Volume 201 THE Number 1 BIOLOGICAL BULLETIN AUGUST 2001 Published by the Marine Biological Laboratory http://www.biolbull.org Made to my exact Let's address my specs first, specifically lugh resolution, contrast and infinity-corrected optics. They've all reached Olympian standards thanks to Olympus. But even more to the point, here's how the BX2's modular design came through for me. First, the eight-position universal condenser offers the flexibility to choose from brightfield, darkfield and phase as well as DIC. Next, its assortment of prisms makes it possible to match the optical image shear to the specimen, achieving the optimal balance of contrast and resolution. Finally, the Plan APO objectives, with superb chromatic correction and contrast, provide extraordinary detail. Now let's move on. And yours Picture yourself sitting here, looking into your Olympus BX2 research microscope, your fluorescence requirements having been met. Specifically: The aspherical collector lens produces a fluorescence intensity that's twice as bright as others and more even across the field. The unique excitation balancers improve visualization of multiple labels by revealing details that would otherwise be unseen. The six-posi- tion filter turret makes single and multiband imaging faster and simpler. And the rectangular field stop, another Olympus exclusive, protects the specimen by exposing only the precise area being imaged in addi- tion to enhancing the S/N ratio. Time to see what's next. OLYMPUS FOCUS ON LIFE Visit us at www.olympusamerica.com or call 1-800-446-5967. specifications And yours. And yours. Here, imaging and automation is a must. And here, the BX2 responds as a high-performance, highly efficient, digital imaging machine. The motorized nosepiece, Z-drive, condenser, illuminator and filter wheels are fully integrated through the user-friendly software package. It's you who commands this automated imag- ing system with your PC, optional keypad or preset buttons located on the microscope frame itself. Digital images can now be acquired, processed and analyzed faster than before. And reports and documentation have never been this easy to generate. Which leaves one more set of specs. Now modularity really is in high gear as the Olympus FLUOVIEW 500 is added, resulting in a complete confo- cal laser scanning microscope system. It provides five imaging channels and has an intuitive operation that makes it readily available to everyone so that productivity is greatly enhanced. By the way, the BX2 is the only microscope that offers a Metal Matrix Composite frame the ultimate in static and thermal rigidity making it the optimal solution for frequent use of 3D microscopy, time-lapse observations and high-end digital imaging. So you see, with all this mod- ularity and flexibility, my BX2 microscope is also your BX2 microscope. Research Microscope Series 2001 Olympus Am Cover About 3.5 million years ago (Ma), rising sea levels opened the Bering Strait, and the North Atlantic Ocean was invaded by hundreds of taxa from the North Pacific. Among the invaders was the seastar genus Asterias. At present, two species of Asterias are recognized in the North Atlantic: A. forbesi on the west coast of the Atlantic, from Cape Cod south to Cape Hatteras, and A. rubens, a European species that ranges from southern France to Norway and Iceland, but also occurs in the northwestern Atlan- tic, mainly from Cape Cod north. Representatives of these species are shown on the cover, as is a specimen of A. amurensis, which inhabits the North Pacific from British Columbia to Japan. After entering the Atlantic, populations of Asterias were separated, and speciation subsequently occurred. The timing of the separation is critical, for it deter- mined, in part, the mechanism involved in the specia- tion, and it is the basis for the present geographic distribution of Asterias species in the North Atlantic. However, as the map on the cover illustrates, the timetable of these events was constrained by habitat and oceanographic instability during the Pleistocene glaciation. 1 In particular, most of the current North American habitat of Asterias rubens was repeatedly covered by a kilometer of ice and was unavailable to this seastar until about 15,000 years ago long after the opening of the Bering Strait. 1 The map on the cover is a polar view of the North Atlantic and Pacific Oceans during the Wisconsin glacial maximum, about 20.000 years ago. The solid blue line marks the average glacial margin; the dashed blue lines show the extent of sea ice in summer (upper) and winter (lower); the dotted black line illustrates how lower sea levels during glacial maxima altered the Atlantic coastline; and the shades of blue and green represent isotherms, highly compressed in the north- western Atlantic, and producing a strong temperature gradient. The speciation of Asterias in the Atlantic has been explained by two hypotheses. Either the event oc- curred recently, with strong natural selection pre- cluding hybridization; or the speciation into North American and European species occurred shortly after Asterias entered the North Atlantic, with a recolonization of the northwestern coast of the At- lantic by A. rubens taking place in recent times. The second hypothesis implies that speciation was due to prolonged isolation and was independent of ob- served adaptations to different water temperatures. As reported in this issue (p. 95), John P. Wares has collected genetic sequence data from populations of A. forbesi, A. rubens, and A. amurensis and used them in phylogenetic and population genetic analy- ses to test the two hypotheses. He concludes that, although changes in climate and ocean currents particularly the formation of the Labrador Cur- rent were concomitant with the separation of As- terias populations in the North Atlantic 3 Ma, permanent colonization of New England and the Canadian Maritimes by A. rubens occurred very recently. (Credits: map from B. Frenzel, M. Pecsi, and A.A. Velichko, eds., 1992, Atlas of Paleociimates and Paleoenvironments of the Northern Hemi- sphere, Geographical Research Institute, Hungarian Academy of Sciences, Budapest, p. 43; images of Asterias forbesi and A. rubens from the George M. Gray Museum collection, formerly administered by the Marine Biological Laboratory, now at the Peabody Museum of Natural History of Yale Uni- versity; image of A. amurensi from a photograph by Jan Haaga, provided online by the Alaska Fish- eries Science Center/National Marine Fisheries Ser- vice; cover design by Beth Liles, MBL.) THE BIOLOGICAL BULLETIN AUGUST 2001 Editor Associate Editors Section Editor Online Editors Editorial Board Editorial Office MICHAEL J. GREENBERG Louis E. BURNETT R. ANDREW CAMERON CHARLES D. DERBY MICHAEL LABARBERA The Whitney Laboratory, University of Florida Grice Marine Biological Laboratory, College of Charleston California Institute of Technology Georgia State University University of Chicago SHINYA INDUE, Imaging and Microscopy Marine Biological Laboratory JAMES A. BLAKE, Keys to Marine Invertebrates of the Woods Hole Region WILLIAM D. COHEN, Marine Models Electronic Record and Compendia PETER B. ARMSTRONG ERNEST S. CHANG THOMAS H. DIETZ RICHARD B. EMLET DAVID EPEL GREGORY HINKLE MAKOTO KOBAYASHI ESTHER M. LEISE DONAL T. MANAHAN MARGARET MCFALL-NGAI MARK W. MILLER TATSUO MOTOKAWA YOSHITAKA NAGAHAMA SHERRY D. PAINTER J. HERBERT WAFTE RICHARD K. ZIMMER ENSR Marine & Coastal Center, Woods Hole Hunter College, City University of New York University of California, Davis Bodega Marine Lab., University of California, Davis Louisiana State University Oregon Institute of Marine Biology, Univ. of Oregon Hopkins Marine Station, Stanford University Cereon Genomics, Cambridge, Massachusetts Hiroshima University of Economics, Japan University of North Carolina Greensboro 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 PAMELA CLAPP HINKLE VICTORIA R. GIBSON Managing Editor Staff Editor ~ <-, 'ceanoo _ ,. . , . CAROL SCHACHINGER Editorial Associate WENDY CHILD AUG 3 1 2001 Subscription & Advertising Secretary Published by MARINE BIOLOGICAL LABORATORY WOODS HOLE, MASSACHUSETTS http://www.biolbull.org Genomic Research Leaders Choose Microway Scalable Clusters E os Biotechnology, Marine Biological Laboratory, Millennium Pharmaceu- ticals, Mount Sinai Medical School, NIH, Pfizer, and Rockefeller University All Choose Microway Custom Clusters and Workstations for Reliability, Superior Technical Support and Great Pricing. 1.4 GHz Dual Athlon, 1.7 GHz Pentium 4, or 1 GHz Dual Pentium III in 1U or 2U Clusters Dual Alpha 833 MHz Clusters and Towers For maximum price/performance choose our Alpha 1U 833 MHz, 4 MB DDR Cache CS20, 4U UP2000+ or 4V 264DP RuggedRack Myrinet, Gigabit Ethernet or Dolphin Wulfkit High Speed Low Latency Interconnects RAID and Fibre Channel Storage Solutions Microway' 1 " Screamer Dual Alpha UP2000* 833 MHz. 4MB Cache in RuggedRack Chassis with RRR Redundant Power Supply Microway has earned an excellent reputation since 1982. If you need a quality product that is fine tuned and built to last, from a company that will be around to support you for years to come, Microway is The Number One Choice. .Microway has delivered high-performance computing prod- ucts since 1982, when our pioneering software made it pos- sible to use an 8087 in the IBM-PC. In 1987 we created the world's first PC parallel processing systems. Since then, our QuadPuter architecture has migrated from Transputers to i860s and finally to Alphas in 1995. Over the past three years, we engineered and delivered over 300 clusters that utilized MPI running on Linux. As a software developer and hard- ware manufacturer, we know the value of extensive testing and validation. We are experts at configuring and validating the low latency interconnects we employ in our clusters. Our technical support is legendary the systems we sell arrive at your site and WORK! Los Alamos chose Microway to maintain and upgrade its 144 node Alpha Avalon Cluster because of our reputation. Large clusters we have sold include 400+ nodes at the University of Wisconsin and 250+ nodes at Rockefeller University. Microway Technology you can count on m Microway offers three Athlon/ Pentium enclosures 1U, 2LI and tower, plus five Alpha configura- tions 1 LI, 3U, 4U RuggedRack, QuadPuter and full tower. Our 264DP includes two 21264's with up to 4 GB of memory in our cus- tom 4U RuggedRack, which features front accessible redundant power supplies and hard disks. This rugged configuration was chosen by the U.S. Navy for onboard use. We also offer a dual Alpha UP2000+ running at 833MHz with 2GB of memory. Our QuadPuter chassis holds 4 Alpha processors and up to 4GB memory. The 1U CS20 dual Alpha (at right) is the highest density computational platform available. Microway is API-Networks' Top North American ^B Channel Partner. "Most Powerful, Highest Density Computational Platform On the Planet" Microway Scalable 25 Node 50 Processor Cluster Using CS20 Dual 833 MHz Alphas and Myrinet Interconnect Yielding Peak Throughput of 82.5 Gigaflop: I have ordered numerous Alpha and Intel-based servers and work- stations from Microway running both Tru64 UNIX and Linux. We have been very happy with both the performance and great value of Microway's products. The major UNIX vendors don't come close to Microway in this regard, and we have also found that Microway provides better value than other Linux hardware vendors. I have also used Microway's tech support and was pleased with their response. We've been using their systems for over a year and have had only a couple of minor incidents which were dealt with promptly." David Kristofferson, Ph.D., MBA, Director of Information Systems, Eos Biotechnology, Inc. Find out why over 75% of Microway's sales come from repeat customers. Please call 508-746-7341 for a technical salesperson who speaks your language! Visit us at www.microway.com A Research Park Box 79, Kingston, MA 0236* 508-746-7341 info@microway.com AUG 3 1 2001 CONTENTS VOLUME 201. No. 1: AUGUST 2001 RESEARCH NOTE Seibel, Brad A., and David B. Carlini Metabolism of pelagic cephalopods as a function of habitat depth: a reanalysis using phylogenetically in- dependent contrasts NEUROBIOLOGY AND BEHAVIOR Herberholz, Jens, and Barbara Schmitz Signaling via water currents in behavioral interac- tions of snapping shrimp (Alpheus heterochaelis) .... PHYSIOLOGY AND BIOMECHANICS Reddy, P. Sreenivasula, and B. Kishori Methionine-enJkephalin induces hyperglycemia through evestalk homiones in the estuarine crab Stylla sermta . . . Mogami, Yoshihiro, Junko Ishii, and Shoji A. Baba Theoretical and experimental dissection of gravity- dependent mechanical orientation in gravi tactic micro- organisms 26 SYMBIOSIS AND PARASITOLOGY Hanten, Jeffrey J., and Sidney K. Pierce Synthesis of several light-harvesting complex I polypep- tides is blocked by cycloheximide in symbiotic chloro- plasts in the sea slug, Elysia chlorotica (Gould): A case for horizontal gene transfer between alga and animal?. . . McCurdy, Dean G. Asexual reproduction in Pygospio elegans Claparede (Annelida, Polychaeta) in relation to parasitism by Lepocreadium setiferoides (Miller and Northup) (Platy- helminthes, Trematoda) 17 34 DEVELOPMENT AND REPRODUCTION Stewart-Savage, J., Aimee Phillippi, and Philip O. Yund Delayed insemination results in embryo mortality in a brooding ascidian 52 CELL BIOLOGY Ballarin, Loriano, Antonella Franchini, Enzo Ottaviani, and Armando Sabbadin Momla cells as the major immunomodulatory hemo- cytes in ascidians: evidences from the colonial species Botnllm schlosseri 59 ECOLOGY AND EVOLUTION Halanych, Kenneth M.. Robert A. Feldman, and Robert C. Vrijenhoek Molecular evidence that Sclerolinum brattstromi is closely related to vestimentiferans, not to frenulate pogonophorans (Siboglinidae. Annelida) 65 Ponczek, Lawrence M., and Neil W. Blackstone Effect of cloning rate on fitness-related traits in two marine hydroids 76 Meidel, Susanne K., and Philip O. Yund Egg longevity and time-integrated fertilization in a tem- perate sea urchin (Stnmgylocenfrotus droebachiensis) .... 84 Wares, J. P. Biogeography of Astmas: North Atlantic climate change and speciation 95 SYSTEMATICS Gershwin, Lisa-ann Systematics and biogeography of the jellyfish Aurelia labiata (Cnidaria: Scyphozoa) 104 45 Annual Report of the Marine Biological Laboratory. ... Rl ANNOUNCEMENT The Marine Biological Laboratory is pleased to announce that it has entered into an agreement with HighWire Press of Stanford University to publish The Biological Bulletin electronically. The online journal will be launched on 23 August 2001. It will be available free of charge to subscribers and the general public for the next six months. Subsequently, subscribers to The Biological Bulletin will receive both the print and electronic versions of the journal. We invite you to visit The Biological Bulletin online at http://www.biolbull.org 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): $235 for libraries; $95 for individuals. Subscription per volume (three issues): $1 17.50 for libraries; $47.50 for individuals. Back and single issues (subject to availability): $40 for libraries; $20 for individuals. Communications relative to manuscripts should be sent to Michael J. Greenberg. Editor-in-Chief, or Pamela Clapp Hinkle, Managing Editor, at the Marine Biological Laboratory, 7 MBL Street, Woods Hole, Massachusetts 02543. Telephone: (508) 289-7428. FAX: 508-289-7922. E-mail: pclapp@mbl.edu. http://www.biolbull.org 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 Ceo 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 2001, by the Marine Biological Laboratory Periodicals postage paid at Woods Hole, MA, and additional mailing offices. ISSN 0006-3 185 INSTRUCTIONS TO AUTHORS The Biological Bulletin accepts outstanding original research reports of general interest to biologists throughout the world. Papers are usually of intermediate length (10-40 manuscript pages). A limited number of solicited review papers may be accepted after formal review. A paper will usually appear within four months after its acceptance. Very short, especially topical papers (less than 9 manuscript pages including tables, figures, and bibliography) will be published in a separate section entitled "Research Notes." A Research Note in The Biological Bulletin follows the format of similar notes in Nature. It should open with a summary paragraph of 150 to 200 words comprising the introduction and the conclusions. The rest of the text should continue on without subheadings, and there should be no more than 30 references. References should be referred to in the text by number, and listed in the Literature Cited section in the order that they appear in the text. Unlike references in Nature, references in the Research Notes section should conform in punctuation and arrangement to the style of recent issues of The Biological Bulletin. Materials and Methods should be incorpo- rated into appropriate figure legends. See the article by Loh- mann et 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. Digital art: The Biological Bulletin will accept figures sub- mitted in electronic form; however, digital art must conform to the following guidelines. Authors who create digital images are wholly responsible for the quality of their material, including color and halftone accuracy. Format. Acceptable graphic formats are TIFF and EPS. Color submissions must be in EPS format, saved in CMKY mode. Software. Preferred software is Adobe Illustrator or Adobe Photoshop for the Mac and Adobe Photoshop for Windows. Spe- cific instructions for artwork created with various software pro- grams are available on the Web at the Digital Art Information Site maintained by Cadmus Professional Communications at http:// cjs. cadmus.com/da/home. html Resolution. The minimum requirements for resolution are 1200 DPI for line art and 300 for halftones. Size. All digital artwork must be submitted at its actual printed size so that no scaling is necessary. Multipanel figures. Figures consisting of individual parts (e.g., panels A, B, C) must be assembled into final format and submitted as one file. Hard copy. Files must be accompanied by hard copy for use in case the electronic version is unusable. Disk identification. Disks must be clearly labeled with the following information: author name and manuscript num- ber; format (PC or Macintosh); name and version of software used. 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-publication 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, numbered with consecutive Roman numerals, and placed after the Literature Cited. Figure legends should contain enough information to make the figure intelligible separate from the text. Legends should be typed double spaced, with consec- utive Arabic numbers, on a separate sheet at the end of the paper. Footnotes should be limited to authors' current ad- dresses, acknowledgments or contribution numbers, and expla- nation 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 a/., 1980. Personal communications and ma- terial in preparation or in press should be cited in the text only, with author's initials and institutions, unless the material has been formally accepted and a volume number can be supplied. The list of references following the text should be headed 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 abbre- viations should normally follow those of the U. S. A. Standards Institute (USASI). as adopted by BIOLOGICAL ABSTRACTS and CHEMICAL ABSTRACTS, with the minor differences set out below. The most generally useful list of biological journal titles is that published each year by BIOLOGICAL ABSTRACTS (BIOSIS List of Serials; the most recent issue). Foreign authors, and others who are accustomed to using THE WORLD LIST OF SCIENTIFIC PERI- ODICALS, may find a booklet published by the Biological Coun- cil 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 additional abbreviations. The following points of reference style for THE BIOLOGICAL BULLETIN differ from USASI (or modified WORLD LIST) usage: A. Journal abbreviations, and book titles, all underlined (for italics) B. All components of abbreviations with initial capitals (not as European usage in WORLD LIST e.g., J. Cell. Comp. Physiol. NOT J. cell. comp. Physiol. ) C. All abbreviated components must be followed by a period, whole word components must not (i.e., J. Cancer Res.) D. Space between all components (e.g., J. Cell. Comp. Physiol.. not J. Cell. Comp. Physiol.) E. Unusual words in journal titles should be spelled out in full, rather than employing new abbreviations invented by the author. For example, use Rit Vfsindafjelags Is/endinga without abbreviation. F. All single woid journal titles in full (e.g., Veliger, Ecology, Brain). G. The order of abbreviated components should be the same as the word order of the complete title (i.e., Proc. and Trans, placed where they appear, not transposed as in some BIOLOGICAL ABSTRACTS listings). H. A few well-known international journals in their pre- ferred forms rather than WORLD LIST or USASI usage (e.g.. Nature, Science, Evolution NOT Nature, Loud.. Science, N.Y.; Evolution, Lancaster, Pa. ) 6. Reprints, page proofs, and charges. Authors may purchase reprints in lots of 100. Forms for placing reprint orders are sent with page proofs. 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 publication. They will be charged the current cost of printers' time for corrections to these (other than corrections of printers' or editors' errors). Other than these charges for au- thors' alterations, The Biological Bulletin does not have page charges. Reference: BiW. Bull. 201: I-?. (August 2001) Metabolism of Pelagic Cephalopods as a Function of Habitat Depth: A Reanalysis Using Phylogenetically Independent Contrasts BRAD A. SEIBEL 1 * AND DAVID B. CARLINI 2 1 Monterey Bay Aquarium Research Institute, 7700 Sandholdt Road, Moss Landing, California 95039; and 'Department of Biology, 101 Hurst Hall, American University, 4400 Massachusetts Avenue, NW, Washington, DC 20016-8007 Metabolic rates of deep-living animals have been in- tensely studied (1 ). Within pelagic fishes, crustaceans, and cephalopods, a strong decline in rates of mass-specific metabolism with depth has been observed. Childress and Mickel (2) put/onward the visual interactions hypothesis to explain this general pattern. Their hypothesis states that reduced metabolic rates among manv deep-sea pelagic tax- onomic groups result from relaxed selection for strong locomotory abilities for visual predator-prey interactions in the light-limited deep sea. This pattern has, however, been tested using mean metabolic rates for species as individual data points. Felsenstein (3) warned that, because species are descended in a hierarchical fashion from common an- cestors, they generally cannot be considered as independent data points in statistical analyses. Statistical methods have recently been developed that incorporate phylogenetic in- formation into comparative studies to create phvlogeneti- cally independent values that can then be used in statistical analyses. Reliable independent phylogenetic information has only recently become available for some deep-sea or- ganisms. The present contribution reanalyzed the metabolic rates (4, 5) of pelagic cephalopods as a function of, for consistency with previous studies, MDO (minimum depth of occurrence) using phylogenetic independent contrasts de- rived from a recent molecular phytogeny (6). This analysis confirms the existence of a significant negative relationship benveen metabolism and minimum habitat depth in pelagic cephalopods but suggests that phylogenetic history also has Received 29 August 2000; accepted 12 April 2001. * To whom correspondence should be addressed. E-mail: bseibel@ mbari.org considerable influence on the metabolic rates of individual species. Childress ( 1 ) argued against a phylogenetic basis for the observed relationships between metabolism and depth. He based the argument on the identification of convergence of metabolic rates at a given depth among distantly related taxa (fishes, crustaceans, cephalopods) as well as divergence within closely related groups as a function of depth. This pattern strongly suggests that species experience similar selective regimes at any given depth and that rates of metabolism are evolved in response to that selection. Seibel e t al. ( 5 ) further argued, on the basis of an analysis of higher nodes, that most of the variation in metabolic rates among cephalopods is between families within an order, as opposed to between genera within a family or species within a genus. Therefore, families are more appropriate units for compar- ison. A decline in metabolic rates with increasing habitat depth was also observed when families were used as inde- pendent data points (5). Nevertheless, the degrees of free- dom used for statistical analyses in these studies are ele- vated, to varying degrees, due to phylogenetic non- independence of the data. Felsenstein (3) proposed computing weighted differences ("contrasts") between the character values of pairs of sister species nodes, or both, as indicated by phylogenetic topol- ogy, thereby estimating an ancestral character value (e.g., the ancestral states of log-transformed depth and metabolic data presented in Fig. 1 ). Insofar as the ancestral nodes are correctly determined, each of these contrasts is independent of the others in terms of the evolutionary changes that have occurred to produce differences between the two members of a single contrast (7). Felsenstein's (3) method requires knowledge of the cladistic relationships between the species B. A. SEIBEL AND D. B. CARLINI A. 1.85, -0.50 | 1.79. -0.40F 1.98, -0.36. 2.03, -0.28 2.02, -0.26_T 2.12, -0.28 2.10, -0.08. 1.91,0.10 1.72,0.27 2.14, -0.25 2.15, -0.27 1.00,0.73 1.00,0.75 1.32,0.63 2.74, -0.99 2.30, 0.07 | 2.82, -0.81 1 2.89, -0.78 2.71, -0.83 .Cranchia 1.00, -0.43 .Liocranchia 2.70, -0.57 .Leachia 1.70, -0.25 .Helicocranchia 2.48, -0.23 .Histioteuthis 2.18,0.01 .Octopoteuthis 2.00, -0.21 .Joubiniteuthis 2.70, -0.39 .Gonatus 2.00,0.82 Jllex 1.00,0.95 .L. pealei 1.00,0.81 ,L. opalescens 1.00,0.68 .Sepioteuthis 1.00,0.71 .Chtenopleryx 1.70,0.37 .Bathyteulhis 2.90, -0.23 .Heteroteuthis 2.04,0.63 .J. diaphana 2.85, -0.82 .J. heathi 2.78, -0.80 .Eledonella 2.99, -0.74 .Amphitretus 2.48, -0.89 .Vampyroteuthis 2.78, -1.22 .Nautilus 2.18, -0.30 B. 1.60, -0.38 1.85, -0.50 r 1.91, -0.31 1.88,0.23 2.09, -0.23 2.15, -0.28 1.35, -0.26 r 2.18, -0.25r 1.74, -0.27 r 1.00, 0.72r 1.00, 0.75 r 2.00, 0.33 1.70,0.27 1.70,0.47 2.76, -0.17r 2.48. -0.1 5 r 2.00, 0.68 r 2.00. -0.06 r 2.18, -0.07 r 0.00, 0.86 2.92, -0.77 2.82, -0.81 r 1.23, -0.17 2.35, -0.78 Cranchia 1.00, -0.43 Liocrancha 2.70, -0.57 L. dislocata 1.00, -0.26 L. pacifica 1.70, -0.25 Galliteuthis 2.48, -0.27 Megalocranchia 1.00, -0.27 Helicocranchia 2.48, -0.23 L. pealei 1.00,0.81 L. opalescens 1.00,0.68 Sepioteuthis 1.00,0.71 Chtenopteryx 1.70,0.37 Bathyteuthis 2.90, -0.23 Heteroteuthis 2.04, 0.63 A.felis 1.70,0.33 A. pacificus 1.70,0.21 Enoploteuthis 1.70, 0.70 Pterygioteuthis 1.70, 0.43 C. calyx 2.48, -0.17 C. imperator2A8,-Q.\2 Valbyteuthis 2.95, -0.18 G. om.'* 2.00, 0.82 G.pwos 2.00,0.53 O. deletron 2.00, 0.09 O. nielseni 2.00, -0.21 H. heteropsis 2.18, -0.14 //. hoy lei 2.18,0.01 llex 1.00,0.95 Todarodes 1.00,0.76 Onychoteuthis o.oo, 0.76 oubiniteuthis 2.70, -0.39 Mastigoteuthis 2.57, -0.23 . diaphana 2.85, -0.82 /. heathi 2.78, -0.80 Eledonella 2.99, -0.74 Amphitretus 2.48, -0.89 Oc\thoe 1.00, 0.44 Octopus 1.00, 0.44 Vampyroteuthis 2.78, -1.22 Nautilus 2.\&, -0.30 Figure 1. Phylogenetic trees used for calculating independent contrasts on metabolic rate data. Log- transformed minimum depth of occurrence (MDO) and metabolic rates, in that order, are shown to the right of taxon names. Ancestral states of log-transformed MDO and metabolic rate data (i.e.. weighted differences or "contrasts." see text), calculated using the CAIC software application ( 18), are also shown at the internal nodes. (A) A 21-taxa tree for which both COI sequences and metabolic rate data are available. Branch lengths INDEPENDENT CONTRASTS FOR CEPHALOPOD METABOLISM being analyzed. Several studies have attempted to construct phylogenies for cephalopods. However, only a single reli- able family-level phylogeny exists that includes deep-water fauna. One previous phylogenetic analysis relied exclu- sively on morphological characters that are associated with buoyancy and locomotion and are thus confounded with metabolism and depth (8). We therefore felt that analysis was unsuitable for use in the present study. Other analyses have been unable to obtain sufficient resolution for familial relationships (9) or have included only shallow-living taxa (10, 11). Carlini and Graves (6) recently analyzed the higher level phylogenetic relationships of extant cephalopods by using a 657-bp sequence of the mitochondrial cytochrome c oxidase (COI) gene. The molecular sequence data from Carlini and Graves (6) provide an opportunity to test the visual interactions hypothesis directly, using a more valid statistical approach. An additional analysis based on actin gene sequences (12) was not included, primarily because there was very little overlap between taxa for which actin gene sequences were available and those for which meta- bolic data are available. Furthermore, the actin study pro- vides a more accurate reconstruction of gene family evolu- tion within the cephalopods than of specific relationships among taxa. The phylogenetic trees presented here from which the independent contrasts were calculated include only those species for which metabolic data are available. Similar trees were constructed including species for which enzymatic data are available. Although it may have been preferable to "prune" the complete COI tree rather than reconstruct trees using only taxa for which metabolic data are available, we decided to calculate new trees so that we could include taxa for which COI sequences were obtained after the publica- tion of the COI paper (6). The species we added were Amphitretus pelagicus, Helicocranchia pfefferi, and Jape- tella heathi. Pruning the tree would have had only a small effect on the values of the standardized contrasts and would not have significantly altered our conclusions. A second requirement of Felsenstein's (3) method is knowledge of branch lengths in units of expected variance of change. Ideally, branch lengths should represent expected units of evolutionary change (gradual model). For this ap- proach to be valid, independent contrasts must be ade- quately standardized so that they will receive equal weight- ing in subsequent regression analyses. We plotted the absolute value of each standardized independent contrast, generated from the fully resolved tree (Fig. la), versus its standard deviation (7) and found no relationship between the two variates (data not shown). Thus, the contrasts were adequately standardized and properly weighted in regres- sion analysis. However, even if a particular phylogenetic tree is well resolved and well supported, branch lengths are always estimates and are thus subject to error. A less optimal approach, but one that involves fewer assumptions about the evolutionary relationships of the taxa in question, is to assume that every branch in the phylogeny is the same length (punctuated model). The advantage of this approach is that it can be used for poorly resolved trees or for data sets where branch lengths cannot be estimated, such as those derived from both molecular (6) and morphological (13, 14) data. This allows more contrasts to be performed, increasing the power of subsequent statistical tests. On the other hand, the punctuated model is unrealistic for most data sets, as there is likely to be significant heterogeneity with respect to the evolutionary rates of the taxa under study. In any case, use of a punctuated model is far superior to any method that treats species values as independent data points. In the present study we employed both gradual and punc- tuated models in constructing trees for comparison. The gradual model tree is depicted in Figure la (21 taxa, met- abolic rates as a function of MDO). A similar tree was constructed including species for which enzymatic data are available (not shown, 18 taxa, enzymatic activities as a function of MDO). A tree constructed using the punctuated model for contrasts involving all taxa for which data are available is depicted in Figure Ib (39 taxa, metabolic rates versus MDO). A similar tree was constructed including species for which enzymatic data are available (not shown. 32 taxa, enzymatic activities versus MDO). Independent contrasts for log-transformed, normalized mean oxygen consumption rates (4, 15-18) were produced, for both gradual and punctuated models, using CAIC v. 2.0.0 (19), and were regressed against those produced for (molecular clock enforced) were calculated from the strict consensus of two most-parsimonious trees (Tree Length = 1432 steps; Consistency Index = 0.348: Retention Index = 0.334) derived from parsimony analysis of the COI data in PAUP* (28). (B) Partially resolved 39-taxa tree representing relationships between all pelagic taxa for which metabolic rate data are available. The conservative tree topology is based on a consensus of molecular and morphological evidence. In this case, branch lengths are unknown and a punctuated model of change was assumed; that is. all branches are of equal length. For example, the ancestral character state for log-transformed metabolic rate for the Cranchia-Liocranchia node, assuming a punctuated model of change, is calculated assuming a branch length equal to one and taking an average of the two species (0.43 + 0.57/2 = 0.50. corresponding to a calculated ancestral oxygen consumption rate of 0.61 /j,m O ; g 'h *). Determination of ancestral character states, assuming a gradual model of change, requires calculation of branch length using the CAIC software. B. A. SEIBEL AND D. B. CARLINI c o a, E 0.2-r 0-- u -0.4- 00 6 -- 64 orj 2 -0.8- -I- -t- -r- H- 0.2 0.4 0.6 0.8 1 Contrast: Log (Minimum Depth of Occurrence) Figure 2. Standardized contrasts of log-transformed oxygen consump- tion data plotted as a function of standardized contrasts of log-transformed minimum depth of occurrence calculated from the 39-taxon tree (Fig. Ib; punctuated model ). Contrasts for the three sister-species groupings within the cranchiid family (Cranchia-Liocranchia; Leachia dislocata-L. paci- fica; Galliteuthis-Megalocranchia; Fig. Ib) are indicated with open sym- bols and are included in the plotted regression. The slope of the regression is significant (P < 0.01 1. See Table 1 and text for equation and related statistics. MDO (Fig. 2, Table 1). We produced similar regressions for contrasts of activities of citrate synthase (CS) and octopine dehydrogena.se (ODH) (5, 20-22), indicators of aerobic and anaerobic metabolic potential, respectively (Table 1). We tested the validity of log transformation by using a method suggested by Purvis and Rambaut (19). authors of the CAIC package. Regressions of the absolute values of the contrasts on the estimated nodal values were performed, and none had slopes significantly different from zero. We also performed regressions of the absolute values of the contrasts against the standard deviations of the contrasts and detected no relationship in any case. These two tests ensure that we did not violate any of the assumptions of Felsenstein's (3) model of evolution of continuous characters as a random walk process. Relationships between contrasts of metabolism and depth are summarized in Table 1. A significant decline in oxygen consumption rate with habitat depth was observed when all taxa were included and a punctuated model was assumed (Fig. 2: v = -0.36.Y -- 0.02, P = 0.01). A similar relationship was observed using the gradual model ( v = -0.59.V - 0.049. P = 0.03). but only when the Cranchia versus Liocranchia contrast was excluded (see below). CS and ODH activities were weakly correlated with habitat depth when a gradual model was assumed, even with the Cranchia versus Liocranchia contrast excluded from anal- ysis (Table 1;CS, v = -l.Ol.v + 0.46, P = 0.06; ODH, v = -1.26.x - 0.22, P = 0.099). Contrasts performed using the punctuated model for the CS and ODH data indicated a significant negative relationship between enzy- matic activity and habitat depth with the Cranchia versus Liocranchia contrast excluded from analysis (Table 1: CS, v = -0.64.V + 0.08, P = 0.01; ODH, y = -1.02* - 0.04. P = 0.005). Although these results suggest a negative trend in metab- olism with increasing depth independent of phylogeny, there are clearly phylogenetic influences on the data. For example, members of the family Cranchiidae (including Cranchia and Liocranchia, the contrast excluded from sev- eral of the analyses) have low metabolic rates regardless of Metabolism of pelagic cephalopods as a function of habitat depth Table 1 Parameter Model All contrasts included MO, Punctuated 39 -0.36 -0.02 0.29 0.01 Gradual 21 n.s. CS Punctuated 32 n.s. Gradual 18 n.s. ODH Punctuated 32 n.s. Gradual 18 n.s. Cranchia vs. Liocranchia contrast excluded M0 2 Punctuated not performed Gradual 21 -0.59 -0.05 0.27 0.03 CS Punctuated 32 -0.64 0.08 0.32 0.01 Gradual 18 -1.01 0.46 0.26 0.06 ODH Punctuated 32 -1.02 0.04 0.40 0.005 Gradual 18 -1.26 -0.22 0.21 0.099 Log-transformed contrasts (y) of oxygen consumption rates (MO 2 = /j.mole O, g 'h ') and enzymatic activities (citrate synthase, CS. and octopine dehydrogenase. ODH, units g~'l of pelagic cephalopods were regressed against minimum depth of occurrence (.v), expressed as y = mA" + b. Number of taxa (n). regression coefficients (R 2 ) and P values are also presented. INDEPENDENT CONTRASTS FOR CEPHALOPOD METABOLISM habitat depth. The Cranchiidae is a very diverse family, and our data set is slightly biased toward cranchiid species (/; = 7 out of 39. MO 2 . punctuated model. Fig. Ib). Although many cranchiid species undergo ontogenetic vertical migra- tions in which successive developmental stages occupy progressively greater depths (12, 23), some species appear to remain near the surface until sexual maturity (24. 25). Seibel ct ul. (4) argued that the use of transparency (26) by the cranchiids reduces detection distances (27) at all depths and therefore allows them to employ sit-and-wait predation strategies, facilitating low metabolic rates, even in well-lit epipelagic waters. With the Cranchia-Liocranchia contrast removed, we consistently observed a much stronger rela- tionship between metabolism and habitat depth. Several sources of depth-related variation in metabolism, such as buoyancy and body mass, exist in addition to phylogeny. These have been discussed elsewhere (4. 5). Literature Cited i. Childress, J. J. 1995. Are there physiological and biochemical ad- aptations of metabolism in deep-sea animals? Tree 10: 30-36. 2. Childress, J. J., and T. J. Mickel. 1985. Metabolic rates of animals from the hydrothermal vents and other deep-sea habitats. Bio/. Soc. Wash. Bull. 6: 249-260. 3. Felsenstein, J. 1985. Phylogenies and the comparative method. Am. Nat. 126: 1-25. 4 Seibel. B. A., E. V. Thuesen, J. J. Childress. and L. A. Gorodezky. 1997. Decline in pelagic cephalopod metabolism with habitat depth reflects differences in locomotory efficiency. Biol. Bull. 192: 262-278. 5. Seibel, B. A., E. V. Thuesen, and J. J. Childress. 2000. Light- limitation on predator-prey interactions: consequences for metabolism and locomotion of deep-sea cephalopods. Biol. Bull. 198: 284-298. 6. Carlini. D. B., and J. E. Graves. 1999. Phylogenetic analysis of cytochrome c oxidase I sequences to determine higher-level relation- ships within the coleoid cephalopods. Bull. Mar. Sci. 64: 57-76. 7. Garland. T. J., P. H. Harvey, and A. R. Ives. 1992. Procedures for the analysis of comparative data using phylogenetically independent contrasts. Syst. Biol. 41: 18-32. 8. Clarke, M. R. 1988. Evolution of buoyancy and locomotion in recent cephalopods. Pp. 203-213 in The Mollusca. Vol. 12: Paleon- tology and Neontology of Cephalopods. M. R. Clarke and E. R. Trueman. eds. Academic Press. San Diego, CA. 9. Young, R. E., and M. Vecchione. 1996. Analysis of morphology to determine primary sister-taxon relationships within coleoid cephalo- pods. Am. Malacol. Bull. 12: 91-112. 10. Boucher-Rodoni, R., and L. Bonnaud. 1996. Biochemical and molecular approach to cephalopod phylogeny. Am. Malacol. Bull. 12: 79-85. 11. Bonnaud. L., R. Bouchvr-Rodoni, and M. Monnerot. 1997. Phy- logeny of cephalopods inferred from mitochondria] DNA sequences. Mol. Phylogenet. Evol. 1: 44-54. 12. Carlini, D. B., K. S. Reece, and J. E. Graves. 2000. Actin family gene evolution and the phylogeny of coleoid cephalopods (Mollusca: Cephalopoda). Mol. Biol. Evol. 17: 1353-1370. 13. Voss, N. 1988. Evolution of the cephalopod family Cranchiidae (Oegopsida). Pp. 293-314 in The Mollusca, Vol. 12: Paleontology and Neontology of Cephalopods. M. R. Clarke and E. R. Trueman, eds. Academic Press. San Diego. CA. 14. Young, R. E., and R. F. Harman. 1988. Phylogeny of the "Eno- ploteuthid" families. Smithson. Contrib. Zool. 586: 257-270. 15. Segawa, S. 1995. Effect of temperature on oxygen consumption of juvenile oval squid, Sepioteuthis lessoniana. Fish. Sci. 61: 743-746. 16. O'Dor, R. K., and M. J. Wells. 1987. Energy and Nutrient Flow in Cephalopod Life Cycles, P. R. Boyle, ed. Academic Press, London. Pp. 109-133. 17. DeMont, M. E., and R. K. O'Dor. 1981. The effects of activity, temperature and mass on the respiratory metabolism of the squid. I/lex il/ecebrosus. J. Mar. Biol. Assoc. UK 64: 535-543. 18. Wells, M. J., and J. Wells. 1985. Ventilation and oxygen uptake by Nautilus. J. E\p. Biol. 118: 297-312. 19. Purvis, A., and A. Rambaut. 1995. Comparative analysis by inde- pendent contrasts (CAIC): An Apple Macintosh application for ana- lyzing comparative data. Com/nit. Appl. Biosci. 11: 247-251. 20. Hochachka, P. W., K. B. Storey, and J. Baldwin. 1975. Squid muscle citrate synthase: Control of carbon entry into the Krebs cycle. Com/'. Biochem. Physiol. 52B: 193-199. 21. Baldwin, J. 1982. Correlations between enzyme profiles in cepha- lopod muscle and swimming behavior. Pac. Sci. 36: 349-356. 22. Ballantyne, J. S., P. W. Hochachka, and T. P. Mommsen. 1981. Studies on the metabolism of the migratory squid, Loligo opalescens: enzymes of tissues and heart mitochondria. Mar. Biol. Lett. 2: 75-85. 23. Young, R. E. 1975. Transitory eye shapes and the vertical distribu- tion of two midwater squids. Pac. Sci. 29: 243-255. 24. Young, R. E. 1975. Leachia pacifica (Cephalopoda, Teuthoidea): Spawning habitat and function of the brachial photophores. Pac. Sci. 29: 19-25. 25. Voss, N. A., and R. S. Voss. 1983. Phylogenetic relationships in the cephalopod family Cranchiidae (Oegopsida). Malacologia 23: 397- 426. 26. Seapy, R. R., and R. E. Young. 1986. Concealment in epipelagic pterotracheid heteropods (Gastropoda) and cranchiid squids (Cepha- lopoda). J. Zool. Lond. 210: 137-147. 27. Johnsen, S., and E. A. Widder. 1998. Transparency and visibility of gelatinous zooplankton from the Northwestern Atlantic and Gulf of Mexico. Biol. Bull. 195: 337-348. 28. Swofford, D. L. 1998. PAUP*: Phylogenetic Analysis Using Parsi- mon\ ( * and Other Methods}. Version 4.0. Sinauer. Sunderland, MA. Reference: Biol. Bull. 201: 6-16. (August 2001) Signaling via Water Currents in Behavioral Interactions of Snapping Shrimp (Alpheus heterochaelis) JENS HERBERHOLZ 1 * AND BARBARA SCHMITZ 2 1 Georgia State University, Department of Biology, P.O. Box 4010, Atlanta, Georgia 30302; and 2 Lehrstuhl fur Zoologie, TU Miinchen, Lichtenbergstr. 4. 85747 Garching, Germany Abstract. The snappping shrimp Alpheus heterochaelis produces a variety of different water currents during in- traspecific encounters and interspecific interactions with small sympatric crabs (Eurypanopeus depressus). We stud- ied the mechanisms of current production in tethered shrimp and the use of the different currents in freely behaving animals. The beating of the pleopods results in strong pos- teriorly directed currents. Although they reach rather far, these currents show no distinctions when directed toward different opponents. Gill currents are produced by move- ments of the scaphognathites (the exopodites of the second maxillae) and can then be deflected laterally by movements of the exopodites of the first and second maxillipeds. These frequent but slow lateral gill currents are most probably used to enhance chemical odor perception. The fast and focused, anteriorly directed gill currents, however, represent a powerful tool in intraspecific signaling, because they reach the chemo- and mechanosensory antennules of the opponent more often than any other currents and also because they are produced soon after previous contacts between the animals. They may carry chemical information about the social status of their producers since dominant shrimp release more anterior gill currents and more water jets than subordinate animals in intrasexual interactions. Introduction Alpheus heterochaelis of the family Alpheidae (Deca- poda, Caridea) is one of the largest snapping shrimp, reach- ing a body length of up to 55 mm. It shows a large, modified Received 27 November 2000; accepted 10 April 2001. * To whom correspondence should be addressed. E-mail: biojhh@ panther.gsu.edu snapper claw on one (left or right) side and a small pincer claw on the other side in both sexes (Williams, 1984). The snapper claw allows the animals to produce an extremely fast water jet (of up to 25 m/s; Versluis et al., 2000) by rapid claw closure after cocking the claw in the open position (Ritzmann. 1974). The high velocity of the water jet results in a pressure drop below vapor pressure that causes a cavitation bubble to grow to a size of about 3.5 mm in front of the snapper claw. The collapse of this bubble (and not as previously supposed the mechanical contact of both claw surfaces) causes the extremely loud (up to 215 dB re 1 ;u,Pa at 1 m distance; Schmitz, 2001) and short (about 500 ns) snapping sound (Versluis et al., 2000). The strong effect of the water jet and the cavitation bubble collapse can be seen during interspecific encounters. Small prey (e.g., worms, goby fish, or shrimp) can be stunned or even killed by the jet (MacGinitie, 1937; MacGinitie and MacGinitie, 1949; Mor- ris et al., 1980; Suzuki, 1986; Downer, 1989), and interspe- cific opponents (e.g., small sympatric crabs, Eurypanopeus depressus) can be injured at interaction distances of on average 3 mm (Schultz et al., 1998). Toward conspecifics the water jet was not observed to cause any damage but functions as a communicative signal (Herberholz and Schmitz, 1999), both opponents ensuring an interaction distance of on average 9 mm (Schmitz and Herberholz, 1998), which is far enough away from danger caused by implosion of the cavitation bubble. This hydrodynamic sig- nal is analyzed by the receiving shrimp predominantly with the help of mechanosensory hairs on the snapper claw, and may contain information about the strength, motivation, and sex of the snapper (Herberholz and Schmitz, 1998; Herber- holz, 1999). The still rather small interaction distance of less than 1 cm in agonistic encounters between two snapping shrimp WATER CURRENTS IN SNAPPING SHRIMP 7 also favors the exchange of chemical signals between the opponents. The literature on chemical orientation and com- munication in snapping shrimp is limited: Hazlett and Winn ( 1962) tested aggressive and defensive responses of Svnal- pheus lu'inphilli to crushed male or female extract, and Schein (1975) and Hughes ( 1996) investigated the choice of Alpheus heterochaelis toward extracts of male or female water in Y-maze experiments without clear-cut results. On the other hand, ablation of the chemosensitive antennules in Alpheus edwardsii strongly reduced pair formation and sex recognition, which may be due to impeded distant or contact chemoreception since the pairing frequency remained high when only the antennae were ablated (Jeng, 1994). The importance of olfactory signals during hierarchy formation was shown in male American lobsters (Karavan- ich and Atema. 1998a). In these experiments, the recogni- tion of urine-carried chemical signals, which were received by the antennules, allowed the subordinate animal to avoid the familiar dominant shrimp, and therefore reduced the duration and aggression of fights. The exchange of chemical signals is also assumed to play a major role in individual recognition and memory in male and female Homarus americamts (Karavanich and Atema, 1998b; Berkey and Atema, 1999). In lobsters, urine is released through a paired set of nephropores on the ventral sides of the basal segments of the second antennae (Parry, 1960). Agonistic behavior in lobsters causes an increase in the probability and volume of urine release (Breithaupt et al., 1999). The released urine is then carried by the powerful anteriorly directed gill currents and may therefore transfer chemical information from one animal to another (Atema, 1985). In recent studies (Zulandt Schneider et al., 1999; Zulandt Schneider and Moore. 2000), chemical cues were also described as an important source for recognition of the dominance status or stress condition of conspecifics in another crustacean, the red swamp crayfish (Procambarus clarkii). In light of these examples, a similar mechanism of chem- ical signal exchange via gill currents in snapping shrimp seems likely. We cannot, however, exclude the possibility that the animals also exchange hydrodynamic signals. In fact, it has been shown that the antennules of crayfish (Mellon, 1996) and lobsters (Guenther and Atema, 1998; Weaver and Atema, 1998) are equipped with both chemical and mechanosensory receptors, and detailed morphological studies of antennule sensory hairs favor the same situation in snapping shrimp (Schmitz, unpubl. obs.). Therefore, snapping shrimp may also perceive hydrodynamic stimuli as well as chemical stimuli with their antennules. Previous studies (Herberholz and Schmitz, 1998. 1999) have shown that the transfer of hydrodynamic signals is realized by the powerful water jet that is formed by rapid closure of the large claw. In contrast, the much weaker gill currents appear to be more suitable for transferring chemical information. Suspended plastic particles were successfully used to visualize and quantify biological flow fields in lobsters and crayfish in a series of experiments by Breithaupt and Ayers ( 1996, 1998). Small floating particles of the same density as seawater were added to the aquarium water and illuminated in a horizontal or vertical plane in the vicinity of a tethered animal. Flow fields were then analyzed by tracking individ- ual particles. It was shown that both lobsters and crayfish produce a great variety of flow fields by using the exopo- dites of the maxillipeds and by fanning the pleopods. The latter was also discussed with respect to chemical commu- nication: male American lobsters commonly fan their pleo- pods at the second entrance of their shelter, thus creating a strong current that may contain chemical information about the female positioned at the first entrance (Atema, 1985, 1988). The pleopod fanning frequencies in males correlate with the frequencies of females checking the shelter. The existence of pheromones that control female choice and molting as well as male aggression was therefore assumed (Cowan and Atema, 1990; Atema, 1995; Bushman and Atema. 1997). The possible exchange and use of different water currents during agonistic encounters has rarely been studied; but see Rohleder and Breithaupt (2000) for a preliminary study in the crayfish Astacus leptodactylus. To test the possibility that snapping shrimp use guided water currents as signals, we visualized and analyzed all water currents that the shrimp produced during their encounters with conspecifics of the same or different sex and in encounters with sympa- trically living mud flat crabs (Eurypanopeus depressus). Materials and Methods We analyzed the behavior of 12 adult specimens of Alpheus heterochaelis. a species of snapping shrimp (6 males, 6 females; body size: 3.9 0.4 cm. mean SD). Each animal was tested in an encounter with a conspecific of equal size of either the same or different sex, as well as in an encounter with a small crab (Eurypanopeus depressus; mean length and width of carapace: 1.6 0.2 X 1.2 0.2 cm, mean SD). All animals were caught in waters of the Gulf coast of Florida at the Florida State University Marine Laboratory near Panacea. Prior to the experiments the ani- mals were labeled with small numbers designated for mark- ing queen bees and were kept individually in perforated plastic containers ( 1 1 X 11 X 15 cm) containing gravel and oyster shells for shelter. The containers were placed within a large tank (90 X 195 X 33 cm) with 330 1 of circulating filtered seawater (salinity: 23%c^28%o; temperature: 22- 23C). Proteins were removed from the water, and pH. carbonate, oxygen, CO 2 . and NO 3 were regularly con- trolled. The shrimp were exposed to an illumination cycle of 12 h light/ 12 h dark and fed frozen shrimp, fish, or mussels three times a week. For visualization of the different water currents, we pre- J. HERBERHOLZ AND B. SCHMITZ pared the aquarium water (temperature: 22-24"C, water level: 5 cm) with small, floating plastic panicles (ABS- particles, Bayer, Leverkusen, diameter: 500-710 jum; spe- cific weight: 1.03 kg/1). The aquarium (30 X 24 X 24 cm; floor covered with black cloth to facilitate walking) was positioned on a platform isolated from vibrations (Breit- haupt et at., 1995). At the level of the interacting animals, the seawater was illuminated from one side by a slide projector holding a slide with a thin horizontal slit. Before each experiment fresh seawater and particles were added, and two animals (two snapping shrimp or one snapping shrimp and a crab) were placed in the aquarium for 10 min for acclimatization: the animals were separated by an opaque divider to prevent visual, tactile, and directed-chem- ical contact. After the partition was removed, all interac- tions between the animals during the following 20 min were videotaped from above (camera: Panasonic AG 455; video recorder; Panasonic AG 7355; monitor: Sony Trinitron). The reflexive characteristics of the suspended particles then allowed a precise tracking using standard video-frame anal- ysis. Each experiment (interactions between two snapping shrimp of the same or different sex or between a snapping shrimp and a crab) was characterized by the number of physical contacts between the opponents, regardless of their duration and strength, as well as by the number of water jets. Three different water currents were characterized, in- cluding a lateral gill current, an anterior gill current, and a pleopod current (Fig. la). The pleopod current was mea- sured only when the shrimp was not in locomotion, because this current is also likely to be used in supporting the animal's walking. Moreover, no current was included in our analysis unless the single-frame video analysis gave clear evidence that it had moved two or more plastic particles. The following parameters were evaluated for all visualized water currents: frequency, duration (time between onset of movement of the first floating particle and end of movement of the last particle), range (total distance covered by an identified particle due to a certain current: possibly under- estimated when the current hit an opponent or an aquarium wall), velocity and target of the currents, their potential to transfer chemical information (i.e.. entering the area of chemical perception at the receiver's side), the temporal correlation between currents and previous physical contacts, and the correlation between produced currents and water jets in winners and losers during intrasexual interactions. To determine a winner or loser, we counted the number of aggressive acts and the number of submissive acts after each physical contact between the conspecitic opponents throughout the encounter. Aggressive acts include behav- iors such as approach, aggressive stance, and grasping and opening of the claws. Submissive acts include moving back- wards and turning and tail flipping away from the opponent. These definitions are largely adopted from Nolan and Salmon (1970). In 11 out of 12 experiments, one animal produced more aggressive acts and fewer submissive ones than its opponent and was therefore determined to be the winner while the opponent was determined to be the loser. Statgraphics Plus 6.0 (Manugistics Group, Inc.) and SPSS 6.0.1. (SPSS Science Software GmbH) were used for statistics. Mean and standard deviation were calculated for each variable of interest for each tested individual, and only one value per individual (grand mean) is included in each statistical test. The behavior of the respective opponents (male and female snapping shrimp, and crabs) was not analyzed and is not included in our results (exception: data presented in Fig. 7). If not otherwise stated, the Friedman rank test for repeated measurements (sample size >2) or the Wilcoxon rank test (sample size = 2) were used, and values with P < 0.01 and P < 0.05 are indicated in the text. We used nonparametric statistical tests because most of the data did not fulfill the requirements for the use of parametric tests i.e., normality or equal variance. To gain more insight into the mechanism of gill current production and redirection, two snapping shrimp were teth- ered upside down in a small petri dish filled with seawater and floating plastic particles, and the activity of the different mouth parts, which produced or deflected the currents, was videotaped using a CCD camera (Sony XC-77CE) mounted on a binocular microscope with high magnification. In ad- dition, small drops of black ink (Brilliant Black 4001. Pelikan) were placed between the third and fourth walking legs of these shrimp as well as of animals tethered dorsal side up to a vertical holder and standing on a platform so that the gill currents could be visualized. (Fig. Ib). Results Visualization of water currents in tethered shrimp A unique feature of snapping shrimp is the production of an extremely rapid water jet by fast closure of a specialized snapper claw. Apart from this water jet. the snapping shrimp Alpheus heterochaelis is able to produce four kinds of water currents (Fig. 1), which can be subdivided into two main categories. Fanning of the pleopods causes a strong, poste- riorly directed pleopod current, and a gill current is pro- duced by rhythmically beating the scaphognathites as re- vealed by our visualization experiments in two tethered shrimp. Beating of the scaphognathites produces a depres- sion in the gill chamber; water is therefore sucked into this chamber and subsequently released anteriorly through two small openings in the carapace. This "normal" gill current can be visualized with ink in tethered animals, but it is too slow and weak to move floating particles and was therefore not analyzed during encounters of snapping shrimp and their opponents. It can, however, be accelerated and de- flected into a lateral gill current (see Fig. IB) by the exopodites of the second and third maxillipeds. The exopo- WATER CURRENTS IN SNAPPING SHRIMP 'normal" gill current pleopod current lateral gill current antennule anterior gill current Figure 1. (A) Schematized drawing (lateral view) of a snapping shrimp modified after Kim and Abele ( 1988) showing four different water currents (gray arrows): the "normal" gill current, the lateral gill current, the anterior gill current, and the pleopod current. Black arrows show the direction of water entering the gill chamber. (B) Frontal view of an A/pheiis helerochaelis snapping shrimp, tethered to a vertical holder by means of a plastic nut glued to the carapace and standing on a textile platform. Black ink was placed with a syringe between the third and fourth left pereiopods (see ink trace) to visualize the gill currents. The shrimp is fanning the exopodites of the right second and third maxillipeds. thus producing an ink-stained lateral gill current to the right. dites of the first maxilliped do not participate in this process. Fanning of the left exopodites results in acceleration and deflection of the released gill current to the left side, and fanning of the right exopodites results in deflection to the right side. Tethered snapping shrimp never beat the exopo- dites of both sides simultaneously, and this was also never observed during interactions in which the illuminated par- ticles were directed to only one side at a time. Interestingly, 10 J. HERBERHOLZ AND B. SCHMITZ D 1-gc a-gc homo hetero type of interaction inter Figure 2. Frequency of three different water currents (1-gc, lateral gill current, a-gc, anterior gill current, pc, pleopod current) produced by Al- pliens heterochaelis snapping shrimp in interactions with another shrimp of the same sex (homo), of different sex (hetero), and with a Eurypanopeus depressus crab (inter). Grand means and standard deviations for 12 snap- ping shrimp each are shown. Significant differences within interaction types with P < 0.01 are indicated by two asterisks (**). a (fast) anterior gill current was restricted to encounters of freely moving animals; it could not be elicited in tethered shrimp. Its production obviously requires physical, chemi- cal, or visual contact between the animals. As a result, we were not able to analyze the producing mechanism; that is, we did not identify the involved mouth parts. General characteristics of released water currents Encounters between two snapping shrimp of different sex (hetero) are characterized by a significantly higher number of physical contacts (23.9 8.3, /; = 287; P < 0.01) than seen in encounters between two shrimp of the same sex (homo; 13.8 6, n = 165), or between a snapping shrimp and a crab (Eurypanopeus depressus) (interspecific; 12.7 5.3. n = 157). On the other hand, snapping (water jet production) of the tested shrimp is significantly increased after a contact with a crab (38% 16<7r; P < 0.01) when compared to snapping after hetero and homo contacts (5% 4% and 11% 11%, respectively). These differences in mind, we first evaluated the number of water currents (lateral gill currents, anterior gill currents, and pleopod currents) in each experiment. Figure 2 shows that there are no essential differences between interaction types (homo, hetero, or interspecific). Within each interac- tion type, however, the number of lateral gill currents sig- nificantly (P < 0.01 ) exceeds that of anterior gill currents as well as that of pleopod currents. In addition, in interspecific encounters with a crab, the frequency of anterior gill cur- rents is significantly lower than the frequency of pleopod currents (P < 0.01). The duration of the different water currents (Fig. 3A) tends to be longest for lateral gill currents, with no signif- icant differences regarding the type of the opponent. The duration of anterior gill currents is generally shorter, with similar values in intraspecific interactions, yet almost twice as long as in interactions with a small crab. Anterior gill currents in interspecific encounters are significantly shorter in duration than lateral gill currents (P < 0.05). Pleopod currents, in contrast, reveal very consistent values for all types of interactions. Figure 3B shows the range of the different currents in all interaction types. Regardless of the opponent, the snapping shrimp tend to produce lateral gill currents with small ranges. Anterior gill currents generally cover larger dis- tances in intraspecific interactions, whereas the mean value is reduced in interactions with a crab. The most powerful current is the pleopod current, which covers long distances in all interaction types. Range differences within interaction types are significant at P < 0.05 and P < 0.01, respectively. The velocity of the water currents during the first 120 ms (6 video frames) was evaluated for 10 examples for each current and interaction type (Fig. 3C). There are no signif- icant differences in the velocities within and between dif- ferent types of interactions. The lateral gill current shows the slowest velocities in all encounters. The anterior gill current and the pleopod current show similar values and are both more powerful than the lateral gill current. Initial velocities are higher, but their analysis has not proved satisfactory because of the standard video time resolution of 20 ms (50 frame/s). Temporal relation of water currents to physical contact Figure 4 compares the frequency of water currents that were elicited within 10 s after a physical contact between the opponents with those that were "spontaneously" pro- duced that is, emitted more than 10 s after a preceding contact. As shown in Figure 4A. in all interaction types the lateral gill current is significantly more often produced spontaneously than following a physical contact (P < 0.01 ). In homo interactions it occurs in only 6.2% of all cases (n = 10 of 162) shortly after a contact. During hetero interactions this current is elicited by a contact in 11.5% of all cases (n = 2\ of 183); in interactions with a crab, the lateral gill currents occur within 10 s after a contact in only 8.5% of all cases (n = 13 of 153). The analysis of the anterior gill current reveals a com- pletely different frequency pattern, with more elicited cur- rents than spontaneous ones (Fig. 4B). In homo interactions the anterior gill current is produced in 65.5% of all cases (/; = 19 of 29) within 10 s after a preceding contact. Similarly, in hetero interactions this gill current is elicited by a contact in 62.5% of all cases (n = 15 of 24). Finally, during interactions with a crab, anterior gill currents are WATER CURRENTS IN SNAPPING SHRIMP II 1 u -a 25 20 15 10 B 1 U M I CJ 10s) homo hetero type of interaction inter 25 i/i 1 20 -i C ^ o 1 5 JO B B I o D 40 3 C 1-gc (ha) D 1-gc (ot) ** ** 1 homo hetero inter type of interaction a-gc (ha) D a-gc (ot) homo hetero inter type of interaction pc(<10s) D pc(>10s) urrents s DC C ** ** ** O 4-1 O 1 A 4 JO 1 9 c 2 J. hoi , - no hetero inter type of interaction 4. Mean number ol' lulcriil gill currents (A), anterior gill cur- rcnls (If), and |iloo|ioil cuiu-nls (C'l \\ithin 10 s (lilack tolunin-.) cu more than 10 s (white columns) alter a physical contact between the opponents in inlet, K lions ol Iwo siuppmg sin imp ol ilie same sex (homol, ot tlitTercnt se\ (heleio). anil ol a snapping shrimp ami a crab (inter), (iraiul means and standard deviations for 12 shrimp are shown. Significant differences with /' < 0.01 are indicated In luo asterisks ('*). V) *- o u- OJ I 3 C pc (ha) D pc (ot) homo hetero type of interaction inter ' 5. Mean number of lateral gill currents (A), anterior gill currents (B). and pleopod currents (C) hitting the antennules of the opponent (black columns, ha) or reaching other targets (white columns, ot) in interactions of two snapping shrimp ol the same se\ (homo), of different sex (hetero), and of a snapping shrimp and a crab (inter). Grand means and standard deviations for 12 shrimp are shown. A significant difference with P < 0.05 is indicated by tine asterisk (*) and with P < 0.01 by two asterisks (**). WATER CURRENTS IN SNAPPING SHRIMP 13 homo hetero number of a-gc to ^ O^ 00 C 1 1 / "/i 1 2 4 6 8 10 number of jets 2 4 6 8 10 number of jets 1 1 1 3246 number ot jets i i Figure 6. Correlation between the number of water jets and the num- ber of anterior gill currents produced in interactions of two snapping shrimp (A) of the same sex (homo; Spearman's coefficient of rank corre- lation r, = 0.9, P < 0.01). (B) of different sex (hetero). and (C) of a snapping shrimp and a crab (inter). Data of \2 shrimp each some data points overlap. that of lateral gill currents: the undirected currents signifi- cantly exceed the antennule-directed ones in each interac- tion type (P < 0.05 or 0.01, respectively; Fig. 5C). In homo interactions an average of only 1 1.5% (n = 6 of 52) of all pleopod currents are projected towards the chemoreceptive antennules. and during hetero interactions 16.7% (/; = 8 of 48) of all pleopod currents reach the antennule area. Finally. in interspecific interactions no pleopod current is aimed towards the antennules of the crab, but all (;i = 54) are directed elsewhere. Anterior gill currents and water jets In view of the prominent role of the anterior gill current with respect to its timing after a physical contact and the increased possibility of chemosensory information transfer, we tested the correlation between these gill currents and emitted water jets (Fig. 6). As mentioned before, in com- parison to intraspecific interactions, encounters with crabs are characterized by an increased number of water jets and a reduced number of anterior gill currents (Fig. 6C). In addition, more water jets are emitted in homo interactions between snapping shrimp (Fig. 6A) than in hetero encoun- ters (Fig. 6B). Thus, the number of anterior gill currents significantly increases with an increasing number of water jets only in interactions between two snapping shrimp of the same sex (Spearman rank correlation coefficient: >\ = 0.9, P < 0.01: Fig. 6A). This is not the case in interactions between two shrimp of different sex d\ = 0.5, P > 0.05), though a noticeable trend is shown and the overall low number of water jets may have prevented a significant result. An even lower degree of correlation is seen in interactions with a crab (r v = 0.4, P > 0. 1 ). As shown in Figure 7, winners of homo interactions (as defined by aggressive and submissive acts see Materials and Methods) not only produce a significantly higher mean number of water jets (N = II, P < 0.01) but also a significantly higher mean number of anterior gill currents than losers produce (N = 1 1; P < 0.01 ). Discussion Snapping shrimp (Alpheus hetemcluielis) produce two main water currents, a strong posteriorly directed pleopod current and an anteriorly directed gill current. We show that the "normal" anteriorly directed gill current can be modified and redirected into a lateral and a fast anterior gill current. The production of the latter is restricted to social interac- tions, in which it represents a powerful tool for chemical signaling. Moreover, the use of the fast anterior gill currents varies for the winners and losers of individual encounters. Mechanisms of gill current production Our experiments in tethered snapping shrimp show that water is sucked into the gill chamber due to a depression elicited by the beating scaphognathites (Fig. 1A). A "nor- mal" gill current is then released anteriorly with low veloc- ity through two small openings of the carapace. Once the left or right expodites of the second and third maxillipeds start fanning, the current is accelerated and deflected later- ally to that side (Fig. IB). As previously described in winner loser Figure 7. Frequency of water jets (jets, black columns) and anterior gill currents (a-gc, white columns I lor winners and losers in interactions of two snapping shrimp of the same sex. The significant differences between winners and losers with P ^ 0.01 are indicated by two asterisks (**). 14 J. HERBERHOLZ AND B. SCHMITZ lobsters (Homarus americanus), the exopodites of the first maxillipeds do not contribute to these lateral gill currents in snapping shrimp, whereas in crayfish (Procambarus clarkii) these appendages are also involved (Breithaupt. 1998). The production mechanism of the fast anterior gill current remains unclear, since this behavior obviously requires physical, chemical, or visual contact during intra- or inter- specific encounters of snapping shrimp, and thus was never seen in tethered animals. From our knowledge about the lateral gill current, we assume that the fast anterior gill current is created by high-frequency beating of the scapho- gnathites without contribution of the exopodites of the sec- ond and third maxillipeds. Since it is difficult to video- record the mouth parts with high magnification during social interactions, we are currently testing other methods of monitoring scaphognathite beating frequencies during en- counters to verify this hypothesis. Role of the fast anterior gill current during social interactions The analysis of the fast anterior gill current revealed the most surprising and interesting results. Although anterior gill currents were observed and well described in lobsters (Atema, 1985. 1995) and crayfish (Breithaupt, 1998), we found decisive differences in snapping shrimp. First of all, Alpheus heterochaelis produces different types of anterior gill currents. The "normal" anterior current is a slow, weak release of water, which was sucked through the gill cham- ber, as opposed to the fast, strong, anteriorly directed gill current, which occurs during social interactions. The pro- duction of the fast anterior gill current is rare (Fig. 2) but strongly linked to previous contacts with a conspecific or a crab (Fig. 4B). Among the observed currents, only the fast anterior current is created shortly after a preceding contact, regardless of the type of opponent. In fact, this current never occurred before the first contact. Moreover, we show that only this current is suited to transfer chemical information towards the other animal (Fig. 5B): it reaches the antennules of the opponent in nearly 50% of all cases. Of all analyzed currents, only the fast anterior gill current shows some peculiarities with respect to the shrimps' op- ponent. The number, duration, and range is smaller in en- counters with a crab than in interactions with conspecifics (Figs. 2, 3). We assume that the shrimp collect information about the genus of their opponent and reduce the effort to communicate accordingly, if it is a crab. Role of lateral gill currents during social interactions During social interactions between snapping shrimp and conspecifics of the same or different sex as well as during interactions with small crabs, the lateral gill currents are most prominent and significantly outnumber all other ob- served currents (i.e., pleopod currents and fast anterior gill currents; Fig. 2). Moreover, they are produced for long intervals but have a short range and a low velocity (Fig. 3). They are barely elicited by physical contact (Fig. 4A) and hardly ever reach the antennules of their opponents (Fig. 5A). These properties of the lateral gill currents do not change with different opponents but appear to result from a stereotyped form of production. Thus, obviously lateral gill currents are not predestinated to play a prominent role in active (chemical) signaling between the animals. Still, their function needs explanation. From our obser- vations we conclude that the lateral gill current is used to improve the shrimps' ability to sense possible odor signals that occur at close distance. By redirecting the "normal" gill current, the shrimp refreshes the area around its chemical receptors from its own smell (released by the slow and permanent gill current) and thereby improves the detection of the chemical surrounding. This idea is supported by our knowledge that Alpheus heterocliaelis naturally inhabits small, oyster-shell-covered areas with little water flow and that individuals of the species appear to be rather stationary within that area (Herberholz and Schmitz, pers. obs.). The lateral gill current produced by snapping shrimp seems to be used to remove water from the area around the antennules and to a much lesser extent to draw water toward that region as proposed for the posteriorly or laterally redirected gill currents of lobsters and crayfish (Atema, 1995; Breithaupt. 1998). In contrast to lobsters and crayfish, snapping shrimp were never observed to fan simultaneously with appendages on both sides. Instead, they beat the exopodites of one side at a time, and there are no obvious movements of particles from the opposite side toward the animal's anterior region. Role of pleopod currents during social interactions In lobsters (Homarus americanus), pleopod currents are used for chemical (possibly pheromonal) communication during courtship at a shelter (Atema. 1985. 1988. 1995; Cowan and Atema, 1990: Bushman and Atema, 1997). The snapping shrimp Alpheus heterocliaelis, in addition to using its pleopods for locomotion and to provide an oxygen sup- ply for attached eggs, uses them for shelter digging, fanning the substrate (sand or muddy-sand) backward behind it (Nolan and Salmon, 1970). These authors also mention (pleopod) fanning as an aggressive act, with a shrimp vig- orously beating its pleopods and directing a water current posteriorly quite close to another shrimp. The frequency of pleopod fanning is not noted by Nolan and Salmon (1970), but the behavior was described to occur between two fe- males at the entrance of a shelter. In our experiments, we did not provide a shelter, and all shrimp were in the middle of their molt cycle. In view of the finding that the actual impact of pleopod currents in lobsters depends to a high degree on the molt state of the animals as well as on their readiness to mate (Cowan and Atema, 1990), these condi- WATER CURRENTS IN SNAPPING SHRIMP 15 tions may have affected our results. Though pleopod cur- rents were rather often produced (Fig. 2) and (in comparison to gill currents) show an average duration, a large range, and high velocity (Fig. 3). there is a lack of correlation with previous contacts (Fig. 4C) and a low precision in hitting the antennules of the opponent (Fig. 5C). There are hardly any differences in the characteristics of these currents towards different opponents. All this indicates that pleopod currents are of little relevance for (chemical) signaling or commu- nication among snapping shrimp and between shrimp and sympatric crabs under our conditions. A specialized gill current for chemical .signaling and communication ? The transfer of chemical signals between interacting lob- sters (see e.g., Atema. 1995; Bushmann and Atema. 1997) and crayfish (Breithaupt et al., 1999) has been described in detail. In lobsters these signals can evoke long-term indi- vidual recognition (Karavanich and Atema, 1998a, b), and in crayfish they communicate dominance status or stress condition (Zulandt Schneider et al., 1999; Zulandt Schnei- der and Moore, 2000). In all cases, urine-borne signals were assumed to be the source of chemical signaling (Breithaupt et al., 1999; Breithaupt, pers. comm.). Since the urine is released through a paired set of nephropores on the ventral sides of the basal segments of the second antennae (Parry. 1960). it can be carried toward an opponent by the anterior gill current. Moreover, agonistic behavior in catheterized lobsters increases the probability and volume of urine re- lease (Breithaupt et al., 1999). In the present study we show for the first time that the pattern of water current production actually changes with respect to the social situation of an aquatic animal. Although snapping shrimp have the ability to produce "normal" an- terior gill currents, they create different, more powerful, anteriorly directed gill currents shortly after contacting their interaction partner. These elicited currents are then more likely to reach the opponents' area of chemical perception. The same may hold true for lobsters and crayfish, but their currents have not yet been quantified during social interac- tions. On the other hand, we still have to prove that the fast anterior gill current in snapping shrimp actually carries chemical signals toward the opponent. Although the data presented favor this assumption, we cannot exclude the possibility that hydrodynamic signals transferred by the gill currents participate in the communication between the ani- mals. Judging by their sensory equipment, snapping shrimp like crayfish (Mellon. 1996) and lobsters (Guen- ther and Atema, 1998; Weaver and Atema. 1998) are most likely to perceive hydrodynamic stimuli as well as chemical stimuli with their antennules (Schmitz, unpubl.). We plan to test this possibility by deactivating the chemical receptors only. In any case, the production of the fast anterior gill current may play a critical role during hierarchy formation in snap- ping shrimp. We show that in intrasexual encounters the numbers of water jets and anterior gill currents are posi- tively correlated (Fig. 6) and that both are significantly higher in the winner than in the loser (Fig. 7). In the present study, winner and loser met in only a single 20-min exper- iment. Preliminary experiments show that repetitive pairing of winners and losers reduces the number of water jets and anterior gill currents (Obermeier and Schmitz. unpubl.). This supports the finding that these behaviors are most probably correlated with dominance and social status in snapping shrimp. Although the strength of the water jet represents the strength of the animal (see Herberholz and Schmitz, 1999), the signal transferred by the gill current may then allow recognition of the sender. This, in turn, can prevent two Alpheus heterochaelis shrimp of the same sex from engaging in more severe fighting during subsequent encounters, thus reducing the number of the "costly" water jets. Acknowledgments We would like to thank Maren Laube for help in data analysis. The experiments comply with the current laws of Germany. Supported by a grant of the Deutsche For- schungsgemeinschaft (Schm 693/5-2). The work of J.H. was additionally supported by a NIH grant (NS26457) to Donald H. Edwards at Georgia State University. Literature Cited Atema, J. 1985. Chemoreception in the sea: adaptations of chemorecep- tors and behaviour to aquatic stimulus conditions. Soc. Exp. Biol. Svm/. Ser. 39: 387-423. Atema, J. 1988. Distribution of chemical stimuli. Pp. 29-56 in Sensory Bio/ogv of Aquatic Animals, J. Atema, R. R. Fay, A. N. Popper, and W. N. Tavolga, eds. Springer Verlag, New York. Atema, J. 1995. Chemical signals in the marine environment: dispersal, detection and temporal signal analysis. Proc. Nail. Acad. Sci. USA 92: 62-66. Berkey, C., and J. Atema. 1999. Individual recognition and memory in Homarus americanus male-female interactions. Biol. Bull. 197: 253- 254. Breithaupt, T. 1998. Flow generation by specialized appendages in lobsters and crayfish. Pp. 185-186 in BIONA-Repori 13. R. Blickhan. A. Wisser, and W. Nachtigall. eds. Gustav Fischer Verlag. Stuttgart. Breithaupt. T.. and J. Avers. 1996. Visualization and quantitative anal- ysis of biological flow fields using suspended particles. Pp. 1 17-129 in Zooplankton: Sensory Ecology and Physiology, P. H. Lenz, D. K. Hartline. J. E. Purcell. and D. L. Macmillan, eds. Gordon Breach Publishers. Amsterdam. Breithaupt, T., and J. Avers. 1998. Visualization and quantification of biological flow fields through video-based digital motion-analysis tech- niques. Mar. Fi~esln\'. Bt'lniv. Phvsiol. 31: 5561. Breithaupt, T., B. Schmitz, and J. Tautz. 1995. Hydrodynamic orien- tation of crayfish (Procambarus clarkii) to swimming fish prey. J. Comp. Physio/. A 177: 481-491. Breithaupt, T., D. P. Undstrom, and J. Atema. 1999. Urine release in 16 J. HERBERHOLZ AND B. SCHMITZ freely moving catheterised lobsters (Homarus ainericanus) with refer- ence to feeding and social activities. / Exp. Biol. 202: 837-844. Bushman, P. J., and J. Atema. 1997. Shelter sharing and chemical courtship signals in the lobster Hinnarus americwws. Can. J. Fish. Aqnat. Set. 54: 647-654. Cowan, D. F., and J. Atema. 1990. Moult staggering and serial monog- amy in American lobsters, Homarus wnericaniis. Aniin. Behav. 39: 1199-1206. Downer, J. 1989. Mir den Augen der Tiere. Teil: Laute tier Natitr. [video recording). Universum. BBC and WDR (Westdeutscher Rundfunk). Guenther, C. M., and J. Atema. 1998. Distribution of setae on the Homarus ainericanus lateral antennular rlagellum. Biol. Bull. 195: 182-183. Hazlett. B. A., and H. E. Winn. 1962. Sound production and associated behavior of Bermuda crustaceans (Paniiliriis, Gonodactylus, Alpheus. and Syna/phetis}. Crustaceana 4: 25-38. Herherholz, J. 1999. Die Bedeutung hydrodynamischer Signale in intra- und interspezifischen Interaktionen von Pistolenkrebsen (Alpheus het- erochaelis). Ph.D. thesis. Technische Universitat Miinchen, Germany. Herberholz, J., and B. Schmitz. 1998. Role of mechanosensory stimuli in intraspecific agonistic encounters of the snapping shrimp (Alpheus hetemchae/is). Biol. Bull. 195: 156-167. Herberholz, J., and B. Schmitz. 1999. Flow visualization and high speed video analysis of water jets in the snapping shrimp (Alpheus helemchaelis). J. Co/up. Phvsiol. A 185: 41-49. Hughes, M. 1996. The function of concurrent signals: visual and chem- ical communication in snapping shrimp. Anim. Behav. 52: 247-257. Jeng, M.-S. 1994. Effect of antennular and antennal ablation on pairing behavior of snapping shrimp Alpheus edwardsii (Audouin). J. Exp. Mar. Biol. Ecol. 179: 17I-17S. Karavanich, C., and J. Atema. 1998a. Olfactory recognition of urine signals in dominance fights between male lobsters, Homarus america- nits. Behaviour 135: 719-730. Karavanich, C., and J. Atema. 1998b. Individual recognition and mem- ory in lobster dominance. Anim. Behav. 56: 1553-1560. Kim, W., and L. Abele. 1988. The Snapping Shrimp Genus Alpheus from the Eastern Pacific (Decapoda: Caridea: A/pheidae). Smithso- nian Institution Press. Washington, DC. 1 19 pp. MacGinitie, G. E. 1937. Notes on the natural history of several marine Crustacea. Am. Midi Nat. 18: 1031-1036. MacGinitie, G. E., and N. MacGinitie. 1949. Natural History of Ma- rine Animals. McGraw Hill. New York. 473 pp. Mellon, DeK. 19%. Behavioral responses of crayfish antennules to odor- ant and hydrodynamic stimuli. Soc. Neurosci. Abstr. 22: 1406. Morris, R. H., D. P. Abbott, and E. C. Haderlie. 1980. Intenidal Invertebrates of California. Stanford University Press, Stanford. 690 PP Nolan, A. N., and M. Salmon. 1970. The behaviour and ecology of snapping shrimp (Crustacea: Alpheus heterochelis and Alpheus nor- manni). Forma Function 2: 289-335. Parry, G. 1960. Excretion. Pp. 341-366 in The Physiology of Crustacea, Vol I: Metabolism and Growth. T. H. Waterman, ed. Academic Press, New York. Ritzmann, R. E. 1974. Mechanisms for the snapping behavior of two alpheid shrimp. Alpheus californiensis and Alpheus heterochelis. J. Comp. Physio/. 95: 217-236. Rohleder, P., and T. Breithaupt. 2000. Visualizing chemical commu- nication in crayfish. Zoology: Analysis of Complex Systems Suppl. Ill: 103: 72. Schein, H. 1975. Aspects of the aggressive and sexual behaviour of Alpheus helerochaelis Say. Mar. Behav. Phvsiol. 3: 83-96. Schmitz, B. 2001. Sound production in Crustacea with special reference to the Alpheidae. Pp. 521-533 in The Crustacean Nervous System, K. Wiese. ed. Springer- Verlag, Berlin. Schmitz, B., and J. Herberholz. 1998. Snapping behaviour in intraspe- ciric agonistic encounters in the snapping shrimp (Alpheus heterochae- lis). J. Biosci. 23: 623-632. Schultz, S., K. \\uppermann, and B. Schmitz. 1998. Behavioural interactions of snapping shrimp (Alpheus heterochaelix} with conspe- cifics and sympatric crabs (Eurypanopeus depressus). Zoology: Anal- ysis of Complex Systems Suppl. I: 101: 85. Suzuki, D. 1986. Stunning sounds. Series: The Nature of Things, [video recording). CBC (Canadian Broadcasting Corporation). Versluis, M., B. Schmitz, A. von der Heydt, and D. Lohse. 2000. How snapping shrimp snap: through cavitating bubbles. Science 289: 2114- 2117. Weaver, M., and J. Atema. 1998. Hydrodynamic coupling of lobster antennule motion to oscillatory water flow. Biol. Bull. 195: 180-182. Williams, A. B. 1984. Shrimps, Lobsters, and Crabs of the Atlantic Coast of the Eastern United States. Maine to Florida. Smithsonian Institution Press, Washington. DC. / nl. inch Schneider, R. A., and P. A. Moore. 2000. Urine as a source of conspecific disturbance signals in the crayfish Procambarus clarkii. J. E\p. Biol. 203: 765-771. Zulandt Schneider, R. A., R. W. S. Schneider, and P. A. Moore. 1999. Recognition of dominance status by chemoreception in the red swamp crayfish, Procambarus clarkii. J. Chem. Ecol. 25(4): 781-794. Reference: Biol. Bull. 201: 17-25. (August 201)1) Methionine-Enkephalin Induces Hyperglycemia Through Eyestalk Hormones in the Estuarine Crab Scylla serrata P. SREENIVASULA REDDY* AND B. KISHORI Department of Biotechnology, Sri Venkateswara University, TIRUPAT! - 517 502, India Abstract. The hypothesis is tested that methionine-en- kephalin. a hormone produced in and released from eyestalk of crustaceans, produces hyperglycemia indirectly by stim- ulating the release of hyperglycemic hormone from the eyestalks. Injection of methionine-enkephalin leads to hy- perglycemia and hyperglucosemia in the estuarine crab Scylla serrata in a dose-dependent manner. Decreases in total carbohydrate (TCHO) and glycogen levels of hepato- pancreas and muscle with an increase in phosphorylase activity were also observed in intact crabs after methionine- enkephalin injection. Eyestalk ablation depressed hemo- lymph glucose (19<7r) and TCHO levels (22%), with an elevation of levels of TCHO and glycogen of hepatopan- creas and muscle. Tissue phosphorylase activity decreased significantly during bilateral eyestalk ablation. Administra- tion of methionine-enkephalin into eyestalkless crabs caused no significant alterations in these parameters when compared to eyestalk ablated crabs. These results support the hypothesis that methionine-enkephalin produces hyper- glycemia in crustaceans by triggering release of hypergly- cemic hormone from the eyestalks. Introduction In decapod crustaceans, hemolymph sugar level is regu- lated by hyperglycemic hormone. Abramowitz et al. ( 1944) were the first to demonstrate that injection of eyestalk extract induced hyperglycemia in Callinectes. Since then, hyperglycemia as a response to injection of eyestalk extract Received 14 July 2000; accepted 6 March 2001. * To whom correspondence should be addressed. E-mail: psreddy@vsnl.com has been observed in almost all groups of crustaceans (see review by Keller, 1992). This neurohormone is stored in and released from the sinus gland. The chemical nature, mode, and site of action of hyperglycemic hormone has been extensively studied in a number of crustaceans (see reviews by Keller et al., 1985; Sedlmeier, 1985). The amino acid sequence of hyperglycemic hormones has been determined from a large number of crustaceans (see La Combe et al., 1999, for review). The gene for hyperglycemic hormone was also cloned from crabs (Kegel et al., 1989), lobster (Tensen et al., 1991), prawn (Ohira et al., 1997). isopod (Martin et al., 1993). and crayfish (Kegel et al., 1991; Huberman et al., 1993; Yasuda et al., 1994). Recently, we reported the expression of hyperglycemic hormone gene at different molt stages in Homarus americanus, the American lobster (Reddy et al.. 1997). Since the discovery of opioid peptides in decapod crus- taceans by Mancillas et al. (1981). several workers have attempted to determine the physiological function of these peptides, but the results are fragmentary. Sarojini et al. (1995. 1996. 1997) provided evidence that methionine- enkephalin slowed ovarian maturation in the fiddler crab Uca pugilator and the crayfish Procanibarus clarkii, and suggested that methionine-enkephalin produces this effect indirectly by stimulating the release of gonad-inhibiting hormone from eyestalks. In Uca pugilator, methionine- enkephalin appears to stimulate release of the concentrating hormones for black and red pigment cells (Quackenbush and Fingerman. 1984) and the dark-adapting hormone for distal retinal pigment cells (Kulkarni and Fingerman. 1987). We reported a neurotransmitter role for methionine-en- kephalin in regulating the hemolymph sugar level of the freshwater crab O-ioielphusa senex senex, and hypothesized that methionine-enkephalin produces hyperglycemia indi- 17 18 P. S. REDDY AND B. KISHORI rectly by stimulating release of hyperglycemic hormone (Reddy, 1999). The objectives of the present study were threefold: (a) by extending our studies to the estuarine crab Scylla serrata, to test our hypothesis, generated by the study of Oziotelphusa senex senex, that methionine-enkephalin produces hyper- glycemia in decapod crustaceans; (b) to determine the changes in levels of tissue carbohydrates and phosphorylase activity during methionine-enkephalin treatment; and (c) to provide evidence that supports the triggering of release of hyperglycemic hormone during methionine-enkephalin treatment. Materials and Methods Individuals of Scylla serrata (15 2 cm in carapace width; 110 5 g wet weight) were collected from the Chennai coast, India. They were kept in large aquaria with continuous aeration and acclimatized to laboratory condi- tions for one week under constant salinity (25 1 ppt), pH (7.2 0.1 ), and temperature (23 2C). During this period the crabs were fed fish flesh. Feeding was stopped 24 h before the beginning of the experiments, and no food was given during experimentation. Only intermolt (Stage C 4 ), intact, male crabs were used in the present study. Methionine-enkephalin (Sigma Chemical Co.) was dis- solved in physiological saline (Pantin, 1934). In these ex- periments, each of the 10 groups of crabs used consisted of 10 individuals. The first group served as normal and re- ceived no treatment. A second group served as control, with each crab in this group receiving an injection of 10 /il of physiological saline (Pantin, 1934) through the base of the coxa of the 3rd pair of the walking legs. In groups 3-5 respectively, each crab received an injection of 10~ 7 , 10~ s , and 10~ y mole methionine-enkephalin in 10 jal volume. Both eyestalks were ablated from all the crabs in groups 6-10. The eyestalks were extirpated by cutting them off at the base, without prior ligation but with cautery of the wound after operation. Twenty-four hours after eyestalk ablation, these groups were used for experimentation. Crabs in group 6 served simply as eyestalkless animals, and crabs in group 7 received 10 ju,l crustacean Ringer solution and served as eyestalkless controls. In groups 8-10 respectively, each crab was injected with 10~ 7 , 10~ x , and 10~ 9 mole methionine-enkephalin in 10 /xl volume. Based on prelim- inary kinetic studies, the crabs were sacrificed for analysis 2 h after injection (Figs. 1. 2). Hemolymph (500 jul) was aspirated by syringe, through the arthrodial membrane of the coxa of the 4th pair of walking legs. The other tissues (hepatopancreas and muscle from chela propodus) were then quickly dissected out. weighed, and analyzed by the procedures outlined below. Hemolvmph total carbohydrate level. Hemolymph total carbohydrate (TCHO) levels were estimated in trichloroace- tic acid supernatant (10% TCA w/v) according to the method of Carroll et al. ( 1956). Hemolymph glucose level. For measurement of glucose, 100 /u,l of hemolymph was mixed with 300 ju.1 of 95% ethanol. After deproteinization (4 C, 14,000 X g, 10 min), the sample was combined with a mixture of glucose enzyme reagent (glucose-6-phosphate dehydrogenase and NADP) and color reagents (phenazine methosulfate and iodo- nitrotetrazolium chloride) (kit from Sigma). After 30 min, the intensity of the color was measured at 490 nm and quantified with standards. Tissue TCHO and glycogen levels. TCHO levels in the tissues (hepatopancreas and muscle) were estimated in the 10% TCA supernatant (5% w/v), and glycogen was esti- mated in the ethanolic precipitate of TCA supernatant, ac- cording to the method of Carroll et al. ( 1956). To 0.5 ml of clear supernatant was added 5.0 ml of anthrone reagent, and the combination was boiled for 10 min in a water bath. The samples were then immediately cooled. A standard sample containing a known quantity of glucose solution was always tested along with the experi- mental samples. Absorbance was measured at 620 nm against a reagent blank. Tissue phosphorylase activity. Phosphorylase activity was assayed in hepatopancreas and muscle by colorimetric determination of inorganic phosphate released from glu- cose- 1 -phosphate by the method of Cori et al. (1955). First, 0.4 ml of the enzyme was incubated with 2.0 mg of glyco- gen for 20 min at 35 C, then the reaction was initiated by the addition of 0.2 ml of 0.016 M glucose- 1 -phosphate (G-l-P) to one tube (phosphorylase a) and a mixture of 0.2 ml of G-l-P and 0.004 M adenosine-5-monophosphate (phosphorylase ah) to another tube. The reaction was incubated for 15 min for determining total phosphorylase and for 30 min for active phosphor- ylase. The reaction was terminated by the addition of 5.0 ml of 5 N sulfuric acid. Released inorganic phosphate was estimated by the method of Taussky and Shorr (1953). Protein determination. Total protein levels in the enzyme source were estimated following the method of Lowry et al. (1951) using bovine serum albumin as standard. MET-ENKEPHALIN-INDUCED HYPERGLYCEMIA IN CRAB 19 Table 1 Effect of eyestalk ablation fESX) (24-h post-ablation) and injection of methionine-enkephalin into intact and ablated crabs on hemolymph total sugar aiui glucose levels in Scylla serrata No treatment Ringer injection 10~ 7 mol/crab 10 " mol/crab 10~ 9 mol/crab Dunnet's comparison test Total Sugar (mg/100 ml) Intact (Group 1 ) ESX (Group 2) 12.11 1.01 9.41 1.13" (-22.22) 12.73 1.84 a (5.12) 9.34 1.03 b ' c (-0.74) 28.8 2.18 h (126.23) 9.41 1.13 b ' c (0.74) 19.64 I.41 h (54.28) 9.43 1.01 b - c (0.96) 16.52 1.94 h (29.77) 9.21 1.08 b ' c (-1.39) F (4-45) = 137.160 F (4 . 45 , = 0.099 Intact (Group 1 ) ESX (Group 2) Two-way ANOVA: F, w (Between groups) = 772.002, P < 0.001; F 4 9n (Among treatments) = 98.747, P < 0.001; F 490 (Interaction) = 94.552, P < 0.001. Glucose (mg/100 ml) 6.55 0.76 a 12.07 1.34" 11.44 1.28" 9.13 0.78" F (4 45l = 75.613 (2.16) (84.27) (74.65) (39.38) 5.52 0.81 Kc 5.19 1.01 b - c 5.21 0.91 bc 5.44 0.77 Kc F, 445 , = 0.387 (6.35) (-5.97) (-5.61) (-1.44) 5.19 1.01 h (-19.03) Two-way ANOVA: F l 90 (Between groups) = 440.810. P < 0.001; F 4 90 (Among treatments) = 40.092, P < 0.001: F.,,,,, (Interaction) = 44.753, P < 0.001. Values are mean SD of 10 individual crabs. Values in parentheses are percent change from control. For calculation of percent change for eyestalk-ablated (ESX) crabs and Ringer-injected intact crabs, intact crabs served as control; for met-injected crabs. Ringer-injected crabs served as control. a Not significant compared with intact crabs. b /> < 0.001 compared to intact crabs. L Not significant compared io eyestalkless crabs. Statistical analysis. Statistical analysis of the results was made using a two-way ANOVA test followed by Dunnet's multiple range test (preceded by one-way ANOVA), using SPSS version 10.0 (SPSS Inc., Chicago, ID. the possible mobilization of glucose molecules from hepa- topancreas and muscle to hemolymph. Phosphorylase (both total and active) activity levels were significantly increased in both hepatopancreas and muscle Results Effects of methionine-enkephalin on carbohvdrate metabolism of intact crabs Injection of methionine-enkephalin into intact crabs re- sulted in significant hyperglycemia and hyperglucosemia in a dose-dependent manner (Table 1 ). whereas injection of physiological saline had no effect on hemolymph carbohy- drate levels. At doses between 10~ 9 mol/crab (36.41%) and lO" 6 mol/crab (147.81%), the effect of methionine-en- kephalin was statistically significant. For doses lower than 10~ 9 mol/crab, however, methionine-enkephalin did not elicit a hyperglycemic response (Fig. 1 ). A time course for methionine-enkephalin-induced hyperglycemia is shown in Figure 2 for a IO" 7 mol/crab dose, which is a nearly saturating dose. The hemolymph glucose level increased significantly within 30 min of methionine-enkephalin injec- tion, reached a peak at 2 h, then declined gradually. Hepatopancreas glycogen and TCHO levels in crabs that received methionine-enkephalin were significantly lower than those of control crabs (Table 2). Decreases in muscle glycogen and TCHO levels were also significant after the injection of methionine-enkephalin (Table 3), suggesting 36 ~ 30 IIS MS SALINE- _1Q INJECTED 10 -9 '0 10 10 [Methioninc -Enkcphatin] (mol/crab) -6 10 10* Figure 1. Dose-dependent effect of methionine-enkephalin on the hemolymph glucose levels in intact Scylla serrata. Two hours after injec- tion of saline (10 /nl/animal) or methionine-enkephalin at the doses indi- cated, hemolymph was withdrawn from crabs for glucose determination. Each bar represents a mean SD (n = 10). Numbers in parentheses indicates the percent increase from the normal values. * Significant differ- ence from normal crabs at P < 0.001. NS Not significant. 20 P. S. REDDY AND B. KISHORI Time after injection (h) Figure 2. Time course of methionine-enkephalin-induced hyperglyce- mia. After injection of methionine-enkephalin (10~ 7 moL/crab). hemo- lymph was withdrawn from intact crabs at the time points indicated for glucose determination. Each point represents a mean SD (n = 10). Numbers in parentheses represent percent change from zero time controls. * Significant difference from zero time control at P < 0.001. ** Signif- icant difference from zero time control at P < 0.001. NS Not significant from zero time control. kephalin, indicating conversion of inactive to active phos- phorylase. Effects of bilateral e\estalk ablation and injection of methionine-enkephalin into ablated crabs on carbohydrate metabolism Bilateral eyestalk removal caused a significant decrease in hemolymph carbohydrate level (Table 1 ). Enhancement of TCHO level of hepatopancreas and muscle was also significant in eyestalk-ablated crabs (Tables 2, 3). The in- crease was greater in muscle. Glycogen level in hepatopan- creas increased significantly in eyestalkless crabs. A similar pattern was observed in muscle. Tissue phosphorylase ac- tivity levels decreased significantly in eyestalk-ablated crabs (Tables 4, 5). Injection of methionine-enkephalin into eyestalkless crabs did not significantly change hemolymph carbohydrate levels compared to Ringer-injected eyestalkless crabs (Ta- ble 1 ). The levels of tissue TCHO and glycogen and activity levels of total and active phosphorylase were also not sig- nificantly altered in eyestalkless crabs after methionine- enkephalin injection (Tables 2-5). after the injection of methionine-enkephalin (Tables 4. 5). The ratio of active to total phosphorylase also increased in the tissues of crabs after the injection of methionine-en- Discussion The effect of eyestalk hormones on tissue carbohydrate levels and phosphorylase activity has been extensively stud- Table 2 Effect of eyestalk ablation (ESX) (24-h post-ablation) and injection of methionine-enkephalin into intact ami ablated crabs on hepatopancreas total carbohydrate (TCHOl and glycogen levels in Scylla serrata Intact (Group 1) ESX (Group 2) No treatment Ringer injection 10 7 mol/crab 10~ s mol/crab 10 9 mol/crab Dunnet's comparison test TCHO (mg/g) 13.66 1.54 13.84 1.6T' 8.47 0.97 b 9.01 1.51 h 9.47 1.49 b ^,4.45, = 38.033 (1.32) (-38.80) (-34.89) (-31.57) P < 0.001 17.87 1.94 h 18.01 \.91 h -' 17.44 1.43 hx 17.X1 1.62 bx 17.93 1.59 b ' c F I44S , = 0.229 (30.96) (0.67) (-0.74) (-0.96) (-1.39) Two-way ANOVA: F, ,, (Between groups) = 566.317. P < 0.001; F 4 , m (Among treatments) = 19.027. P < 0.001; F 49(1 (Interaction) = 14.896. P < 0.001. Glycogen (mg/g) Intact 1.22 : t 0.10 1.23 0.09 a 0.58 ().14 h 0.61 0.13 h 0.64 0.2 l h F, 4 . 45 , = 148.477 (Group 1 ) (0.82) (-52.84) (-50.40) (-47.96) P < 0.001 ESX 2.04 0.29 h 2.06 0.31 h.c 2.11 1.1 8"^ 2.09 0.21 ht 2.07 0.28 b ' c F, 44 ,, = 0.230 (Group 2) (67. 21) (0.98) (2.42) (1.45) (0.48) Two-way ANOVA: F, g,, (Between groups) = 1658.593, P < 0.001; F 490 (Among treatments) = 24.964, P < 0.001; ^4.90 (Interaction) = 27.016. P < 0.001. Values are mean SD of 10 individual crabs. Values in parentheses are percent change from control. For calculation of percent change for ESX crabs and Ringer-injected intact crabs, intact crabs served as control; for met-injected crabs. Ringer-injected crabs served as control. J Not significant compared with intact crabs. * P < 0.001 compared to intact crabs. c Not significant compared to eyestalkless crabs. MET-ENKEPHALIN-INDUCED HYPERGLYCEMIA IN CRAB Table 3 Effect of eyestalk ablation treatment Ring er injection 10~ 7 mol/crab 10~ K mol/crab 10"' mol/crab Dunnet's comparison test TCHO (mg/g) Intact (Group 1 ) ESX (Group 2) 4.39 6.26 (4 0.53 0.71 h 2.59) 4.41 0.49 a (0.46) 6.31 0.8 l bx (0.80) 2.94 0.3 l h (-33.33) 3.01 0.37 h (-31.74) 6.25 O.S4 1 " (-0.95) 3.12 0.92 h (-29.25) 6.33 0.92 b - c (0.31) M4.45) P < ^14.45) ~ 30.829 .001 0.045 6.31 0.76 b ' c (0) Intact (Group 1) ESX (Group 2) Two-way ANOVA: F, go (Between groups) = 579.612. P < 0.001; F 4 (Among treatments) = 8.707, P < 0.001; F 4 QO (Interaction) = 9.1 14, P < 0.001. Glycogen (mg/g) 0.66 0.06 F<4.45> = 45.114 P < 0.001 F t 4.45)= 0.188 0.64 0.09" 0.34 0.09 b 0.37 0.06" 0.41 0.08 h (-3.03) (-46.87) (-42.18) (-35.31) 1.01 0.09 b 1.02 O.ll"- c 0.99 0.14 b ' c 1.07 0.33 bx 1.03 0.2 l hx (53.03) (0.99) (-2.94) (4.90) (0.98) Two-way ANOVA: F, gn (Between groups) = 422.031. P < 0.001; F 4 9() (Among treatments) = 6.391, P < 0.001; F 41 ,,, (Interaction) = 6.713. P < 0.001. Values are mean (mg glucose/g tissue) SD of 10 individual crabs. Values in parentheses are percent change from control. For calculation of percent change for ESX crabs and Ringer-injected intact crabs, intact crabs served as control; for met-injected crabs. Ringer-injected crabs served as control. a Not significant compared with intact crabs. b /> < 0.001 compared to intact crabs. 1 Not significant compared to eyestalkless crabs. led (Keller, 1965; Ramamurthi et al.. 1968; Sagardia, 1969). Eyestalk removal inactivates the phosphorylase system and activates uridine-diphosphate-glucose glycogen transglu- cosylase (glycogen synthetase) (Keller, 1965; Ramamurthi ct nl.. 1968). Ramamurthi et al. ( 1968) also observed stim- ulation of uptake and incorporation of glucose I4 C into the glycogen fraction of muscle tissue after eyestalk removal; this stimulation was accompanied by a decrease in hemo- lymph sugar level. Injection of eyestalk extract reversed these changes. The hyperglycemic hormone of eyestalks of the crab Oziotelphusa senex sene.x and the prawn Penaeus monodon enhances the activity of the phosphorylase system (Reddy et al.. 1982, 1984; Reddy, 1992). An increase in phosphorylase activity and a decrease in glycogen and TCHO levels in hepatopancreas and muscle of Scylla serrata, followed by hyperglycemia after the injec- tion of methionine-enkephalin, indicate glycogenolysis and mobilization of sugar molecules from tissues to hemo- lymph. This is in agreement with other findings (see review by Reddy and Ramamurthi, 1999). Though the hormone that elevates hemolymph sugar is conventionally called crustacean hyperglycemic hormone (CHH). Hohnke and Scheer (1970) suggested that the primary function of the CHH is not to elevate hemolymph sugar level, but to elevate intracellular glucose through the degradation of glycogen by activating the enzyme phosphorylase. The conversion of phosphorylase from its inactive to active form results in glycogenolysis, and the resultant glucose molecules leak into the hemolymph, causing hyperglycemia. This view has been supported by Telford (1975). Our results clearly demonstrate that methionine-enkepha- lin is involved in the regulation of carbohydrate metabolism in the crab Scylla serrata. In the present study, we show that methionine-enkephalin elicited a hyperglycemic response in S. serrata in a dose-dependent manner (Fig. 1 ). Methionine- enkephalin-induced hyperglycemia has been similarly dem- onstrated in the freshwater crab Oziotelplutsa senex senex (Reddy, 1999) and the brackish-water prawns Penaeus in- dicus and Metapenaeits monocerus (Kishori et al., 2001). The doses of methionine-enkephalin that induced hypergly- cemia ranged from 10 9 to 10~ 6 mol/animal (Fig. 1 ), which is comparable to those reported for O. senex senex (Reddy, 1999). Our observation that methionine-enkephalin was in- effective in inducing hyperglycemia in eyestalk-ablated S. serrata (Table 1) is also consistent with those obtained in crabs (Reddy, 1999) and prawns (Kishori et al., 2001 ) and suggests that the hyperglycemic effect of methionine-en- kephalin results from an enhanced release of CHH (Keller, 1992; Soyez. 1997). Injection of methionine-enkephalin into intact S. serrata also has two other effects. It activates the phosphorylase system, which causes degradation of glycogen. It also re- sults in accumulation of sugar molecules in the tissues; these molecules are ultimately mobilized to hemolymph. 22 P. S. REDDY AND B. KISHORI Table 4 Effect of evestalk ablation I ESX) (24 h post-ablation) and injection of methionine-enkephalin into intact and ablated crabs on hepatopancreas phosphorylase activity levels in Scylla serrata No treatment Ringer injection 10~ 7 mol/crab 10~ 8 mol/crab 10 " mol/crab Dunnet's comparison test Phosphorylase a 3.63 0.34 h 3.60 0.42" F, 4 45 , = 28.430 (35.95) (34.83) P < 0.001 1.81 0.22 b - c 1.84 0.31 b ' c F, 445 , = 1.473 (8.38) (10.17) Two-way ANOVA: F, w , (Between groups) = 716.848. P < 0.001; F 4 91 , (Among treatments) = 23.852, P < 0.001; F 4 , (Interaction) = 18.208, P < 0.001. Intact 2.62 0.29 2.67 0.33" 3.87 0.46 b (Group 1) ESX 1.72 0.3 l h (1.91) 1.67 0.29 b - c (44.94) 1.69 O.ll"- 1 - (Group 2) (-34.35) (-2.33) (1.19) Phosphorylase ab Intact (Group 1) ESX (Group 2) 4.52 0.41 4.56 0.44 a 5.81 0.67 b (0.89) (27.41) 4.06 0.44 b 4.08 0.41 hL 4.10 0.39"-' (-10.18) (0.49) (0.49) 5.69 0.52" (24.78) 4.12 0.34 h - c (0.98) 5.56 0.73" (21.92) 4.09 0.51 b ' c (0.24) F (4 . 45) = 15.846 P < 0.001 F, 44 ,, = 0.044 Two-way ANOVA: F, uo (Between groups) = 169.103, P < 0.001; F 4 w (Among treatments) = 11.291, P < 0.001: F 4 ,,,, (Interaction) = 9.985. P < 0.001. Values are mean (iP released/mg protein/h) SD of 10 individual crabs. Values in parentheses are percent change from control. For calculation of percent change for ESX crabs and Ringer-injected intact crabs, intact crabs served as control; for met-injected crabs. Ringer-injected crabs served as control. a Not significant compared with intact crabs. b P < 0.001 compared to intact crabs. c Not significant compared to eyestalkless crabs. causing hyperglycemia. Methionine-enkephalin might have elevated the phosphorylase system in intact crabs in several different ways for example, by triggering release of hy- perglycemic hormone or by mimicking the action of this hormone. However, because methionine-enkephalin was not able to produce these changes in eyestalkless crabs, it seems most likely that methionine-enkephalin exerted its hyperglycemic effect by triggering release of hyperglyce- mic hormone from the sinus gland of eyestalks. This sup- ports our earlier results that sinus glands in the eyestalks of crabs are the main release site for hyperglycemic hormone (Reddy and Ramamurthi, 1982). The mechanisms whereby methionine-enkephalin causes release of neurohormones are still uncertain. In mammals, endogenous opioid peptides are involved in regulating the release of neurohypophysial peptides (Bicknell et al.. 1988; Yamada et al., 1988; Sasaki et ai, 2000). In crustaceans, opioid-peptide-like (methionine-enkephalin-like, leucine- enkephalin-like and /?-endorphin-like) hormones were iso- lated and characterized from X-organ sinus gland com- plexes of eyestalks (Fingerman et ai, 1983, 1985). However, there is little information about the effect of opioid peptides on release of neurohormones in crustaceans. Sarojini et al. (1995, 1996). using highly selective opioid antagonists, provided evidence that methionine-enkephalin exerts its effect by acting through delta-type opioid recep- tors in regulating ovarian maturation in Procambarus clarkii. In vivo studies with tissues of P. clarkii showed that methionine-enkephalin exerted its effect by at least modu- lating the release of eyestalk peptide hormone (Sarojini et al., 1997). Recently, we provided evidence for a neurotrans- mitter role for methionine-enkephalin in causing hypergly- cemia in the crab O. senex senex (Reddy, 1999). Methio- nine-enkephalin also triggers the release of red-pigment- concentrating hormone, black-pigment-dispersing hormone (Quackenbush and Fingerman, 1984). and dark-pigment- adapting hormone (Kulkarni and Fingerman, 1987). Three facts make it seem likely that this hyperglycemic action of methionine-enkephalin in the present study on S. serrata is also indirect and involves stimulation of release of CHH. Methionine-enkephalin-like material is present in the neu- roendocrine complex of the eyestalk of crustaceans (Finger- man et al., 1983. 1985). Methionine-enkephalin mediation of release of neurohormones has been demonstrated (Reddy, 1999). In cases where methionine-enkephalin has been found to stimulate neurohormone release, it does not act in the absence of neuroendocrine organs. As further support for the conclusion, eyestalk extract from methio- nine-enkephalin injected prawns showed significantly less activity than the normal eyestalk extract in inducing hyper- glycemia (Kishori et al., 2001 ). Although the mechanisms that trigger release of CHH are still unknown, it is noteworthy that 5-hydroxytryptamine (5-HT), or serotonin, triggers CHH release in the crayfish MET-ENKEPHALIN-INDUCED HYPERGLYCEMIA IN CRAB 23 Table 5 Effect of evesta/k ablation I ESX) (24 h post-ablation) and injection of methinine-enkephalin into intact and ablated crabs on muscle phosphor/lose activin levels in Scylla serrala No treatment Ringer injection 10~ 7 mol/crab 10 ~ s mol/crab 10~" mol/crab Dunnet's comparison test Phosphorylase a Intact 1 .92 0.09 1.94 0.1 4 a 3.26 0.36 b 3.01 0.12' 1 3.02 0.26 h F, 4 . 451 = 84.853 (Group 1 ) (1.04) (68.04) (55.15) (55.67) P < 0.001 ESX 0.99 0.08 h 1.02 O.ll hl 1.01 0.09 Kc 1.04 O.I3 hc 1.06 0.2 1 KC F |44 ,, = 0.368 (Group 2) (-48.44) (3.03) (-0.74) (1.96) (3.92) Two-way ANOVA: F, 91 , (Between groups) = 1711.188. P < 0.001; F 4 M1 , (Among treatments) = 58.745. P < 0.001; F 4 w (Interaction) = 52.927, P < 0.001. Phosphorylase ab Intact 2.49 0.45 2.52 0.49 a 3.49 0.4 l h 3.46 0.44 h 3.44 0.51 h F ( 44S) = 16.086 (Group 1) (1.21) (38.49) (37 .30) (36.50) P < .001 ESX 2.22 * 0.32 b 2.18 0.31 bc 2.22 0.34 b - c 2.24 0.42 hc 2.25 0.41 b.t F, 4.45) = 0.061 (Group 2) (-11 .65) (-0.91) (1.83) (2. 75) (3.21) Two-way ANOVA: F, ,, (Between groups) = 136.048, P < 0.001; F 4 , m (Among treatments) = 10.259, P < 0. 001; " 4, MO (Interaction) = 8.734, P < 0.001. Values are mean (iP released/mg protein/h) SD of 10 individual crabs. Values in parentheses are percent change from control. For calculation of percent change for ESX crabs and Ringer-injected intact crabs, intact crabs served as control; for met-injected crabs. Ringer-injected crabs served as control. a Not significant compared with intact crabs. * P < 0.001 compared to intact crabs. c Not significant compared to eyestalkless crabs. Orconectes limosus (Keller and Bayer, 1968), Astacus lep- todactylus (Strolenberg and Van Herp, 1977), and Procam- barus clarkii (Lee et ai. 2000). Strolenberg and Van Herp (1977). working with A. leptodactylus, and Martin (1978). working with Porcellio dilatatits, found that the sinus glands of specimens injected with 5-HT show increased numbers of exocytotic profiles, suggestive of increased CHH release. Exocytosis in A. leptodactylus was maximal 2 h after 5-HT was injected, and the hemolymph glucose concentration peaked 4 h after the injection (Strolenberg and Van Herp. 1977). In P. dilatatus, hyperglycemia in- duced by 5-HT is mediated by 5-HT,- and 5-HT : -like receptors in triggering release of CHH (Lee et ai, 2000). In summary, we have shown that methionine-enkephalin is a potent hyperglycemic regulator in the crab Scylla ser- nita. The most likely site of action of methionine-enkepha- lin is the eyestalks. where the X-organ-sinus glands may respond to methionine-enkephalin stimulation by releasing CHH. Based on these results, experiments are being con- ducted to determine whether methionine-enkephalin en- hances the release of CHH in crustaceans. Acknowledgments We thank Prof. Armugam, University of Madras. Chen- nai, for supplying Sc\lla serrata and providing necessary laboratory facilities, and Dr. K. V. S. Sharma, Professor, Department of Statistics, Sri Venkateswara University, for analyzing the data. We also thank the anonymous reviewers whose comments improved our manuscript. We are grateful to Prof. R. Ramamurthi. Department of Zoology, for his encouragement. Mr. S. Umasankar and Miss B. Prema Sheela provided skilled technical assistance. This work was carried out with the financial assistance from Department of Science and Technology research grant (SP/SO/CO4/96) to Dr. PSR. We also thank the staff. Department of Biotech- nology, Sri Venkateswara University, for their invaluable assistance. Literature Cited Abramowitz, A. A., F. L. Hisaw, and D. N. Papandrea. 1944. The occurrence of a diabetogenic factor in the eyestalks of crustaceans. Bil. Bull. 86: 1-5. Bicknell, R. J., G. Leng, J. A. Russell, R. G. Dyer, S. Mansfield, and B. G. Zhao. 1988. Hypothalamic opioid mechanisms controlling oxytocin neurons during parturition. Brain Res. Bull. 20: 743-749. Carroll, N. V., N. N. Longley, and J. G. Roe. 1956. Glycogen deter- mination in liver and muscle by use of anthrone reagent. J. Biol. Chem. 220: 583-593. Cori, G. T., B. Illingworth, and P. G. Keller. 1955. Muscle phosphor- ylase. Pp. 200-207 in Methods in En-ymology. Vol. I. S. P. Colowick and N. O. Kaplan, eds. Academic Press, New York. Fingerman. M., M. M. Hanumante. and L. L. Vacca. 1983. Enkepha- lin-like and substance P-like immunoreactivity in the eyestalk neuroen- 24 P. S. REDDY AND B KISHORI docrine complex of the fiddler crab, Uca pugilator. Soc. Neurosci. Abstr. 9: 439. Fingerman, M.. M. M. Hanumante. (.. K. Kulkarni, R. Ikeda, and L. L. Vacca. 1985. Localization of substance P-like, leucine-en- kephalin-like. methionine-enkephalin-like. and FMFR amide-like im- munoreactivity in the eyestalk of the fiddler crab, Uca pugilator. Cell Tissue Res. 241: 473-477. Hohnke. L., and B. T. Scheer. 1970. Carbohydrate metabolism in crustaceans. Pp. 147-166 in Chemical Zoology, Vol. 5. M. Florkin and B. T. Scheer. eds. Academic Press. New York. HII!M i in. in. A., M. B. Aguilar, K. Brew, J. Shabanowitz, and D. F. Hunt. 1993. Primary structure of the major isomorph of the crusta- cean hyperglycemic hormone (CHH-1) from the sinus gland of Mex- ican crayfish Procambarus bouvieri (Ortmann): interspecies compari- son. Peptides 14: 7-16. Kegel, G., B. Reichwein, S. Weese, G. Gaus, J. Peter-Katalinic, and R. Keller. 1989. Amino acid sequence of the crustacean hyperglycemic hormone (CHH) from the shore crab, Carcinus maenas. FEBS Lett. 255: 10-14. Kegel, G., B. Reichwein, C. P. Tensen, and R. Keller. 1991. Amino acid sequence of crustacean hyperglycemic hormone (CHH) from the crayfish Orconectes limosus. Emergence of a novel neuropeptide fam- ily. Peptides 12: 909-913. Keller, R. 1965. Uber eine hormonale Kontrolle des Polysaccharidstoff wechsels beim Flusskiebs Cambarits maenas. Z. Vgl. Pliysiol. 51: 49-59. Keller, R. 1992. Crustacean neuropeptides: structures, functions and comparative aspects. Experientia 48: 439-448. Keller, R., and J. Beyer. 1968. Zur hyperglykamischen Wirkung von Serotonin und Augeststielextrakt beim Flusskrebs Orconectes limosus. Z. Vgl. Physio/. 59: 78-85. Keller, R., P. Jaros. and G. Kegel. 1985. Crustacean hyperglycemic neuropeptides. Am. Zool. 25: 207-221. Kishori. B., B. Premasheela, R. Ramamurthi, and P. S. Reddy. 20(11. Evidence for hyperglycemic effect of methionine-enkephalin in prawns Penaeus indicus and Metapenaeus monocerus. Gen. Comp. Endocri- nnl. (in press I. Kulkarni, G. K., and M. Fingerman. 1987. Distal retinal pigment of the fiddler crab, Uca pugilator. Release of the dark-adapting hor- mone by methionine-enkephalin and FMRF-amide. Pigm. Cell Res. 1: 51-56. Lacombe, C., P. Greve, and G. Martin. 1999. Overview on the sub- grouping of the crustacean hyperglycemic hormone family. Neuropep- tides 33: 71-80. Lee, C. Y., S. M. Yau, C. S. Liau, and W. J. Huang. 2000. Serotogenic regulation of blood glucose levels in the crayfish. Procambarus clarkii: site of action and receptor characterization. J. E.\p. Zool. 286: 596- 605. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with Folin phenol reagent. J. Biol. Chem. 193: 265-275. Mancillas, J. R., J. F. McGinty, A. I. Selverston, H. Karten, and F. E. Bloom. 1981. Immunocytochemical localization of enkephalin and substance P in retina and eyestalk neurones of lobster. Nature 293: 576-578. Martin, G. 1978. Action de la serotonine sur la glycemie et sur la liberation des neurosecretions contenues dans la glande du sinus de Porcellio di/atatus Bradt (Crustace. Isopode, Oniscoide). C. R. Seances Soc. Biol. Fil. 172: 304-307. Martin, G., O. Sorokine, and A. Van Dorsselaer. 1993. Isolation and molecular characterization of hyperglycemic neuropeptide from the sinus gland of the terrestrial isopod Armadillidium vulgare (Crustacea). Em: J. Biochem. 211: 601-607. Ohira, T., T. Watanabe, H. Nagasawa, and K. Aida. 1997. Cloning and sequence analysis of a cDNA encoding a crustacean hyperglycemic hormone from the Kuruma prawn Penaeus japonicus. Mot. Mar. Biol Biotechnol. 6: 59-63. Pantin, C. F. A. 1934. The excitation of crustacean muscle. ./. Exp. Biol. 11: 11-27. Quackenbush, L. S., and M. Fingerman. 1984. Regulation of neuro- hormone release in the fiddler crab. Uca pugilator: Effects of gamma- aminobutyric acid, octopamine, met-enkephalin and beta-endorphin. Comp. Biochem. Physiol. 79C: 77-84. Ramamurthi, R., M. W. Mumbach. and B. T. Scheer. 1968. Endo- crine control of glycogen synthesis in crabs. Comp. Biochem. Phvsiol. 67B: 437-445. Reddy, P. S. 1992. Changes in carbohydrate metabolism of Pila globosa in response to crustacean hyperglycemic hormone. Arch. Int. Phvsiol. Biochim. Biophys. 100: 281-283. Reddy. P. S. 1999. A neurotransmitter role for methionine-enkephalin in causing hyperglycemia in the fresh water crab O-iotelplntsa senex senex. Curr. Sci. 76: 1126-1128. Reddy, P. S., and R. Ramamurthi. 1982. Neuroendocrine control of carbohydrate metabolism in the rice field crab (Oziotelphusa senex sene.\). J. ReprocL Biol. Comp. Endocrinol. 2: 49-57. Reddy, P. S., and R. Ramamurthi. 1999. Recent trends in crustacean endocrine research. Proc. Indian Natl. Sci. Acad. B65: 15-32. Reddy, P. S., A. Bhagyalakshmi, V. Chandrasekharam, and R. Ra- mamurthi. 1982. Hyperglycemic activity of crab and scorpion hor- mones in grasshopper. Poecilocerus pictus. Experientia 38: 811-812. Reddy, P. S., A. Bhagyalakshmi, V. Chandrasekharam, and R. Ra- mamurthi. 1984. Differential responsiveness of phosphorylase sys- tem and fat body carbohydrates of Poecilocerus pictus (Insecta) to hyperglycemic factors of crab (Crustacea) and scorpion (Arachnida). Gen. Comp. Endocrinol. 54: 43-45. Reddy, P. S., G. D. Prestwich, and E. S. Chang. 1997. Crustacean hyperglycemic hormone gene expression in the lobster Homants americanus. Pp. 51-5 in Advances in Comparative Endocrinology, Vol. 1, S. Kawashima and S. Kikuyama, eds. Monduzzi Editore, Bologna. Italy. Sagardia, F. 1969. The glycogen phosphorylase system from the muscle of the blue crab, Callinectes danae. Comp. Biochem. Physiul. 28: 1377-1385. Sarojini, R., R. Nagabhushanam, and M. Fingerman. 1995. Evidence for opioid involvement in the regulation of ovarian maturation of the fiddler crab, Uca pugilator. Comp. Biochem. Physiol. 111A: 279-282. Sarojini, R., R. Nagabhushanam, and M. Fingerman. 1996. In vivo assessment of opioid agonists and antagonists on ovarian maturation in the red swamp crayfish. Procambarus clarkii. Comp. Biochem. Physiol. 115C: 149-153. Sarojini, R., R. Nagabhushanam, and M. Fingerman. 1997. An in vitro study of the inhibitory action of methionine enkephalin on ovarian maturation in the red swamp crayfish, Procambarus clarkii. Comp. Biochem. Physiol. 117C: 207-210. Sasaki, T., K. Shimada. and N. Saito. 2000. Regulation of opioid peptides on the release of arginine vasotocin in the hen. J. Exp. Zool. 286: 481-486. Sedlmeier, I). 1985. Mode of action of the crustacean hyperglycemic hormone. Am. Zoo/. 25: 223-232. Soyez, D. 1997. Occurrence and diversity of neuropeptides from the crustacean hyperglycemic hormone family in arthropods. Ann. NY Acad. Sci. 814: 319-323. MET-ENKEPHALIN-INDUCED HYPF.RGLYCEM1A IN CRAB 25 Slrolenberg, G. E. C.. and F. Van Herp. 1977. Mise en evidence du phenomene d'exocytose dans la glande du sinus il'A.itacus leptodactv- lus (Nordmannl sous I'inrluence d'injection de serotonine. C. R. Acad. Sci. Ser. Ill Sci. Vie 284D: 57-59. Taussky, H. M., and E. Shorr. 1953. A micro colorimetric method tor determination of inorganic phosphate. J. Biol. Cliem. 202: 675-682. Telford, M. 1975. Blood glucose in crayfish III. The sources of glu- cose and role of eyestalk factor in hyperglycemia of Cambarus robus- tns. Catnp. Biocht'in. Pln'siol. 51: 69-73. Jensen. C. P.. D. P. V. De Kleijn, and F. Van Herp. 1991. Cloning and sequence analysis of cDNA encoding two crustacean hyperglycemic hormones trom the lobster Homctrus americanus, Eur. J. Bioi'hein. 20(1: 103-106. Vamada, T., K. Nakao, H. Itoh. N. Morii, S. Shiono, M. Sakamoto, A. Sugawara, Y. Saito, M. Mukoyama, H. Aral, M. Eigyo, A. Matushita, and H. Imura. 1988. Inhibitory action of leumorphin on vasopressin secretion in conscious rats. Endocrino/ogv 122: 985- 990. Yasuda, A., Y. Yasuda, T. Fujita, and Y. Naya. 1994. Characterization of crustacean hyperglycemic hormone from the crayfish (Procambarus clarkii): multiplicity of molecular forms by stereoinversion and diver- sion functions. Gen. Comp. Endocrinol. 95: 387-398. Reference: Biol. Bull. 201: 26-33. (August 2001) Theoretical and Experimental Dissection of Gravity-Dependent Mechanical Orientation in Gravitactic Microorganisms YOSHIHIRO MOGAMI,* JUNKO ISHII.t AND SHOJI A. BABA Depanment of Biology, Ochanomizit University. Otsuko, Tokyo 1 12-8610, Japan Abstract. Mechanisms of gravitactic behaviors of aquatic microorganisms were investigated in terms of their mechan- ical basis of gravity-dependent orientation. Two mechanical mechanisms have been considered as possible sources of the orientation torque generated on the inert body. One results from the differential density within an organism (the grav- ity-buoyancy model) and the other from the geometrical asymmetry of an organism (the drag-gravity model). We first introduced a simple theory that distinguishes between these models by measuring sedimentation of immobilized organisms in a medium of higher density than that of the origanisms. Nr + -immobilized cells of Paramecium caitda- tuin oriented downwards while floating upwards in the Percoll-containing hyper-density medium but oriented up- wards while sinking in the hypo-density control medium. This means that the orientation of Paramecium is mechan- ically biased by the torque generated mainly due to the anterior location of the reaction center of hydrodynamic stress relative to those of buoyancy and gravity; thus the torque results from the geometrical fore-aft asymmetry and is described by the drag-gravity model. The same mechan- ical property was demonstrated in gastrula larvae of the sea urchin by observing the orientation during sedimentation of the KCN-immobilized larvae in media of different density: like the paramecia, the gastrulae oriented upwards in hypo- density medium and downwards in hyper-density medium. Immobilized pluteus larvae, however, oriented upwards re- gardless of the density of the medium. This indicates that the orientation of the pluteus is biased by the torque gen- erated mainly due to the posterior location of the reaction center of gravity relative to those of buoyancy and hydro- Received 14 July 2000; accepted 30 March 2001 * To whom correspondence should he addressed. E-mail: mogami@ cc.ocha.ac.jp t Died on 1 March 1999. dynamic stress; thus the torque results from the fore-aft asymmetry of the density distribution and is described by the gravity-buoyancy model. These observations indicate that, during development, sea urchin larvae change the mechanical mechanism for the gravitactic orientation. Evi- dence presented in the present paper demonstrates a definite relationship between the morphology and the gravitactic behavior of microorganisms. Introduction Many swimming microorganisms, including ciliate and flagellate protozoa and the planktonic larvae of some inver- tebrates, are negatively gravitactic; that is, they tend to swim preferentially upwards in water columns despite being heavier than water. This behavior requires the organism to orient upwards in relation to the gravity vector. Several mechanisms have been postulated for the gravitactic orien- tation of aquatic microorganisms (Chia el al., 1983; Bean, 1984; Machemer and Braucker. 1992). From a physical point of view and taking account of the mechanical prop- erties of these microorganisms, it has been postulated that the interaction of gravitational and hydrodynamic forces may cause them to orient with fore end upward. In addition to the mechanical basis, gravitactic orientation might also be explained on the physiological basis of gravity percep- tion. To modulate the propulsive activity, some mechano- sensitive devices that sense gravity (for example, statocysts) might be needed. Although functional statocysts have been found in some unicellular organisms (Fenchel and Finlay, 1984, 1986), a line of evidence for gravity-dependent mod- ulation of propulsion has been accumulated for Paramecium (Machemer et al., 1991; Ooya et al., 1992) and Eitglenu (Machemer-Rohnisch el al., 1999), which have no stato- cyst-like structure. The present paper focuses on the mechanical properties of 26 MECHANICAL BIAS OF MICROBIAL GRAVITAXIS 27 microorganisms, which, irrespective of propulsion, generate the torque to orient the organisms either upwards or down- wards. This mechanical torque should bias the gravitactic orientation, even if the organisms have active physiological mechanisms of gravitaxis. According to Roberts (1970), two mechanical mechanisms have been considered as possible sources of the orientation torque. These are reconsidered, in the present paper, as two mechanical models, the gravity-buoy- ancy model and the drag-gravity model. The gravity-buoyancy model was first postulated by Ver- worn ( 1 889. cited in Machemer and Braucker. 1992) for the negative gravitaxis of Paramecium. This model is based on the differential density within an organism. If the internal density of the organism is not homogeneous, the center of mass (the center of gravity) does not necessarily coincide with the centroid (the center of buoyancy). Posterior accu- mulation of the mass would result in the upward orientation of the organisms, and anterior accumulation would result in the downward orientation. The drag-gravity model was postulated by Roberts (1970) on the basis of the low Reynolds number hydrodynamics of the swimming of microorganisms that have a geometrical fore-aft asymmetry. This model is characterized by a dumb- bell with two spheres of unequal diameter but homogeneous density, which could mimic the fore-aft asymmetry of the microorganisms. According to Stokes' drag formula, the larger sphere of the dumbbell can sink faster than the smaller, at the rate of the square of the ratio of diameters. The applicability of this model has been confirmed by scale-model experiments (Roberts, 1970). Organisms, in general, possess some asymmetry both in internal density and in external geometry. It is therefore pos- sible that these two mechanical models operate independently to generate the gravity-induced orientation torque. Since the mechanical properties for gravitactic orientation are indepen- dent of propulsive thrust, we can assess the mechanical influ- ence by measuring the orientation of immobilized organisms sinking under gravity. Both models predict that, when immo- bilized, an organism orients upwards when sinking in a me- dium with a density lower than its own. In the present paper, we show that the above two models can be distinguished by observing what happens to an organism placed in a medium whose density is higher than its own. We show the results of the experiments on the gravitactic orientation of Paramecium and sea urchin lar- vae, both of which are known to perform typical negative gravitaxis (Mogami et /., 1988: Ooya et al.. 1992). Theory The external forces acting on the body of an aquatic micro- organism due to gravity acceleration are gravitational (F c ) and buoyant forces (F B ), each of which is generated as the product of the volume and density of the body or of the external fluid. The vector sum of the forces encounters the hydrodynamic force (F H ). Since the Reynolds number of an aquatic micro- organism in translational motion is significantly less than unity (of the order of 10 2 ). F H is generated in proportion to the velocity (Happel and Brenner, 1973; Vogel, 1994). F c , F B , and F H act on the center of mass (G), the centroid (B), and the reaction center of hydrodynamic stress (//), respectively. For an immobilized microorganism sinking in the fluid, these three forces are balanced as B + FH = 0. (1) Each term in the equation (positive in upward direction) is described as F G = - Vp,g, F B = Vpg. and (2) (3) (4) where V and p, are the total volume and the average density of the organism, p and g the density of the external fluid and the acceleration due to gravity, and K and 5 the coefficient of hydrodynamic drag and the sinking velocity. We assume in the present paper that a microorganism has a body of rotating symmetry on its fore-aft axis. The sim- plest case of this approximation is that the body has fore-aft symmetry, such as a prolate spheroid. When a prolate spher- oid with uniform density is sinking in the fluid, the three forces act on the same point and therefore do not generate any torque to rotate the body (Fig. la). If. however, the body of a prolate spheroid has a region of higher density in the rear half of the body, as postulated in the gravity-buoyancy model, G is located posterior to B and H (Fig. Ib). This generates the torque (7\,; subscript V is after Verworn) which is given by TV = sin 0, (5) where L G is the distance between G and B (and/or H), and 6 is the orientation angle of the fore-aft axis of the body to the vertical. The fore-aft asymmetry of the external geometry, as postulated in the drag-gravity model, also separates the reaction centers of the forces. If a microorganism of homo- geneous density has a larger radius of revolution around the fore-aft axis in the posterior part (Fig. Ic). H is located anterior to B and G, according to the analogy of a fore-aft asymmetrical dumbbell of homogeneous density (Happel and Brenner. 1973). The torque (T R : subscript R is after Roberts) by the anterior shift of the center of hydrodynamic force is given by T R = -F H L H s'm = (F c + F B )L H sin 6, (6) where L H is the distance between H and G (and/or B). Provided that the Reynolds number of rotational motion is sufficiently small, all torques should be proportional to 28 Y. MOGAMI ET AL. Figure 1. Schematic drawings illustrating the mechanical (physical) basis for the generation of gravity- dependent orientation torque. Gravity (F G ), buoyancy (F B ), and hvdrodynamic force (F H ) are balanced in sinking microorganisms; these forces act at the center of mass (G), the centroid (B). and the reaction center of hydrodynamic stress (//), respectively, (a) Three forces act at the same point in the body of prolate spheroid with uniform density, (b) The center of mass is deviated to the rear end of the body of prolate spheroid, which generates the torque in proportion to F a and the sine of the orientation angle to the gravity vector (W). (c) The reaction center of hydrodynamic stress is deviated to the front end of the body with fore-aft asymmetry but with uniform density, which generates the torque in proportion to the vector sum of F Cl and F H and the sine of the orientation angle. the first power of rotational velocity (dQIdt). In such cases equations of rotational motion are given by -flTj~ = T v orT K , (7) where R is the coefficient of resistance for rotational motion and T) is the viscosity of the external fluid. From these equations the rotational velocity of each model is given as a common form of dO -=3sin0. (8) where the proportional factor is the instantaneous rate at = 90 degrees, and given by (9) (10) Rj] V(p,-p)gL H Rr, for the gravity-buoyancy and drag-gravity models, respec- tively. Equations 9 and 10 indicate that /3 r is insensitive to changes in the density of the external medium (p), whereas f3 K reverses the sign as p exceeds the density of organisms (p,-). This means that the two models can be distinguished by increasing p greater than p,. When im- mobilized organisms are immersed in the hyper-density medium (p > p,), they would orient upwards during floating upwards if they obeyed the gravity-buoyancy model, whereas they would orient downwards if they obeyed the drag-gravity model. The gravity-buoyancy and drag-gravity models are the two extremes of these conditions that can generate the orientation torque depending on the different physical mechanisms. Passive orientation of the organisms (Eq. 8), in fact, would be explained as a result of combining the two models, because none of three forces would necessarily have a common reaction center. In order to extract the origin of the mechanical bias of the orientation. Equation 8 should be examined by measuring |3 by the sedimentation experi- ment using media of different p. If j3 is constant independent of p, the gravity-buoyancy model is the only mechanism for generating the orientation torque. Otherwise, the drag-grav- ity model may play a part in the generation of the torque. A negative value of in the hyper-density medium indicates MECHANICAL BIAS OF MICROBIAL GRAVITAXIS 29 that the drag-gravity model is the major mechanism in passive gravitactic orientation. Materials and Methods Microorganisms and experimental solutions Paramecium caudatum was grown at 24 C in a hay infusion in Dryl's solution (2 mM sodium citrate, 1.2 mM Na,HPO 4 . 1 .0 mM NaH 2 PO 4 , 1 .5 mM CaCU, pH 7.2). Cells grown to the early stationary phase (14-20 d after incuba- tion) were collected and adapted in the experimental solu- tion (KCM; 1.0 mM KC1, 1.0 mM CaCU, 1.0 mM MOPS, pH 7.2). After the adaptation, cells gravitactically accumu- lating beneath the water surface were collected and immo- bilized in the KCM containing 5 mM NiCU. Hyper-density KCM (P-KCM) was prepared by substituting a colloidal solution of Percoll (Sigma) for water up to 60% (v/v) in KCM. At 24 C, the specific gravity and relative viscosity of KCM were 1.00 and 1.02, respectively; those of P-KCM were 1.06 and 1.57. Specific gravity of the experimental solutions was determined by weighing the known volume, and viscosity was measured by means of an Ostwald vis- cometer. Larvae of the sea urchin Hemicentrotus pulcherrimus were grown in the laboratory at 17 C (Degawa et ai, 1986). Larvae at the mid- to late gastrula stage and the early pluteus stage (ca. 24 and 48 h after insemination, respec- tively) were collected by hand centrifuge and washed once with artificial seawater (ASW; 450 mM NaCl. 10 mM KC1, 10 mM CaCl 2 , 25 mM MgCl 2 . 28 mM MgSO 4 , 10 mM Tris-HCl, pH 8.0). For immobilization, larvae were im- mersed in ASW containing 2 mM KCN. Hyper-density ASW (P-ASW) was prepared by substituting Percoll for water up to 22% (v/v) in ASW. At 25 C, the specific gravity and relative viscosity of ASW were 1.01 and 1.07, respectively; those of P-ASW were 1.04 and 1.14. Recordings and analyses of gravity-dependent orientation Ni 2+ -immobilized Paramechtm cells and KCN-immobi- lized sea urchin larvae were transferred, with experimental solutions to be tested, into a chamber made of a slide and coverslip and silicone rubber spacer (inner dimension 12 X 24 X 1 mm for Paramecium and 1 6 X 1 6 X 1 mm for sea urchin larvae) and kept air bubble-free without any partic- ular sealant. The chamber was set on a horizontal micro- scope equipped with a rotating stage. After trapping immo- bilized specimens at the bottom or the top of the chamber (depending on the density of the medium), the chamber was rotated upside down, and the orientation motion during vertical movement due to gravity was recorded with a video camera (XC-77, Sony, Tokyo) and a videotape recorder. To avoid the hydrodynamic interactions between nearby mov- ing objects, we chose organisms moving down (or up) far from neighbors (>1 mm, about 5 body lengths, apart). For measuring the orientation angle, we selected recordings in which the orientation motion was observed in a single focal plane. The orientation angle as a function of time (0, t) was measured directly on the video monitor. The rotational velocity as a function of orientation angle (dOldt, 6) was obtained as an average velocity ((fl,+ i - 0,)/A?) at the angle of geometrical average ((0, + 0, + ,)/2) between every successive datum of inclination angle versus time. /3 in Equation 8 was obtained by nonlinear least-squares re- gression of the velocity data (dO/dt, 0) to the equation d9 ~dt = /3 sin (9 + a). (ID where a is a factor to adjust the angle between the morpho- logically defined fore-aft axis and the mechanically defined axis. Results The drag-gravity model is the major mechanism of Paramecium When Paramecium was immobilized by Ni 2 + , it main- tained an anterior-thinner cell shape. This shape was pre- served in P-KCM as well as in KCM; cells showed no significant changes in axial length (162 17 jam [/; = 30] and 163 16 jim [n = 21]. P = 0.64, for cells in KCM and P-KCM, respectively) or in maximum width (47.2 6.9 p.m and 46.5 4.7 p.m, P = 0.69). Thus it is highly likely that rotational motion of the immobilized cell occurs with the same coefficient of resistance in both media. Typical recordings of gravity-dependent orientation of immobilized paramecia in the hypo- and hyper-density me- dia are shown in Figure 2a and b. In KCM (p < p,.), paramecia oriented upwards during sinking due to gravity, whereas in P-KCM (p > p,) they oriented downwards during floating up. As shown in Figure 2c. plots of orien- tation rates (d6/dt) against orientation angle (0) fit well to the sinusoidal function of Equation 1 1 . Values for |3 ob- tained by least-square regression were positive in the con- trol hypo-density medium and negative in the hyper-density medium (Table 1 ). Negative values of (3 in the hyper- density medium indicate that the drag-gravity model is the major mechanism of mechanical gravitactic orientation in Paramecium. Sea urchin lamie change the mechanical mechanism of gravitactic orientation during development When sea urchin larvae were treated with KCN, their cilia ceased beating and stood nearly perpendicular to the larval surface. The outer morphology of the larvae was observed to be well preserved in P-ASW as well as in ASW: for gastrulae, axial length was 151 7.6 p_m (n =~- 16) and 145 6.1 jum (n = 13), P = 0.19, in ASW and P-ASW, 30 Y. MOGAMI ET AL. a T3 0.15 0.10 0.05 -0.05 -0.10 -0.15 9 (rad) Figure 2. Typical examples of gravity-dependent orientation of Ni 2 + immobilized Paramecium caudutiim. (a, h) Sequential images of gravity- dependent orientation of a cell in KCM (a) and of another in P-K.CM (h), in which recorded images are superimposed at l-s intervals and the time sequence of the motion is illustrated by cyclic change in tone (dark medium > light). In each figure the anterior end of the cell is located to the right, and the gravity vector is towards the bottom of the figure. Scale bar. 0.1 mm. (c) Orientation rates (iltt/ilo as a function of the inclination angle (0). Data from the cells shown in a (KCM) and b (P-KCM) are plotted with open and closed circles, respectively. Sinusoidal curves were obtained by the least-squares fitting to Equation 1 1. respectively, and the maximum width was 135 3.7 JLUTI and 132 5.7 p,m, P = 0.06; for plutei, axial length was 235 19 ju.m (H = 26) and 240 13 /am <;i = 18), P = 0.29, in ASW and P-ASW, respectively, and the maximum width was 175 13 jam and 175 12 juni, P = 0.98. This may justify the common basis for drag coefficients in rota- tion in the different density media, as in Parumeciiim. The gravity-dependent orientation of immobilized larvae is shown in Figure 3a to d, which demonstrates the clear difference between gastrula and pluteus. In ASW (p < p,). both gastrula and pluteus oriented upwards while sinking; in hyper-density P-ASW, however, gastrula oriented down- wards but pluteus upwards while floating up. As shown in Figure 3e and f, the orientation rate appears to be a sinu- soidal function of the orientation angle; although data from larvae fitted less closely to Equation 1 1 than did those from Paramecium, this was probably due to the uncertainty in measuring the orientation angle of the larvae. We some- times observed that larvae rotated slowly around the fore-aft axis during sedimentation. This slow axial rotation made it difficult to determine the fore-aft axis of the larvae. As shown in Table 1. values of |3 obtained from gastrula larvae were positive in the control medium and negative in the hyper-density medium. Thus, in gastrulae as in Para- mecium, the drag-gravity model is the major mechanism of passive gravitactic orientation. However, pluteus larvae have positive values of |3 both in the control and in the hyper-density medium (Table 1). The relatively weak de- pendency of |3 of plutei on the density of the external medium indicates that the gravity-buoyancy model is the major mechanism of passive gravitactic orientation in these larvae. These results indicate that sea urchin larvae change the mechanical mechanism of gravitactic orientation during development. Discussion Estimation of the contribution of the mechanical models in the gravitactic orientation The Reynolds number of rotational motion (Re,) of the microorganisms is defined as Re, = / : o>p (12) where / is a characteristic body length and w is the angular velocity of rotation (Happel and Brenner. 1973). From the maximum velocity of rotation (cu. 0.2 rad s~'. Table 1), Re, of Paramecium or sea urchin larvae is calculated to be about 2 X 10~\ which is sufficiently smaller than unity. This means that the linear assumption of Equation 7 (see the Theory section) is valid to formulate the rotational motion of these microorganisms. The orientation torque generated as a result of the com- bination of the torque originating from different mechanical sources causes the passive orientation of the immobilized organisms. It is difficult to formulate the combination, be- cause we know little about the density distribution within an organism and its geometrical asymmetry. The simplest as- sumption for the combination of the rotational torque is that G. B. and H are located on the geometrical fore-aft axis of the organisms. This gives a sinusoidal function as a linear summation of the sinusoidal equations, each of which is deduced from the gravity-buoyancy and drag-gravity model, respectively. As a result, the orientation rate is given as MECHANICAL BIAS OF MICROBIAL GRAVITAXIS Table 1 Orientation rate t{5), in rad ' s . measured in different densitv media 31 Normal medium Percoll-containing medium Organism Mean SD Range n Mean SD Range Paramecium 0.090 0.033 0.043 - 0. 183 23 -0.104 0.058 -0.257 - -0.041 14 Sea urchin larvae Gastrula 0.140 0.032 0.107 - 0. 197 8 -0.120 0.020 -0.150- -0.090 7 Pluteus 0.157 0.03 1 0.105 - 0. 190 9 0.1 10 0.013 0.097 - 0.137 7 sintf. 13) This simple linear assumption seems to be supported by the fact that a in Equation 1 1 was calculated on average as nearly zero (0.00 0.26 rad (n = 37) for Paramecium, 0.03 0.18 (;i = 15) for gastrula, and 0.06 0.21 (// = 16) for pluteus). Therefore, it is likely that the morpholog- ically defined fore-aft axis almost coincides with the me- chanically defined axis. According to the assumption above, |3 5 obtained in the different density media are given by Vp ig L V(p, - p N )gL H ~ ~~^ ~ _ PP V(p,-p P )gL H (15) where /3 ;V is the maximum orientation velocity measured in the normal density (p N ) medium (KCM or ASW) of the viscosity of TJ^, and fB P is that measured in the hyper- density (p p ) medium (P-KCM or P-ASW) of the viscosity of T] P . Equations 14 and 15 give L H . the distance from B to H, as PP - Ps Vg ' and. thus, f3 R and j8 v are given by: = /3. v - For Paramecium, p v = 1.00, p p = 1.06 and p, = 1.03 g cm" 3 (Ooya et ai, 1992), and T) P lr\ N = 1.53. For sea urchin larvae, p N = 1.01. p p = 1.04, and p, = 1.03 and 1.03 g cm~ 3 , for gastrula and pluteus, respectively (values were obtained by sedimentation equilibrium experiments: data not shown), and TJ/./TJ^. = 1.07. Using these values and /3 V and P P in Table 1. Equations 17 and 18 can be used to obtain values for the contribution of the two mechanisms to negative gravitaxis in normal-density medium. The up- ward orientation of Paramecium in KCM, corresponding to f3 N = 0.09 rad s~ ', is the result of an upward drag-gravity component ( J3 R = 0. 1 2 rad s ' ) combined with a smaller downward gravity-buoyancy component (/3 V = -0.03 rad s '). The situation is similar for sea urchin gastrulae. The upward orientation with |3 A , = 0.14 rad s~' results from an upward drag-gravity component ({$ R = 0.18 rad s" 1 ) combined with a small downward gravity-buoyancy component (j8 v = -0.04 rad s~'). However, the upward orientation of pluteus larvae with fi N = 0.16 rad s" 1 reflects a very different situation. The gravity-buoyancy component has reversed direction from downward to up- ward, and has increased to j8 v - = 0.13 rad s~ ' . The upward drag-gravity component has diminished greatly, to f3 K = ( 14) 0.03 rad s , so that it now makes only a small contribu- tion to the upward orientation. The mechanical property of 'Paramecium There have been several investigations on the mechanical basis of the passive upward orientation of Paramecium. Most of them favored the gravity-buoyancy model as a major mechanism of gravitactic orientation. Fukui and Asai (1980) reported that Triton-treated immobilized cells ori- ented mostly upwards at the sedimentation equilibrium in sucrose density gradient. This upward orientation was evi- dent in well-fed cells but not in starved cells. The upward- orienting posture was found under centrifugal forces in Ni 2 + -immobilized cells in the isodensity medium (Taneda et ai, 1987) and also in the cells swimming at isopycnic level in the density gradient with Ficoll or Percoll (Kuroda and Kamiya. 1989). It was also reported that upward orien- tation was induced by centrifugal force effectively in the cells at the early culture phase but not in those at the late phase, which showed little or no gravitaxis. These results appear to conform with the conclusion that the upward orientation of Paramecium is strongly biased by the torque resulting from the higher density of the posterior part of the organism: the increased density is mainly due to the accu- mulation of food vacuoles (Fukui and Asai, 1985). It should be noted, however, that the results of the sedi- mentation equilibrium experiments were ascribed only to the function of the gravity-buoyancy model and not to the contribution of the drag-gravity model, since F H = with buoyancy artificially balanced with gravity. Furthermore, it (16) 17) 32 Y. MOGAMI ET AL. 1/5 TO 0.20 0.15 010 0.05 o -0.05 -0.10 -0.15 -0.20 0.25 0.20 0.15 0.10 0.05 -0.05 n/2 6 (rad) O n/2 6 (rad) Figure 3. Typical examples of gravity-dependent orientation of KCN- immobilized sea urchin (Hemicentrotus pulcherrimus) larvae, (a-d) Se- quential images of gravity-dependent orientation of the single different larvae at the gastrula (a and b) and the pluteus (c and d) stages. Movements of a larva in ASW (a and c) and of another in P-ASW (b and d) are shown at 3-s intervals in the same way as in Fig. 2a and b. In each figure the animal pole of the larva (leading end in forward swimming) is located to the right, and the gravity vector is towards the bottom of the figure. Scale bar. 0.1 mm (e. f) orientation rates (ilti/dt) as a function of the inclination angle (D). measured from gastrula (e) and pluteus (f). In e. data from the gastrulea shown in a (ASW) and b (P-ASW) are plotted with open and closed circles, respectively. In f, data from the plutei shown in c (ASW) and d (P-ASW) are plotted with open and closed circles, respectively. Sinusoidal curves were obtained by the least-squares fitting to Equation 1 1. seems likely that the gravity-buoyancy component of the orientation torque might be enhanced in these experiments. Since the center of gravity would shift in relation to the content and the distribution of organelles such as food vacuoles, it is probable that in the sedimentation equilib- rium experiments, the intracellular distribution of the or- ganelle was reorganized by gravity during long-lasting sed- imentation of Triton-permeabilized cells through the sucrose density gradient (Fukui and Asai, 1980), or by a large centrifugal acceleration ( 100 X g, Taneda et al., 1987; 300-400 x g, Kuroda and Kamiya, 1989). This may result in accumulation of organelles in the rear part of the cell, and may cause upward orientation, even if the cells originally have a slightly top-heavy organelle distribution that gives a negative j3 v / as estimated above. These facts suggest that the results of previous experiments are still equivocal for the contribution of the drag-gravity model in the gravitactic orientation of Parameciiini. The evidence presented in the Results, on the contrary, indicate that the drag-gravity model makes a major contri- bution to generating a torque for the gravitactic orientation. Although the possibility of a minimal contribution cannot be ruled out, it is clear that the gravity-buoyancy model cannot solely explain the alteration of the sign of the rota- tional torque in the hyper-density medium. In addition, paramecia were observed in P-KCM to swim mostly down- wards (data not shown). Swimming cells changed the net direction of their helical swimming trajectory gradually downwards and accumulated at the bottom of the chamber against the strong floating bias. Positive gravitaxis of Par- ameciitm in the hyper-density medium can be explained by the drag-gravity model, not by the gravity-buoyancy model. Developmental clmnges in the mechanical property in sea urchin lan'ae In the present paper we demonstrated a change in the mechanical basis for gravitactic orientation during the de- velopment of sea urchin larvae: from the drag-gravity model in gastrulae to the gravity-buoyancy model in plutei. Gas- trulae have a thicker posterior part, similar to that of Par- ciiiieciiiin. which is required for the drag-gravity model to function. Plutei. on the other hand, have a thicker anterior part. Therefore they may orient the rear end upwards if the rotational torque is generated according to the drag-gravity model. This was not the case for plutei. Regardless of the remarkable fore-aft asymmetry in morphology, plutei obeyed the gravity-buoyancy model. Gravitactic orientation by different mechanisms was also revealed in the gravitactic swimming behavior of the larvae in P-ASW. In spite of the strong floating bias, gastrulae swam preferentially down- wards (positive gravitaxis) and accumulated at the bottom of the chamber, whereas plutei swam upwards (negative gravitaxis) and accumulated at the top of the chamber (data not shown). Mogami et al. ( 1 988) found that sea urchin larvae change their gravitactic behavior during development. Larvae at the blastula stage to the early gastrula stage swim preferentially MECHANICAL BIAS OF MICROBIAL GRAVITAXIS 33 upwards. This may be explained by a major upward drag- gravity component of orientation torque. The negative gravitatic behavior becomes less remarkable in prism lar- vae: they tend to swim in random directions independent of the gravity vector. This transient disappearance of gravi- taxis may correspond to the alteration of the orientation mechanism revealed in the present paper. At the pluteus stage, larvae again show negative gravitaxis as they acquire the orientation mechanism with a major upward gravity- buoyancy component. A strong separation between the cen- ters of gravity and buoyancy may develop in association with the growth of skeletal structures. Rudiments of spicules initiated in the early gastrula fully extend to give rise to the specific shape of the pluteus larva. The spicule is made of magnesian calcite with a density about three times higher than the average density (Okazaki and Inoue, 1976). As spicules grow, they may change the density distribution to shift the center of gravity toward the rear of the cell. If plutei hereafter maintained the rear-end-heavy mass distribution, they could maintain negative gravitactic behavior irrespec- tive of pronounced morphological changes during the late larval stages. Although the functional role of the drag-gravity model has been accepted in theory, it was not experimentally demonstrated in the orientation movement of organisms. In the present paper we present the first evidence that external geometry is actually important to the gravitactic behavior of aquatic microorganisms. The morphology-dependent inter- action of the organisms with the external fluid seems to be more complicated than hypothesized in the Theory section of this paper. The slow axial rotation observed in sediment- ing sea urchin larvae indicates a hydrodynamic coupling between translational and rotational motion (Happel and Brenner, 1973). Therefore, it is probable that the hydrody- namic coupling secondarily functions to drift the swimming direction upwards, as argued in previous researches (Winet and Jahn. 1974; Nowakowska and Grebecki, 1977). In conclusion, the present study on the mechanical prop- erties of gravitactic orientation in the gravity field demon- strates a relation between the morphology of microorgan- isms and their gravitactic behavior. This relationship might be instructive in researching cases of microbial gravitaxis whose mechanism is still disputed. Acknowledgments This study was carried out as a part of "Ground Research Announcement for Space Utilization" promoted by Japan Space Forum. Literature Cited Bean, B. 1984. Microhial geotaxis. Pp. 163-198 in Membrane and Sensory Transduction, G. Colombetti and F. Lenci, eds. Plenum Press, New York. Chia. F-S.. J. Buckland-Nicks, and C. M. Young. 1983. Locomotion of marine invertebrate larvae: a review. Can. J. Zool. 62: 1205-1222. Degawa, M., Y. Mogami, and S. A. Baba. 1986. Developmental changes in Ca"^ sensitivity of sea-urchin embryo cilia. Comp. Bio- chem. Physiol. 82A: 83-90. Fenchel, T., and B. Finlay. 1984. Geotaxis in the ciliated protozoan Loxodes. J. Exp. Biol. 110: 17-33. Fenchel, T., and B. Finlay. 1986. The structure and function of Muller vesicles in loxodid ciliates. J. Protozool. 33: 68-76. Fukui, K., and H. Asai. 1980. The most probable mechanism of the negative geotaxis of Paramecium caudatum. Pmc. Jpn. AcuJ. 56(B): 172-177. Fukui, K., and H. Asai. 1985. Negative geotactic behavior of Parame- ciiim caudatum is completely described by the mechanism of buoyan- cy-oriented upward swimming. Bioph\s. J. 47: 479-482. Happel, J., and H. Brenner. 1973. Low Reynolds Number Hydrody- namics. Noordhoff International Publishing, Leyden. Kuroda, K., and N. Kamiya. 1989. Propulsive force of Paramecium as revealed by the video centrifuge microscope. Exp. Cell Res. 184: 268-272. Machemer, H., and R. Braucker. 1992. Gravireception and gravire- sponses in ciliates. Acta Protozool. 31: 185-214. Machemer, H., S. Machemer-Riinisch, R. Braucker, and K. Takahashi. 1991. Gravikinesis in Paramecium: theory and isolation of a physi- ological response to the natural gravity vector. J. Comp. Physiol. A 168: 1-12. Machemer-Rb'hnisch, S., U. Nagel, and H. Machemer. 1999. A grav- ity-induced regulation of swimming speed in Euglena graci/is. J. Comp. Physiol. 185: 517-522. Mogami, Y., C. Oobayashi, T. Yamaguchi, Y. Ogiso, and S. A. Baba. 1988. Negative geotaxis in sea urchin larvae: a possible role of mechanoreception in the late stages of development. J. Exp. Biol. 137: 141-156. Nowakowska, G., and A. Grehecki. 1977. On the mechanism of orien- tation of Paramecium caudatum in the gravity field. II. Contributions to a hydrodynamic model of geotaxis. Acta Protozool. 16: 359-370. Okazaki, K., and S. Inoue. 1976. Crystal property of the larval sea urchin spicule. De\: Growth Differ. 18: 413-434. Ooya, M., Y. Mogami, A. Izumi-Kurotani, and S. A. Baba. 1992. Gravity-induced changes in propulsion of Paramecium caudatum: a possible role of gravireception in protozoan behaviour. J. Exp. Biol. 163: 153-167. Roberts, A. M. 1970. Geotaxis in motile micro-organisms. J. E.xp. Biol. 53: 687-699. Taneda, K., S. Miyata, and A. Shiota. 1987. Geotactic behavior in Paramecium caudatum II. Geotaxis assay in a population of the spec- imens. Zool. Sci. 4: 789-795. Verworn, M. 1889. Psychophysiologische Protistenstudien. Gustav Fischer, Jena. (Cited by Machemer and Braucker. 1992.) V'ogel. S. 1994. Life in Moving Fluids: The Physical Biology of Flow. 2nd ed. Princeton University Press, Princeton. NJ. Winet, H., and T. L. Jahn. 1974. Geotaxis in protozoa. I. A propulsion gravity model for Tetrahymena (Ciliata). J. Theor. Biol. 46: 449-465. Reference: Bio/. Bull. 201: 34-44. (August 2001) Synthesis of Several Light-Harvesting Complex I Polypeptides Is Blocked by Cycloheximide in Symbiotic Chloroplasts in the Sea Slug, Elysia chlorotica (Gould): A Case for Horizontal Gene Transfer Between Alga and Animal? JEFFREY J. HANTEN 1 ' 2 AND SIDNEY K. PIERCE 2 * 1 Department of Biology. University of Man-land, College Park, Maryland 20742; and 2 Department of Biology, University of South Florida. Tampa, Florida 3362G Abstract. The chloroplast symbiosis between the asco- glossan (=Sacoglossa) sea slug Elysia chlorotica and plas- tids from the chromophytic alga Vaitcheria litorea is the longest-lived relationship of its kind known, lasting up to 9 months. During this time, the plastids continue to photosyn- thesize in the absence of the algal nucleus at rates sufficient to meet the nutritional needs of the slugs. We have previ- ously demonstrated that the synthesis of photosynthetic proteins occurs while the plastids reside within the diver- ticular cells of the slug. Here, we have identified several of these synthesized proteins as belonging to the nuclear- encoded family of polypeptides known as light-harvesting complex I (LHCI). The synthesis of LHCI is blocked by the cytosolic ribosomal inhibitor cycloheximide and proceeds in the presence of chloramphenicol, a plastid ribosome inhibitor, indicating that the gene encoding LHCI resides in the nuclear DNA of the slug. These results suggest that a horizontal transfer of the LHCI gene from the alga to the slug has taken place. Introduction Most alga-animal symbioses are extracellular associa- tions between two genetically distinct organisms. The alga is usually located extracellularly or enclosed within vacu- Received 22 September 2000; accepted 19 April 2001. * To whom correspondence should be addressed. E-mail: pierce@chumal.cas.usf.edu Abbreviations: CAP, chloramphenicol: CHX. cycloheximide; FCPC, fucoxanthin chlorophyll ale binding proteins; LHC. light-harvesting com- plex; PSI; photosystem I. oles inside the animal's cells. Rarer, but not uncommon, are intracellular symbioses occurring with intact algal chloro- plasts that are captured by specialized cells within the animal. In particular, several species of ascoglossan ( = Sacoglossa) (Opistobranchia) sea slugs capture intact, functional plastids from their algal food source and retain them within specialized cells lining the mollusc's digestive diverticula. This phenomenon has been termed chloroplast symbiosis (Taylor. 1970) or kleptoplasty (Clark et al.. 1990). The sequestered plastids continue to photosynthesize for periods ranging from a few days to a few months, depending on the species (Greene, 1970: Hinde and Smith, 1974; Graves et al., 1979; Clark et al., 1990). The longest such association, lasting as long as 9 months, is found in Elysia chlorotica (Gould), which obtains sym- biotic plastids from the chromophytic alga Vaucheria lito- rea (C. Agardh) (West, 1979; Pierce et al.. 1996). The association begins at metamorphosis of the slug from plank- tonic veliger to juvenile. In laboratory cultures, filaments of V. litorea must be present for metamorphosis to take place (West et al.. 1984). Veligers home in, attach to the fila- ments, and metamorphose into juvenile slugs over the next 24 h. The juveniles eat the algal filaments and sequester the chloroplasts within one of at least two morphologically distinct types of epithelial cells lining the walls of the digestive diverticula (West et al., 1984). Once the plastids are sequestered, the slugs can sustain photosynthesis at rates sufficient to satisfy the nutritional needs for the complete life cycle of the slug, when provided with direct light and carbon dioxide (Mujer et al., 1996; Pierce et al., 1996). 34 SYMBIOTIC PLASTID GENES IN SLUGS 35 Even in nature the slugs obtain most of their energy from photosynthesis (West. 1979). The longevity of this relationship in E. chlorotica makes it especially interesting. Photosynthesis requires the contin- uous synthesis of a variety of chloroplast proteins because many of them, including those used in light harvesting, are rapidly degraded and must be replaced (Greenberg et cil.. 1989; Mattoo et ai, 1989; Barber and Andersson, 1992; Wollman et ai. 1999). Furthermore, photosynthesis re- quires the interaction of as many as 1000 proteins, only about 13% of which are coded in the plastid genome (Mar- tin and Herrmann, 1998). In the plant cell, substantial nu- clear input is required to sustain photosynthetic function, in the form of direct coding of the proteins as well as providing the means for their intracellular transport and regulation (Berry-Lowe and Schmidt. 1991; Wollman et ui, 1999). Considering the level of nuclear and extra-plastid input required, it is not surprising that the longevity of the plastids in most kleptoplastic slugs is relatively short. However, several photosynthetic proteins are synthesized in the se- questered plastids of E. chlorotica (Pierce et ai, 1996), including the large subunit of RuBisCO, Dl, D2, CP43, cyt /and others (Pierce et ai, 1996; Mujer et ai, 1996; Green et ai, 2000). Although all of the synthesized plastid proteins identified to date are plastid encoded (Mujer et ai, 1996; Pierce et ai, 1996; Green et ai, 2000). two groups of synthesized plastid proteins can be distinguished pharma- cologically: those inhibited by cycloheximide (CHX). an SOS cytosolic ribosome inhibitor (Obrig et ai, 1971). and those inhibited by chloramphenicol (CAP), which inhibits protein synthesis on 70S plastid and mitochondrial ribo- somes (Lamb et ai, 1968: Stone and Wilke. 1975). Because the inhibition by CHX suggests that the genes for several plastid proteins must reside in the nuclear DNA, we have done some experiments to identify these proteins and test that possibility. Our present study reports the iden- tification of several of the CHX-blocked proteins as mem- bers of the light-harvesting complex 1 (LHCI). a family of pigment-binding proteins responsible for collecting radia- tion energy from sunlight and transferring it to photosystem I (PSI). LHCI proteins are encoded by the Lhca genes in the nuclear genome of all the plants and algae whose genomes have been examined (Jansson, 1994. 1999; Green and Durn- ford, 1996; Durnford et ai, 1999; Wollman et ai, 1999). This result suggests that the LHCI genes have been some- how transferred from the algal nucleus to the slug's DNA. Materials and Methods Animals and alga Specimens of Elysici chlorotica were collected in both the spring and fall from an intertidal marsh near Menemsha Pond on the island of Martha's Vineyard, Massachusetts. The slugs were maintained in 10-gallon aquaria at 10 C in aerated, artificial seawater (ASW: Instant Ocean. 925-1000 mosm) on a 16/8-h light/dark cycle (GE cool-white fluores- cent tubes. 15 W). Sterile cultures of Vaucheria litorea were maintained in enriched ASW (400 mosm) [modified from the F/2 medium (Bidwell and Spotte. 1985)]. The alga was grown at 20 C on a 16/8-h light/dark cycle (GE cool-white fluorescent tubes; 40 W). and the medium was changed weekly. Inhibitor treatments and plastid protein labeling All reagents used were molecular bio-grade (DNase-, RNase-. and protease-free) purchased from Sigma unless otherwise noted. Effective concentrations of CHX and CAP were determined empirically with initial dose-response curves (Pierce et ai, 1996). CHX (2 mg ml" 1 ) was used to inhibit protein synthesis on SOS cytosolic ribosomes; CAP (160 /o,g ml , stock concentration 50 mg ml" 1 in absolute ethanol) was used to inhibit translation on 70S plastid and mitochondrial ribosomes. Two to four slugs, total wet weight about 1.25 g, were placed into glass scintillation vials containing ASW (1000 mosm) and the appropriate inhibitor, and incubated under intense light (150 W, GE Cool Beam incandescent indoor flood lamp) at 20 C in a gently agitating water bath. After 1 h. 20 /iCi ml" 1 [ 35 S]- methionine (0.7 MBq ml" 1 , trans-[ 35 S]-methionine, ICN) was added, and the slugs were incubated for an additional 6 h. previously demonstrated to provide ample time to incorporate radioactive label into the plastid proteins (Pierce et ai, 1996). Additional slugs were incubated in 0.025% ethanol/ASW (v/v) solution plus [ 35 S]-methionine to serve as a control for the carrier in CAP treatments. Chloroplast isolation and protein separation Chloroplasts were isolated from slugs by using a centrif- ugation protocol. The slugs were homogenized in the pres- ence of the mucolytic agent N-acetyl-cysteine (500 mM). and the homogenate was filtered successively through cheesecloth, Miracloth (Calbiochem), and then nylon mesh (60 ;um to 10 jitm) to remove large debris and the copious amount of mucus the animals produce. The plastids were purified on a pre-formed. 25% Percoll (v/v) gradient, which provides a very pure fraction containing large numbers of intact plastids (Pierce et ai, 1996). In this experiment, the lowest green band containing labeled plastids was isolated from the gradient by using a flamed Pasteur pipette, and residual Percoll was removed by centrifugation. The puri- fied chloroplast pellets were resuspended, lysed by freeze- thawing. and stored at 20C until use. The incorporation of radioactive label was determined by a liquid scintillation counter (Beckman LS60001C). and the protein content was determined using the modified Lowry assay (Peterson, 1977). The resulting specific activity was calculated as counts per minute (cpm) (jug protein)^'. Chlorophyll con- 36 J. J. HANTEN AND S. K. PIERCE tent was determined by extracting the pigment in 80% acetone, then measuring the extract absorbance spectropho- tometrically at 652 nm. The results were calculated as micrograms per microliter according to standard equations (Joyard et al., 1987). Sodium dodecyl sulfate-polyacrylamide gel electro- phoresis (SDS-PAGE) autoradiography was used to assess the effects of CHX and CAP on the pattern of protein synthesis. The plastid lysates obtained from the above pro- cedure were boiled for 2 min in Tris-HCl (pH 6.8)-10% SDS (w/v) buffer containing 5% /3-mercaptoethanol (|3- ME) (v/v). The solubilized proteins were loaded in equal amounts onto 15% SDS-polyacrylamide gels and separated by electrophoresis (Laemlli, 1970). The gels were stained with Coomassie brilliant blue, dried, and exposed to film (Kodak Biomax MR) for 2 to 30 days at -80C, depending on the level of radioactive label present. Approximate mo- lecular masses of the proteins were determined by compar- ison to the migration distances of known molecular weight standards (BioRad, broad-range kaleidoscope) run in adja- cent lanes on each gel. Immunoblot identification of plastid proteins After the plastid isolation and protein separation via SDS-PAGE as described above, the proteins were electro- phoretically transferred (30 V, 4 C, overnight) to PVDF membranes (Immobilon-P; Millipore) (Towbin et til.. 1979). As additional controls, V. litorea chloroplasts [iso- lated and purified using a 30% to 75% Percoll step gradient as previously described (Pierce et al., 1996)] and thylakoids from the red alga Porphyridium cmentum (generously do- nated by Professor Elisabeth Gantt, University of Mary- land), were lysed. and the proteins were separated electro- phoretically and transferred to membranes as above. The membranes were blocked with 5% (w/v) dehydrated milk dissolved in Tris-buffered saline (TBS) (Tris-base 50 mM. NaCl 0.9%, pH 7.5) for 1 h at room temperature, washed twice in TBS for 10 min, and treated with primary antibody for 1 h. In this case, the primary antibody was a polyclonal antibody to LHCI which was produced in a rabbit using a 22-kDa, recombinant LHCI polypeptide produced from a clone of the LhcaRI gene of P. cmentum (Grabowski et al., 2000) (also provided by Professor Gantt) ["/?/" indicating it is a rhodophyte gene (Tan et al., 1997a)] as the antigen combined with Freund's adjuvant in a standard immuniza- tion procedure. After binding of the primary antibody, the membranes were washed twice as above and incubated with secondary antibody, anti-rabbit conjugated hydrogen perox- idase, for 1 h. After washing, the bands were visualized with a 4-chloro-l-napthol and hydrogen peroxide reaction ac- cording to manufacturer's instructions. The immunolabeled western blots were exposed to film as described above to identify the coincidence of antibody binding and radioactive incorporation in the presence of each inhibitor. As a control to confirm that the CAP was blocking plastid-directed protein synthesis and that CHX was not, parallel measurements were run to monitor cytochrome / (cyt/ ) synthesis. Earlier experiments conducted on E. clilo- rotica have demonstrated that cyt / is synthesized in the slugs and is encoded in the plastid DNA (Green et al., 2000). Thus, if CHX and CAP are working as expected, their effect on cyt / and any nuclear-encoded proteins should be opposite. Anti-cyt/, raised to P. cmentum cyt/, was also a gift of Professor Gantt. Immunoprecipitations Immunoprecipitations were conducted to confirm the identity of the radioactive immunolabeled bands on the western blots, using a modified version of the protocol previously used to precipitate proteins from isolated E. chlorotica plastids (Pierce et al., 1996). Plastid proteins were solubilized in lysing buffer ( 10 mM Tris-HCl, 10 mM EDTA, 150 mM NaCl, 1 mM PMSF, 1% (v/v) Nonidet P-40, pH 8.0), using equal amounts of chlorophyll per sample, mixed with a small amount of Protein-A Sepharose beads to eliminate nonspecific binding, and incubated on ice with occasional agitation. The beads were removed by cen- trifugation and discarded, the supernatant was saved, and the appropriate antibody was added to the lysate and rotated overnight (4 C). Protein-A beads, swelled in washing buffer (50 mM Tris-HCl, 5 mM EDTA, 150 mM NaCl, 1 mM PMSF, 0.1% (v/v) Nonidet P-40, pH 8.0), were added the following morning and rotated (3 h, room temperature). The antigen-antibody-protein-A Sepharose bead com- plexes were washed several times in washing buffer and removed by centrifugation. In the case of cyt/, the antigen- antibody-protein-A Sepharose bead complexes were resus- pended in 10.0 M urea. 10% SDS (w/v), 5% |3-ME (v/v), pH 12.5, and boiled for 10 min to liberate the cyt/ antigen. The solution was centrifuged, the supernatant was removed, and the beads were discarded. The supernatant proteins were separated by SDS-PAGE as described above, and the gel was autoradiographed. The LHCI antibody-antigen complex could not be broken efficiently with any treatment, which prevented the visual- ization of the labeled LHCI proteins via SDS-PAGE. Al- though this was unexpected, it is not unusual and may have been caused by a number of factors. The presence of several different LHCI polypeptides with varying isoelectric points, ranging between 4.5 and 9.5 (De Martino et al., 2000). makes it very difficult to create optimal reaction conditions for each one. The polyclonal antibody molecules bind to all the LHCI polypeptides as well as to each other, creating a large antigen-antibody complex with a core inaccessible to the chemicals necessary to liberate the antigen. Very few SYMBIOTIC PLASTID GENES IN SLUGS 37 researchers have attempted LHC immunoprecipitations be- cause of the pitfalls involved in precipitating inner-mem- brane proteins (Anderson and Blobel, 1983). Instead, other protocols have been designed using mild detergents to ex- tract intact photosystem holocomplexes from the thyla- koids. followed by protein separation on sucrose density gradients (Fawley and Grossman, 1986; Buchel and Wil- helm, 1993: Wolfe end., 1994; Schmid et id.. 1997). These isolations require large amounts of starting material (Schmid et id.. 1997) that greatly exceed what is available to us in the slugs. So, instead, we used the LHCI antibody to demonstrate that LHCI had incorporated radioactivity. Following the procedure described above, the protein A Sepharose beads were reacted with anti-LHCI and then with a radiolabeled plastid protein extract. The antigen-anti- body-protein-A Sepharose bead complexes were repeatedly washed by centrifugation until the radioactivity in the su- pernatant was reduced to background. The washed antigen- antibody-protein-A Sepharose bead complexes were resus- pended in optifluor (Packard), and radioactivity was determined by a scintillation counter. Controls for nonspe- cific binding to protein-A Sepharose beads were conducted with the same procedure, but without the addition of the LHCI antibody. Counts per minute resulting from nonspe- cific binding were subtracted from experimental values for each inhibitor treatment and controls, and the final data were converted to cpm (jug chlorophyll)" ' (ju,g protein)" 1 . The normalized data were averaged and expressed in terms of percent of control for each inhibitor. Results Plastid protein synthesis and identification The Coomassie-stained SDS-PAGE gels of protein ex- tracts from isolated slug plastids were similar to controls regardless of the inhibitor present, either CHX or CAP, indicating no difference in the protein composition of the plastids after treatment (Fig. 1). However, autoradiograms of SDS-PAGE gels of plastid proteins extracted from slugs incubated in the presence of [ 35 S]-methionine indicate that very different patterns of protein synthesis occur in the slugs between controls and inhibitors as well as between inhibi- tors (Fig. 2). CHX has a profound effect on protein synthe- sis, preventing synthesis of the majority of the protein bands labeled in the absence of inhibitor (Fig. 2, CON), whereas the synthesis of many more labeled bands occurs in the presence of CAP. Furthermore, these protein bands differ from those visualized in the CHX treatments (Fig. 2). Verification of inhibitor effects Cyt/ antibodies reacted with a protein band synthesized in the presence of CHX on western blots at approximately 36 kDa (Fig. 3). Immunoprecipitations using anti-cyt / (kDa) CON CHX CAP 218_l 43.5 33.9 _ 17.4_ 7.6_ Figure 1. Coomassie brilliant blue-stained 15% SDS-PAGE gel of proteins extracted from isolated Elysia chloroplasts. The protein bands visualized are identical regardless of the inhibitor treatment, CHX or CAP (CON refers to control). Approximate molecular weights are indicated to the left. identify a band with a molecular weight corresponding to cyt/, confirming its identity (Fig. 4). Autoradiograms of the same gels show [ 15 S]-methionine incorporation into cyt/ in the presence of CHX. but not in the presence of CAP (Fig. 5). The anti-LHCI we made to Porphyridium cruentum re- combinant LHCI recognized both the recombinant LHCI antigen (Fig. 5 A, lane 1 ) and the LHCI polypeptides from P. cruentum thylakoids (Fig. 5B, lane 2). Six polypeptide bands were identified in P. cruentum, ranging in approxi- mate molecular weights from 19 to 24 kDa (Fig. 5B, lane 2), sizes consistent with those previously described for the LHCI polypeptides in this species (Tan et al, 1995). The antibody bound onto western blots of plastid proteins from Vciucheria litorea and Elysia chlorotica, with or without the CHX and CAP treatments (Fig. 5C, lanes V. lit.. CON, 38 J. J. HANTEN AND S. K. PIERCE (kDa) CON CHX CAP 126- 90. 43.5 Discussion LHCI, a family of plastid polypeptides essential for pho- tosynthesis, is synthesized while Vaucheria litorea chloro- plasts reside within the cells of the digestive diverticula of Elysia chlorotica. In addition, our data indicate the LHCI polypeptides are probably the products of genes located in the host-cell nuclear genome because their synthesis is inhibited by the cytosolic ribosome inhibitor, CHX, but not by the presence of the plastid ribosome inhibitor, CAP. This remarkable result would not be surprising in a plant or algal species since the LHCI polypeptide family's genes, Lhcal- Lhca6, reside in the nuclear DNA of all plants and algae examined to date (Jansson, 1994; Green and Durnford, 33.9_ (kDa) 126_ 90 B 17.4 7.6 Figure 2. Autoradiograph of plastid proteins separated by SDS-PAGE gel run under the same conditions as those depicted in Figure 2. The plastid proteins incorporating [ 35 S]-methionine label differ following treatment with CHX or CAP. The control (CON) represents chloroplast proteins isolated from slugs without inhibitor treatment. Arrows identify the ap- proximate positions of cyt/ (large arrow) and the LHCI (small arrows) proteins. CHX, CAP). As expected, the six polypeptide bands bound by the anti-LHCI in V. litomi and E. chlorotica plastids have a slightly greater size range 18 to 32 kDa than those identified in P. cnicntiiin. These same antibody-la- beled bands from E. chlorotica plastid proteins incorporate radioactive label in the presence of CAP. but incorporation is blocked by the presence of CHX (Fig. 6). The amount of radiolabel precipitated by anti-LHCI from the slug plastid extracts following CHX treatment is only 2% of the control level, indicating a reduction in LHCI synthesis (Fig. 7). In contrast, the LHCI proteins in CAP- treated slugs incorporated [ 35 S]-methionine at 92% of con- trol rates, more than 40-fold higher than the level found in CHX treated animals (Fig. 7). 43. 5 _ 33.9 _ 17.4_ 7.6 _ Figure 3. Immunoblot labeled with antibody to cyt / (A), and its corresponding autoradiograph (B). The slugs were exposed to CHX and the proteins were labeled as described in the methods. Anti-cyt / binds at approximately 36 kDa, coincident with a radiolabeled protein. The arrow indicates the autorudiograph band corresponding to the position of cyt /. SYMBIOTIC PLASTID GENES IN SLUGS 39 (kDa) CONTROL^ CHX CAP 16.8_ CBB Auto CBB Auto CBB Auto Figure 4. Immunoprecipitation of cyt/. Coomassie brilliant blue (CBB)-stained gels of proteins precipitated with anti-cyt / from chloroplast extracts from slugs subjected to no inhibitor (Control), to CHX. or to CAP. and their corresponding autoradiographs (Auto). The arrow indicates the position of cyt/. Large bands above and below cyt/ are the heavy and light chains of the antibody, respectively. The radioactivity corresponding to the antibody bands in control and CHX is probably undissociated cyt /. 1996: Durnford et at., 1999: Jansson. 1999; Wollman et a/.. 1999). However, the synthesis of LHCI directed by an animal's genome indicates that genes have been transferred into the slug DNA. Although surprising, the site of synthesis and the identi- fication of LHCI seem to be without question as long as inhibitor and antibody specificity are not problems. Both CHX and CAP have been used in a wide array of studies, and their sites of action are well established. In fact, they have been used, exactly as we have done here, to establish that the site of synthesis of the "light harvesting chlorophyll protein" (=LHCI) occurs on 80s cytoplasmic ribosomes in Phaeodactyliini tricomutum (Fawley and Grossman, 1986). There are several reasons to conclude that our antibody is specific. We raised the antibody against the red alga LHCI not only because it was available, but also because the chromophytes, the taxonomic group of V. litorea, probably arose through a secondary symbiosis from a red alga (Rieth, 1995; Green and Durnford, 1996; Palmer and Delwiche, 1996; Martin and Herrmann, 1998: Delwiche, 1999). Fur- thermore. Porphyridium cnientum LHCI possesses both sequence homologies and immunological relatedness to the chromophytic light-harvesting proteins (Wolfe et ai, 1994; Rieth, 1995; Tan et al.. 1997b). Thus, a polyclonal antibody raised to a rhodophyte LHCI should have a good chance of specifically recognizing the LHCI polypeptides in V. lito- rea. Our results indicate that the anti-LHCI binds the P. cnientum recombinant LHCI, the antieenic source for the antibody, as well as all six of the native P. cnientum LHCI proteins (Tan et al., 1995; Grabowski et al., 2000) in control immunoblots of extracted thylakoids. The anti-LHCI immu- noblots of E. chlorotica and V. litorea also identified six protein bands with a greater size range than the LHCI proteins identified in P. cnientum. Those bands are consis- tent with the sizes of LHCI polypeptides from many species (Gantt. 1996; Jansson. 1999: Wollman et ai, 1999), and no other bands were labeled by the antibody. Seeing six LHCI proteins is not surprising, because LHCI is typically found in multiple homologs in algae, ranging from two in one species of Xanthophyceae (Buchel and Wilhelm, 1993) to at least six paralogs in some rhodophytes (Tan et al., 1995), and as many as eight in the chromophyte Heterosigma carterae (Durnford and Green, 1994). With few exceptions [such as in Euglena gracilis (Jansson, 1994)], each is en- coded by a separate, nuclear gene belonging to the Lhc super-gene family (Jansson, 1999). Thus, location of the gene aside, the presence of six LHCI proteins in the endo- symbiotic plastids in E. chlorotica is not surprising. It seems clear that each of the bands immunodecorated by anti-LHCI corresponds to a single LHCI polypeptide and not a dimer. LHCI dimers can result from their association with other LHC proteins and their respective photosystems /;; situ, and they do not always readily dissociate under the denaturing conditions of SDS-PAGE (Tan et al., 1995). If LHCI dimers were present here, they should have minimum molecular weights of about 36 kDa, corresponding to dou- 40 J. J. HANTEN AND S. K. PIERCE A B C (kDa) 1 (kDa) 2 (kDa) V. lit CON CHX CAP 33.9 29.0 33.9 17.4 18.2 17.4 Figure 5. Immunoblots testing the antibody raised to Porphyridium inientum LHCI. (A) Anti-LHCI binds the recombinant 22 kDa Llica RI product from P. cruentwn (lane i ). Its appearance as a 28-30 kDa protein in SDS-PAGE and subsequent immunoblots results from the addition of a 33 amino acid N-termina! fusion in the recombinant protein (Grabowski el at., 2000). (B) Anti-LHCI binds LHCI polypeptides extracted from P. cruentum thylakoids (lane 2). (C) Vauclieria litorea (lane V. lit.) and Ely\ia chlorotica plastid proteins have six bands binding the anti-LHCI identical in size to each other. All six proteins are present in the slugs regardless of the inhibitor treatment [lanes CON (control). CHX and CAP]. Molecular weights are indicated to the left of (A). (B). and (C). ble the molecular weight of the smallest immunolabeled band. However, the largest of the six immunolabeled bands present in the gels is about 32 kDa, seemingly too small to be an LHCI dimer. Other dimers might form with a number of photosystem I (PSD proteins due to the close association of LHCI with the PSI subunits that compose the PSI-LHCI holocomplex (Wollman et al.. 1999; Jansson. 1999). This also does not seem to be the case here. Anti-PSI. raised against the cyanobacteria PSI holocomplex (again, courtesy of Profes- sor Gantt), binds a single 10-kDa protein band on western blots of E. chlorotica plastid proteins (data not shown). The combination of this PSI polypeptide with any of the three smaller bands (18-20 kDa) that react with the anti-LHCI could form a dimer with molecular weights comparable to each of the three larger polypeptides (28-32 kDa). How- ever, since anti-PSI and anti-LHCI do not co-label any bands, an LHCI-PSI dimer is unlikely. An additional possibility might be that one of the bands could be another LHC-type protein possessing immunolog- ical similarities to LHCI, such as the fucoxanthin chloro- phyll ale binding proteins (FCPC) found in chromophytes or light-harvesting complex II (LHCII) proteins. In fact, our previous work has demonstrated the presence of FCPC in plastids of both E. chlorotica and V. litorea. However, the size of the FCPC protein identified there does not corre- spond to the weights of the proteins bound by the anti-LHCI used here (Pierce et al.. 1996; Green et til., 2000). Further- more, previous attempts to demonstrate FCPC synthesis with radioactive labels in the slugs have not yielded positive results (Pierce et ai. 1996). The LHCII family of polypeptides is closely related to LHCI, performing similar functions in photosystem II to those performed by LHCI in PSI. The LHC II genes are in the same nuclear-encoded Lhc super-gene family (Jansson, 1999) and share sequence homologies with those genes encoding LHCI (Durnford et al., 1999; Jansson. 1999; Wollman et al., 1999). There is, however, a clear separation in the phylogenies of LHCI and LHCII (Durnford et al.. 1999), indicating some degree of dissimilarity between the two proteins. Nevertheless, the possibility seems to remain that the proteins bound by our antibody could be from LHCII. Of the LHCII components, CP24. CP26, and CP29 con- tain the most sequence similarities to the LHCIs (Green and Durnford. 1996) and have molecular weights. 25-30 kDa (Wollman et al., 1999). that roughly correspond to these of the three largest polypeptides identified in our anti-LHCI immunoblots of E. chlorotica and V. litorea plastid proteins (28-32 kDa), which appear to be slightly larger than most LHC proteins in chromophytes (Green and Durnford, 1996). An LHCII antibody derived from pea (generously donated by Dr. Kenneth Cline, University of Florida) was unreactive in our iinmunoblotting protocol (data not shown). This SYMBIOTIC PLASTID GENES IN SLUGS 41 CAP CHX Figure 6. Immunoblot (IB) of LHCI synthesized in the presence of CAP and 35 [S]-methionine, and its corresponding autoradiograph (CAP). The arrows indicate radiolabeled bands coinciding to LHCI immunola- beled bands shown in (IB). The bands in (CAP) are not labeled in the presence of CHX (CHX). result seems to indicate that the polypeptides are not LHCII, but since the similarity between the green plant and chro- mophyte LHC proteins is relatively low (Green and Durn- ford, 1996; Durnford et al., 1999), we probably cannot completely eliminate the possibility that the anti-LHCI is binding LHCII polypeptides. However, just like LHCI, all of the LHCII genes are nuclear encoded in the plants and algae where they have been found (Jansson, 1994, 1999; Wollman et al.. 1999), and even if we have identified LHCII, the conclusion is still the same: that an algal LHC gene has been transferred to the DNA of the slug. The immunoprecipitations provide additional evidence that the LHCI polypeptides are being synthesized on the cytoplasmic ribosomes in the slug. The high amount of radioactivity precipitated by the antibody in the presence of CAP compared to that precipitated in the presence of CHX demonstrates that the proteins recognized by the anti-LHCI are indeed synthesized in the slugs. Since the amount of radioactivity incorporated varied from slug to slug and from experiment to experiment, we had to normalize the immu- noprecipitation data as percent of control in order to com- pare them. However, in a typical experiment, the values for the amount of radioactive material incorporated into the precipitate in the presence of CAP ranged from 5000 to 25,000 cpm, whereas those in the presence of CHX ran from 150 to 400 cpm, which may give a clearer picture of the level of material bound by the antibody. The results of the pharmacological experiments, the im- munoblots, and the immunoprecipitations. taken together, provide substantial evidence that LHCI is the identity of some of the plastid proteins that are synthesized in the presence of CAP. The inhibition of LHCI synthesis by CHX suggests that the algal Lhca genes have somehow been transferred to the slug. To be certain that a gene transfer has occurred, direct evidence of the gene in the genomic DNA of the slug must be found, and we are pursuing this confirmation. However, in addition to the results presented here, other circumstantial evidence for the transfer of the LHCI genes between alga and slug is available in several characteristics of the asso- ciation. First, although the turnover rate of LHCI in E. chlomtica is unknown, the fact that it is synthesized indi- cates that it is not an unusually robust protein LHCI replacement is necessary for plastid function to proceed. Second, Lhca genes have not been found in the plastid genomes of any organism (Durnford et /., 1999), including other Vaucheria species (Linne von Berg and Kowallik, 1992). Of course, if LHCI were present in the plastid genome, it would be synthesized with CHX present, as is the case with the cyt / controls; but it is not. Third, the V. litorea plastid genome is 119.1 kb (Green et al., 2000), which is similar in size to those of other algae, including V. 125- o 75 -i O c 0) 50- 0) - 25- , i CONTROL CHX CAP Inhibitor-Treatment Figure 7. CHX inhibits synthesis of LHCI. In the presence of CHX. anti-LHCI precipitated only 2% of control radioactivity incorporated into LHCI compared to 92% of control in the presence of CAP. Control rates were defined as 100%. and inhibitor rates were calculated as a mean percent of control (>i = 6). 42 J. J. HANTEN AND S. K. PIERCE sessilis and V. hursata (Linne von Berg and Kowallik, 1988, 1992), hut small relative to those of other plants (Martin and Herrmann. 1998). Even though the plastid genomes of chro- mophytic algae have a greater coding capacity, relative to their size, than other algae because of fewer introns and inverted repeats (Rieth, 1995). they are too small to carry sufficient genetic information to encode all of the enzymes required for photosynthesis and plastid protein targeting. Fourth, transfer of algal DNA remnants or a nucleomorph- type structure during plastid capture seems unlikely. To date, nucleomorphs have been found only in the Crypto- phyta and Chlorarachniophyta (Delwiche. 1999; Zauner et til.. 2000) and have not been identified in any chromophyte (Maier et a!., 1991; Delwiche, 1999). Although DNA of this type would probably be transcribed on nucleomorph SOS ribosomes (Douglas et al, 1991) and blocked by CHX. neither substantial electron microscopy (Kawaguti and Ya- masu, 1965; Graves et til., 1979; Mujer et al., 1996) nor molecular testing (Green et al., 2000) has so far produced evidence for either nucleomorphs or algal nuclear remnants in E. chlorotica. Furthermore, if algal DNA remnants were present somewhere in the slug cells, the likelihood is remote of their containing the correct genes and being present in all of the plastid-containing cells in all of the slugs in the populations year after year. Finally, others have suggested that some of the proteins necessary to maintain photosyn- thesis may be encoded in the mitochondria! genome and are redirected to the chloroplast (Rumpho et al.. 2000). Al- though dual targeting of proteins has been demonstrated in Arabidopsis (Chow et al.. 1997: Menand et al.. 1998), it seems highly unlikely with LHC1. LHCI has never been found associated with mitochondria in any organism; and CAP, which inhibits the mitochondria! ribosomes in addi- tion to those associated with the plastids. would prevent its synthesis anyway. The horizontal transfer of DNA from the endosymbiont to the nucleus of the host cell provides the basis for the theory of the endosymbiotic origin of eukaryotic organelles. This movement of the symbiont's genes to the host enabled the host to incorporate the organelle's function into its own biochemistry and to faithfully replicate it in subsequent generations. The remnants of eubacterial genes in the mi- tochondria! and plastid genomes of modern eukaryotes probably resulted from such events (Martin and Herrmann. 1998). Most of the discussions regarding the evolution of plastids focus on the horizontal gene transfer resulting from the primary endosymbiotic event in which a primitive pro- karyote engulfed a cyanobacteria (Palmer. 1993; Reith. 1995; Palmer and Delwiche, 1996; Martin et al., 1998; Tengs et al., 2000). Other hypotheses propose a secondary endosymbiosis, probably involving a eukaryote that en- gulfed a red or green alga (Gibbs, 1981; Palmer and Del- wiche, 1996; Martin et al.. 1998: Zhang et al., 1999; Del- wiche, 1999; Tengs et al.. 2000), that produced the plastids of the chromophytic algae and their relatives. In many of these cases, the identity of the initial host, symbiont, or both is unknown. In the case of E. chlorotica and V. litorea, the origin of LHCI is known; if the gene has been transferred, the transfer occurred between two multicellular eukaryotes and represents a case of tertiary endosymbiosis. Finally, the mechanism by which such a gene transfer could occur may be found in the viruses that appear in each generation of the slugs at the end of their life cycle. The viruses have several features in common with Retroviridae and seem to be endogenous (Pierce et al.. 1999). Retrovi- ruses are capable of transferring genes between organisms; if they are incorporated in the germ cells, they are trans- ferred to the subsequent generations as Mendelian genes (Scharfman et al.. 1991). Thus, resolving the relationships between the slugs, alga, plastids, and viruses may have profound implications for both cell and evolutionary biol- ogy. Acknowledgments Research support was provided by a National Science Foundation award (IBN-9604679) to SKP. We thank Elisa- beth Gantt and Beatrice Grabowski for their helpful sug- gestions. Literature Cited Anderson, I). J., and G. Blobel. 1983. Immunoprccipitations of proteins from cell-free translations. Methods Enzvmol. 96: I 1 1-120. Barber, J., and B. Andersson. 1992. Too much of a good thing: Light can be bad for photosynthesis. Trends Biochem. Sei. 17: 61-66. Berry-Lowe, S. L., and G. W. Schmidt. 1991. Chioroplast protein transport. Pp. 257-302 in The Molecular Biology of Plastids: Cell Culture mill Somatic Cell Genetics of Plants. Vol. 7A, L. Bogorad and I. K. Vasil. eds. Academic Press. New York. Bidui'll. J. P.. and S. Spotte. 1985. Artificial Seawaters: Formulas and Method*. Jones and Bartlett. Boston. Pp. 305-306. Buchel, C., and C. YVilhelm. 1993. Isolation and characterization of a photosystem I-associated antenna (LHCI I and a photosystem I-core complex from the chlorophyll i-containing alga Pleurochloris ineirin- xcnsis (Xanthophyceae). ./. Photochem. Photohiol. B Bio/. 20: 87-93. Chow, K-S., D. P. Singh. J. M. Roper, and A. G. Smith. 1997. A single precursor protein for ferrochelatase-I from Arabiilopsis is imported in vitro into both chloroplasts and mitochondria. J. Biol. Cliem. 272: 27565-2757 1 . Clark, K. B., K. R. Jensen, and H. M. Strits. 199(1. Survey of func- tional kleploplasty among West Atlantic Ascoglossa ( =Sacoglossa) (Mollusca: Opistobranchia). Veliger 33: 339-345. Delwiche, C'. F. 1999. Tracing the thread of plastid diversity through the tapestry of life. Am. Nat. 154: S164-S177. I)c Martino, A., D. Douady, M. Quinet-S/ely, B. Rousseau, F. Crep- ineau, K. Apt, and L. Caron. 2000. The light-harvesting antenna of brown alga: highly homologous proteins encoded by a multigene family. Eur. J. Biochem. 267: 5540-5549. Douglas, S. E., C. A. Murphy, D. F. Spencer, and M. W. Gray. 1991. Cryptomonad algae are evolutionary chimaeras of two phylogenetically distinct unicellular eukaryotes. Nature 350: 148-151. Durnford. D. G., and B. R. Green. 1994. Characterization of the SYMBIOTIC PLASTID GENES IN SLUGS 43 light-harvesting proteins of the chromophytic alga. Olisthodiscus luteus (Ht'ierosigma carterae). Biochim. Bio/thy*. Ada 1184: 118-126. Durnfiird, I). (;., J. A. Deane, S. Tan, G. I. McFadden, E. Gantt, and B. R. Green. 1999. A phylogenetic assessment of the eukaryotic light-harvesting antenna proteins, with implications for plastid evolu- tion. J. Ml. t'.vol. 48: 59-68. Fawley, M. \V., and A. R. Grossman. 1986. Polypeptides of a light- harvesting complex of the diatom. Phaeodactylum tricornutum are synthesized in the cytoplasm of the cell as precursors. Plant Pli\\iot. 81: 144-155. Gantl. E. 1996. Pigment protein complexes and the concept of the photosynthetic unit: chlorophyll complexes and phycobilisomes. Pho- tosyn. Re.',. 48: 47-53. Gihbs. S. P. 1981. The chloroplasts of some algal groups may have evolved from endosymbiotic eukaryotic algae. Ann. NY Acad. Sci. 361: 193-207. Grabowski, B., S. Tan, F. X. Cunningham, Jr., and E. Gantt. 200(1. Characterization ot the Porphyridium cntentum chl (/-binding LHC by in vitro reconstitution: LlicaRI binds 8 chi-o molecules and propor- tionately more carotenoids than CAB proteins. Photosyn. Res. 63: 85-96. Graves, D. A., M. A. Gibson, and J. S. Bleakney. 1979. The digestive diverticula of Aldcria modesta and Elvsiu chlorotica. Veliger 21: 415-422. Green, B. J., W-Y. Li. J. R. Manhart, T. C. Fox, E. L. Summer, R. A. Kennedy, S. K. Pierce, and M. E. Rumphn. 2000. Mollusc-algal chloroplast endosymbiosis. Photosynthesis, thylakoid protein mainte- nance, and chloroplast gene expression continue for many months in the absence of the algal nucleus. Plant Physiol. 124: 331-342. Green, B. R., and D. G. Durnford. 1996. The chlorophyll carotenoid proteins of oxygenic photosynthesis. Annit. Rev. Plant. Physiol. Plant. Mol. Bini 47: 6S5-7I4. Greenberg, B. M., V. Gaba, O. Cananni, S. Malkin, A. K. MaHoo, and M. Edelman. 1989. Separate photosensitizers mediate degradation of the 32 kDa photosystem II reaction center protein in the visible and UV spectral regions. Proc. Natl. Acad. Sci. USA 86: 6617-6620. Greene, R. \V. 1970. Symbiosis in sacoglossan opistobranchs: functional capacity of symbiotic chloroplasts. Mar. Biol. 7: 138-142. Hinde. R., and I). C. Smith. 1974. "Chloroplast symbiosis" and the extent to which it occurs in Sacoglossa (Gastropoda: mollusca). Biol. J. Li in,. Soc. 6: 349-356. Jansson, S. 1994. The light-harvesting chlorophyll //>-binding proteins. Biochim. Biopliys. Ada 1184: 1-19. Jansson, S. 1999. A guide to the Lhc genes and their relatives in Arahidopsis. Trends Plant Sci. 4: 236-240. Joyard, J., A-J. Dome, and R. Douce. 1987. Use of thermolysin to probe the cytosolic surface of the outer envelope membranes from plastids. Pp. 195-206 in Methods in Enzymo/agy. Vol. 148, L. Packer and R. Douce, eds. Academic Press. San Diego, CA. Kawaguti. S., and T. Vamasu. 1965. Electron microscopy on the sym- biosis between an elysioid gastropod and chloroplasts of a green alga. Biol. J. Okayama Univ. 11: 57-65. I .M mill. I'. K. 1970. Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227: 608-685. Lamb, A. J., G. D. Clark-Walker, and A. VV. Linnane. 1968. The biogenesis of mitochondria: the differentiation of mitochondria! and cytoplasmic protein synthesizing systems in vitro by antibiotics. Bio- chim. Biophys. Ada 161: 415-427. Linne von Berg, K.-H., and K. V. Kovvallik. 1988. Structural organi- zation and evolution of the plastid genome of Vaucheria sessilis (Xan- thophyceae). Biosystems 21: 239-247. Linne von Berg. K.-H., and K. V. Kowallik. 1992. Structural organi- sation of the chloroplast genome of the chromophytic alga Vaucheria hursata. Plant Mol. Biol. 18: 83-95. Maier, U.-G., C. J. B. Hoffman, E. Eschbach. J. Wolters, and G. L. Igloi. 1991. Demonstration of nucleomorph-encoded eukaryotic small subunit ribosomal RNA in cryptomonads. Mol. Gen. Genet. 230: 155-160. Martin, W., and R. G. Herrmann. 1998. Gene transfer from organelles to nucleus: How much, what happens and why? Plant Physiol. 118: 9-17. Martin, VV., B. Stoebe, V. Goremykin. S. Hansmann, M. Hasegawa, and K. V. Ko\\allik. 1998. Gene transfer to the nucleus and the evolution of chloroplasts. Nature 393: 162-165. Mattoo, A. K., J. B. Marder, and M. Edelman. 1989. Dynamics of the photosystem II reactions center. Cell 5: 241-246. Menand, B.. L. Marechal-Drouard, W. Sakamoto, A. Dietrich, and H. Wintz. 1998. A single gene of chloroplast origin codes for mitochon- drial and chloroplastic methionyl-tRNA synthetase in Arabidopsn thaliana. Proc. Null. Acad. Sci. USA 95: 11014-11019. Mujer, C. V., D. L. Andrews, J. R. Manhart, S. K. Pierce, and M. E. Rumpho. 1996. Chloroplast genes are expressed during intracellular symbiotic association of Vaucheria litorea plastids with the sea slug Elysia chlorotica. Proc. Natl. Acad. Sci. USA 93: 12333-12338. Obrig, T. G., W. J. Culp, W. L. McKeehan, and B. Hardesty. 1971. The mechanism by which cycloheximide and related glutarimide anti- biotics inhibit peptide synthesis on reticulocyte ribosomes. J. Biol. diem. 246: 174-1X1. Palmer. J. D. 1993. A genetic rainbow of plastids. Nature 364: 762-763. Palmer, J. D., and C. F. Delwiche. 1996. Second-hand chloroplasts and the case of the disappearing nucleus. Proc. Natl. Acad. Sci. USA 93: 7432-7435. Peterson, G. L. 1977. A modification of the protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem. 83: 346-356. Pierce, S. K., R. W. Biron. and M. E. Rumpho. 1996. Endosymbiotic chloroplasts in molluscan cells contain proteins synthesized after plas- tid capture. J. Exp. Biol. 199: 2323-2330. Pierce, S. K.. T. K. Maugel, M. E. Rumpho, J. J. Hanten, and W. L. Mondy. 1999. Annual viral expression in a sea slug population: life cycle control and symbiotic chloroplast maintenance. Biol. Bull. 197: 1-6. Rieth, M. 1995. Molecular biology of rhodophyte and chromophyte plastids. Annu. Rev. Plan! Physiol. Plant Mol. Biol. 46: 549-575. Rumpho, M. E., E. J. Summer, and J. R. Manhart. 2000. Solar- powered sea slugs. Mollusc/algal chloroplast symbiosis. Plant Physiol. 123: 29-38. Scharfmann, R., J. H. Alexrod, and I. M. Verma. 1991. Long-term in vivo expression of retrovirus-mediated gene transfer in mouse fibro- blast implants. Proc. Natl. Acad. Sci. USA 88: 4626-4630. Schmid, V. H. R., K. V. Cammarata, B. U. Bruns, and G. W. Schmidt. 1997. In vitro reconstitution of the photosystem I light-harvesting complex LHCI-730: Heterodimerization is required for antenna pig- ment organization. Proc. Natl. Acad. Sci. USA 94: 7667-7672. Stone, A. B., and D. Wilkie. 1975. Loss of cytochrome oxidase in Sacchromvces cerevisiae during inhibition of mitochondrial proteins synthesis by erythromycin and chloramphenicol. J. Gen. Microbiol. 91: 150-156. Tan, S., G. R. Wolfe, F. X. Cunningham, Jr., and E. Gantt. 1995. Decrease in the PSI antenna complex with increasing growth irradiance in the red alga Porphyridiwn t rucntitm. Photosyn, Res. 45: 1-10. Tan, S., F. X. Cunningham, Jr., and E. Gantt. 1997a. LhcaRl of the red alga Porph\ridini cnientiim encodes a polypeptide of the LHC I complex with seven potential chlorophyll a-binding residues that are conserved in most LHCs. Plant Mol. Biol. 33: 157-167. 44 J. J. HANTEN AND S. K. PIERCE Tan, S., A. Ducret, R. Aebersold, and E. Gantt. 1997b. Red algal LHC I genes have similarities with both Chi ale- and a/obinding proteins: A 21 kDa polypeptide encoded by LhcaR2 is one of the six LHC I polypeptides. Photosvn. Res. 53: 129-140. Taylor, D. L. 1970. Chloroplasts as symbiotic organelles. Int. Rev. C\tol. 27: 29-64. Tengs. T., O. J. Dahlberg, K. Shalchian-Tabrizi, D. Klaveness, K. Rudi, C. F. Delwiche, and K. S. Jakobson. 2000. Phylogenetic analyses indicate that the 19' hexanoloxy-fucoxanthin-containing dinoflagellates have tertiary plastids of haptophyte origin. Mol. Bio/. Evol. 17: 718-729. Towbin, H., T. Staehelin. and J. Gordon. 1979. Electrophoretic trans- fer of proteins from polyacrylamide gels to nitrocellulose sheets: pro- cedure and some applications. Proc. Nail. Acad. Sci. USA 76: 4350- 4354. West, H. H. 1979. Chloroplast symbiosis and the development of the ascoglossan opistobranch Elysia chlomticii. Ph.D. dissertation. North- eastern University. Boston. MA. West, H. H., J. Harrigan, and S. K. Pierce. 1984. Hybridization of two populations of marine opistobranch with different developmental pat- terns. Veliger 26(3): 199-206. Wolfe, G. R., F. X. Cunningham, D. Durnford, B. R. Green, and E. Gantt. 1994. Evidence for a common origin of chloroplasts with light-harvesting complexes of different pigmentation. Nature 367: 566-568. Wollman, F.-A., L. Minai, and R. Nechushtai. 1999. The biogenesis and assembly of photosynthetic proteins in thylakoid membranes. Bio- cliiin. Bi,i/>li\s. Acta 1411: 21-85. Zauner, S., M. Fraunholz, J. Wastl. S. Penny, M. Beaton, T. Cavalier- Smith, U.-G. Maier, and S. Douglas. 2000. Chloroplast protein and centrosomal genes, a tRNA intron, and odd telomeres in an unusually compact eukaryotic genome, the cryptomonad nucleomorph. Proc. Nail. Acad. Sci. USA 97: 200-205. Zhang, Z., B. R. Green, and T. Cavalier-Smith. 1999. Single gene circles in dinoflagellate Chloroplast genomes. Nature 400: 155- 159. Reference: Bin/. Bull. 201: 45-51. (August 2001) Asexual Reproduction in Pygospio elegans Claparede (Annelida, Polychaeta) in Relation to Parasitism by Lepocreadium setiferoides (Miller and Northup) (Platyhelminthes, Trematoda) DEAN G. McCURDY* Coastal Studies Center, 6775 College Station, Bowdoin College. Brunswick, Maine 04011-8465 Abstract. Life-history theory predicts that parasitized hosts should alter their investment in reproduction in ways that maximize host reproductive success. I examined the timing of asexual reproduction (fragmentation and regeneration) in the polychaete annelid Pygospio elegans experimentally exposed to cercariae of the trematode Lepocreadium setiferoides. Con- sistent with adaptive host response, polychaetes that became infected by metacercariae of trematodes fragmented sooner than unexposed controls. Parasites were not directly associated with fission in that exposed polychaetes that did not become infected also fragmented earlier than controls. For specimens of P. elegans that were not exposed to trematodes, new frag- ments that contained original heads were larger than those that contained original tails, whereas original head and tail frag- ments did not differ in size for infected polychaetes. In infected specimens, metacercariae were equally represented in original head and tail fragments and were more likely to be found in whichever fragment was larger. Despite early reproduction, parasitism was still costly because populations of P. elegans exposed to parasites were smaller than controls when mea- sured 8 weeks later and because exposure to cercariae reduced survivorship of newly divided polychaetes. Taken together, my results suggest that early fragmentation is a host response to minimize costs associated with parasitism. Introduction Hosts respond to parasitism in a number of ways, which include avoidance of parasites in space or time (e.g.. mi- Received 19 October 2000; accepted 10 April 2001. * Current address: Department of Biology, Albion College, Albion, Michigan 49224. E-mail: dmccurdy20,000 m~ 2 ), but Ilyanassa obsoleta and its associated cercarial parasites were rare (<0.25 snails m" 2 ), minimiz- ing the likelihood that polychaetes used in experiments were already infected. I collected polychaete tubes in the mid- intertidal zone by sieving the top 5 cm of mud (500-jum mesh) and transported tubes to the nearby running-seawater laboratory at the Coastal Studies Center of Bowdoin Col- lege for sorting. I retained only undamaged, entire adult polychaetes (>2 mm) that were not about to fragment (detectable because P. elegans constricts just prior to fis- sion; Gibson and Harvey. 2000). To obtain cercarial trematodes for experiments. I col- PARASITISM AND ASEXUAL REPRODUCTION 47 lected specimens of /. obsoleta from throughout the inter- tidal zone at Strawberry Creek. Great Island, Maine (4349'N. 6958'W). This mudflat is located 2.5 km from the Wyer-Orr's mudflat and supports high densities of /. obsoleta (>10 m~ 2 ). In the laboratory. I housed 550 mud snails in separate 9-oz plastic cups with 125 ml of filtered seawater (55 ^im, 31 ppt, 23 C). I retained only large snails (>15 mm. tip of apex to lip of siphonal canal) because previous studies have shown that the prevalence of Lep o- creadium setiferoides increases with shell height of snails (Curtis. 1997; McCurdy el ai. 2000c). After 30 h. I exam- ined each cup for cercariae of L. setiferoides (identified using McDermott, 1951). combined cercarial-infested sea- water from cups of six snails that had shed cercariae, and pipetted 20 ml of the solution into each dish that contained a polychaete that was to be exposed. Unexposed polychaetes each received 20 ml of seawater from six cups that contained snails that did not shed cercariae (confirmed by dissection, as cercarial release is a poor indicator of infection status; Curtis and Hubbard. 1990). Experiments To investigate the impact of parasites on the timing of asexual reproduction, I individually housed 52 adult speci- mens of P. elegans in 150-ml custard dishes filled with unfiltered seawater with or without cercariae (18 C. 16 h light day" 1 ). After 24 h. I transferred each polychaete to a new dish filled with seawater and lined with defaunated mud (prepared by passing mud through a 425-ju.m sieve and heating it to 70 C). Every 24 h, I suspended each dish from a harness and determined the status of each polychaete by observing its tube (or tubes) through the bottom of its dish with the aid of a fiber-optic illuminator and 10X magnifying loupe. Polychaetes could easily be observed because they constructed tubes that opened against the bottoms of their dishes. Polychaetes were fed the pea-flower- based supple- ment Liquifry Marine (Interpet Inc.; Brown el ai, 1999) every 3 days (concentration = 1 drop 1 ~ ' ) following a complete change of water. I removed polychaetes from the experiment when they died or fragmented, and I measured the relaxed length of all fragments with an ocular microme- ter (nearest 0.1 mm; Gudmundsson. 1985). I then dissected each fragment to determine if it was infected by trematode metacercariae and compared median time-to-fragmentation among exposed but uninfected. exposed and infected, and unexposed polychaetes. In making this comparison. I sepa- rated exposed but uninfected polychaetes from unexposed ones because of the possibility that host response might be associated with indirect cues associated with parasitism (i.e., response might not require an actual infection to oc- cur). To compare time-to-fragmentation. I applied a non- parametric Kruskal-Wallis ANOVA because the residuals for all groups were non-normal. I then applied Dunn's method to compare differences among medians (Zar, 1996). To investigate how exposure to parasites affected the reproductive success of P. elegans, I randomly housed 18 sets of 10 polychaetes (hereafter referred to as populations of polychaetes) in separate dishes and exposed half of the sets to cercariae of trematodes (housing conditions for polychaetes were as described above). Because the infection status of polychaetes that died during this experiment could not be determined without disturbing surviving polychaetes, I assessed rates of experimental and background infection by randomly removing two sentinel populations after 3 days: a population of polychaetes that had been exposed to cercariae, and a population of unexposed polychaetes. Rates of infection at that time represented maximum levels that could occur because cercariae of L. setiferoides survive for less than 48 h outside a host (Stunkard, 1972). After 8 weeks. I removed the remaining dishes and processed each population by counting the number of polychaetes retained after sieving (425-jam mesh) and dissecting each polychaete to determine its infection status. To assess survivorship and regenerative ability of newly divided polychaetes in relation to parasitism. I cut 59 polychaetes into two fragments and exposed 30 pairs of fragments to cercariae. Cutting each polychaete resulted in a smooth, clean blastema similar to that resulting from sublethal predation or asexual fragmentation (Gibson and Harvey. 2000; pers. obs.). To mimic conditions in nature, where newly fragmented polychaetes generally remain in the same burrow during regeneration (Gudmundsson, 1985; Gibson and Harvey, 2000). I individually housed original head and tail fragments together in a dish with seawater and mud (housing conditions as described above). To avoid disturbing fragments (as above), I assessed initial rates of infection at 3 days after exposure or non-exposure by re- moving and dissecting randomly chosen sentinel pairs of exposed fragments (/; = : 10 polychaetes) and unexposed fragments (n = 10 polychaetes). At 10 days after exposure or non-exposure. I removed all remaining fragments, mea- sured their lengths, and determined their infection status. Results Parasitism and host fragmentation In the experiment investigating the impact of trematodes on the timing of asexual reproduction in Pygospio elegans, parasite prevalence was low (42.3% of polychaetes exposed became infected; n = 26). Asexual fragmentation always yielded two fragments; one containing the original head and thorax and a second containing the original tail (see Gibson and Harvey, 2000, for a description of body components). In all cases, polychaetes fragmented within 24 h of observable constrictions. Time-to-fragmentation differed between ex- posed and infected, exposed but uninfected. and unexposed 48 D. G. McCURDY 30 a -o F 2-1 S l b -= 1 < c b tr 1 6 , i i s i i Unexposed !x posed/ Exposed/ LtninlcLlcd Infected Treatment Figure 1. Median ( quartiles) numbers of days for asexual reproduc- tion to occur in individuals of Pygospio elegans that were experimentally infected, exposed but not infected, and not exposed to cercariae of the trematode Lepocreadiwn setiferoides. Polychaetes and parasites were col- lected from mudflats in Harpswell, Maine, and housed in the laboratory. Median-, with the same letter do not differ significantly from each other. polychaetes (// |2 . 52) = 10.56. P < 0.01: Fig. 1). Specif- ically, polychaetes that were exposed to cercariae but did not become infected fragmented earlier than unexposed polychaetes (Q = 2.99, P < 0.005), as did polychaetes that were exposed and became infected (Q = 2.16, P 0.05). Of all polychaetes that were exposed to cercariae, however, infection status did not affect time-to-fragmenta- tion (Q = 0.49, NS). For unexposed polychaetes and exposed polychaetes that remained uninfected. fragments that contained original heads were larger than those that contained original tails, whereas lengths of original head and tail fragments did not differ for infected polychaetes (Table 1 ). In infected polychaetes. parasites were just as likely to be found in fragments that contained original heads (n = 5) as those that contained original tails (n 5) (an additional polychaete harbored a metacercaria in each new fragment). For infected polychaetes, infected fragments were signifi- cantly larger than uninfected fragments (infected fragments: x s = 2.0 0.2 mm; uninfected fragments: x s = 1.4 0.2 mm; paired r (l)) = 2.28. P < 0.05). and in 9 of 10 cases, metacercariae were found in the larger fragment (Xf, > = 6.4, P = 0.01). Cercariae were not observed to penetrate segments that comprised, or were adjacent to, planes of fission. Parasitism and host asexual reproductive success At 3 days post exposure, 17 of 20 fragments (8.5 of the original 10 polychaetes) were alive in the sentinel popula- tion that was exposed to cercariae. Only one fragment in this population was infected by trematodes a living tail frag- ment infected with a single metacercaria. In the sentinel population that was not exposed to cercariae, 18 of 20 fragments were alive after 3 days and no parasites were found (one fragment, containing an original head, was lost during processing). At 8 weeks after exposure or non- exposure, I saw no evidence of recent fission in polychaetes as all fragments had complete or nearly complete heads and tails. Therefore. I considered all fragments equally when measuring population sizes at that time. Populations of polychaetes that were exposed to cercariae were smaller than those that were not exposed (exposed populations: A 5 = 17.3 2.4 polychaetes; unexposed populations: x s = 29.8 3.7 polychaetes; / (14) = 2.84, P = 0.01). When dissected, only seven polychaetes in exposed popu- lations were infected (one polychaete in each of three pop- ulations and two polychaetes in each of two populations), and none of the polychaetes in any of the unexposed pop- ulations was infected. Considering sentinel polychaetes that had been cut into two pieces, 2 of 10 polychaetes exposed to cercariae were infected at 3 days post-exposure. In each case, the infection was in the original head fragment and by a single metacer- caria. None of the 10 unexposed polychaetes was infected. When examining the remaining polychaetes 7 days later, I found that both head and tail fragments of exposed polychaetes were less likely to be alive than the respective fragments of unexposed polychaetes (head fragments: = 8.07, P < 0.005; tail fragments: , , = 12.22, P < 0.001 : Fig. 2). Only two exposed polychaetes were infected by metacercariae (one polychaete had an infected tail frag- ment and another an infected head fragment; /; = 20). and no unexposed polychaetes were infected (n == 19). In all cases, regeneration of "lost" components was nearly com- plete by 10 days, and lengths of original head and tail fragments did not differ in relation to exposure (unexposed heads: A SE = 2.65 0.15; exposed heads: x SE = 2.56 0.26; r (26) = 0.32. NS: unexposed tails: x SE = Table 1 Si;cs af fraxiiicnt* produced hy uxe.\iial fission of Pygospio elegans in relation to panisitism Fragment length (mm| Heads Tails Paired / test Unexposed 2.1 0.2 1.57 0.1 /, 2 ,, = 2.7. P = 0.01 Exposed but uninfected 2.4 0.2 1.57 0.2 f,, 4l = 2.3, P = 0.04 Exposed and infected 1.9 0.2 1.71 0.2 ?,,, = 0.8. P = 0.44 Data are means and standard errors for lengths of fragments containing original heads and those containing original tails of polychaetes that were experimentally infected, exposed but not infected, and not exposed to cercariae of the trematode Lepocreadiwn setiferoides. The last column shows results from paired t tests for lengths of original head versm, tail fragments. PARASITISM AND ASEXUAL REPRODUCTION 49 I cC A Heads Tails Original fragments Figure 2. Proportions (95% confidence intervals) of original head and tail fragments of individuals of Pygospio elegans that survived for 10 days in the laboratory following exposure or non-exposure to cercariae of the trematode Lepocreatl/iuii .fciiti'i-niilfs. Sample sizes are shown above the bars. 2.71 0.19: exposed tails: x SE = 2.49 0.28; f (22) = 0.63. NS). Discussion Parasitism and host fragmentation In support of the hypothesis of adaptive host response I found that specimens of Pygospio elegans infected by meta- cercariae of Lepocreadium setiferoides hastened their onset of asexual reproduction relative to unexposed controls. By doing so, polychaetes may be expected to achieve greater reproductive success than if they had failed to respond because of increasing costs associated with parasitism over time (Forbes, 1993). However, my observation that early fragmentation also occurred in exposed polychaetes that remained uninfected complicates this interpretation. In a study that separated hosts by exposure and infection status, Minchella and Loverde ( 1981 ) found that freshwater snails of the species Biomphalaria glahrata increased their rates of early egg laying when infected by Schistosoma mansoni, but that the rates for exposed but uninfected individuals and unexposed controls did not differ. These authors argued that only infected snails responded because successful parasit- ism was associated with a high cost to future reproduction (castration). For individuals of P. elegans exposed to, but not infected by, cercariae. early reproduction could still be an adaptive host response if exposure to cercariae in nature is a reliable indicator that costly infections will soon result (Minchella. 1985). Support for this idea comes from the observation that Ilyanassa obsoleta infected by L. setiferoides, although uncommon across mudflats, can remain for several months in small patches where some P. elegans are found (Mc- Curdy et til., 20()0c). As a result, thousands of cercariae are shed in areas where infections are most likely to occur. Additional information on the infection process of L. setif- crnitles is necessary to determine whether polychaetes de- tect cercariae, and whether the exposure-related response resulted from the presence of cercariae or from failed at- tempts at penetration. There is evidence from other parasite- host systems that invertebrates can detect and exhibit anti- parasite behaviors to minimize the likelihood of infection (e.g.. Leonard et al.. 1999). Early fragmentation of P. elegans is unlikely to be a parasite adaptation, because it apparently does not increase transmission rates for cercariae or metacercariae. Specifi- cally, fragmentation was not associated with increased sus- ceptibility to parasitism: most polychaetes fragmented after free-living cercariae would have (>48 h; Stunkard, 1972). For metacercariae. residing in small fragments would not appear to benefit transmission to final hosts, because floun- der select prey at larger sizes relative to conspecifics, and even small differences in prey size preference can pro- foundly influence the energy budgets of predators foraging on mudflats (MacDonald and Green. 1986: Boates and Smith, 1989; Keats. 1990). To assess whether early frag- mentation is actually adaptive for parasites or hosts, the consequences of early fragmentation could be further ex- plored by constructing a model derived from empirical observations of parasites, their intermediate hosts, and the predators that are their final hosts. This approach was used recently to show that the early onset of receptivity to mating observed in females of the amphipod Corophium volntator infected by the trematode G\naecotyla adunca resulted in greater reproductive success for the amphipods than if they had waited to become receptive at the optimal time for uninfected females (McCurdy et al., 2001). I found no evidence that fragmentation of P. elegans served to isolate or remove metacercariae, in that fission produced only two fragments, the smaller of which almost never contained metacercariae. It is unclear whether the greater presence of metacercariae in larger fragments is adaptive for the parasite or its host or whether larger frag- ments merely represent larger targets for parasites. Meta- cercariae might benefit from residing in larger fragments because of the availability of additional resources for para- site development or the possibility of a greater transmission rate to final hosts (as stated above, flounder tend to select larger prey). If residing in larger fragments is parasite- mediated, the observation that metacercariae develop near the site of initial penetration (Stunkard. 1972; pers. obs) indicates that the mechanism does not involve movements by metacercariae through the host coelom and into larger fragments. Fragmentation could also be interpreted as a host response: If larger fragments are better able to tolerate stresses associated with parasitism, the result would be a net reproductive benefit to hosts. In fact, host response need not 50 D. G. McCURDY be exclusive of benefits to parasites, depending on the timing of altered behavior of infected hosts (McCurdy et ai. 1999). Simulated parasites such as Sephadex beads (Suwan- chaichinda and Paskewitz. 1998) could be used to help separate effects mediated by the parasite from those medi- ated by the host. Experiments with simulated parasites would provide cues to the host that it has become infected while removing the possibility of parasite manipulation. Across all experiments, I found no evidence for onset of sexual reproduction, observing neither eggs nor spermato- phores. Seasonal constraints may have precluded sexual reproduction, which usually occurs only during the winter in P. elegans (Rasmussen, 1953; Gudmundsson, 1985; Wil- son, 1985). However, even if the polychaetes had shown evidence of sexual reproduction, this tactic might be ex- pected to increase reproductive success only if mates were available; an unlikely event given the rarity of parasites in natural populations of P. elegans (above). Parasitism and host asexual reproductive success I found that even a low level of exposure to cercariae (on average, 8% of cercariae that a single snail sheds in 30 h) reduced the asexual reproductive success of P. elegans (45%, measured in populations 8 weeks after exposure). In a related finding from another experiment, both head and tail fragments were less likely to survive to complete regen- eration than were unexposed fragments. Direct effects of parasitism are not sufficient to account for these results given that few exposed polychaetes actually became in- fected in either experiment. One possibility is to explain the reduced reproductive success of exposed but uninfected hosts as the result of a trade-off between host reproductive effort and costly activities associated with defenses against parasites. Recent work has shown that hosts exposed to parasites may trade off energy used in reproduction for behaviors or immune responses to resist parasites (Sheldon and Verhulst, 1996; Leonard et ai, 1999). Regardless of the underlying causes, the dramatic reduc- tion in reproductive success of P. elegans after exposure to cercariae has implications for natural populations of this species and for soft-bottom intertidal communities. Pygos- pio elegans often dominates such communities, and thus can directly affect the distribution and abundance of other infauna (Wilson, 1983: Brey, 1991; Kube and Powilleit, 1997). In addition, it is possible that parasitism of P. elegans may influence the structure of intertidal communities by altering or creating engineering functions in hosts. Engi- neering functions are those that produce new habitat as a result of changes in behaviors or life history associated with parasitism (Thomas et ai, 1999). Clearly, researchers should consider the impacts of parasites on the ecology and evolution of the reproductive strategies of marine inverte- brates and on the structure of infaunal communities. Acknowledgments I thank Glenys Gibson for her suggestions on experimen- tal design and Mark Forbes for our many discussions about life-history theory. Funding was provided by postdoctoral fellowships from the Natural Sciences and Engineering Re- search Council of Canada and the Coastal Studies Center, Bowdoin College. Literature Cited Anger, V. 1984. Reproduction in Pygospio elegans (Spionidae) in rela- tion to its geographical origin and to environmental conditions: a preliminary report. Fortschr. Zoo/. 29: 45-51. Anger, K.. V. Anger, and E. Hagmeier. 1986. Laboratory studies on larval growth of Polydora ligni, Polydora ci/iata, and Pygospio elegans (Polychaeta: Spionidae). Helgol. Meeresunters. 40: 377-395. Boates, J. S., and P. C. Smith. 1989. Crawling behaviour of the am- phipod. Corophium vohitator and foraging by semipalmated sandpip- ers. Culidri.\ pus ilia. Can. J. Zoo/. 67: 457-462. Born, M. A. 1999. Tidelog: Northern New England Edition. Pacific Publishers. Bolinas. CA. Brey, T. 1991. Interactions in soft bottom benthic communities: quanti- tative aspects of behaviour in the surface deposit feeders Pygospio elegans (Polychaeta) and Macoma balthica (Bivalvia). Helgol. Meere- sunters. 45: 301-316. Brown, R. J.. M. Conradi, and M. H. Depledge. 1999. Long-term exposure to 4-nonylphenol affects sexual differentiation and growth of the amphipod Corophium volutator (Pallas, 1766). Sri'. Total Environ. 233: 77- X8. Chia, F. S., G. Gibson, and P. V. Qian. 1996. Poecilogony as a reproductive strategy of marine invertebrates. Oceanol. Acta 19: 203- 208. Curtis, L. A. 1987. Vertical distribution of an estuarine snail altered by a parasite. Science 235: 1509-151 1. Curtis, L. A. 1997. Ilvanassa obsoleta (Gastropoda) as a host for trema- todes in Delaware estuaries. J. Parasitol. 83: 793-803. Curtis, L. A., and K. M. K. Hubhard. 1990. Trematode infections in a gastropod host misrepresented by observing shed cercariae. J. Exp. Mar. Bwl. Ecol. 143: 131-137. Folstad, I., A. C. Nilssen, O. Halvorsen, and J. Andersen. 1991. Par- asite avoidance: the cause of post-calving migrations in Rangifer? Can. J. Z<>/. 69: 2423-2429. Forbes, M. R. L. 1993. Parasitism and host reproductive effort. Oikos 67: 444-450. Gibson, G. I)., and J. M. L. Harvey. 2000. Morphogenesis during asexual reproduction in Pygospio elegans Claparede (Annelida. Polychaeta). Biol. Bull. 199: 41-49. Gudmundsson. H. 1985. Life history patterns of polychaete species of the Family Spionidae. J. Mar. Biol. Assoc. UK 65: 93-1 11. Hobson, K. D.. and R. H. Green. 1968. Asexual and sexual reproduc- tion of Pygospio elegans (Polychaeta) in Barnstable Harbor. Massa- chusetts. Biol. Bull. 135: 410 (abstract). Hurst, L. D., and J. R. Peck. 1996. Recent advances in understanding of the evolution and maintenance of sex. Trends Ecol. Evol. 11: 46-52. Keats, D. VV. 1990. Food of winter flounder Pseudopleuronectes ameri- ciinii.i in a sea urchin dominated community in eastern Newfoundland ( Canada 1. Mar. Ecol. Prog. Ser. 60: 13-22. Kraaijeveld, A. R., and C. J. Godfray. 1997. Trade-off between para- sitoid resistance and larval competitive ability in Drosophi/a melano- t>astcr. Nature 389: 278-280. Kube. J., and M. Powilleit. 1997. Factors controlling the distribution of Maren;.elleria cf. viridis. Pvgospio e/egans and Strehlospio shrubsoli PARASITISM AND ASEXUAL REPRODUCTION 51 (Polychaeta: Spionidae) in the southern Baltic Sea, with special atten- tion for the response to an event of hypoxia. Aquat. Ecol. 31: 1 87-198. Leonard, N., M. R. Forbes, and R. L. Baker. 1999. Effects of Limno- chares americana (Hydrachnidia: Limnochandae) mites on life history traits and grooming behaviour of its damselfly host, Enallagma ebrium (Odonata: Coenagrionidae). Can. J. Zool. 77: 1615-1622. Lively, C. M. 1996. Host-parasite coevolution and sex: Do interactions between biological enemies maintain genetic variation and cross-fer- tilization? Bioscience 46: 107-1 14. Macdonald, J. S., and R. H. Green. 1986. Food resource utilization by live henthic feeding fish in Passamaquoddy Bay, New Brunswick (Canada). Can. J. Fish. Aquat. Sci. 43: 1534-1546. Martin. \V. E. 1938. Studies on trematodes of Woods Hole: the life cycle of Lepocreadium setiferoides (Miller and Northup). Allocreadi- idae. and the description of Ct-rcuria cumingiae N. Sp. Biol. Bull. 75: 463-474. McCurdy. D. G., M. R. Forbes, and J. S. Boates. 1999. Testing alternative hypotheses for variation in amphipod behaviour and life history in relation to parasitism. Int. J. Parasitol. 29: 1001-1009. McCurdy, D. G., M. R. Forbes, and J. S. Boates. 2000a. Male amphi- pods increase their mating effort before behavioural manipulation by trematodes. Can. J. Zool. 78: 606-612. McCurdy, D. G., J. S. Boates, and M. R. Forbes. 2000b. Reproductive synchrony in the intertidal amphipod Corophium volutator. Oikos 88: 301-308. McCurdy, D. G., J. S. Boates. and M. R. Forbes. 2000c. Spatial distribution of the intertidal snail llyanassa obsolete! in relation to parasitism by two species of trematodes. Can. J. Zoo/. 78: 1 1 37-1 143. McCurdy, D. G., J. S. Boates, and M. R. Forbes. 2001. An empirical model of the optimal timing of reproduction for female amphipods infected by trematodes. J. Parasitol. 87: 24-31. McDermott, J. J. 1951. Larval trematode infection in Nassa obsoleta (Say), from New Jersey waters. Ph.D. dissertation. Rutgers University. Newark, NJ. Minchella. D. J. 1985. Host life-history variation in response to parasit- ism. Parasitology 90: 205-216. Minchella, D. J., and C. M. LoVerde. 1981. A cost of increased early reproductive effort in the snail Biomphalaria glabrata. Am. Nat- 118: 876-881. Morgan, T. S., A. D. Rogers, G. L. J. Paterson, L. E. Hawkins, and M. Sheader. 1999. Evidence for poecilogony in Pygospio elegans (Polychaeta: Spionidae). Mar. Ecol. Prog. Ser. 178: 121-132. Perrin, N., P. Christie, and H. Richner. 1996. On host lite-history response to parasitism. Oikos 75: 317-210. Polak, M., and W. T. Starmer. 1998. Parasite-induced risk of mortality ele\ ates reproductive effort in Drosophila. Proc. R. Soc. Land. B 265: 2197-2201. Poulin. R. 1995. "Adaptive" changes in behaviour of parasitized ani- mals: a critical review. Int. J. Parasitol. 25: 1371-1383. Poulin. R.. J. Brodeur, and J. Moore. 1994. Parasite manipulation of behaviour: should hosts always lose? Oikos 70: 479 1-84. Rasmussen. E. 1953. Asexual reproduction in Pygospio elegans Clapa- rede (Polychaeta sedentaria). Nature 171: 1 161-1 162. Sheldon, B. C., and S. Verhulst. 1996. Ecological immunology: costly parasite defenses and trade-offs in evolutionary ecology. Trends Ecol. Evfil. 11: 317-321. Stunkard, H. VV. 1972. Observations on the morphology and life-history of the digenetic trematode, Lepocreadium setiferoides (Miller and Northup. 1926) Martin. 1938. Biol. Bull. 142: 326-334. Suwanchaichinda. C., and S. M. Paskewitz. 1998. Effects of larval nutrition, adult body size, and adult temperature on the ability of Anopheles gambiae (Diptera: Culicidae) to melanize Sephadex beads. J. Med. Entomol. 35: 157-161. Thomas, F., R. Poulin. T. de Meeiis, J. Guegan, and F. Renaud. 1999. Parasites and ecosystem engineering: what roles could they play? Oikos 84: 167-171. Wilson. \V. H., Jr. 1983. The role of density dependence in a marine infaunal community. Ecologv 64: 295-306. Wilson. W. H., Jr. 1985. Food limitation of asexual reproduction in a spionid polychaete. Int. J. Invertcbr. Reprod. Dev. 8: 61-65. Woodin, S. A. 1982. Browsing: important in marine sedimentary envi- ronments? Spionid polychaete examples. J. Mar. Biol. Assoc. UK 60: 35-45. Zajac. R. N. 1995. Sublethal predation on Polydora comuta (Polychaeta: Spionidae): patterns of tissue loss in a field population. predator functional response and potential demographic impacts. Mar. Biol. 123: 531-541. Zar, J. H. 1996. Biostatistical Analysis. Prentice-Hall, Upper Saddle River. NJ. Zuk, M. 1990. Reproductive strategies and disease susceptibility: an evolutionary viewpoint. Pamsitol. Today 6: 231-233. Reference: Bio/. Bull. 201: 52-58. (August 2001) Delayed Insemination Results in Embryo Mortality in a Brooding Ascidian J. STEWART-SAVAGE 1 *, AIMEE PHILLIPPI 2 , AND PHILIP O. YUND 2 'Department of Biological Sciences, University' of New Orleans, New Orleans. Louisiana 70148; and 2 Darling Marine Center, School of Marine Sciences, University of Maine, Walpole, Maine 04573 Abstract. We explored the effects of temporal variation in sperm availability on fertilization and subsequent larval development in the colonial ascidian Botryllus schlosseri. a brooding hermaphrodite that has a sexual cycle linked to an asexual zooid replacement cycle. We developed a method to quantify the timing of events early in this cycle, and then isolated colonies before the start of the cycle and insemi- nated them at various times. Colony-wide fertilization lev- els (assayed by early cleavage) increased from zero to 100% during the period when the siphons of a new generation of zooids were first opening, and remained high for 24 h before slowly declining over the next 48 h. Because embryos are brooded until just before the zooids degenerate at the end of a cycle, delayed fertilization might also affect whether em- bryos can complete development within the cycle. Conse- quently, we also determined the effect of delayed insemi- nation on successful embryo development through larval release and metamorphosis. When fertilization was delayed beyond the completion of siphon opening, there was an exponential decline in the percentage of eggs that ultimately produced a metamorphosed larva at the end of the cycle. Thus, even though the majority of oocytes can be fertilized when insemination is delayed for up to 48 h, the resulting embryos cannot complete development before the brooding zooids degenerate. Introduction Field experiments have contributed greatly to current understanding of fertilization processes in free-spawning marine invertebrates (reviewed by Levitan and Petersen, 1995; Yund, 2000). In response to the evidence of potential Received 20 October 2000; accepted 8 March 2001. * To whom correspondence should he addressed. E-mail: jssavage@ uno.edu sperm limitation reported in some field studies, many lab- oratory studies have started to explore diverse related as- pects of invertebrate reproductive biology such as gamete viscosity (Thomas, 1994a,b). egg size and sperm swimming speed (Levitan, 1998), egg longevity (Meidel and Yund, 2001 ), sperm morphology (Eckelbarger et at., 1989a,b), and the kinetics of fertilization (Young. 1994; Levitan. 1998; Powell et a/., 2001 ). However, results from laboratory stud- ies have in turn led some authors to question the extent to which simple field fertilization experiments adequately mimic the details of fertilization processes in nature (e.g., Thomas, 1994a,b; Meidel and Yund, 2001). Field experi- ments may often circumvent aspects of reproductive strat- egies that have evolved to mitigate sperm limitation (Yund, 2000). Hence laboratory experiments still play a vital role in understanding reproductive strategies, and field fertilization studies should endeavor to incorporate the details of the fertilization process gleaned from laboratory work. Performing realistic field experiments with marine inver- tebrates that brood embryos presents challenges that are very different from those faced when dealing with broadcast spawners. The biggest challenge with field fertilization studies of broadcasters is interpreting results obtained by artificially holding eggs in a concentrated group (e.g., Levi- tan and Young. 1995; Wahle and Peckham, 1999) or by removing them from the water column after only a brief interval (Levitan, 1991; Coma and Lasker, 1997). This issue is moot with brooders, who by definition retain eggs and have internal fertilization. However, a different set of prob- lems merits further consideration. The precise timing of egg viability, sperm release, and fertilization itself is often less well understood than in broadcasters. Sperm function may be regulated by the female through sperm chemotaxis (Miller, 1985), activation (Bolton and Havenhand, 1996), or storage (Bishop and Ryland, 1991). In the latter case, the 52 EFFECTS OF DELAYED INSEMINATION 53 temporal pattern of fertilization within a female may be uncoupled from the pattern of sperm release by males. In hermaphrodites, the potential for self-fertilization is a con- cern, and genetic analyses of paternity may be required to conclusively exclude selting in some taxa (Yund and Mc- Cartney, 1994). For some brooders, the actual path of sperm access to eggs is poorly understood. Information on all of these topics is critical both to the design of more realistic field fertilization studies and to the interpretation of existing studies. The colonial ascidian Botryllus schlosseri is a useful model for field fertilization studies (Grosberg, 1991; Yund and McCartney, 1994; Yund, 1995, 1998). Fertilization is internal, and embryos are brooded until released as tadpole larvae (Milkman, 1967). When colonies are grown on glass surfaces, egg production can be quantified non-destructively (Yund et cil.. 1997), thus permitting estimation of fertiliza- tion levels by comparing egg and embryo counts (Yund, 1995, 1998). Although the general time of fertilization within the life cycle (i.e., temporal resolution on the order of a day) has long been known (Milkman, 1967), the finer- scale timing (temporal resolution on the order of hours) has not been explored. Many authors have assumed that the apparent temporal separation of fertilization and sperm re- lease prevents self-fertilization (e.g., Milkman, 1967; Gros- berg, 1987; Yund and McCartney, 1994), but we have recently shown (Stewart-Savage and Yund, 1997) that sperm release commences several days earlier than previ- ously thought. Although sperm storage has been demon- strated in another colonial ascidian (Bishop and Ryland. 1991; Bishop and Sommerfeldt. 1996), past workers have implicitly assumed that storage is unlikely in B. schlosseri (Milkman, 1967; Grosberg. 1991; Yund, 1995, 1998). To the best of our knowledge, this assumption has never been explicitly tested. To address this interrelated set of issues, this paper explores the effect of variation in the timing of fertilization on fertilization levels and subsequent larval development in B. schlosseri. and compares those results with published information on the timing of sperm release. Materials and Methods Study 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 zooids periodically undergo a synchronous asexual zooid replacement cycle in which a new generation of zooids, termed buds, forms between the existing zooids (Berrill, 1941; Izzard. 1973). At the end of the life span of adult zooids (about 8 days at 16C; cycle length is temperature dependent), the buds expand, take over the function of the previous generation of zooids (which are quickly resorbed). and then commence their sexual reproductive cycle. The sexual 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 developing embryos, which are released just before the zooids degenerate at the end of the cycle (Milkman, 1967). Standard methods The colonies of B. schlosseri that were employed in this study were collected from the Damariscotta River. Maine. Animals were grown on glass microscope slides in the flowing seawater system at the University of Maine's Dar- ling Marine Center. Field-collected colonies that had been established in laboratory culture were divided to provide clonal replicates (ramets) of genotypes. Colonies employed in all experiments were monitored for the approach of takeover (the transition between zooid generations). When colonies were about to commence takeover (late stage 5 through early stage 6 by the criteria of Milkman, 1967), they were isolated in 50 ml of sperm-free (aged >24 h) seawater. Isolated colonies were housed in an incubator at 16C (range: 14-18 C) and fed phytoplankton (Duniella sp.) at densities of approximately 10 5 cells/ml. Water and food were changed twice daily. Colonies were monitored for siphon opening and then isolated in individual 250-ml con- tainers with algae (water and food were changed daily) until exposed to sperm. Sperm exposure was accomplished by placing colonies in a flowing seawater tank in proximity to numerous male-phase colonies (>24 h after siphon open- ing; Stewart-Savage and Yund, 1997) for 1 h. After insem- ination, colonies were rinsed with aged seawater and re- turned to isolation. Experimental protocols To standardize insemination times, we first had to accu- rately quantify the start of the reproductive cycle (i.e., the functional opening of siphons). Inhalant siphons are formed early in the takeover process, but the common exhalant siphon of a system generally does not form until near the end. However, it is difficult to ascertain functional siphon opening on morphological criteria alone. In the course of other work, we observed that the consumption of green algae immediately turned the digestive systems of actively feeding zooids (i.e., those that must have open siphons) green. Consequently, we used algal uptake as an assay for siphon opening. To establish the temporal pattern of siphon opening, we isolated 14 colonies and briefly exposed them to algae three to four times during the process of takeover. At each sample interval we recorded the percentage of siphons that were open (% of zooids with green digestive systems). From these data we calculated an average rate of siphon opening. This approach subsequently allowed us to 54 J. STEWART-SAVAGE ET AL. make single observations of the percentage of siphons that were open and back-calculate the time of the first siphon opening. Both of our other experiments use this approach to estimate the time of initial siphon opening, and the timing of insemination is expressed relative to this event. To examine the effect of the timing of fertilization on fertilization levels, we exposed colonies to sperm through a range of different times after siphon opening (0.5 to 96 h; n =- 79). Colonies with about 20 eggs (mean of 20.0 standard error of 11.6) were utilized throughout, and all eggs and embryos in a colony were surgically removed 10-18 h after insemination and scored for successful de- velopment. Initial studies indicated that embryos should be in the 8-cell to the 32-cell stages during this time range. Uncleaved eggs were scored as unfertilized, as were em- bryos with an abnormal cleavage pattern (arrested cleavage, abnormal cell number or shape). A few embryos at ad- vanced developmental stages (e.g., gastrula) were excluded from the data set since fertilization was by either contami- nating or self sperm. To examine the effect of timing of fertilization on sub- sequent development and metamorphosis, colonies were initially fertilized in sets of multiple ramets per genotype. For each genotype, one ramet was left unfertilized (to assess the level of sperm contamination or self-fertilization), one ramet was fertilized about 22 (2) h after the beginning of siphon opening (when results from the previous experiment indicated that all siphons should be open), and remaining ramets (2-3) were fertilized at various times up to 85 h after initial siphon opening. Because fertilization was consis- tently minimal in unfertilized controls and the availability of genotypes with multiple egg-bearing ramets was often lim- ited, later trials were conducted without the control treat- ment. Before takeover, we counted the number of eggs produced by each colony (minimum egg production was set at 25 eggs). After insemination, colonies were returned to isolation until all ramets of a genotype had been fertilized and at least 24 h had elapsed since the last insemination. Colonies were subsequently housed in a flowing seawater table with an independent seawater supply while embryonic development proceeded: they were re-isolated at stage tour (Milkman. 1967). After each isolated colony had started the next reproductive cycle, all metamorphosed juveniles in the isolation container were counted. Data from colonies that died or became visibly unhealthy during the experiment were discarded. Results Timing of siphon opening Feeding did not begin until after the organization of zooids into new systems and formation of the common exhalant siphon. Although the rate of siphon opening varied among colonies (Fig. 1: range of 3.0%/h 17.8%/h), the 100 n o o N so H 40 - -o u u 20 - 4 Time from Initial Observation (h) Figure 1. Rate of siphon opening in colonies of Botryllus schlosseri as assayed by the presence of algae in the digestive system. Colonies were isolated in 50 ml aged seawater with 2 x 10 5 algae/ml and monitored at intervals of from 1 to 12 h. Zero time is the first observation of algae in the gut. Temporal patterns for 14 individual colonies are shown. Differences in the v-intercept simply reflect how far the takeover process had proceeded when colonies were first observed; slopes indicate the rate of siphon opening. average rate of siphon opening of the colonies was 7.8%/ h 4.5%/h (X SD). We used the average rate of siphon opening to normalize the time of sperm exposure to the start of siphon opening for colonies in the other two experiments. Effect of timing of insemination on fertilization levels To determine the time frame during which eggs can be fertilized within the female, we exposed virgin females to a 1-h pulse of sperm at various times after the beginning of siphon opening and assayed successful fertilization by the percentage of normally cleaved embryos present (Fig. 2). When virgin females were exposed to sperm during the period in which their siphons were opening (first 24 h), the level of fertilization increased with time (Fig. 2B). In col- onies fertilized during siphon opening, there was no spatial relationship between fertilized and unfertilized eggs either within or among systems; it was common to find both in the same zooid. Because the rate of increasing fertilization (5.4%/h) is similar to the rate of siphon opening (7.8%/h 4.5%/h), we conclude that fertilization of the eggs within a zooid occurs shortly after the opening of the siphon. After the completion of siphon opening, fertilization suc- cess remained high (>90%) for 24 h and then declined over the next 48 h with a 7" 500 , of 72 h (Fig. 2B). In a subset of genotypes where multiple ramets were inseminated at dif- ferent times in the same reproductive cycle, thus controlling for potential genotype and cycle effects, the effect of EFFECTS OF DELAYED INSEMINATION 55 72 96 B 24 48 72 96 Insemination Time (h after start siphon opening) Figure 2. Effect of insemination pulse timing on fertilization levels. Colonies were isolated before the start of siphon opening, monitored for the timing of siphon opening, and exposed to sperm for 1 h: the number of cleaving embryos was determined 10-18 h later. (A) Fertilization levels in different ramets of seven genotypes fertilized at different points in the same reproductive cycle. (Bl Overall effect of insemination time on fertilization success in ramets from 25 genotypes. The line represents a polynomial regression of the data (R 2 = 0.580). delayed insemination on fertilization varied by genotype (Fig. 2 A). Of the seven genotypes in which different ramets were inseminated at different times, five genotypes had a decline in fertilization that mirrored the population data. In the other two genotypes, fertilization levels declined rapidly in one. but remained relatively stable over 60 h in the other. Excluding the genotype that exhibited little decline in fer- tilization, the average T 50Vf for the reduction of fertilization was 62 15 h, a value similar to the population-wide regression. Effect of liming of insemination on embr\o development and metamorphosis The maximum duration of gestation is fixed by the length of the asexual zooid replacement cycle. Since eggs could be fertilized well after siphon opening, but the time of embryo release is fixed, we examined the effect of delayed insem- ination on reproductive success. Successful embryo meta- morphosis was selected as an assay of reproductive success because it integrates possible effects on fertilization, devel- opment, larval behavior, and settlement. In five trials that included unfertilized (low control), insemination at 22 h (high control), and ramets inseminated at different times after siphon opening, the percentage of eggs that success- fully developed through metamorphosis consistently de- creased with the time of insemination (Fig. 3A). The unfer- tilized controls resulted in either zero or very low (<5%) levels of larval metamorphosis (Fig. 3 A). However, the percent of eggs developing through metamorphosis varied substantially among 22-h insemination controls (Fig. 3A). Because of the low levels of successful metamorphosis in two genotypes fertilized at 22 h. we calculated the T 50% relative to the maximum value for each genotype. The relative T 50C7c for the reduction of metamorphosis success was 41 6 h after the start of siphon opening (about 19 h after the completion of siphon opening). When data from all 12 trials were combined (Fig. 3B), larval metamorphosis exhibited an exponential decline with fertilization time be- yond 22 h. No larval metamorphosis occurred when colo- nies were fertilized more than 78 h after the start of siphon opening. Two outliers (both ramets of the same genotype) had disproportionately high levels of metamorphosis when fer- tilized about 48 h after siphon opening (Fig. 3B. open squares). Independent evidence (i.e.. observations of suc- cessful embryo development in isolated colonies) suggested that this genotype may sometimes be able to self-fertilize. Alternatively, the high fertilization levels in these two col- onies may be the result of sperm contamination. Because these inconsistent values are limited to one genotype, we have excluded these values from the regression in Figure 3B. Inclusion of the two points in the regression has little effect on the equation parameters, but it substantially re- duces the coefficient of determination. Note that many other ramets of this genotype were employed in this experiment (Fig. 3B, open squares) and produced results consistent with those of the other genotypes. Discussion Although more than 50% of Boti-yllus schlosseri eggs can be fertilized 38 to 48 h after the completion of siphon opening (Fig. 2), few viable larvae are produced unless fertilization occurs within the first 19 h (Fig. 3). The de- crease in embryo production after delayed fertilization could be caused by either egg aging or limitations on the duration of brooding. As in most invertebrates, the time required to complete development is a function of temper- ature in B. schlosseri. Since the asexual zooid replacement 56 J. STEWART-SAVAGE ET AL. _c =0 .5 o " p. I o g E Q - 100 r 75 - 50 25 ~^^T! \ . " Tj HI 3 24 48 72 Unfert o - a? . o -- .2 e- o o | Q 2 100 r 75 - 50 25 - B 24 48 72 96 Insemination Time (h after start siphon opening) Figure }. Effect of insemination pulse timing on embryo development and larval metamorphosis. Colonies with quantified egg production were isolated before the start of siphon opening, monitored for the timing of siphon opening, and exposed to sperm for 1 h: the number of settled juveniles was determined 5-7 days later. (A) Developmental success of different ramets from five genets. In three of the genets, one ramet was never exposed to sperm (unfertilized, solid symbols). (B) Overall effect of insemination time on successful development. The open squares are the ramets from the putative self-fertilizing genotype; closed symbols repre- sent the other 1 1 genotypes. The line is an exponential regression of the data except for two outliers at 48 h (R 2 = 0.713). cycle is also a function of temperature (Grosberg, 1982). delayed fertilization could cause the brooding zooids to degenerate before the embryos have become competent to undergo metamorphosis. The deleterious effects of egg ag- ing have been demonstrated in mammals (Juetten and Bavister. 1983; Xu et <(/., 1997), but such effects are usually manifested early in development. Since early development was normal in all but one colony with delayed fertilization (pers. obs.), the decreased gestational duration caused by delayed fertilization is more likely to be responsible. Nev- ertheless, additional work on the mechanism by which de- layed fertilization decreases larval production could more fully resolve this issue. In spite of the narrow temporal window in which both fertilization and development are likely to be successful (Figs. 2 and 3), field experiments indicate that colonies of B. schlosseri are very adept at acquiring sperm. A single male- phase colony can fertilize most eggs of a nearby female- phase colony with very few sperm (Yund, 1998). If several males are present, they compete to fertilize eggs (Yund. 1995. 1998), and closer males can be successful at the expense of more distant males (Yund and McCartney, 1994). Although sperm transfer usually occurs among nearby colonies (Yund. 1995). sperm can also be obtained from very distant locations when insufficient local sperm are available (Yund, 1998). Even eggs of colonies isolated from the nearest natural populations by tens of meters can be fertilized at appreciable levels (Yund and McCartney, 1994). The apparent ease of fertilization under field condi- tions, in spite of a very limited temporal window for suc- cessful fertilization and development, suggests that the pro- cess of sperm capture by colonies must be extremely efficient. Nevertheless, in low-density populations where sperm may be in short supply (Yund. 1998). or in marginal habitats in which sperm production is suppressed (Stewart- Savage et a/.. 2001). our work suggests that reproductive failure may occur in spite of successful fertilization if fer- tilization occurs too late in the reproductive cycle. Recent field sampling has demonstrated this phenomenon in natural populations near the end of the annual reproductive season (Yund and Phillippi, unpubl. data). Unlike the colonial ascidian Diplosoma listerianum, in which fertilization can be temporally disassociated from sperm exposure and colonies can store sperm for up to one month (Bishop and Ryland. 1991 ; Bishop and Sommerfeldt, 1996). B. schlosseri colonies apparently cannot store sperm. The evidence for this conclusion is, first, that colonies isolated in sperm-free seawater were not fertilized until we experimentally supplied a sperm pulse, indicating that sperm are not stored and transferred from one asexual generation of zooids to the next. The apparently complete resorption of all zooid tissue at the end of the cycle further suggests that transmission between cycles is unlikely. Sec- ond, the tight temporal relationship between siphon opening and fertilization (Fig. 2B) suggests that sperm cannot enter until the new generation of zooids opens its siphons and starts to feed. Third, the narrow window of time in which fertilization is both possible (Fig. 2) and results in viable offspring (Fig. 3) eliminates any apparent fitness advantage to sperm storage within a single asexual generation. The route of sperm access to eggs in B. schlosseri is unknown, but there are at least two possible points of entry (Ryland and Bishop. 1993): sperm enter through the EFFECTS OF DELAYED INSEMINATION 57 inhalant siphon and cross the pharyngeal basket to reach the eggs, or sperm enter through the exhalant siphon and then swim to the eggs. During takeover in B. schlosseri, the exhalant siphon of each system is formed before the inhal- ant siphons of all of the component zooids open, and the precise timing of exhalant siphon formation varies among systems (pers. obs.). If sperm enter via the exhalant siphon, fertilization levels in the early time intervals of our fertili- zation timing experiment should have varied among sys- tems, but should not have varied within a system. However, we routinely found mixtures of fertilized and unfertilized eggs within the same system, suggesting that sperm entry to each zooid required an open inhalant as well as exhalant siphon. Although further work is required to determine the route of sperm entry into Botryllus colonies, we think it is unlikely that sperm enter via the exhalant siphon. Hermaphroditism creates another challenge for success- ful reproduction in B. schlosseri. Inbreeding depression (Sabbadin, 1971) is likely to exert selective pressure to prevent self-fertilization, even though selting would be a possible mechanism to assure fertilization in the narrow time window in which fertilization can produce functional embryos. When the data in this paper are combined with previous data on the timing of sperm release (Stewart- Savage and Yund. 1997). it is apparent that the male and female phases of the reproductive cycle overlap in B. schlosseri (Fig. 4). Sperm release overlaps for about 48 h with the window for successful fertilization, but there is substantially less overlap with the narrower window in which fertilization results in viable embryos (Fig. 4). Con- sequently, B. schlosseri is not a true sequential hermaphro- dite (Milkman, 1967), but the male and female phases are functionally separated in time. This functional segregation of the reproductive phases probably plays some role in 24 48 7: 96 120 144 168 192 216 240 Time From Completion of Siphon Opening (h) Figure 4. Relationship between male and female reproductive phases in Botryllus schlosseri. Data collected at different temperatures have been normalized to a 10-day cycle length. The zero time point is the completion, rather than the initiation (as in Figs. 2 and 3), of siphon opening. The sperm release curve is redrawn from Stewart-Savage and Yund (1997) with permission. ensuring that few metamorphosing embryos result from self-fertilization. However, the very success of our experi- mental protocols indicates that one or more additional mechanisms to prevent self-fertilization must exist. Eggs of colonies isolated in small volumes of water until points in the reproductive cycle at which substantial self-sperm should have been present (Fig. 4) nevertheless remained unfertilized until we introduced a pulse of sperm (with the possible exception of the two outliers in Fig. 3B). Conse- quently, some form of self-incompatibility, as described in other colonial and solitary ascidians (Rosati and De Sands. 1978: Bishop, 1996), appears likely in B. schlosseri (see also Scofield et /., 1982). Acknowledgments Financial support was provided by the National Science Foundation (OCE-97-30354). This is contribution number 366 from the Darling Marine Center. Literature Cited Berrill, N. J. 1941. The development of the bud in Bolryllus. Biol. Bull. 80: 169-1X4. Bishop, J. D. D. 1996. Female control of paternity in the internally fertilizing compound ascidian Diplosoma listeriannm. I. Autoradio- graph investigation of sperm movements in the female reproductive tract. Proc. R. Soc. Land. B 263: 369-376. Bishop, J. D. D.. and J. S. Ryland. 1991. Storage of exogenous sperm by the compound ascidian Diplosoma listeriannm. Mar. Biol. 108: 111-118. Bishop. J. D. D.. and A. D. Sommerfeldt. 1996. Autoradiographic investigation of uptake and storage of exogenous sperm by the ovary of the compound ascidian Diplosoma listerianum. Mar. Biol. 125: 663- 670. Bolton. T. F.. and J. N. Havenhand. 1996. Chemical mediation of sperm activity and longevity in the solitary ascidians Ciona intestinalis and Ascitlit'lla aspersa. Biol. Bull. 190: 329-335. Coma, R., and H. R. Lasker. 1997. Effects of spatial distribution and reproductive biology on in situ fertilization rates of a broadcast-spawn- ing invertebrate. Biol. Bull. 193: 20-29. Kckelbarger. K. J., C. M. Young, and J. L. Cameron. 1989a. Modified sperm ultrastructure in four species of soft-bodied echinoids (Echino- dermata: Echinothuriidae) from the bathyal zone of the deep sea. Biol. Bull. 177: 230-236. F.ckelharger, K. J., C. M. Young, and J. L. Cameron. 1989b. Ultra structure and development of dimorphic sperm in the abyssal echinoid Phrissoc\sti.t multispina (Echinodermata: Echinoidea): implications for deep sea reproductive biology. Biol. Bull. 176: 257-271. Grosberg, R. K. 1982. Ecological, genetical, and developmental factors regulating life history variation within a population of the colonial ascidian Botryllus schlosseri (Palas) Savigny. Ph.D. dissertation. Yale University. New Haven, CT. Grosberg. R. K. 1987. Limited dispersal and proximity-dependent mat- ing success in the colonial ascidian Botryllus schlosseri. Evolution 41: 372-384. Grosberg, R. K. 1991. Sperm-mediated gene flow and the genetic struc- ture of a population of the colonial ascidian Botryllus schlosseri. Evolution 45: 130-142. Izzard. C. S. 1973. Development of polarity and bilateral asymmetry in the palleal bud of Botryllus schlosseri (Pallas). ./. Morphol. 139: 1-26. 58 J. STEWART-SAVAGE ET AL Juetten, J., and B. D. Bavister. 1983. Effects of egg aging on in vitro fertilization and first cleavage division in the hamster. Gamete Res. 8: 219-230. Levitan, D. R. 1991. Influence of body size and population density on fertilization success and reproductive output in a free-spawning inver- tebrate. Biol. Bull. 181: 261-268. Levitan. D. R. 1998. Does Bateman's principle apply to broadcast- spawning organisms? Egg traits influence in situ fertilization rates among congeneric sea urchins. Evolution 52: 1043-1056. Levitan, D. R., and C. Petersen. 1995. Sperm limitation in the sea. Trends Ecol. Evol. 10: 228-231. Levitan, D. R., and C. M. Young. 1995. Reproductive success in large populations: experimental measures and theoretical predictions of fer- tilization in the sea biscuit Clypeasler rosaceus. J. Exp. Mar. Biol. Ecol 190: 221-241. Meidel, S. K., and P. O. Yund. 2001. Egg longevity and time-integrated fertilization in a temperate sea urchin (Strongylocentrotus droebachien- sis). Biol. Bull. 201: 000-000. Milkman. R. 1967. Genetic and developmental studies on Botrylhis schlosseri. Biol. Bull. 132: 229-243. Miller, R. L. 1985. Sperm chemo-oriemation in the metazoa. Pp. 275- 337 in Biology of Fertilization. Vol. 2, C. B. Metz and A. Monroy. eds. Academic Press, Orlando. FL. Powell, I). K., P. A. Tyler, and L. S. Peek. 2001. Effect of sperm concentration and sperm ageing on fertilization success in the Antarctic soft-shelled clam Laternula elliptica and the Antarctic limpet Nucella concinna. Mar. Ecol. Prog. Ser. 215: 191-200. Rosati, F., and R. De Santis. 1978. Studies on fertilization in the ascidians. I. Self-sterility and specific recognition between gametes of dona intestinalis. Ev/>. Cell Res. 112: 1 1 1-1 19. Ryland, J. S., and J. D. D. Bishop. 1993. Internal fertilization in hermaphroditic colonial invertebrates. Oceanogr. Mar. Biol. Annit. Rev. 31: 445-477. Sabbadin, A. 1971. Self- and cross-fertilization in the compound ascid- ian Botryllus schlosseri. Dev. Biol. 24: 379-391. Scofield, V. L., J. M. Schlumpberger, and I. L. Weissman. 1982. Colony specificity in the colonial tunicate Bony/Ins and the origins of vertebrate immunity. Am. Zoo/. 22: 783-794. Stewart-Savage, J., and P. O. Yund. 1997. Temporal pattern of sperm release from the colonial ascidian. Botryllus schlosseri. J. Exp. Zooi 279: 620-625. Stewart-Savage, J.. A. Stires, and P. O. Yund. 2001. Environmental effect on the reproductive effort of Botrylhis schlosseri. Pp. 3 1 1-3 14 in Biology of Ascidians. H. Sawada. H. Yokosawa, and C. C. Lambert, eds. Springer-Verlag, Tokyo, Japan. Thomas, K. I. M. 1994a. Physical properties of gametes in three sea urchin species. J. E.\p. Biol. 194: 263-284. Thomas. F. I. M. 1994b. Transport and mixing of gametes in three free-spawning polychaete annelids, Phragmatopoma californica (Fewkes), Sabellaria cementarium (Moore), and Schi-obranchia insig- nis (Bush). J. Exp. Mar. Biol. Ecol. 179: 1 1-27. Wahle, R. A., and S. H. Peckham. 1999. Density-related reproductive trade-offs in the green sea urchin. Strongylocentrotus droebachiensis. Mar. Biol. 134: 127-137. Xu, Z., A. Abbott, G. S. Kopf, R. M. Schultz. and T. Ducibella. 1997. Spontaneous activation of ovulated mouse eggs: time dependent effects on M-phase exit, cortical granule exocytosis, maternal messenger ri- bonucleic acid recruitment and inositol 1,4,5-trisphosphate sensitivity. Biol. Reprod. 57: 743-750. Young, C. M. 1994. The biology of external fertilization in deep-sea echinoderms. Pp. 179-200 in Reproduction. Lan-al Biology and Re- cruitment of the Deep-sea Benthos. C. M. Young and K. J. Eckelbarger, eds. Columbia University Press. New York. Yund, P. O. 1995. Gene flow via the dispersal of fertilizing sperm in a colonial ascidian (Botryllus schlosseri): the effect of male density. Mar. Biol. 122: 649-654. Yund, P. O. 1998. The effect of sperm competition on male gain curves in a colonial marine invertebrate. Ecology 79: 328-339. Yund, P. O. 2000. How severe is sperm limitation in natural populations of marine free-spawners? Trends Ecol. Evol. 15: 10-13. Y'und, P. O., and M. A. McCartney. 1994. Male reproductive success in sessile invertebrates: competition for fertilizations. Ecologv 75: 2151-2167. Yund, P. O., Y. Marcum, and J. Stewart-Savage. 1997. Life-history variation in a colonial ascidian: broad-sense heritabilities and tradeoffs in allocation to asexual growth and male and female reproduction. Biol. Bull. 192: 290-299. Reference: Biol. Bull. 201: 59-64. (August 20(11) Morula Cells as the Major Immunomodulatory Hemocytes in Ascidians: Evidences From the Colonial Species Botryllus schlosseri LORIANO BALLARIN 1 -*, ANTONELLA FRANCHINI 2 , ENZO OTTAVIANI 2 , AND ARMANDO SABBADIN 1 1 Department of Biologv, University of Padova, via U. Bassi 58/B, 35100 Padova. Italy: and ^Department of Biologv, University of Modena and Reggio Emilia, via Campi 213/D, 41100 Modena, Italy Abstract. Immunocytochemical methods were used to study the presence and distribution of IL-1 -a- and TNF-a- like molecules in the hemocytes of the colonial ascidian Botryllus schlosseri. Only a few unstimulated hemocytes were positive to both the antibodies used. When the hemo- cytes were stimulated with either mannan or phorbol 12- mono-myristate. the phagocytes were not significantly changed in their number, staining intensity, or cell morphol- ogy. In contrast, stimulated morula cells were intensely labeled, indicating that these cells play an important immu- nomodulatory role. Introduction Phagocytes and morula cells are two types of circulating hemocytes that play a key role in ascidian immunobiology. Phagocytes can easily recognize and ingest non-self cells and particles (Smith, 1970; Anderson. 1971; Fuke and Fu- kumoto. 1993: Ballarin et ai. 1994: Ohtake et ai, 1994; Dan-Sohkawa et ai, 1995; Cima et ai, 1996) and are able to synthesize and release opsonic agglutinins (Coombe et ai, 1984; Kelly et ai, 1992; Ballarin et ai. 1999). Morula cells, a ubiquitous hemocyte type among ascidians. take part in a variety of biological functions of irnmunological rele- vance, such as hemolymph clotting, tunic synthesis, and Received 18 July 2000; accepted 10 May 2001. * To whom correspondence should be addressed. E-mail: ballarin@civ.bio.unipd.it Abbreviations: FSW. filtered seawater; HA. hyaline amoebocytes: IL. imerleukin; MLC. macrophage-like cells: PMM. phorbol 12-mono-myris- tate; TNF. tumor necrosis factor. encapsulation of foreign bodies (Endean, 1955b; Smith. 1970; Anderson, 1971; Chaga, 1980; Wright. 1981: Za- niolo. 1981). They are by far the most frequent circulating ascidian cell-type (Endean. 1955a: Andrew. 1961; Smith. 1970; Kustin et ai, 1976; Ballarin et ai, 1995). and their abundance suggests direct involvement in other important defense reactions. Although most of their roles in ascidian immune responses still remain unclear, morula cells can induce cytotoxicity after recognition of foreign molecules or cells (Parrinello. 1996; Cammarata et ai. 1997: Ballarin et til.. 1998), and they are also required for phagocytosis (Smith and Peddie. 1992). Cytokines are soluble molecules that mediate communi- cation among various immunocyte types in vertebrate im- mune systems. In the last decade, much evidence has accu- mulated indicating that cytokine-like molecules are also involved in invertebrate immune responses, and their pres- ence has been demonstrated in hemocytes of molluscs, annelids, arthropods, echinoderms, and tunicates (Beck and Habicht. 1991; Ottaviani et ai. 1995a.b. 1996: Franchini et iii. 1996). Cytokine-like molecules stimulate cell prolifer- ation, increase hemocyte motility and phagocytic activity, and induce nitric oxide synthase (Raftos et ai, 1991: Otta- viani ft ai, 1995b). As regards ascidians, the activities of interleukin-l (IL-1 )- and IL-2- but not tumor necrosis factor (TNF)-like molecules have been revealed in various spe- cies, either solitary or colonial (Beck et ai. 1989). Tunicate IL-1 -like molecules modulate immune responses and are secreted by hemocytes in response to exogenous stimuli (Raftos et ai. 1991. 1992. 1998: Beck et ai, 1993; Kelly et ai, 1993). 59 60 L. BALLARIN ET AL. We have studied in hemocytes of the colonial ascid- ian Botryllus schlosseri the presence and distribution of molecules that are immunoreactive to antibodies raised to human IL-l-a and TNF-a. The results indicate that these immunoreactive molecules are mainly detectable in stim- ulated morula cells, suggesting that these cells have a role in immunomodulation. Moreover, previous results in other ascidian species are supported (Smith and Peddie, 1992). Materials and Methods Animals Wild colonies of Botryllus schlosseri from the lagoon of Venice, Italy, were used. They were kept in aerated aquaria, attached to glass slides, and fed with Liquifry Marine (Liquifry Co., England) and algae. Hemocyte monolayers Colonies were rinsed in filtered seawater (FSW), pH 7.5, containing 10 mM L-cysteine as anticoagulant. The tunic marginal vessels were then punctured with a fine tungsten needle, and hemolymph was collected with a glass micropi- pette. Hemolymph was centrifuged at 780 X g for 10 min, and pellets were resuspended in FSW to a final hemocyte concentration of 8-10 X 10 6 cells/ml. Samples of the he- mocyte suspension (50-100 /xl) were cytocentrifuged onto slides with a Shandon Instrument Cytospin II running at 500 rpm for 2 min. Hemocytes were then stained with May Griinwald-Giemsa for morphological examination with a Leitz Dialux 22 light microscope. Hemocyte stimulation Cell suspensions were placed in 1-ml tubes on a revolv- ing mixer, and hemocytes were stimulated by incubation for 5, 15, 30, and 60 min with mannan at 5 mg/ml or phorbol 12-mono-myristate (PMM) at 20 nM in FSW containing 10 mM L-cysteine to prevent cell clotting. Mannan. a quite common microbial polysaccharide, is easily recognized by mannose receptors, the presence of which has been indi- rectly interred on the surface of Botryllus phagocytes (Bal- larin et al., 1994). PMM is a well-known activator of protein kinase C that mimics the action of diacylglycerol (Wolfe, 1993). The above-reported concentrations of the two com- pounds were previously demonstrated as the most effective in stimulating Botryllus phagocytes and the related respira- tory burst (Ballarin et al.. 1994; Cima et al.. 1996). FSW was used for controls. The viability of hemocytes, after the incubation, was assessed by the trypan blue exclusion assay (Gorman et al., 1996). Immunocytochemistry The immunocytochemical procedure described by Otta- viani et al. (1990) was performed. The following two pri- mary antibodies were used: polyclonal anti-human IL-l-a (1:250, 1:500, 1:1000) (Santa Cruz Biotech., USA) and monoclonal anti-human TNF-a (1:25, 1:50, 1:100) (Neo- Markers, USA). Cells were incubated with primary antibod- ies overnight at 4C, and reactivity was revealed by immu- noperoxidase staining using avidin-biotin-peroxidase complex (Hsu et al.. 1981). The best results were obtained with anti-IL-1-a and anti-TNF-a diluted 1:500 and 1:25, respectively. In control preparations, the primary antibodies were either substituted with non-immune sera or absorbed with homologous antigen (i.e., human IL-l-a and TNF-a) before addition to hemocyte monolayers. Moreover, a poly- clonal antibody raised against Botryllus agglutinin (BA) (Ballarin et nl., 2000) was also assayed as a control for specificity. Nuclei were counterstained with hematoxylin. The frequency of positive hemocytes, phagocytes, and morula cells was reported as the percentage of the total hemocyte number, which was determined by counting at least 600 cells in 10 fields under the light microscope. Statistical analysis All experiments were repeated in triplicate, and statistical analysis was performed using the chi-square test (^ 2 ). Results Morphology of cytocentrifuged Botryllus hemocytes The main hemocyte types present in B. schlosseri hemo- lymph were identifiable under the light microscope after cytocentrifugation. Lymphocyte-like cells, representing 2%-4% of circulating hemocytes. contain a large round nucleus surrounded by a thin layer of basophilic cytoplasm. Phagocytes, which include hyaline amoebocytes (HA; ac- tively phagocytosing cells) and macrophage-like cells (MLC) (Ballarin et al.. 1994). have roundish nuclei and neutrophilic cytoplasm which, in the case of MLC, sur- rounds one or more vacuoles containing ingested material (Fig. la, b). Phagocytes constitute 30%-40% of circulating blood cells. Morula cells, the frequency of which is 30%- 50% of total hemocytes, are characterized by the presence of several yellowish-green vacuoles (Fig. 2a, c). Nephro- CYTOKINE-LIKE MOLECULES IN BOTRYLLUS 61 LL . HA MLC * N a Figure 1. Cytocentrifuged Botry/liis schlosseri hemocytes stained with May Griinwald-Giemsa solution, (a) Lymphocyte-like cell (LL) and hyaline amebocyte (HA); (b) macrophage-like cell (MLC; n: nucleus; v: vacuole); (cl nephrocyte (N) with several empty vacuoles (arrowheads). Bar = 10 /xm. cytes and pigment cells (6%-10% of circulating hemocytes) were not well preserved after cytocentrifugation; they ap- peared as giant cells with empty vacuoles (Fig. Ic). Response of unstimulated hemocytes to anti-cytokine antibodies Using anti-IL-1-a and anti-TNF-a, only some phago- cytes and a few morula cells were labeled after immuno- peroxidase staining (Table 1). Thus, most HA, MLC, and morula cells were not immunoreactive with either antibody (Fig. 3). Moreover, no other cell-types stained positively for stimulated cvtes to antibodies raised to human cytokines Antibodies' 1 Cell type Anti-iL-1-a Anti-TNF-a Phagocytes'" Morula cells 0.4 0.3 1.1 0.9 0.9 0.4 4.5 1.2 a Values are percentage of total hemocytes plus or minus the standard deviation. h Phagocytes include hyaline amoebocytes and macrophai -li; 62 L. BALLARIN ET AL. anti-BA l&r d e anti-cytokine Figure 3. Immunocytochemistry on Botryllus schlosseri hemocytcs with anti-BA (a, h), and anti-cytokine (c-e) antibodies, (a) Positive HA; (hi negative morula cells; (c) unlabeled. unstimulated HA; (d) stimulated HA positive for IL-l-a; (e) stimulated MLC positive for TNF-a. Bar = 1? /xm. Discussion In the present work, we demonstrate that molecules rec- ognized by antibodies raised to human IL-l-o and TNF-o are present in immunocytes of the compound ascidian Bot- tyllus schlosseri. After stimulation, only morula cells, among all hemocytes, show a marked and significant in- crease in immunoreactivity. The increase in the number of immunoreactive cells depends on the length of the time of hemocyte incubation with the stimulating agents. In con- trast, among unstimulated hemocytes, only some morula cells and a few phagocytes are immunoreactive. Therefore, although the ligands recognized by the antibodies used are unknown and notwithstanding that serological cross-reac- tivity is not sufficient proof of evolutionary homology be- tween those ligands and vertebrate cytokines. still our data indicate that the morula cells have an important immuno- modulatory role in ascidian blood. We hypothesize that morula cells are the main source of cytokine-like molecules in Botryllus hemolymph, which can better explain their abundance in the circulation. Indeed, these cells are able to encapsulate foreign bodies (Anderson, 1971; Wright, 1981; De Leo el al, 1996) and are involved in clotting after blood vessel damage (Vallee, reported by Wright, 19X1). In many ascidian species, they can also induce cytotoxicity after recognition of foreign molecules or cells (Parrinello, 1996; Cammarata el al.. 1997: Ballarin ct uL. 1998). All these events can be modulated by cytokine- like molecules produced by activated cells. In agreement with this view, TNF-a-like molecules are involved in insect encapsulation (Franchini et ui, 1996), and IL-1-like mole- cules have been shown to stimulate echinoderm coelomo- cyte aggregation, which occurs in encapsulation (Beck and Habicht, 1991 ). Moreover, in vertebrates, both TNF-a and IL-l-n stimulate immune and inflammatory responses, and TNF-a is required for blood coagulation (Abbas et al.. 1991). The induction of cytokine-like molecules in hemocytes after stimulation has already been reported in bivalve mol- luscs and insects: in all these cases, phagocytes are the immunoreactive cells (Hughes et al.. 1990; Franchini et al., 1996). Analogously, in vertebrates, mononuclear phago- cytes are the main source of both IL-l-a and TNF-a (Abbas ct al.. 1991 ). Nevertheless, the situation in Botryllus appears peculiar in that positivity to anti-cytokine antibodies is absent from the majority of phagocytes without significant differences in its distribution between unstimulated and stimulated cells. Although morula cells have no phagocytic activity, they are reported to promote phagocytosis by ascidian phago- cytes (Smith and Peddie, 1992). Thus, the stimulatory effect on phagocytes and the enhancement of phagocytosis by morula cell lysates (Smith and Peddie, 1992) may easily be explained by the immunomodulatory role of the cytokines they produce. This idea is strongly supported by the obser- CYTOK1NE-L1K.E MOLECULES IN BOTRYLLUS 63 35- 25- f 15 60 15 30 time (min) 60 Figure 4. Morula cells positive to anti-IL-1-a and anti-TNF-a, ex- pressed as percentage of total hemocytes. after stimulation with either mannan at 5 mg/ml (circles) or PMM at 20 nM (triangles) for 5, 15, 30, and 60 min. *P < 0.001 vs. control (unstimulated hemocytes, t = 0). vation that the time-dependent increase of immunoreactive morula cells closely resembles the time-dependent increase in the frequency of phagocytizing hemocytes in in vitro assays (Ballarin et al., 1997). The opsonic role of tunicate IL-1-like molecules reported by Kelly et at. (1993) is in agreement with this view. Acknowledgments The authors wish to thank Mr. M. Del Favero, Mr. R. Mazzaro, and Mr. C. Friso for their technical assistance. This work was supported by a grant from the University of Padova to one of us (L.B.). Literature Cited Abbas, A. K., A. H. Lichtman, and J. S. Pober. 1991. Cellular and Molecular Immunology. W. B. Saunders, Philadelphia. Anderson. R. S. 1971. Cellular responses to foreign bodies in the tuni- cate Molgula manhatlensis (DeKay). Biol Bull. 141: 91-98. Andrew, \V. 1961. Phase microscope studies of living blood-cells of the tunicates under normal and experimental conditions, with a description of a new type of motile cell appendage. Q. J. Microsc. Sci. 102: 89-105. Ballarin. I... F. Cima, and A. Sabbadin. 1994. Phagocytosis in the colonial ascidian Botryllus schlosseri. Dev. Comp. Immunol. 18: 467- 481. Ballarin. L., F. Cima. and A. Sabbadin. 1995. Morula cells and histo- compatibility in the colonial ascidian Botryllus schlosseri. Zoo/. Sci. 12: 757-764. Ballarin, L., F. Cima, and A. Sabhadin. 1997. Calcium homeostasis and yeast phagocytosis in the compound ascidian Botryllus schlosseri. Comp. Biochem. Physiol. 118A: 153-158. Ballarin, L., F. Cima, and A. Sabbadin. 1998. Phenoloxidase and cytotoxicity in the compound ascidian Botryllus schlosseri. Dev. Comp. Immunol. 22: 479-492. Ballarin, L.. C. Tonello, L. Guidolin, and A. Sabbadin. 1999. Purifi- cation and characterization of a humoral opsonin. with specificity for D-galactose, in the colonial ascidian Botryllus schlosseri. Comp. Bio- chem. Phy.iiol. 123B: 115-123. Ballarin, L., C. Tonello, and A. Sabbadin. 2000. Humoral opsonin from the colonial ascidian Botryllus schlosseri as a member of the galectin family. Mar. Biol. 136: 823-827. Beck. G., and G. S. Habicht. 1991. Primitive cytokines: harbingers of vertebrate defense. Immunol. Toda\ 12: 180-183. Beck, G., G. R. Vasta, J. J. Marchalonis, and G. S. Habicht. 1989. Characterization of interleukin- 1 activity in tunicates. Comp. Biochem. Physiol. 92B: 93-98. Beck, G., R. F. O'Brien, G. S. Habicht, D. L. Stillman, E. L. Cooper, and D. A. Raftos. 1993. Invertebrate cytokines III: Invertebrate interleukin- 1 -like molecules stimulate phagocytosis by tunicate and echinoderm cells. Cell. Immunol. 146: 284-299. Cammarata. M., V. Arizza, N. Parrinello, G. Candore, and C. Caruso. 1997. Phenoloxidase dependent cytotoxic mechanism in ascidian (Sryela plicata) hemocytes against erythrocyte and K562 tumor cells. Eur. J. Cell. Biol. 74: 302-307. Chaga, O. Y. 1980. Ortho-diphenoloxidase system of ascidians. Tsi- tologiya 22: 619-625 (in Russian). Cima, F., L. Ballarin, and A. Sabbadin. 1996. New data on phagocytes and phagocytosis in the compound ascidian Botryllus schlosseri (Tu- nicata: Ascidiacea). Ital. J. Zoo/. 63: 357-364. Coombe, D. R., P. L. Ey, and C. R. Jenkin. 1984. Particle recognition by haemocytes from the colonial ascidian Botrylloides leachi: evidence that the B. leachi HA-2 agglutinin is opsonic. J. Comp. Phvsiol. 154: 509-521. Dan-Sohkawa, M., M. Morimoto, and H. Kaneko. 1995. In vitro reactions of coelomocytes against sheep red blood cells in the solitary ascidian Halocynthia roret-i. Zoo/. Sci. 12: 411 117. De Leo, G.. N. Parrinello, D. Parrinello. G. Cassara, and M. A. Di Bella. 1996. Encapsulation response of dona intestinalis (Ascidia- cea) to intratunical erythrocyte injection. I. The inner capsular archi- tecture. J. Invertebr. Pcithol. 67: 205-212. Endean, R. 1955a. Studies of the blood and tests of some Australian ascidians. I. The blood of Pyitrci stolnnifera (Heller). Aust. J. Mar. Freshwater Res. 6: 35-59. Endean, R. 1955b. Studies of the blood and tests of some Australian ascidians. III. The formation of the test of Pyura stolonifera (Hellerl. Aust. J. Mar. Freshwater Res. 6: 157-164. Franchini, A.. J. A. Mivan. and E. Ottaviani. 1996. Induction of ACTH- and TNF-a-like molecules in the hemocytes of Calliphora vomitoria (Insecta. Diptera). Tissue Cell 28: 587-592. Fuke, M., and M. Fukumoto. 1993. Correlative fine structural, behav- ioral, and histochemical analysis of ascidian blood cells. Ada Zoo/. IStockh.) 74: 61-71. Gorman. A., J. McCarthy, D. Finucane. \V. Reville, and T. Cotter. 1996. Morphological assessment of apoptosis. Pp. 1-20 in Tech- niques in Apoptosis: A User's Guide. T. G. Cotter and S. J. Martin, eds. Portland Press, London. Hsu, S. M., L. Raine, and H. Fanger. 1981. Use of avidin-biotin- peroxidase complex (ABC) immunoperoxidase techniques: a compar- ison between ABC and unlabeled antibody (PAP) procedures. J. His- tochem. C\tochem. 29: 577-580. 64 L. BALLARIN ET AL. Hughes, T. K., E. M. Smith, R. Chin, P. Cadet, J. Sinisterra, M. K. Leung, M. A. Shipp, B. Scharrer, and G. B. Stefano. 1990. Interactions of immunoreactive monokines (interleukine-1 and tumor necrosis factor) in the bivalve mollusc Mytilu.i ednlis. Proc. Natl. Acail. Sci. USA 87: 4426-4429. Kelly, K. L., E. L. Cooper, and D. A. Raftos. 1992. Purification and characterization of a humoral opsonin from the solitary urochordate Styela clava. Comp. Biochem. Physiol. 103B: 749-753. Kelly, K. L., E. L. Cooper, and D. A. Raftos. 1993. Cytokine-like activity of a humoral opsonin from the solitary urochordate Sryela clava. Zoo/. Sci. 10: 57-64. Kustin, K., D. S. Levine, G. C. McLeod, and W. A. Curby. 1976. The blood of Ascidia nigra: blood cell frequency distribution, morphology, and the distribution and valence of vanadium in living blood cells. Bio/. Bull. 150: 426-441. Ohtake, S.-I., T. Abe, F. Shishikura, and K. Tanaka. 1994. The phagocytes in hemolymph of Halncynthia roretzi and their phagocytic activity. Zoo/. Sri. 11: 681-691. Ottaviani, E., F. Petraglia, G. IVlontagnani, A. Cossarizza, D. Monti, and C. Franceschi. 1990. Presence of ACTH and /3-endorphin im- munoreactive molecules in the freshwater snail Planorbarius corneus (L.) (Gastropoda. Pulmonatal and their possible role in phagocytosis. RcKiil. Pent. 27: 1-9. Ottaviani, E., E. Caselgrandi, and C. Franceschi. 1995a. Cytokines and evolution: in vitro effects of IL-la. IL-1/3. TNF-a and TNF-/3 on an ancestral type of stress response. Biochem. Biophys. Res. Commun. 207: 288-292. Ottaviani, E., A. Franchini, S. Cassanelli, and S. Genedani. 19951). Cytokines and invertebrate immune responses. Biol. Cell. 85: 87-91. Ottaviani, E., A. Franchini, D. Kletsas, and C. Franceschi. 1996. Presence and role of Cytokines and growth factors in invertebrates, /fa/. J. Zoo/. 63: 317-323. Parrinello, N. 1996. Cytotoxic activity of tunicate hemocytes. Pp. 190- 217 in Invertebrate Immunology. B. Rinkevich and W. E. G. Muller, eds. Springer-Verlag. Berlin. Raftos, D. A., E. L. Cooper, G. S. Habicht, and G. Beck. 1991. Invertebrate cytokines: tunicate cell proliferation stimulated by an interleukin l-like molecule. Proc. Natl. Acad. Sci. USA 88: 9518- 9522. Raftos, D. A., E. L. Cooper, D. L. Stillman, G. S. Habicht, and G. Beck. 1992. Invertebrate cytokines II: release of interleukin-1-like mole- cules from tunicate hemocytes stimulated with zymosan. Lymphakine Cytokine Rex. 11: 235-240. Raftos, D. A., D. L. Stillman, and E. L. Cooper. 1998. Chemotactic responses of tunicate ( Urochordata. Ascidiacea) hemocytes in vi'fro. J. Invertebr. Patliol. 72: 44-49. Smith, M. J. 1970. The blood cells and tunic of the ascidian Halocynthia auranlium (Pallas). I. Hematology, tunic morphology, and partition of cells between blood and tunic. Biol. Bull. 138: 354-378. Smith, V. J., and C. M. Peddie. 1992. Cell cooperation during host defense in the solitary tunicate Ciona intestinalis (L). Biol. Bull. 183: 211-219. Wolfe, S. L. 1993. Molecular and Cellular Biology. Wadsworth, Bel- mont, CA. Wright, R. K. 1981. Urochordates. Pp. 565-626 in Invertebrate Blood Cells. Vol. 2. N. A. Ratcliffe and A. F. Rowley, eds. Academic Press. London. Zaniolo, G. 1981. Histology of the ascidian Botryllus schlosseri tunic: in particular, the test cells. Boll. Zoo/. 48: 169-178. Reference: Bio/. Bull. 201: 65-75. (August 2001) Molecular Evidence that Sclerolinum brattstromi Is Closely Related to Vestimentiferans, not to Frenulate Pogonophorans (Siboglinidae, Annelida) KENNETH M. HALANYCH 1 *. ROBERT A. FELDMAN 2 , AND ROBERT C. VRIJENHOEK 1 1 Biology Department MS 33, Woods Hole Oceanographic Institution. Woods Hole, Massachusetts 02543; : Molecular Dynamics, Inc.. part of Amersham Pharmacia Biotech, 928 East Arques Ave., Sunnyvale, California 94086-4250; and 3 Monterey Bay Aquarium Research Institute. 7700 Sandholdt Road. Moss Landing, California 95039 Abstract. Siboglinids. previously referred to as pogono- phorans, have typically been divided into two groups, frenu- lates and vestimentiferans. Adults of these marine proto- stome worms lack a functional gut and harbor endosymbiotic bacteria. Frenulates usually live in deep, sedimented reducing environments, and vestimentiferans inhabit hydrothermal vents and sulfide-rich hydrocarbon seeps. Taxonomic literature has often treated frenulates and vestimentiferans as sister taxa. Sclerolinum has traditionally been thought to be a basal siboglinid that was originally regarded as a frenulate and later as a third lineage of siboglinids. Monilifera. Evidence from the 18S nuclear rDNA gene and the 16S mitochondria! rDNA gene pre- sented here shows that Sclerolinum is the sister clade to vestimentiferans although it lacks the characteristic mor- phology (i.e.. a vestimentum). The rDNA data confirm the contention that Sclerolinum is different from frenulates, and further supports the idea that siboglinid evolution has been driven by a trend toward increased habitat specialization. The evidence now available indicates that vestimentiferans lack the molecular diversity expected of a group that has been argued to have Silurian or possibly Cambrian origins. Introduction Siboglinids were formerly called pogonophorans and in- clude two groups of marine protostomes, frenulates and vestimentiferans, that are commonly referred to as beard- Received 22 November 2000: accepted 1 1 April 2001. * To whom correspondence should be addressed. E-mail: khalanvch@ whoi.edu worms and tubeworms, respectively. Both groups lack a functional gut as adults and rely on endosymbiotic bacteria for nutrition. They have a closed circulatory system and possess a metamerized tail region called the opisthosoma. Vestimentiferans are distinguished from frenulates by the presence of a vestimentum, a winged region near the ante- rior of the organism. Both taxa occur in reducing environ- ments and typically are found at depths below several hundred meters. Due to the limited availability of samples and the difficulty of retrieving live specimens, several as- pects of their biology (e.g.. reproduction, physiology) are still poorly understood. Vestimentiferans, in general, have been better studied than frenulates because they are key- stone species in eastern Pacific hydrothermal vent habitats and in Pacific and Caribbean seeps. The taxonomic literature concerning frenulate and vesti- mentiferan siboglinids has a colorful and confusing history. One taxonomic scheme recognizes frenulates (aka pogono- phorans sensu stricto) and vestimentiferans as distinct phyla (Jones, 1985). Alternatively, vestimentiferans have also been recognized as a class within the phylum Pogonophora (Jones, 1981; Ivanov, 1994). Others place frenulates and vestimentiferans within the phylum Annelida (Land and N0rrevang, 1977; Kojima et al.. 1993; Bartolomaeus. 1995; McHugh, 1997; Rouse and Fauchald. 1997; also see South- ward, 1988). The latter hypothesis has been supported by recent morphological (Rouse and Fauchald. 1995. 1997). embryological (Young et al., 1996; Southward, 1999), and molecular analyses (Kojima et al., 1993: McHugh, 1997; Blacker*-//., 1997; Kojima, 1998; Halanych et al.. 1998). To further complicate matters, a ranked classification scheme has produced different names for the same clade of organ- 65 66 K. M. HALANYCH. R. A. FELDMAN. AND R. C. VRUENHOEK isms. Vestimentiferans have been called Vestimentifera (Jones, 1981). Obturata (Jones, 1981; Southward, 1988; Southward and Galkin, 1997). and Afrenulata (Webb. 1969). Frenulates have been called Pogonophora (Jones. 1985), Frenulata (Webb, 1969), Perviata (Southward, 1988), and originally Siboglinidae (Caullery, 1914). Hereafter we apply the following nomenclature: (1) Ves- timentifera are equated with Obturata and Afrenulata; (2) Frenulata are equated with Perviata and Pogonophora (sensu Jones, 1985); (3) Monilifera is a third monogeneric clade that includes Sclerolinum; and (4) Siboglinidae refers to the clade that includes Vestimentifera, Frenulata, and Monilifera. We recognize that the term "Pogonophora" is more commonly used and that rules of priority for nomen- clature do not apply to higher taxa. However, we have opted to use the term "Siboglinidae" throughout this manuscript to emphasize that this group of organisms represents derived annelids (McHugh, 1997; Rouse and Fauchald. 1997). We restrict the term "pogonophoran" to common usage. Even among siboglinids, there has been one group. Sclerolinum, that has been particularly problematic in terms of phylogenetic position. Unlike most frenulates that live in the mud, Sclerolinum species can live on decaying organic material like wood or rope made from natural fibers (Webb, 1964a; Southward, 1972). This taxon was originally con- sidered a member of the frenulate family Polybrachiidue (Southward. 1961 ). but Webb ( 1964b), mainly citing differ- ences in the postannular region, argued that Sclerolinum could not be ascribed to either of the two orders (Theca- nephria and Athecanephria) of siboglinids recognized at the time (vestimentiferans had not been discovered yet). He erected a new family, Sclerolinidae, that he states should "have order rank." Ivanov ( 1991 ) more formally recognized the unique nature of Sclerolinum, and in 1 994 he proposed that Frenulata (= Perviata), Monilifera (= Sclerolinidae), and the Vestimentifera be regarded as three taxa with equal rank (i.e.. classes within the phylum Pogonophora). Addi- tionally, Ivanov ( 1994) further suggested that Monilifera are allied to the Vestimentifera on the basis of the common absence of several characters (e.g.. spermatophores, teloso- mal diaphragm, metasoma preannular and postannular re- gions) relative to the Frenulata. Southward (1999) sug- gested that Monilifera might be similar to the ancestral siboglinid form, thus predicting that it should occupy a basal position in siboglinid phylogeny. Distinguishing between these hypotheses on the placement of Sclerolinum will allow us to test the notion of Black el al. (1997) that habitat preference or specificity may be an important factor in siboglinid evolution. If Black et al. are correct, Sclerolinum is expected to occupy a position between frenulates and vestimentiferans (which may be consistent with Ivanov's ideas), and not a position basal to the frenulate-vestimen- tiferan clade. To date, molecular studies that include siboglinids have either focused on vestimentiferans (Williams et al.. 1993; Black et al.. 1997; Kojima et al., 1997; Halanych et al., 1998) or have addressed siboglinid origins (Winnepen- ninckx et al., 1995a; Kojima et al., 1993; Kojima, 1998; McHugh, 1997). Most studies have included only one frenu- late representative. Although Black et al. (1997) included two "frenulate" siboglinids, one of these, the Loihi worm, was undescribed. Additionally, several 18S sequences were reported in a symposium contribution (Halanych et al., 1998) for which page limitations did not permit detailed analyses or explanation. Herein we extend these previous analyses by increasing the sampling of frenulates. including Sclerolinum, and using novel 18S rDNA and 16S rDNA data. The present findings support the notion that habitat requirements have been important in siboglinid evolution. Additionally, frenulates are sister to a Sclerolinum-vesti- mentiferan clade, the latter of which showed limited diver- sity suggestive of a recent radiation within the clade. Materials and Methods Taxa employed Table 1 lists the species analyzed and GenBank accession numbers for the rDNA sequences used in this study. The frenulate and vestimentiferan operational taxonomic units (OTUs) included in this study represent all of the currently recognized genera available to the authors. The addition of closely related species within a genus would have increased OTUs without increasing the phylogenetic signal for the issues under examination and were therefore excluded. For example, there are no nucleotide differences observed in the 18S rDNA of Escarpia spicata (Guaymas Basin) and E. laminata (Florida Escarpment). Limiting the number of OTUs also reduced computation time, allowing for more thorough analyses. Unless otherwise noted, collection local- ities correspond to those given in Black et al. (1997). Siboglinum ekmani, S. fiordicum. and Sclerolinum brattstromi were collected near Bergen, Norway, and iden- tified by Eve Southward. Marine Biological Association of the United Kingdom. Identification of the frenulates Spiro- brachia and Polybrachia were made by Eve Southward on the basis of animal and tube morphology. Both specimens were collected by TVGrab from the Aleutian Trench (5727.394'N, 14800.013'W) at a depth of 4890 m on the German research vessel Sonne. The non-siboglinid annelid OTUs for the 18S data were chosen to represent a diversity of lineages for which se- quences were available. The arthropod (Anemia) sequence was designated as the most distant outgroup for rooting purposes. Based on both morphology (e.g., Eernisse et al., 1992) and molecular studies (e.g.. Halanych et al.. 1995; Winnepenninckx et al., 1995a; Aguinaldo et al., 1997; Eernisse, 1997), arthropods are clearly outside of the proto- Timi used in rDNA anal\ses SIBOGLINID EVOLUTIONARY HISTORY TABLE 1 67 Organism GenBank Accession' 1 GenBank Accession 11 18S rDNA I6S rDNA Organism 18S rDNA 16S rDNA Pogonophora Frenulata Galalheiiliiuiin brachiosum AF168738 Polybrachia sp. AF 168739 Siboglinum fiordicinn GB X79876 h Siboglinum fiordicwn AF3 15060 Siboglinum ekmani AF3 15062 Spirobriit-liiti sp. AF 168740 Vestimentifera Escarpia spicata AF 168 741 Escarpiid n. sp. Lumellihriichia barhami AF168742 Oiisisia alvinae AF168743 Ridgeia piscesae AF 168 744 Ridgeia piscesae GB X79877 h Chaetopterida Chaetopterus variopedatus U67324 C AF3 15040 Hirudinea AF3 15037 Haemopis sanguisuga X91401 J Hinulu nit'dk-iniilix AF3 15058 AF3 15039 Oligochaete AF3 15038 Enchytraeus sp. Z83750 d AF315036 Phyllodocida Glycera americiina U19519 e AF3 15041 Polynoidea AF3 1 5053 Lepidonotopodium fimbriutum AF3 1 5056 AF3 15043 Branchipolynoe symmytilida AF3 15055 AF3 15044 Sabellida AF315045 Sabella piminniti U67144' AF3 15047 Tubificidae AF3 15052 Tubifex sp. AF3 15057 AF3 15048 Echiura AF3 15051 Ochetostoma erythrogrammon X79875 h AF3 15054 Urechis sp. AF3 15059 Sipuncula Riftia pachyptila AF 168745 AF3 15049 Phascolosoma gnnuilaiiini X79874 b AF3 15050 Nemertea Tevnia jerichonana AF168746 AF3 15042 Linens sp. X79878" Monilifera Mollusc Sclerolinum brattstromi AF3 15061 AF3 15046 Scutopux ventrolineatus X91977' Annelida Priapulida Alvinellidae Priiipuliix caudatus X80234 1 Puralvinella pabniformis AF 168747 Arthropod Anemia salina X01723 h a Unless otherwise noted, sequences were obtained in this study. b Sequence from Winnepenninckx et al. ( 1995a). c Sequence from Nadot and Grant (unpublished). J Sequence from Kim el al. ( 1996). Sequence from Halanych et al. ( 1995). ' Sequence from Winnepenninckx et al. ( 1996). g Sequence from Winnepenninckx et al. ( 1995b). h Sequence from Nelles et al. ( 1984). stome worm radiation. Because siboglinids are not closely related to molluscs and because of rate heterogeneity prob- lems within the Mollusca, only a single representative (the aplacophoran Scutopus) was used. Due to alignment limi- tations, outgroups employed in the 16S analyses a leech, an oligochaete, two polynoid polychaetes, and an echiu- rid were more limited (see Table 1). Because different investigators collected the data at different times, there was not a 1:1 correspondence in OTUs between data sets. We felt it more important to present all the relevant data rather than trim taxa from the data sets. The aligned data sets are available at the journal's Supplement's page (http:// www.mbl.edu/BiologicalBulletin/VIDEO/BB.video.html) and at TREEBASE (http://phylogeny.harvard.edu/treebase). Data collection Total genomic DNA was extracted using a modified hexadecyl-trimethyl-ammonium bromide (CTAB) protocol (Doyle and Dickson. 1987). The entire 18S nuclear rDNA gene was amplified via PCR (polymerase chain reaction), using the universal metazoan oligonucleotide primers 18e and 18P (Halanych et al.. 1998). A region of the 16S mitochondria! rDNA was amplified using 16Sar-5' and 16Sbr-3' primers (Palumbi, 1996). Each 50 /xl reaction consisted of about 50 ng of template DNA, 0.5 /u,A/ of each primer, 2.5 mM MgCl 2 , 200 pM dNTPs, 5 ju.1 of manufac- turer's 10X reaction buffer, and 1.5 U Tag polymerase (Promega Inc.. Wisconsin). Cycling profiles were as fol- 68 K. M. HALANYCH. R. A. FELDMAN, AND R. C. VRIJENHOEK lows: 18S initial denaturation at 95 C for 3 min, 35 cycles of amplification (denaturation at 95 C for 1 min, annealing at 50 C for 2 min, extension at 72 C for 2 min 30 s), and a final extension at 72 C for 5 min: 16S initial denaturation at 94 C for 2 min, 40 cycles of amplification (denaturation at 94 C for 30 s, annealing at 46 C for 30 s, extension at 72 C for 1 min), and a final extension at 72 C for 7 min. PCR products were purified using the QIAEX II gel extraction kit (Qiagen Inc., California). Approximately 60 ng of purified PCR product was used in sequencing reactions according to the manufacturer's instructions (FS Dye Termination Mix or Big Dye, Applied Biosystems Inc., California). The reaction profile was 25 repetitions of de- naturation at 94 C for 30 s, annealing at 50 C for 15 s, and extension at 64 C for 4 min. Dye-labeled fragments were separated by electrophoresis on a Perkin Elmer ABI 373A or 377 DNA sequencer. Both strands of the PCR product were sequenced. In addition to the PCR primers, the oligo- nucleotide primers used for sequencing are given in Halanych el al. (1998) or Hillis and Dixon (1991). The sequences were assembled and verified using the AutoAs- sembler and Sequence Navigator programs (Applied Bio- systems Inc., California). The terminal primer regions were not included in the sequences submitted to GenBank or in the phylogenetic analyses. Phylogenelic analyses Sequence alignment was produced with a Clustal W program (Thompson el al., 1994) and subsequently cor- rected by hand using the protostome secondary structure models available through the Ribosomal Database project (http://rdp.cme.msu.edu/html/). Regions that could not be unambiguously aligned (e.g., divergent loop domains) were excluded from analyses. Tree reconstructions were imple- mented with the PAUP* 4.0b4b2 program (Swofford, 2000), and MacClade 3.06 (Maddison and Maddison, 1992) was used for character and tree analyses. Neighbor-joining (NJ), parsimony, and maximum likelihood (ML) analyses were performed and yielded similar results. In the interest of brevity, results and discussion will focus on ML analyses. NJ trees were reconstructed under Jukes-Cantor, Kimura- 2-parameter, Tamura-Nei. general-time-reversible, and log/ det models. All except log/det were examined under equal rates of among-site rate variation using the empirically derived gamma shape parameter, a, of 0.3 (see Swofford el al.. 1996, for summary of different assumptions used in these models). A Kishino-Hasegawa ( 1989) likelihood eval- uation of the resulting topologies revealed no significant differences between models for either the 16S or the 18S data. Kishino-Hasegawa evaluations estimated a six-substi- tution-type rate matrix for which nucleotide base frequen- cies were set to empirical values and a was estimated. NJ bootstraps consisted of a log/det correction (model was arbitrarily chosen) with 1000 iterations. Parsimony analyses consisted of heuristic searches with 100 random sequence additions and tree-bisection-reconnection (TBR) branch swapping. Transitions (Ti) and transversions (Tv) were given equal weighting. ML evaluation of parsimony topol- ogies was the same as for NJ topologies. One thousand iterations were used for parsimony bootstrap analyses. When using likelihood to search for the "best" tree (as opposed to evaluating given trees), computation time was limiting. Therefore, we used a nucleotide model with two substitution types where the Ti/Tv ratio was set to the value estimated for the best parsimony tree (empirical base fre- quencies were used). ML searches were heuristic with 10 random sequence-addition replicates. ML bootstraps em- ployed the "Faststep" option with 100 iterations. Results The 18S rDNA data set consisted of 26 OTUs and 1935 nucleotide positions. Of the 1614 nucleotide positions that could be unambiguously aligned. 34.6% (559 positions) were variable and 18.7% (303 positions) were parsimony informative. Figure 1 shows the single best likelihood tree (Ln likelihood = -8260.55148) recovered. All search methods in all analyses found a monophyletic siboglinid clade (bootstrap support was >98% for all methods). Res- olution within the vestimentiferan clade, as well as between annelid groups, was poor, however. The moniliferan Sclerolinum brallslromi falls out with the vestimentiferan taxa in all analyses (bootstrap S 98%). The remaining trenulates form a distinct sister-clade to the Sclerolinwn- vestimentiferan clade with >99%> bootstrap support. Resolution among annelid taxa and within the vestimen- tiferans was poor due to the lack of phylogenetic signal. Because this paper does not focus on the annelid radiation, we did not try to enhance resolution among all annelid taxa. However, we did attempt to boost the signal within the vestimentiferan clade by employing a less inclusive taxo- nomic alignment. For metazoan 18S sequences, inclusion of broader taxonomic diversity can often create larger regions of ambiguous alignment that should not be included in analyses, due to poor assumptions about positional homol- ogy. Thus by reducing the taxonomic breadth examined, the phylogenetic signal can potentially be increased by a "bet- ter" alignment (Halanych, 1998). Unfortunately, even when just the siboglinids were aligned, little genetic diversity was observed, and the vestimentiferan taxa were still poorly resolved (not shown). The exception was Lamellibrachia harhami, which was consistently placed as the most basal vestimentiferan. Table 2 shows the logdet/paralinear dis- tances (below diagonal) and absolute distances (above di- agonal) for this less-inclusive, siboglinid-only alignment (in which most divergent domains could be unambiguously aligned). Even though the distance values for the siboglinid- SIBOGLINID EVOLUTIONARY HISTORY 69 e; 99 100 100 1 100 ipirooracnia I Polybrachia _l g1 Galathealmum ^ Siboglinumekmani 100 r Siboghnum fiordicum GB Siboglinum fiordicum 86 58 (D 96 59 Escarpia Ridgeia RidgeiaGB Oasisia Riftia Tevnia Lamellibrachia Sderolinum ^ Enchytraeus -Oligochaete ^^^^^^ ^^ Haemopis -Leech Moniliferan Sabella - Polychaete Paralvinella - Polychaete Phascolosoma -Sipunculid Ochetostoma - Echiund Chaetopterus - Polychaete Glycera Polychaete ^^ Lineus Nemertean Scutopus Mollusk Artemia - Arthropod Priapulus- Priapulid 0-01 substitutions/site Figure 1. Results of 18S rDNA phylogenetic analyses. The single best likelihood tree (Ln likelihood = 8260.551481 found. Analysis details are given in the text. Maximum likelihood bootstrap values of >50% are given in bold. Parsimony (italicl and neighbor joining (underlined) values are also given for the major nodes of interest (values for other nodes were omitted in the interest of space). Branch lengths are drawn proportional to the inferred amount of change along the branch (scale shown). only alignment are only slightly greater than the full align- ment values, the greatest distance within vestimentiferans was only 0.02 (with a maximum of 25 nucleotide differ- ences), revealing that there was very little 18S genetic diversity within this group. The 16S rDNA data set consisted of 24 OTUs, each with 497 nucleotide positions. Of the 465 nucleotide positions that could be unambiguously aligned, 60.4% (281 positions) were variable and 47.7% (222 positions) were parsimony informative. The reconstructed topology (Ln likelihood = -3967.21062). Figure 2, was qualitatively similar to the 18S topology. Siboglinids are divided into two major lin- eages: vestimentiferans plus the moniliferan Sderolinum brattstromi (bootstrap support 83% for ML and 100% for NJ and parsimony) and a frenulate sister-clade (bootstrap support >94% in all analyses). Again. 5 1 . brattstromi was basal to the vestimentiferans. In a departure from the 18S analyses, Riftia pachyptila, not Lamellibrachia barhami. often fell out as the most basal vestimentiferan. However, this was never supported by >54% bootstrap support; ML analyses that excluded the non-siboglinid outgroups re- vealed that the base of the Vestimentifera was poorly re- solved with 16S data. A comparison of genetic divergence values (Table 3) indicates that there was limited genetic 70 K. M. HALANYCH. R. A. FELDMAN, AND R. C. VRIJENHOEK TABLE 2 Paim'ise distances for the siboglinid-only IKS rDNA data set: absolute distances above diagonal and log/det distances below diagonal 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1 Spirobrachia _ 109 113 87 132 131 124 122 125 124 131 125 121 120 2 Polybrachia 0.07 9 104 138 137 139 140 140 140 147 142 140 136 3 Galathealinum 0.07 0.01 106 142 141 142 143 143 143 150 145 143 139 4 Siboglinum ekniani 0.05 0.06 0.06 116 117 113 1 III 113 110 121 112 107 112 5 Siboglinum fiordicum 0.08 0.09 0.09 0.07 5 14(1 136 136 140 143 138 134 139 6 Siboglinum fiordicum GB 0.08 0.08 0.09 0.07 0.00 143 139 139 143 146 141 137 140 7 Escarpia 0.08 0.09 0.09 0.07 0.09 0.09 7 14 10 19 6 13 31 8 Ridgeia 0.08 0.09 0.09 0.07 0.09 0.09 0.00 8 7 17 4 12 32 9 Ridgeia GB 0.08 0.09 0.09 0.07 0.09 0.09 0.01 0.00 14 21 11 19 38 10 Oasisia 0.08 0.09 0.09 0.07 0.09 0.09 0.01 0.00 0.01 20 7 14 32 1 1 Riftia 0.08 0.09 0.09 0.07 0.09 0.09 0.01 0.01 0.01 0.01 16 25 39 12 Tevnia 0.08 0.09 0.09 0.07 0.09 0.09 0.00 0.00 0.01 0.00 0.01 1 1 30 1 3 Lamellibrachia 0.07 0.09 0.09 0.07 0.08 0.09 0.01 0.01 0.01 0.01 0.01 0.01 28 14 Sclerolintiin 0.08 0.09 0.09 0.07 0.09 0.09 0.02 0.02 0.02 0.02 0.02 0.02 0.02 variation within vestimentiferans (<0. 1 1 log/del distance; a maximum of 47 nucleotide differences). As for the frenulate clade, neither 18S or 16S supported a monophyletic Siboglinum: but because only two Siboglinuin species were examined, additional taxa are needed to verify the status of this frenulate taxon. Additionally, we performed Kishino-Hasegawa ( 1989) likelihood evaluation for both genes to test the monophyly of the frenulate and vestimentiferan- Sclerolinum clades. To this end, we used the constraints option in PAUP* 4.0b4b2 to conduct parsimony heuristic searches (specifics same as above) to find the best trees that were consistent and inconsistent with the monophyly of these clades. Both the 16S and the 18S data significantly support the monophyly of both groups ( 1 8S frenulates average ML score supporting monophyly = -8244.69, non-monophyly score = -8278.135. P value < 0.01; 16S frenulates monophyly = -3894.889. non-monophyly = -3927.49, P value < 0.005; 18S vestimentiferan-Sc/ero/znMW! monophyly = 8244.69, non-monophyly = -8271.922, P value < 0.05; 16S vestimen- tiferan-Sclerolinum monophyly = 3894.889, non-mono- phyly = -391 1.802, P value < 0.05). Discussion The monophyly of siboglinids (aka, Pogonophora sensu hit 11) is supported by morphological (Southward, 1988, 1993; Rouse and Fauchald, 1995; Rouse. 2001). embryo- logical (Southward, 1999), and molecular (Winnepenninckx et ui. 1995a; Black el ai. 1997; McHugh. 1997; Halanych c/ ai. 1998. this study) evidence. Thus, in agreement with others (Southward, 1988, 1999; Ivanov, 1994; McHugh, 1997), we see no support for the recognition of vestimen- tiferuns and frenulates as having fundamentally different body plans (i.e.. "phyla" sensu Jones. 1985). The assertion made by Webb !l964b) and later by Ivanov (1991. 1994) that Sclerolinitm was notably different from frenulates is validated by the present data. Moreover, we found that Sclerolinitm brtittstroini is closely allied to the vestimenti- ferans. and does not occupy a position basal to a frenulate- vestimentiferan clade, confirming Ivanov's (1991; 1994; Ivanov and Selivanova, 1992) ideas that moniliferans oc- cupy a position intermediate between vestimentiferans and frenulates. Southward (1993) also suggested a possible evolutionary link between Sclerolinum and vestimentiferans. This con- tention is confirmed by the present analysis, as well as a recent morphological cladistic analysis (Rouse, 2001). Us- ing 44 morphological characters coded for all recognized siboglinid genera. Rouse found support for the monophyly of Frenulata. Vestimentifera, and the Sclerolinum-vestimen- tiferan clade. However, our use of nomenclature differs from Rouse with regard to the term Monilifera, which he applies to the Sr/e>w/i'w-vestimentiferan clade. Because this term was originally (Ivanov and Selivanova, 1992) applied to only Sclerolinitm, and because of the morpho- logical differences from vestimentiferans. Rouse's use of the term will inject confusion into the literature. Although we acknowledge that Monilifera, as defined here, is redun- dant with the generic name Sclerolinum, several aspects of siboglinid evolution and taxonomy are in need of additional study. Thus, we have chosen not to name this clade until more is understood about siboglinid evolution. The placement of Sclerolinum was especially interesting in the context of the evolution of habitat preference. Previ- ous studies of vestimentiferans (Black et ui. 1997). clams (Peek et til.. 1997), mussels (Craddock et ai. 1995). and shrimp (Shank et ai, 1999) reveal that vent-endemic organ- isms are related to. and possibly derived from, species associated with hydrocarbon seeps that occur near subduc- tion zones and continental margins. Furthermore, recent observations (Feldman et ai. 1998; Baco et ai. 1999; Distel SIBOGLINID EVOLUTIONARY HISTORY 71 94 e: i Spirobrachia ! 100 r Polybrachia cum * Galathealmum 700 100 I 79 1 Escarpia 1- Escarpiid n. sp. 1- - Tevnia Ridgeia 1 \ 75 ' - Ridgeia 2 53 Ridgeia 3 / 80 , ' 93 56 Oasisia i 80 r Lamellibrachia 1 1 *~ Lamellibrachia 2 5 r 67 Lamellibrachia 3 (0 no Lamellibrachia 4 83 1 99 I- Riftia 1 ; iXY 1 " Riftia2 1 'II Monilife*' 01 " 1 100 1 Lepidonotopodium - Polychaete 61 ifex - Oligochaete (D I (0 ^^ 0.05 substitutions/site Figure 2. Results of 16S rDNA phylogenetic analyses. The best likelihood tree (Ln likelihood = -3967.21062) found. Another tree with a Ln likelihood score of -3967.25739 was found in the same search. The trees differed in relationships within the Ridgeia clade. Analysis details are given in the text. Maximum likelihood (ML) bootstrap values of >50% are given in bold. Parsimony (italic) and neighbor joining (under- lines) values are also given for the major nodes of interest (values for other nodes were omitted in the interest of space). In the ML bootstrap analysis, Lamellibrachia and Sclerolinum formed a clade in 55% of the iterations. That is not shown above because it is incompatible with the "best" ML tree. Branch lengths are drawn proportional to the inferred amount of change along the branch (scale shown). c/ ill.. 2000) reveal that several symbiont-bearing clams, vestimentiferan tubeworms, and mussels can survive on rotting organic material, such as wood or a whale carcass. The moniliferan S. brattstromi and related species (e.g., S. javanicum, S. minor, and 5. major) are typically found growing on decaying organic material such as wood or rope (Webb, 1964a, b; Southward, 1972; Ivanov and Selivanova, 1992). Other members of the genus, (e.g., S. sibogae and S. magdalenae) lived buried in mud (Southward, 1972). These habitat preferences suggest that affinity for a mud or silt habitat was ancestral in siboglinids, allowing us to speculate that a pattern of evolution from low-oxygen, sedimented habitats to decaying organic material to hydrocarbon seeps to hydrothermal vents has occurred within the Sclerolinum- vestimentiferan clade. Although neither the 18S nor the 16S data clearly resolve relationships within the Vestimentifera, the cytochrome c oxidase subunit I (COD data of Black et al. (1997) show 72 K. M. HALANYCH, R. A. FELDMAN, AND R. C. VRIJENHOEK -t -t r- I- so CN m -^ . -f -t rr . ~ P*-, pr, P*-, -t pr, P*-, - 3 OC O OO ON - t- -t >r, -f -f i/ r*~i ON sO C ?S | r-i - pr, \G P - r-i ON m OsC-fsI 3 OC OC vC ^C - ON -^- OC C N 1 d ri -t r -t r-i c: C 3 I/-, _ ON in r* NO VO r*~i t d d '/-, - HI I = ,___ ~ sC oc \o r-j p Oi - CS | C 3 IO rl 1 r*-i ^f 3 d O ~l ON ^ ri r ; -t oc so si 3 Tj- r-j rr, p- -r p*-. r - p- O oc ri C NO | d C 3 -f rn 3 d d ON ~ r i doc 3 NO f) 3 d d >c ?! c - ri ri ri i/ ; ri ci O P d d d c 1 sC OC 3 d d oc N g -' ^ ^ S 3 d d d c 3 NO 1^ 3 d d -t -r -t r^ > 3 rj ON rj- C 3 ifi v~. l^ O N P- in ON - - 1- ON "-, NO r-l 1 c - r-l -f | O C 3 O O O C 3 O O - - OC 3 j co o in - - *c *c in -t - - : ^ Tj ^ ; gj S S S 3 : - -t doc SSSS 3 d d d c 3 r*-< oo -T o 3 d d 3 vO m I r*"i t ^ - = - r*-, ri p* , co ^ en en rr, ri rr, m ON C 3 rn Cfv T C 3 10 "^ 2 ON - r ON OC r*-j r i c -*< v-- r- s~> r\ n < 5 - C ; -t \c n -^ ; r* t- p- p- rj in O C -, r i i o d d d c 3 o' d d c 3 O O H .g -" in c P; 2C g C N CfN CfN f^l ^ - r~ -f g o O a 3 in ON / -. ., ^ | 2 S = 3 O O O C 3 d d 1 - - 883 S FN OC OC OO ON C 2 O' O O O C - r-- f, ' diagonal and OO Q. I 0. 50 - o Growing Restricted n=9 n=10 O O o 100 200 300 400 500 Total protein per colony (ng) Figure 3. Bivariate scatter plots of the number of polyps in a colony and its total protein content. Linear regression using combined data from both treatments yields the equation v = 0.507.x - 2.44 (^-squared = 0.98). This intercept is not significantly different from zero (T = 0.568. P > 0.58). Regression lines for growing and restricted colonies do not differ in slope (ANCOVA, F = 0.84. df = 1. 15, P > 0.37) or elevation (F = 0.97. df = 1. 16. P > 0.34). D) c E o co g 0) _i D. CXI O (a) P carnea o o o o o Growing Restricted n=12 n=13 (b) H. symbiolongicarpus o o Growing n=6 Restricted n=6 500 1000 1500 2000 2500 3000 Colony size (ug protein) Figure 4. Bivariate scatter plots of oxygen uptake rate of growing and restricted colonies, (a) Data for Podocoryna cornea. The slopes of the regression lines for the growing and restricted treatments do not differ (ANCOVA, F = 1.75. df = 1. 21, P > 0.20), but an elevation difference was found (F = 20.54, df = 1, 22, P < 0.0002). These relationships were strengthened by omission of a single outlying data point from the growing data set (slope: F = 0.10. df = 1, 20, P > 0.76; intercept: F = 41.06, df = I. 21. P < 0.0001). (b) Data for Hydractinia svmbiolongicarpus. The slopes of the regression lines for the two treatments were not significantly different (ANCOVA. F = 0.83. df = 1. 8, P > 0.39), and neither were the intercepts (F = 0.98, df = 1. 9. P > 0.35). Although no significant difference in oxygen consumption rate was found between treatments for H. symbiolongicar- pus (Fig. 4b), a trend may be discerned in the data that would indicate agreement with the result found for P. car- nea. The sample size is too small to render this trend statistically significant, however. Characterization of colonv morpholog\ Growing colonies of both species had a more runner-like morphology than their restricted counterparts (Fig. 5; H. symbiolongicarpus, F = 12.56, df = 1, 20, P < 0.002; P. carnea. F = 6.16, df = 1, 22, P < 0.0212). Growth rate after regression A growth assay was performed 4 to 6 months after the pronounced winter regression. At this time, no significant EFFECT OF CLONING RATE IN HYDROIDS 81 30 (Runner-like) o o (Sheet-like) n=12 T H symb/olongicarpus P camea Figure 5. Comparison of growing and restricted colonies after 18 months of experimental treatment in terms of colony morphology as given by the shape metric (colony perimeter)/\ (colony area). Means and stan- dard errors are represented. difference was detected between treatments in either species for growth as measured by total colony polyp counts (Fig. 6a; H. symbiolongicarpus, F = 0.06, df = 1, 16, P > 0.806; P. carnea, F = 0.44, df = 1, 18, P > 0.516. data for both analyses log-transformed) or by total colony protein (Fig. 6b; H. symbiolongicarpus, F = 0.17. df = 1. 16. P > 0.689; P. carnea, F = 1.04. df = 1, 18, P > 0.321; data for both analyses log-transformed). Discussion Two experimental treatments were used in this study of hydroid colonies. One group of replicates was allowed to completely overgrow and remain undisturbed on 12-mm coverslips ("restricted" colonies); a second group was re- peatedly cloned as vegetative growth continued, without being allowed to enter into a gamete-producing sexual phase ("growing" colonies). A clear difference in growth rate was found between treatments in both species studied. with restricted colonies exceeding growing colonies in growth rate during controlled assays. Since only one clone was used per species, this result is not replicated at the level of the species. Nevertheless, at a higher level (i.e., species within family), the two clones provide replication of this primary result. Assays of the oxygen uptake rate between treatments revealed that the growing colonies of Podocoryna carnea exceeded the restricted ones in oxygen consumption. Al- though no significant statistical difference was found for Hydractiniu symbiolongicarpus, the sample size was small, and a trend seems to be discernible in the data that would suggest agreement with the result for P. ciirneu. Such a result may seem counterintuitive: the colony that uses more oxygen might also be expected to grow faster. On the other hand, higher oxygen uptake may be correlated with lower growth rate if the former indicates greater metabolic expen- diture on, for instance, somatic maintenance. Such a hy- pothesis is not entirely implausible. These hydroid colonies are ecologically space-limited, typically inhabiting small hermit crab shells. It is likely that selection favors rapid sequestration of available space to prevent the settlement of competitors; colonies may maximally allocate energy re- sources to growth until the available space is covered. Under such conditions of intense metabolic demand, cellu- lar metabolism may generate high levels of reactive oxygen species (Allen, 1996; Chiueh, 2000). These reactive species can cause various defects in macromolecules. so continu- ously growing colonies might experience defects in the mechanisms of oxidative phosphorylation or allocate greater resources to production of anti-oxidant enzymes (e.g., Blackstone, 2001). Thus the data are consistent with the hypothesis that growing colonies expend more energy on functions other than somatic growth, although further study of this issue is needed. Our interpretation of these results is that the restricted colonies are metabolically more efficient and so can allocate more energy to growth (Lowell and Spiegelman, 2000). g- 40-, ! | 30- to & (a) T H symbiolongicarpus P camea 3 60- o 30 - Q. "ro 5 20 - H 10 (b) H symbiolongicarpus P camea Figure 6. Growth rate comparisons of growing and restricted Hydrac- tinia symbiolongicarpus and Pntlt>ntiti curnea colonies from the ass;iy performed after 32-35 months of experimental treatment. Means and standard errors are represented, (a) Number of polyps per colony (M Total colony protein content. 82 L. M. PONCZEK AND N. W. BLACKSTONE The widespread tissue regression that occurred appar- ently reset to zero the growth rate difference that had been entrained by the experimental treatments. By this view, the physiological basis of the difference prior to regression was transmitted to the clonal fragments of the growing colonies, becoming enhanced over time as shown by the decreasing colony growth rate. This may suggest an epigenetic basis for the phenomenon, wherein a particular state of gene activity underlies the increased rate of oxygen consumption coupled with the reduced growth rate. During the regression event. all colonies lost most of their living tissue, effecting a cell population bottleneck. The elimination of the growth rate difference could perhaps be due to sampling error in the cells that escaped death during the regression, or to some dedifferentiation process involving a return to a metabolic ground state. In any case, cells of similar condition and gene activity seem to have survived the regression. Periodic regressions of this kind have been observed in some clonal taxa and are possibly related to senescence (Bayer and Todd, 1997; Gardner and Mangel. 1997). The life span of the modules (polyps) that make up a colony may be ex- tended through cycles of degeneration and regeneration (Hughes. 1989). Comparing absolute growth rates of colonies undergo- ing both treatments early in the experiment (Fig. 1) with those measured some two years later (Fig. 3) reveals a consistent decline. Furthermore, the growth rate equal- ization after regression occurred not by the growing colonies recovering a rapid growth rate but by the faster growing restricted ones assuming a similarly diminished rate. This reduction in growth rate over time may be considered to be a manifestation of colony senescence (Bell, 1988). By this criterion, growing colonies senesced more rapidly than restricted ones prior to the tissue regression event, suggesting that a high cloning rate accelerates colony senescence relative to uncloned colo- nies. After regression, the degree of clonal senescence (measured by growth rate) became equalized. Hydractiniid hydroid colonies fragment to produce po- tentially viable clonal modules, thus enlarging and dis- persing the genet asexually (Cerrano et al.. 1998). The colony fragmentation rate (equivalent to the cloning rate considered in this study) presumably could vary with the physical environment in which the hydroids are found. In aquaria. Cerrano et al. ( 1998) found that clonal colonies arising from fragments of Podocoryna exigna colonies can grow on a sandy-bottom substratum and that hermit crabs with naked shells placed into this environment were colonized within a few days. If such a process occurs naturally in P. exigna and other hydractiniid hydroids. such as the species used in this study, a genet might extend itself naturally by fragmentation. Clonal lineages may vary in fragmentation rate and growth rate of colo- nial ramets. This study shows that cloning rate could possibly affect the growth rate of a ramet within a lineage through negative feedback, since variation in growth rate may be passed on through some epigenetic mechanism such as cytosine methylation (but see Tweedie and Bird, 2000; and Amedeo et al., 2000). Nevertheless, histocom- patibility data (Grosberg et al.. 1996; Mokady and Buss, 1996) suggest that in at least some populations of H. symbiolongicarpus the rate of fragmentation is low rela- tive to the rate of sexual recruitment. The alteration in morphology with variation in cloning rate might have a bearing on the ecological functioning of a hydroid colony (McFadden et al., 1984; Yund, 1991; Brazeau and Lasker, 1992). Intraspecific competition is common between Hydractinia colonies (Buss and Black- stone, 1991). The present study has shown that a high cloning rate can produce a more runner-like colony mor- phology, thus tending towards a form associated with a "guerrilla" ecological strategy (Jackson et al., 1985). Such a clone might have more limited direct competitive ability, but might also be dispersed to more locations due to its greater rate of fragmentation. Asexual reproduction is an essential part of the life his- tory of all clonal organisms and is thus an important factor in their evolution and ecology. In some taxa, fragmentation rate depends on morphological characters, which are at least in part genetic and thus subject to selection. The fragmen- tation rate of clones of branching coral reef demosponges was found to depend on branch thickness (Wulff, 1985). A coral of the genus Plexaura has evidently evolved morpho- logical characters that make fragmentation more common in this species than in its congeners and produce some popu- lations in which more than 90% of the individuals are clonemates (Lasker, 1990). A possible difference in growth rate dependent on cloning rate would have to be taken into account when considering the demographic impact of frag- mentation. The effects of the two experimental treatments on the clonal replicates of both hydroid species indicate that fre- quently fragmenting colonies exhibit reduced colony growth rates, hence diminished reproductive potential and compromised competitive ability in the space-limited hab- itats in which they are typically found. Moreover, a within- species difference in colony morphology was found be- tween unfragmented colonies and those maintained in a constant state of vegetative growth by repeated cloning (fragmenting); this difference could affect the ecological functioning of the colonies in nature. However, these dis- crepancies may disappear if a large-scale regression of colony tissue occurs. Regardless of the specific physiolog- ical mechanisms producing these differential effects, frag- mentation rate can be important to various aspects of the biology of clonal organisms. EFFECT OF CLONING RATE IN HYDROIDS 83 Acknowledgments Comments were provided by K. Gasser, B. Johnson- Wint, and P. Meserve. The National Science Foundation (IBN-94-07049 and IBN-00-90580) provided support. Literature Cited Allen, J. F. 19%. Separate sexes and the mitochoncirial theory of aging. ./. Thcor. Binl. 180: 135-140. Amedeo, P.. V. Habu. K. Afsar, O. Mittelstein Scheid. and J. Pasz- ko ski. 2000. Disruption of the plant gene MOM releases transcrip- tional silencing of methylated genes. Nature 405: 203-206. Bayer, M., and C. D. Todd. 1997. Evidence for zooid senescence in the marine bryozoan Electro pilosa. Invertebr. Bio/. 116: 331-340. Bell. G. 1982. The Masterpiece of Nature: The Evolution and Genetics of Sexuality: Croom Helm. London and Canberra. Bell. G. 1988. Sex and Death in Protozoa: The History of an Obsession. Cambridge University Press. Cambridge. Blackstune. N. \V. 2001. Redox state, reactive oxygen species, and adaptive growth in colonial hydroids. J. Exp. Biol. 204: 1845-1853. Blackstone, N. VV., and L. \V. Buss. 1991. Shape variation in hydrac- tmiid hydroids. Biol. Bull. 180: 344-405. Brazeau. D. A., and H. R. Lasker. 1992. Growth rates and growth strategy in a clonal marine invertebrate, the Caribbean octocoral Bria- reum asbestiiiitm. Biol. Bull. 183: 269-277. Brusca, R., and G. Brusca. 1990. Invertebrate Biologv. Sinauer Asso- ciates. Sunderland, MA. 922 pp. Buss. L. VV., and N. W. Blackstone. 1991. An experimental exploration of Waddington's epigenetic landscape. Philos. Trans. R. Sue. Lond. 332: 49-58. Cerrano, C., G. Bavestrello, S. Puce, and M. Sara. 1998. Biological cycle ofPodocoryna exigua (Cnidaria: Hydrozoa) from a sandy bottom of the Ligurian sea. J. Mar. Biol. Assoc. U.K. 78: 1101-11 1 1. Chiueh, C. C., ed. 2000. Reactive Oxygen Species. Ann. N. Y. Acad. Sci. 899: 425 pp. Dudgeon, S., A. Wagner, J. R. Vaisnys, and L. W. Buss. 1999. Dy- namics of gastrovascular circulation in the hydrozoan Podocoryne camea: the one-polyp case. Biol. Bull. 196: 1-17. Edwards. C". 1972. The hydroids and the medusae Podocoryne areolata. P. borealis. and P. camea. J. Mar. Biol. Assoc. U. K. 52: 97-144. Gardner, S. N., and M. Mangel. 1997. When can a clonal organism escape senescence' Am. \at. 150: 462-490. Grosberg, R. K., D. R. Levitan, and B. B. Cameron. 1996. Evolution- ary genetics of allorecognition in the colonial hydroid Hydractinia symbiolongicafpus. Evolution 50: 2221-2241). Hall, V. R., and T. P. Hughes. 1996. Reproductive strategies of modular organisms: comparative studies of reef-building corals. Ecology 77: 950-963. Hughes, R. N. 1989. ,4 Functional Biology of Clonal Animals. Chapman and Hall. New York. Hughes, R. N., and J. M. Cancino. 1985. Ecological overview of cloning in metazoa. Pp. 153-186 in Population Biology and Evolution of Clonal Organisms. J. B. C. Jackson. L. W. Buss, and R. E. Cook, eds. Yale University Press, New Haven. Jackson, J. B. C. 1977. Competition on marine hard substrata: the adaptive significance of solitary and colonial strategies. Am. Nat. Ill: 743-767. Jackson, J. B. C., L. W. Buss, and R. E. Cook, eds. 1985. Population Biology and Evolution of Clonal Organisms. Yale University Press. New Haven. Larwood, G., and B. Rosen, eds. 1979. Biology and Systematics of Colonial Organisms. Academic Press. London. Lasker. H. R. 1990. Clonal propagation and population dynamics of a gorgonian coral. Ecology 71: 1578-1589. Lowell, B. B., and B. M. Spiegelman. 2000. Towards a molecular understanding of adaptive thermogenesis. Nature 404: 652-660. McFadden. C. S.. M. J. McFarland, and L. VV. Buss. 1984. Biology of hydractiniid hydroids. 1. Colony ontogeny in Hydractinia echinata (Flemming). Biol. Bull. 166: 54-67. Medawar, P. B. 1952. An Unsolved Problem of Biology. H. K. Lewis. London. Mokady, O., and L. VV. Buss. 1996. Transmission genetics of allorecog- nition in Hydractinia symbiolongicarpus (Cnidaria: Hydrozoa). Genet- ics 143: 823-827. Schmidt-Nielsen, K. 1997. Animal Physiology: Adaptation and Envi- ronment. Cambridge University Press, Cambridge. Stearns, S. C. 1992. The Evolution of Life Histories. Oxford University Press, Oxford. Tweedie. S.. and A. Bird. 2000. Mutant weed breaks silence. Nature 405: 137-138. Wulff, J. L. 1985. Variation in clone structure of fragmenting coral reef sponges. Binl. J. Linn. Sac. 27: 311-330. Yund, P. O. 1991. Natural selection on hydroid colony morphology by intraspeciric competition. Evolution 45: 1564-1573. Zamer, W. E., J. M. Shick, and D. W. Tapley. 1989. Protein measure- ment and energetic considerations: comparisons of biochemical and stoichiometric methods using bovine serum albumin and protein iso- lated from sea anemones. Limnol. Oceanogr. 34: 256-263. Reference: Biol. Bui!. 201: 84-94. (August 2001) Egg Longevity and Time-Integrated Fertilization in a Temperate Sea Urchin (Strongylocentrotm droebachiensis) SUSANNE K. MEIDEL* AND PHILIP O. YUND School of Marine Sciences, Darling Marine Center, University of Maine, Walpole. Maine 04573 Abstract. Recent tield experiments have suggested that fertilization levels in sea urchins (and other broadcast spawners that release their gametes into the water column I may often be far below 100%. However, past experiments have not considered the potentially positive combined ef- fects of an extended period of egg longevity and the release of gametes in viscous fluids (which reduces dilution rates). In a laboratory experiment, we found that eggs of the sea urchin Strongylocentrotus droebachiensis had high viability for 2 to 3 d. Fertilization levels of eggs held in sperm- permeable egg baskets in the field and exposed to sperm slowly diffusing off a spawning male increased significantly with exposure from 15 min to 3 h. In a tield survey of time-integrated fertilizations (over 24, 48, and 72 h) during natural sperm release events, eggs held in baskets accrued fertilizations over as much as 48 h and attained fairly high fertilization levels. Our results suggest that an extended period of egg longevity and the release of gametes in viscous fluids may result in higher natural fertilization lev- els than currently expected from short-term field experi- ments. Introduction Recent work has started to explore the fertilization dy- namics of free-spawning marine organisms that release one or both gametes into the water column (e.g., algae: Pearson and Brawley, 1996; corals: Lasker et al., 1996; starfish: Babcock et al.. 1994; sea urchins: Levitan et al.. 1992; ascidians: Yund, 1998; fish: Petersen etui, 1992). Although the details of scientific approaches vary, studies can be Received 22 September 2000; accepted 26 April 2001. * To whom correspondence should be addressed. E-mail: meidel@maine.edu broadly grouped into experiments in which a limited num- ber of manipulated organisms are induced to spawn, and surveys of natural spawning events (Levitan. 1995; Yund, 2000). Experimental studies that control spawning syn- chrony and spatial relationships to test specific mechanistic hypotheses generally suggest that fertilization levels may be limited by sperm availability unless males and females spawn simultaneously, at close range, or under nearly ideal flow conditions (see Levitan and Petersen, 1995; and Yund, 2000, for reviews). In contrast, many surveys of natural spawns report fairly high fertilization levels, at least at the times and places in which most members of a population spawn (Yund, 2000). However, comparisons between exist- ing experiments and surveys are complicated by two major factors. First, results from experimental studies can success- fully predict fertilization levels in natural spawns only if experimental conditions (both biotic and abiotic) accurately mimic natural spawning conditions; however, experiments often circumvent reproductive strategies that may have evolved to enhance fertilization (Yund, 2000). Second, ex- periments and surveys are rarely conducted with the same species, so it is virtually impossible to distinguish between taxonomic and methodological effects in existing studies. Echinoderms have proven to be a particularly valuable model system for short-term field experiments, and experi- mental fertilization data from echinoderms generally sup- port the paradigm of severe sperm limitation under a wide range of flow and population conditions (e.g., Pennington, 1985; Levitan, 1991; Levitan et al., 1992; Wahle and Peck- ham, 1999; but see Babcock et a!., 1994). However, there are no published surveys of fertilization levels in natural spawns of echinoderms. The absence of survey data is probably due in part to a lack of information on temporal spawning patterns and the proximate environmental cues that initiate spawning (though multiple cues have been 84 SEA URCHIN FERTILIZATION DYNAMICS 85 proposed and investigated: Hirnmelman, 1975; Starr et al., 1990. 1992. 1993). Two interrelated adaptations that have been largely by- passed in previous experimental studies may have consid- erable effects on fertilization levels in natural spawns of temperate echinoderms. The first is an extended period of egg viability, which potentially allows fertilizations to ac- crue over time. Short-term experiments make one or both of the following assumptions: that most eggs are fertilized within the first few seconds of release (Denny and Shibata. 1989; Levitan et al., 1991) and that gametes are quickly diluted to concentrations below which fertilization can oc- cur. Consequently, extended egg viability has implicitly been presumed to have little influence on fertilization levels in the field. Meanwhile, recent estimates of egg longevity have steadily extended what was presumed to be a relatively short period of viability. Pennington ( 1985) reported a min- imum viability period of 24 h for eggs of the temperate sea urchin Strongylocentrotiis droebachiensis (Muller), and eggs of a West coast sea urchin are now known to be viable for up to 2 wk when stored under axenic conditions (Epel et ai. 1998). If eggs can be fertilized for a long period of time, extended or repeated exposure of eggs to sperm during long-duration spawning events (or events in which multiple males spawn successively) could result in high time-inte- grated levels of fertilization, even if sperm are limiting in the short term. A second adaptation that may interact with extended egg longevity to increase fertilization levels is the release of gametes in viscous fluids, which reduces gamete dilution rates and potentially increases the duration of egg exposure to sperm. Thomas ( 1994) has shown that three species of sea urchins (Tripneustes gratilla, Echinometra miitliaei, and Colobocentrotus at rants) release gametes in such viscous fluids that eggs and sperm remain on the test and spines at current speeds less than 0.13 m s~'. When the current speed increases, gametes are transported away from this reservoir in long (3-4 cm) strings or clumps, which led Thomas (1994) to hypothesize that sea urchins may achieve high fertilization levels if gametes encounter each other in these structures. Sperm concentrated in clumps presumably also have greater longevity because of a reduction in the respiratory dilution effect (Chia and Bickell, 1983). In con- trast to natural sperm release, fertilization experiments often mimic "males" with syringes from which diluted gametes are extruded at a fixed (and fast) rate, thus circumventing the potentially beneficial effect of "sticky" sperm that cling to the test and spines and slowly diffuse away. In this study, we investigate the effects of these two aspects of sea urchin reproductive biology on fertilization levels in Strongylocentrotiis droebachiensis. We initially determine the duration of egg viability at two points during the reproductive season. We then explore whether extended (3 h) exposure of eggs to sperm diffusing off a male sea urchin enhances fertilization levels relative to short-term ( 15 min) contact at various downstream distances. Finally, we use the full period of egg viability to assay time-inte- grated fertilization levels during natural sperm release events in small populations and use the distribution of developmental stages in these field samples to evaluate the temporal distribution of fertilization events. Materials and Methods General procedures To obtain fresh eggs and sperm for use in experiments and field sampling, sea urchins (Strongylocentrotus droe- bachiensis) were injected through the peristomial mem- brane with 0.2-2.0 ml of 0.5 M KC1. Females spawned into 50-ml glass beakers containing chilled seawater that had been aged (~ 15-20 h; hereafter referred to as aged seawa- ter) to eliminate ambient sperm. Female spawn was checked to confirm the absence of immature oocytes (as indicated by the presence of a large nucleus and nucleolus) and then washed three times with aged seawater. Dry sperm was pipetted directly from the aboral surface of spawning males and kept refrigerated until use (maximum 2 h). To assay fertilization levels in the field, unfertilized eggs were deployed in sperm-permeable containers. These egg baskets consisted of a 0. 1-m-long frame of PVC pipe (in- ternal diameter 0.05 m) with the sides (90% of circum- ference) cut away, covered with 35-jim Nitex mesh (after Wahle and Peckham, 1999, as modified from Levitan et ai. 1992). and two Styrofoam floats attached for positive buoy- ancy. Baskets were suspended from the surface or deployed on the bottom in different spatial arrangements as described in the following sections. Egg longevity To determine the viability period of eggs of Strongylo- centrotus droebachiensis, we performed laboratory experi- ments at the beginning (experiment 1 : February 28 to March 2. 2000) and in the middle (experiment 2: March 28 to April 1, 2000) of the spawning season along the coast of Maine (March to May, Cocanour and Allen, 1967). In each exper- iment. 120 JJL\ of freshly spawned eggs (mean SE of 1651 69 eggs) from each of four females were added to 10 ml aged seawater (aerated for 1 h prior to use) in 20-ml glass scintillation vials. At the start of each experiment (0 h) and after 24. 48. 72. and 96 h (experiment 2 only), eggs in each of four replicate vials per female (only one replicate per female at h in experiment 1 ) were fertilized with 20 /xl of a 10-fold sperm dilution ( 10 jul fresh dry sperm from 3 males. 90 /u,l aged seawater). Vials were gently agitated three times during a 15-min period, following which the fertilization process was stopped with the addition of 2.5 ml 37% formaldehyde. At each time point, one additional via! 86 S. K. MEIDEL AND P. O. YUND per female was fixed without fertilization, as a control for false fertilization envelopes (from causes such as egg dam- age or low egg quality). Vials were kept at ambient seawater temperature ( 1-3C) during both experiments. Fertilization levels were calculated as the percentage of a random sub- sample of 300 eggs with a fertilization envelope. Two-way analyses of variance (ANOVA) with the fixed factors Female (four levels) and Time (three levels in ex- periment 1; five in experiment 2) were used to analyze variation in fertilization levels (% fertilization). To achieve homogeneity of variances, percent fertilization values were arcsine transformed for experiment 1 (O'Brien's test, F = 1.20. P > 0.32) but not transformed for experiment 2 (O'Brien's test, F 1.35, P > 0.19). The Student- Newman-Keuls (SNK) test was used for post-hoc compar- isons of levels within main effects in the absence of a significant interaction effect. Cumulative fertilization in the field: 15 min vs 3 h In this experiment, we determined whether extended (3 h) exposure of eggs in baskets to sperm from a spawning male enhanced fertilization levels relative to short-term (15 min) exposure. We constructed a fertilization platform that was mounted on a concrete block (L X W X H: 0.36 m X 0.33 m X 0.14 m) deployed by a rope. The platform consisted of a pine board (1.59 m X 0.24 m X 0.02 m) bolted to the concrete block so that it extended 0.31 m upstream of the block and 0.92 m downstream. The board housed one male and two female stations. The male station was simply a surface-mounted PVC plate (0.08 m X 0. 12 m X 0.003 m), located 0.30 cm from the upstream end of the board, to which a spawning male could be fastened. Female stations consisted of eyebolts anchoring ropes that extended to the surface and were located 0.3 and 1.0 m downstream of the male station. Experiments were performed on a sandy substratum be- low the dock of the University of Maine's Darling Marine Center in the Damariscotta River estuary (ME. 4350'N, 6933'W) at a depth of 4.30 m at mean low water (MLW). For each trial (n = 8), four egg baskets (two side by side 0.05 m above the platform at each of two female stations) containing 500 ju.1 freshly spawned eggs (mean SE: 7613 455 eggs) from one female were attached to the eyebolts. A male was induced to spawn by injection of 2.5-4.5 ml 0.5 M KC1 and then attached to the male station with rubber bands. The fertilization platform was then im- mediately deployed. In addition to the platform, two mobile female stations (baskets on weighted lines with the lower basket 0.35 m above the substratum) were deployed 2 m upstream (control for ambient sperm: one basket) and 2.60 m downstream (two baskets spaced 0.1 m apart vertically, omitted from trial 1 ) from the male station. After 15 min, one egg basket from each of the three downstream female stations was retrieved without disturbing the remain- der of the array, by pulling it to the surface on its own line. The remaining baskets were retrieved after 3 h, and the presence or absence of sperm on the aboral surface of the male was recorded. Eggs were immediately collected and fixed with formaldehyde. To determine fertilization levels, 300 eggs per vial (200-300 in five cases, 154 in one case) were randomly sampled and scored for the presence or absence of a fertilization envelope. Where sufficient num- bers of eggs were retrieved (82% of baskets), small sub- samples were taken before fixation and scored after about 15-20 h for the presence or absence of later developmental stages. During trials 2 through 8, current velocity was recorded with a 3D- ACM acoustic-doppler current meter (Falmouth Scientific). Each trial took place around mid-tide (i.e., com- menced 1.5 h after high [or low] water and ended 1.5 h before low [or high] water) to minimize variation in the flow regime. Three laboratory controls (held at 3C), consisting of 200 ju.1 freshly spawned eggs in 10 ml aged seawater, were assayed for ( 1 ) fertilization at the start of each trial; (2) fertilization at the end of each trial; and (3) the presence of false fertilization envelopes, scored twice (after retrieval of 15 min and 3 h samples). Laboratory controls were scored in the same manner as field samples. A two-way ANOVA with the fixed factors Time (two levels) and Distance (three levels) was used to determine differences in fertilization levels (%) in field samples. Per- cent fertilization values were arcsine transformed prior to analysis to achieve homogeneity of variances (O'Brien's test. F = 0.94, P > 0.47). Sperm availability in nature We measured cumulative (over 24, 48, or 72 h) fertiliza- tion levels of eggs retained in baskets during natural spawn- ing events of Strongylocentrotus droebachiensis. This sam- pling design is a hybrid between an experiment and a true survey of natural spawns, because any sperm present were naturally released, but egg locations were under experimen- tal control. Sampling started in mid-February and ended in early April in 1999 and 2000 but varied in intensity (both spatial and temporal) during the two years. In 1999, samples were collected at a single station at Christmas Cove (ChC, mouth of the Damariscotta River estuary); in 2000, samples were collected from three stations at ChC and four stations at Clarks Cove (C1C. 1 km seaward of the Darling Marine Center and -9 km from the ChC site). Both sites were relatively sheltered with a sandy substratum, and surveys of the immediate surroundings indicated the absence of sea urchin populations other than those sampled (pers. obs.). A small population of 5. droebachiensis ( 150 animals in 1999, -60 in 2000) occurred naturally at ChC. At C1C, we SEA URCHIN FERTILIZATION DYNAMICS 87 released about 350 sea urchins on a rock ledge around the lower low water line on January 29, 2000, but this popula- tion appeared to have declined to about 30 animals by April 7, 2000. At each site, multiple stations were positioned to provide samples at different nominal distances from the sea urchins. At ChC, station 1 was within 1 m of a rock wall that was inhabited by sea urchins during the autumn months; station 2 was on the shoreward end of a floating dock, 5 m straight offshore of the wall: and station 3 was on the seaward end of the same dock, about 13 m from the wall. The shallow depth of station I ( 1 .4 m at MLW) allowed sampling at only one depth (0.15 to 0.35 m above the substratum). At stations 2 and 3, we sampled the surface waters during each interval (1.4 to 6.2 m above the substratum, depending on the tidally variable water depth): at times of anticipated sperm pres- ence (based on 1999 results) we also sampled the bottom water 0.15 to 0.35 m above the substratum. During 1999. only station 3 was sampled, and egg baskets were deployed only near the surface. Because the sea urchins were free to move, the positions of our stations relative to spawning males could not be known precisely. However, likely loca- tions can be inferred from sea urchin movement patterns. In 1999, sea urchins mainly remained on the rock wall or wandered between stations 1 and 2. whereas in 2000 many animals spent the spawning season on a piling adjacent to station 2. We employed a similar sampling scheme at C1C. with minor modifications to accommodate local dock structures. Station 1 was within 1 m of the rock ledge to which sea urchins were transplanted: station 2 was 1 m straight off- shore of station 1 (along a fixed wooden dock): and stations 3 and 4 were on floating docks about 12 m from station 1, at 45 angles to either side of the transect from stations 1 to 2. Because of minimal water depth ( 1 .0 to 1 .4 m at MLW). all stations were sampled at only a single depth (stations 1 and 2: 0.15 to 0.35 m above the substratum; stations 3 and 4: 0.4 to 3.5 m above the substratum, depending on the tidally variable water depth). Stations 3 and 4 were sampled only when sperm were expected to be present. At each site, sets of three replicate egg baskets (spaced 0.1 m apart vertically) were deployed at each station and depth and retrieved 24 h (1999 only), 48 h. or (on only three occasions) 72 h later. In 1999. baskets contained 500 fil of eggs ( 7600 eggs) from one female, and in 2000 they contained 800 /j.1 of eggs (mean number SE: 11216 787 eggs) pooled from two to three females. Laboratory controls (200 /xl of eggs in 10 ml aged seawater) were used to determine the incidence of fertilization membranes prior to basket deployment (presumably reflecting sperm contam- ination) and at the time of retrieval (presumably reflecting false membranes). To determine fertilization levels. 300 eggs per basket or vial were randomly subsampled and scored in three categories: unfertilized, presence of a fertil- ization envelope, or development through a later stage (2-64 cells, unhatched/hatched blastula. gastrula). Eggs with fertilization envelopes present were judged to have been fertilized only if the sample also contained later de- velopmental stages. From 41% of baskets (181 out of 441 ). fewer than 300 eggs were retrieved; in these cases, all retrieved eggs were scored. For the calculation of mean fertilization levels, only baskets with more than 50 retrieved eggs were used, resulting in a loss of replicates at some sites and times. We estimated the approximate distribution of fertilization events during a sample interval from the distribution of developmental stages in a sample and the known rate of development to each stage. We used Stephens' (1972) de- velopmental times for S. droebachiensis at 4C from fertil- ization to 32-cell stage (2-cell: 5 h: 4-cell: 8 h: 8-cell: 10.5 h; 16-cell: 14 h; 32-cell: 18 h). From the 64-cell stage to gastrulation, we used our own observations of develop- mental times (64-cell: 21 h; blastula: 24 h; hatching: 40 h; early gastrula: 48 h). We calculated the distribution of fertilizations (%) in time as the percent at each stage (i.e., of a certain age, in h) of all embryos detected (pooled from three replicate baskets). To establish the extent to which spawning had occurred during the 2000 sampling period, we collected sea urchins for analysis of gonad index (wet weight of gonads as a percentage of total wet body weight) from ChC l/i = 10) and C1C (;; = 1 1 ) on April 7 and 11. 2000, respectively. Results Egg longevity Egg viability in aged seawater in the laboratory (as as- sayed by fertilization with fresh sperm) varied significantly among time intervals and females in both experiments (Fig. 1). In experiment 1 (February 28 to March 2, 2000). the effects of both Female (F 3 . 3f) = 5.68, P = 0.003) and Time (F 2 36 = 8.94, P < 0.001 ) were significant, but the interaction between the two main factors was not (F 6 36 = 1.74. P = 0.14). Post-hoc comparisons revealed that fertilization levels were significantly lower for female 2. but similar for females 1. 3. and 4 (SNK-test. P < 0.05: Fig. 1A). Fertilization levels were highest at h, similar at 24 and 48 h (SNK. P > 0.05 ). and significantly lower by 72 h (SNK, P < 0.05). In experiment 2 (March 28 to April 1, 2000), there were again significant Female (F 3 60 = 18.0. P < 0.001 ) and Time (F 4 60 = 273. P < 0.001 ) effects, as well as a significant interaction between the two main factors (F,, 60 = 32.9, P < 0.001). Fertilization of eggs from females 1 and 4 remained relatively high at 72 h. while levels declined markedly for females 2 and 3 (Fig. IB). For females 1 and 2. fertilizations dropped to very low levels by 96 h. while fertilizations for females 3 and 4 were higher at 96 h than at 72 h (Fig. IB). Of a total of 36 control sample- 88 S. K. MEIDEL AND P. O. YUND A) February 28 - March 3. 2000 loo-,. , i, li IL LL ,L ED" 1148 D : B) March 28 - Apnl 1.2000 LL Female Figure 1. Mean ( +SE) fertilization levels (%) over time of eggs from four female sea urchins (A) at the beginning (experiment 1 ) and (B) in (he middle (experiment 2) of the spawning season. Replication is four vials for eac I). me expermen o e spawnng season. epcaon s each female/time combination (except experiment 1 at h: replication = ( 16 and 20 in experiments 1 and 2. respectively), 5 had 0.3% false fertilization envelopes and 1 had 0.7%. In spite of the significant variation among sample times and females in both experiments, egg viability was basically quite high for 48 to 72 h (Fig. 1). With the exception of female 2 in experiment 1. more than 75% of eggs held in aged seawater in the laboratory were viable for 48 h (Fig. 1 ). At 72 h, viability was in the 50%-75% range for eggs from 6 of the 8 females (Fig. 1 ). Cumulative fertilization level (15 min vs 3 li) When eggs in baskets were exposed to a continuous sperm supply from a spawning male, fertilization levels increased from 15 min to 3 h at distances of 0.3 and 1.0 m downstream from the male, but remained similar over time at 2.6 m (Fig. 2). In the 15-min samples, fertilization de- creased with distance from 0.3 to 1.0 m. but remained similar between 1 and 2.6 m (Fig. 2). In the 3-h samples, fertilization decreased monotonically with distance. The two-way ANOVA indicated significant Time (F, , 9 = 31.3, P < 0.001) and Distance (F 2 39 - 40.1, ' P < 0.001 ) effects, as well as a significant interaction between the two main factors (F 2 __, 9 = 4.87, P = 0.013). In 5 out of 8 trials, the male still had sperm on its test at the end of the 3-h deployment, suggesting that fertilization would have continued well beyond the end of our sample interval. Upstream controls for ambient sperm levels (Fig. 2) gen- erally had SO. 3% fertilization except in trials 1, 6, and 7 when fertilization levels reached 5.3%, 9.0%, and 2.0%, respectively. We attribute fertilizations in trial 6 to a large boat wake that probably created oscillatory water motion and transported sperm towards the upstream control sample immediately before retrieval of the 15-min samples, and we attribute fertilizations in trial 7 to false envelopes (see below). Fertilizations in trial 1 could not be attributed to any obvious cause, and the recorded value was subtracted from the fertilization levels recorded in experimental baskets for that trial. The apparent absence of a decline in fertilization between the 1- and 2.6-m samples at 15 min and the lack of an increase in fertilization between the 15-min and 3-h samples at 2.6 m are both attributable to one exceptional sample. During trial 5. we recorded a fertilization level of 48% at 2.6 m at 15 min. while values in other trials ranged only from 0.0%' to 3.3% (mean SE %: 1.4% 0.5%; n = 6) at 15 min and from 3.7%- to 15.3% (6.9% 1.8%; n = 6) at 3 h. If this outlier is excluded, fertilization declines from 1 to 2.6 m at 15 min and increases from 15 min to 3 h at 2.6 m. In laboratory controls, fertilization levels were always very high at the beginning (mean SE: 94.6% 1.7%; /; = 8) and the end (94.8% 1.6%; n = 8) of a trial. Controls for sperm contamination or false fertilization en- velopes mostly indicated 0% envelopes (15 min, 0.3% 0.2%; 3 h, 0.5%. 0.4%; n = 8) except in trial 7 where n n i. Distance from male (m) * Current direction Figure 2. Fertilization as a function of distance and duration of sperm exposure in the field experiment. Mean ( +SE) fertilization levels (%) are reported for each time/distance combination. Spawning male is located at 0.0 m mark. Upstream basket was retrieved after 3 h (hatched bar); downstream baskets after 15 min (stippled bars) or 3 h. Replication is 8 trials, except 7 trials for 2.6 m after 15 min. and 6 trials for 2.6 m after 3 h. SEA URCHIN FERTILIZATION DYNAMICS 89 \.T7c and 3.3% envelopes were found after 15 min and 3 h, respectively. These percentages were subtracted from the fertilization levels recorded in the Held for that trial. Current velocities varied widely during trials 2 through 7 and ranged mainly from 0.08 to 0.20 m s~ ' (Fig. 3). Mean velocities varied 5-fold among trials during the initial 15- min period (from 0.026 to 0.130 m s~') but were quite similar over 3 h (from 0.121 to 0.155 m s~"). Sperm availability in nature In both years of the survey (1999, 2000) and at both sites (ChC, C1C), no fertilizations were recorded during most of the sample intervals. However, in both years several sperm- release events of variable magnitude were detected. In 1999 at ChC (only station 3 surface was sampled), fertilizations occurred on March 5 (mean time-integrated fertilization level 4.7% ). March 23 (57.3%). March 31 (6.6%). and April 1 (24.69r ). In 2000 at ChC, fertilizations occurred on Feb- ruary 19 (station 1 only. 39.5%), March 10 (station 1. 10.3%; station 2, 9.3% surface; no bottom samples were deployed and no fertilization was detected at station 3), March 19 (station 1, 62.3%; station 2. 34.3% surface and 11.3% bottom; station 3. 30.4% surface and 5.3% bottom), and March 29 (station 1, 3.4%-; station 2, 4.6% surface; station 3, 4.5% surface; no bottom samples were deployed). At C1C (sampled only in 2000), fertilizations were detected on March 10 (station 1, 24.1%; no fertilization was detected at station 2; stations 3 and 4 were not sampled), March 17 (station 1, 27.7%: station 2. 10.4%; station 3. 26.2%; station 4, 3.7%), and April 3 (station 1, 6.9%; station 2, 3.3%; stations 3 and 4 were not sampled). In laboratory controls, fertilization levels were always very high at the start of each sample interval (mean SE %: 1999, 96.7% 0.7%, // -- 19; 2000, 93.8% 0.9%, n = 20). Controls for false fertilization envelopes (stored in the laboratory and fixed upon retrieval of the corresponding field sample) had very low levels of false envelopes (1999, 0.8% 0.5%, n = 16; 2000, 0.2% 0.1%; n = 20). Based on the distribution of developmental stages (two- cell to early gastrula) at the time of collection, we estimated that the temporal fertilization pattern varied markedly among the major sperm release events that we detected (Figs. 4-6). Because the discrete developmental stages that we scored are separated by longer time intervals later in development, the 24-h sample interval utilized in 1999 at ChC produced far better resolution of the time of fertiliza- tion (~3 h) than did the 48- to 72-h intervals employed in 2000 ( 3-h resolution for the 24 h immediately preceding sample collection, but 10 h for the portion of the interval >24 h prior to collection). In 1999. fertilizations occurred in fairly continuous trickles over about 48 h (March 3-5; Fig. 4A) or 24 h (March 22-24; Fig. 4B) or in two distinct pulses of similar magnitude about 24 h apart (March 30-April 1; 8 ^ C 1 . Iceland (64N. 22W) 1 Voucher specimens are being maintained in the marine invertebrate collections of C. W. Cunningham at Duke University. the two Atlantic species is entirely due to long-term isola- tion. Thus, subsequent physiological adaptations to warmer water in A. forbesi (Franz et ai, 1981 ) are independent of the speciation event. Essentially, the distinction between these species reflects either primary divergence due to se- lection or secondary contact following vicariance (Endler. 1977). In this study, mitochondrial and nuclear sequence data were collected from populations of A.forbesi and A mbens throughout North America and Europe, as well as from populations of the Pacific sister taxon A. tinnirensis (Clark and Downey. 1992). Phylogenetic and population genetic assays were used to test the hypotheses described above. It appears that Worley and Franz (1983) were remarkably accurate in suggesting a Pliocene speciation followed by a recent invasion of A. mbens from Europe, even in their prediction of details of timing, mechanisms, and effects. Although selection may have driven some of the diver- gence, it now seems clear that the initial separation of A. mbens and A. forbesi is due to late Pliocene changes in climate and ocean current flow, whereas North American populations of A. mbens are very recent arrivals. Materials and Methods Asterias specimens were collected from intertidal sites listed in Table 1 . Tube feet were immediately placed in 95% ethanol or DMSO buffer (0.25 M EDTA pH 8.0. 20% DMSO, saturated NaCl; Seutin et cil., 1991). Species were identified on the basis of key morphological characters described in Clark and Downey (1992) and Hay ward and Ryland (1995). DNA extraction and amplification DNA was phenol-extracted from each specimen follow- ing the protocol in Hillis et ul. (1996). These extractions were stored at 80C. PCR amplification of an approxi- mately 700-bp portion of the mitochondrial cytochrome c oxidase I (COI) protein-encoding gene was performed using the primers LCO1490 and HCO 2198 from Folmer et al. (1994). Amplification was performed in 50-;u,l reactions containing 10-100 ng DNA, 0.02 mM each primer, 5 jul Promega 10X polymerase buffer, 0.8 mM dNTPs (Pharma- cia Biotech), and 1 unit Taq polymerase (Promega). Reac- tions took place in a Perkin-Elmer 480 thermal cycler with a cycling profile of 94 : (60 s) -40 (90 s) -72 ( 150 s) for 40 cycles. The internal transcribed spacer (ITS) region was amplified under similar conditions, with an annealing tem- perature of 50C and with primers ITS4 and ITS5 (White et ul.. 1990). For each individual, sequences were obtained for three to four clones, and the consensus sequence was ob- tained to eliminate Taq error. PCR products were prepared for sequencing and were cycle-sequenced as in Wares (2001) using both PCR prim- ers. COI sequences representing each individual in this study have been deposited with GenBank (AF240022- 240081 ); ITS sequences were only obtained for 10 individ- uals, representing each species and region, and are also accessible in GenBank (AF346608-AF346617). Sequences were aligned and edited for ambiguities using complemen- tary fragments in Sequencher 3.0 (Genecodes Corp., Cam- bridge, MA). No gaps or poorly aligned regions occurred in the COI alignment, but missing characters were trimmed from the ends of the alignment to produce equal sequence lengths for all individuals. In the ITS alignment, all missing or ambiguous characters, including gaps, were removed. Consensus sequences were exported as a NEXUS file for subsequent analysis in PAUP*4.0b4a (Swofford. 1998). Phylogenetic analysis A heuristic search for the set of most-parsimonious trees based on the COI data was performed using PAUP*4.0b4a (Swofford, 1998). Trees were rooted using Leptasterias polaris (Asteriinae) and individuals of A. tinnirensis. Start- ing trees were obtained via stepwise addition, with simple addition sequence. Tree-bisection-reconnection was used for branch swapping, and branches were collapsed if the maximum branch length was zero. Maximum-likelihood (ML) phylogenies were also gener- ated in PAUP*. The best-fit model for all likelihood anal- yses (HKY with F-distributed rate variation; Hasegawa et ai, 1985; Yang, 1994) was determined by adding parame- ters until the likelihood description of the neighbor-joining tree did not significantly improve (Goldman. 1993; Cun- ningham et 0.10), though the substitution rate is significantly different (P < 0.05). Bootstrap replicates of the ITS data also indicate strong support for differentiation among these species. The ITS data do not reject a molecular clock model. Divergence among these species is indicated in Table 2. HKY -I- F distances in the COI fragment indicate that A. amurensis, A. forbesi, and A. nibens have been isolated from each other for a similar amount of time; assuming trans-Arctic isolation around 3.5 Ma. A. nibens and A. forbesi have been separated for at least 3.0 Ma. Although the estimated divergence date is higher when all codon positions are included (Table 2), and these data do not reject a molecular clock, neutrality tests (see below) suggest that some second-position substitutions may be under selection. Therefore, third-position sites may be more appropriate for the divergence estimate. The estimated divergence time is also higher when the ITS data are used; however, there is no reason to believe that speciation predated the appearance of Asterias in the North Atlantic, and the long branch leading to A. forbesi is not easily explained since it appears in both phylogenies (one using a protein-coding gene, one using untranslated spacer region data). This longer branch appears to influence the age estimates of the COI (all positions) and ITS data sets strongly. A McDonald-Kreitman test (McDonald and Kreitman, 1991) rejects a pattern of neutral substitution between A. nibens and A. forbesi (P < 0.01 , Table 3). Despite branch lengths that do not reject the molecular clock model, there is an excess of amino acid replacement substitutions be- tween the Atlantic species. The replacement substitutions between A. nibens and A. forbesi do not include any first- position substitutions. Half (8/16) of the amino acid substi- tutions do not involve a change in charge or polarity, whereas almost half (7/16) of the changes substitute a basic residue for an uncharged or nonpolar residue. However, there does not seem to be an obvious pattern to these changes between A. nibens and A. forbesi. Other species comparisons do not reject the neutral model of substitution (Table 3). Within each species, Tajima's (1989) test is nonsignificant (A. amurensis, D = 0.837, P > 0.10; A. forbesi, D = -0.705, P > 0.10; A. nibens, D = -1.482, P > 0.10), indicating that there is no reason to suspect non-neutral evolution in the intraspecific comparisons. Additionally, Bayesian analysis (Templeton el at., 1992; Clement ct ai, 2000) of the COI data within A. rubens indicates greater than 95% confidence that the intraspecific gene tree is parsimonious. The ML root haplotype is found on both coasts of the Atlantic (Fig. 1A, Haplotype B), and this haplotype is at least an order of magnitude more likely to be the ancestral haplotype than any other haplotype of A. rubens (likelihood index = 0.857). All North American haplotypes are also found in Europe; the unique haplotypes found in Europe contribute to a significantly higher allelic diversity (P < 0.0 1 . Table 4). The ITS data are consistent with the COI data in that there is no allelic diversity among North American and European individuals of A. nibens (n = 6). Discussion Understanding the mechanisms that are responsible for the divergence of Asterias nibens and A. forbesi first re- quires that the timing of their divergence be estimated. Estimates based on the molecular calibrations reported here suggest that these species last shared a common ancestor at least 3.0 Ma (Table 2), not long after the genus first arrived in the North Atlantic (around 3.5 Ma; Worley and Franz, 1983; Vermeij, 1991 ). Note, however, that asterozoan skel- etons are rarely preserved in the fossil record, because they lack rigidly articulated skeletons and rapidly disintegrate (Barker and Zullo, 1980); indeed, fossils of A. forbesi have been reported only twice, each time in Pleistocene intergla- cial sediments. Thus, little direct evidence points to the first appearance of Asterias in the North Atlantic (Durham and MacNeil, 1967; Worley and Franz, 1983), and the biogeo- graphic data used in this paper is therefore based on con- BIOGEOGRAPHY OF ASTERIAS 99 1 A. forbesi 10C> -L 1 H Norway Norway Haplotype A (n=16) 1 Ireland _l rubens 100 Haplotype B (=14) - Ireland 1 Ireland A Haplotype C (n=l) - France - France Norway Haplotype D (=2) e and 100 100 ^ r _jT 100 * 1~ 0.01 substitutions/site B r 100 1 Ireland Iceland i Iceland Newfoundland A. ruutis 100 Maine Maine -1 A. forbesi 99 0.005 substitutions/site Figure 1. Phylogenetic trees for Asierias generated using the best-tit maximum likelihood model in each data set (COI: HKY + T; ITS: F81 ). (A) Cytochrome c oxidase I phylogeny of inter- and intraspecific Asterias relationships. Here all characters (first, second, and third position) are included; an identical topology is found using parsimony or distance methods, or looking at third-position characters alone. Bootstrap support for each species is indicated by the numbers below each branch. These data do not reject a molecular clock model. The divergence across the Arctic (between A. amurensis and the Atlantic speciesl is considered to be 3.5 Ma; this generates an estimate of about 3.0 Ma for the divergence between A. rubens and A. amurensis (see Table 2 and Discussion). Haplotypes A D of A. rubens are found on both the North American and European coasts (A: Maine (n = 8), Nova Scotia (n = 2), Newfoundland (n = 2), Iceland in = 1 I. Norway (H = 2), Ireland (H = I ); B: Maine (n = 2). Nova Scotia (n = 4). Newfoundland (n = 3). Iceland (H = 1 ). Ireland (H = 2). France (n = 2); C: Maine (n = 1 ). Norway (n = 3), Ireland (n = 2). and France (n = 1 ); D: Ireland (n = 1 ), and Maine (n = 1 )). Amphi-Atlantic haplotype B is the maximum likelihood root (index = 0.857). (B) Internal transcribed spacer ( ITS ) phylogeny of inter- and intraspecific Asterias relationships. Likelihood ratio tests do not reject a hypothesis of proportional branch lengths (P > 0. 10) suggesting that, aside from substantial differences in substitution rate, the two phylogenies are equivalent representations of interspecific differentiation. A nearly identical phylogeny is reconstructed when indels are included in the ITS data. sistent fossil evidence from other cold temperate species that participated in the trans-Arctic exchange. Nevertheless, there is reason to believe that Asterias also spread from the Pacific to the Atlantic at about 3.5 Ma (Worley and Franz. 1983). Miocene and early Pliocene temperatures were around 5-6C warmer in the North Atlantic and Arctic, permitting the initial trans-Arctic passage of temperate spe- cies (Berggren and Hollister. 1974; Vermeij, 1991 ), but then two dramatic changes were initiated around 3.0 Ma that appear to play a role in speciation within the North Atlantic. 100 J. P. WARES Table 2 Internal branch lengths (based on best-fit likelihood model} separating Asterias species (lower triangle*, all 3 matrices) All characters A. atnurensis A. rubens A. forbesi A. anmrensis tL = 1.954 x 10~ 8 8.63 x 10" 9 IJL = 2.665 X 10~ 8 9.59 X 10~" A. nibens 0.13678 0.06044 3.59 Ma A. forbesi 0.18658 0.06715 0.16576 0.04595 3rd position only A. anmrensis A. rubens A. forbesi A. amurensis ju, = 6.689 x 10"" 3.36 x 10~ 8 M, = 9.751 x 10~ 8 3.74 x lO"" A. rubens 0.48084 0.2352 2.96 Ma A. forbesi 0.68254 0.26168 0.49270 0.15661 ITS-1 A. anmrensis A. rubens A. forbesi A. amurensis p. = 5.142 x 10~ 9 2.04 x 10~ 9 /M = 7.188 x 10~ 9 2.40 X 10~" A. nibens 0.0361 0.0143 3.84 Ma A. forbesi 0.0500 0.0168 0.0470 0.0163 The calibration date of 3.5 Ma is used to obtain the mutation rate ^. for comparisons between A. amurensis and the Atlantic species. The estimated divergence time between A. rubens and A. forbesi is based on the mean of this calibrated mutation rate (cytochrome c oxidase I [COI] all positions, top; COI 3rd position only, middle; internal transcribed spacer (ITS) 1. bottom). * In each matrix, the lower triangle containing the internal branch lengths is made up of the matrix cells below the diagonal line of empty cells representing comparisons within the same value. The upper triangle contains the estimated mutation rates and estimated divergence data. At that time, warm North Atlantic currents were dis- placed by the formation of the cold-water Labrador Current. This event created a significant thermal gradient in the North Atlantic, and tropical-temperate faunas were abruptly replaced with polar and subpolar faunas on the continental Table 3 McDonald-Krehman tests on each Asterias species pair using cytochrome c oxidase I (COI) translated data Species pair Fixed differences Polymorphisms A. rubens-A. forbesi Synonymous 39 19 Nonsynonymous 16 ; P < 0.001 Synonymous 36 21 Nonsynonymous 12 1 P > 0.05 A. forbesi-A. amurensis Synonymous 44 15 Nonsynonymous 14 1 /' > 0.15 Only the comparison between A. rubens and A. forbesi indicates a significant departure from neutral evolution. A two-tailed Fisher's exact test was used for each set of comparisons. shelf off Nova Scotia and the rest of New England (Berg- gren and Hollister, 1974; Worley and Franz, 1983; Cronin, 1988). As Northern Hemisphere glaciation began, (lie present-day latitudinally controlled faunal provincialization was established as well (Berggren and Hollister, 1974). This dramatic cooling of the northwestern North Atlantic prob- ably initiated the separation of North Atlantic Asterias into European and North American populations with very little genetic contact (Worley and Franz, 1983). Subsequent Pleistocene glaciation would have prevented the long-term Table 4 Comparisons of haplotype diversity (\\, see eqn. 8.4 in Nei 1987. calculated in DNAsp 3.50, ROMS and Rozas 1999) for the cytochrome c oxidase I fragment in each species and population of A. rubens Species/Population Haplotype diversity (H) cr Asterias rubens 0.793 0.00138 North America 0.597 0.00395 Europe 0.893 0.00143 Asterias forbesi 0.964 0.00596 Asterias amurensis 0.999 0.03125 European populations of A. rubens have significantly higher allelic diversity than North American populations (P < 0.01); this finding is supported by nonparametric haplotype sampling in Wares (2000). BIOGEOGRAPHY OF ASTERIAS 101 establishment of populations in New England, as most of the North American coast from Long Island Sound north- ward was covered by a kilometer of ice during glacial maxima (Kelley et til., 1995). Pacific and Atlantic populations of other species appear to have had more recent trans-Arctic genetic contact than the estimates above would suggest for Asterias (Palumbi and Kessing. 1991; van Oppen el al., 1995). Moreover, rapid climatic fluctuations (Cronin, 1988: Roy et al.. 1996) during the Pleistocene could have permitted large-scale changes in the geographic range of cold temperate species. However, both the sea urchin Strongylocentrotus pal/idus (Palumbi and Kessing, 1991) and the red alga Phycodrys nihens (van Oppen et al., 1995) appear to have greater tolerance for Arctic waters than Asterias does. Worley and Franz (1983) report that expansion of Asterias populations into habitats as far north as Greenland only occurs period- ically, and that these populations cannot tolerate colder waters (Franz et al.. 1981 ). However, the indirect morpho- logical and paleontological evidence is bolstered by the molecular evidence, which strongly suggests that A. rubens and A. forbesi diverged shortly after their ancestral lineage separated from the Pacific A. amurensis. The estimates of mutation rate presented here are very similar to other esti- mates for both the COI fragment (Knowlton and Weigt, 1998; Schubart et al.. 1998; Wares, 2001; Wares and Cun- ningham, in press) and the ITS fragment (Schlotterer et al.. 1994; van Oppen et al.. 1995). Thus these data strongly support earlier inferences of a late Pliocene trans-Arctic passage and subsequent speciation within the Atlantic. An analysis of genealogical patterns within A. rubens confirms that the North American populations of this spe- cies are descendants of a recent colonization from Europe that probably followed the most recent glacial maximum (about 20.000 BP, Holder et al.. 1999). The genealogical data presented here fit several important patterns that sug- gest a recent range expansion (Wares. 2000). All North American haplotypes are identical to the most-common European haplotypes (Fig. 1A). Generally, invading haplo- types are the most deeply nested haplotype in the European (putative source) population. This is to be expected, because deeply nested ancestral haplotypes are often the most com- mon (Castelloe and Templeton. 1994), and therefore have a higher probability of participating in long-distance dispersal events. Haplotype B (Fig. 1 A) is a good illustration of this expectation it is closely related to each other haplotype and has a high copy number in both European and American populations. These observations contribute to the high like- lihood (85.7<7r, more than an order of magnitude greater likelihood than any other haplotype) that this is the ancestral allele in A. rubens. Additionally, allelic diversity is significantly lower in North American A. rubens than in Europe (Table 4), a signal of recent range expansion (Hewitt, 1996; Austerlitz et al.. 1997). However, the North American colonization is diffi- cult to date because there are no unique haplotypes in North America; ancestral allelic polymorphism tends to inflate indirect estimates of population size and age (Kuhner el al.. 1998: Edwards and Beerli, 2000). The lack of unique di- versity in North America also prevents the meaningful use of other phylogeographic methods; for instance, statistics of the geographic dispersion of haplotypes (for review see Templeton, 1998) are uninformative (Wares, unpubl. data). This is primarily because even closely related individuals (identical haplotypes) are distributed across the entire geo- graphic range of A. rubens. It is possible that the multiple shared alleles between Europe and North America represent a multiple-invasion history; Asterias larvae are planktotro- phic and disperse in the water column for 6 or more weeks (Clark and Downey, 1992). There is evidence that natural selection has played some role in the overall divergence between these species. A significant number of amino acid replacement substitutions distinguish A. rubens from A. forbesi (Table 3), all of them reflecting second- or third-position nucleotide substitutions. There is no obvious pattern to the amino acid replacements, as most of them involve substitutions among uncharged or nonpolar amino acids. Two of the three species in the genus Asterias are found in cold-temperate waters, while A. forbesi is found in the warmer mid- Atlantic region (Schopf and Murphy, 1973; Franz et al., 1981 ). Many of the phys- iological differences between A. rubens and A. forbesi (Franz et al., 1981 ) reflect this latitudinal distribution. How- ever, the possibility that these amino acid substitutions are related to physiological differences in the warm-temperate A. forbesi has never been tested. The difference in temper- ature between the habitats of A. rubens and A. forbesi is unlikely to contribute to differences in metabolic rate that could accelerate the mutation rate (for review see Rand, 1994). Nevertheless, this hypothesis is worth examination, because A. forbesi is supported by relatively long branches in both the COI and the non-coding ITS region (Table 2, Fig. IB). If natural selection is playing a role in the amino acid divergences of the mitochondrial COI gene between A. rubens and A. forbesi, there is no reason why a noncoding nuclear sequence should reflect the same increase in diver- gence rate. In conclusion, the biogeographic response of Asterias to late Pliocene climatic and oceanographic change fits a pat- tern predicted by Worley and Franz (1983). Following the arrival of Asterias in the North Atlantic around 3.5 Ma (Worley and Franz, 1983; Venneij, 1991). populations were established on both the European and North American coasts during a period when the North Atlantic was as much as 5-6C wanner (Berggren and Hollister. 1974). The for- mation of the Labrador Current 3.0 Ma rapidly changed the faunal composition of the intertidal Canadian Maritimes and New England coast, and Asterias populations in this region J, P. WARES probably went extinct. An American population survived under the conditions of the mid-Atlantic coast and Gulf Stream waters (A.forbesi), and the European population (A. rubens) has recently recolonized the cold-temperate shores of New England and the Canadian Maritimes. Thus, the zone of sympatry between these two species appears to be a zone of secondary contact. Hybridization is considered rare between these species (Schopf and Murphy, 1973; Worley and Franz, 1983), but whether behavioral mechanisms (Franz et ai, 1981) or gametic recognition mechanisms (Hellbergand Vacquier, 1999; Fernet, 1999) are responsible is unclear. The genetic data presented here illustrate a strong con- cordance between paleoceanographic changes and indirect estimates of speciation between the North Atlantic Asterias species. The species boundaries are phylogenetically quite distinct, and the divergence estimates based on these genetic data appear to support a late Pliocene, rather than late Pleistocene or Holocene, separation. A better understanding of the balance between oceanographic and climatic changes in the late Pliocene and Pleistocene, and of the response of species based on varying life-history characters to these changes, will enable us to predict the responses of other taxa (Cunningham and Collins, 1998: Wares and Cunningham, in press). Acknowledgments 1 thank G. Manchenko, A. Ingolfsson, J. Maunder, B. O'Connor, D. Garbary, D. M. Rand, and C. Damiani for assistance in the field collecting seastars. The manuscript was greatly improved thanks to discussions with T. Turner and the suggestions of two anonymous reviewers. These analyses were done in the laboratory of C. W. Cunningham, whose aid during this and other projects was invaluable. A National Science Foundation Dissertation Improvement Grant (NSF DEB-99-72707) to J. P. W. funded this study. Literature Cited Austerlitz, F., B. Jung-Muller, B. Godelle, and P.-H. Gouyon. 1997. Evolution of coalescence times, genetic diversity and structure during colonization. Theor. Popul. Biol. 51: 148-164. Barker, L., and V. A. /ullo. 1980. Asterias forbesi (Desor) (Asterozoa. Asteroidea) from the Pleistocene "Coquina" at Fort Fisher, New Hanover County, North Carolina. J. Elislui Mitchell Sci. Soc. 96: 39-44. Berggren, W. A., and C. D. Hollister. 1974. Paleogeography, paleobio- geography, and the history of circulation in the Atlantic Ocean. Soc. Econ. Paleontol. Mineral. 20: 126-186. Castelloe, J., and A. R. Templeton. 1994. Root probabilities for in- traspecific gene trees under neutral coalescent theory. Mo!. Phylogenet. E\-ol. 3: 102-113. Clark, A. M., and M. E. Downey. 1992. Starfishes of the Atlantic. Chapman & Hall. London. Clement, M., D. Posada, and K. A. Crandall. 2000. TCS: a computer program to estimate gene genealogies. Mol. Ecol 9: 1657-1660. Collins, T. M., K. Frazer, A. R. Palmer, G. J. Vermeij, and W. M. Brown. 1996. Evolutionary history of northern hemisphere Nucella (Gastropoda. Muricidael: molecular, morphological, ecological, and paleontological evidence. Evolution 50: 2287-2304. Crandall, K. A., and A. R. Templeton. 1996. Applications of intraspe- citic phylogenetics. Pp. 81-99 in New Uses for New Phytogenies, P. H. Harvey. A. J. Brown. J. Maynard Smith, and S. Nee, eds. Oxford University Press, Oxford. Cronin, T. M. 1988. Evolution of marine climates of the U. S. Atlantic coast during the past four million years. Philos. Trans. R. Soc. Land. B 318: 661-678. Cunningham, C. W., and T. M. Collins. 1998. Beyond area relation- ships: extinction and recolonization in molecular marine biogeography. Pp. 297-321 in Molecular Ecology and Evolution: Approaches and Applications. B. Schierwater, B. Streit, G. Wagner, and R. DeSalle, eds. Birkhauser Verlag. Basel. Cunningham, C. W., H. Zhu, and D. M. Hillis. 1998. Best-fit maxi- mum-likelihood models for phylogenetic inference: Empirical tests with known phylogenies. Evolution 52: 978-987. Durham, J. W., and F. S. MacNeil. 1967. Cenozoic migrations of marine invertebrates through the Bering Strait region. Pp. 326-349 in The Bering Land Bridge. D. M. Hopkins, ed. Stanford University Press, Stanford. Edwards, S. V., and P. Beerli. 2000. Perspective: Gene divergence, population divergence, and the variance in coalescence time in phylo- geographic studies. Evolution 54: 1839-1854. Endler, J. A. 1977. Geographic Variation. Speciation. and Clines. Princeton University Press, Princeton, NJ. Felsenstein, J. 1988. Phylogenies from molecular sequences: inference and reliability. Annu. Rev. Genet. 22: 521-565. Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. 1994. DNA primers for amplification of mitochondria! cytochrome c oxidase subunit I from diverse metazoan invertebrates. Mol. Mar. Biol. Bio- icclmol. 3: 294-299. Franz, D. R., E. K. Worley, and A. S. Merrill. 1981. Distribution patterns of common seastars of the middle Atlantic continental shelf of the northwest Atlantic (Gulf of Maine to Cape Hatteras). Biol. Bull. 160: 394-418. Goldman. N. 1993. Statistical tests of models of DNA substitution. J. Mol. Evol. 36: 182-198. Gosling. E. M. 1992. Systematics and geographical distribution of Myti- lus. Pp. 1-20 in The Mussel Mytilus: Ecology. Physiology. Genetics, and Culture. E. M. Gosling, ed. Elsevier, Amsterdam. Gosner, K. L. 1978. A Field Guide to the Atlantic Seashore From the Ba\ of Fundv to Cape Hatteras. Houghton Mifflin. Boston. Griffiths, R. C., and S. Tavare. 1994. Sampling theory for neutral alleles in a varying environment. Philos. Trans. R. Soc. Land. B 344: 403-410 Hasegawa, M., H. Kishino, and T. Yano. 1985. Dating of the human- ape splitting by a molecular clock of mitochondria! DNA. J. Mol, Evol. 21: 160-174. Hayward, P. J., and J. S. Ryland. 1995. Handbook of the Marine Fauna of North-West Europe. Oxford University Press, Oxford. Hellberg, M. E., and V. D. Vacquier. 1999. Rapid evolution of fertil- ization selectivity and lysin cDNA sequences in teguline gastropods. Mol. Biol. Evol. 16: 839-848. Hewitt, G. M. 1996. Some genetic consequences of ice ages, and their role in divergence and speciation. Biol. J. Linn. Soc. 58: 247-276. Hillis. D. M., B. K. Mable, A. Larson, S. K. Davis, and E. A. Zimmer. 1996. Nucleic acids IV: Sequencing and cloning. Pp. 321-384 in Molecular Systematics. D. M. Hillis. C. Moritz. and B. K. Mable, eds. Sinauer, Sunderland. MA. Holder, K., R. Montgomerie, and V. L. Friesen. 1999. A test of the glacial refugium hypothesis using patterns of mitochondrial and nu- BIOGEOGRAPHY OF ASTERIAS 103 clear DNA sequence variation in rock ptarmigan (Lagoons inutus}. Evolution S3: 1936-1951). Kelley, J. T., A. R. Kelley, and S. Appollonio. 1995. Landtorms of the Gulf of Maine. Pp. 19-38 in From Cap? Cod to the Buy of Fiuu/v: an Environmental Atlas of the Gulf of Maine. P. W. Conkling. ed. MIT Press, Cambridge, MA. Knimlton. N.. and L. A. \Veigt. 1998. New dates and new rates for divergence across the Isthmus of Panama. Proc. R. Soc. Loud. B 265: 2257-2263. Kuhner, M. K., J. Vamato, and J. Felsenstein. 1998. Maximum like- lihood estimation of population growth rates based on the coalescent. Genetics 149: 429-43-4. McDonald, ,). H., and M. Kreitman. 1991. Adaptive protein evolution at the Adli locus in Drosophila. Nature 351: 652-654. Menge, B. A. 1979. Coexistence between the seastars Asterias vulgaris and A. forbesi in a heterogeneous environment: a non-equilibrium explanation. Oecologia 41: 245-272. Nei, M. 1987. Molecular Evolutionary Genetics. Columbia University- Press. New York. Nei, M., and \V.-H. I,i. 1979. Mathematical model for studying genetic \ anation in terms of restriction endonucleases. Proc. Natl. Acait. Sci. USA 76: 5269-5273. Palumbi. S. R.. and B. D. Kessing. 1991. Population biology of the trans-Arctic exchange: mtDNA sequence similarity between Pacific and Atlantic sea urchins. Evolution 45: 1790-1805. Fernet, B. 1999. Gamete interactions and genetic differentiation among three sympatric polychaetes. Evolution 53: 435-446. Posada. D.. and K. A. Crandall. 1998. Modeltest: testing the model of DNA substitution. Bioinfonnatics 14: 817-818. Rand, D. 1994. Thermal habitat, metabolic rate, and the evolution of mtDNA. Trends Ecol. Evol. 9: 125-131. Reid. D. G., E. Rumbak, and R. H. Thomas. 1996. DNA. morphology and fossils: phylogeny and evolutionary rates of the gastropod genus Littorina. Philcs. Trans. R. Soc. Land. B 351: 877-895. Roy, K., J. W. Valentine, D. Jablonski, and S. M. Kidwell. 1996. Scales of climatic variability and time averaging in Pleistocene biotas: implications for ecology and evolution. Trends Ecol. Evol. 11: 458-463. Rozas, J.. and R. Rozas. 1999. DnaSP version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinfonnatics 15: 174-175. Schlotterer, C., M. Hauser, A. von Haeseler, and D. Tautz. 1994. Comparative evolutionary analysis of rDNA ITS regions in Drosoph- ila. Mol. Biol. Evol. 11: 513-522. Schopf, T. J. M., and L. S. Murphy. 1973. Protein polymorphism of the hybridizing seastars Asterias forbesi and Asterias vu/garis and impli- cations for their evolution. Biol. Bull. 145: 589-597. Schubart, C. I)., R. Diesel, and S. B. Hedges. 1998. Rapid evolution to terrestrial life in Jamaican crabs. Nature 393: 363-365. Seutin. G., B. N. White, and P. T. Boag. 1991. Preservation of avian blood and tissue samples for DNA analyses. Can. J. Zool. 69: 82-90. Swofford, D. L. 1998. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4. Sinauer Associates, Sunderland. MA. Tajima, F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 123: 585-595. Templeton, A. R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Mol. Ecol. 7: 381-397. Templeton. A. R., K. A. Crandall. and C. F. Sing. 1992. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cla- dogram estimation. Genetics 132: 619-633. van Oppen, M. J. H., S. G. A. Draisma, J. L. Olsen, and W. T. Stam. 1995. Multiple trans-Arctic passages in the red alga Phycoctrys rubens: evidence from nuclear rDNA ITS sequences. Mar. Biol. 123: 179-188. Vermeij, G. 1991. Anatomy of an invasion: the trans-Arctic interchange. Paleobiology 17: 281-307. Wares, J. P. 2000. Abiotic influences on the population dynamics of marine invertebrates. Ph.D. dissertation. Duke University. Durham. NC. Wares, J. P. 2001. Patterns of speciation inferred from mitochondria! DNA in North American Chthamaliis (Cirripedia: Balanomorpha: Chthamaloidea). Mol. Phylogenet. Evol. 18: 104-116. Wares, J. P., and C. W. Cunningham. Phylogeography and historical ecology of the North Atlantic intertidal. Evolution (in press). White, T. J., T. Bruns, S. Lee, and J. Taylor. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. Pp. 315-322 in PCR Protocols: A Guide to Methods and Applications. M. A. Innis. D. H. Gelfand. J. J. Snmsky. and T. J. White, eds. Academic Press. San Diego. Worley, E. K., and I). R. Franz. 1983. A comparative study of selected skeletal structures in the seastars Asterias forbesi. A. vulgaris, and A. rubens with a discussion of possible relationships. Proc. Biol. Soc. Wash. 96: 524-547. Yang, Z. 1994. Maximum likelihood phylogenetic estimation from DNA sequences with variable rates over sites: Approximate methods. J. Mol. Evol. 39: 306-314. Reference: Biol. Bull. 201: 104-119. (August 2001) Systematics and Biogeography of the Jellyfish Aurelia labiata (Cnidaria: Scyphozoa) LISA-ANN GERSHWIN Cabrillo Marine Aquarium, San Pedro, California 90731 and CSU Northrid?>e, California 91330 Abstract. The hypothesis that the common eastern North Pacific Aurelia is A. aurita is falsified with morphological analysis. The name Aurelia lahiata is resurrected, and the species is redescribed, to refer to medusae differing from A. aurita by a suite of characters related to a broad and elon- gated manubrium. Specifically, the oral arms are short, separated by and arising from the base of the fleshy manu- brium. and the planulae are brooded upon the manubrium itself, rather than on the oral arms. Aurelia aurita possesses no corresponding enlarged structure. Furthermore, the num- ber of radial canals is typically much greater in A. lahiata, and thus the canals often appear more anastomosed than in A. aurita. Finally, most A. labiaia medusae possess a 16- scalloped bell margin, whereas the margin is 8-scalloped in most A. aurita. Separation of the two forms has previously been noted on the basis of allozyme and isozyme analyses and on the histology of the neuromuscular system. Partial 18S rDNA sequencing corroborates these findings. Three distinct moiphotypes of A. lahiata, corresponding to sepa- rate marine bioprovinces, have been identified among 17 populations from San Diego. California, to Prince William Sound. Alaska. The long-undisputed species A. limhata may be simply a color morph of A. labiata, or a species within a yet-unelaborated A. lahiata species complex. The first known introduction of Aurelia cf. aurita into southern Cal- ifornia waters is documented. Although traditional jellyfish taxonomy tends to recognize many species as cosmopolitan or nearly so, these results indicate that coastal species, such as A. labiata, may experience rapid divergence among iso- lated populations, and that the taxonomy of such species should therefore be scrutinized with special care. Received 16 December 1998; accepted 5 April 2001. Current address: Dept. of Imegrative Biology, University of California. Berkeley. CA 94720. E-mail: gershwin@socrates.berkeley.edu Introduction Perhaps had Darwin not been afflicted with seasickness, he might have noticed the bewildering array of geographi- cally varying jellyfish morphologies. Some of his contem- poraries documented species separated by only short dis- tances but differing greatly in appearance (Eschscholtz. 1829; Brandt. 1835, 1838; Agassiz, 1862; Haeckel, 1879, 1880). Morphological distinctions have since been reported for populations of Cassiopea from separate islands of the Caribbean (Hummelinck, 1968), Mastigias in different lakes of Palau (Hamner and Hauri, 1981), and Aurelia scyphistomae from various parts of the Thames estuary (Lambert, 1935). In his studies of the genus Cyanea, Brewer ( 1 99 1 ) reported distinct morphotypes that could be corre- lated with isolated locations in Long Island Sound, USA; these observations resurrected a long-standing argument about species distribution and recognition criteria of North Atlantic Cyanea. Nineteenth-century taxonomists recog- nized different species, corresponding to a latitudinal gra- dation, on both sides of the Atlantic. Cyanea arctica Peron and Lesueur, 1810, was known as the boreal species from Europe to North America. In the western Atlantic. C. fulva L. Agassiz. 1 862, was found along the mid-Atlantic states, while the form south of the Carolinas was recognized as C. versicolor L. Agassiz. 1862. In the eastern Atlantic, C. capillata (Linnaeus. 1746) was established as the northern European species, while C. lamarckii Haeckel, 1880. was identified in warmer southern European waters. This pattern of biodiversity was largely overlooked by twentieth-century taxonomists. who often lumped the forms and recognized only C. capillata (Mayer. 1910; Bigelow. 1914; Stiasny and van der Maaden. 1943; Kramp, 1961; Calder, 1971; Larson, 1976). The scarcity of biogeographic studies of jellyfishes may be, in part, attributable to the unclear systematics of these 104 SYSTEMATICS AND BIOGEOGRAPHY OF AURELIA LAB/ATA 105 Figure 1. Original illustrations of Aurelia labiata, showing greatly enlarged munubrium: (A) lateral view of medusa; (B) oblique view of subumbrella. (Reprinted from Chamisso and Eysenhardt. 1821). animals. Color differences, patterns of pigmentation, and anatomical variation led to the description of many nominal species during the expeditions of the nineteenth century (see Mayer, 1910; Kramp, 1961). The range of variation in jellyfishes is not well understood, and species definitions are often vague, focusing only on the few most obvious char- acters. For example, if one sees a flat, whitish medusa with four horseshoe-shaped gonads. most tend to think it must be Aurelia aiirita. The details of anatomy have not been scru- tinized closely. Therefore, significant morphological differ- ences have not been detected, and inappropriate identifica- tions and erroneous conclusions regarding biogeography have been made. The systematic tangle and biogeographic mistakes are common throughout the medusan taxa, though I focus herein on Aurelia. Mayer (1910) recognized 13 unique forms of Aurellia (the spelling was later formally changed back to Aurelia by Rees, 1957), and sorted these forms into three morpholog- ical groups: 1. A. aiirita (Linnaeus, 1746) sensu Lamarck, 1816, and its seven varieties, described as A. cniciata Haeckel. 1880, A. colpota Brandt, 1835 [sensu Gotte, 1886] (as =A. coerulea von Lendenfeld, 1884), A. flavidula Peron and Lesueur. 1810 [incorrectly listed as 1809) (as =A. habanensis Mayer, 1900). A. hyalina Brandt. 1835. A. dubia Vanhfiffen, 1888. A. vitiana Agassiz and Mayer. 1899. and A. imirginalis L. Agassiz. 1862 2. A. labiata Chamisso and Eysenhardt, 1821 [incor- rectly listed as 1820[. with three varieties, described as A. clausa Lesson, 1829, A. limbata (Brandt. 1835) [incorrectly listed as 1838], and A. inaldivensis Big- elow. 1904 3. A. solida Browne, 1905 Mayer distinguished A. labiata and its varieties from the other two groups based primarily on the degree of scallop- ing of the bell margin, being 16-notched in the former and 8-notched in the latter. He subsequently found a specimen of A. iinritti at Tortugas. Florida, closely resembling A. labiata, leading him to conclude that A. labiata was prob- ably derived as a mutation from A. aiirita (Mayer, 1917). Kramp also wavered on the validity of A. labiata, first recognizing the species in his 1961 synopsis, then later regarding it as doubtful (1965, 1968). Most recently, au- thors such as Russell (1970). Larson (1990), and Arai ( 1997) have recognized two valid species: A. limbata, which is primarily arctic and has a conspicuous brown bell margin, and A. auritci. whose name has been treated as the senior synonym of all others. Russell (1970) followed Kramp (1965, 1968) in regarding all other species as varieties, whereas Larson (1990) and Arai (1997) simply did not mention any other species. The source of this confusion is unclear, as the original description of A. labiatu was quite specific. Translated from Latin, "It differs from A. aiirita by its very long oral lips. Marginal tentacles were not observed, but are without a doubt present. Arms appressed to the bell. Diameter of the bell nearly a foot" (Chamisso and Eysenhardt. 1821). The focus of the description and its accompanying illustrations is the strikingly unique elongated manubrium (Figs. 1,2). 106 L. GERSHWIN Figure 2. Aurelia labiata. adult medusa, from Monterey Bay. Califor- nia. although this character is rarely mentioned in later revisions. Furthermore, the characteristically short oral arms arising from the base of the manubrium were mentioned as being held close to the bell, a trait that is readily apparent in live specimens. Ironically, the commonly accepted character of 16 marginal scallops is not mentioned, although it is subtly illustrated. It is unclear why certain key characters of the original description have been ignored by later workers. Disorder in the nomenclature of Aurelia worldwide has caused confusion about the identity of the species in the eastern North Pacific. Depending on the author, one to three species have been recognized. Most authors have applied the name A. aurita to all forms. Some distinguish ,4. lim- bata, although this appears to have been occasionally con- fused with A. labiata (Zubkoff and Lin, 1975; Greenberg et al., 1996). When A. labiata has been recognized, it has been separated from A. aurita only by the doubling of marginal scallops (Hand. 1975; Kozloff, 1974). Although A. labiata was originally described from California, most reports of the species (apparently incorrectly) are from regions outside the eastern North Pacific. Throughout all the confusion, several studies have re- ported differences between the eastern North Pacific Aurelia and those of other regions, yet failed to elaborate the sys- tematics. Chia et al. ( 1984) found that the muscle system in Puget Sound polyps is distinct from that of polyps from Plymouth. England. Zubkoff and Lin (1975) observed pe- culiar banding in the isozyme patterns of Aurelia scyphis- tomae from Puget Sound, Washington, that caused them to wonder whether this population may belong to a species other than A. aurita. Similarly. Greenberg et al. (1996) could distinguish two groups on their allozyme patterns: one group consisted of two populations of A. "aurita" from Japan (one from Tokyo Bay, and one aquarium-raised) plus a population that was apparently introduced to San Fran- cisco Bay; and the second group consisted of wild medusae from Monterey Bay, California, and Vancouver, British Columbia. They further distinguished the two groups on the basis of morphology, using manubrium length and the highly anastomosed condition of the radial canals. To test the hypothesis that the common eastern North Pacific Aurelia is A. aurita, I compared the morphology of 17 populations of Aurelia from San Diego, California, to Prince William Sound, Alaska, to the morphology of A. aurita from Europe, and A. flavidula from the eastern United States, as described and figured by Agassiz (1862), Mayer (1910), Kramp (1961). Russell (1970), and many of the references therein. The conclusions that I have drawn on morphological characters are consistent with those emerg- ing from the enzyme analyses of Zubkoff and Lin (1975) and Greenberg et al. (1996), the neuromuscular study of Chia et al. (1984), and the DNA sequencing results of J. Lowrie of the Cnidarian Research Institute (pers. comm., June 2000) that is, that the common eastern North Pacific Aurelia is not A. aurita. However, it does match the de- scription of the species previously described as Aurelia labiata Chamisso and Eysenhardt, 1821. Thus, I propose a revalidation of A. labiata, and herein offer a redescription and designate a neotype. In scrutinizing the morphology of A. labiata. I further found that each population possesses unique characters that cluster into three morphotypes cor- responding to well-demarcated biogeographic provinces. The purposes of this paper are to describe the morphological and geographical variation in A. labiata and to stabilize the nomenclature for the species. This is necessary as a basis for further systematic investigation, for ongoing biodiversity studies, and for proper management of species introduc- tions. Materials and Methods Aurelia aurita and other fonns Literature-based comparisons were made using the Euro- pean form, Aurelia nuriia, and are denoted traditionally (e.g., Aurelia aurita). The full breadth of literature used for comparison is too massive to list here, but can be found in the synonymies of Mayer (1910), Kramp (1961), and Rus- sell (1970). Literature-based comparisons were made with A. flav- idula from the eastern United States, primarily following Agassiz (1862) and the references in the synonymy of Mayer (1910). Literature-based comparisons were made to the boreal A. limbata using Brandt (1835. 1838). Vanhoffen (1906), Kishinouye (1910), Bigelow (1913, 1920), Uchida (1934), Bigelow (1938), Kramp (1942), Stiasny and van der Maa- den (1943), Naumov ( 1961 ), Uchida and Nagao (1963), and Faulkner (1974). SYSTEMATICS AND BIOGEOGRAPHY OF AURELIA LABIATA 107 Comparisons were made using live, captive medusae descended from a Japanese population (cultured at Cabrillo Marine Aquarium); although the phylogenetic relationship between the European and Japanese forms is still in ques- tion, they are structurally similar that is. they both lack the enlarged manubrium characteristic of A. labiata. Comparisons were also made on some live, wild medusae from Spinnaker Bay, Long Beach, California, which pos- sessed the A. aurita body form, and on the descriptions of Greenberg et al. (1996) for the introduced San Francisco Bay form. Live representatives of Greenberg's population at Foster City could not be found. References made to forms that possess the A. aurita body type but are of uncertain taxonomic affiliation