Lauder(1995, 1996). Results As swimming speed increased, so did tailbeat frequency (Fig. 1). For swimming at both 10 and 20C, the increase of tailbeat frequency with swimming speed is well fit by a linear regression for the range of steady swimming speeds. The swimming speed at which initial muscle recruitment occurred varied significantly with both muscle fiber type and longitudinal position (Table II). At 20C, muscle from more posterior positions was recruited at a lower swimming speed than muscle from anterior positions. At the MID position, red muscle and pink muscle were initially re- cruited at the lowest steady swimming speed, 2.5 BL s~' (Fig. 2). At the ANT-2 position, red muscle was also re- cruited at the lowest steady swimming speed, but pink muscle was not recruited until at least 3.0 BL s~', and not until 4.0 BL s~' in some fish (Fig. 2). At the ANT-1 position, red muscle was initially recruited at 2.5-3.0 BL s , whereas pink muscle was not initially recruited until an average of almost 4 BL s~'. Similar trends hold for 10C. Both red muscle and pink muscle at the MID position were initially recruited at the lowest steady swimming speed, 1.5 BL s" ' (Fig. 3). At the ANT-2 position red muscle was also recruited at this speed, but the swimming speed of initial recruitment for pink muscle was significantly faster (1.9 BL s~',t = 3.67, P = 0.021 withdf = 4. Fig. 3). At the ANT-1 position, sample size was limited (n = 2). In these fish, red 6 - 5 - => CT 0) tl 4 CD CD .0 3 - 10C 20C Linear Regression 2 3 Swimming Speed (BL s~' Figure 1. Tailbeat frequency as a function of swimming speed at two temperatures. The regression equations are as follows: tailbeat frequency = 0.30 + 1.45 (swim speed), r = 0.99 and P = 0.012 at 10C; and tailbeat frequency = 1.679 + 0.938 (swim speed), r = 0.97 and P < 0.001 at 20C. 148 D. J. COUGHLIN AND L. C. ROME Table II Swimming speed I in body lengths per second SE) at which aerobic muscle in swimming scup is initially recruited varies with both fiber Type tinil longitudinal position Longitudinal Position Fiber ANT-1 ANT-2 MID POST type (n = 5) ( = 7) (n = 5) (n = 5) Pink 3.8 0.1 3.3 0.2 2.6 0.1 _ Red 2.7 0.1 2.5 0.0 2.5 0.0 2.5 0.0* Initial recruitment speeds were determined at 20C through automated analysis of EMG bursts: the recruitment speed for a given position and fiber type is the minimum speed at which bursts were detected. Sample size (n) is given for each position. Longitudinal position (F = 15.4. P < 0.001 ) and fiber type (F = 48.9, P < 0.001) both had an effect on recruitment speed. There was also an interactive effect (longitudinal position x fiber type; F = 8.3. P = 0.001). * There is no identifiable pink muscle layer at the POST position, just a scattered distribution of pink muscle fibers within the red muscle layer (Zhang et al.. 1996). The value given for red muscle represents unpub- lished data. muscle was initially recruited at 2.0 BL s ', and pink muscle was not recruited until 2.25 BL s~'. White muscle recruitment (at the ANT-2 position) was observed in all fish at a speed of 4.5 BL s~ ' at 20C and at a speed of 3.0 BL s" 1 at 10C. The stimulation of red muscle during each tailbeat cycle occurred at the same time as or just prior to stimulation of pink muscle for most swimming conditions at 20C (Fig. 4, Table III). Phase difference (red vs. pink) was positive for all three positions at 4 BL s"'. This was true for the MID position at all swimming speeds. However, at lower swim- ming speeds, the time of electromyographic activity onset for pink and red muscle was the same for the more anterior positions; that is, the phase difference of red and pink muscle was near zero (Fig. 4). Swimming speed and longi- tudinal position significantly affected the onset time for red v.?. pink muscle during each tailbeat cycle (two-way ANOVA: for swimming speed, F = 3.484, P = 0.009; for longitudinal position, F = 3.640, P = 0.028). Pink muscle and red muscle are both activated prior to the beginning of shortening during each cycle of length change (Table III). The onset time for red muscle is before that of pink muscle, as described above. The phases of red and pink muscle are closest to one another at anterior positions and at lower swimming speeds. Generally similar results were obtained for swimming at 10C, but sample sizes were limited at that temperature. The duty cycle of muscle activity during swimming at 20C was affected significantly by fiber type (F = 4.34. P = 0.04; Table III). Red muscle had longer duty cycles than pink muscle. When data from the POST position are in- cluded, duty cycle was significantly affected by longitudinal position (F = 12.571. P < 0.001). This agrees with many other studies, such as Jayne and Lauder (1995) on bass, Wardle and Videler (1993) for both mackerel and saithe, and Rome et al. (1993) for scup red muscle. Discussion The recruitment and activity of aerobic swimming muscle depends on several variables: the type of muscle fiber, the longitudinal position on the fish, and the swimming speed. Pink muscle, with its faster kinetics, is used differently than red muscle. Whereas red muscle is active at most positions at all swimming speeds, pink muscle activity is restricted on the basis of both longitudinal position and swimming speed. Swimming speed of initial recruitment and duty cycle Fiber type affects the patterns of recruitment during swimming. Initial recruitment of pink muscle occurs at swimming speeds the same as or higher than recruitment speed of red muscle for all longitudinal positions. Pink muscle recruitment occurs at the lowest swimming speeds for the MID position only; at other positions, pink muscle is initially recruited at higher speeds than red muscle. Pink muscle is used minimally at low swimming speeds, when the tailbeat frequency and the frequency of oscillation of muscle are lowest. This correlates with the low power output of pink muscle at low oscillation frequencies. Aerobic muscle recruitment is also affected by longitu- dinal position on the fish. Positions more posterior on the fish body are recruited before anterior ones, particularly for pink muscle. Although pink muscle generally is recruited at the lowest steady swimming speed at the MID position, it is not recruited until higher swimming speeds at more anterior positions. At ANT-1, pink muscle is not consistently re- cruited until the near-maximum steady swimming speed. The pattern is less obvious for red muscle. At the lowest steady swimming speed, red muscle was active at the POST (Rome, unpubl. data), MID. and ANT-2 positions. At ANT-1, red muscle was not always recruited at the mini- mum steady swimming speed. This pattern makes sense from a functional viewpoint. ANT-1 pink muscle and red muscle undergo very low strain during swimming, even when the muscle is recruited. For instance, although all aerobic muscle fibers are recruited at 4.0 BL s' ' in scup, red muscle strain is around 2.0% for the ANT-1 position (Rome et til., 1993), and red and pink muscle strain at the ANT-2 position is less than 3% (Table III). At low swim- ming speeds, the combination of low strain and low oscil- lation frequency limits power production by the aerobic muscle (Coughlin and Rome, 1996). Power output of cycli- cally active muscle is a function of work per cycle and cycle frequency. If work per cycle is limited by very low muscle strain and if cycle frequency is low, power output will be low as well. In scup, the anterior aerobic musculature (pink PINK AND RED MUSCLE ACTIVITY IN SCUP 149 2.5BLs' 3.0 BLs' 3.5 BL s ANT-2 Red Pink MID Red Pink o o 4.0 BL s" 4.25 BL s' ANT-2 0.5s MID Red Pink Red Pink Figure 2. Electromyograms from red (R) and pink (P) muscle at the ANT-2 and MID positions for a scup swimming at 20C. Each trace is 1.0 s long. BL = body length. In this fish, both red and pink muscle are recruited at a swimming speed of 2.5 BL s" ' at the MID position. Red muscle at the ANT-2 position is recruited at this speed as well, although pink muscle at the ANT-2 position is not recruited until 4.0 BL s~' (as determined by computer-driven burst analysis). and, to a lesser extent, red muscle) is not consistently recruited at low swimming speeds when power production would be low or negative. A recent report on muscle function in swimming eels (Gillis, 1998) describes a pattern of recruitment for red muscle similar to that in scup. Posterior red muscle is recruited at the lowest steady swimming speeds, and red muscle from more anterior positions is initially recruited at higher steady swimming speeds. The results presented here for recruitment of red, pink, and white muscle in scup agree in a general sense with the work of Johnston and colleagues. Johnston et al. (1977) reported that pink muscle recruitment occurs at swimming speeds intermediate between those of red and white muscle. This seems to be due to the intermediate kinetics of pink muscle relative to the slower red and the faster white muscle (Coughlin et al.. 1996). In scup, the initial speed of pink muscle recruitment is most clearly intermediate at the ANT-2 position (Table II). At the MID position, recruitment of red and pink muscle occurs at the same swimming speed. At the ANT-1 position, pink muscle is often not recruited until the maximum steady swimming speed, just before recruitment of white muscle. That red and pink muscle at one longitudinal position can be recruited to power swim- ming independently raises an interesting question about the innervation patterns of the aerobic swimming musculature. Fiber type has a significant effect on duty cycle. For most swimming speeds and at most positions, the duty cycle of red muscle is longer than that of pink muscle. The simul- taneous offset of electromyographic activity for red and pink fibers in scup holds true across the range of swimming speeds reported here. Interestingly, the approximately si- multaneous offset of all red muscle on one side of the fish has been observed in a number of fish species, such as scup 150 D. J. COLIGHLIN AND L C. ROME 1.5BL.S 1 2.0 BLs" 2.5 BLs" ANT-2 Red Pink MID Red p ink o o 1.0s Figure 3. Electromyograms from red (R) and pink (P) muscle at the ANT-2 and MID positions for a scup swimming at 10C. Each trace is 1.5 s long. BL = body length. In this fish, both pink and red muscle are recruited at a swimming speed of 1.5 BL s~ ' at the MID position, although pink muscle recruitment is relatively weak. At the ANT-2 position, red muscle is weakly recruited at 1.5 BL s '. but pink muscle is not recruited until 2.0 BL s" 1 (as determined by computer-driven burst analysis). Swimming Speed (BL s" 1 ) Figure 4. Phase difteren e t red and pink muscle for three longitu- dinal positions across the range of steady swimming speeds at 20C. The difference in the timing of onset electromyographic activity in red and pink muscle during each length ch.i, t ,. \ , |e is expressed as a proportion of the period of one cycle. Positive values of relative phase indicated that red muscle is stimulated before pink i> iixcle during each length change cycle. (Rome et al.. 1993; Coughlin and Rome. 1996), carp (van Leeuwen et ai, 1990), bass (Jayne and Lauder, 1995), and mackerel and saithe (Wardle and Videler, 1993). Alterna- tively, more elongate fish that swim with greater body curvature do not have a simultaneous offset of the red muscle electromyographic activity. Instead, the offset of red muscle activation progresses from anterior to posterior, just as the wave of the onset of activation progresses (Wardle et til., 1995; Hammond et al., 1998). Not enough is known about the nervous control of swimming muscle activity to explain the variety of aerobic fiber activation patterns seen in swimming fish. Phase of muscle activity Both red muscle and pink muscle are activated prior to muscle shortening for most swimming conditions, such that there is a negative phase shift of electromyographic activity relative to shortening. This agrees with most previous work on red muscle in steady swimming fish such as bass (Jayne and Lauder, 1995, 1996), mackerel and saithe (Wardle and Videler, 1993), carp (van Leeuwen et ai. 1990; van Leeu- wen. 1995), eel (Williams etui. 1989; Wardle et ai. 1995). adult rainbow trout (Hammond et al.. 1998), and with previous work on scup (Rome et al.. 1993). All of these PINK AND RED MUSCLE ACTIVITY IN SCLIP 151 Table III Patterns of in vivo muscle activity for scup swimming at 2.5 and 4 BL s ' at 20C with a tai/beat frequencv of 4.0 and 6.0 Hz. respectively 2.5 BL s 4.0 BL Fiber type ANT-2 (n = 4) MID(;i = 3) ANT-2 1 ;i = 9) MID |/i = 4) Strain Pink 3.69 0.31 4.06 0.20 2.62* 3.43* Red 3.6* 5.5* 2.9* 4.8* Duly cycle Pink 0.310 0.062 0.367 0.021 0.412 0.021 0.331 0.037 Red 0.395 0.040 0.403 0.040 0.421 0.016 0.391 0.014 Phase Pink -14.5 11.4 -52.1 26.4 -14.5* -17.6* (-0.040) (-0.145) (-0.040) (-0.049) Red -16.1 20.7 -65.2 27.7 -32.7* -40.8* (-0.045) (-0.181) (-0.091) (-0.113) Strain is the percent change in muscle length during the oscillatory length change cycle. Duty cycle is the duration of the burst of electromyographic activity during each oscillatory cycle expressed as a proportion of the oscillation period. Phase is the timing of the onset of the activity burst relative to the onset of shortening, expressed in degrees (one oscillation cycle = 360). Negative values indicate muscle activity preceding muscle shortening. For comparison, phase expressed as a proportion of the period oscillation is provided in parentheses. Data are not provided for the ANT-1 position because the data set for that position at 2.5 BL s"' is not complete. Sample size (n) is given in parentheses at the top of each column. * Red muscle data from Rome et al. (1993); pink muscle data from Coughlin and Rome (1996). studies report results similar to those presented here a smaller phase shift for red muscle at more anterior positions. In a few fish, such as rainbow trout smolts (Williams et al., 1989) and, at some swimming speeds, bass (Johnson et al., 1994; Jayne and Lauder. 1995), a positive phase of red muscle activity has been reported for anterior positions. In these fish, muscle shortening occurs prior to muscle stimu- lation. In general, this was not observed in scup, although for a few fish, red muscle at the ANT-1 position did have a positive phase at low swimming speeds (unpubl. data). However, technical difficulties inherent in measurements of the muscle length change at this position lead to consider- able uncertainty of the phase. For rainbow trout, at least, a positive phase shift of muscle activation results in net negative power production (Coughlin and Burdick. 1996). Anterior muscle that under- goes either a very small negative phase shift or a positive phase shift will not contribute significantly to powering swimming (Rome et cil., 1993: Coughlin and Rome, 1996). For this reason. Hammond et al. (1998) suggest that the phase data from Williams et al. (1989) cannot reflect red muscle activity during steady swimming in trout. However, the many species of fish that show a relatively low negative phase in the anterior red muscle probably swim similarly to scup: most of the power is generated by the posterior musculature. Curiously, the data of Gillis (1998) suggest the same may be true for eels: at low steady swimming speeds, only the posterior red muscle is recruited to power swim- ming; the anteriormost red muscle is not recruited until the highest steady swimming speeds. Furthermore, when it is recruited, the negative phase shift is smaller than in poste- rior muscle (Gillis. 1998). This small negative phase shift may limit power production by anterior eel red muscle during swimming and. therefore, be associated with the lack of recruitment of this muscle at lower swimming speeds. No previous research has reported on the relative phase differences in the stimulation of the aerobic muscle fibers of fish during each length change cycle of swimming. During each cycle of length change, pink muscle is stimulated after red muscle consistently (Fig. 4), except in the more anterior positions at lower swimming speeds. This correlates with the faster activation kinetics of pink muscle (Coughlin et al., 1996): since it activates more rapidly, its stimulation can occur nearer to the onset of muscle shortening. Otherwise, it would reach high force levels while the muscle was still lengthening and might then generate net negative work during the length change cycle. Jayne and Lauder (1994) report similar results for red and white muscle in bass swimming at high, unsteady speeds. In those fish, white muscle stimulation lagged behind red muscle stimulation at a given longitudinal position. However, the lag between red and white activation was the same for both anterior and posterior positions; this finding differs from our results. We found that the phase difference between red and pink mus- cle increased at the more posterior positions of steady swimming scup. Pink muscle function in swimming fish Pink muscle is employed effectively by swimming scup. At maximal steady swimming speeds at both 10 and 20C, pink muscle is heavily recruited and allows fish to swim steadily at speeds greater than possible with red muscle alone. At swimming speeds at which the performance of red muscle is near maximal, pink muscle produces considerable power for swimming (Coughlin and Rome. 1996). In addi- 152 D. J. COUGHLIN AND L. C. ROME tion. across a range of swimming speeds, the activation patterns of pink muscle (i.e. the delayed onset of activation) are well timed to take advantage of this muscle's faster rate of activation. At submaximal swimming speeds, the pink muscle at anterior body positions is a relatively ineffective source of power, and its recruitment is correspondingly limited. Acknowledgments We thank D. Leavitt, B. Lancaster, and B. Tripp and the rest of the staff of the Coastal Research Center at the Woods Hole Oceanographic Institution for facilitating our work. We thank Linn Nguyen for laboratory assistance. This work was supported by NIH AR38404 and NSF IBN-95 14383 to LCR and by an NIH postdoctoral fellowship to DJC. Literature Cited Bone, Q. 1989. Evolutionary patterns of axial muscle systems in some invertebrates and fish. Am. Zoo!. 29: 5-18. Coughlin, D. J., and J. W. Burdick. 1996. Muscle mechanics and power production in swimming rainbow trout, Oncorhynchus mykiss. Am. Zool. 36: 61 A. Coughlin, D. J.. and L. C. Rome. 1996. The roles of pink and red muscle in powering steady swimming in scup, Stenotonuis chrysops. Am. Zool. 36: 666-677. Coughlin, D. J., G. Zhang, and L. C. Rome. 1996. Contraction dynam- ics and power production of pink muscle of scup (Stenotomus chry- sops). J. Exp. Bio/. 199: 2703-2712. Gillis, G. B. 1998. Neuromusculur control of anguilliform locomotion: patterns of red and white muscle activity during swimming in the American eel Anguilla rostrata. J. Exp. Binl. 201: 3245-3256. Hammond, L., J. D. Altringham, and C. S. Wardle. 1998. Myotomal slow muscle function of rainbow trout. Oncorhynchus mykiss during steady swimming. J. E.\p. Biol. 201: 1659-1671. Jayne, B. C., and G. V. Lauder. 1994. How swimming lish use slow and fast muscle fibers: implications for models of vertebrate muscle recruitment. J. Comp. Physiol. A 175: 123-131. Jayne, B. C., and G. V. Lauder. 1995. Red muscle motor patterns during steady swimming in largemouth bass: effects of speed and correlation with axial kinematics. J. Exp. Biol. 198: 657-670. Jayne, B. C., and G. V. Lauder. 1996. New data on axial locomotion in fishes: how speed affects diversity of kinematics and motor patterns. Am. Zool. 36: 642-655. Johnson, T. P., D. A. Syme, B. C. Jayne, G. V. Lauder, and A. F. Bennett. 1994. Modeling red muscle power output during steady and unsteady swimming in largemouth bass. Am. J. Physiol. 267: R481- R488. Johnston, I. A., W. Davison, and G. Goldspink. 1977. Energy metab- olism of carp swimming muscles. J. Comp. Physio/. 114: 203-216. Josephson, R. K. 1985. Mechanical power output from striated muscle during cyclical contraction. J. Exp. Biol. 114: 493-512. Rome. L. C. 1995. A device for synchronizing biomechanical data to cine film. J. Bimnech. 28: 333-338. Rome, L. C., P. T. Loughna, and G. Goldspink. 1984. Muscle fiber recruitment as a function of swim speed and muscle temperature in carp. Am. J. Physiol. 247: R272-R270, Rome, L. C.. R. P. Funke, and R. M. Alexander. 1990. The influence of temperature on muscle velocity and sustained performance in swim- ming carp. J. Exp. Biol. 154: 163-178. Rome, L. C., I. Choi, G. Lutz, and A. A. Sosnicki. 1992. The influence of temperature on muscle function in fast swimming scup. I. Shortening velocity and muscle recruitment during swimming. J. Exp. Biol. 163: 259-274. Rome, L. C., D. Swank, and D. Corda. 1993. How fish power swim- ming. Science 261: 340-343. van Leeuwen, J. L. 1995. The action of muscles in swimming fish. Exp. Physiol. 80: 177-191. van Leeuwen, J. L., M. J. M. Lankheet, H. A. Akster, and J. W. M. Osse. 1990. Function of red axial muscles of carp (Cyprinus carpioY. recruitment and normalized power output during swimming in different modes. J. Zoo/. (Lend.) 220: 123-145. Wardle, C. S., and J. J. Videler. 1993. The timing of the electromyo- gram in the lateral myotomes of mackerel and saithe at different swimming speeds. J. Fish Biol. 42: 347-359. Wardle, C. S., J. J. Videler, and J. D. Altringham. 1995. Tuning in to fish swimming waves: body form, swimming mode and muscle func- tion. J. Exp. Biol. 198: 1629-1636. Williams, T. L., S. Grillner, V. V. Smoljaninov, P. Wallen, S. Kashin, and S. Rossignol. 1989. Locomotion in lamprey and trout: relative timing of activation and movement. J. Exp. Biol. 143: 559-566. Zhang, G., D. M. Swank, and L. C. Rome. 1996. Quantitative distri- bution of muscle fiber types in the scup. J. Morph. 229: 71-81. Reference: Bio/. Bull. 196: 153-162. (April 1999) Osmoregulation and FMRFamide-Related Peptides in the Salt Marsh Snail Melampus bidentatus (Say) (Mollusca: Pulmonata) HAMID R. KHAN 1 '*. DAVID A. PRICE 2 , KAREN E. DOBLE 2 , MICHAEL J. GREENBERG 2 . AND A.S.M. SALEUDDIN 1 ^Department of Biology. York University, Toronto, Ontario, Canada M3J IPS: and 2 Whitney Laboratory, University of Florida, St. Augustine. Florida 32086 Abstract. The pulmonate snail Melampus bidentatus oc- cupies the high intertidal zone of salt marshes in a nearly terrestrial environment. The hemolymph osmolarity of the snails collected in the field paralleled that of the adjacent water and was affected by the tides and precipitation. The snails initially gained or lost weight when submerged in hypo- or hyperosmotic media, respectively, but returned to their original weight after 24 h. The content of their immu- noreactive (IR)-FMRFmide-/\elated Peptide.v (FaRPs) was measured in various tissues by radioirnmunoassay, and IR- FaRPs were found in every tissue analyzed. The subesopha- geal part of the central nervous system (CNS) contained more IR-FaRPs than the supraesophageal part, and the kid- ney and the tissues of the reproductive tract contained more than other peripheral tissues. The levels of IR-FaRPs in the CNS, kidney, and hemolymph were higher in snails that were immersed in higher concentrations of seawater. Many IR neurons are present in all ganglia of the CNS except the pleural ganglia, and IR neurites are extensively distributed within the CNS and its connective tissue sheath. The vis- ceral nerve from the visceral ganglion is immunoreactive and could be seen to innervate the kidney, which contains IR-varicosities. An osmoregulatory role for the FaRPs is suggested. Received 7 January 1998; accepted 27 January- 1999. * To whom correspondence should be addressed. E-mail: hamid(S' luring. sci.yorku.ca Abbreviations: One-letter abbreviations of the amino acids are used to state the peptide sequences. ASW. artificial seawater; BSA. bovine serum albumin; CNS. central nervous system; FaRPs. FMRFmnide-related />ep- tide.s; IM. incubation medium; IR. immunoreactive: NGS. normal goat serum: PBS. phosphate-buffered saline: RIA. radioirnmunoassay. Introduction Melampus bidentatus is a common amphibious pulmo- nate snail. Its habitat is the high intertidal zone of salt marshes, and it ranges from Nova Scotia, Canada, to the coast of the Gulf of Mexico in Texas (Apley. 1970; Hilbish. 1981 ). Melampus belongs to the primitive family Ellobiidae and is believed to be related to an ancestral marine strain that colonized the intertidal habitat, giving rise to the land snails and then to the freshwater snails (Morton, 1955: Russell-Hunter. 1978). Environmental factors such as tem- perature, salinity, and tides are extremely variable in the high littoral habitat. But the adult snails are adapted to this variation, and can survive for several days at temperature extremes of -12C to 40C. submergence in 25%-100% seawater, and desiccation at 0% relative humidity for 27-36 h (Price, 1980; McMahon and Russell-Hunter, 1981). Al- though adults of M. bidentatus can endure a terrestrial life, its planktonic veliger larvae are restricted to the aquatic habitat of estuaries (Russell-Hunter et al.. 1972). The natural history of M. bidentatus suggests that in estuarine and semiterrestrial molluscs, osmotic and volume regulation must be particularly responsive. Moreover, scat- tered evidence suggests that the family of neuropeptides related to FMRFamide is involved in this regulation. First. in a freshwater pulmonate snail, Helisoma duryi, FMRF- amide causes water retention in the kidney in vitro (Saleud- din et al.. 1992). Further, the level of immunoreactive, FMRFamide-related peptides (IR-FaRPs) in the kidney of Helisoma trivolvis kept in hypoosinotic medium is lower than that in the kidneys of snails kept in isosmotic medium (Madrid et al., 1994). The IR-FaRPs have also been local- ized in the kidneys of the terrestrial pulmonate snail Helix 153 154 H. R KHAN ET AL. aspersa (Lehman and Price, 1987), and in the central ner- vous system (CNS) of the veligers of M. bidentatns (Mof- fett, 1992). Finally, the regulation of hypoosmotic cell vol- ume by cardiac muscles of the clam Mercenaria mercenaria is potentiated by FMRFamide (Deaton. 1990). Since M. bidentatns inhabits highly variable osmotic con- ditions, it is an appropriate species with which to further test the proposal that the FaRPs have a role in osmoregulation. In this paper, we have studied, both in the field and the laboratory, the influence of the ambient osmotic concentra- tion on that of the hemolymph. We have then compared the levels of IR-FaRPs in the CNS, hemolymph, and kidney of snails maintained in media of different osmotic concentra- tions. We also report IR-FaRP-staining of neurites in the kidney, and of neurons in the CNS of adult snails. Materials and Methods Animals and media During January to April, M. bidentatns adults (shell lengths, 10.0 2.0 mm) were collected at random from an area 100 m X 20 m in a salt marsh on the Matanzas river estuary near Crescent Beach, Florida (294()' N : 8113' W). Experiments with the freshly collected snails were per- formed nearby, at the Whitney Laboratory of the University of Florida. At York University, the animals were maintained at 22C, 95% relative humidity, and a photoperiod of 14 h light to 10 h dark, in glass tanks (30 cm L x 15 cm W x 20 cm H) covered with nylon screens. The floor of each tank was covered with a layer of crushed oyster shells (poultry feed) graded in thickness from 0.5 cm at one end of the tank to 3.0 cm at the other. A volume of 50% artificial seawater (ASW; Instant Ocean, Aquarium Systems, Ohio) was added to cover about one-third of the crushed oyster shells. Slanted against the walls of each tank were 4-6 broken, irregularly shaped pieces of clay plant-pot (about 10 cm in diameter). Most snails crawled to terrestrial conditions above the water level on the walls of the tank and on the dark side of the moist pieces of clay plant-pot. Distilled water was added daily to maintain osmolarity and water level, and the snails were fed fish-food and boiled lettuce ad libitem. The snails were acclimated to laboratory conditions for at least 2 weeks before experimentation. The kidney is embedded just under the dorsal surface of the mantle tissue, and the CNS is located under the skin, between the tentacles. Before dissection, the snails were chilled on ice for 5-10 min, which reduced their movements and mucus secretion; no other anesthetic was used. Deshelled snails were pinned on a dish lined with Sylgard (Dow Corning, Canada) and dissected in filtered isosmotic ASW. We adjusted the osmolarity of the ASW with distilled water to be isosmotic with that of the hemolymph, which varies with the medium or relative humidity (described later). The kidney and CNS were removed from a snail within 3-5 min, and were used for FaRP studies. Osmotic conditions Snails that had been maintained under terrestrial condi- tions (described above) were submerged for various periods (maximum, 4 days; minimum 3 h) in 10%, 50%, or 100% ASW (dilutions made with distilled water) in 250-ml plastic containers, each with a nylon screen. The levels of IR- FaRPs were measured (described later) after 24 h of osmotic treatments. Hemolymph was collected as follows. The snails were pricked in the head area with a sharp needle. Each animal was then quickly placed in a 0.5-ml polypro- pylene microcentrifuge tube with a hole in its bottom. This tube was then placed in a larger (1.5 ml) polypropylene microcentrifuge tube. The nested tubes and the snail were spun at 500 X g for 2-5 s. Then the inner tube containing the snail was removed, and the outer tube containing the hemolymph was spun for an additional 5 min. About 30-50 /id of hemolymph could be collected from each snail. The osmotic pressures of 10-jul samples were measured in a Wescor Model 5300 vapor pressure osmometer (Logan, Utah). After the bleeding, all of the snails fed and lived for many days thereafter. To measure the effects of osmotic water exchange on weight, the snails were weighed with a Mettler AE163 balance after the visible water and mucus had been removed from their shells and feet with adsorbent tissue. Radioimmnnoassay Pulmonate molluscs contain two major classes of FaRPs: the tetra-FaRPs (FMRFamide and FLRFamide) and the hepta-FaRPs (XDP[F/Y]LRFamide, where the N-terminal residue X is pQ, S, N, or G). The hepta-FaRP analog with the glycyl residue occurs only in snails of the subclass Basommatophora (like M. bidentatiis), and these animals lack the analog with the pyroglutamic acid (pQ) residue, which occurs only in stylommatophorans (Price et al, 1987a, b). To ensure that all of the FaRPs would be de- tected, we used S253 antiserum in the radioimnumoassay (RIA). The antiserum was raised to thyroglobulin-peptide (synthetic analog YGGFMRFamide) conjugate in a rabbit. It has high affinities for both FMRFamide and GDPFLRF- amide. and was used in the assays at a dilution of 1 : 10,000. lodinated pQYPFLRFamide was used as tracer (for details see: Price et ai, 1990; Lesser and Greenberg. 1993). The protocol of Madrid et al. (1994) was followed in processing the tissues. After dissection, each tissue sample was placed in a plastic tube ( 1.5 ml) containing HPLC grade acetone (1 part tissue: 4 parts acetone) and frozen immedi- ately at -80C for at least 24 h. The acetone extracts were spun for 5 min at 5,000 X g, and the supernatants were collected and then dried in a Speed-Vac centrifuge. Each OSMOREGULATION AND FMRFAMIDE IN A SNAIL 155 pellet was dissolved in 100 ju.1 RIA buffer |().()1 M sodium phosphate with 1% bovine serum albumin (BSA), 0.9% sodium chloride. 0.01% merthiolate. and 0.025 M sodium EDTA|. Aliquots of 10 /j.1 were placed in glass tubes to which 100 y,\ tracer ( 10.000 cpm) (in RIA buffer) and 100 /Ltl diluted antiserum (in RIA buffer) were added. The tubes were stored at 4C overnight: on the following morning. 1 .0 ml charcoal suspension (0.25% charcoal, 0.025% dextran. 0.01% merthiolate in 0.1 M sodium phosphate. pH 7.5) was added to each tube. After 10 min. the mixture was centri- fuged at 2500 X g for 15 min at 4C, and the supernatant was counted in an LKB MiniGamma counter. All statistical comparisons were made with a one-way analysis of vari- ance (ANOVA); when the ANOVA showed a significant difference, the effect was followed with the Bonferroni multiple comparison post test using a statistical software package (InStat. GraphPad Software, San Diego, Cali- fornia). Immunocytochemistry Isolated CNS and kidney tissues from animals in labora- tory terrestrial condition were either embedded in paraffin and sectioned, or examined in whole mount. Sections were prepared as follows. The tissues were fixed for 12-18 h in modified Bouin-Hollande [7% picric acid. 2.5% copper acetate. 2% formaldehyde (freshly prepared from parafor- maldehyde), and 1.5% glacial acetic acid, in phosphate- buffered saline (PBS) (0.2 M NaCl. 0.003 M KC1, 0.002 M KH 2 PO 4 . 0.02 M Na 2 HPO 4 , 0.001 M CaCL. 0.001 M MgCL, pH 7.5)], then dehydrated in graded ethanol solu- tions and embedded in paraffin. Serial sections (10 ^im) were cut and mounted on coverslips (22 mm X 22 mm) coated with 0.5% gelatin. 0.5% chrome alum, and 0.01% formaldehyde. The deparaffmized sections were treated with 100% methanol and 0.1% H 2 O 2 for 5 min. hydrated in a graded series of ethanol solutions, kept 30 min in an incubation medium (IM) consisting of PBS containing 2% Triton X 100 (Sigma Chem., St. Louis, MO), 5% normal goat serum (NGS). and 1% BSA. The sections were incu- bated for 1 h at 22C with hepta-FaRP-specific CK anti- serum, which was raised in a rabbit to the peptide CKQD- PFLRFGK (a gift from Dr. G.A. Cottrell. University of St. Andrews. Scotland) (Cottrell et al, 1994). The primary antiserum was diluted 1 :200 in IM. After many rinses in IM for 10 min, the sections were incubated for 1 h at 22C in fluorescein-conjugated goat antirabbit-serum (Sigma) di- luted 1:80 in IM (secondary antibody). The coverslips were rinsed well in PBS and mounted on glass slides in 5% poly vinyl alcohol. 30% glycerol, and 0.1 7r phenylenedi- amine in PBS (mounting medium). Whole mounts were prepared as follows. The tissues were treated with proteases before and after fixation (Long- ley and Longley, 1986). The tissues were first treated in 0.05% pronase in isosmotic saline for 10 min, then fixed in Bouin-Hollande (see above) for 12 h. After fixation, the tissues were rinsed in saline for 10 min, then treated in 0. 1 % trypsin in saline for 30 min, and rinsed well in saline for 10 min. A preincubation in IM for 6-12 h at 4C was followed by incubation in primary antibody (hepta-FaRP-specific CK antiserum; 1 :200 in IM) for 24 h. After rinsing several times in 1% Triton X 100 in PBS for 1 h, the tissues were incubated in secondary antibody (1:80 in IM) for 6 h, followed by repeated rinses in IM. When whole-mount tissues were incubated for more than 60 min. this was done in the dark at 4 "C. The tissues were mounted in mounting medium (see above) between two coverslips (one rectangu- lar. 60 mm X 22 mm; and one circular, 18 mm). To prevent tissue distortion, the weight of the coverslip was supported by four pieces of broken coverslip (0.5 mm diameter). The larger coverslip was affixed with adhesive tape on an aluminum slide (36 mm X 80 mm X 1 mm) with a central 25-mm circular window. The smaller circular coverslip was placed face down in this window, so these preparations could be flipped for viewing from either side. For controls, either pre-immune rabbit serum ( 1 : 200 in IM) or IM alone was used in the primary incubation; in either case, the tissues were then stained with secondary antibody as described above. The controls showed no stain- ing. The specimens were viewed and photographed with either a Leitz epifluorescence microscope or a Bio-Rad MRC 600 confocal microscope. The IR-FaRP cells were mapped from the serial sections and from whole-mount images, and were measured with the calibration marker from the confocal images. Results Osmoregulation The high-tide location of the marsh where M. bidentatus was collected has an uneven surface with a shallow slope, and contains abundant decaying organic matter of plant and animal origin. The osmolarity and pH of the waters in the habitat varied with time and location in the observation area. During the sunny days of April, the average osmoiarities were 1200 26 mosm/kg H 2 O (mean standard error of mean, n = 50) (range = 1015-1650 mosm/kg FLO), and the pH ranged from 6.0-7.2: at 1200-1400 h, the ambient air temperature was 31 4C, and the soil surface and water temperatures were 34.2 1.3C. At the same time, the tidal seawater coming through the Matanzas inlet had a consis- tent osmolarity of 980 5 mosm/kg FLO. a temperature of 22 2C, and a pH of 7.5 0. 1 . The variables that altered the osmoiarities of the water in the habitat and hemolymph of snails were tide, temperature, and precipitation. Most snails in the field were out of the water, on grass stems or higher points of the uneven ground, while a much smaller number were creeping and feeding underwater. The osmo- 156 H R, KHAN ET AL. larities of the hemolymph of snails from different locations and conditions and during high and low tides were mea- sured within 30 min after collection and compared with the osmolarities of the adjacent media. The hemolymph osmo- larities ranged from 1 100 to 1500 mosm/kg H 2 O and ap- peared to parallel that of the adjacent medium. A few hours of heavy rainfall reduced the osmolarity of the water in the habitat to 750 40 mosm/kg H 2 O; the hemolymph osmo- larities also declined to 850 50 mosm/kg H 2 O. Most snails in the laboratory crawled to the dry parts of the tank (see methods section) and stayed away from the water; thus they sought terrestrial conditions. But a few snails were also seen feeding and crawling underwater. The hemolymph osmolarity of the laboratory snails that sought terrestrial conditions was much lower (450 25 mosm/kg H 2 O) than that of field snails (1200 100 mosm/kg H 2 O) collected away from the water. When laboratory snails from terres- trial conditions were held for 24 h in 10% or 50% ASW, the osmolarity of the hemolymph paralleled, and was about 150 mosm/kg H 2 O hyperosmotic to, that of the medium. Snails in 10% and 50%- ASW gained weight during the first 3-6 h, but over 24 h, gradually returned towards their original weight. In contrast, snails kept in 100% ASW for 24 h first lost weight and then returned towards their original weight (Fig. 1). Levels of IR-FaRPs Immunoreactive-FaRPs were detected in every tissue studied; the CNS had higher levels than the other organs. The supraesophageal portion of the CNS (the buccal and Weight changes 100% ASW 50% ASW 10% ASW Table I. Immunoreactive FaRP content detected by S253 antisemm in various tissues of animals from terrestrial condition in laboratory 7.5- I 3- CL o: ,, ro 2 U. ri o 1 Q. IR-FaRPs in Hemolymph 0.4- S 0.3H Q. OL "? 0.2- "o |o.i- n n- 1 T T 2 CO ^p o m 2 CO o t Terrestrial CO o IR-FaRPs in Kidney i $ CO 1.2 l 1 -,- ro In 0) ^ 1 k_ o 3 o o m Figure 2. IR-FaRPs in the CNS. Note that the level of IR-FaRPs in the terrestrial condition is higher than that in the three ASW conditions (P < 0.05 for 100% ASW and 50% ASW; P < 0.01 for 10% ASW. F = IX.02) (n = 5; mean standard error of mean). Figure 3. IR-FaRPs in the hemolymph. The levels of IR-FaRPs in the terrestrial and 100% ASW conditions are significantly greater than those in 50% and 10% ASW conditions (P < 0.01 and P < 0.001. respectively. F = 10.50) (n = 5; mean standard error of mean). Figure 5. Optical section of the whole mount of the kidney (k) and adjacent mantle (m) tissue showing IR-fibers and varicosities (arrows) in the kidney. Scale bar: 50 /^m Figure 6. Longitudinal section of the kidney showing IR-staining (thin arrows) possibly as innervations to smooth muscles within the kidney tissue and also in the periphery, kl. kidney lumen; me, mantle cavity. Scale bar: 300 ^m.. The distribution of IR-neurons in the CNS is mapped in Figure 7. All commissures, connectives, and major nerves in the CNS display IR-fibers, and all ganglia except the pleu- rals contain IR-cell bodies. Visceral and right parietal gan- glia contain both large (20 5 /am) and small ( 10 5 jam) IR-neurons, whereas mostly small and only a few large IR-neurons are found in the cerebral and pedal ganglia (Figs. 7-9, 12). Cerebral ganglia. Two groups of large (20 5 jam) and small (10 5 /am) IR-neurons are present at the anterior and posterior location of the cerebral commissure (Figs. 7-9). The middle part of each cerebral ganglion contains a large (20 5 /am) IR-cell surrounded by small IR-cells (Fig. 9). In the lateral lobes, 2-4 large (20 3 jam) cells and 8-12 small ( 12 4 jam) IR-cells are seen (Figs. 7, 10). Figure 4. IR-FaRPs as detected by antiserum S253 in the kidney. The level of IR-FaRPs in the terrestrial condition is significantly (P < 0.05. F = 8.17) higher than that of 10% ASW (n = 5; mean standard error of mean). 158 H. R. KHAN ET AL. Ptn Figure 7. Showing the schematic distribution of the IR-cells in the CNS. BG. buccal ganglia; CBC. cerebro-buccal connective; CC, cerebral commissure; Cn, cutaneous nerve; CPC. eerebro-pedal connective; CPIC. cerebro-pleural connective; DB, dorsal bodies; LCG, left cerebral gan- glion; LL. lateral lobe; Ln, labial nerves; LPdG, left pedal ganglion; LP1G. left pleural ganglion; LPrG, left parietal ganglion; On, optic nerve; PC, pedal commissure; Pdn. pedal nerve; Ptn, peritentacular nerve; Prn. parietal nerves; PVn, parieto-visceral nerve; RCG. right cerebral ganglion; RPdG. right pedal ganglion; RP1G, right pleural ganglion; RPrG. right parietal ganglion; SPC. sub-pedal commissure; Tn, tentacular nerve; VG. visceral ganglion; Vn, visceral nerve. Not to scale. Buccal ganglia. In each buccal ganglion. 3 intermediate sized (18 2 /xnii IR-cells, and many IR-fibers and vari- cosities are seen (Fig. 111. Pedal ganglia. Groups of 6-10 small IR-neurons are seen at the periphery of each pedal ganglion (Figs. 7, 12). Parietal ami visceral ganglia. More large and small IR- cells were seen in the right parietal and visceral ganglia than in any other ganglion (Figs. 7, 13). The left parietal ganglion contains only small IR-neurons (Fig. 7). The IR-cells are arranged in anterior and posterior groups in both ganglia (Figs. 7. 14). Numerous IR-neurites emanate from the CNS into the surrounding connective tissue and appear to termi- nate in varicosities (Fig. 15). Discussion In both the field and the laboratory, most individuals of M. bidentatus were emergent, living under semiterrestrial conditions. But some members of the population were al- ways submerged, suggesting that they crawl in and out of water intermittently. The osmotic concentration of the he- molymph was correlated with that of the adjacent aqueous medium. Moreover, IR-FaRPs are present in the CNS, kid- ney, and hemolymph, and the levels of these peptides are also correlated with the osmotic concentration of the ambi- ent medium. Osmoregulation and kidney When the osmolarity of the external medium changes, the hemolymph osmolarity of M. bidentatus also changes due to inward and outward movements of water from its body. Such conditions must be offset by various adjustments: regulation of urine production by its kidney is one of them. The IR-staining of the kidney was seen as a diffuse pattern on the kidney cells, as an intense localized pattern in areas with neural arborizations and varicosities, and in other small areas that are probably bundles of smooth muscles. The smooth muscles of the kidney of the basommatophoran snail Helisoma duiyi are innervated by FMRFamide-immu- noreactive axons (Saleuddin et ai. 1992). In this species the /'/( vitro contraction of kidney tissues was influenced by FaRPs (unpubl. obs., A.S.M. Saleuddin). The FaRPs are well known for their contractile activity on visceral and smooth muscles in various species. They are also known to regulate salivary glands and ion channels (Bulloch et al., 1988; Green et ai, 1994; Price and Greenberg, 1994; Lingueglia et al.. 1995). In the pulmonate kidney they are probably involved in regulating smooth muscle contractions for peristalsis and the production of hydrostatic pressure for urine formation, as well as in regulating ion channels for selective secretion and reabsorption of ions from urine. Osmoregulation and IR-FaRP levels The CNS and hemolymph displayed the highest and lowest levels of IR-FaRPs, respectively. Organs containing involuntary muscles, such as the reproductive tracts and kidney, have high levels of IR-FaRPs. These organs appar- ently receive FaRPs through direct innervation by neurons from the CNS. Thus, in these target tissues, the FaRPs may act as neuromodulators or as paracrine agents. Numerous IR-FaRP processes terminating in varicosities in the con- nective tissue around the CNS may constitute a diffuse OSMOKMiH . \IION AM) I-MRI- \\IIIM IN A SNA 1 1 159 Figure 8. Groups of IR-cells (large arrows! in a cross section of the left and right cerebral ganglia; thin arrows point to IR-material in the commissure (cm) and neuropile. Scale bar: 100 fj.ni. Figure 9. Optical section of the left cerebral ganglion showing a large IR-cell (large arrow) among small IR-FMRFamide cells and their neurites (small arrows). Scale bar: 20 /urn. neurohemal area and account for the hemolymph IR-FuRP levels. The levels of IR-FaRPs in the CNS, hemolymph, and kidney varied under different osmotic conditions. The IR- FaRP levels were reduced in hypoosmotic conditions com- pared with those in hyperosmotic or terrestrial conditions. Terrestrial or hyperosmotic salt-water conditions produced threats of desiccation and salt loading in snails, whereas hypoosmotic conditions had the opposite effect. The changes of IR-FaRP levels in the tissues of M. bidentatus by osmotic conditions may reflect osmoregulation by the snail. Since IR-FaRP levels increased in the CNS, hemolymph. and kidney under conditions of increased water losses such as in terrestrial or 100% ASW conditions, the FaRPs may have an antidiuretic function. //; vitro culture of the kidney tissues of Helisoma with synthetic FMRFamide increased its intracellular water uptake, and FMRFamide has been suggested to have an antidiuretic role (Saleuddin et al., 1992; Madrid et al, 1994). Antidiuretic and diuretic activ- ities of FMRFamide and GDPFLRFamide respectively have been demonstrated in nephridia of a leech (Salzet et al., 1994). Basommatophoran pulmonates such as Melampus contain GDPFLRFamide (unpub. obs., D. A. Price). Both tetra- and hepta-FaRPs are present in the osmoregulatory tissues of Helisoma (Madrid et al., 1994). The tetra- and hepta-FaRPs are expressed in exclusive neurons and may have different actions in a target tissue (Greenberg and Price. 1992; Benjamin and Burke, 1994; Price and Green- berg, 1994). The precise mechanism of the tetra- and hepta- FaRP actions in molluscan kidneys awaits future studies. In molluscs, osmoregulation by the kidney appears to be a combined effect of many physiological systems regulated by different hormones. In Lymnaea stagnalis, a neuropep- tide that stimulates sodium uptake by the skin has been sequenced; furthermore, axons that release immunoreactive sodium-influx-stimulating peptides are present in this snail (De With et al., 1994). In Aplysia califomica, R15al neu- ropeptide causes water retention, thus having an osmoreg- ulatory role (Weiss et al., 1989). Additionally, in this spe- cies various kidney functions are modulated by L10 and LUQ neurons, and the latter neurons are immunoreactive to FMRFamide (Koester and Alevizos, 1989; Giardino et al., 1996). Su,ch systems include increased circulation of hemo- lymph for ultrafiltration, subsequent secretion and reabsorp- tion of ions and organic matters, and expulsion of final urine (Khan and Saleuddin, 1979). Bioassay studies assessing the effects of FaRPs on kidney functions will enhance under- standing of hormonal control of osmoregulation. Location of IR-FaRP cells The neurosecretory cells of the CNS of M. bidentatus have been described by using histochemical staining (Price, 1977; Ridgway, 1987). The IR-FaRP cells described in this paper are some of these neurosecretory cells; their pattern of distribution is basically similar to that of other basommato- phoran pulmonates such as L\mnaea stagnalis and Heli- soma duryi (Schot and Boer. 1982; Buckett et al., 1990; Murphy et al., 1985; Saleuddin et al.. 1992). In the cerebral ganglia of M. bidentatus, the IR-cells are present in groups located anteriorly and posteriorly. Similar groups are also seen in the cerebral ganglia of H. duryi near the endocrine growth-regulating mediodorsal cells and the ovulation-reg- ulating caudodorsal cells. In the latter species, however, an additional large group of small IR-cells is seen in the left cerebral extension; such cells are absent in the correspond- ing location in M. bidentatus. The lateral lobes of the cerebral ganglia of M. bidentatus contain several large and small IR-cells, whereas those of H. duryi contain only two IR-cells. and those of L. stagnalis have none (Saleuddin et 160 H. R. KHAN ET AL. Saleuddin et /., 1994; Saleuddin and Ashton, 1996). In the buccal ganglia of M. bidentatus, three IR-cells have been seen; in those of H. dur\i two cells occur: and in those of//. Figure 10. Optical section of the right lateral lobe and a part of right cerebral ganglion (eg) showing large (large arrows) and small (small arrows) IR-cells. Note that the large cells are intensely stained and the small cells are less intensely stained; the cells can be recognized by the lack of staining in their nuclei. Orientation of the tissue is shown by the two axes showing anterior (A) to posterior (P). and left (L) to right (R) sides. Scale bar: 20 jum. Figure 11. Three cells (arrow heads) and many IR-neurites and vari- cosities (thin arrows) are shown in the whole mount of right buccal ganglion. Orientation of the tissue is shown by the axes showing anterior (A) to posterior (P). and left (L) to right (R) sides. Scale bar: 50 /urn. Figure 12. Cross section of the pedal ganglia showing a small IR-cell (large arrow) and many IR-neurites (thin arrows), ao, aorta. Scale bar: 250 fim. ul.. 1992; Schot and Boer, 1982). The lateral lobes of L stagnalis and H. diuyi are known to regulate both growth and reproductive centers of the CNS (for review, see Figure 13. Cross section of the anterior region of the visceral (v) and mid region of the right parietal (p) ganglia showing IR-cells (thin arrows). Scale bar: 50 p.m. Figure 14. Cross section of the mid region of the visceral (v) and posterior region of the right parietal (p) ganglia showing IR-cells (thin arrows). Scale bar: 50 /jm. Figure 15. Whole mount of the cerebral commissure (com) showing extensive IR-neurites and varicosities within the connective in adjacent area (arrows). Orientation of the tissue is shown by the two axes showing anterior (A) to posterior (P). and left (L) to right (R) sides. Scale bar: 50 ^m. OSMOREGULATION AND FMRFAMIDE IN A SNAIL 161 t rival vis about 30 smaller cells have been seen (Murphy et ul., 1985: Saleuddin et al.. 1992). The IR-cells in the pari- etal and visceral ganglia of M. bidentatus are arranged in anterior and posterior groups, which appear to be similar to those in H. duryi and L sttignalis (Schot and Boer. 1982; Bucket! et ul., 1990; Saleuddin et al., 1992). The location of IR-cells in the bassommatophoran pulmonates such as H. duryi and L. sui^nulis is close to or within important endo- crine centers that regulate growth and reproduction, and the lateral lobes of the cerebral ganglia that regulate both growth and reproduction. Numerous IR-ribers emanate from the CNS and terminate as varicosities in the nearby connec- tive tissue, suggesting that the FaRPs are released into the hemolymph that perfuses the CNS. The FaRPs may also participate in regulating other endocrine centers of the CNS. Acknowledgments This study was supported by grants from York University (to HRK). NSERC. Canada (to ASMS), and NIH, USA (to MJG). We thank Dr. B. G. Loughton. York University, for his helpful comments on the manuscript and Ms. Mary-Lou Ashton, York University, for her technical help. Literature Cited Apley. M. L. 1970. Field studies on life history, gonadal cycle and reproductive periodicity in Melampus bidentanis (Pulmonata: Ellobi- idae). Malacologia 10: 381-397. Benjamin, P. R., and J. F. Burke. 1994. Alternative mRNA splicing of the FMRFamide gene and its role in neuropeptidergic signaling in a defined neural network. Bioessays 16: 335-342. Buckett, K. J., M. Peters, G. J. Dockray, J. van Minnen, and P. R. Benjamin. 199(1. Regulation of heartbeat in Lymnaea by motor neu- rons containing FMRFamide-like peptides. J. Ncurophysiol. 63: 1426- 1435. Bulloch, A. G. M., D. A. Price, A. D. Murphy, T. D. Lee, and H. N. Bowes. 1988. FMRFamide peptides in Helisoma: identification and physiological actions at a peripheral synapse. J. Neurosci. 8: 3459- 3469. Cottrell, G. A., E. M. Lutz, D. A. Price, and J. Sommerville. 1994. N-terminally extended FMRFamide related peptides of Helix aspersa: processing of the precursor protein and distribution of the released peptides. Mol. Cell. Neurosci. 5: 632-641. Deaton, L. E. 1990. Potentiation of hypoosmotic cellular volume regu- lation in the quahog. Mercenaria mercenaria. by 5-hydroxytryptamine, FMRFamide. and phorbol esters. Biol. Bull. 178: 260-266. De With, N. D., H. H. Boer, A. B. Smit, and R. C. Van Der Schors. 1994. Neurosecretory yellow cells and hydromineral regulation in the pulmonate freshwater snail Lymnaea stagnalis. Pp. 81-84 in Perspec- tive\ in Comparative Endocrinology. K. G. Davey. R. E. Peter, and S. S. Tobe. eds. National Research Council of Canada. Ottawa. Giardino, N. D., R. S. Aloyz, M. Zollinger, and M. W. Miller. 1996. L5-67 and LUQ- 1 peptide precursors of Aplysia californica: distribu- tion and localization of immunoreactivity in the central nervous system and in peripheral tissues. J. Comp. Neural. 374: 230-245. Green, K. A.. S. VV. P. Falconer, and G. A. Cottrell. 1994. The neuropeptide Phe-Met-Arg-Phe-NH, (FMRFamide) directly gates two ion channels in an identified Helix neurone. Pfliigers Arch. 428: 232- 240. Greenberg, M. J., and I). A. Price. 1992. Relationships among the FMRFamide-like peptides. Pro K . Bruin Kes. 92: 25-37. Hilbish, T. J. 1981. Latitudinal variation in freezing tolerance of Mel- innpiis bidenliiltis (Say) (Gastropoda: Pulmonata). J. E.\p. Mar. Biol. Ecol. 52: 283-297. Khan, H. R., and A. S. M. Saleuddin. 1979. Effects of osmotic changes and neurosecretory extracts on kidney ultrastructure in the freshwater pulmonate Helisoma. Can. ./. Zool. 57: 1256-1270. Koester, J., and A. Alevizos. 1989. Innervalion of the kidney ofAplysiii by L10 and LUQ cells, and an identified peripheral motoneuron. J. Neurosci. 9: 4078-4088. Lehman. H. K., and I). A. Price. 1987. Localization of FMRFamide- like peptides in the snail Helix aspersa. J. Exp. Biol. 131: 37-53. Lesser, W., and M. J. Greenberg. 1993. Cardiac regulation by endog- enous small cardioactive peptides and FMRFamide-related peptides in the snail Helix aspersa. J. Exp. Biol. 178: 205-230. Lingueglia, E., G. Champigny, M. Lazdunski, and P. Barbry. 1995. Cloning of the amiloride sensitive FMRFamide peptide-gated sodium channel. Nature 378: 730-733. Longley. R. D., and A. J. Longley. 1986. Serotonin immunoreactivity of neurons in the gastropod Aplysia californica. J. Neurobiol. 17: 339- 358. Madrid. K. P., D. A. Price. M. J. Greenberg, H. R. Khan, and A. S. M. Saleuddin. 1994. FMRFamide-related peptides from the kidney of the snail, Helisoma trivolvis. Peptides 15: 31-36. McMahon, R. F., and W. D. Russell-Hunter. 1981. The effects of physical variables and acclimation on survival and oxygen consump- tion in the high littoral salt-marsh snail, Melampus bidentanis Say. Biol. Bull. 161: 246-269. Moffett. S. B. 1992. Patterns of neurolransmitter-like immunoreactivity in veligers of the pulmonate gastropod Melampus bidentatus. Am. Zool. 32: 131A. Morton, J. E. 1955. The evolution of Ellobiidae with a discussion on the origin of the Pulmonata. Proc, Zool. Soc. Land. 125: 127-168. Murphy. A. D., K. Lukowiak, and W. K. Stell. 1985. Peptidergic modulation of patterned motor activity in identified neurons of Heli- soma. Proc. Nail. Acad. Sci. USA 82: 7160-7166. Price, C. H. 1977. Morphology and histology of the central nervous system and neurosecretory cells in Melampus bidentatus Say (Gas- tropoda: Pulmonata). Tran.\. Am. Micros. Soc. 96: 295-312. Price, C. H. 1980. Water relations and physiological ecology of the salt marsh snail. Melampus bidentatus Say. J. Exp. Mar. Biol. Ecol. 45: 51-67. Price, D. A., and M. J. Greenberg. 1994. Comparative aspects of FMRFamide gene organization in molluscs. Neth. J. Zool. 44: 421- 431. Price, D. A., N. W. Davies, K. E. Doble, and M. J. Greenberg. 1987a. The variety and distribution of FMRFamide-related peptides in mol- luscs. Zool. Sci. 4: 395-410. Price, D. A., C. G. Cobb, K. E. Doble, J. K. Kline, and M. J. Green- berg. 1987b. Evidence for a novel FMRFamide-related heptapeptide in the pulmonate snail Sip/ionaria pectinata. Peptides 8: 533-538. Price, D. A., W. Lesser, T. D. Lee, K. E. Doble, and M. J. Greenberg. 1990. Seven FMRFamide-related and two SCP-related cardioactive peptides from Helix. J. Exp. Biol. 154: 421-437. Ridgway, R. L. 1987. Alcian blue-alcian yellow mapping of neurose- cretory cells in the central nervous system of the salt marsh pulmonate snail Melampus hijfiitutiis. Comp. Biochem. Physio/. 87A: 295-303. Russell-Hunter, W. D. 1978. Ecology of freshwater pulmonates. Pp. 335-383 in Pulmonates. Vol. 2A. V. Fretter and J. F. Peake. eds. Academic Press. New York. Russell-Hunter, W. D., M. L. Apley, and R. D. Hunter. 1972. Early life-history of Melampus and the significance of semilunar synchrony. Biol. Bull. 129: 623-656. 162 H. R KHAN ET AL .Saleuddin, A. S. M., and M. L. Ashton. 1996. Neuronal pathways of three neurosecretory cells from the lateral lobes in Helisoma (Mollus- ca): innervation of the dorsal body. Tissue Cell 28: 5362. Saleuddin, A. S. M., H. R. Khan, M. L. Ashton, and B. Griffond. 1992. Immunocytochemical localization of FMRFamide in the central ner- vous system and the kidney of Helisoma duryi (Mollusca): its possible antidiuretic role. Tissue Cell 24: 179-189. Saleuddin, A. S. M., S. T. Mukai, and H. R. Khan. 1994. Molluscan endocrine structures associated with the central nervous system. Pp. 257-263 in Perspectives in Compararive Endocrinology, K. G. Davey, R. E. Peter, and S. S. Tobe, eds. National Research Council of Canada. Ottawa. Salzet, M., P. Bulet, C. Wattez, and J. Malecha. 1994. FMRFamide- related peptides in the sex segmental ganglia of the pharyngobdellid leech Erpobdella ocloculata. Identification and involvement in the control of hydric balance. Eur. J. Biochem. 221: 269-275. Schot, L. P. C., and H. H. Boer. 1982. Immunocytochemical demon- stration of peptidergic cells in the pond snail L\mnaea stagnalis with an antiserum to the molluscan cardioactive tetrapeptide FMRFamide. Cell Tissue Res. 225: 347-354. Weiss, K. R., H. Bayley, P. E. Lloyd, R. Tenenbaum, M. A. Gainowicz Kolks, E. C. Cropper, S. C. Rosen, and I. Kupfferman. 1989. Purification and sequencing of neuropeptides contained in neuron R15 of Aplysia califomica. Proc. Nail. Acutl. Sci. USA 86: 2913-2917. Reference: Bioi Bull. 196: 163-176. (April 1999) Behavioral Physiology of Four Crab Species in Low Salinity I. J. MCGAW, C. L. REIBER, AND J. A. GUADAGNOLI Department of Biological Sciences, University of Nevada, Las Vegas, 4505 Maryland Parlovay, Las Vegas, Nevada 89154-4004 Abstract. Reports focusing on the behavioral responses of crabs to exposure to low salinity have involved choice chamber experiments or quantification of changes in activ- ity. In addition to describing changes in locomotor activity in four species of crabs of differing osmoregulatory ability. the present study describes six behaviors: increased move- ment of the mouthparts, cleaning of the mouthparts with the chelae, cleaning of the antennae and antennules with the maxillipeds, flicking of the antennae, retraction of the an- tennules, and extension of the abdomen. Callinectes sapidns and Carcinus maenas are classed as efficient osmoregula- tors, and in general, showed an increase in these behaviors with decreasing salinity. Cancer magister, a weak regulator, and Libinia emarginata, an osmoconformer, exhibited these behaviors to a lesser degree and became inactive in the lower salinities, tending to adopt an isolation-type response. The differences in behaviors between the species correlated closely with previously reported changes in cardiovascular function and hemolymph flow. These overt reactions are discussed in relation to the osmoregulatory physiology and ecology of each crab species. Introduction The osmoregulatory physiology of a number of crusta- cean species has been studied extensively during the past four decades (see Mantel and Farmer, 1983; Pequeux, 1995, for references). There are few reports on behavioral reac- tions to salinity variation, and these have involved either salinity choice experiments or quantification of locomotor activity. The anomuran crab Porcellana platycheles displays dis- Received 16 July 1998; accepted 11 January 1999. Abbreviations: SW = seawater. CW = carapace width. E-mail: imcgaw@ccmail.nevada.edu criminatory behavior in low salinity, but only outside its limits of physiological tolerance. In choice chamber exper- iments, this species shows rapid avoidance of salinities below 40% seawater (SW), but cannot distinguish between pairs of salinities above 40% SW (Davenport, 1972; Dav- enport and Wankowski, 1973). The amphipod Corophium volutator has a preference for salinities in the range of 10-30 ppt (30-90% SW). but only discriminates between pairs of salinities outside this range (McLusky, 1970). Com- parable behavioral reactions are reported for Marinogam- marus marinus, which has a preferred salinity range of 80%-100% SW, although it is able to survive in more dilute concentrations (Bettison and Davenport. 1976). Carcinus maenas, the green shore crab, increases its locomotor ac- tivity in low salinity, a behavior defined as halokinesis (Taylor and Naylor. 1977; Thomas et ai, 1981; Bolt and Naylor, 1985; Ameyaw-Akumfi and Naylor, 1987). Its pre- ferred salinity range, as determined by choice chamber experiments, is 27-41 ppt (82-125% SW; Thomas et al., 1981) or 17-40 ppt (51-121% SW; Ameyaw-Akumfi and Naylor, 1987), and it is able to discriminate between salin- ities separated by a difference as little as 0.5 ppt (McGaw, 1991). In addition, salinity choice behavior in this species is affected by the coloration of the individual and by prior acclimation salinity (McGaw and Naylor, 1992a), as well as by the availability of shelter (McGaw and Naylor, 1992b). Low salinity is also known to re-entrain and modulate the endogenous tidal locomotor rhythm in this species (Bolt and Naylor, 1985; McGaw and Naylor. 1992O. A number of other species can control the osmotic pres- sure of body fluids by behavioral selection of different salinities. The coconut crab Birgus latro (Gross, 1995), the lined shore crab Pachygrapsus crassipes (Gross, 1957), and the hermit crab Pa gurus bernhardits (Davenport et 0.05; Table I). Locomotor activity was more 1200 -, 1000 - CO O 800 - > 600 - | 400 - V) O 200 - - C. sapidus C. maenas C. magister L. emarginata 20 40 60 80 100 Percentage seawater Figure 1. Hemolymph osmolality (mean SEM) of 8 male Callinectes sapidus, Carcinus maenas. Cancer magister. and Libinia emarginata in seawater concentrations of 100%-25%, shown in relation to iso-osmotic line for each seawater concentration. 166 I. J. McGAW ET AL. 6 -i o 8 Carcinus maenas 3 i 1 - J 30 60 90 120 150 180 Time (min) Cancer magister E 3- o o o o o -J 30 60 90 120 150 180 Time (min) Libinii emarginata 30 60 90 120 Time (min) 150 180 Figure 2. Locomotor activity of 16 crabs (mean SEM) during 3-h exposure to seawakr Loncentrations of 100%, 75%, 50%, and 25% sea- water; (a) Carcinus inarmis, (b) Canter magister, and (c) Libitria emar- ginata. pronounced in Carcinus nuieiius. There was a significant increase after 15-30 min in 50% and 25% seawater (Fig. 2a; F = 22.45, P < 0.000), and activity remained elevated above levels in 100% and 75% seawater. for the duration of the experiment. Both Cancer magister (Fig 2b) and Libinia emarginata (Fig. 2c) showed an immediate and significant increase in activity as the salinity was lowered (F = 7.46 and 7.60, P < 0.000). However, the pattern was somewhat different than in Callinectes sapidus and Carcinus maenas. The activity levels of Cancer magister declined during the first hour of low-salinity exposure, and Libinia emarginata was largely inactive after 45 min. Individuals of both spe- cies buried themselves in the gravel and moved infrequently thereafter. Each species also responded to a decrease in salinity with an increase in frequency of mouthpart movements; each set of mouthpart movements (3-8 movements per set) was associated with a ventilatory reversal (not shown). In Cal- linectes sapidus, there was a clear increase in mouthpart movements as the seawater was lowered to 25% (Fig. 3a; F = 23.29, P < 0.000). In 100% seawater this species only occasionally opened its mouthparts, but in 50% and 25% seawater the frequency of mouthpart movements was ele- vated for the 3-h experimental period. Carcinus maenas exhibited a similar behavior pattern, with increasing mouth- part movements in the lower salinities (Fig. 3b; F = 20.1 1. P < 0.000). In 50% and 25% SW there was a significant increase in frequency of mouthpart movements compared to levels in 100% and 75% SW (Table I); this was maintained for the duration of the experiment. In Cancer magister (Fig. 3c), there was a significant increase in mouthpart move- ments in all salinities below 100% seawater (F = 4.55, P < 0.01 ); but unlike Callinectes sapidus and Carcinus maenas, Cancer magister did not maintain the increase. The number of mouthpart movements quickly declined, reaching levels equivalent to those in 100% SW after 30-45 min. In Libinia emarginata (Fig. 3d), a similar trend was observed, with a short-term increase in frequency of mouthpart movements in 25% SW (F = 3.25, P < 0.05), which decreased within an hour to levels comparable to those in 100% SW. The overall trend was a decrease in the frequency of mouthpart movements with decreasing osmoregulatory ability of each crab species (Fig. 1, Fig. 3a-d). The crabs used their chelae to scrape the third maxillipeds and exopodites of the mouthparts; this behavior was only observed during the first hour of low-salinity exposure. Both Callinectes sapidus (Fig. 4a) and Carcinus maenas (Fig. 4b) cleaned their mouthparts only in the lowest salinity tested (25% SW; F = 12.26 and 11.65, P < 0.000). In Callinectes sapidus this behavior stopped after 90 min (Fig. 4a), whereas in Carcinus maenas it was not observed after 45 min (Fig. 4b). Only a small percentage of Cancer ma- gister individuals actually cleaned their mouthparts (Fig. 4c), and the increase in this behavior was significant only in 50% seawater (F = 4.6, P < 0.01; Table I). Libinici emar- ginata (Fig. 4d) showed a significant increase in mouthpart cleanina in 50% and 25% SW (F = 13.37. P < 0.000; Table BEHAVIOR OF CRABS IN LOW SALINITY 167 Table I Student-Newman-Keuls painvise tests for significant differences in behavior of each crab species, beru'een each of the four salinities tested Percentage Seawaler Comparison 100 vs 75 100 vs 50 100 vs 25 75 vs 50 75 vs 25 50 vs 25 I'lillnit'cii"* Mi/nilus Locomotor activity NS NS NS NS NS NS Mouthpart movement NS * * * * * Mouthpart cleaning NS NS * NS * * Antennae cleaning NS * * NS * NS Antennae flicking NS NS NS NS NS NS Antennule retraction * * * NS NS NS Abdomen extension NS NS NS NS NS NS Can-inns maenas Locomotor activity NS * * NS * * Mouthpart movement NS * * * * NS Mouthpart cleaning NS NS * NS * * Antennae cleaning * * * * * NS Antennae flicking * * * NS NS NS Antennule retraction NS * NS NS NS NS Abdomen extension NS NS * NS * * Cancer magister Locomotor activity * * * NS NS NS Mouthpart movement * * * NS NS NS Mouthpart cleaning NS * NS NS NS * Antennae cleaning NS NS NS NS NS NS Antennae flicking NS NS * NS NS NS Antennule retraction * * * * * * Abdomen extension NS NS * NS * * Libiniu enuirginata Locomotor activity * * * NS NS NS Mouthpart movement NS NS * NS NS NS Mouthpart cleaning NS * * * * * Antennae cleaning NS NS NS NS NS NS Antennae flicking NS NS NS NS NS NS Antennule retraction * * * NS NS NS Abdomen extension NS NS * NS * * NS = not significant; * = significant (P < 0.05). I), and as with the other species, this behavior ceased after 45 min. Both Callinectes sapidus (Fig. 5a) and Carcinus maenas (Fig. 5b) cleaned the antennae and antennules with the palps of the third maxillipeds, and this behavior increased with decreasing salinity (F == 10.85 and 11.86, P < 0.000). Callinectes sapidus showed an increase in frequency of cleaning in 50% and 25% SW, whereas Carcinus maenas increased antennae and antennule cleaning in 75% SW, as well as in the two lowest salinities (Table I). Antennae or antennule cleaning essentially did not occur in Cancer ma- gister or Lihiniti emarginata (not shown): only one or two individuals of each species exhibited this behavior, hence the insignificant value (P > 0.05). In Callinectes sapidus the antennae were flicked up and down, on average. 2-3 times per minute, and this did not change with salinity (Table I; F = 2.35. P > 0.05). There was also no significant change in antennae flicking in Li- binia emarginata (F = 0.83, P > 0.05), since this behavior was only observed in two individuals in 25% SW. In Car- cinus maenas there was an increase in antennae flicking in all salinities below 100% SW (F = 7.04, P < 0.000; Table I), which increased in frequency as the salinity decreased. In Cancer magister, there was a similar increase in frequency of antennae flicking; however, rates were significantly ele- vated only in 25% SW (F = 3.04, P < 0.05; Table I). In each case there was a large variation in this behavior be- tween each of the salinities. The antennules of all the crabs were flicked rapidly while orientated in different directions, but were also retracted into the carapace for periods of time. In 100% SW each species retracted the antennae for about 5-20 s of every minute (10%-30%), but distinct differences between the species occurred in the lower salinities (Fig. 6). Callinectes sapidus (Fig. 6a) retracted its antennules upon initial expo- sure to 25% SW, but after 10 min the antennules were rarely 168 I. J. McGAW ET AL 100 -| 80 - 60 - 100% SW A-A-A Callinectes sapidus 7S * SW OO O 50% SW - ~ 25% SW 0) o E 40 - CO Q. | 20 O J 12 n .5 10 - c 8 o E o S 6 E t! 4 CO CL Z 2H J 40 Carcinus maenas 30 60 90 120 150 180 Time (min) Cancer magisler 0) E = 20 - k_ IB O. 10 - o o J 12 1 E i- 4 - IB Q. f = 2 - J 30 60 90 120 Time (min) Libmia emargmata 150 180 60 90 120 Time (min) 150 180 30 60 90 120 Time (min) 150 180 Figure 3. Mouthpurt movements of 16 crabs (mean SEM) during 3-h exposure to seawater concentrations of 100%, 75%, 50%, and 25% seawater; (a) Callinectes sapidus, (b) Carcinus maenas, (c) Cancer magisler, and (d) Libinia emarginata. retracted; in 75% and 50% SW the antennules remained extended for the entire experimental period (F = 24.24. P < 0.000). Although the pattern was similar in Carcinus mae- nas (Fig. 6b), with the antennules exposed for longer peri- ods in all salinities below 100%, this was statistically sig- nificant only in 50% SW (F = 4.48, P < 0.01). In Cancer magister and Libinia emarginata, the opposite response was seen: in low salinities the animals retracted the antennules. Cancer magister (fig. 6c) showed a stepwise and significant increase in antennule retraction (F = 27.32. P < 0.000) in decreasing salinities, with the antennules more-or-less re- tracted for 100% of the time in 25% SW. Libinia einar- ginata (Fig. 6d) also retracted the antennules to a greater degree (60%- 80% of the time) in all salinities below 100% SW (F = 33.88, P < 0.000), but there was no significant difference in retraction times between 75%, 50%, and 25% SW, as occurred in Cancer magisler (Fig. 6c). Crabs extended their abdomens in 50%' and 25% SW only, although this behavior was also observed in some animals when returned to 100% SW in the holding tanks. Callinectes sapidus did not extend the abdomen to any significant degree (F = 1.77, P > 0.05): only three animals were observed to extend the last segment of the abdomen in 25% SW, for short periods of time (not shown). Carcinus maenas extended the entire abdomen during the first hour in 25% SW and to a lesser degree in 50% SW (Fig. 7a), and this was usually accompanied by slow fanning movements of the abdomen; however, this was significantly different from control levels only in 25% SW (F == 12.03, P < 0.000). Cancer magister (Fig. 7b) also extended the abdo- men in 50% and 25% SW, although the time course for this behavior was more erratic. Again, this behavior was only significantly different from the control in 25% SW (F = 9.95, P < 0.000). Abdomen extension increased steadily (F = 12.35. P < 0.000) in Libinia einarginata (Fig 7c) after 1 h exposure to 25% SW. This appeared to be a passive Bl-HAVIOR OF CRABS IN LOW SALINITY 169 6 -i VJ , <- 4 - hi n Q. o E ju O 2 - J Calltnectes safiidus 1.5 re o. a o E ,0.5 - nj JH O J 30 60 90 120 Time (min) Cancer magtsler 150 180 -i-- . 60 90 120 Time (min) 150 180 O Carcmus maenas 4 i - 3, o. 1 2 H o J 60 90 120 Time (min) emarginata 150 180 30 60 90 120 Time (min) Figure 4. Cleaning of the mouthparts in 16 crabs (mean SEM) during 3-h exposure to 100%, 75%. 50%, and 25% seawater: (a) Callinecles sapitlus. (b) Curcinus maenas, (c) Cancer magister, and (d) Libinia emarginata. process, rather than the extension and fanning of the abdo- men seen in the other species. Discussion Callinectes sapidus hyper-regulated its body fluids in all salinities, including 100% SW (Fig. 1); this has been re- ported previously (Tan and Van Engel, 1966). In all the other species, hemolymph was iso-osmotic with 100% SW. Carcimts maenas, which is classed as an efficient osmoreg- ulator, had hemolymph osmolality levels (Fig. 1 ) similar to previous reports (Lucu et ai, 1973; Rankin and Davenport. 1981). All Carcinus used in the present study were green- colored male crabs. Red-colored individuals, which are in a prolonged intermolt (McGaw et ai, 1992), are poorer os- moregulators (Reid et ai. 1989: McGaw. 1991) and have different behavioral responses to low salinity (McGaw and Naylor. 1992a. c). Cancer magister is classed as a weak osmoregulator (Jones, 1941) and the levels of hemolymph osmolality reported here (Fig. 1 ) agree closely with the data of Hunter and Rudy (1975). Cancer magister is able to survive in salinities as low as 12 ppt (36% SW; Cleaver, 1957); however, in the present study all survived short-term exposure (24 h) to 25% SW (8 ppt) (Fig. 1). Libinia emar- ginata is an osmoconformer, with ion levels closely follow- ing those of the external medium (Gilles. 1970). This was also seen in the present study (Fig. 1 ), except in the lowest salinity. Libinia emarginata can withstand dilution of the medium only to 40% SW (Gilles, 1970). Mortality was high after 12-h exposure to 25% SW, and the short acclimation time (10 h) used would not have allowed hemolymph os- molality to decline to stable levels (Fig. 1). Most reports on the activity of crabs in low salinity pertain to Carcinus. Taylor and Naylor (1977) report that Carcinus maenas responds to a lowering of salinity with an increase in locomotor activity, defined as halokinesis. This has been confirmed by a number of other studies (Taylor et 170 I. J. McGAW ET AL T = 7- E a Callmectas sapidus 75% P W O O O 6 - 0) 1 5 " Q) 1 1 50X SW - 1 4 " ; I CD flj 3 - c i ^t , ...... c d> c 2 - o> C 1 ~ i \, i '1 1 *1 "--i ' M / ' '' '.. 1 ) 1 i' \ ^ --4 7"V.... i 'E n s ^ r- "' ^_ -* * s Vi - -^""^-ii _ = c - J 30 60 90 120 Time (min) Cancer magister 4... 150 180 100 - 80 - .2 60 -I o IB I 1 1 1 30 60 90 120 Time (min) o J 60 90 120 Time (min) Libinia emarginata 150 180 150 180 30 60 90 120 Time (min) 150 180 Figure 6. Percentage time of antennule retraction of 16 crabs (mean SEM) during 3 h in seawater concentrations ranging from 100%-25%: (a) Callinectes saniiliis. (b) Carcinus maenas, (c) Cancer magister. and (d) Libinia emarginata. 1964, 1966). Evidence suggests that the posterior gills have the highest Na-K-ATPase activity (Florkin and Schoffe- niels, 1969; Neufeld et ai. 1980: Siebers et al., 1982, 1983, 1985, 1986). and therefore a ventilatory reversal would bring water into contact with the pumps of the posterior gills, enhancing active ion uptake. In support of this con- cept. Callinectes sapidus, which is the most efficient osmo- regulator tested, exhibits the highest frequency of mouthpart movements (and hence ventilatory reversals: Fig. 3a), and this behavior decreases with the declining osmoregulatory ability of the species (Fig. 3), with the weak osmoregulator Cancer magister (Fig. 3c) and the stenohaline Libinia emar- ginata (Fig. 3d) showing substantially less ventilatory re- versals than the two efficient osmoregulators (Fig. 3a. b). In addition, both Callinectes sapidus (Fig. 3a) and Carcinus maenas (Fig. 3b) showed a stepwise increase in mouthpart movements with decreasing salinity; this increase was sus- tained in 507r and 257c SW. for the 3-h experimental period. In both Cancer magister and Libinia emarginata (Fig. 3c, d: Table I), mouthpart movements increased only during the first 30 min. Thereafter, the crabs kept the mouth- parts sealed, isolating the branchial chamber. This sealing of the branchial chambers has been reported previously for Cancer magister (Sugarman et at.. 1980). Isolation of the branchial chambers, in conjunction with a decrease in gill blood flow via a reduced cardiac output (McGaw and Mc- Mahon, 1996: Cornell. 1973. 1974). would help reduce the gradient for water uptake and diffusive salt loss. There are also differences in hemolymph flow to the muscles of the mouthparts, which are supplied by the sternal artery and branches of the anterolateral arteries (Pearson. 1908: McLaughlin. 1983). In Callinectes sapidus, hemo- lymph flow through the sternal artery and anterolateral arteries is elevated for 2-4 h in 25% SW (McGaw and Reiber, 1998), which corresponds to the period of increased mouthpart movement (Fig. 3a). In contrast, blood flow through the sternal artery and anterolateral arteries of Can- cer magister decreases (McGaw and McMahon, 1996), and 172 I. J. McGAW ET AL 30 - a Carcinus maenas 25 - 100% SW A-A-A 20 - Try fill f\ O O 15 - , 50% SW - 25% SW ^7~ K7 V7 10 - :l i 1 i t> 5 - ,'; j ' i . A 3 60 90 120 150 180 15 - 10 - 5 - J 60 -i * SO- 40 - 30 - LI 20 H . 1976. Osmotic and ionic regulation in the Dungeness crab Cancer magister (Dana). Comp. Biochem. Phvsiol. 51A: 439-447. Jones, L. L. 1941. Osmotic regulation in several crabs of the Pacific coast of North America. J. Cell. Coinp. Physiol. 18: 79-91. Kamemoto, F. I. 1976. Neuroendocrinology of osmoregulation in deca- pod Crustacea. Am. Zoo/. 6: 1 4 1 - 1 50. Krijgsman, B. J., and N. E. Krijgsman. 1954. Osmorezeption in Jasus lalamlii. Z. Vgl. Physiol. 37: 78-81. Lagerspetz, K., and M. Mattila. 1961. Salinity reactions of some fresh and brackish water crustaceans. Biol. Bull. 120: 44-53. Lucu. C., D. Siebers, and K. R. Sperling. 1973. Comparison of osmo- regulation between Adriatic and North Sea Carcinus. Mar. Biol. 22: 85-95. Lynch, M. P.. K. L. Webb, and W. A. Van Engel. 1973. Variations in serum constituents of the blue crab Callinectes sapidus. Chloride and osmotic concentration. Comp. Biochem. Phvsiol. 44A: 719-734. Malley, D. F. 1977. Salt and water balance of the spiny lobster. Panuliris argus: the role of the gut. 7. Exp. Biol. 70: 231-245. Mantel, L. H.. and L. L. Farmer. 1983. Osmotic and ionic regulation. Pp. 54-142 in Biology of Crustacea. Vol. 5. internal Anatomy and Physiological Regulation. L. H. Mantel, ed. Academic Press. New York. Mangum, C. P., and L. M. Amende. 1972. Blood osmotic concentra- tion of blue crabs (Callinectes sapidus} found in freshwater. Chesa- peake Sci. 13: 318-320. McGaw, I. J. 1991. Behavioural responses of the shore crab Carcinus maenas to salinity variation. Ph.D. Thesis. University of Wales. McGaw, I. J., and B. R. McMahon. 1995. The FMRFamide-related peptides Fl and F2 alter hemolymph distribution and cardiac output in the crab Cancer magister. Biol. Bull. 188: 186-196. McGaw, I. J., and B. R. McMahon. 1996. Cardiovascular responses resulting from variation in external salinity in the Dungeness crab Cancer magister. Physio/. Zoo/. 69: 1384-1401. McGaw, I. J.. and B. R. McMahon. 1998. Endogenous rhythms of haemolymph flow and cardiac performance in the crab Cancer magis- ter. J. Exp. Mar. Biol. Ecol. 224: 127-142. McGaw, L J., and B. R. McMahon. 1999. Actions of putative cardioin- hibitory substances on the in vivo decapod crustacean cardiovascular system. J. Crust. Biol. (in press). McGaw, I. J., and E. Naylor. 1992a. Salinity preference of the shore crab Carcinus maenas in relation to coloration during intermoult and to prior acclimation. 7. Exp. Mar. Biol. Ecol. 155: 145-159. McGaw, I. J., and E. Naylor. 1992b. The effect of shelter on salinity preference behaviour of the shore crab Carcinus maenas. Mar. Behav. Physiol. 21: 145-152. McGaw, I. J., and E. Naylor. 1992c. Distribution and rhythmic loco- motor patterns of estuarine and open shore populations of Carcinus maenas. J. Mar. Biol. Assoc. UK 72: 599-609. McGaw, I. J., and C. L. Reiber. 1998. Circulatory modification in the blue crab Callinectes sapidus. during exposure and acclimation to low salinity. Comp. Biochem. Physiol. 121A: 67-76. McGaw, I. J., M. J. Kaiser, E. Naylor, and R. N. Hughes. 1992. Intra-specific morphological variation related to the moult-cycle in colour forms of the shore crab Carcinus maenas. 7. Zoo/. lLoml.1 228: 351-359. McGaw, I. J., C. N. Airriess, and B. R. McMahon. 1994a. Peptidergic modulation ot cardiovascular dynamics in the Dungeness crab Cancer magister. J. Comp. Physiol. 164: 103-1 1 1. McGaw, I. J., C. N. Airriess, and B. R. McMahon. 1994b. Patterns of haemolymph flow variation in decapod crustaceans. Mar. Biol. 121: 53-60. McGaw, I. J., J. L. Wilkens, B. R. McMahon, and C. N. Airriess. 1995. Crustacean cardioexcitatory peptides may inhibit the heart in vivo. J. Exp. Biol. 198: 2547-2550. McLaughlin, P. A. 1983. Internal anatomy. Pp. 1-41 in Biology of the Crustacea. Vol. 5. Internal Anatomy and Physiological Regulation. L. H. Mantel, ed. Academic Press, New York. McLusky, D. S. 1970. Salinity preference in Corophium volutator. J. Mar. Biol. Assoc. UK 50: 747-752. Neufeld, G. J., C. VV. Holliday, and J. B. Pritchard. 1980. Salinity adaptation of gill Na.K-ATPase in the blue crab, Callinectes sapidus. J. Exp. Zoo/. 211: 215-224. Pearson, J. 1908. Cancer (the edible crab). Proc. Trans. Liverpool Biol. Soc. 22: 291-499. Pequeux, A. 1995. Osmotic regulation in crustaceans. J. Crustac. Biol. 15: 1-60. Phillips. J. E., J. Hanrahan. M. Chamberlin. and R. B. Thomson. 1986. Mechanisms of control of reabsorption in insect hindgut. Adv. Insect Physiol. 19: 329-422. Rankin, J. C., and J. Davenport. 1981. Animal Osmoregulation: Ter- tian Level Biology. Blackie and Son. Glasgow. Rasmussen, A. D., and P. Bjerregaard. 1995. The effect of salinity and calcium concentration on the apparent water permeability of Cherax destructor, Astacus astacus and Carcinus maenas (Decapoda, Crusta- cea). Comp. Biochem. Physiol. 111A: 171-175. Reid, D. G., P. Abello, I. J. McGaw, and E. Naylor. 1989. Phenotypic variation in sympatric crab populations. Pp. 89-96 in Phenotypic Responses and Individuality in Aquatic Ectotherms, J. C. Aldrich. ed. JAPAGA Ashford, WickJow, Ireland. Schmidt. M. 1989. The hair-peg organs of the shore crab, Carcinus maenas (Crustacea. Decapoda): ultrastructure and function properties of sensilla sensitive to changes in seawater concentration. Cell Tissue Res. 257: 609-621. Shelton, R. G., and M. S. Laverack. 1970. Receptor hair structure and function in the lobster Homarus gammurus (L.). 7. Exp. Mar. Biol. Ecol. 4: 201-210. Siebers, D., K. Leweck, H. Markus, and A. Winkler. 1982. Sodium regulation in the shore crab Carcinus maenas as related to ambient salinity. Mar. Biol. 69: 37-43. Siebers, D., A. Winkler, K. Leweck, and A. Madian. 1983. Regulation of sodium in the shore crab Carcinus maenas. adapted to environments of constant and changing salinities. He/go/. Meeresunters. 36: 303-312. Siebers, D., A. Winkler, C. Lucu, G. Thiedens, and D. Weichert. 1985. Na-K-ATPase generates an active transport potential in the gills of the hyperregulating shore crab Carcinus maenas. Mar. Biol. 87: 185-192. Siebers, D., C. Lucu, A. Winkler. L. Dalla Venezia, and H. W ille. 1986. Active uptake of sodium in the gills of the hyperregulatini: shore crab Carcinus maenas. Helgol. Meersunters. 40: 151-160. Spaargaren, D. H. 1974. Measurements of the relative rate of blood flow in the shore crab Carcinus maenas at different temperatures and salin- ities. Neth. J. Sea Res 8: 398-406. Spaargaren, D. H. 1982. Cardiac output in the crab. Carcinus maenas 176 I. J. McGAW ET AL. (L.) in relation to solute exchange and osmotic pressure. Mar. Biol Lett. 3: 231-2411. Sugarman, P. C., W. H. Pearson, and D. L. Woodruff. 1980. Detec- tion of thresholds and behavioral responses to salinity changes by the Dungeness crab Cancer magister. Am. Zoo/. 20: 922. Tan, E. (.'., and W. A. Van Engel. 1966. Osmoregulation in the adult blue crab Callinectes sapidus, Rathbun. Chesapeake Sci. 7: 30-35. Taylor, A. C. 1977. Respiratory responses of Carcinus maenas to changes in environmental salinity. / Exp. Mar. Biol. Ecol. 29: 197- 210. Taylor, A. C., and E. Naylor. 1977. Entrainment of the locomotor rhythm of Carcinus maenas by cycles of salinity change. J. Mar. Biol. Assoc. UK 57: 273-277. Taylor, E. W., P. J. Butler, and A. Al-Wassia. 1977. The effect of a decrease in salinity on respiration, osmoregulation and activity in the shore crab Carcinus maenas at different acclimation temperatures. J. Com/;. Physiol. 119: 155-170. Tazaki. K. 1975. Sensory units responsive to osmotic stimuli in the antennae of the spiny lobster. Panuliris japonicus. Comp. Biochem. Physiol. 51A: 647-653. Thomas, N. J., T. A. Lasiak. and E. Naylor. 1981. Salinity preference behaviour in Carcinus. Mar. Behav. Physiol. 7: 277-282. Van-Weel, P. B., and J. P. Christofferson, 1966. Electrophysiological studies on various stimuli in the antennulae of certain crabs. Phvsiol. Tool. 39: 317-325. Warrnan, C. G., P. Abello, and E. Naylor. 1991. Behavioural responses of Carcinus mediterraneus Czemaivsky. 1884. to changes in salinity. Sci. Mar. 55: 637-643. Reference: Biol. Bull. 196: 177-186. (April Protein Metabolism in Lecithotrophic Larvae (Gastropoda: Haliotis rufescens) JAY VAVRA AND DONAL T. MANAHAN* Department of Biological Sciences, University of Southern California, Los Angeles, California 90089-0371 Abstract. Rates of protein depletion, synthesis, and turn- over were measured in larvae of the abalone Haliotis rufe- scens as an approach to understanding macromolecular me- tabolism during lecithotrophic development. Protein content decreased linearly during development to metamorphic competence, with 34% of the initial protein in eggs depleted during the 8-day larval life span. Fractional rates of protein synthesis (percentage of total body-protein synthesized per day) decreased during development, from 40% (1-day-old trochophore larva) to 14% (7-day-old veliger larva). Sepa- ration of proteins by one-dimensional gel electrophoresis showed that protein pools in larvae are dominated by two high-molecular-weight protein classes (88 and 121 kDa). When the proteins of 1- and 3-day-old larvae were labeled with a mixture of 3S S-methionine and cysteine. the pattern on two-dimensional gels showed that the turnover process (protein synthesis and degradation) involved hundreds of different proteins. The energy gained from loss of protein could account for 20% of the protein turnover rates for trochophore larvae and 79% of the lower turnover costs for late-stage veligers. Lecithotrophic larvae of H. rufescens maintained high biosynthetic activities, with up to 40% of their whole-body protein being turned over each day. Such dynamic processes during development of nonfeeding lar- vae would contribute significantly to maintenance metabo- lism. Introduction Nonfeeding (lecithotrophic) larval forms are widely rep- resented in the life cycles of many marine invertebrate taxa (Thorson. 1950; Strathmann, 1978; Pearse, 1994). Never- Received 28 May 1998: accepted 6 January 1999. * To whom correspondence should he addressed. E-mail: manahan@ usc.edu theless. our understanding of metabolic processes during lecithotrophic larval development is very limited compared to what is known about metabolism and the dynamics of biosynthetic processes in species with planktotrophic devel- opment (e.g., protein synthesis in sea urchin embryos: Goustin and Wilt, 1981; Bedard and Brandhorst. 1983). Most studies of the biochemistry and physiology of lecitho- trophic larvae have focused on the composition of energy reserves in the early stages of development (Turner and Rutherford. 1976; Jaeckle and Manahan, 1989a; Anger, 1996). The importance of maternally endowed energy sup- ply to larval life span has also been studied (Shilling and Manahan, 1994; Ben-David-Zaslow and Benayahu. 1998), as have respiration rates and amino acid transport (Jaeckle and Manahan. 1989a. b; Shilling et ai, 1996; Hoegh- Guldberg and Emlet, 1997). The aim of this study is to quantify the energetics of protein degradation and synthesis (turnover) in a lecithotro- phic larval form. Protein metabolism was chosen as the focus of this study because protein, not lipid. has been reported (Jaeckle and Manahan, 1989a) to be the major endogenous energy reserve utilized by larvae of Haliotis rufescens (red abalone), a species with lecithotrophic devel- opment. Materials and Methods Larval culture Adult abalone (Haliotis rufescens) were spawned and fertilized at a commercial hatchery ( Ab Lab, Port Hueneme, California). All cultures were started using gametes from 1 male and 2-3 females (different adults were used for each culture). Zygotes and larvae were maintained at 14-15C in unstirred, UV-irradiated seawater that had been passed through a 5-jum (pore size) filter; the larvae were reared on 177 178 J. VAVRA AND D. T. MANAHAN 80-/u,m mesh screens. Under these culturing conditions, it took 6-8 days for larvae to reach metamorphic competence, as determined by ability to undergo metamorphosis when induced with 7-aminobutyric acid (GAB A). Change in total protein Throughout development, samples (Culture 1. n = 4-5 independent samples; Culture 2, H = 3 independent sam- ples) containing 50-75 individuals were taken for measure- ment of protein content (Bradford assay, 1976; as modified by Jaeckle and Manahan, 1989a). The rate of protein loss during development was calculated from the linear regres- sion of total protein per individual with time, during the period from egg to metamorphic competence. Protein con- tent of eggs (day zero) was determined only for Culture I. Rates of ami no acid transport To determine absolute rates of protein synthesis, rates of umino acid transport by larvae were measured during the 14 C-labeling experiments (see Manahan, 1983, for meth- ods). All physiological rates were measured at 15 0.1C. Each transport assay used 5000 larvae in 10 ml of filtered (pore size 0.2 /nm) seawaterto which was added 1 /zCi ml ' of M C-(U)-glycine (New England Nuclear; 110 juCi /amor 1 ). The rate of isotope transport was measured (500 IJL\ sample every 2-3 min) during a 16-18-min exposure of larvae to 9 ^M glycine in seawater. Tau Rules of protein synthesis A modified radioisotope-labeling protocol, based on that of Fry and Gross ( 1970) for sea urchin embryos, was used to measure absolute rates of protein synthesis in abalone larvae. Our modifications included trace labeling over a short time interval (<18 min) to minimize possible effects of perturbations, caused by high transport rates of exoge- nous 14 C-labeled amino acid, of the free amino acid pool. A short exposure also avoided the interconversion of 14 C- glycine, the tracer used in our measurements, into other amino acids (confirmed by analysis with high-performance liquid chromatgraphy, HPLC). Short experiments also min- imized any effects from protein degradation (i.e., reutiliza- tion of 14 C-labeled amino acids due to protein breakdown). Glycine was used as the tracer because it is transported at a high rate and is a measurable fraction of the free amino acid pool in larvae of H. ntfescens (Fig. 1 ). These characteristics allowed for accurate HPLC measurement of the change, with time, in the specific activity of glycine in the free amino acid pools of larvae (Fig. 2 A). Also, glycine in acid-hydrolyzed, whole-body protein extracts of H. rufe- scens larvae (Table I) could be measured with HPLC; this value is required for calculation of absolute rates of protein synthesis (see below). 1 03 H R M 10 20 30 Elution time (minutes) B 20 -i "o o 5. S ~ 15 - o 2 E if 3= CD C "5 P c S 'o _>, O 10 - 5- 02468 Age (days) Figure 1. Glycine in the free amino acid pool of larvae of Haliotis rufescens. (A) Chromatogram showing the separation of amino acids extracted from 3-day-old veliger larvae by reverse-phase high-performance liquid chromatography: D, aspartic acid; E, glutamic acid; N, asparagine; S. serine; H. histidine; G. glycine; R. arginine; Tau, taurine; A. alanine; Y. tyrosine; M, methionine; F, phenylalanine; I, isoleucine; L, leucine; K, lysine (unknown peaks are unlabeled). (B) Change in the amount of glycine in free amino acid pools through larval development. Error bars repre- sent 1 SE of the mean (n = 6 for each stage). The equation for the linear regression is pmol glycine larva" 1 = 1.51 x + 5.16; where .v = age in days (r = 0.50; n = 48). Rates of protein synthesis were determined by measuring the rate of incorporation of l4 C-glycine into protein [defined as the 5% trichloroacetic acid (TC A (-insoluble fraction of larval homogenates, shell included]. For measurements from the free amino acid pools, the intracellular contents of larvae were extracted overnight into 707r ethanol. The spe- cific activity of glycine was then determined with HPLC (Welborn and Manahan, 1995) for each larval stage studied. After detection of fluorescence (i.e., moles of glycine: Fig. 1 A, B), post-column samples of eluent were collected every 30 s (LKB fraction collector) and mixed with scintillation cocktail to determine the amount of radioactivity in the PROTEIN METABOLISM IN ABALONE LARVAE 179 co O lOOOi .c w II w c c ^ O 'm 600 O) Q. 400 4 8 12 Time (minutes) 16 Figure 2. Determination of the absolute rate of protein synthesis in 7-day-oId veliger larvae of fialiotis rufescens. (A) Specific activity of intracellular glycine in the free amino acid pool following transport of ' 4 C-glycine. Each point represents the specific activity determined by high-performance liquid chromatography using fluorescence detection (moles of glycine) and measuremenl by liquid scintillation counting of radioactivity in the glycine peak collected from the chromatographic eluent. (B) Radioactivity in the irichloroacetic acid (TCA)-insoluble frac- tion (protein). (C) Rate of protein synthesis after correcting the TCA- msoluble traction for the change in intracellular specific activity of L4 C- glycine. the mole percent of glycine in protein, and the average mole- percent corrected molecular weight of amino acids in larval protein (Table I). glycine peak. All measurements of radioactivity (as counts per min. CPM) were corrected for quenching and converted to total disintegrations per min (DPM). By correcting for the change in the specific activity of glycine in the free amino acid pool (Fig. 2 A), the incorporation rate of ' 4 C-glycine (Fig. 2B) could be converted to the total amount of glycine incorporated into protein (both I2 C- and 14 C-glycine). The value for the incorporation of total glycine into protein was then converted to an absolute rate of protein synthesis (Fig. 2C) by determining the mole-percent of glycine in larval protein and the mole-percent-corrected molecular weight of all amino acids in larval protein of H. rufescens (Table I). The components required for the calculation of the absolute rate of protein synthesis are MW _ where Jt 5 is the rate of protein synthesis, M W /t is the mole- percent-corrected molecular weight of amino acids in larval protein, S,,, is the mole-fraction of glycine in protein. S f , is the amount of radioactivity in protein, and S /M , is the spe- cific activity of glycine in the free amino acid pool, which changed during the time course of exposure to isotope (Fig. 2 A). The above equation was solved for each sampling time interval during which the incorporation of C-glycine into protein was measured (Fig. 2B|. Patterns of protein synthesis For the analysis of the electrophoretic patterns of protein synthesis, each labeling experiment used 10.000-15,000 larvae in 14 ml of filtered seawater with 200 /u.Ci of 35 S- methionine/cysteine (1 100-1200 Ci mmo!"'. New England Nuclear). Non-radiolabeled methionine was added to sea- water (final concentration of 500 nM) to increase substrate concentration and, hence, transport rates. After a 3-h incu- bation, unincorporated isotope was removed with three suc- cessive seawater washes after the larvae were pelleted in 15-ml conical tubes with a hand centrifuge. Protein samples for electrophoresis were prepared by ultrasonicating em- bryos or larvae in 50 mM Tris-HCl at pH 7.2. Samples were kept on ice during this process to prevent any rise in temperature. Homogenates were centrifuged to remove lar- val shell (30 min at 15.000 X g, 4C) prior to electrophore- sis. The amount of ?5 S-methionine/cysteine incorporated into protein was determined by TCA (5%) precipitation of 10-/J.1 aliquots (n = 2) of the homogenate. Tissue solubilizer (0.5 ml. Scintigest. Fisher Co.) was added to these precip- itates before the radioactivity was counted. These data were required for loading of equal amounts of radioactivity on each gel to permit comparisons of different larval stages (see Results). The amount of protein loaded on each gel was also determined. 180 J. VAVRA AND D. T. MANAHAN Table I Atninit ticiil composition (mole-percent) oj proteins in larval stages of Haliotis rutescens Amino acid Age (days after fertilization) Mean ( SEM) Glutamic acid & Glutamine* 9.9 10.0/10.6 8.6 10.5/10.5 10.7/10.5 10.5 10.2(0.2) Alanine 8.4 8.3/8.0 9.0 9.1/8.) 8.1/8.3 8.4 8.4(0.1) Glycine 7.0 7.5/7.7 9.9 8.2/8.1 8.0/7.8 8.1 8.0(0.2) Aspartic acid & Asparagine* 7.4 7.2/7.3 7.9 6.3/6.4 7.8/7.9 7.6 7.3 (0.2) Lysine 6.5 6.7/6.5 6.2 6.9/8.1 6.8/7.7 7.4 7.0(0.2) Serine 7.0 7.2/7.1 6.3 7.2/7.1 6.9/6.7 7.1 7.0(0.1) Leucine 6.8 6.8/6.7 6.8 6.6/7.4 6.6/6.8 6.2 6.8(0.1) Valine 6.4 6.9/6.9 6.3 7.2/7.0 6.3/6.2 5.7 6.5 (0.2) Isoleucine 6.8 6.8/6.7 6.8 3.9/7.4 6.6/6.8 6.2 6.5 (0.3) Threonine 6.4 5.9/7.0 6.2 6.3/5.4 5.8/5.5 5.6 6.0(0.2) Arginine 6.0 6.1/5.7 5.8 6.5/5.8 5.8/5.7 6.2 6.0(0.1) Proline 5.9 6.4/5.8 5.2 5.1/5.8 5.0/4.8 4.9 5.4(0.2) Tyrosine 4.7 4.0/3.9 5.0 4.0/3.3 4.7/4.6 4.4 4.3 (0.2) Phenylalanine 3.9 4.0/3.8 3.9 4.0/4.7 4.2/4.2 4.0 4.1 (0.1) Methionine 3.6 2.8/3.3 3.3 6.1/2.7 3.5/3.3 4.4 3.7(0.3) Histidine 2.4 2.6/2.3 1.9 1.7/1.6 2.5/2.4 2.2 2.2(0.1) Cysteine 0.8 0.9/0.8 0.8 0.3/0.7 0.8/0.7 1.2 0.8(0.1) Mole-percent 135.67 134.9/ 136.2/ corrected A/W r t 1 36. 1 135.1 133.7 135.0 1 36.5 136.8 135.5(0.3) All values are calculated as mole percents: where two values are separated by a slash (e.g., 10.0/10.6), this represents two replicate chromatographic analyses for that sample. * During acid hydrolysis, asparagine and glutamine form aspartic acid and glutamic acid, respectively. t Mole-percent corrected molecular weights (A/W ) represent the average molecular weight of amino acids in larval protein determined by multiplying each mole percent value by the appropriate molecular weight (e.g.. glycine: 0.08 * 75.1 g mol ~'l, and then taking the sum of all of the mole-percent corrected molecular weights of each amino acid. Protein extracts of larvae were separated with one- and two-dimensional polyacrylamide gel electrophoresis (PAGE). One-dimensional sodium dodecyl sulfate (SDS)- PAGE was conducted according to Laemlli (1970). with equal amounts of radiolabeled protein (1 X 10 6 DPM) loaded per lane, totaling about 30 /xg protein. The proteins resolved by electrophoresis were stained with Coomassie brilliant blue (Biorad) and were exposed to X-ray film (XOMAT AR, Eastman Kodak) for 24 h. Two-dimensional gel electrophoresis was conducted with pre-cast gels on the Multiphor system (LKB-Pharmacia) following the manu- facturer's protocol for the first dimension Immobiline Dry Strip Kit and the second dimension ExcelGel SDS (LKB- Pharmacia: publication 1X-1038-63, edition AA). For two- dimensional gels, samp;, s were loaded with equal amounts of radiolabeled protein (3 X 10 6 DPM), totaling about 100 jug protein per loading. The two-dimensional gels of the different larval stages, each loaded with equivalent amounts of radioactivity, were exposed to X-ray film for an identical time period (8 days). Results Change in total protein There was a linear decrease in protein during develop- ment of Haliotis rufescens (Fig. 3). The rate of protein loss was not statistically different between Culture 1 (23.7 5.7 ng day~' [ 1 SE of the slope]) and Culture 2 (22.5 5.7 ng day '); ANOVA: F\\^,\ for comparison of slopes = 0.1 7"^ and for comparison of y-intercepts = 0.17 ns . All data were combined to calculate a rate of protein loss from egg to metamorphic competence (nanograms of protein per larva = 21.2 .v + 498.6. where x = age in days; SE of slope = 3.52; n = 68). By the end of larval development, 8-day-old larvae still had 66% of their initial (egg) protein content (from regression analysis. Table II). Glvcine transport Larvae had higher rates of glycine transport near meta- morphosis (Culture 1: day 6 = 9.6 pmol glycine larva"' h" 1 ; Culture 2: day 8 = 9.3 pmol glycine larva" 1 h"'; Fig. PROTEIN MF.TABOLISM IN ABALONE LARVAH 181 Age (days) Figure 3. Total protein content during development of Haliotis rufe- scens. Data from two cultures are shown: Culture I (n = 4-5 independent samples per data point: O) and Culture 2 (n = 3 independent samples per data point; ). Protein content of eggs was determined only for Culture 1. Rates of protein loss were not statistically different between cultures (see text), and all data were combined to calculate the rate of protein loss during development: ng protein individual" 1 = 21. 2 .v + 498.6. where .v = age in days (SE of slope = 3.52. n = 68). Error bars are 1 SEM. 4). Combining all glycine transport data for both cultures revealed that transport rates were significantly higher at metamorphic competence (based on larvae before and after day 6: / |24| = 2.81. P < 0.005). Table II Protein synthesis and fractional rates of protein turnover during lamil development of Haliotis rufescens Fractional rates of protein turnover Protein synthesis rate (% total protein Age Total protein (ng protein synthesized synthesized (days) (ng larva" 1 ) larva"' day"') day"') 1 477 193 40 2 456 150 33 3 435 115 26 4 414 87 21 5 393 67 17 6 371 54 15 7 350 49 14 Total protein amounts (nanograms of protein per larva) were calculated from the least-squares linear regression of the change in protein content calculated for two different cultures during development (Fig. 3: ng protein larva"' -21.2 x + 498.6. where .r = age in days). Changes in the absolute rate of protein synthesis during development were calculated from the second-order polynomial (Fig. 5: ng protein synthesized day"' = 3.76 x 2 - 54.09.V + 243.52, where .r = age in days). Fractional rates of protein turnover (percentage of whole-body protein synthesized per larva per day) were calculated from the ratio of the rate of protein synthesis to total protein content at the corresponding stage of development. 10- ^ 8- O O lycine transport glycine (larva)" 1 (I * a> o o O O A Ia- o- c Metamorphic competence I 2 4 6 Age (days) 8 B -" 8- !*- CD :=, ' * A * 0) C 'o >. 4 " -> cn : . la- Metamorphic competence c I 2 4 6 8 Age (days) Figure 4. Transport of ' 4 C-glycine (9 \iM~t from seawater by larvae of Haliotis rufescens. Each data point represents the rate of transport calcu- lated from a regression of a separate time-course experiment (4-5 samples per experiment, with about 250 larvae per sample; all assays had r values > 0.96). Data are shown for two cultures: Culture 1 (A) and Culture 2 (B). Arrows show the time when metamorphic competency was reached for larvae from each culture. Glycine content in free amino acid pools and protein The amount of glycine in the free amino acid pool in- creased significantly during development (ANOVA: F [146] = 45.43. P < 0.0005; Fig. IB). The linear increase in glycine during larval development is represented by the following equation: picomoles of glycine per larva = 1.51 A + 5.16. where .v = age in days (r = 0.50; n = 48). Over the larval life span, glycine only represented about 1% of the total free amino acid pool (see Fig. 1A), with taurine being the dominant amino acid. The mole-percent of gly- cine in protein did not change during development (from glycine data in Table I: ANOVA of the regression slope of protein-glycine over time was not significantly different from zero, F |K7| == 0.44" s ). The mean mole-percent of 182 J. VAVRA AND D. T. MANAHAN I 1.1 l-g 100 H O , in larvae of Huliolis rufescens. (A) Coomassie-stained protein pools. (B) Corresponding autorudhr.i.ii, -I proteins labeled with 15 S-methionine/cysteine (i.e., newly synthesized proteins). Each lane contains a dilt. rein larval stage (day 1. 2, 3. 5, and 7) and equal radioactivity ( 1 X 10" DPM loaded per lane). Bunds marked P, and P, are high-molecular-weight proteins present at each stage of larval development as shown in the Coomassie-stained gel (A) and are not radiolabeled in the autoradiogram (B). Molecular-weight standards arc represented with markers from 97 to 14 kDa. PROTEIN METABOLISM IN ABALONE LARVAE 183 . kDa 97 66 45 31 21 14 7.0 4.0 Pi Figure 7. Autoradiogram of proteins in 3-day-old veliger larvae of Haliotis rufescens labeled with 35 S- methionine/cysteine and separated by two-dimensional gel electrophoresis. Area bounded by the dashed box is displayed in expanded form in Figure 8B. Molecular-weight standards correspond to Coomassie-stained gel (not shown) with markers from 97 to 14 kDa. Isoelectric points (pi) fall within a linear pH gradient from 7.0 to 4.0. Patterns of protein synthesis At the analytical resolution of one-dimensional gel elec- trophoresis, similar patterns of Coomassie-stained proteins were evident during development (1- to 7 -day-old larvae). Two classes of high-molecular-weight protein (Fig. 6A: labeled as P, and P 2 at 121 and 88 kDa, respectively) dominated the Coomassie-stained protein pool at each larval stage analyzed. The patterns of the existing pool of proteins (Fig. 6A: Coomassie-stained proteins) were different than those of the newly synthesized proteins (Fig. 6B: autora- diogram of labeled proteins), with no detectable synthesis of the two high-molecular-weight protein classes at any larval stage (note lack of signal in Fig. 6B autoradiogram for P, and P 2 ). The patterns of synthesis of individual proteins were compared, using two-dimensional gel electrophoresis (Figs. 7. 8). for the trochophore stage (1-day-old) and a veliger stage (3-day-old). The two-dimensional separation resolved over 300 proteins that were being synthesized in larvae (Fig. 7). The results of a qualitative analysis (visual absence or presence) are given in Figure 8 to show some of the differ- ent proteins being synthesized by 1- and 3-day-old larvae. The main point illustrated in Figures 7 and 8 is that protein synthesis in these larval forms is not limited to any single group (MW or pi) of proteins, but involves many proteins and complex patterns. Discussion Larval stages of Haliotis rufescens lost protein reserves continuously during development (Fig. 3). By day 8, velig- ers that were competent to metamorphose contained 34% less protein than the egg. The average daily loss of protein was 21.2 ng protein larva" 1 day" 1 (average rate of both cultures. Fig. 3). equivalent to 509 juJ day" 1 (24.0 kJ g~' protein; Gnaiger, 1983). During this period of protein loss, the absolute rates of protein synthesis decreased fourfold, from 193 to 49 ng protein synthesized larva"' day"' for 1-day-old trochophores and 7-day-old veliger larvae, re- spectively (Fig. 5; Table II). Note that the rate of protein loss was linear with time (Fig. 3). whereas the decrease in the rate of protein synthesis was nonlinear (Fig. 5). The relationship of these ontogenetic changes in protein loss and synthesis was analyzed further by calculating the fractional rates of protein synthesis during development. When kDa ,T 4 c + T T I A * 45 * R~l /* * f ' * R2 A f T i . * \ * A A A A A 1 31 B R3 4* 21 5.8 4.8 Pi kDa Rl * A T <^ P9 ? R2 31 ' ; ' _: R3 A *' R4 4 21 5.8 4.8 Pi Figure 8, Autoradiograms of two-dimensional separation of "S-labeled proteins in (A) I -day-old tro- chophuiv larvae ;md (B) 3-day -old veliger larvae of Haliotis rufescens (see Fig. 7 for the entire two-dimensional pattern of pro;. -in synthesis for the veliger stage). Each gel was loaded with equal radioactivity (3 x I0 h DPM) and exposed u> X-ray film for 8 days. Four different proteins that were labeled in each developmental stage are marked as reference points (R1-R4) for comparison of proteins in each gel. Arrows show proteins synthesized at one stage but not the other (by visual comparison of overlaid autoradiograms of trochophore and veliger). Orientation of arrows is not intended to indicate greater or lesser amounts of protein, just presence or absence. Molecular-weight standards correspond to Coomassie-stained gel (not shown) with markers from 45 to 2\ kDa. Isoelectric points (pi) fall within a linear pH gradient from 5.8 to 4.8. 184 PROTKIN METABOLISM IN ABALONE L.ARVAE 185 expressed as the percentage of a larva's total (whole-body) protein content that was synthesized per day. the fractional rates ranged from 40% (1 -day-old trochophore) to 14% (7-day-old veliger) (Table II). These rates of synthesis rep- resent the turnover rates of whole-body protein in larvae of H. rufescens. When there is no net deposition (growth) of total protein, as in the lecithotrophic larvae of H. rufescens. the rate of protein synthesis is the measure of the rate of protein turnover (Waterlow et /., 1978). Protein turnover, the continual breakdown and replacement of cellular pro- teins, is a significant component of maintenance metabolism in adult marine invertebrates (Hawkins. 1991 ). Few data are available in the literature for rates of protein turnover during development of marine invertebrates, making comparisons with our own data difficult. However. Berg and Merles (1970) measured a rate of protein turnover in sea urchin embryos (Lvtechinux ananiesits at 19C) of 23% day" 1 , a value that falls w ithin the range we determined for larvae of H. rufescens ( 149r-40% at 15C, Table II). These synthesis rates in abalone larvae are not limited to a few specific proteins. The electrophoretic patterns of protein synthesis showed widespread incorporation of 35 S-label, indicating a high synthesis rate of multiple proteins (Figs. 6B. 7. 8). as has been found in developing stages of other species of marine invertebrates (e.g., sea urchin embryos: Bedard and Brandhorst, 1983). These complex patterns of synthesis show that many proteins are involved with the synthesis side of the turnover process (synthesis and degradation). The specific proteins that are degraded to provide the pre- cursors for new synthesis have not been identified in this study, but the lack of synthesis of two high-molecular- weight proteins (Fig. 6B: PI and P2 at 88 and 121 kDa. respectively) suggests that these proteins might be degraded to support the new synthesis of other proteins. Amino acids dissolved in seawater are unlikely to have been a source of precursors for synthesis in our experiments, because the very high numbers of larvae (500 ml" ' ) that we used would have quickly depleted any substrates present at low concen- tration in natural seawater. The energy cost of protein turnover ( = synthesis) in larvae of H. rufescens can be calculated for comparison with the energy made available from the degradation of protein (Fig. 3). The cost of protein synthesis appears to be constant in different animals and stages of development [e.g.. Reeds et til.. 1985: mammal = 11.52 1.12 J (mg protein syn- thesized)" 1 : Hawkins et ai. 1989: juvenile marine bi- valve = 1 1.38 8.88 J (mg protein synthesized)" 1 ; Vavra ft id., unpubl.: veliger larvae of the bivalve Crassostrea gigas = 13.2 4.2 J (mg protein synthesized)" 1 ]. Using our value for molluscan larvae, the protein turnover rate measured for trochophore larvae of H. rufescens of 193 ng protein would equate to a requirement of 2548 ju.J larva day" 1 , decreasing in late-stage veliger larvae (7-day-old) to 647 fiJ larva" 1 day" 1 (turnover rate of 49 ng protein; Table II). The energy gained from the loss of protein (24.0 kJ g ') during development of H. rufescens was constant at 509 p.} day . Assuming complete oxidation of the protein de- pleted, this input of energy from the loss of protein could account for only 20% (509/2548) of the cost of protein turnover in 1 -day-old larvae. Obviously, during early devel- opment most of the costs of protein turnover have to be supplied from sources other than protein degradation. How- ever, later in development (near metamorphosis) when rates of protein turnover decrease, the energy made available from protein reserves could account for a large percentage (79%) of the costs of turnover. Protein turnover plays a major role in establishing the metabolic rate and physiological state of animals (Waterlow et til.. 1978: Reeds et al.. 1985; Hawkins, 1991). For leci- thotrophic larval forms, little is known about rates of mac- romolecular synthesis during nonfeeding development. Our findings with larvae of H. rufescens show that these pro- cesses proceed at a high rate and undergo dynamic ontoge- netic changes, with up to 40% of a larva's whole-body protein being turned over per day compared to depletion rates of only 4%-6%. If abalone larvae are typical of other lecithotrophs. at least in their biochemical and physiological activities, then biosynthetic rates in nonfeeding larval forms may be much higher than expected. Acknowledgments We thank John McMullen and Mike Machuzak of the Ab Lab (Port Hueneme, California) for supplying animals and supporting this project. Our thanks also to Nicholas Appel- mans for his help in rearing the larvae. This research was supported by a grant from NOAA, Office of Sea Grant (U.S.C. Sea Grant) and by a grant from the Office of Naval Research ( NOOO 1 4-90- J- 1740). Literature Cited Anger, K. 1996. Physiological and biochemical changes during larval development and early juvenile growth in the northern stone crab, Lithodes maja (Decapoda: Anomura). Mar. Biol. 126: 283-296. Bedard. P. A., and B. P. Brandhorst. 1983. Patterns of protein synthe- sis and metabolism during sea urchin embryogenesis. Dev. Bid. 96: 74-83. Ben-David-Zaslott, R., and V. Benayahu. 1998. Competence and lon- gevity in planulae of several soft corals. Mar. Ecol. Prog. Ser. 163: 235-243. Berg, VV. E., and D. H. Merles. 1970. Rates of synthesis and degrada- tion of protein in the sea urchin embryo. .v/>. Cell Res. 60: 218-224. Bradford, 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. Fry, B. J., and P. R. Gross. 1970. Patterns and rates of protein synthesis in sea urchin embryos: I. Uptake and incorporation of amino acids during the first cleavage cycle. Dev. Biol. 21: 105-124. 186 J. VAVRA AND D. T. MANAHAN Gnaiger, E. 1983. Calculation of energetic and biochemical equivalents of respiratory oxygen consumption. Pp. 337-345 in Polarographic Oxygen Sensors, E. Gnaiger and H. Forstner. eds. Springer-Verlag, Berlin. ( .(piisiin. A. S., and F. H. Wilt. 1981. Protein synthesis, polyribosomes, and peptide elongation in early development of Strongylocentrotus />ttii\iiii.\. Dcr. Biol. 82: 32-40. llnu kins. A. J. S. 1991. Protein turnover: a functional appraisal. Fund. Ecnl. 5: 222-233. Hawkins, A. J. S., J. Widdows, and B. I,. Bayne. 1989. The relevance of whole-body protein metabolism to measured costs of maintenance and growth in Myrilus eilulis. Physiol. Zool. 62: 745-763. Hoegh-Guldberg, O., and R. B. Emlet. 1997. Energy use during the development of a lecithotrophic and a planktotrophic echinoid. Biol. Bull. 192: 27-40. Jaeckle, W. B., and D. T. Manahan. I989a. Growth and energy im- balance during the development of a lecithotrophic molluscan larva (Haliotix rufescenx). Biol. Bull. 177: 237-246. Jaeckle, W. B., and D. T. Manahun. 1989b. Feeding by a "nonfeeding" larva: uptake of dissolved umino acids from seawater by lecithotrophic larvae of the gastropod Halioii\ nifescens. Mar. Biol. 103: 87-94. Laemmli, U. K. 1970. Cleavage of structural proteins during the assem- bly of the head of bacteriophage T4. Nutiiiv 227: 680-685. Manahan, D. T. 1983. The uptake and metabolism of dissolved amino acids by bivalve larvae. Biol. Bull. 164: 236-250. Pearse, J. S. 1994. Cold-water echinoderms break "Thorson's Rule." Pp. 26-43 in Reproduction, Larval Bi(>tt>^\: and Recruitment in the Deep- sea Benthos, K. J. Ecklebarger and C. M. Young, eds. Columbia University Press. New York. Reeds, P. J., M. F. Fuller, and B. A. Nicholson. 1985. Metabolic basis of energy expenditure with particular reference to protein. Pp. 46-57 in Substrate and Energy Metabolism, J. S. Garrow and D. Halliday. eds. John Lihbey. London. Shilling, F. M., and D. T. Manahan. 1994. Energy metabolism and amino acid transport during early development of antarctic and tem- perate echinoderms. Biol. Bull. 187: 398-407. Shilling, F. M., O. Hoegh-Guldberg, and D. T. Manahan. 1996. Sources of energy for increased metabolic demand during metamor- phosis of the abalone Haliotis riifexcens (Mollusca). Biol. Bull. 191: 402-412. Strathmann, R. R. 1978. The evolution and loss of feeding larval stages of marine invertebrates. Evolution 32: 894-906. Thorson, G. 1950. Reproductive and larval ecology of marine bottom invertebrates. Biol. Rev. 25: 1-45. Turner, R. L., and J. C. Rutherford. 1976. Organic, inorganic, and caloric composition of eggs, pentactulae, and adults of the brooding sea cucumber Cuiumaria cnrala Cowles (Echinodermata: Holothuroidea). J. Ev/>. Mm: Biol. Etal. 24: 49-60. Waterlow, J. C., P. J. Garlick, and D. J. Millard. 1978. Protein Turnover in Mammalian Tissues and in the Whole Body. North-Hol- land. Amsterdam. Welborn, J. R., and D. T. Manahan. 1995. Taurine metabolism in larvae of marine molluscs (Bivalvia. Gastropoda). J. .v/>. Bid. 198: 1791-1799. Reference: Binl. Bull. 196: IS7-I9S. (April Settlement and Metamorphosis of Capitella Larvae Induced by Juvenile Hormone-Active Compounds Is Mediated by Protein Kinase C and Ion Channels WILLIAM J. BIGGERS' AND HANS LAUFER 12 * ^Department of Molecular ami Cell Biolag\, The University of Connecticut, Starrs, Connecticut 06268; and 'The Marine Biological Laborator\, Woods Hole, Massachusetts 02543 Abstract. The signal transduction pathway by which ju- venile hormone-active compounds induce settlement and metamorphosis of metatrochophore larvae of the polychaete annelid Capitella sp. 1 was investigated. The known protein kinase C (PKC) activator phorbol-12.13-dibutyrate was an active inducer of settlement and metamorphosis, whereas H-7. an inhibitor of PKC. inhibited settlement and meta- morphosis in response to juvenile hormone III (JH III). JH III and methyl farnesoate (MF) also directly activated, in vitro, both a PKC-like enzyme present in Capitella homog- enates and PKC purified from rat brain. In addition, binding studies using the fluorescent PKC inhibitor RIM-1 revealed the presence of a PKC-like enzyme in intact Capitella larvae and juveniles. Settlement and metamorphosis of the larvae was also stimulated by membrane-depolarizing con- centrations of KC1. This response to KC1 was inhibited by tetraethylammonium. The potassium channel blocker 4-aminopyridine induced settlement and metamorphosis, whereas settlement and metamorphosis in response to JH III was inhibited by the potassium channel ionophore nigericin. Settlement and metamorphosis induced by JH III was in- hibited by the calcium channel blockers Ni 2 + . Zn 2 + , and verapamil. whereas settlement and metamorphosis was in- duced by the calcium ionophore A23187. These results suggest that in mediating this response, juvenile hormones may cause activation of PKC. leading to subsequent mod- ulation of potassium and calcium channels. Introduction The chemoreception by marine invertebrate larvae of chemical "cues" that are present in the ocean environment Received 6 August 1998; accepted 29 December 1998. * To whom correspondence should be addressed. and induce settlement and metamorphosis is important for the recognition of habitats that favor growth and reproduc- tion (Chia and Rice. 1978; Rittschof and Bonaventura. 1986; Scheuer. 1990). These settlement signals appear to be specific for different species, as evidenced by findings that larvae of the abalone Haliotis nifescens respond to specific chemicals in red algae (Morse el al. 1984). larvae of the nudibranch Phestilla sibogae respond to chemicals in corals (Hadfield. 1978. 1984). larvae of the polychaete annelid Phragmatopoma californica respond to chemicals present in the burrows of adult worms (Pawlik, 1988, 1990; Jensen and Morse. 1990), and larvae of the sand dollars Dendraster excentricus (Burke. 1984) and Echinarachniits parma (Pearce and Scheibling, 1990) respond to chemicals pro- duced by adult sand dollars. In previous studies, we have found that juvenile hor- mones (JH). which are known morphogens that regulate reproduction and development of insects and crustaceans (Laufer and Borst, 1983, 1988; Laufer et al.. 1987), as well as chemicals with juvenile hormone activity in insect cuticle bioassays, are able to induce settlement and metamorphosis of metatrochophore larvae of the polychaete annelid Capi- tella sp. I (Biggers and Laufer. 1992. 1996). which is a subspecies member of the Capitella polychaete complex (Grassle and Grassle. 1976). In nature, larvae of Capitella sp. I are stimulated to settle and metamorphose (Fig. 1) when they come into contact with chemical inducers present in sediments (Butman et al., 1988). although the identity of these chemicals remains in debate (Cuomo. 1985; Dubilier. 1987). they appear to have JH-activity (Biggers. 1994). We have now investigated the signal transduction process through which the Capitella larvae respond to JH-active compounds. Our results presented in this paper indicate that JH-active compounds stimulate settlement and metamor- 187 188 W. J. BIGGERS AND H, LAUFER prostomium prototroch Life Cycle of the Polychaete Annelid Capitella sp. I telotroch "N Free swimming Metatrochophore larva Settlement of larva into sediments, metamorphosis and growth Formation of brood tube, Incubation and development of eggs and larvae Figure 1. Diagram of the life cycle and metamorphosis of Capitella sp. I larvae. In parental brood tubes, trochophore larvae develop into segmented metatrochophore larvae. After hatching and release from the brood tubes, in response to chemical settlement cues, swimming meta- trochophore larvae settle and metamorphose into juvenile worms by losing their cilia needed for swimming and developing capillary setae and hooded hooks necessary for crawling through sediments. phosis of these larvae through the activation of protein kinase C (PKC) and subsequent modulation of ion channels. Materials and Methods Capitella larval settlement bioassays Stock cultures of Capitella sp. I were maintained at 18C in large plastic containers containing artificial seawater (Utikem Co.) and washed sea sand (Fisher Scientific) and were fed Tetramin fish food flakes. Brood tubes containing adult females along with their developing eggs and larvae were then separated from the cultures and placed into 60-mm glass petri dishes containing seawater. The dishes were checked daily for hatched, swimming metatrocho- phore larvae to be used for bioassays. All test chemicals, except for methyl farnesoate (MF) which was synthesized in our laboratory, were purchased from Sigma Chem. Co. Stock solutions of juvenile hormone III (JH III), MF, phor- bol-12,13 dibutyrate (PDBU), l-(5-isoquinolinyl-sulfonyl)- 2-methylpiperazine (H-7), arachidonic acid, elaidic acid, verapamil, 4-aminopyridine, and nigericin were prepared in 95% ethanol. Stock solutions of KC1. NiCl 2 , ZnCK. and tetraethylammonium chloride (TEA) were prepared in dis- tilled water. Settlement and metamorphosis bioassays were conducted at 18C using 60-mm glass petri dishes that were pre-baked at 250C to remove contaminants (Biggers and Laufer, 1996). For the assays, 10 to 100 ;ul of the test chemical stock solutions were added by micropipet into petri dishes containing 10 metatrochophore larvae less than 1 day old ( 1 day post-release), and 10 ml of artificial seawater. The dishes were then observed. After 1 h, the amount of settlement and metamorphosis was assessed by placing each dish under a dissecting microscope and count- ing the number of settled larvae crawling on the bottom of the dish. Metamorphosis after 1 hour was also more criti- cally assessed by using a compound microscope and noting the loss of cilia, elongation, and development of capillary setae. Protein kinase C assavs Assays for the presence of protein kinase C were carried out essentially as described by Yasuda et a/. (1990), by measuring phosphorylation of an 1 1 -residue synthetic pep- tide from myelin basic protein (MBP 4 _ U ). This method is specific for measurement of PKC, and permits selective measurement in crude tissue preparations. The PKC assay was first standardized using purified rat brain PKC (calcium and phospholipid dependent) from Calbiochem Co. Phos- phorylation of MBP 4 _, 4 was measured using 2-10 ng of enzyme, and over a time period of 30 min. Reactions were carried out in plastic Eppendorf tubes, with 50-ju.l reaction mixtures containing 20 mM Tris/HCl. pH 7.5, 5 mM mag- nesium acetate, 100 /iA/ CaCl : , 25 juM MBP 4 I4 (Sigma Chem. Co.), 50 ng diolein and 500 ng phosphatidylserine, 2-10 ng of rat brain PKC, and 10 H.M ATP (containing 5-6 X 10 s CPM gamma 32 P-ATP). Reactions were started by addition of enzyme, and incubations were carried out at 30C for up to 30 min. The reactions were then stopped by placing the tubes on ice and spotting the reaction mixtures onto P-81 anion exchange paper discs (Whatman Co.), which were then immediately immersed in 75 mM H,PO 4 , and washed eight times in 100 ml of the same solution. The filter discs were placed in Ecolume cocktail for analysis with a scintillation counter. Other PKC assays using rat brain PKC were carried out in the same manner, except with the replacement of diolein and phosphatidylserine with test chemicals as indicated. PKC activity in Capitella larval homogenates was deter- mined in the same manner, except with the inclusion of the larval homogenates (20-100 jug protein) instead of the rat brain PKC. Capitella larvae were collected within 1 day of release from the brood tubes, and were frozen at 20C. After thawing, the larvae were centrifuged at 5000 rpm for 1 min in a microfuge, the seawater was discarded, and larval homogenates were prepared by homogenizing 1000 larvae in 1 ml 20 mM Tris/HCL, pH 7.5, containing 0.5 mM phenylmethylsulfonyl fluoride (protease inhibitor) using small glass homogenizers. The homogenates were then cen- trifuged in a microfuge at 5000 rpm for 1 min to remove large cellular debris, and the PKC activity of the supernatant was assessed as previously described for studies with rat SIGNAL TRANSDUCTION IN CAPITELIA BY PKC 189 brain PKC. Protein concentrations of supernatants were determined using a microprotein assay (Sigma Chem. Co.) and a standard curve of increasing concentrations of bovine serum albumin. Localisation of PKC by KIM- 1 analysis Rhodamine-conjugated bisindolylmaleimide (RIM-1) (Cal- biochem). a fluorescent PKC inhibitor (Chen and Poenie, 1993), was used to visually locate PKC in Capitella larvae and juveniles. Metatrochophore larvae less than 1 day old and 1 -day-old juveniles were briefly exposed for 1 min at room temperature to 200 nm RIM-1 in seawater ( 100 ^il) in depres- sion slides covered with aluminum foil. Larvae or juveniles were then fixed in 2% formaldehyde containing seawater for 15 min, transferred into methanol for 15 min to permeabilize the membranes, and then transferred three times (5 min each wash) into 1 ml fresh seawater to rinse away excess unbound RIM-1. Fixed larvae or juveniles were then transferred in 10% glycerol-seawater onto microscope slides and visualized by either light microscopy or fluorescent microscopy with a Ni- kon confocal fluorescent microscope equipped with a rhoda- mine filter. Results Effects of protein kinase C modulators on Ian-til settlement and metamorphosis We have previously found that the juvenile hormones MF. JH I, and JH III, as well as compounds with juvenile hormone activity in insect cuticle bioassays. including ara- chidonic acid, are able to induce settlement and metamor- phosis of Capitella sp. I metatrochophore larvae (Biggers and Laufer, 1992). The first indication that the larvae sense the presence of JH III in the seawater is the onset of excited swimming. This response, which is typical of normal settlement behavior (Butman et ai. 1988; Pechenik and Cerulli. 1991), is faster than normal, with intermittent spiral-corkscrew movements, gradual body elongation, and periodic touching of the bot- tom of the petri dish. The larvae then settle and metamor- phose into normal juvenile worms within 1 h. Both MF and JH III are very potent inducers of settlement and metamor- phosis of the Capitella larvae at micromolar concentrations, whereas control larvae do not start to spontaneously settle and metamorphose until after 24 h (Fig. 2). In preliminary investigations of the signal transduction process that mediates JH-induced settlement and metamor- phosis of the Capitella larvae, we found compounds that elevate intracellular cAMP concentrations to be ineffective inducers of settlement and metamorphosis, indicating that cAMP does not act as a second messenger in this signal transduction process (Biggers and Laufer. 1992). Based upon the ability of JH I, JH III, and arachidonic acid 120 -/ o .c Q. 6 E (0 100 - 80- 60- 40- 20- uM 1 uM 10 uM 25 uM Figure 2. Concentration-dependent stimulatory effects of juvenile hor- mone III (JH III) and methyl farnesoate (MF) on Capitella larval settle- ment. Metatrochophore larvae, less than I day after release from parental brood tubes, were exposed to several concentrations of JH III or MF in seawater, ranging from 1 to 25 /j,M. Percent settlement and metamorphosis is shown for each concentration after I h. Each bar represents the percent settlement and metamorphosis for each concentration (mean SEM, n = 6. representing 60 larvae/concentration). Controls (0 p.M) received 50 /nl 95% ethanol. Asterisks indicate significant differences (P < 0.001, F > 50) between control and experimental values at each concentration tested by one-way ANOVA with a Bonferroni correction for multiple comparisons where a = 0.017. to activate PKC in other species including insects (Yamamoto et al.. 1988; Holian et al. 1989; Sevala and Davey, 1989; Shearman et ai, 1989a; Shinomura <>?/.. 1991), and on reports that PKC activation is involved in the mediation of settlement and metamorphosis of other species of marine invertebrate larvae (Muller, 1985; Baxter and Morse, 1987; Leitz and Klingman, 1990; Morse, 1990), we further investigated the involvement of PKC activation in mediating settlement and metamorphosis of the Capitella larvae by testing the effects of other known PKC modulators. The phorbol ester phorbol- 12.13-dibutyrate (PDBU) is a well-studied activator of PKC, and PKC has been found to be the actual cellular receptor for PDBU in some cells (Castagna et al, 1982; Vandenbark et al.. 1984). Experiments with PDBU on the Capitella larvae showed that it is also a very potent inducer of settlement and metamorphosis (Fig. 3A). indicating that PKC activation is involved in mediating this process. The effect of a PKC inhibitor was next studied to deter- mine if it could inhibit settlement and metamorphosis in- duced by JH. The PKC inhibitor l-(5-isoquinolinyl-sulfo- nyl)-2-methylpiperazine, abbreviated H-7 (Hidaka et al.. 1984), caused a concentration-dependent inhibition of Cap- itella settlement and metamorphosis induced by 25 /u,A/ JH III, when the larvae were pre-exposed to H-7 for 3 h (Fig. 3B). This result again indicates the involvement of PKC activation in mediating the effects of JH on settlement and metamorphosis. 190 W. J. BIGGERS AND H. LAUFER 120- A B 100 Figure 3. Effects of protein kinase C (PKC) modulators on larval settle- ment and metamorphosis. (A) Induction of settlement and metamorphosis by 1 2. 13 dibutyrate (PDBLI): Cupilella metatrochophore larvae, less than 1 day after release, were exposed to the PKC activator phorhol PDBLI added to the seawater at concentrations ranging from 1 to 25 jiM, and settlement and metamorphosis was noted after 1 h. Controls received 50 /al 95% ethanol. Bars represent the percent settlement and metamorphosis for each concentration (mean SEM. n = 6. representing (SO larvae/concentration). Asterisks indi- cate significant differences (P < 0.001. F > 25) between control and exper- imental values at each concentration tested by one-way ANOVA with a Bonterroni correction for multiple comparisons where a = 0.017. (Bl Inhibi- tion of juvenile hormone (JH) Ill-induced settlement and metamorphosis by H-7. Cn/iitclUi metatrochophore larvae, less than 1 day after release, were pre-exposed lor 3 h to concentrations of H-7. a PKC inhibitor, in seawater ranging from 1 to 100 /xA/. before being exposed to JH III (20 /jjV/) lor I h. Controls (0 fj.A'7) received 100 M l 45<7 r ethanol followed by JH III. Bars represent the percent settlement and metamorphosis for each concentration after the 4-h time period (mean SEM, n = 6. representing 60 larvae/ concentration). Asterisks same as in (A). Activation of u protein kimiae C-like enzyme in Capitella larvae b\ JH-active chemicals Although PKC is considered to be a ubiquitous enzyme present throughout the animal kingdom and has been found in marine sponges (Muller ct a/.. 19H7) and Dictwsti'Iimu discoideum (Jimenez et a/., 1989; Luderus et al., 1989). its presence in polychaetes has so far not been documented. The presence of a PKC-like enzyme in Capitella larvae was therefore investigated, as was also the ability of JH-active compounds to activate this enzyme. In these investigations, we used an assay specific for PKC. Developed by Yasuda et til. ( 1990), this assay is based upon the specific phosphor- ylation by PKC of an 11-amino acid peptide sequence of myelin basic protein, which is not phosphorylated by either cAMP-dependent protein kinase (PKA), casein kinases I and II, Ca" + /calmodulin-dependent protein kinase II, or phosphorylase kinase. The results demonstrate that a PKC- like enzyme does exist in Capitella (Fig. 4). PKC activity was linear for up to 15 min for homogenate concentrations up to 100 ;ag (Fig. 4A, B). The specific activity of PKC in the larval homogenates was calculated as being 6.7 fmoles P-incorporated per minute per microgram of protein. In- cubations were also done without the PKC substrate MBP 4 _| 4 to monitor endogenous larval protein phosphory- lation, of which only a very small amount could be detected. Experiments were next conducted to determine whether 14000 12000 CL Q 10000 A 8000 6000 = 4000 20 40 60 Larval protein (ug) 80 100 Q. Q Bo 1 10 15 20 Time (minutes) 25 Figure 4. Protein kinase C (PKC) activity in Capitella larvae. PKC activity was detected in larval homogenates after activation with phospha- tidylserine and diolein. ' ; P-incorporation into MBP 4 ]4 was linear with increasing protein concentrations ranging from 20 to 100 ^xg (A) and was also linear for 15 min using 100 |U,g protein (B). SIGNAL TRANSDUCTION IN CAP1TELH BY PKC 191 the PKC-like enzyme present in the Capitella lar\ r ae could he activated by JH-active compounds. Incubations were carried out using 100 ^g of the larval homogenates for IS inin at 30 "C in the presence of either phosphatidylserine/ diolein (PS/DO). 10 ^M arachidonic acid (AA), 10 ^M JH III, 10 juA/ MF. or 10 M A/ elaidic acid (EA). The results of this experiment indicate that JH III, MF, and AA are able to activate CapiteUa PKC //; \-itn> (Fig. 5 A). In comparison with activation by PS/DO, arachidonic acid was the stron- gest activator W< activation), followed by MF (73% ac- tivation) and JH III (51% activation). Elaidic acid, a trans- isomer of oleic acid that does not activate rat brain PKC (Shearman ct ai. 1989a) and does not induce settlement and metamorphosis of the Capitellci larvae, did not activate the Capitella PKC. These results therefore again indicate the involvement of PKC activation in mediating settlement and metamorphosis of the Capitella larvae. Since it is possible that the crude larval homogenates contain enzymes, receptors, and substrates of the diacyl- glycerol pathway, such as phospholipase C and guanine- binding proteins through which indirect activation of PKC may occur, we also tested whether JH could directly activate a purified preparation of rat brain PKC. Polyunsaturated fatty acids such as arachidonic acid can activate bovine brain PKC (Shearman et til.. 1989a) and rat brain PKC (Holian et al., 1989; Shinomura et ai, 1991); therefore, juvenile hormones might also be able to cross species lines and activate rat brain PKC. Incubations were carried out using 5 ng purified rat brain PKC (Calbiochem Co.) in the presence of either PS/DO, 10/uAf AA, lO^A/JHIII. \() ^M MF. or 10 JU.A-/ EA, at 37C for 15 min. The results show that insect juvenile hormones and the crustacean juvenile hor- mone MF are able to directly activate rat brain PKC in the absence of phosphatidylserine and diolein (Fig. 5B). Ara- chidonic acid was the most potent activator tested (93% activation relative to PS/DO), whereas elaidic acid was inactive, confirming earlier work reported by Shearman et cil. ( 1989a). MF caused 59% activation of the rat brain PKC and was more active than JH III (28% activation). Locali-ution of PKC in Capitella />v RIM-1 To visualize the location of PKC in the CapiteUa larvae and juveniles and to identify possible chemosensory cells that would rapidly take up external chemicals or chemicals in the environment, larvae and juveniles were briefly ( 1 min) exposed to a fluorescently labeled protein kinase C inhibitor, rhodamine-conjugated bisindolylmaleimide (RIM-1), which has proven useful as a fluorescent probe for PKC (Chen and Poenie. 1993). After exposure to this in- hibitor, the larvae or juveniles were fixed, permeabilized. rinsed to remove excess unbound RIM-1. and viewed under a fluorescent microscope. In metatrochophore larvae, dis- tinct cells in the ciliary bands of the prototroch and telotroch 120 -f 100 - A i B PS/DO Figure 5. Activation of an enzyme resembling protein kinase C (PKC) in CtipiteHa larval homogenates and purified rat brain PKC by JH-active chemicals. (A) Activation of the PKC-like enzyme present in larvae. PKC assays using Capitella larval homogenates were carried out in the presence of either phosphatidylserine/diolein (PS/DO). 10 \j.M arachidonic acid (AA). 10 /nA/ juvenile hormone (JH) III. 10 juiW trans, trans methyl farnesoate (MF). or 10 pM elaidic acid (EA). Incorporation of 32 P into MBP 4 _ I4 for AA. JH III. MF. and EA is shown expressed as the percentage incorporation relative to that found for the control (PS/DO). (B) Activation of purified rat brain PKC. PKC assays using 5 ng purified rat brain PKC were carried out in the presence of either phosphatidylserine/diolein (PS/ DO). 10 fj-M arachidonic acid (AA). 10 /xM JH III. 10 ^M trans, tram, methyl farnesoate (MF), or 10 p.M elaidic acid (EA). Incorporation of 12 P into MBP 4 _, 4 for AA, JH III, MF, and EA is shown expressed as the percentage incorporation relative to that found for the control (PS/DO I. and isolated cells in the apical region of the prostomium and the rest of the body displayed RIM-1 binding, indicating that these cells possess PKC (Fig. 6B). Juvenile Capitella exposed in the same manner displayed RIM-1 binding to cells in the apical region of the prostomium and scattered throughout the rest of the body (Fig. 6D). These results provide more evidence for the presence of a PKC-like enzyme in CupitelUi larvae and juveniles, and are consistent with a chemosensory function for apical cells in regions of the prostomium as previously noted by Eckelbarger and Grassle (1987). 192 W. J. BIGGERS AND H. LAUFER B Figure 6. Localization of protein kinase less than 1 day after release from parental brc a rhodamine-conjugated PKC inhibitor. (A) cenee microscopy of metatrochophore larva well as isolated apical cells located in the Cupiti'lln prostomium (400X ). (D) Fluoresce of RIM- 1 to apical cells (arrow) (400X). C (PKC) in Ctipitellu by RIM-1 analysis. Metatrochophore larvae >od tubes and 1 -day-old juveniles were exposed for 1 min to RIM-1. Light microscopy of metatrochophore larva (400X). (B) Fluores- showing binding of RIM-1 to ciliary cells of the prototroch (p) as prostomium (arrow) (800X). (C) Light microscopy of juvenile nee microscopy of juvenile Capilella prostomium showing binding Effects of potassium chiinnel modulators In mediating cellular responses. PKC activation has in many cases been found to result in the modulation of potassium and calcium channels (Kaczmarek, 1987; Shear- man et ai, 1989b). The involvement of potassium channels in mediating the settlement and metamorphosis of several types of marine invertebrate larvae has also been previously demonstrated (Baloun and Morse, 1984; Yool et ai, 1986; Leitz and Klingman. 1990; Carpizo-Ituarte and Hadfield, SIGNAL TRANSDUCT1ON IN CAPITELLA BY PKC 193 Table I Effects of ion channel modulators on settlement uiii/ metamorphosis of Capitclla .s/>. / lamu Chemical % Larvae settled and metamorphosed (after I h) Potassium channel modulators KCI (20 mA/) KCI (20 mA/) + TEA ( KM) mA/) 4-Aminopyridine (100 fj.A/1 JH III (38 vM) JH III (38 /iA/> + nigericin (500 ng/ml) Calcium channel modulators A23187 (400 nM) JH 111 (38 |U.A/) JH III (38 ijM) + NiCl, (10 mA/) JH III (38 /MA/) + ZnCI, (10 mA/) JH 111 (38 /aA/l + Verapamil (17 /j.M) 100 5 101) 100 (I 101) 100 JH = juvenile hormone. 1998). Studies were therefore carried out to determine if the modulation of potassium channels may also mediate settle- ment and metamorphosis of the Capitella larvae. Increased external KCI concentrations in the seawater induced settle- ment and metamorphosis of the Capitella larvae in a con- centration-dependent manner, with an added concentration of 20 mM (30 mM total including seawater) inducing 100% settlement and metamorphosis in 1 h (Table I). These effects of KCI appear to be mediated by K + ions and not Cl~ since addition of NaCl had no effect on settlement and metamor- phosis. The effect of tetraethylammonium chloride (TEA), a known blocker of potassium currents, was next studied to determine its effect on settlement and metamorphosis. TEA did not induce settlement and metamorphosis of the Capi- tella larvae, but instead inhibited the stimulatory response of the larvae to KCI, with a concentration of 100 mM TEA almost completely inhibiting settlement and metamorphosis (Table 1 1. Settlement and metamorphosis of the Capitella larvae was, however, stimulated in a concentration-depen- dent manner by 4-aminopyridine (4-AP), another potassium channel blocker. which blocks outward rectifying potassium currents. A concentration of 100 ju,jW4-AP stimulated 100% settlement and metamorphosis of the larvae within 1 h (Table 1). The effects of the potassium channel ionophore, nigericin. was next studied. Pre-exposure of the larvae to nigericin for 30 min inhibited the response of the larvae to JH III in a concentration-dependent manner, with 500 ng/ml nigericin causing 100<7r inhibition of settlement and meta- morphosis in response to JH III in 1 h (Table I). Effects of calcium channel modulators Calcium channels are in many cases activated in response to membrane depolarization and in response to PKC acti- vators (Kaczmarek. 1987; Shearman et u/.. 1989b). Calcium channel modulators were therefore studied for their effects on settlement and metamorphosis of the Capitella larvae. Pre-exposure of the larvae to the known calcium channel blockers Nr + at a concentration of 10 mM, 7.n 2+ at a concentration of 10 mA/. and verapamil at a concentration of 17 JU.A/ completely inhibited the settlement- and metamor- phosis-inducing effects of JH III (Table I). Next, to deter- mine whether an influx of calcium could stimulate settle- ment and metamorphosis, the effect of the calcium channel ionophore A23187 was tested. A23187 proved to have a potent, concentration-dependent effect on larval settlement and metamorphosis (Fig. 7), with a concentration of 400 nM stimulating 100% of the larvae to settle and metamorphose in 1 h (Table I). These results therefore suggest that JH HI activation of PKC may lead to the opening of calcium channels. Discussion The ability of juvenile hormones and compounds with JH activity to induce settlement and metamorphosis in Capi- tella larvae raises the possibility that these compounds may act on the larvae through a mechanism similar to that by which they affect insect metamorphosis and reproduction. In insects, juvenile hormones have multiple mechanisms for regulating metamorphosis and reproduction. For example, they may act through nuclear receptors and transcriptional regulation (Jones et al.. 1993: Palli et al., 1994; Jones and 50 100 1 50 200 250 A23187 (nM) 300 350 400 Figure 7. Effect of the calcium ionophore A23187 on settlement and metamorphosis of Capitella larvae. Metatrochophore larvae less than I day after release, exposed for 1 h to concentrations of A23 1 87 ranging from 25 to 400 nM in seawater. were stimulated to settle and metamorphose in a concentration-dependent manner. 194 W. J. BIGGERS AND H. LAUFER Sharp. 1997). or by affecting mRNA stability (Jones et al.. 1993) and mRNA translation Ulan el til., 1972). JH 1 affects follicle cell patency in Rliodniits prolixus through another well-studied mechanism (Sevala and Davey, 1989) a sig- nal transduction cascade involving a membrane receptor, activation of PKC, and the subsequent activation of a Na/ K-ATPase. Other studies (Yamamoto et nl.. 1988) have demonstrated that PKC activation and opening of calcium channels is involved in the mechanism by which JH III induces protein synthesis in Drosophilu male accessory glands. Our studies indicate that PKC activation and ion channel modulation are also involved in the process of chemosensory signal transduction by which JH and com- pounds with JH activity induce settlement and metamorpho- sis of larvae of Capitella sp. I. In addition to JH I and JH III. other known activators of PKC can also induce settlement of Capitella larvae. One example is arachidonic acid, which has shown JH activity in insect cuticle bioassays (Slamu, 1962) and strongly acti- vates PKC from bovine brain (Shearman ct ai. 1989a) and rat brain (Holian et ul.. 1989; Shinomura et at., 1991 ). The potency of the phorbol ester PDBU in inducing settling and metamorphosis of Capitella larvae (Fig. 3A) is particularly good evidence for the activation of a protein kinase C-like enzyme, since it is well established that PDBU directly activates PKC in mammalian tissues (Castagna et ai. 1982; Nishizuka. 1984; Vandenbark et al.. 1984; Parker et ai. 1986). The PKC inhibitor H-7 is able to inhibit the settle- ment and metamorphosis effects of JH III (Fig. 3B). Al- though H-7 at higher concentrations also inhibits other protein kinases such as cyclic AMP-dependent protein ki- nase (PKA) (Hidaka et ul., 1984). the effects of H-7 on the Capitella larvae are probably due to the inhibition of PKC and not PKA. since PDBU. which induces settlement and metamorphosis, activates only PKC and not PKA (Castagna et al., 1982). PKC is regarded as a ubiquitous enzyme present through- out the animal kingdom (Nishizuka. 1984). having been demonstrated in Dictyostelium (Luderus et ai. 1989; Jime- nez etui., 1989) and sponges (Muller e? /., 1987) as well as in mammalian tissues. Other studies have also demonstrated that PKC activation is involved in the signal transduction processes that mediate settlement and metamorphosis of marine invertebrate larvae of the coelenterate Hydractinia (Muller. 1985; Leitz and Klingman, 1990) and the mollusc Haliotis nifescens (Baxter and Morse. 1987; Morse. 1990). It is therefore likely that Capitella also possesses a PKC- like enzyme that is involved in mediating larval settlement and metamorphosis. Our data indicate that Capitella does possess such an enzyme, since crude homogenates of the larvae show PKC activity in a selective PKC assay (Fig. 4). More studies are needed to further characterize the enzyme and determine its requirement for calcium. JH appears to activate PKC in the Capitella larvae di- rectly, much like the mechanism of action of phorbol esters. Our results show that MF, JH III, and arachidonic acid can, /'// vitro, directly activate the PKC-like enzyme in Capitella as well as purified PKC from rat brain in the complete absence of the normal membrane inducers phosphatidyl serine and diacylglycerol (Fig. 5). These data therefore indicate that juvenile hormones, as well as JH-active com- pounds such as arachidonic acid, are able to bind to the lipid-binding site in PKC (Parker et ai. 1986; Ziesel, 1993). This is perhaps not surprising given that phorbol esters and juvenile hormones are both terpenoid compounds; phorbol esters are diterpenoids and juvenile hormones are sesqui- terpenoids. In sensing JH-active compounds in the seawater. these lipophilic compounds presumably pass through the mem- brane of ciliary epithelial chemosensory cells similar to those reported to transduce chemical signals for settlement and metamorphosis in larvae of the abalone Haliotis (Trapido-Rosenthal and Morse. 1986), the cnidarian Hy- dractinia (Schwoerer-Bohning et ai, 1990). the polychaete Phragmatopoma culifornica (Amieva et ul., 1987), and the sea star Lnidia senegalensis (Komatsu et al., 1991). Our studies with the fluorescent PKC inhibitor RIM-1 provide more evidence that Capitella larvae possess a PKC-like enzyme and that PKC is present in chemosensory cells. The RIM-1 presumably was able to directly enter chemosensory cells of the larvae that allow rapid uptake of external chem- icals, since the larvae were exposed to RIM- 1 only briefly ( 1 min). The intense RIM-1 binding in isolated cells of the prostomium (Fig. 6) may represent the presence of PKC in apical cilia that are thought, on the basis of evidence from electron microscopy (Eckelbarger and Grassle, 1987) to serve a chemosensory function. However, entry of RIM-1 into the larvae through other processes cannot be ruled out, since RIM-1 binding was found in cells throughout the body, especially in ciliary cells of the prototroch and telotroch, and in both metatrochophore larvae and juveniles. It is likely that JH-active compounds, after passing through the membrane of chemosensory cells, bind with an inactive PKC present in the cytosol, which, as in other species (Nishizuka. 1984). then becomes active and is trans- located to the membrane. It is evident from our studies that micromolar concentrations of JH-active compounds are able to activate PKC and thereby induce settlement and meta- morphosis. Concentrations of 10 p.M JH III. MF, and ara- chidonic acid activate both Capitella PKC and rat brain PKC (Fig. 4), and the same concentrations of these chem- icals dissolved in the seawater also induce settlement and metamorphosis of the larvae. The inability of elaidic acid to activate PKC in vitro and to induce settlement and meta- morphosis is further evidence that PKC activation is in- volved in mediating settlement and metamorphosis of the Capitella larvae. In Capitella larvae. PKC activation may cause several SIGNAL TRANSDLICT1ON IN CAPITELLA BY PKC 195 Initial State Voltage gated Ca++ channel Protein Kinase C (inactive) Illfl JJJJJ JHA/Vw, .- ..-..,,.,... - - . -, ! JHA-activation ; K+ channel blockage Jttft***** Protein Kinase C (active) f Ca++ channel activation Membrane depolarization; closed open Figure 8. Model for the juvenile hormone (JH) signal-transduction pathway in Capitella sp. I larvae. In this proposed model, JH induces settlement and metamorphosis by directly interacting with and activating cytosolic protein kinase C (PKC), which acts as the cellular receptor in the larval chemosensory cells. This activation of PKC then causes the closure of potassium channels by phosphorylation. invoking membrane depolarization, and the subsequent opening of voltage-activated calcium channels. PKC and calcium may in subsequent steps activate transcription factors and protein kinases, leading to changes in gene activity, and also cause neurotransmitter release. cellular events that transduce the external JH signal and lead to settlement and metamorphosis. One well-characterized effect of PKC activation is the modulation of ion transport- ers such as a Na+/K+ ATPase (Ilenchuk and Davey. 1987; Sevala and Davey. 1989), Na+/H+ exchangers (Berridge, 1984), and ion channels (Kaczmarek, 1987; Shearman etai, 196 W. J. BIGGERS AND H. LAUFER 1989b). The modulation of potassium channels, resulting in membrane depolarization and neural relay of external chem- ical settlement cues, has been demonstrated to be involved with the settlement and metamorphosis of several types of marine invertebrate larvae, including those of the abalone Haliotis rufescens (Baloun and Morse. 1984), the nudi- branch Phestilla sibogae and the prosobranch Astraea un- dosa (Yool et /., 1986). the cnidarian Hydractinia (Leitz and Klingman. 1990), and the polychaetes Phragmatopoma califomica (Yool et <;/., 1986), and Hydroides elegans (Carpizo-Ituarte and Hadfield, 1998). Pheromone reception by insects and non-insect species has also in some instances been demonstrated to be mediated by PKC activation and the concommitant modulation of ion channel activity (Zu- fall and Hatt, 1991; Stengl, 1993). Like those studies, our investigations suggest that ion channel modulation is also involved. We found that the addition of excess KC1 to the seawater stimulated settlement and metamorphosis, whereas this effect was negated by simultaneous addition of the potassium channel blocker tetraethylammonium chloride (TEA). These results indicate that TEA blocks settlement and metamorphosis by preventing the entry of excess po- tassium. These results are similar to those reported by Carpizo-Ituarte and Hadfield (1998) in their investigations with larvae of the polychaete Hydroides elegans. TEA did not induce settlement and metamorphosis of the Capitella larvae at the concentrations tested; however, the larvae were induced to settle and metamophose in response to another potassium channel blocker, 4-aminopyridine (4-AP). which causes blockage of outward rectifying potassium currents (Alkon et at.. 1986). These data indicate that blockage of outward rectifying potassium channels can induce settle- ment and metamorphosis of the Capitella larvae. Since studies by Leitz and Klingman (1990) demonstrated that potassium channel closure was involved in mediating set- tlement and metamorphosis of Hydractinia larvae in re- sponse to PKC-activating diacylglycerols, we tested the possibility that JH induces settlement of the Capitella larvae through closure of potassium channels in response to PKC activation. Our findings that the potassium channel iono- phore nigericin inhibits settlement and metamorphosis in- duced by JH III and that 4-AP can directly stimulate settle- ment and metamorphosis provide evidence that JH stimulates settlement and metamorphosis through the clo- sure of potassium channels. Since potassium channel closure may cause membrane depolarization, it is possible that voltage-gated calcium channels are activated during this process and participate in mediating settlement and metamorphosis. Our studies sup- port this idea, since the calcium channel blockers Ni 2 + , Zn 2 + , and verapamil inhibited settlement and metamorpho- sis induced by JH III, whereas the calcium channel iono- phore A23187 induced settlement and metamorphosis. Our interpretation of these data is that juvenile hormones and JH-active chemicals induce settlement and metamor- phosis of Capitella larvae through activation of PKC, which causes closure of potassium channels; the reduced efflux of potassium depolarizes the membranes of the chemoreceptor cells, leading to the opening of voltage-activated calcium channels (Fig. 8). In eliciting metamorphosis, PKC activa- tion by JH may activate transcription factors such as nuclear factor-kB (NF-kB) or stimulate the mitogen-activated pro- tein kinase (MAP-kinase) pathway. Membrane depolariza- tion and calcium influx in PC 12 cells has been demonstrated to activate the MAP kinase pathway (Rosen et al., 1994; Rosen and Greenberg, 1996). PKC modulation also plays a role in regulating neural differentiation of Xenopus (Durston and Otte, 1991), epithelial patterning in Hydra (Shenk and Steele, 1993), and skin morphogenesis in chickens (Noveen ct nl., 1995). The molecular mechanisms by which PKC and calcium may be regulating settlement and metamorphosis of the Capitella larvae are now the subject of our ongoing research. Acknowledgments The authors thank Dr. Judith Grassle at Rutgers Univer- sity for kindly supplying cultures of Capitella sp. I, Dr. David Knecht at the University of Connecticut for help and use of the fluorescent microscope, and Mrs. Mary Jane Spring of the University of Connecticut for providing illus- trations. This research was supported in part by a research fellowship to WJB from the Marine Science Institute of the University of Connecticut and by the Sea Grant College Program, NOAA. Literature Cited Alkon. D. L., M. Kubota, J. T. Neary, S. Naito, D. Coulter, and H. Rasmussen. 1986. C-kinase activation prolongs Ca 2 * -dependent in- activation of K + currents. Biochem. Biophys. Res. Commiin. 13-1: 1245-1253. Amieva, M. R., C. G. Reed, and J. R. Pawlik. 1987. infrastructure and behavior of the larva of Phragmatopoma califomica (Polychaeta: Sa- bellariidae): identification of sensory organs potentially involved in substrate selection. Mar. Biol. 95: 259-266. Baloun, A. J., and I). E. Morse. 1984. Ionic control of settlement and metamorphosis in larval Haliotis rufescens (Gastropoda). Biol. Bull. 167: 124-138. Baxter, G., and D. E. Morse. 1987. G protein and diacylglycerol reg- ulate metamorphosis of plunktonic molluscan larvae. Proc. Natl. Acacl. Sci. USA 84: 1867-1870. Berridge, M. J. 1984. Inositol trisphosphate and diacylglycerol as sec- ond messengers. Biochem. J. 220: 345360. Biggers, W. J. 1994. Effects of juvenile hormone-active compounds on larval settlement and metamorphosis of the polychaete annelid Capi- tella sp. I. Dissertation. University of Connecticut. Storrs. CT. Biggers, VV. J., and H. I.aufer. 1992. Chemical induction of settlement and metamorphosis of Cnpilella capiiala sp. 1 (polychaetal larvae by juvenile hormone-active compounds. Invert. Reprod. De\'. 22: 39-46. Biggers, W. J., and H. Laufer. 1996. Detection of juvenile hormone- active compounds by larvae of the marine annelid Capitella sp. I. Arch. Insect Biochem. Phwtiol. 32: 475-484. SIGNAL TRANSDUCTION IN CAP1TELL\ BY PKC 197 Burke. R. I). 1984. Pheromonal control of metamorphosis in the Pacific sand dollar. Dcmlrnster c\centricns. Science 225: 442 44 .V Butmun. C. A.. J. P. Grassle, and C. M. Webb. 1988. Substrate choices made by marine larvae settling in slill water and in a Hume flow, \iitnre 333: 771-773. Carpizo-Ituarte, K.. and M. G. Hadn'eld. 1998. Stimulation of mela- morphosis in the polychaete Hniroidcs elcguns Huswell (Serpulidae). Bit 'I. Bull. 194: 14-24. Castagna, M., V. Takai, K. Kaibuchi. K. Sano. I'. Kikkana. and Y. Nishizuka. 1982. Direct activation of calcium-activated, phospho- lipid-dependent protein kinase C by tumor-promoting phorbol esters. J. Bial. C/iem. 257: 7847-7851. Chen. C.-S., and M. Poenie. 1993. New fluorescent probes for protein kinase C. J. Biol. C/iem. 268: 15812-15822. (.'hia. F. S.. and M. E. Rice. eds. 1978. Settlement and Metamorphosis of Marine Invertebrate Larvae. Elsevier, New York. C'lKimo. M. C. 1985. Sulfide as a larval settlement cue for Capitella sp. I. Biogeocheinistry 1: 169-181. Oubilier. N. 1987. H 2 S- A settlement cue or toxic substance for Capi- lella sp. 1 larvae'' Biol. Bull. 174: 30-38. Durston. A. J.. and A. P. Otte. 1991. A hierarchy of signals mediates neural induction in Xenopns laevis. Pp. 109-127 in Cell-Cell Interac- tions in Early Development. 3. Gerhart. ed. Wiley-Liss. New York. Eckelbarger, K. J., and J. P. Grassle. 1987. Interspecific variation in genital spine, sperm, and lar\al morphology in six sibling species of Capitella. Biol. Soc. Wash. Bull. 7: 62-76. Grassle. J. P.. and J. F. Grassle. 1976. Sibling species in the marine pollution indicator Capirella (Polychaeta). Science 192: 567-569. Hadn'eld, M. G. 1978. Metamorphosis in marine molluscan larvae: an anaKsis of stimulus and response. Pp. 165-175 in Settlement and Metamorphosis of Marine Invertebrate Larvae. F. S. Chia and M. E. Rice, eds.. Elsevier, New York. Hadn'eld. M. G. 1984. Settlement requirements of molluscan larvae: new data on chemical and genetic roles. Aquaculture 39: 2S3-298. Hidaka. H., M. Inagaki. S. Kawamoto, and Y. Sasaki. 1984. Koquino- linesulfonamides. novel and potent inhibitors of cyclic nucleotide de- pendent protein kinase and protein kinase C. Biochemistry 23: 5036- 5041. Holian, ().. R. Kumar, and L. M. Nyhus. 1989. Fatty acyl esters potentiate fatty acid induced activation of protein kinase C. Biochem. Biophys. Rex. Commnn. 160: 1110-1116. Ilan, J.. J. Ilan, and N. Patel. 1972. Regulation of messenger RNA translation mediated by juvenile hormone. Pp. 43-68 in Insect Juvenile Hormones, Chemistry and Action. J. Menn and M. Beroza. eds. Aca- demic Press. New York. Ilenchuk, T. T.. and K. G. Davey. 1987. Effects of various compounds on Na/K-ATPase activity. JH I binding capacity and patency response in follicles of Rhodnius prolixus. Insect Biochem. 17: 1085-1088. Jensen, R. A., and D. E. Morse. 1990. Chemically induced metamor- phosis of polychaete larvae in both the laboratory and ocean environ- ment. J. Chen,. Ecol. 16: 91 1-930. Jimenez, B., A. Pestana, and M. Fernandez-Renart. 1989. A phos- pholipid-stimulated protein kinase from Dictyoste/ium discoidcinn. Biochem. J. 260: 557-561. Jones. G., and P. A. Sharp. 1997. Ultraspiracle: an invertebrate nuclear receptor for juvenile hormones. Proc. Nat. Acad. Sci. USA 94: 1 3499- 13503. Jones, G., V. Venkalaraman, M. Manczak, and D. Schelling. 1993. Juvenile hormone action to suppress gene transcription and influence message stability. Dev. Gen. 14: 323-332. Kaczmarek, L. K. 1987. The role of protein kinase C in the regulation of ion channels and neurotransmitter release. Trends Nenrosci. 10: 30-34. Komatsu. M.. F.-S. Chia, and R. Koss. 1991. Sensors neurons ot the hipmnaria larva ol the sea star, Litidia senegalensis. Invenebr. Repniil. Dev. 19: 203-2 1 1 . Laut'er, H.. and I). \V. Burst. 1983. Juvenile hormone and its mecha- nism of action. Pp. 203-216 in Invertebrate Endocrinology, Vol. I. Endocrinologv of /mni- / I I I Figure 3. Schematic depiction of ciliary-driven movement of water (shown by arrows) over colony of Me mbranipora membranacea. Ciliary feeding currents produce feeding zones through which water is drawn into the lophophore. Water exits the bases of lophophores between the tenta- cles. Because lophophores are closely spaced, this excurrent flow is di- rected below the array of extended lophophores until it finds a gap, or chimney, where lophophores are not extended; it then escapes through this gap as a jet from colony surfaces. The relatively high velocity of the excurrent jet carries excurrents to some height above the colony surface, which minimizes refiltration since jets are then carried away by boundary layer flow (Lidgard, 1981). Results Flow measurements Flows differed at the three sites as expected, with Site 1 consistently experiencing the lowest flow speeds and Site 3 the highest (Fig. 4). The period of inflow (into Lough Hyne) was much shorter, and thus higher flows were sustained, than during the longer period of slower outflow. The upper plot of Figure 4 shows the overall mean velocities in the \, v, and - directions. In nearly all cases apart from several instances around the time of slack water, the major contrib- utor to flow was in the .v (downstream) direction due to the incoming and outgoing flows, as can be seen by the ranges of flows in each direction for each site during the inflow and outflow periods (Table I). Turbulent kinetic energy density was consistently highest in Site 3 and lowest in Site 1 (see Fig. 4, lower panel), as might be expected from the flow velocities at the three sites, and was greater during inflow. Bryozonn responses to flow M. inemhninucea showed morphological variation for several traits along the flow gradient. Mean zooid lengths decreased from Sites 1 to 4, in association with the increase in mean (and maximum) flow velocities at these sites, although this trend was not quite significant (Kruskal-Wallis test: // ad| = 7.49 1 . df = 3, 0.05 < P < 0. 10; Fig. 5a). There was no change in mean zooid widths among the four sites (Kruskal-Wallis test: W ad| = 0.184, df = 3. P S> 0.05: Fig. 5b). Mean zooid elongation (zooid length/zooid width) sig- nificantly decreased with increased flow (F 3 _ 30 = 4.106. P = 0.015; Fig. 6a), but mean zooid areas (zooid length X zooid width) did not differ among the four sites (F, , = 1.411, P = 0.259: Fig. 6b). These results indicate that zooids become less elongate largely through a decrease in FLOW AND SUSPENSION FEEDING 209 2.5 i 9:36 12:00 14:24 16:48 19:12 21:36 0:00 Time Figure 4. Two characterizations of ambient flow regimes at the algal canopy level for Sites 1-3. Sites are denoted by numbers, with arrows pointing to relevant data. Overall mean velocities (OMV) are plotted in the top graph and were calculated from flow measurements in the .v, y. and ; directions over 5-min intervals during inflow (from 1024-1450) and out- flow (from 1513-2311) at Sites 1-3 (see Fig. I). Bars represent one standard deviation. Values of turbulent kinetic energy density (TKED) are shown in the lower graph. See Materials and Methods section for details on flow measurements and characterization of the flow regimes. length. There must also, however, be some slight increase in width even though zooid width on its own did not vary significantly among sites (but see trend towards increasing width with increased flows; Fig. 5b) since less elongate zooids in sites exposed to faster flows have areas similar to those of zooids in slower flow sites. Although we made no explicit attempt to compare zooid shapes, other than assess- ing zooid elongation, it was apparent that zooids in colonies exposed to the extremely rapid flows at Site 4 were irregular in outline. They were more hexagonal, often with wavering walls, than the normal elongate, fairly rectangular zooid with straight walls (see Fig. 2). This suggests that calcula- tions based on data for maximum length and width may overestimate the zooid areas of colonies under conditions of very rapid flows. Such irregularly shaped zooids may be a character indicative of exposure to flow in M. incm- hranacea. Table I Ranges of mean velucilii'.t nicu\nrcJ in tin 1 \. y. and z directions denoted as u. v. ami u mea\ured at Sites 1-1 during inflow and outflow tidal periods Tidal Period Site Mean u (m s ') Mean v (m s~ Mean w (m s ) 1 Inflow 2 3 0.082-0.564 0.411-0.866 0.548-1.652 0.023-0. 1 26 (-0.026J-0.133 0.022-0.186 0.002-0.117 0.028-0.191 0.199-0.436 1 (-0.056J-O.I33 0.013-0.178 ( -0.0004 )-0.052 Outflow 2 0.089-0.485 0.053-0.181 0.018-0.082 3 (-0.230J-0.954 0.057-0.351 0.024-0.245 Data represent minimum and maximum values obtained over all 5-minute sampling periods conducted at each site during inflow and out- flow. The mean diameter of lophophores decreased signifi- cantly with increased flow (Kruskal-Wallis test: // adj = 13.657. df = 3, P < 0.005: Fig. 7a). Coupled with the decrease in lophophore diameter was a decrease in the number of tentacles per lophophore (G = 48.388, df = 6. P < 0.001 ) (Fig. 8). Frequency analysis on data for number A) c re I T5 B) 1.1 0.9 0.8 0.7 0.6 ^ 0.36 T (10) (11) (10) (3) <"> do, N ' 32 (10) (3) n 0.28 to > O n IA - 1 1 1 1 Site Figure 5. (A) Overall mean zooid lengths (ZL) per colony from colonies collected from Sites I -4. (B) Overall mean zooid widths (ZW) per colony from colonies collected from Sites 1-4. Bars represent one standard deviation: n values in parentheses. 210 B. OKAMURA AND J. C. PARTRIDGE A) o 3.5 T 3 -- (10) (11) (10) C o 1 0) o i (3) 2.5- I N y . i i 1 1 B) 1000 T (10) CM n i) 1 800 - (10) C U I n (i) } 1 600- 3 Ann - i 1 1 spaced and smaller lophophores before they are ejected as jets from colony surfaces. For monitoring growth rates, we were able to relocate 6 of the 8 colonies photographed in Site 1 , and 7 of the 8 colonies photographed in Site 2, but we failed to find any of the 13 colonies photographed in Site 3. We noted that the holes punched for attaching cable ties to algal fronds were often beginning to split along the length of the fronds, presumably due to drag forces, causing the cable ties to cut through and split apart the fronds. Indeed, in several in- stances we re-attached the cable ties to different locations on fronds when we noticed these splits developing. Thus it is likely that the drag forces in Site 3 were great enough to result in the loss of all cable ties and surveyor's tape used to mark the mapped and photographed colonies. ANOVA indicated no significant differences in the initial sizes of colonies retrieved for growth analysis (Site 1: mean initial colony size = 317.5 mm 2 , SD = 186.1, n = 6; Site 2: mean initial colony size = 368.4 mm 2 , SD = 240.7. n = 7; F, ,, = 0.1802, P = 0.6794). Analysis of the growth increments of these colonies revealed no significant differ- ences in the difference between mean final and initial col- Site Figure 6. (A) Overall mean zooid elongation (zooid length/zooid width) per colony from colonies collected from Sites 1-4. (B) Overall mean zooid area (zooid length x zooid width) per colony from colonies collected from Sites I -4. Bars represent one standard deviation; n values in parentheses. of tentacles per lophophore entailed pooling frequencies in the highest and lowest categories to avoid low expected values. Such pooling ensured that no more than 20% (in our case 8.33%) of expected values were less than 5 (in our case one expected value was 3.5) and therefore that the sampling distribution provided a good approximation of the x 1 dis- tribution (Cochran, 1954. cited in Siegel and Castellan, 1988). Smaller lophophores with fewer tentacles should also be shorter, although we made no direct measurements of lophophore height. This expectation is supported by the positive relationship between tentacle number and lo- phophore height in Best and Thorpe's study (1986) of a variety of British bryozoan species, including Membra- nipora, as well as Ryland's overall finding (1975) that all lophophore parameters are related to lophophore size and therefore are intercorrelated. We also found significant de- creases in mean interchimney distances along the flow gra- dient (F^ 2g = 39.442, P < 0.001; analysis on log-trans- formed data to ensure homogeneity of variances; see Fig. 7b). Thus, as ambient flows increase, colonies produce less elongate zooids with smaller lophophores, and excurrent flows travel shorter distances below these more closely A) E - 75 i E. (10) 5> 0.7 - a E ra 0.65 - a 'a. 0.6 - (11) (10) o \ c 055- I i s ft C _ 1 1 1 t 1 1 1 1 1 1 B) -=- 8 T (10) nterchim. Distant 10 * (T [ (11) (10) 1 c Oi 5 . 1 1 1 1 Site Figure 7. (A) Overall mean lophophore (LophJ diameter per colony from colonies collected from Sites 1-4. (B) Overall mean interchimney (Interchim.) distances per colony from colonies collected from Sites 1-4. Bars represent one standard deviation; ;i values in parentheses. 80 -, o O> 20 a N o 0) o> 5 80^ O o S? 60 " (18) c S 40- 5 mS 20 - (18) rt - 1 1 1 1 > 1 Site Figure 10. Mean percentage cover of laminarian fronds by Memhra- nipnrti membranacea colonies in Sites 1-3. Bars represent one standard deviation; n values in parentheses. branacea colonies growing in Sites 1 and 2 suggest that these morphological changes are adaptive, in that they are expressed in colonies that achieve similar growth rates under different flow conditions, presumably because of sim- ilar rates of feeding. At some point, boundary layers will become so thin that further miniaturization is not feasible: however, the continuous decline in size shown by M. mem- branacea colonies over the four sites studied suggests that this critical point had not been reached. Previous field investigations have obtained evidence that increased turbulence can have beneficial effects on growth (Wildish and Kristmanson, 1979; Patterson, 1984; Frechette and Bourget, 1985; Frechette et ai, 1989). However, while increased levels of turbulent kinetic energy density charac- terized the ambient flow regimes measured at the level of the algal canopy at Site 2. these were not reflected in increased growth rates of M. membranacea colonies. It may be that the bryozoans themselves do not experience the higher turbulence because the flexibility of algal fronds in response to three-dimensional flows, and the thickness of the boundary layer, reduces the relative movements of water over frond surfaces. Although Koehl and Alberte (1988) found that increased blade flapping in Nereoc\stis luet- keaiui, the giant bull kelp, stirs water near the blade surface, this effect was greatest at slow ambient flows and is unlikely to apply to the extreme flow conditions experienced by laminarian fronds in Lough Hyne. Alternatively, turbulence may have had no effect on growth in our study if food levels were not limiting. It is notable that the percent cover of M. membranacea colonies decreased dramatically along the flow gradient: values near 100% on some fronds in Site 1 declined to 0% on most fronds by Site 3. Beyond Site 4. where flow velocities were very great (and not measured), no colonies were evident. These patterns could reflect pre-settlement. settlement, or post-settlement events. The lack of colony development in sites exposed to extremely fast flows may be due to detrimental effects of flow on feeding that can no longer be overcome by increased miniaturization. It is also conceivable that larvae settle at much lower rates with increased exposure to flow, either because they actively choose not to settle or because drag forces prevent most of them from adhering to frond substrata. Yet another expla- nation that the higher cover in Site 1 results from differ- ential settlement patterns controlled by proximity to a source population of larvae within Lough Hyne is unlikely because the larvae of M. membranacea are long-lived planktotrophs. Our finding that growth rates were equivalent at Sites 1 and 2 contrasts with Eckman and Duggins' study ( 1993) in which the growth of very young colonies of M. memhrana- cea was reduced at increased flows through pipes. This lack of concordance might be explained by several factors. First, the flows experienced by colonies within pipes could have been greater than those experienced by colonies in the Rapids since the former were growing on acrylic plastic strips whose edges were oriented into the flow through the center of the pipes and which were 2 cm long in the along-stream dimension. However, calculations of turbulent boundary layer thicknesses (Vogel. 1981 ) for the two stud- ies suggest that this was not the case. Turbulent boundary layer thicknesses for slow (mean flow = 0.02 m s~') and fast (mean flow = 0.15 m s~') flows in pipes were 21 mm and 14 mm respectively, assuming a downstream distance of 10 mm from the leading edge of their plastic growth strips. Corresponding thicknesses in Lough Hyne would range from 28 mm to 2 1 mm for slow (mean flow = 0.22 m s" 1 : Fig. 4) and fast (mean flow = 0.87 m s~'; Fig. 4) flows respectively, assuming a downstream distance of 0.5 m from the leading edge of fronds to the colonies. This indi- cates that in all situations, boundary layer thicknesses were much greater than the heights of lophophores; colonies were thus hidden from mainstream flows in both studies. The variation in growth may therefore be related to differences in the steepness of the velocity gradients within these boundary layers. A second, and probably equally important, difference in the two studies is that Eckman and Duggins assessed the growth of juvenile colonies that were less than 2 mm in diameter at the start of the experiments (and hence were probably at the stage of having from 2-8 zooids). Colonies in our studies were not at an early astogenetic stage but had already become established in the field. These differences are illustrated by the very different growth rates in the two studies. Mean growth rates of the small colonies in Eckman and Duggins' study ranged from about 5 to 20 mm 2 d~ ' (see fig. 4 in Eckman and Duggins, 1993) while the equivalent R.OW AND SUSPENSION FEEDING 213 mean growth rates of the larger colonies in our study were 86mnr d~' (SD = 55 mm 2 d~') in Site 1 and I03mnr d~' (SD = 45 mm 2 d~ ' ) in Site 2. It may be that early stages of colony development are more sensitive to flow conditions in microhabitats because arrays of lophophores and excurrent jets are not yet developed. Eckman and Duggins (1993) obtained large differences in growth rates in replicates for each of the two slower flow speeds studied (tig. 4. Eckman and Duggins. 1993). If the higher values obtained for each of these replicates are regarded as outliers due to anomalous conditions in these pipes, there is much less suggestion of a flow effect on growth across the range of the five flow speeds studied. Nevertheless, since at some point drag forces must begin to inhibit feeding, flow will eventually limit growth as Eckman and Duggins' evidence indicates. We feel the most likely explanation for our contrasting results is that colonies differed in developmental stages and experienced different velocity gradients within boundary layer microhabitats. Other consequences of ambient flow regimes As discussed above, miniaturization appears to be an effective means of achieving similar growth rates when colonies happen to become established in sites exposed to rapid flows. However, another consequence of living in such sites is that colonies occur at reduced densities. Thus they will experience a reduction in intraspecific competition for space, allowing individual colonies to attain larger size: it was certainly notable that, at high percent cover, colonies tended to be smaller, although we did not collect data to document this. A possible detrimental effect of low cover is a reduction in the likelihood of achieving fertilization. M. memhranacea releases sperm packets into the water column, where they become entrained by feeding currents. Once within the lophophore region they gain access to maternal zooids through specialized intertentacular organs of the lo- phophore. and subsequently effect internal fertilization of eggs within maternal zooids (Temkin. 1994). Although proximity to neighboring colonies may increase fertilization success, it is also possible that release of sperm packets into the rapid and turbulent flows in the Rapids swamps this effect, particularly since sites of high and low density are within meters of each other. We have no data on the reproductive output of colonies growing in the different ambient flow conditions to resolve this issue. Although miniaturization allows similar growth rates, it could con- ceivably inhibit or promote reproduction if more or less energy were allocated to growth. Also unknown is the extent to which these effects may be offset or otherwise influenced by the possibility of reaching larger colony sizes due to decreased intraspeciric competition for space. Colony hydraulic* Venting of excurrent flows as jets through chimney re- gions allows M. memhranacea to avoid or at least reduce refiltering of particle-depleted water. Chimneys in M. mem- branacea represent small regions in which zooids degener- ate and are not replaced. Thus the potential contribution to feeding made by these zooids is sacrificed, presumably because the energetic gain conferred by minimizing refil- tration exceeds the cost of zooid degeneration (Lidgard. 1981). One mechanism postulated to explain the formation of chimneys is that chimneys develop when the pressure beneath the extended array of lophophores becomes greater than the array can withstand (Dick. 1987). The decrease in interchimney distances that accompanies the miniaturiza- tion of zooids in M. memhranacea is consistent with this explanation. Smaller and shorter lophophores provide less space for excurrent water below the lophophore array. Thus, pressure develops more quickly and excurrents must be vented sooner. The notably regular spacing of chimneys in M. membranacea is also consistent with a hypothesis of pressure build-up under the lophophore sheet (Dick. 1987). This postulated mechanism of pressure build-up provides a simple explanation for how M. membranacea can main- tain functional colony-wide hydraulics across a range of ambient flow conditions, as suggested by the evidence pre- sented here. The explanation that chimneys arise at partic- ular developmental stages in colonies (i.e., are astogeneti- cally determined) is less satisfactory. If chimneys are astogenetically determined, interchimney distances might be expected not to vary when growth is equivalent. How- ever, while growth for colonies at Sites 1 and 2 was not significantly different, chimneys were produced much more frequently in Site 2. Implications of stuclv We have found that bryozoans show flow-dependent morphologies that result in a miniaturization of colonies with exposure to rapid flows. These morphological changes likely promote exposure to similar flow microhabitats. thereby allowing the maintainenance of similar rates of feeding, and hence growth, at least up to a point. For several reasons, we suggest that these morphological changes are the result of phenotypic plasticity rather than selection for colonies with particular morphologies. Firstly, tidal flushing through the Rapids ensures that algal fronds will be subject to recruitment by the same larval pool. Secondly, there is evidence that flow regime limits active choice in habitat selection by larvae (Butman. 1987), and as all sites were subject to considerable flow, it is unlikely that larvae could discriminate amongst them. Thirdly, post-recruitment inter- actions are also unlikely to have selected for certain types. Colonies were chosen to be free from overgrowth interac- tions, and there was no evidence of predation. 214 B. OKAMURA AND J. C. PARTRIDGE The only obvious factor that could result in post-settle- ment mortality was flow regime itself: increased flow could have selected for miniaturized colonies. Unfortunately, we did not conduct transplant experiments; thus selection can- not be distinguished from phenotypic plasticity. However, in general, algal fronds should offer great variation in ex- posure to flow for encrusting bryozoans: as fronds grow, the colonies are placed in different positions both along the length of the frond and in the water column. This is partic- ularly true since larvae of M. membranacea preferentially recruit to the younger regions of laminarian fronds, near the holdfasts of the plants (Brumbaugh el /., 1994). Further- more, colonies may be buffered from flows when they are growing on dense fronds within macroalgal stands but may be more exposed to flows when growing on fronds along the edges of stands. These arguments suggest that genetic spe- cialization to narrow ranges of flow regimes is an unlikely strategy to have evolved. Rather, colonies should be able to tolerate a wide range of flow conditions since it cannot be predicted what conditions will be encountered during their lifetime. This tolerance can be achieved by phenotypic plasticity. In the Introduction, we cited some of the many studies of feeding from flow for a variety of suspension feeders; these have provided important insights about feeding opportunities (e.g., increased turbulent mixing) and con- straints (e.g.. increased drag forces) presented by in- creased flow. However, a common practice in these stud- ies has been to assess feeding across a range of flows by suspension feeders collected from particular sites (but see Okamura, 1992, and Anthony, 1997). Hence flow-depen- dent feeding rates of suspension feeders which them- selves inhabit a range of flow environments have re- ceived little investigation, and therefore the extent to which suspension feeders may be locally adapted to flow conditions is poorly known. Thus, although studies to date reveal important information about the general ef- fects of flow on feeding for a given situation, their results may not depict what happens in the real world since they do not take into account the longer term responses to flow by the animals in the field. However, even if many suspension feeders show plasticity in growth in response to flow, flow and feeding studies of suspension feeders collected from relatively restricted ranges of flow envi- ronments are nonetheless informative. Such studies should provide predictions about what kinds of morpho- logical change might be adaptive under different flow conditions and may lead to insights about the significance and consequences of observed morphological variation. Acknowledgments We thank Linda Teaglc . v her great help in setting up and collecting data using the ultrasonic flow meter and for assisting in the field; Allison Myles for help collecting colonies during the short periods of low flow in the Rapids; the Department of Zoology of the University College Cork for laboratory space; and the National Parks & Wildlife Service, Office of Public Works, Dublin, for permission to work at Lough Hyne. This work was partially supported by a grant from the Royal Society to B. Okamura. which provided funds to purchase the ultrasonic flow meter. Com- ments from two anonymous referees have helped to im- prove the manuscript. Literature Cited Anthony, K. R. N. 1997. Prey capture by the sea anemone Metridiuin senile (L.): effects of body size, flow regime, and upstream neighbors. Biol. Bull. 192: 73-86. Bassindale, R., F. J. Ebling, J. A. Kitching, and R. D. Purchon. 1948. The ecology of Lough Ine Rapids with special reference to water currents. I. Introduction to hydrography. J. Ecol 36: 305-322. Best, B. A. 1988. Passive suspension feeding in a sea pen: effects of ambient flow on volume rate and filtering efficiency. Biol. Bull. 175: 332-342. Best, M. A., and J. P. Thorpe. 1986. Effects of food concentration on feeding current velocity in six species of marine Bryozoa. Mar. Biol. 93: 255-262. Brumbaugh, D. R., J. M. West, J. L. Hintz, and F. E. Anderson. 1994. Determinants of recruitment by an epiphytic marine bryozoan: field manipulations of flow and host quality. Pp. 287-301 in Reproduction and Development of Marine Invertebrates. W. H. Wilson. Jr.. S. A. Strieker, and G. L. Shinn, eds. The Johns Hopkins University Press, Baltimore, MD. Butman, C. A. 1987. Larval settlement of soft-sediment invertebrates: the spatial scales of pattern explained by active habitat selection and the emerging role of hydrodynamical processes. Oceanogr. Mar. Biol. /twin. Rev. 25: 113-165. Clifford, N. J., and J. R. French. 1993. Monitoring and modelling turbulent flows: historical and contemporary perspectives. Pp. 1-34 in Turbulence: Perspectives in Flow and Sediment Transport. N. J. Clif- ford. J. R. French, and J. Hardisty, eds. John Wiley, Chichester, UK. Dick, M. H. 1987. A proposed mechanism for chimney formation in encrusting bryozoan colonies. Pp. 73-80 in Br\o~oa: Present and Past. J. R. P. Ross, ed. Western Washington University, Bellingham. WA. Eckman, J. E., and D. O. Duggins. 1993. Effects of flow speed on growth of benthic suspension feeders. Biol. Bull. 185: 28-41. Eckman, J. E., C. H. Peterson, and J. A. Cahalan. 1989. Effects of flow speed, turbulence, and orientation on growth of juvenile bay scallops Argopecten irradians concentricus (Say). J. Exp. Mar. Biol. Ecol. 132: 123-130. Frechette, M., and E. Bourget. 1985. Food-limited growth of M\tilus edulis (L.) in relation to the benthic boundary layer. Can. J. Fish. Ai/uut. Sci. 42: I 166-1 170. Frechette, M., C. A. Butman, and W. R. Geyer. 1989. The importance of boundary-layer flows in supplying phytoplankton to the benthic suspension feeder, Mytilus edulis L. Limnol. Oceanogr. 34: 19-36. Grizzle, R. E., R. Langdon, and W. H. Howell. 1992. Growth re sponses of suspension-feeding bivalve molluscs to changes in water flow: differences between siphonate and non-siphonate taxa. J. Exp. Mm: Biol. Ecol. 162: 213-228. Griinbaum, D. 1995. A model of feeding currents in encrusting bryo- zoans shows interference between zooids within a colony. J. T/ieor. Biol. 174: 409-425. Judge, M. L., and S. F. Craig. 1997. Positive flow dependence in the FLOW AND SUSPENSION FEEDING 215 initial culoni/alion of a fouling community: results from in situ water current manipulations. J. Exp. Mar. Bin/. Ecol. 210: 209-222. Kirbj -Smith. \V. \V. 1972. Growth of the bay scallop: the influence of experimental water currents. J. Exp. Mar. Bin/. Ecol. 8: 7-18. Korhl. M. A. R. 1977. Effects of sea anemones on the flow forces they encounter. J. Exp. Bio/. 69: 87-105. Koehl, M. A. R.. and R. S. Alberte. 1988. Flow, flapping, and photo- synthesis of Nereocystis luetkeana: a functional comparison of undu- late and flat blade morphologies. Mar. Bin/. 99: 435-444. I.enihan, H. S., C. H. Peterson, and J. M. Allen. 1996. Does flow speed also have a direct effect on growth of active suspension-feeders: an experimental test on oysters. Limnol. Oceanogr. 41: 1359-1366. Leonard, A. B., J. R. Strickler, and N. D. Holland. 19X8. Effects of current speed on filtration during suspension feeding in Oligometra serripinna (Echinodermata: Crinoidea). Mar. Bin/. 97: 111-126. I.idgard, S. 1981. Water flow, feeding, and colony form in an encrusting cheilostome. Pp. 135-142 in Recent and Fossil Bryozoa, G. P. Larwood and C. Nielsen, eds. Olsen and Olsen. Fredensborg, Denmark. McFadden, C. S. 1986. Colony fission increases particle capture rates of a soft coral: advantages of being a small colony. J. Exp. Mar. Bin/. Ecol. 103: 1-20. Nowell. A. R. M., and P. A. Jumars. 1984. Flow environments of aquatic benthos. Anna. Re\: Ecol. Syst. 15: 303-328. Okamura. B. 1984. The effects of ambient flow velocity, colony size, and upstream colonies on the feeding success of Bryozoa. I. Bugula stolonifera Ryland. an arborescent species. J. Exp. Mar. Bin/. Ecol. 83: 179-193. Okamura. B. 1985. The effects of ambient flow velocity, colony size, and upstream colonies on the feeding success of Bryozoa. II. Conopeum reticulum (Linnaeus), an encrusting species. J. Exp. Mar. Bml. Ecol. 89: 69-80. Okamura, B. 1990. Particle size, flow velocity, and suspension-feeding by the erect bryozoans Bugula neriiina and B. stolonifera. Mar. Bin/. 105: 33-38. Okamura. B. 1992. Microhabitat variation and patterns of colony growth and feeding in a marine bryozoan. Ecology 73: 1502-1513. Palumbi. S. R. 1984. Tactics of acclimation: morphological changes of sponges in an unpredictable environment. Science 225: 1478-1480. Patterson, M. R. 1984. Patterns of whole colony prey capture in the octocoral, A/cyoniiim siilerium. Bin/. Bull. 167: 613-629. Ryland, J. S. 1975. Parameters of the lophophore in relation to popula- tion structure in a bryozoan community. Pp. 363-393 in Proceedings of the 9"' European Marine Biology Symposium. H. Barnes, ed. Aberdeen University Press, Aberdeen. Scotland. Seed, R. 1976. Observations on the ecology of Membranipora (Bryozoa) and a major predator Doridella steinhergae (Nudibranchiata) along the fronds of Laminaria saccharine! at Friday Harbor. Washington. J. Exp. Biol. Mar. Ecol. 24: 1-17. Siegel, S., and N. J. Castellan, Jr. 1988. Nonparametric Statistics for the Behavioral Sciences. McGraw-Hill, New York. 399 pp. Sponaugle, S.. and M. LaBarbera. 1991. Drag-induced deformation: a functional feeding strategy in two species of gorgonians. J. Exp. Mar. Biol. Ecol. 148: 121-134. Temkin, M. H. 1994. Gamete spawning and fertilization in the gymno- laemate bryozoan Membranipora membranacea. Biol. Bull. 187: 143- 155. Vogel, S. 1981. Life in Moving Fluids: the Physical Biology of Flow. Princeton University Press. Princeton, NJ. 352 pp. Wildish, D. J., and D. D. Kristmanson. 1979. Tidal energy and sublit- toral macrobenthic animals in estuaries. J. Fish. Kes. Board Can. 36: 1197-1206. Wilkinson, C. R., and J. Vacelet. 1979. Transplantation of marine sponges to different conditions of light and current. J. Exp. Mar. Biol. Ecol. 37: 91-104. Reference: Binl. Bull. 196: 216-226. (April 1999) Genetic Relationships (RAPD-PCR) Between Geographically Separated Populations of the "Cosmopolitan" Interstitial Polychaete Hesionides gohari (Hesionidae) and the Evolutionary Origin of the Freshwater Species Hesionides riegerorum HARTMUT SCHMIDT AND WILFRIED WESTHEIDE* Spezielle Zoologie, Biology/Chemistry, University of Osnahriick. 49069 Osnabriick, Germany Abstract. In an analysis of the population genetics of the tiny meiofaunal polychaete Hesionides gohari. the RAPD- PCR method was applied to 49 specimens from 7 collecting sites far apart on three continents: French Atlantic coast, Mediterranean (Majorca, Giglio, Crete), Red Sea, Indian Ocean (Phuket), and U.S. Atlantic coast (Florida). In the band patterns produced with 14 arbitrary decamer primers, 496 genetic characters were detected. Genetic distances between the H. gohari populations vary between 0.55 and 0.70. The data were evaluated by three cluster programs; in the almost congruent phenograms, three clades were found with high bootstrap values: ( 1 ) European Atlantic-Mediter- ranean-Red Sea. (2) Indian Ocean, (3) Western Atlantic. In all cluster analyses. Hesionides riegerorum from a U.S. east coast river system is shown as genetically nearest to the Florida specimens of H. gohari, making it most probable that this freshwater species of the genus originated from a Western Atlantic H. gohari population. The genetic dis- tances detected between the H. gohari specimens from the three continents are almost identical to those found between morphologically similar interstitial polychaete species pairs. Thus, the degree of genetic consistency is considered not to be high enough to corroborate the notion of a cosmopolitan distribution pattern, but rather suggests that the three clades represent different species. Received 9 September 1998; accepted 29 January 1999. * To whom correspondence should he addressed. E-mail: westheide(S' mail.biologie.uni-osnabrueck.de Introduction The many species of Hesionides Friedrich, 1937, are highly characteristic meiofaunal polychaetes with a great variety of morphological, reproductive, and behavioral ad- aptations to the habitat of interstitial crevices in marine surf-beaten sand beaches (Westheide, 1967, 1971, 1984). Several of the Hesionides species are known to be charac- terized by another pecularity of the marine interstitial fauna a broad geographic distribution, comprising inter- tidal localities throughout the world (Westheide, 1971 : Ster- rer. 1973; Giere, 1993). Within the genus, H. arenaria Friedrich, 1937, and H. gohari Hartmann-Schroder, 1960, appear to have the most cosmopolitan distribution (West- heide, 1977); they occur on all continents except Antarctica. H. gohari was discovered by Adolf Remane in littoral sands near Ghardaqa (today: Hurghada) in the Red Sea (Remane and Schulz, 1964); it was described by Hartmann-Schroder (1960). Westheide (1970, 1972a, 1972b) investigated the morphology, reproductive biology, and local distribution pattern of populations occurring in beaches of northern Tunisia (Mediterranean Sea). This tiny species (max. length 1.5 mm) is typical of warm seas and has a range extending as far north as Arcachon, on the Atlantic coast of France (Westheide. 1972/73). New localities for the species are still being recorded worldwide (Hartmann-Schroder, 1991; Westheide. 1992). Its most characteristic diagnostic features are considered to be its length, the proportions of the head appendages, the position of penis papillae, the dentation of notopodial chaetae, and especially the shape of the anal appendages. Among the other Hesionides species of comparable size, one inhabits 216 POPULATION GENETICS OF A MEIOFAUNAL POLYCHAETE 217 Figure 1. Records (open circlesl and sampling sites (solid circles) for Hesionides gohari and sampling site for H. riegemrum (triangle). fresh water, H. riegerorum Westheide, 1979: it can be distinguished from H. gohari by various morphological features. The demonstrated almost cosmopolitan distribution of these Hesionides species, together with that of many other species of various meiofaunal taxa, gave rise to a still- controversial hypothesis of Sterrer (1973) (see also Rao, 1972) based on the notion that speciation of these forms is extraordinarily slow. It is postulated that they were already present on an old supercontinent and were distributed over their present vast range by the drifting apart of the conti- nental plates. This explanation could apply only if the genetic changes in the separated populations were relatively small over long periods of time, and mostly had no effect on the phenotype. Such minimal genetic variability can be explained by assuming a complete constancy of the ecolog- ical factors in the habitat of these animals, the sand beaches. Sterrer' s hypothesis appeared necessary because long- range, transoceanic dispersal of these meiofaunal organisms seemed inconceivable: they are not capable of active swim- ming and, with few exceptions, have no larval dispersal stages. However, it gradually became evident that, even without dispersal stages, many of these species do succeed in colo- nizing geologically young islands far from any coast (West- heide. 1991): furthermore, dispersal of meiofaunal individ- uals in the seawater column along shores has been observed and demonstrated experimentally (Hagerman and Rieger, 1981; Palmer, 1988; Armonies. 1989). These observations led to the proposal that the cosmopolitan distribution pattern is ascribable not to geologically ancient processes but rather to occasional contemporary events (Sterrer, 1973: Gerlach. 1977) involving long-range dispersal by birds, on drifting material, or in ballast sand or tanks in ships. Lately, arguments for one or the other view have been supported by genetic investigations. Todaro el al. (1996) expressed the conviction, based on their restriction-frag- ment length polymorphism analyses, that the proposed cos- mopolitan species Xenotrichula intermedia (Gastrotricha), apparently present both in the Mediterranean Sea and along the coast of North America (Ruppert, 1977), consists of more than one taxon. Similarly, with RAPD-PCR (random applied polymorphic DNA-polymerase chain reaction) Soosten et til. (1998) found genetic differences between specimens of the polychaete Petitia amphophthalma from Europe and North America, although these differences are considerably smaller than those between certain morpho- logically distinct species of other taxa. In the present study of Hesionides gohari collected from seven intertidal local- ities on three continents (Europe, North America, Southeast Asia; Fig. 1 ), the RAPD-PCR method was likewise chosen, to evaluate the results obtained for P. amphophthalma with 218 H. SCHMIDT AND W. WESTHEIDE Table I Collection information for H. gohari and H. riegerorum Species Locality Abbreviation Number of specimens H. gohari Atlantic Ocean France Arcachon (intertidal) A 11 H. gohari Mediterreanean Sea Spain. Majorca Palma (subtidal) M 8 H. gohari Mediterranean Sea Italy, Giglio Campese (subtidal) G 8 H. gohari Mediterranean Sea Greece, Crete Heraklion (intertidal) C 8 H. gohari Red Sea Egypt Hurghada (intertidal) H 4 H. gohari Indian Ocean Thailand Phuket (intertidal) T 4 H. gohari Atlantic Ocean Florida Ft. Pierce (intertidal) F 6 H. riegerorum Chowan River North Carolina Edenhouse (riverbank) NC 5 those of other cosmopolitan species studied in our labora- tory (Schmidt and Westheide, 1998; Schmidt unpubl. re- sults). For general advantages of the RAPD-PCR method in molecular systematic and ecological studies, see Hadrys et al. (1992). Schierwater (1995). and Schirmacher et al. (1998). Materials and Methods A total of 49 individuals of He sionides gohari Hartmann- Sehroder, 1960, from seven localities and 5 specimens of//. riegerorum Westheide, 1979, from North Carolina were examined genetically (Table I). The meiofaunal animals were extracted from small samples of sand by the magne- sium chloride method (Westheide, 1990). After sorting and washing in doubly distilled water, the specimens were either frozen in LTE-buffer at -60C or dried at temperatures below 90C. Both methods of preservation guaranteed re- producible results even when samples were stored for months prior to actual DNA isolation. Total DNA was isolated by a modification of the method of Kocher et al. (1989). The DNA was extracted from frozen tissues ( 70C) in 1.5-ml microtubes by digestion in 100 /xl 100 mM Tris HC1, pH 8.0, 10 mM EDTA, 100 mM NaCl, 0.1% SDS, 50 mM dithiothreitol and proteinase K (0.5 p-g/ml) incubated for 2 h at 50C in a heating block. SDS was precipitated by 40 /nl 3 M potassium acetate cooled down to 0C for 10 min. The DNA was purified by extracting once with 150 /j,l phenol/chloroform/isoamyl al- cohol (25:24:1, saturated with 10 mM Tris, pH 8.0, and 1 mM EDTA) and once with 150 /xl chloroform/isoamyl alcohol (24:1). After precipitation with 150 jul 100% iso- propanol, the DNA was washed with 100 ^\ 70% ethanol (both refrigerated at 0C). The phases were separated by centrifugation at 14,000 rpm for 10 min at 4C. The dried DNA was dissolved in 1 mMTris HC1, pH 8.0, plus 0.1 mM EDTA and stored at 4C. Fourteen arbitrary 10-mer primers from Operon Technol- ogies (Alameda, California) listed in Table II were used for PCR amplification following the recipes reported by Wil- liams et al. (1990) with minor modifications. PCR was done in a total reaction volume of 25 ju.1 with the following compounds: 10 mM Tris HC1, pH 8.3: 50 mM KC1; 2 mM MgCU: 1% Triton X-100; 100 juM each of dATP, dCTP, dGTP. and dTTP (Boeringer); 5 pM decamer primer; 1-5 ng DNA; and 0.4 U Tbr Polymerase (Biometra Prime Zyme Polymerase). Controls were run with no template DNA. Amplification was done simultaneously by two thermocy- clers: Perkin Elmer Cetus DNA thermal cycler 480 and Biometra personal cycler 48. All synthesized fragments patterns were comparable. Each PCR cycle consisted of denaturation for I min at 94C, hybridization for 1 min at 36C, and extension for 2 min at 72C. This cycle was repeated 45 times followed by a paused file at 4C. Fastest available transitions between temperatures were employed in each case. Reaction products were separated by electro- phoresis in 1.5% agarose gels buffered in 1 X TBE (0.045 M Tris-borate, 1 mM EDTA) at 3 V/cm for 2.5 h, stained with ethidium bromide, and documented under UV light. As size markers, 100-base-pair DNA ladders were used (from Life Table II Decamer primers tint/ their sequences used in this study (all primers from Operon Technologies, Inc., Alameda, CA) Primer Primer sequence 5' to 3' Molecular weight OPB-01 GTTTCGCTCC 2961 OPB-02 TGATCCCTGG 3010 OPB-03 CATCCCCCTG 2915 OPB-04 GGACTGGAGT 3099 OPB-05 TGCGCCCTTC 2946 OPB-06 TGCTCTGCCC 2946 OPB-07 GGTGACGCAG 3048 OPB-08 GTCCACACGG 3004 OPB-10 CTGCTGGGAC 3035 OPB-11 GTAGACCCGT 3019 OPB-12 CCTTGACGCA 2779 OPB-15 GGAGGGTGTT 3130 OPB-17 AGGGAACGAG 3117 OPB-18 CCACAGCAGT 2988 POPULATION GENETICS OF A MEIOFALINAL POLYCHAETE 219 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Figure 2. RAPD fingerprinting of Hesionides gohari specimens from Phuket (2. 31, Florida (8-1 1), Crete (12-151. Giglio(16-18). Arcachon ( 19-23). and H. riegerorum specimens from North Carolina (4-7); 1 and 24; 100 bp marker. Primer OPB 6. Technologies, Inc., Berlin, Germany, or Pharmacia Biotech, Uppsala, Sweden). The reproducibility of the data was checked at regular intervals on different levels throughout all three parallel experiment lines. First, at the beginning of each series the optimal DNA concentration for PCR was determined for the individual animal, by testing the PCR with three DNA concentrations per animal (cci. 1 ng, 3 ng, and 10 ng per 25 jul reaction volume). In pilot experiments with larger annelids, it became clear that the results were reproducible with up to 25 ng/25 /u,l reaction volume. With DNA concentrations as high as ca. 50 ng/25 /u.1 reaction volume, reproducibility was distinctly worse. A further in- crease in DNA concentration (to over 100 ng/25 ju.1 reaction volume) brought the PCR to a complete halt, so that no specific DNA fragments could be detected. Reproducibility was tested further throughout each test series by running the same experiment twice in parallel reactions. In addition, for each series one reaction was carried out with a "blind" sample lacking DNA, to check the possibility of contami- nation with foreign DNA. We also looked for differences in the amplification patterns that might derive from the use of different thermocyclers (Biischer et al., 1993; He et al.. 1994) or preservation methods (drying or deep-freezing), but found none. The detected amplification product patterns were exam- ined visually for monomorphic and polymorphic markers (Hadrys et al.. 1992). The degrees of polymorphism are given in percentages. The banding patterns were then trans- lated into a 0/1 -matrix (0 for absence, 1 for presence of a specific DNA marker) and fed into the cluster analysis program TREECON 1.2 (Peer and Wachter, 1994), which also transformed the data into distance values in percentages (Nei and Li, 1979). Cluster analyses were carried out by UPGMA (unweighted pair-group method using arithmetic averages; Sneath and Sokal, 1973), single linkage (Sneath and Sokal. 1973), and neighbor joining (Saitou and Nei, 1987) including bootstrap proportions (Felsenstein. 1985). Results For the seven sampled populations of H. gohari and one population of H. riegerorum, with 14 different primers a total of 496 different DNA fragments were detected, rang- ing in length from 150 bp to 1900 bp (e.g., Fig. 2). The calculated degree of polymorphism is similarly high in all the populations tested: for Arcachon, 96%; Giglio, 94%; Majorca, 95%; Crete, 94%; Hurghada, 89%; Phuket, 89%; Florida, 97%; and for H. riegerorum of North Caro- lina, 86%, the lowest value. For none of the European populations, nor for the population at the Red Sea, could diagnostic DNA fragments be detected. However, if the European and Egyptian animals are considered as a single group, its members are found to exhibit 33 common poly- morphic characters that are absent from the individuals living in Thailand and Florida. With few exceptions, char- acters present in the animals from Thailand are the same as those found in the European and Egyptian individuals. Three diagnostic and five polymorphic DNA fragments were observed in the Phuket population only. Only one characteristic polymorphic character was detectable in the animals from Florida, and none of the primers revealed diagnostic DNA fragments. A surprising result was ob- tained by comparing the population of H. riegerorum with those of H. gohari. Although this freshwater species can be distinguished from the H. gohari populations by 12 diag- nostic DNA fragments and 9 polymorphic characters, all other DNA bands are common to the two groups. There is an especially close match with the animals of the Florida population: of the 191 shared characters, 39 polymorphic DNA fragments are present only in these individuals. The 220 H. SCHMIDT AND W. WESTHEIDE Table III Comparison of genetic distances generated from RAPD data (after Nei and Li, 19791 within and betiveen the seven thf sole population of Hesionides riegerorum Hesionides gohari populations and Locality Arcachon Giglio Majorca Crete Hurghada Florida N. Carolina Phuket H. rieg. Arcachon x = 0.53 0.43-0.6 Giglio x = 0.57 x = 0.50 0.48-0.67 0.38-0.61 Majorca x = 0.58 x = 0.55 x = 0.53 0.51-0.66 0.43-0.67 0.45-0.61 Crete x = 0.56 x = 0.55 x = 0.55 x = 0.50 0.48-0.63 0.46-0.65 0.47-0.61 0.37-0.58 Hurghada x = 0.60 x = 0.59 x = 0.61 x = 0.6 x = 0.5 1 0.5-0.68 0.53-0.64 0.54-0.69 0.54-0.68 0.46-0.61 Florida x = 0.70 x = 0.70 x = 0.68 x = 0.68 x = 0.70 x = 0.58 0.63-0.78 0.6-0.77 0.59-0.75 0.62-0.75 0.65-0.76 0.51-0.56 Phuket x = 0.68 x = 0.65 x = 0.68 x = 0.64 x = 0.69 x = 0.70 x = 0.50 0.62-0.76 0.56-0.75 0.6-0.73 0.57-0.74 0.62-0.76 0.64-0.77 0.47-0.53 N. Carolina x = 0.78 x = 0.79 x = 0.78 x = 0.77 x = 0.77 x = 0.68 x = 0.75 x = 0.40 H lieg. 0.73-0.86 0.70-0.87 0.69-0.78 0.71-0.83 0.72-0.83 0.56-0.77 0.70-0.81 0.28-0.47 Values are means (\1 followed bv min/max ranges. average genetic distances and the range of variation within and between the seven H. gohari populations are shown in Table III, as is their genetic relation to the population of//. riegerorum in North Carolina. The genetic distance values within the European. Egyp- tian, and Thai populations are roughly the same, but a somewhat higher value was obtained for the animals from Florida. The genetic distances among the individuals of H. riegerorum are distinctly less than the values for all the H. gohari populations. The European populations, including the animals from Hurghada. bear the closest genetic resem- blance to one another, and each of them is about the same distance away from the population of Thailand and from that of Florida. Comparison between H. riegerorum and the H. gohari populations gave an unexpected result (Table III): whereas H. riegerorum is. unsurprisingly, relatively distant genetically from the populations of Europe, Egypt, and Thailand, its genetic distance from the population of Florida is less. These findings become particularly clear when the asso- ciated phenograms are constructed (Figs. 3-5). With all three analytical procedures, the individuals of the following geographic regions form groupings with high bootstrap values: (1) European Atlantic coast ( Arcachon )-Mediterra- nean (Majorca, Giglio and Crete )-Red Sea; (2) Indian Ocean (Phuket); (3) Western Atlantic coast (Florida); (4) H. riegerorum (Chowan River, North Carolina). The European H. gohari. including the individuals from the Red Sea, form a common cluster in the phenograms, the common root of which has a bootstrap value ranging from 76 (single link- age) to a maximum of 98 (UPGMA). Only with UPGMA (see Fig. 3) can the animals from the Red Sea be distinguished as a cluster (bootstrap value <70); with neighbor joining (Fig. 4) and single linkage (Fig. 5) they fall within the European cluster but still form a boot- strap-supported cluster of Red Sea animals. The individuals from the French Atlantic coast form a common cluster (bootstrap value <70) only with UPGMA, and this cluster lies between the Mediterranean populations. In all the trees, the animals from Thailand are separate from the European- Egyptian group. All cluster analyses distinguish H. riegerorum from the H. gohari of Florida, but together, the Florida H. gohari and the H. riegerorum always form a cluster that is separate from the individuals of the H. gohari populations. That is, according to the present results, animals living in regions close together geographically are genetically more similar, although they belong to two morphologically distinct spe- cies, than individuals that are considered to be of the same species but originate in regions far apart from one another. Discussion The European populations of H. golmri treated in the present study (from Arcachon into the Mediterranean re- gion) are relatively similar genetically, in some cases form- ing intermingled clusters, but clustering differently in the different analyses. It is therefore extremely likely that these individuals belong to a single species. The animals from the Red Sea form a distinct cluster in only one of these analy- ses; treated with either of the two other procedures, the four individuals in this sample are positioned among the Euro- POPULATION GENETICS OF A MEIOFAUNAL POLYCHAETE 221 0.1 98 r H, /""ft Europe Red Sea Thailand North Carolina Florida .g.Oo [.g.ivio II n IVyfQ H.g.IVlO 11 /~"T H.g.OV H~ /~1 7 .g.OJ H.g.Ul 1 II 7. f> H.g.GZ H.g.lvlZ 96 H a CA i n ri H.g.LiS II \/i i H.g.lVll 1 1 n MT H.g.lVI / QA II n f^ii o, H.g.CO OO | ^-1*7 1 H.g.C / Hn /""ft .g.*_o |I /Tl H.g.LJ 96 Ho C II JT /"'I M.g.cn 97^ II o n 1 *~"\ rl.g.l^x II _ H* 5 i H.g.IVlj M.g.iviri II A l 11 \A 91 H o A1 II A *f H.g.AZ EJ A4E H.g.AS .g.AlU H_ A 1 1 .g.All II n A/L H_ A T .g.A / 93 .g.Ao Hn A Q .g.AV H \JfA .g.M4 H_ LJ1 H.g.H4 H.g.HZ M.g.MJ ..,.,, l| n XI 100 H.g. 1J H.g. I 2 II r. XI n.g. 1 1 1 II T4 H.g. I 4 H^ M/^l .r.rNLI '^ 1 II .- M*^'> 100 ii .. i\r-i 73 1 nn ii t- ivr^4 H-. XJ/^C 98 .g.r5 Hn i?/; .g.ro 71 II n 171 M.g.rj 7 H_ 17/1 .g.r4 3 II n IT1 .g.rZ Figure 3. UPGMA phenogram of Hesiunides riegemriim specimens (H. r.) from North Carolina (NC) and H. goliari specimens (H. g.) from the local populations in Arcachon (A), Giglio, Mediterranean (G). Majorca, Mediterranean (M). Crete, Mediterranean (C), Hurghada. Red Sea (H), Florida (F), Phuket, Thailand (T). Only bootstrap values above 70 are indicated. pean individuals. Their genetic distances are generally not much different from those within the European group, so that they too are very probably members of the same spe- cies. The fact that the Red Sea site where they were col- lected, Hurghada, is the type locality of Hesionides gohari Hartmann-Schroder. 1960, presumably makes it highly likely that all animals of this cluster are of this species. The relatively slight genetic distance between Red Sea animals and Mediterranean animals indicates either that gene exchange occurred between previously isolated popu- 222 H. SCHMIDT AND W. WESTHEIDE 0.1 94_ 94 \ H o MS Europe and Red Sea Thailand H.r.NCJ H r NC2 M ^u P II o Ml 1 1 Ho Ml 1 Ho M1 & Ho M7 07 11 o PS Ho MS H o Mfi U II o P7 ll.g.U/ H HO ft. OR II o P7 oj n.g.v, / R 3 I H o Pfi H.g.CS T; H n P? H P PI 1 II PI OR Ho PS 1 || f 'd H o H4 n.g.n< P II o III n.g.nj 72 1 ii HI H HI Ho At 1 Ho AID H A7 n.g.A / 1 H o Afi H o AO H H o \8 inn n n A? nni ' II o A1 oo n.g./\i H H n A4 M.g.AH n.g./\j H r^ iir?t H.g.Gl ' H o PI II Tl 1 1 Ho T 2 100 M HoTI 1 H o Tl 100 r iH H.r.NCl Hr NPS Carolina mn H- MrM 77 H o F/^ F5 Florida Florida ' 3 I H.g.rO 100 _ H.g.l II rt F4 H tloF3 H o F** H iff Figure 4. Neighbor-joining phenogram of Hesionides riegerorum specimens (H. r.) from North Carolina (NO and H. gohari specimens (H. g.) from the local populations in Arcachon (A), Giglio. Mediterranean (G), Majorca, Mediterranean (M), Crete, Mediterranean (C), Hurghada, Red Sea (H). Florida (F). Phuket. Thailand (Ti. Only bootstrap values above 70 are indicated. lations in these two regions after the two seas were joined by the construction of the Suez Canal, or that one population was isolated from the other by migrating through the Canal after it was opened. The latter possibility is likely, but it is hard to decide whether the migration was from the Red Sea into the Mediterranean and on to the Atlantic (Lessepsian migration; For, 1978) or in the other direction. No finds from areas further east, between Hurghada and Thailand. POPULATION GENETICS OF A MEIOFAUNAL POLYCHAETE 223 0.1 Ha C"f. .g.OO H.g.Uo II n \fttl H.g.lVIO II n IVIfi H.g.IVlS II n f~"7 H.B.G / 96, H e G4 H.g.GS 77 II n f~*Q II H.g.Co - 93 in n rr H. /~"i .g.C / H.g.Oj II r. f~* 1 H.g.Lil II r, f~"y H.g.OZ Tin \A 1 -| '.'^'.ii H.g.M7 II n A 1 r" 1 B 'A4 H.2.A4 94 ii a r H.2.C4 ii 5 r*. H.g.CS 96 H n r 1 ! Europe and 30 r^ H.e.t,! ii S r 1 ? !!- g -IFA Ked Sea 8/1 ii i H.g.Al ii r. A /; H.g.AO II 5 AT H.g.A / II n A ft H.g.AO 1 IT n A O H.g.AV ii n A i n 1 II _ All H.g.Al 1 Hn UA .g.H4 H.g.H2 i ii in H.g.HJ - H.g.AS Hf, \/f> .g.iviz Hn Vi< Hn III 76 H.2.M3 Hi, n. .g.cj Hn Xl 93 .g. 1 J Hn T"> .g. IZ Thailand .g.Tl Ho T4 .g. 14 II t- M/" 1 ! n lit- M/^? 100" ii ,- isjri North "-"-' H.r.lNl.j 100 II i- Nf"M Hi- M*" 1 ^ _ Hn F5 H^ Hn IT/; .g.ro Hn VI I Hn VA Florida Hn IT1 -C .g.r 1 ii n r"> Figure 5. Single-linkage phenogram of Hesionides riegerorum specimens (H. r.) from North Carolina (NC) and H. gohari specimens (H. g.) from the local populations in Arcachon (A I. Giglio, Mediterranean (G). Majorca, Mediterranean (M). Crete, Mediterranean (C). Hurghada. Red Sea (H). Florida (F), Phuket. Thailand (T). Only bootstrap values above 70 are indicated. have yet been analyzed. Migration from the Red Sea would probably imply that in a very short time, about 1 00 years, the species extended its range through the entire Mediter- ranean Sea to the Atlantic. Immigrations of polychaetes can occur extremely rapidly, as has been demonstrated for spe- cies of Marenzellerici that became distributed in the North and Baltic Seas within a decade (Bastrop et al., 1997); however, these spionids have long-lived pelagic larvae 224 H. SCHMIDT AND W. WESTHEIDE (Bochert, 1997), whereas H. gohari is thought to undergo direct development (Westheide, 1970). Along-shore migra- tion is presumably a general feature of the latter species, but the rate of geographic expansion achieved by this means is unknown. H. gohari is considered to be one of the many cosmo- politan interstitial polychaete species (Westheide, 1971; Sterrer. 1973: Westheide and Rao. 1977; Riser. 1981 ; Giere. 1993; Soosten et at., 1998). As yet no morphological dis- tinctions have been found among individuals from the most diverse of the earth's marine regions (see Fig. 1 ), so that the species must temporarily be regarded as distributed world- wide on the coasts of warm seas (Westheide. 1977; Hart- mann-Schroder, 1991) provided that a morphological spe- cies concept is applied. In the present genetic study, all the cluster analyses employed show clear differences between the animals in the regions ( 1 ) European Atlantic/Mediter- ranean/Red Sea. (2) Indian Ocean, and (3) Western Atlantic (North America). Whether the observed genetic distances reflect species differences that is, reproductive barriers cannot be decided on the basis of current observations. In another cosmopolitan morphospecies, Petitia amphoph- rlmlina Slewing. 1956 (Syllidae). the distances found by RAPD analyses between North American and European animals correspond to the values found here: 0.60-0.66 (Soosten et ai, 1998). However, similar distance values (after Nei and Li, 1979) of geographically separated but morphologically differentiated pairs of interstitial poly- chaete species have been revealed by RAPD analyses as well: Nerilla antennata Schmidt, 1848-/V. mediterranea Schlieper, 1925 (Nerillidae): 0.74-0.77 (Schmidt and Westheide. 1998): Microphthalmus carolinensis West- heide and Rieger, 1987-A/. nahuntensis Westheide and Rieger, 1987: 0.77 (unpubl. data). Schirmacher et al. (1998) found a distance value of only 0.17 between two enchytraeid sibling species. Thus the notion of H. gohari representing a cosmopolitan species will have to be aban- doned. It is remarkable in this connection that the Florida pop- ulation, identified as H. gohari, in all analyses formed a sibling cluster with the freshwater species H. ricgeronmi Westheide, 1979. These animals are the nearest neighbors geographically to the latter species, with recorded finds on the East Coast of the United States in Florida (Westheide. 1995) and North Carolina; that is, they live only about 100 km away from the habitat of the freshwater species, which so far has been found only at a site in the sandy bank of the Chowan River in North Carolina. When H. riegerorum was first described, mention was made of its close morphologi- cal resemblance to H. gohari (Westheide. 1979). There is much evidence that the freshwater form separated from the neighboring marine Hesionides population on the seacoast. The genetic distance between H. riegerorum and the marine Hesionides individuals on the Florida coast is appreciable. although considerably less than that between the latter and the gohari individuals from Thailand or Europe. But whereas the freshwater species is also clearly distinguish- able morphologically (e.g.. by the different notopodial chae- tae. penis papillae, and anal lobes; Westheide, 1979), no such diagnostic characters have yet been identified for the populations classified as H. gohari. The small geographic distance between the new freshwater species and its marine population of origin has evidently produced morphological character shifts, some of which may act as isolating mech- anisms for instance, the differently positioned and shaped penis papillae on the prostomium (see also Westheide, 1984, fig. 8). In contrast, the large geographic distance separating the European/Egyptian, Thai, and North Ameri- can individuals from one another is correlated with genetic distinctions but not with any morphological ones that can readily be discerned, perhaps because of the stability of this marine habitat. To continue to regard all these populations as belonging to a single species. H. gohari, is problematic. They would then constitute something like a paraphylum (Lorenzen, 1976). with one population (Florida) more closely related to the distinct sibling species (H. riegerorum) than to the other populations. As soon as possible, therefore, at least the North American "//. gohari" should be given the status of a separate species. However, we prefer not to undertake its description until the animals on the coast of the United States have been further examined in a specific search for morphological diagnostic characters, so that the practical taxonomic work will not be hampered by a species analysis based exclusively on genetic tests and on the geographic situation. Regarding Sterrer's (1973) hypothesis, it is irrelevant whether these genetically derived clades represent distinct species or merely populations of a single, and hence indeed cosmopolitan, species. Testing this question, however, de- mands a larger number of specimens from different sites separated by oceans and sequencing techniques. Acknowledgments Special thanks for the provision of working facilities are due to Mr. Somsak Chullasorn and Dr. Anuwat Natee- wathana (Phuket Marine Biological Center), Dr. Mary Rice (Smithsonian Institution at Harbor Branch, Ft. Pierce. Flor- ida), Dr. Charles H. Petersen and Hal Summerson (Marine Science Institute. Morehead City. North Carolina), Dr. Claude Cazaux (Laboratoire Marine. Institut Universitaire de Biologic marine de Bordeaux. Arcachon), and Mrs. Mar- tina Ubel (Institut fur Marine Biologic. Giglio). Among the various people who helped in the collection of the speci- mens, we especially thank Cornelia von Soosten, Monika C. Miiller, and Alice Westheide. We gratefully acknowledge POPULATION GENETICS OF A MF.IOFAUNAL POLYCHAETE 225 preparation and typing of the manuscript hy Mrs. Anna Stein and Mrs. Andrea Noel. Literature Cited \t mimics. W. 1989. Meiot'aunal emergence from inlertidal sedimenl measured in the Held: significant contribution to nocturnal planktonic hioiiKiss m shallow waters. Helgol. Meeresunters. 43: 29-43. Baslrop, R., M. Riihner, C. Sturnihauer, and K. Jiirs.s. 1997. Where did Marenzelleria spp. (Polychaeta: Spionidae) in Europe come from? Ai/uat. Ecol. 31: I 14-136. Bochert, R. 1997. Maren:elleria riridis (Polychaeta: Spionidae): a re- Mew of its reproduction. Ai/uat. Ecol. 31: 163-175. Biischer, N., E. Zyprian, and R. Blaich. 1993. Identification of grapewme cultivars by DNA analyses: pitfalls of random amplified polymorphic DNA techniques using lOmer primers. Vitis 32: 187-188. Felsenstein, J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 7778-7783. Gerlach, S. A. 1977. Means of meiofauna dispersal. In The Mcinfmuw Species in Time and Spare. W. Sterrer and P. Ax. eds. Workshop Symposium. Bermuda. 1975. Mikrofauna Meeresboden 61: 89-103. Giere, O. 1993. \leiobenihology. The Microscopic Fauna in Aauaiu Sediments. Springer. Berlin. Hadrys, H., M. Balick, and B. Schierwater. 1992. Application of random amplified polymorphic DNA (RAPD) in molecular ecology. Mol. Ecol. 1: 55-63. Hagerman, G. M., and R. M. Rieger. 1981. Dispersal of benthic meiofauna by wave and current action in Bogue Sound, North Carolina. USA. Puhhl. Srn. Zoo/. Napoli (/. Mar. Ecol.) 2: 245-270. Hartmann-Schroder. G. 1960. Polychaeten aus dem Roten Meer Kiele r Mee resfurscli. 16: 69-125. Hartmann-Schriider, G. 1991. Die Polychaeten der subtropisch-tropi- schen bis tropischen Ostkiiste Australiens zwischen Maclean (New South Wales) und Gladstone (Queensland) sowie von Heron Island (GroBes Barrier Riff). In: G. Hartmann-Schroder and G. Hartmann. Zur Kenntnis des Eulitorals der australischen Kiisten unter besonderer Berticksichtigung der Polychaeten und Ostracoden. Teil 16. Min. Hambg. Zool M us. lust. 88: 17-71. He, Q.. M. K. Viljanen, and .1. Mertsola. 1994. Effects of thermocy- clers and primers on the reproducibility of banding patterns in random amplified polymorphic DNA analysis. Mol. Cell. Probes 8: 155-160. Kocher, T. D., W. K. Thomas, A. Meyer, S. V. Edwards, S. Paabo, F. X. Villablanca, and A. C. Wilson. 1989. Dynamics of mitochon- dria! DNA evolution in animals amplification and sequencing with conserved primers. Proc. Nail. Acini. Sci. USA 86: 6196-6200. Lorenzen, S. 1976. Zur Theorie der phylogenetischen Systematik. Verh. Dtsch. Zool. Ges. 1976: 22X Nei, M., and W. H. Li. 1979. Mathematical model for studying genetic- variation in terms of restriction endonucleases. Genetics 76: 5269- 5273. Palmer, M. A. 1988. Dispersal of marine meiofauna: a review and conceptual model explaining passive transport and active emergence with implications for recruitment. Mar. Ecol. Prog. Ser. 48: 81-91. Peer, Y. van de, and R. de Wachter. 1994. TREECON for Windows: a software package for the construction and drawing of evolutionary trees for the Microsoft Windows environment. Compiit. Appl. Biosci. 10: 569-570. For, F. D. 1978. Lessepsian Migration. The Influx of Red Sea Biota into the Mediterranean hy Way of the Sue:. Canal. Springer. Berlin 228 pp. Rao, G. C. 1972. On the geographical distribution of interstitial fauna of marine sand. Proc. Indian Natl. Sci. Acad. 38B: 164-178. Remane, A., and E. Schulz. 1964. Strand/onen des Roten Meeres und ihre Tierwelt. Kieler Meeresforsch. 20: 5-17. Riser, N. W. 1981. The aberrant polychaete Slygocniiitella from some American beaches. Wasnmnn J. Biol. 38: 10-17. Ruppert, K. E. 1977. Zoogeography and speciation in marine Gastro- tncha. in The Meiofauna Species in Time and Space. W. Sterrer and P. Ax. eds. Workshop Symposium, Bermuda. 1975. Mikrofauna Meeres- hodcn 61: 231-251. Saitou, N., and M. Nei. 1987. The neighbour joining method. A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4: 406- 425. Schierwater, B. 1995. Arbitrarily amplified DNA in systematic* and phylogenetics. Electrophoresis 16: 1643-1647. Schirmacher, A., H. Schmidt, and W. Westheide. 1998. RAPD-PCR investigations on sibling species of terrestrial Enchyirueiis (Annelida: Oligochaeta). Binchem. Syst. Ecol. 26: 27-36. Schmidt, H., and VV. Westheide. 1998. RAPD-PCR experiments con- firm the distinction between three morphologically similar species of Nerilla (Polychaeta: Nerillidae). Zool. An:. 26: 277-285. Sneath, P. H. A., and R. R. Sokal. 1973. Numerical Taxonomy. W. H. Freeman, San Francisco. Soosten, C. von, H. Schmidt, and W. Westheide. 1998. Genetic vari- ability and relationships among geographically widely separated pop- ulations of Pelilia amphophthalma (Polychaeta: Syllidae): Results from RAPD-PCR investigations. Mar. Biol. 131: 659-669. Sterrer, W. 1973. Plate tectonics as a mechanism for dispersal and speciation in interstitial sand fauna. Neth. J. Sea Res. 1: 200-222. Todaro, M. A., J. W. Fleeger, Y. P. Hu, A. W. Hrincevich, and D. W. Foltz. 1996. Are meiofaunal species cosmopolitan'.' Morphological and molecular analysis of Xenolrichula intermedia (Gastrotricha: Chaetonotida). Mar. Biol. 125: 735-742. Westheide, W. 1967. Monographic der Gattungen Hesionides Friedrich und Microphthalmus Mecznikow (Polychaeta). Ein Beitrag zu Organi- sation und Biologic psammobionter Polychaeten. Z. Morph. Tiere 61: 1-159. Westheide, W. 1970. Zur Organisation, Biologie und Okologie des interstitellen Polychaeten Hesionides gohari Hartmann-Schroder (He- sionidae). Mikrofauna Meeresboden 3: 1-37. Westheide, W. 1971. Interstitial Polychaeta (excluding Annelida). Smithson. Contrib. Zool. 76: 57-70. Westheide, W. 1972/73. Nouvelles recoltes d'anne'lides interstitielles dans les plages sableuses du Bassin d'Arcachon. Vie Milieu Serie A 23: 365-370. Westheide, W. 1972a. Raumliche und zeitliche Differenzierungen im Verteilungsmuster der marinen Interstitialfauna. Verh. Dtsch. Zool. Ges. 1971: 23-32. Westheide, W. 1972b. La faune des Polychetes et des Archiannelides dans les plages sableuse a ressac de la cote mediterraneenne de la Tunisie. Bull. lust. Natl. Sci. Tech. Oceanogr. Peclie. Salammbo 2: 449-46S. Westheide, W. 1977. The geographic distribution of interstitial polychaetes. In The Meiofauna Species in Time and Space. W. Sterrer and P. Ax. eds. Workshop Symposium, Bermuda. 1975. Mikrofauna Meeresboden 61: 287-302. Westheide, W. 1979. Hexionides riegcrorum n. sp.. a new interstitial freshwater polychaete form the United States. Int. Rev. Gesaintcn Hydrohiol. 64: 273-280. Westheide, W. 1984. The concept of reproduction in polychaetes with small body size: adaptations in interstitial species. In Polvchaete Re- production, A. Fischer and H.-D. Pfannenstiel, eds. Fortsch. Zool. 29:265-287. Westheide, W. 1990. Polychaetes: Interstitial Families. In Synopsis of the British Fauna {New series) No. 44. D. M. Kermack and R. S. K. Barnes, eds. Universal Book Services/W. Backhuys. Oegstgeest. The Netherlands. Westheide, W. 1991. The meiofauna of the Galapagos. A review. Pp. 226 H. SCHMIDT AND W. WESTHE1DE 37-73 in Galapagos Mamie invertebrates. Taxonomy, Biogeography, Hesionides (Polychaeta, Hesionidae from Indian sandy beaches. Cah. and Evolution in Dar\\-in's Islands, M. J. James, ed. Plenum, New Biol. Mar. 18: 275-287. York Westheide, W., and R. M. Rieger. 1987. Systematics of the amphiat- Westheide, W. 1992. Neue interstitielle Polychaeten (Hesionidae, Dor- lantic Microphtlialmus-Iisrensis-species-group (Polychaeta: Hesion- villeidae) aus dem Literal des Golfs von Bengalen. Microfauna Mar. 1: idae): facts and concepts for reconstruction of phylogeny and specia- 147-157. tion. Z. Zoo/. Syst. Evolulionsforsch. 25: 12-39. Westheide, W. 1995. Pisione hartmannschroederae sp.n. (Polychaeta: Williams, J. G. K., A. R. Kubelik, K. J. Livak, J. A. Rafalski, and Pisiomdae) from a Florida sand beach. Mitt. Hambg. Zoo/. Inst. 92: S. V. Tingey. 1990. DNA polymorphisms amplified by arbitrary 77_84. primers are useful as genetic markers. Nucleic Acids Res. 18: 6531- Westheide, W., and G. G. Rao. 1977. On some species of the genus 6535. THE BIOLOGICAL BULLETIN JUNE 1999 Editor Associate Editors Section Editor Online Editors Editorial Board Editorial Office MICHAEL J. GREENBERG Louis E. BURNETT CHARLES D. DERBY MICHAEL LABARBERA RUDOLF A. RAFF SHINYA INDUE, Imaging and Microscopy JAMES A. BLAKE, Keys to Marine Invertebrates of the Woods Hole Region WILLIAM D. COHEN, Marine Models Electronic Record and Compendia PETER B. ARMSTRONG R. ANDREW CAMERON ERNEST S. CHANG THOMAS H. DIETZ RICHARD B. EMLET DAVID EPEL GREGORY HINKLE MAKOTO KOBAYASHI DONAL T. MANAHAN MARGARET McFALL-NoAi MARK W. MILLER TATSUO MOTOKAWA YOSHITAKA NAGAHAMA SHERRY D. PAINTER J. HERBERT WAITE RICHARD K. ZIMMER-FAUST PAMELA CLAPP HINKLE VICTORIA R. GIBSON CAROL SCHACHINGER PATRICIA BURNS The Whitney Laboratory, University of Florida Grice Marine Biological Laboratory, College of Charleston Georgia State University University of Chicago Indiana University Marine Biological Laboratory ENSR Marine & Coastal Center. Woods Hole Hunter College, City University of New York University of California. Davis California Institute of Technology Bodega Marine Lab., University of California, Davis Louisiana State University Oregon Institute of Marine Biology, Univ. of Oregon Hopkins Marine Station, Stanford University Cereon Genomics, Cambridge, Massachusetts Hiroshima University of Economics, Japan University of Southern California Kewalo Marine Laboratory, University of Hawaii Institute of Neurobiology, University of Puerto Rico Tokyo Institute of Technology. Japan National Institute for Basic Biology, Japan Marine Biomed. Inst., Univ. of Texas Medical Branch University of California, Santa Barbara University of California, Los Angeles Managing Editor Staff Editor Editorial Assistant Subscription & Advertising Secretary Published by MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASSACHUSETTS Cover The illustration on the cover comprising a DNA molecule, the earth, and a distant ringed planet high- lights the publication, in this issue (p. 303), of the pro- ceedings of a workshop entitled Evolution: A Molecular Point of View. The workshop was held in Woods Hole, late in 1997, at the Marine Biological Laboratory; it dealt broadly with the origin and evolution of organic mole- cules; the evolution of genes, genomes and metabolic capacities; the phylogeny of early life forms (prokaryotes and ancestral eukaryotes); and the effect of changing environments on these processes. The workshop was sponsored and funded by NASA through its Center for Advanced Studies in the Space Life Sciences at the MBL. Indeed, this forward looking agency has, for more than four decades, been stimulating interest in, and supporting studies of, the origin and evolution of life as general phenomena in our solar system and beyond. Largely as a result of these efforts, such studies are now in the mainstream of science. Readers should know that the cover illustration is actually the logo of the Woods Hole Astrobiology Pro- gram, whose goals are similar to those of NASA and the workshop. This program, which includes projects at both the MBL and the Woods Hole Oceanographic Institution, is itself a node in a network of related programs consti- tuting a nationwide virtual Astrobiology Institute orga- nized recently by NASA. CONTENTS VOLUME 196, No. 3: JUNE 1999 Catherine Henlev (1922-1999) 227 EVOLUTION: A MOLECULAR POINT OF VIEW DEVELOPMENT AND REPRODUCTION Moran, A.L. Intracapsular feeding by embryos of the gastropod genus l.ittorina 229 Pennington, J. Timothy, Mario N. Tamburri, and James P. Barry Development, temperature tolerance, and settle- ment preference of embryos and larvae of the artic- ulate brachiopod Laqueiu californianus 245 PHYSIOLOGY Goffredi, Shana K., Peter R. Girguis, James J. Chil- dress, and Nicole T. Desaulniers Phvsiological functioning of carbonic anhydrase in the hydrotherma] vent tubeworm Riftia pachyptila. . . 257 ECOLOGY AND EVOLUTION Maas, Paula A. Y., Gregory D. O'Mullan, Richard A. Lutz, and Robert C. Vrijenhoek Genetic and morphometric characterization of mus- sels (Bivalvia: Mytilidae) from Mid-Atlantic hydro- thermal vents 265 Davidson, Seana K., and Margo G. Haygood Identification of sibling species of the bryozoan Bugiila neritina that produce different anticancer bryostatins and harbor distinct strains of the bacterial symbiont "(Mndidatus Endobugula sertula" 273 Palmer, A. Richard, Graeme M. Taylor, and Anise Barton Cuticle strength and the size-dependence of safety factors in Cancer crab claws 281 Steele, Craig, Carol Skinner, Catherine Steele, Philip Alberstadt, and Candice Mathewson Organization of chemically activated food search be- havior in Procambarus clarkii Girard and Orconectes rusticus Girard cravfishes . . .... 295 Sogin, Mitchell Introduction 307 Chang, Sherwood Planetary environments and the origin of life 308 Ferris, James P. Prebiotic synthesis on minerals: bridging the prebi- otic and RNA worlds 311 Ellington, Andrew D. Molecular origins and the null hvpothesis: motifs from our Maker? 315 Huang, Faqing, /hi li Yang, and Mike Yarus Self-capping RNA catalysts derived from selection- amplification 320 Bartel, David P. Creation and evolution of new ribozymes 322 Landweber, Laura F. The evolution of cellular computing 324 Weiner, Alan M., and Nancy Maizels The genomic tag hypothesis: modern viruses as mo- lecular fossils of ancient strategies for genomic rep- lication, and clues regarding the origin of protein synthesis 327 Maizels, Nancy, Alan M. Weiner, Dongxian Yue, and Pei-Yong Shi New evidence for the genomic tag hypothesis: ar- chael CCA-adding enzymes and tDNA substrates . . . 331 Ibba, M., A. W. Curnow, J. Bono, P. A. Rosa, C. R. Woese, and D. Soil Archaeal aminoacyl-tRNA synthesis: unique determi- nants of a universal genetic code? 335 Sharp, Paul M., Elizabeth Bailes, David L. Robertson, Feng Gao, and Beatrice H. Hahn Origins and evolution of AIDS viruses 338 Felsenstein, Joseph Coalescents, phylogenies, and likelihoods 343 Cummings, Michael P., Sarah P. Otto, and John Wakeley Genes and other samples of DNA sequence data for phvlogenetic inference 345 Schopf, J. William Deep divisions in the tree of life what does the fossil record reveal? 351 Doolitde, Russell F., Da-Fei Feng, and Glen Cho Determining divergence times with protein clocks. . 356 Gogarten, J. Peter, Ryan D. Murphey, and Lorraine Olendzenski Horizontal gene transfer: pitfalls and promises .... 359 DeLoiig, Edward F., Christa Schleper, Robert Feldman, and Ronald V. Swanson Application of genomics for understanding the evo- lution of hyperthermophilic and nonthermophilic Crenarchaeota 363 Reysenback, A.-L., S. Seitzinger, J. Kirshtein, and E. McLaughlin Molecular constraints on a high-temperature evolu- tion of early life 367 Forterre, Patrick, and Herve Philippe The Last Universal Common Ancestor (LUCA): sim- ple or complex? 373 Doolittle, W. Ford Rethinking the origin of eukaryotes 378 Patterson, David J., Alastair G. B. Simpson, and Nimalika Weerkaoon Free-living flagellates from anoxic habitats and the assembly of the eukaryotic cell 381 Clark, C. Graham The effect of secondary loss on our views of eukary- otic evolution 385 Hasegawa, Masami, and Tetsuo Hashimoto Phylogenetic position of amitochondriate protists in the evolution of eukaryotes 389 Cavalier-Smith T. Zooflagellate phylogeny and the systematics of pro- tozoa . 393 Douglas, Susan E. Evolutionary history of plastids 397 Gray, Michael W., Gertraud Burger, Robert Cedergren, G. Brian Golding, Claude Lemieux, David Sankoff, Monique Tunnel, and Franz Lang A genomics approach to mitochondrial evolution . . 400 Redfield, Rosemary J. The problem of the evolution of sex 404 Bode, Hans, Daniel Martinez, M. Andrew Shenk, Kerry Smith, Robert Steele, and Ulrich Technau Evolution of head development 408 Mohr, Scott, Jim Freeman, Tom Plasterer, and Temple Smith Patterns of mitochondrial DNA strand asymmetry correlate with phylogeny 411 Margulis, Lynn, Michael Dolan, and Ricardo Guerrero The molecular tangled bank: not seeing the phylog- enies for the trees 413 CONCLUDING REMARKS 415 LIST OF PARTICIPANTS . , 417 Index for Volume 196 421 CONTENTS for Volume 196 No. 1. FEBRUARY 1999 PHYSIOLOGY Dudgeon, Steve, Andreas Wagner, J. Rimas Vaisnys, and Leo W. Buss Dynamics of gastrovascular circulation in the hydro- zoan Poilucorynt carnea: the one-polyp case 1 Kelly, Robert H., and Paul H. Yancey High contents of trimethvlamine oxide correlating with depth in deep-sea teleost fishes, skates, and decapod crustaceans 18 Beaven, Amy E., and Kennedy T. Paynter Acidification of the phagosome in Crassostrea virginira hemocytes following engnlfment of zymosan 26 Smith, Andrew M., Tonya J. Quick, and Rachel L. St Peter Differences in the composition of adhesive and non- adhesive mucus from the limpet Lottia limatula .... 34 Kawaii, Satoru, Keiji Yamashita, Mitsuyo Nakai, Miyuki Takahashi, and Nobuhiro Fusetani Calcium-dependence of settlement and nematocyst discharge in actinulae of the hydroid Tubularia mesembryanlhetnum 45 RESEARCH NOTE Tapley, Da\id W.. Garry R. Buettner, and J. Malcolm Shick Free radicals and chemiluminescence as products of the spontaneous oxidation of sulfide in seawater, and their biological implications 52 DEVELOPMENT AND REPRODUCTION Froggett, Stephan J., and Esther M. Leise Metamorphosis in the marine snail Ilyanti.wi nb\nleln//n>u/\ PHYSIOLOGY 245 Goffredi, Shana K,, Peter R. Girguis, James J. Chil- dress, and Nicole T. Desaulniers Physiological functioning of carbonic anhydrase in the hydrothermal vent tubeworm Riftia pachtptila. . . 257 ECOLOGY AND EVOLUTION Maas, Paula A. Y., Gregory D. O'Mullan, Richard A. Lutz, and Robert C. Vrijenhoek Genetic and morphometric characterization of mus- sels (Bivalvia: Mytilidae) from Mid-Atlantic hydro- thermal vents 265 Davidson, Seana K., and Margo G. Haygood Identification of sibling species fo the bryozoan Bugula nmtina that produce different anticancer bryostatins and harbor distinct strains fo the bacterial symbiont "Candidatus Endobugula sertula" 273 Palmer, A. Richard, Graeme M. Taylor, and Anise Bar- ton Cuticle strength and the size-dependence of safety factors in Cancer crab claws 281 Steele, Craig, Carol Skinner, Catherine Steele, Philip Alberstadt, and Candice Mathewson Organization of chemically activated food search be- harior in I'mrninlxinn, clarkii Girard and Orconectes rustirus Girard crayfishes 295 EVOLUTION: A MOLECULAR POINT OF VIEW Sogin, Mitchell Introduction 307 Chang, Sherwood Planetary environments and the origin of life 308 Ferris, James P. Prebiotic synthesis on mineral: bridging the prebiotic and RNA worlds 311 Ellington, Andrew D. Molecular origins and the null hypothesis: motifs from our Maker? . . 315 CONTENTS: VOLUME 196 Huang, Faqing, Zhili Yang, and Mike Yarus Self-capping RNA catalysts derived from selection- amplification 320 Bartel, David P. Creation adn evolution of new ribozyrnes 322 Landweber, Laura F. The evolution of cellular computing 324 Weiner, Alan M., and Nancy Maizels The genomic tag hypothesis: modern viruses as mo- lecular fossils of ancient strategies for genomic rep- lication, and clues regarding the origin of protein synthesis 327 Maizels, Nancy, Alan M. Weiner, Dongxian Yue, and Pei-Yong Shi New evidence for the genomic tag hypothesis: ar- chael CCA-adding enzymes and tDNA substrates . . . 331 Ibba, M., A. W. Curnow, J. Bono, P. A. Rosa, C. R. Woese, and D. Soil Archaeal aminoacyl-tRNA synthesis: unique determi- nants of a universal genetic code? 335 Sharp, Paul M., Elizabeth Bailes, David L. Robertson, Feng Gao, and Beatrice H. Malm Origins and evolution of AIDS viruses 338 Felsenstein, Joseph Coalescents, phylogenies, and likelihoods 343 Cummings, Michael P., Sarah P. Otto, and John Wakeley Genes and other samples of DNA sequence data for phylogenetic inference 345 Schopf, J. William Deep divisions in the tree of life what does the fossil record reveal? 351 Doolitde, Russell F., Da-Fei Feng, and Glen Cho Determining divergence times with protein clocks . . 356 Gogarten, J. Peter, Ryan D. Murphey, and Lorraine Olendzenski Horizontal gene transfer: pitfalls and promises .... 359 DeLong, Edward F., Christa Schleper, Robert Feldman, and Ronald V. Swanson Application of genomics for understanding the evo- lution of hyperthemophilic and nonthermophilic Crenarchaeota 363 Reysenback, A.-L., S. Seitzinger, J. Kirshtein, and E. McLaughlin Molecular constraints on a high-temperature evolu- tion of early life 367 Forterre, Patrick, and Herve Philippe The Last Universal Common Ancestor (LUCA): sim- ple or complex? 373 Doolitde, W. Ford Rethinking the origin of eukaryotes 378 Patterson, David J., Alastair G. B. Simpson, and Ni- malika Weerkaoon Free-living flagellates from anoxic habitats and the assembly of the eukaryotic cell 381 Clark, C. Graham The effect of secondary loss on our views of eukary- otic evolution 385 Hasegawa, Masami, and Tetsuo Hashimoto Phylogenetic position of amitochondriate protists in the evolution of eukaryotes 389 Cavalier-Smith T. Zooflagellate phylogeny and the systematics of pro- tozoa 393 Douglas, Susan E. Evolutionary history of plastids 397 Gray, Michael W., Gertraud Burger, Robert Cedergren, G. Brian Golding, Claude Lemieux, David Sankoff, Monique Turmel, and Franz Lang A genomics approach to mitochondrial evolution . . 400 Redfield, Rosemary J. The problem of the evolution of sex 404 Bode, Hans, Daniel Martinez, M. Andrew Shenk, Kerry Smith, Robert Steele, and Ulrich Technau Evolution of head development 408 Mohr, Scott, Jim Freeman, Tom Plasterer, and Temple Smith Patterns of mitochondrial DNA strand asymmetry correlate with phylogeny 411 Margulis, Lynn, Michael Dolan, and Ricardo Guerrero The molecular tangled bank: not seeing the phylog- enies for the trees 413 CONCLUDING REMARKS 415 IjST OF PARTICIPANTS 417 Index for Volume 196 421 Reference: Binl. Bull. 196: 227. (June Catherine Henley (1922-1999) We note, with regret, the passing, this year, of Dr. Cathe- rine Henley, whose life and scientific career were closely linked to the Marine Biological Laboratory and, especially, to The Biological Bulletin. Dr. Henley was born in Quantico, Virginia, to Regina Knowles and John Ralph Henley, a colonel in the U.S. Marine Corps. Like all military families, the Henleys moved frequently, so during her early years, their daughter had the experience of living in many different places in the world. Henley received her AB degree in 1943 from the University of North Carolina at Chapel Hill, the MS from Johns Hop- kins in 1947 (with B. H. Willier). and returned to UNC for the Ph.D.. which was completed in 1949 with D. P. Costello. It was as an undergraduate in Chapel Hill that Henley met Prof. Donald P. Costello, the teacher and mentor who would provide the connection to the Marine Biological Laboratory. In the summer of 1944, she came to Woods Hole as a student in the Embryology course, in which Costello was an instructor. As it happened, this was only the first in a series of more than 30 MBL summers; she would return as a Lab Assistant in the Embryology course (1945-47). as an inde- pendent investigator (1950-1976). and finally, as a Trustee of the Laboratory (1971-79). As a Trustee, she is especially remembered for her outspoken espousal of the cause of the non-scientific staff of the Laboratory. From our perspective. Henley's most significant service at the MBL was to The Biological Bulletin. In 1950. when Costello assumed the editorship of this journal, Henley joined him as Assistant to the Editor. The two of them brought out the Bulletin together, migrating with it, each year, between Woods Hole and North Carolina. Their lab space on the first floor of the Lillie building bulged with files and manuscripts, but they enjoyed the view of Great Harbor and of the passing scene and goings on in sundial park. In 1968, they retired from their leadership of the journal, and in the annual report of 1968. the MBL acknowl- edged its "deep sense of gratitude" for their service to The Biological Bulletin. From 1969 to 1979, Henley continued to serve the Bulletin as a member of the Editorial Board. Henley was a member of the Zoology faculty for many years at UNC; she taught there and carried out her research, according to the season, at Chapel Hill or Woods Hole. Her fields were embryology and cytology, and she and Costello collaborated for more than 20 years on studies of gamete physiology, chromatin condensation, patterns of microtu- bules in invertebrate sperm, and related matters. Dozens of papers and abstracts about their work were published in The Biological Bulletin alone. Probably their most lasting con- tribution was a manual: Methods for Obtaining and Han- dling Marine Eggs and Embryos, which was first published by the MBL in 1957 and then revised (Costello and Henley, 1971). This manual is being republished online by Biolog- ical Bulletin Publications and provides most of the data for an online compendium about the breeding seasons and eggs of Woods Hole species (Cohen, 1999). Henley's life began to change significantly in 1975. She left the University of North Carolina to administer the peer review of grant proposals at the National Institutes of Health, the institution that had been the major supporter of her research. The following year, 1976, was her last as a summer researcher at the MBL, and she retired from the NIH in 1990. On February 19, 1999, at the age of 76, Catherine Henley, afflicted with cancer, died in her home at Carolina Mead- ows, Chapel Hill. Literature Cited Cohen, W. D. 1999. Egg characteristics and breeding season for Woods Hole species. In Biological Bulletin Compendia. [Online]. Biological Bulletin Publications. Available: http://www.mbl.edu/BiologicalBulle- tm/ [1999, June]. Costello, D. P., and C. Henley, 1971. Methods for Obtaining and Handling Marine Eggs and Embryos. 2nd ed. Marine Biological Lab- oratory, Woods Hole, MA. 227 Reference: Bio/. Bull 196: 224-244. (June 1999) Intracapsular Feeding by Embryos of the Gastropod Genus Littorina A. L. MORAN 1 Oregon Institute of Marine Biology. University of Oregon, Charleston, Oregon 97420 Abstract. Many gastropod species develop within egg capsules within which larvae are provided with extraemhry- onic nutrients. Species with encapsulated development fre- quently have transitory embryonic organs, such as "larval kidneys." that may represent specializations for consump- tion of intracapsular nutrition. Larvae of Littorina species with nonplanktonic, encapsulated development consume in- tracapsular albumen, but they lack obvious morphological modifications for albumen consumption. To determine the mechanism and location of protein uptake, larvae of seven species of Littorina (L. keenae, L. littorea, L. plena. L. saxatilis, L. scutitlata. L sitkana. L subrotundata) were exposed to solutions of either fluorescently labeled protein (FITC-bovine serum albumen) or ferritin. Under fluores- cence microscopy, larvae of all species with encapsulated, nonplanktonic development displayed strong regional affin- ity for FITC in the ciliated cells of the velum, whereas hatched larvae of planktotrophic Littorina species did not. Transmission electron microscopy of epithelial cells of non- planktotrophic veligers exposed to ferritin supported the interpretation that localized affinity for labeled protein in- dicated endocytotic protein uptake. Planktotrophic Littorina and Littorina with encapsulated, nonplanktonic develop- ment were shown to share equivalent velar width/larval length ratios during early embryonic development, whereas a literature search suggested that in other nonplanktotrophic prosobranchs the velum is relatively smaller than in plank- totrophs. Retention of a large velum in Littorina that de- velop entirely within egg capsules may facilitate feeding on intracapsular protein, in the absence of specialized assimi- lative organs found in other species with encapsulated de- velopment. Received 9 September 1997; accepted 12 February 1999. 1 Present address: Friday Harbor Laboratories, University of Washing- ton, 620 University Road, Friday Harbor. WA 98250. Email: moran fhl.washington.edu Introduction Marine invertebrates exhibit a remarkable variety of re- productive and developmental modes both within and among taxa, and this variation provides a powerful compar- ative means of studying the integration of development, life history, and evolution. One of the best-known dichotomies in invertebrate development is between species with larvae that must feed in the plankton to grow and attain metamor- phic competence (planktotrophic), and species that reach competence without feeding in the plankton (nonplanktotro- phic) (Thorson, 1946; Jablonski and Lutz, 1983; Strath- mann, 1985). Another important distinction can be made between species with entirely planktonic development, and species that spend all or part of development in egg capsules (planktonic. encapsulated nonplanktonic, and mixed devel- opment, respectively) (Thorson, 1946; Pechenik, 1979; Per- ron, 1981). Larvae of many species with encapsulated, nonplanktonic development are morphologically similar to larvae of re- lated planktotrophic species, and retain structures that pre- sumably had an ancestral role in larval swimming and feeding. For example, larvae of most gastropod molluscs with nonplanktonic, encapsulated development possess a velum (e.g., Fretter and Graham, 1962; Buckland-Nicks et ul.. 1973; Strathmann, 1978; Hadfield and laea. 1989), which is the primary larval structure that planktotrophic molluscan larvae use in swimming and food collection (Strathmann and Leise, 1979). The velum may be smaller in species that lack free-living larvae than in planktotrophic species (Jagersten, 1972; Webber, 1977; Rivest, 1983; but see Hadfield and laea, 1989), and it may have different patterns of ciliation (e.g., Lyons and Spight, 1973: Hadfield and laea, 1989). These changes in gastropod larval mor- phology have been interpreted as the loss of complex, ancestral planktotrophic features due to relaxation of stabi- lizing selection, and as functional modifications that en- 229 230 A. L. MORAN hance performance during development in the egg capsule (Fretter and Graham, 1962; Lyons and Spight, 1973; Had- fieldand laea. 1989). Egg capsules protect offspring (Shuto, 1974; Spight, 1977; Pechenik, 1984; Hawkins and Hutchinson, 1988; Rawlings, 1990, 1996) and retain progeny within suitable adult habitat (Wells and Wells, 1962; Chapman. 1965; Pechenik, 1979); in many species, capsules also provide a nutrient-rich environment for developing embryos (Fretter and Graham, 1962; Fioroni, 1977, 1988). Larvae of many gastropods with mixed or entirely encapsulated develop- ment have transitory structures that may represent special- izations for consumption of nutritive materials such as al- bumen or nurse eggs (Portmann. 1955; Portmann and Sandmeier, 1965; Gather and Tompa, 1972; Lyons and Spight, 1973; Rivest. 1983, 1992; Rivest and Strathmann, 1995). However, in many species the mechanisms of con- sumption of intracapsular nutrition are poorly known. Like- wise, the extent to which planktotrophic larval characters have been modified in species with nonplunktonic, encap- sulated larvae is not fully understood, and has not previ- ously been addressed by comparing closely related taxa with contrasting developmental modes. The gastropod genus Littorina, the periwinkle snails, contains ~ 19 species found in the high-shore zone through- out the northern Atlantic and Pacific oceans (Reid, 1989; Reid et al., 1996). The genus contains both planktotrophic and nonplanktotrophic species (see Reid, 1989, for review); molecular phylogenetic evidence supports planktotrophy as the ancestral state within the genus (see Rumbak et al., 1994). All species of Littorina undergo early development in egg capsules. Planktotrophic species have "mixed" de- velopment: early developmental stages are contained in complex pelagic capsules (Fig. 1A), within which each larva is encased in an individual egg envelope (Fig. 1A) from which it emerges just prior to hatching from the capsule as a swimming, planktotrophic veliger. Other spe- cies lack a planktonic stage altogether, metamorphosing from veligers to juveniles within benthic or brooded egg masses. Egg capsules of these nonplanktotrophic species are filled with granular albumen (Fig. IB) that larvae consume during development (Buckland-Nicks et al., 1973). Litto- rina veligers have been reported to lack albumen-absorbing larval kidneys (Rivest, 1981; 1992), and the site and mech- anism of albumen consumption in encapsulated embryos and larvae of nonplanktotrophic Littorina species have not been established. The objectives of this study were to ( 1 ) determine the location of nutrient assimilation by larvae of encapsulated, nonplanktonic species in the gastropod genus Littorina: (2) investigate the extent to which congeneric planktotrophs share similar patterns of assimilation; and (3) compare the larval functional feeding morphologies of planktotrophic B Figure 1. (A) Planktonic egg capsule of Littorina scutulata, a species with planktotrophic development, containing three prehatching veliger- stage larvae surrounded by individual egg envelopes. (B) Albumen-tilled capsule of Littorina saxatilis, a species with encapsulated, nonplanktonic development. Capsule was removed from the oviduct of a gravid female, al. albumen; c. capsule; 1, larva; en, egg envelope. Scale bar = 20(1 [j.m. Littorina and their congeners that develop entirely within egg capsules. Materials and Methods Terminology Because the words embryo and lan-a can have multiple interpretations in taxa with complex life histories, I use the terminology for early developmental stages as defined by McEdward and Janies (1993). Embryo refers to develop- mental stages from fertilized egg to gastrula; larva describes INTRACAPSULAR FEEDING BY LITTORINA 231 any premetamorphic stage with recognizable larval features (e.g., the prototroch), regardless of whether stages are free- living or not. Spawning and larval rearing Seven Littorina species were used in these experiments: four planktotrophs (L. littorea (Linnaeus. 1758), L. keenae Rosewater, 1978, L. plena Gould 1849, L. scutulata Gould, 1849) and three species with encapsulated, nonplanktonic development (L. saxatilis (Olivi, 1792), L. sitkana Philippi, 1846, L. subrotundata (Carpenter, 1864)) (Table I). Egg capsules of planktotrophic species and species with benthic egg masses (L. sitkana, L. subrotundata) were obtained by placing live, freshly collected adult animals into mesh- walled containers (<1 mm diameter mesh size), and im- mersing the containers in vigorously aerated seawater for 1 to 7 days. Egg capsules of the nonplanktotrophic species L. saxatilis. which broods larvae to metamorphosis in the oviduct, were obtained by cracking adult animals with nee- dle-nose pliers and removing capsules from the brood chamber. Larvae of planktotrophic species were reared at concen- trations of about one larva per milliliter in 0.45 jam filtered seawater changed every 4 days. Planktotrophic larvae were fed ad libitum on a mixture of the single-celled algae Isochrysis galbani (CCMP #1324 [T-ISO]) and Dunaliella tertiolecta (CCMP #1320 [DUN]). Egg masses of species with encapsulated, nonplanktonic development were main- tained in glass dishes of filtered (0.45 /u,m) seawater at 12C. Fluorescence microscopy To test for regions of affinity for labeled albumen protein, embryos and larvae of three species with encapsulated, nonplanktonic development (L. saxatilis, L. sitkana, L. sub- rotundata) were removed from their capsules at develop- mental stages from early cleavage to metamorphosis. Fine forceps were used to free embryos and larvae from capsules. Veliger larvae of four planktotrophic species (L. littorea, L planaxis, L. plena, L. scutulata) were examined both before and after they hatched from the egg capsule. Because the egg envelope that surrounded earlier developmental stages of planktotrophs could not be removed without damage to developing larvae, early larval stages of one planktotrophic species (L. plena) were exposed to test solutions while they were still in the egg envelope. Embryos and larvae were placed in solutions of bovine serum albumen labeled with fluoroscein isothiocyanate (FITC-BSA, Sigma #A-9771 ), a solution useful for demon- strating receptor-mediated endocytosis of proteins (Rivest, 1992; Rivest and Strathmann, 1995). FITC-BSA was also made in the laboratory from commercially available BSA and FITC (Sigma #F-7250) using the methods of Rivest (1981). To remove unconjugated FITC, FITC-BSA was dialyzed for 24 h against several changes of filtered (0.45 /am) seawater or treated with excess charcoal for 30 min (Rivest, 1981). Embryos and larvae were placed in test solutions of 10-1000 /u,g/ml FITC-BSA in filtered seawater at 12C for 15 min to 24 h, then rinsed in filtered seawater for periods ranging from 1 to 48 h. Controls were exposed to test solutions containing filtered seawater only, unlabeled BSA, or unconjugated FITC. Experimental and control embryos and larvae were examined with an Olympus epifluorescence microscope fitted with an FITC filter set (Omega Optics stock number XF23, excitation maximum 485 nm. emission 535 nm). Transmission electron microscopy To identify regions active in protein uptake, veligers of one species with encapsulated, nonplanktonic development Table I Species, development and collection information for Littorina utilized in thix study Species. Authority Mode 1 Collection locality Collection habitat Littorina littorea (Linnaeus, 1758) P Woods Hole. MA Rocky shoreline Mystic. CT Littorina keenae Rcsewater, 1978 P Monterey, CA Rocky shoreline Littorina plena Gould, 1849 Littorina scutulata Gould. 1849 P P Charleston, OR Charleston, OR Rocky shoreline Protected estuary Monterey, CA Rocky shoreline Littorina saxatilis (Olivi. 1792) NP Woods Hole, MA Rocky shoreline Mystic. CT Littorina sitkana Philippi, 1846 NP Charleston. OR Estuarine marsh Friday Harbor, WA Rocky shoreline Littorina subrotundata (Carpenter. 1864) NP Charleston, OR Estuarine marsh P = planktotrophic, NP = nonplanktotrophic. 1 D = 232 A. L. MORAN (L. sitkana) were removed from their capsules and prepared for transmission electron microscopy, using the methods of Rivest and Strathmann (1995) with minor modifications. Larvae were placed for 10 min in a solution of 0.05% osmium tetroxide and 3% glutaraldehyde in 0.1 M(pH 7.35) phosphate buffer, with the osmolarity raised to 990 mOsM with sucrose. Next, larvae were placed in a solution of 3% glutaraldehyde in 0.1 M phosphate buffer with the osmo- larity raised to 990 mOsM with sucrose for 1 h, after which an equal volume of 10% EDTA was added (to dissolve the shell) and larvae were fixed for another hour. Larvae were postfixed for 1 h at room temperature in 2% osmium te- troxide in 1.25% sodium bicarbonate, then dehydrated in an ethanol series. Finally, specimens were exchanged in pro- pylene oxide, embedded in epoxy resin, and thin sections were cut on a Reichert Ultracut E ultramicrotome. Sections were picked up on Butvar films on 200-/u,m hex grids, stained with uranyl acetate and lead citrate (after Reynolds, 1963), and examined with a Philips CM 12 electron micro- scope. To determine whether protein was assimilated into velar cells, some larvae were exposed to a solution of 1 mg/ml ferritin (Sigma catalog #F4503) in filtered seawater for 12 h and rinsed in seawater for 2 h prior to fixation. Ferritin is an electron-dense protein useful as a marker for endocytosis (Rivest, 1981). Five ferritin-exposed larvae and four con- trols were examined with transmission ele