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